FactSet's Taking Risk blog

Our most anticipated sessions at GARP's 13th Annual Risk Management Conference

FactSet Risk — 2012-02-01T19:03:17Z

Once again FactSet will be attending a two-day conference hosted by GARP (the Global Association of Risk Professionals) from February 28-29. We hope to see you there!

This year we wanted to share with you the Top 3 Sessions we're looking forward to most this year at the GARP Annual Risk Management Convention.

In the keynotes this year, we were intrigued to see two presentations that seem to piggy-back off one another.

They are The Wall Street Monolith Myth and Medium Term Prospects for Enhanced Financial Stability. The reason I'm intrigued is the first presentation discusses what it was about Bear Stearns corporate culture and process that made it fail while Goldman succeeded. The following presentation with the less interesting title has the distinction of being led by none other than Gerald Corrigan of Goldman Sachs.

Kidding aside, however, after what must be years of research, what does author William Cohan have to say about the avoidable mistakes that Bear Stearns made?Also, from an investor perspective and a managerial perspective what did it take for big banks like Goldman to outlast the conditions of 2008-2009?
The next session that caught my eye was Stress Testing: Computability and Emergent Phenomena. The most interesting point that Professor Timur Gok, who will lead the session, will discuss is whether you can truly manage risk with a "rearview approach." In short, can looking back truly be an effective way to move forward? I'd say yes, in the sense that we can only learn to mitigate risk from looking at the likely investment conditions it creates. And while looking at past crises to anticipate the shock of current ones is by no means a perfect solution, it is one that gets us as close as we can get (short of being fortune tellers) to anticipating the impact of a market shift. Stress testing also has the honored distinction of being deemed valuable to be required by many new accounting standards such as Dodd-Frank.
Finally, Geo/Political Risk Amidst World Turmoil rounds out our top three sessions. This session is likely to be a packed session, considering all we've seen from Europe. I'm particularly interested in hearing about how diversifying on a country basis may now pose unwanted risks. The session, led by Daniel Alpert of Westwood Capital, will also focus on the impact that politics has on the market. It's an extremely relevant topic for the eurozone where several heads of state must agree to come to any resolution. And also, where austerity measures have caused a wide range of reactions in Greece and elsewhere.

Eurozone risk: Managing risk in uncertain times

FactSet Risk — 2012-01-31T19:04:07Z

Recently on FactSet's podcast, we spoke with Laurence Wormald, Head of Research for Sungard APT. In our conversation we discuss one of the topics most on risk managers minds today: the ongoing crisis in Europe.

In our podcast, we asked how investors can gauge expected results from a possible Euro breakup, default of one of the GIIPS countries, and other scenarios for the troubled eurozone. Hear the episode here, or read the recap.

Want more on Eurozone risk? Register for our webcast, Managing Risk in Uncertain Times, with FactSet's Steve Greiner as he discusses how currency risk can increase in tumult-afflicted GIIPS-country companies and discusses options for measuring portfolio risk in multi-currency Eurozone fallout scenarios.

Recap of our interview with Wormald

Registration link for free webcast: Managing Risk in Uncertain Times

Management strategies: Reflections on fundamental vs. quant strategies during and after the credit crisis

Joseph Mezrich — 2012-01-24T22:09:55Z

This week's blog was contributed by guest blogger, Joseph Mezrich, Managing Director and Head of Quantitative Strategies at Nomura Securities International. Mezrich joins FactSet as a keynote speaker at this year's U.S. Investment Process Symposium.

A Conversation with Joseph Mezrich

What was the biggest surprise for fundamental managers during the credit crisis?

In 2007 and 2008 fundamental managers were unaware that value measures (that is to say buying value stocks) were often proxies for credit risk. Many investors were buying cheap stocks because they seemed to be the best buy for the money, but in the credit crisis they were actually priced like junk bonds.

Throughout the past decade, off and on, there are reasons to buy stocks or use particular strategies, but there are always underlying risks that investors aren’t aware of.

How were quants impacted by the crisis?

The famous quant crisis of August 2007 was, in many ways, less important than the fact that momentum was crushed during the crisis, particularly in late 2009. As momentum fell and factors behaved in new and unexpected ways, quants saw their strategies fail. It was only in late 2010 that quant strategies began to fully recover. Fortunately for the quants, their strategies continued to perform increasingly well throughout 2011, particularly if you look at the aggregate results.

What challenges have pointed out the strengths and weaknesses of relative strategies?

There’s been a lot of media coverage about stock correlation last year. Stock correlation was very high throughout 2011, and fundamental managers who rely on stock picking saw the worst year in more than a decade. With high correlations it didn’t matter how skilled of a stock picker you were, upswings and downswings in the market hit you harder as stocks rose and fell together. To contrast, strategies that go beyond stock picking, like quantitative strategies, have fared much better. A portfolio of factor-based strategy was better insulated against high correlation, so we saw a reversal of fortunes for quants.

What is your advice for investors given the market trends we saw in late 2011?

There are several high profile examples of people trying to figure out what’s going to work in the new market conditions. In the past year, correlation has been more a problem than volatility, though both contributed to poor performance. What caught people off guard was the inability to diversify and the inability to deal with market direction or volatility. You have to build into your strategy some methods for achieving these ends. One piece of advice to gravitate towards factor strategies. As important is having some form of risk management. Finally, you have to guard against the diversification problem, that is to say, a lack of diversity in your overall investment universe.

Currency risk forecasting in uncertain times

Steve Greiner — 2011-12-15T19:01:38Z

With the current situation in the Eurozone, combined with the increasing percentage of revenue coming from overseas among the largest stocks in the Global S&P 100, FTSE 100 and S&P 500, there is a simultaneous increase in currency risk for many companies. As a result, we see increasing currency risk for asset managers also, and one example concerns the increased risk for GIIPS country companies.

Suppose you are a publically traded importer of German goods located in Greece and suddenly find your stock trading in Drachmas, while your debt is still denominated in Euros (as it likely would be). In this scenario, the Drachma would be devaluing relative to the Euro, and you’re going to have to pay more for the German import items, while your debt service costs are going up. Thus, business risks are increasing and this uncertainty should be reflected in total risks increasing because there is a new currency exposure for the company (Euros) which it didn’t have before, along with interest rate risk increases due to the inauguration of new Drachma-denominated yield curves. This increased volatility should result in increasing market risk and increasing tracking error to one’s portfolio.

To assess the risk of this hypothetical Greek company, let’s pretend that the Drachma didn’t go away and just converted to some “ether-Drachma” upon Greece joining the Euro. Let’s say that Greece just pegged the ether-Drachma to the Euro at 340.75 Drachma/Euro exchange, which was the “fixed” rate on June 19, 2000 used for conversion to join the Euro. Given that, one could simply apply standard risk model construction methods to regress returns against this fixed FX rate for some rolling time frame (e.g., one year or 250 trading days) from that time until today and look at the beta of the regression which represents the “exposure” of some stock to the Euro, priced in ether-Drachmas. Now, suppose four days ago, the Drachma was re-issued and Greece left the Euro. Then the rolling one-year window of regression done today would have 246 days of fixed exchange rate and four days of floating exchange rate. In this scenario these four days of floating exchange would hardly register any increased Euro risk for the stock since the 246 days of fixed FX would dominate the influence on the exposure as over the one-year regression. So currency risk would be woefully under-estimated in this case and your stock risk estimates would be quite off the mark. You also wouldn’t want to use a shorter time-period for the regressions for many reasons, low statistical significance being just one.

One explicatory demonstration for properly quantifying the currency risk in this scenario, involves using interest rate parity relationships to back-fill the hypothetical foreign exchange between the Greek ether-Drachma and Euro. To see how this works, first let’s assume the German benchmark sovereign debt yield represents the Euro yield, where we’ll use the terms interchangeably. Next, we download the one-year yields of German, Greek, and Italian benchmark sovereign debt and form the interest rate parity relationship:

(1+i(Greek)) / (1+i(Euro)) = FX(forward) / FX(spot)

Substituting i(x) for yields as opposed to interest rates, we’d expect the relationship to be not too far from its theoretical application. Assume in this equation that FX(spot) is the ether-Drachma/Euro exchange rate before the debt crisis and has the fixed value of 340.75. Now, let FX(forward) be that rate the market expects the future FX to be of ether-Drachmas to Euros. From early 2000s until early last year the FX ratio on the right side of this equation was 1.0 simply because nobody expected the (hypothetical) future FX to be different from the spot because the ether-Drachma was fixed to the Euro.

Then, as we move through the Credit Crisis and beyond to today, we might expect the FX(forward) to depart from the FX(spot) and the ratio becoming a multiplier of the original conversion rate. So for example, let i(x) take values of one-year sovereign yields. In examining the equation, yield ratios are directly proportional to FX future to spot ratios and a plot of Greek debt yield over German debt yields will act as a proxy for FX(forward) over FX(spot) of ether-Drachma/Euro.

Now we plot Greek/German yield ratios below for the daily time periods starting March of 2007 until December 2011. We call your attention to the fact that on June 19, 2000 the FX rate of Drachma/Euro was 340.75 and existed before the Credit Crisis at this value. After the crisis, the market is saying, “we believe there really is a floating FX rate of the ether-Drachma/Euro and it’s been moving toward its historical conversion rate but has overshot it recently” which represents a devaluing of this hypothetical ether-Drachma. In essence, using interest rate parity will allow us to go back in time and create a hypothetical time-series FX rate for ether-Drachmas versus the Euro. For the ether-Drachma, we see currency devaluing by a more than a factor of two for the Drachma, moving from 340.75 to over 850 ether-Drachma/Euro meaning the Euro is rising relative to the market’s implied ether-Drachma.

Obviously you could do this for any GIIPS currency for that matter and we show this for the Italian ether-Lira, as plotted below for the same time-period. Though the ether-Lira is devaluing, it’s only been about -6% relative to Euro.

You can use this technique to back-fill any GIIPS currency ether-FX that you can later use to compute currency exposures by regressing stock returns. In this way, the volatility of the FX is more appropriately captured and as of today’s date, regressing a one-year window of this ether-Drachma/Euro exchange rate offers a better estimate of the exposures Greek companies have to the Euro than one would obtain using fixed FX.

For the interest rate risks which are strong influencers for fixed income risk, one would build out a historical yield curve for the ether-Drachma too, later decomposed via principal component analysis at KRD points to measure interest rate risk.

Recently, we read an article in Financial Times that made this statement: “Almost all euro area government debt is also governed by national law (Italian law governs Italian debt, Greece law governs Greece debt, French law governs French debt and so forth). However, almost all cross-border debt issued by Eurozone corporations is governed by English law. So Italy could change the law to make its sovereign debt holders take a haircut, but Italian companies cannot”. An English law firm says private and public companies with Euro denominated debt most probably cannot take advantage of any force majeure issues either.

Thus, we find the need to undertake risk estimates for stress-testing the exposure to the Euro for all companies within GIIPS countries in the event that these countries be expunged from the Eurozone. So here we’ve detailed a way to capture the currency risks inherent in GIIPS countries’ securities, even for non-fungible currencies and we’ve calculated an ether-currency FX rate at which the market might be expected to trade.

What Olympus teaches us about corporate governance and risk

George Hogan — 2011-12-06T15:37:12Z

Today's blog post is contributed by guest blogger George Hogan, who manages Portfolio Sales in FactSet's Asian markets.

While the relationship between corporate governance and investment risk may be somewhat difficult to quantify, the chart below helps to confirm what we all already intuitively know: poor corporate governance can have dramatically negative consequences for investors.

Between October 13 2011 and November 11 2011 Olympus lost more than 80% of its share value after the company ousted its English CEO, Michael Woodford, barely six months into the job. Of course it wasn’t just the fact that the company had booted out its freshman foreign CEO that caused the precipitous drop. Rather, the charges that Mr. Woodford publically leveled against the company and which eventually led to the admission by management that it had been covering up huge investment losses over a period of two decades also played a major role.

While a traditional view of risk might write this incident off to the stock-specific bucket (a risk that could have easily been diversified away), that view ignores at least one very important question. What if this type of corporate governance risk actually has a strong systematic component?

For those investors who have spent a significant portion of their careers with some sort of focus on Japan, this may seem like just the latest in a series of headlining corporate governance scandals to come out of the country. In an October 22 article, for example, The Economist laid out the following short list of similar high profile lapses:

TEPCO	2011	Ineffectual managers kept in place after nuclear disaster
Toyota	2010	Quality problems; board of 29 insiders fails to act. Market cap falls $34 billion in two weeks.
Livedoor	2006	Internet entrepreneur cooks books, is currently in prison.
Nikko Cordial	2006	Accounting scandal; chairman and CEO resign
TEPCO	2002	President resigns over falsified nuclear safety tests

To this we could add any number of similar incidents without too much effort. In a November 14 article in fact, the Nikkei Shimbun published an article (“Ghosts of Yamaichi Haunt Olympus Investors”) that drew links between the current Olympus scandal and the 1997 incident that eventually lead to the failure of Yamaichi Securities, which used to be one of Japan’s most prominent investment banks.

But an even more systematic example here might be the beloved poison pill boom that Japan has enjoyed since 2005. In a May 2008 article the New York Times pinned that number as high as 500 companies, giving each of them “the option to issue stock warrants to dilute the stake of a fund or company launching a hostile takeover.” As hostile bids by foreign funds began to increase in Japan (none of them successful), corporates began to take cover. Quite naturally, this made minority shareholders quite unhappy.

In fact, so bad is Japan’s corporate governance image that it led GovernanceMetrics International to rank Japan 36 out of 39 countries/regions, below the likes of South Korea, Russia, and China (see chart below). To be honest here, I highlight the point not because I agree with the relative rankings, or even to say that I understand the methodology employed by GMI to come up with these stats. Rather, I highlight the ranking to illustrate the point that Japan’s poor image on corporate governance is so pervasive that most who quote the GMI rankings (including the afore mentioned Economist) don’t even bother to qualify them. They just accept that the situation really is that bad.

Alas, all this leads most “Old Japan Hands” to silently gasp in dismay at yet another corporate governance debacle, unless of course they happened to be overweight Olympus at the time the scandal broke (then the reaction is somewhat more vigorous). But the point here is that, insofar as market psychology is important, then it appears that the market views poor corporate governance as a systematic risk factor in Japan. Hence, you can’t simply diversify that risk away.

In a broader context, an international mandate might give you better opportunities to reduce that risk, but unless your mandate is entirely ex-Japan you can’t erase it altogether. And remember, even if you are ex-Japan, Japan isn’t the only country with a poor record for corporate governance. It may be one of the “bad apples” from this perspective, but it isn’t the only bad apple. And this doesn’t even address the question of measurement. Without any type of reliable, objective method for scaling exposure to poor corporate governance (as we can do for things like value, interest rates, or oil) it becomes extremely difficult to have any good sense at all what your real exposure to poor governance actually is.

Nevertheless, there seems to be a strong argument to be made here that poor corporate governance isn’t simply a stock-specific issue, but rather can have a strong systematic component as well. While I’ll be the first to admit there are plenty of practical issues involved here in objectively understanding the scale of this risk, as long as the markets are going to pay attention to it, you as the investor probably ought to as well.

Five easy steps to fixing the ratings agencies (Part 2)

Dan diBartolomeo — 2011-11-21T19:01:55Z

Contributed by guest blogger Dan diBartolomeo, President, Northfield Information Services.

Last week, in part one of this blog, Dan diBartolomeo of Northfield provided the context and background for what is ailing the credit ratings agencies. In part two of his posts, he makes several more suggestions for how to standardize ratings and improve their credibility.

II. Accreditation

As we first suggested in 2009, regulators who oversee rating agencies should have an accrediting board of independent experts who would periodically visit with the rating agencies and determine if their analytical policies and practices are reasonable and within the range of contemporary methods. Such a board would be composed of both academics and practitioners with clear technical expertise. If a given agency’s practices are found to be substandard, they would be suspended from issuing new ratings until the deficiency is resolved. This is no different than the process by which schools, colleges and hospitals are routinely subject to review and accreditation by supervisory agencies set up by regulators or trade associations.

III. Accountability

Providing a rating of the creditworthiness of a borrower is a probabilistic exercise. There is generally a high likelihood of repayment and a small likelihood of default. The hard part is deciding how “high is high” and how “small is small.” In some ways, it is very closely akin to the problem of an insurance company that must decide what future claims will arise from writing a particular set of insurance policies. While predicting claims from something like an earthquake are obviously difficult, even predicting claims from life insurance policies can be problematic in the presence of outbreaks of serious contagious disease, as in the case of AIDS in the 1980s and early 1990s.

Surprisingly, there are relatively few cases of insurance companies getting into financial difficulty because of miscalculating the expected claims from a particular set of insurance policies. We believe there are good reasons for this. Firstly, under the insurance regulations of most jurisdictions, each insurer must appoint a chief actuary who supervises the process of estimating the future claims. The actuary is then required to write periodic reports attesting to their personal involvement in the process and certifying that the process has been carried out in a sound and professional fashion. Secondly, the professional education requirements to be a full actuary in most countries are very substantial, requiring many years of field experience and passing up to nine competency exams. Actuaries whose work falls under any negative suspicion are subject to discipline from both regulators and professional societies, much the same as doctors and lawyers.

Each rating from a rating agency should carry a similar kind of personal certification from the credit analyst who led the rating process. Such a structure would add a layer of professional ethics and discipline to a business model that has been demonstrated to be difficult to regulate effectively.

Professional associations in the finance field such as CFA Institute (CFA), the Global Association of Risk Professionals (GARP), the International Association of Financial Engineers (IAFE), and the Professional Risk Managers International Association (PRMIA) would be very keen to participate in the formulation of rigorous professional standards for credit analysts.

IV. Special Rules for Special Vehicles

The concept of borrowing money from a lending institution or investor is as old as organized society. We need only recall the Biblical discussions of money lenders setting up shop outside temples thousands of years go. However, in the last couple decades there has been a critical change in the nature of fixed income investing. Throughout history, money has been borrowed by entities that actually carried out some purpose other than simply borrowing money. Sovereign bonds have funded the operations of governments, municipal bonds have funded the operations of cities, and corporate bonds (and loans) have financed private businesses that carry out some business activity. In each case, the issuer has some purpose other than to simply engage in some form of financial intermediation.

In contrast, many hundreds of very complex structured product instruments received AAA ratings in recent years while being the financial obligation of a “special purpose vehicle” (SPV). A large fraction of such AAA rated instruments were either severely downgraded or went into default. SPVs are really just corporate shells that act as a legal structure in which a portfolio of assets is financed with a combination of debt instruments (and sometimes an equity tranche). If the assets earn a return higher than the cost of the financing, the organizers of the SPV stand to profit. While such “slicing and dicing” of financial assets can be very useful to investors in order to create new instruments with desirable properties (e.g., customized maturities), there is a clear danger in the operation of the rating process.

Since the financial structure of a given deal [SPV] is essentially arbitrary, the organizers can “game” the rating process.

For example, consider an SPV that holds a portfolio of $100 million of corporate loans that is financed with two bond issues, of which one issue is subordinated to the other. The organizers can effectively ask the rating agencies “How much of the financing has to be in the subordinated bond issue if we want the senior bond issue to get a AAA rating?” If the rating agency says $20 million, you can rest assured that the amount of subordinated bond issue will be $20 million and not a penny more. The deal is structured from the start to be as profitable as possible subject to getting the desired AAA rating for senior debt. From the outset the transaction sits by design on the razor’s edge of a downgrade to the next rating level. This is a very different situation from a traditional borrower where the need to borrow and the creditworthiness of the borrower are largely external outcomes of real world operations.

Given the gaming aspects of such instruments and their high degree of complexity relative to traditional bonds, the ratings of structured products always carry more uncertainty than those of traditional bonds or loans. We suggest that rating agencies be prohibited from issuing their highest rating (i.e., AAA) to any structured product until the credit rating agency has at least ten or more years of experience with that particular type of structured product, so that a meaningful history of rating change experience would be available.

V. External Awareness

Each rating agency has their own process for deciding when the rating of a borrower should be reviewed for possible upgrade or downgrade. Numerous academic studies have shown that well known analytical models have been shown to predict when an upgrade or downgrade is more likely to occur in the near future. It is only human nature that credit analysts, like real estate appraisers are uncomfortable with declaring their own prior analyses as being outdated and invalid. As such, the process of upgrading and downgrading credit ratings tends to be “sticky” and lag behind the flow of actual relevant events. In the words of the stage musical Evita, “it is hard to know which way to go when it is you that you are following.”

There are numerous analytical models of the credit risk of various borrowers and fixed income instruments that are provided for a fee to lenders and bond investors. Service providers such as Moody’s Analytics (separate from the rating business), Starmine, Rapid Ratings, and Kamakura all provide an independent “lender’s perspective” on sovereign and corporate credit risk (Northfield has a proprietary analytical model of corporate credit for use in our portfolio risk models, but does not sell credit risk information as a separate service at this time). We suggest that the rating agencies be required to subscribe to one or more of these services so as to be aware of discrepancies in the perception of credit risk for particular issuers. This requirement does not suggest that the rating agency has any obligation to change their rating in consequence of the view of outside services but rather should be at least aware of major discords as a way to trigger an internal review.

Conclusion

The structure of the ratings business has evolved into the current form for good and substantial reasons.

Credit ratings are deeply embedded in the investing operations and regulation of essentially every bank, insurer and investment fund in the world. While wholesale reform of the ratings business may or may not be desirable, it is clear that relatively simple measures would bring greater strength and consistency to the existing ratings.

We believe these modest reforms are both accessible and sufficient to ensure the quality of the rating process to the extent needed by investors and lenders as well as borrowers.

Five easy steps to fixing the ratings agencies (Part 1)

Dan diBartolomeo — 2011-11-18T16:17:27Z

Contributed by guest blogger Dan diBartolomeo, President, Northfield Information Services.

One of the largest contributing factors to the Global Financial Crisis of 2008-2009 was the huge number of fixed income instruments with very high ratings (e.g., AAA) that were either severely downgraded or went into actual default. All three of the principal rating agencies, Moody’s, Standard and Poor’s, and Fitch were shown to be seriously deficient in their ratings of a variety of debt instruments, particularly complex instruments (e.g., Credit Default Obligations or CDO) and the obligations of major financial institutions demonstrated by the spectacular collapses of firms such as AIG, Bear Stearns and Lehman Brothers.

The first thing we must remember is that the purpose of the rating agencies is not to protect investors and bank lenders from defaults on commercial debts.

In response to such clear failure, there have been a variety of suggestions from economic policy makers and regulators of several countries as to how to restructure the rating business, so as to reduce the potential for ineffective credit ratings to foster systemic risk within the international financial community. In considering such proposals, the first thing we must remember is that the purpose of the rating agencies is not to protect investors and bank lenders from defaults on commercial debts. The purpose of the rating agencies is to provide a mechanism for creditworthy borrowers to demonstrate their creditworthiness, and thereby be able to borrow large sums from a broad set of lenders through a public bond issue. In such cases, no single lender or bond buyer is lending a sufficiently large sum that it would make economic sense to undertake the cost of elaborate credit analysis of the borrower. Instead, such investors depend on diversifying their portfolio across a broad set of debt instruments to mitigate their risk along with the agency ratings.

Of course, it is clear that since the ratings are paid for by the borrower the rating agencies have strong business incentives to be optimistic in their credit assessments, and in some bizarre cases have simply “rubber stamped” ratings with no actual analysis being carried out. Equally perplexing to some financial market participants was the recent downgrading of U.S. Treasury debt from AAA to AA+ by S&P.

The vast size of traded bond and syndicated loan markets around the world are a testament to the fact that the current structure of the ratings business has been largely successful in facilitating capital flows from lenders and investors to borrowers.

During the financial crisis the rating process failed miserably to protect bond investors from downgrades and defaults. However, the rating agencies are paid by the borrower and owe no legal duty or particular allegiance to lenders or investors.

On the one hand, it is hard to fault the rating agencies for failing to carry out a responsibility to lenders that they never accepted as theirs. On the other, the rating agencies have routinely charged large fees to banks and investors for access to the rating data and hence bear significant responsibility for the legitimacy and quality of the information being sold.

The extent to which agency ratings take on the role of a safeguard for investors is a legal construct where all sorts of regulations across banking and investment are dependent on agency ratings, giving private profit-making companies the effective power to be regulators of financial markets. The concept of agency ratings is deeply embedded in international banking regulations (e.g. the Basel accords) wherein rules on balance sheet leverage and capital adequacy are based directly on the ratings of a bank’s asset portfolio. Similarly, the rules of most jurisdictions governing insurance companies, annuities, mutual funds, and private trust management involve the concept of an “investment grade” fixed income instrument, a crucial label bestowed at the discretion of the rating agencies.

Regulatory Fixes?

In response to the problems during the financial crisis where many “safe” fixed income portfolios lost most or all of their value, numerous suggestions have arisen to reform the rating process. Early suggestions in the U.S. focused on regulating the way in which ratings are paid for so as to remove the economic incentives to be optimistic as to the creditworthiness of borrowers. However, if borrowers are not going to pay for the extensive analytic work that should go into a rating, who will? As previously noted, the basic structure of a public bond issue is to allow for many, many investors, none of whom will want to shoulder the cost burden of detailed credit analysis for the relatively small sums they will invest in any one particular bond issue. One possible approach would be to have some kind of tax on bond and syndicated loan issuance and have the tax monies collected used to pay for the rating agencies under contract to a regulatory agency.

Financial regulators in Europe have called for making the global ratings business more competitive by encouraging the establishment of a new rating agency based in Europe, and legal requirements for borrowers to rotate their rating business among rating agencies on a periodic basis so as to reduce the potential for any intentional collusion, and to ensure some business for start-up rating agencies. More radical suggestions have included rewriting all regulation for financial services firms so as to remove or at least reduce dependency on ratings. Of course, if credit ratings are removed from the existing regulations as metrics of the default risk of a debt instrument, it is less clear what mechanisms could be put into place that would better safeguard investors while continuing to promote vigorous fixed income markets around the world.

Rather than try to fix the ratings “business,” we believe the appropriate immediate course of action is to simply put in place some basic rules that would ensure that credit ratings as currently available would be a sufficiently competent metric of creditworthiness.

Investors don’t need to fix the rating business or related regulations. They simply need the ratings to be done with sufficient quality so as to be meaningful measures of economic risks borne by lenders. In this regard we have a series of five suggestions:

I. Consistency and Transparency

Fixed income investors have three key metrics as to the credit risk they are bearing in a particular bond or loan. The first is the expected likelihood of default over some time horizon (Probability of Default or PD). The second is the extent of expected economic loss of value if a default occurs (Loss Given Default or LGD). The third metric is the expected probability that a current rating will be downgraded over some time horizon (Transition Probability or TP). Rating agencies should be required to periodically publish guidelines that describe what each rating level (e.g., “AA+”) represents in terms of expected ranges of PD, LGD, and TP.

By representing the letter rating process in terms of actual economic values, we can improve the rating process in two ways. First, we foster consistency across instrument types. For ratings to be meaningful for investors, the expectations of economic loss arising from a given rating should be universal, not instrument-specific. The credit risk of AA sovereign bonds, AA corporate bonds, AA municipal bonds, and AA structured products should all be comparable within some definable range. In addition, having an economic description of the meaning of ratings would also add consistency through time, so investors can be confident that the loss expectation of a BBB bond in 2011 is the same as it was in 2007. This two-dimensional consistency is absolutely crucial if agency ratings are to be used in bank and insurance regulation.

It should be noted that nothing in this suggestion requires that the various rating agencies agree with one another as to the range of values for PD, LGD or TP. Nor does it require that the rating agencies have a crystal ball such that the rating process would be judged a failure if the realized values of these metrics should fall outside the expected ranges. No reasonable investor would expect the rating agencies to be able to accurately predict wars, recessions, pandemics, or natural disasters that might affect the global economy or the fortunes of a particular issuer. All it requires is that each rating agency be consistent and transparent as to what they believe the economic meaning of a rating actually is.

Please tune in to our blog for the second installment of Dan diBartolomeo's comments on ratings agencies, where he presents four other ways to strengthen these companies.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Is Greece uniting or dividing the rest of Europe?

Daniel Mathon — 2011-11-08T19:06:37Z

This blog is contributed by Daniel Mathon, a Portfolio Analytics Sales Manager in FactSet's London office.

The on-going crisis in Greece continues to make headlines.

By calling for a referendum regarding whether to accept the European financial bail-out, the Greek Prime Minister, Mr. Papandreou, has achieved a rare feat: simultaneously annoying all other European nations. Being the birthplace of democracy (literally, "people power") provides a potential excuse. However, asking your own people whether or not they want to have their quality of life diminished through austerity measures seems verging on politico-economic and emotional blackmail. Following a tumultuous week in the Greek parliament and a quiet chat between Mr. Papandreou and his European peers, the referendum plans have since been scrapped.

However, how has the market relationship between European nations changed during the year in which questions about the future of the Euro have been as prominently featured as mentions of increasing correlations across markets? In this post we take a brief look at the relationship between the French and German markets from a risk point of view, comparing it to the relationship between the non-Euro UK and German market. While the difference between the French and German market has continued to shrink during the year, the relationship between the UK and German market has remained much unchanged. So are the issues in Greece (as well as Portugal, Spain, Italy and Ireland) dividing the Eurozone countries or bringing the better-off Eurozone countries closer together?

The above chart shows the predicted tracking error and factor risk*of the French CAC-40 versus the German DAX index. We used both measures to look at the difference between the two indices in terms of both overall and systematic risk. Both tracking error and factor risk dipped mid-year, but ended up only slightly above where they started the year.

However, active risk numbers, like the predicted tracking error and factor risk, are heavily influenced by overall market volatility, rising along with rising volatility, ceteris paribus. As can be seen in the above graph of the VIX, there was a significant rise in market volatility during the year. This trend might therefore be masking more subtle underlying movements in the active space.

If we now adjust our active risk numbers for this increase in absolute volatility (by dividing them by the absolute risk for each time point), we can see that both the new tracking error and factor risk numbers decreased throughout the year, dropping 50%. This downward movement tells us that on an adjusted basis the two indices track each other more closely, both in terms of overall active risk and factor risk.

Conducting the same exercise on the UK FTSE 100 index versus the DAX, the predicted tracking error rose dramatically in recent months. Looking at the adjusted numbers, this change is only just nullified by adjusting the data for changes in absolute volatility, indicating that the relationship between the UK and German market hasn’t significantly changed during this period.

We have looked at how the relationship between countries within the Eurozone has changed, using the German and French markets as an example, and compared it to the relationship with the main non-Euro country in Europe, the United Kingdom. Based on the above, the connection between the two powerhouses of the Eurozone appears to be strengthening in these difficult times for the Euro. This is in contrast to the relationship between the UK and Germany, where there has been little change over the last year. Is this what you are expecting? Are you taking this into account when constructing your portfolios?

* We used a European, daily-updated, short-horizon, statistical factor model.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Financial Blogs We're Reading

Risk News — 2011-11-02T18:13:24Z

While FactSet's Risk Blog seeks to be a great source of in-depth commentary and perspective on the market, our writers still have some of their own favorite sites to follow.

Favorite Blogs:

Calculated Risk: Retired tech exec Bill McBride helms this blog, which made a name for itself by providing expert commentary prior to and during the mortgage market collapse. The blog is a great source of commentary on "breaking" stories; it releases around three posts a day so the news is always current.
DealBook (New York Times): Focused on single companies and exchanges, we like DealBook because it provides the depth we want from company news. Stories on big bankruptcies, mergers, or record profits are accompanied by historical context and quotes from company insiders. They also aren't stingy about quoting FactSet data sources, either.
iMFdirect: This blog serves as the ongoing voice of International Monetary Fund on the world economy. While its focus is broad, it is a fine resource for global prespective and an understanding of possible sources of international country and currency risk. Posts are detailed, but good for scanning. Also, no one can argue with the tenure of their contributors.
The Call: The blog is written by Ian Bremmer, a brilliant speaker and a keynote at last year's FactSet's Symposium (follow this link for our 2012 Symposium keynotes). It focuses on international politics but paints a more complete picture—a recent article on Putin's argument against Occupy Moscow expands to discuss the future uncertainty of Russia's oil riches.

Honorable mention:
Risk Management Monitor: A bit of a potluck, this blog traces a mishmash of topics, such as the impact of natural disasters and business management best practices. Tucked into its pages, though, are some nicely packaged stories on popular topics in risk management, such as a recent post on why a recent decrease in business ligitation is a trend unlikely to continue.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Sovereign Yield Risk: This time it's (not) different

David Mieczkowski — 2011-10-19T18:02:47Z

Sovereign yield risk is rising all across the globe of late, and people are taking note. An emergent refrain is that there is something fundamentally different about this time. While I agree that sovereign yield risk is on the rise, I don’t agree that “this time it’s different.”

There are three primary risks to owning sovereign bonds: default risk, currency exchange risk, and inflation risk.

Default risk is generally only a risk factor for sovereign bonds issued in a “hard” currency (AKA, a currency they can’t print). For example, Eurozone members can’t print Euros (only the ECB can), so for the Eurozone countries, the spread of a Eurozone country’s sovereign bonds over the Euro-benchmark is driven by default risk. That default risk is driven by leverage and the volatility of economic output. As leverage becomes high, total economic output volatility becomes increasingly important in determining default risk. This is no different than for corporate bonds. Here, contingent claims analysis can provide useful insight (which is what FactSet’s Balanced risk model uses to capture EuroSovereign spread risk).

Click the image below to enlarge.

In the Merton Model, leverage and asset value volatility are the primary drivers of spread risk, with asset volatility becoming increasingly important as leverage becomes high. This model becomes increasingly accurate as GDP volatility and the volatility of the present value of all government revenues become more inexorably linked. Present day GIIPS countries and Argentina in 2002 come to mind as prime examples of countries with both high leverage and high economic output. Actual versus Merton Model implied spread for the GIIPS are graphed below.

FX risk for bond holders in foreign currencies seems pretty obvious for currencies that float. For currencies that are benchmarked to something, the risk is a little more subtle. Any country that links its currency to an exogenous asset (gold, another currency, etc.) effectively turns its currency into a hard currency. Here, currency risk behaves much like default risk. A “default” occurs when domestic reserves of the commodity or linked currency (gold, etc.) run critically low, and the government is forced to let its currency float to the benchmark to avoid a technical default. This causes a sudden devaluation of the currency relative to the benchmark.

Examples of this trend are Mexico in 1994 when the government was forced to float the peso to the dollar, the Asia financial crisis of 1997 (where floating of the Thai Baht led to a domino like effect for Indonesia, Malaysia, Singapore, etc.), and in the US in the 1970’s when Nixon officially abandoned the gold standard. Here, as above, high leverage raises the influence of domestic output volatility, relative to the benchmark, to cause effective (relative to the benchmark) yields to rise.

Finally, for domestic bonds in domestic currency, the risk measured in domestic currency terms is generally limited to inflation risk. For the domestic investors in emerging economies, foreign currency debt default almost immediately leads to very high domestic inflation, which ultimately drives nominal yields higher. The important thing often overlooked about inflation risk for yields is that nominal yields are driven by both inflation expectations as well as the volatility of those expectations. So the spread of nominal yield over real yield is ultimately driven by both the expected inflation rate, as well as the forward looking volatility of inflation.

For domestic investors in emerging economies, foreign currency debt default almost immediately leads to very high domestic inflation, which ultimately drives nominal yields higher.

Using the Equation of Exchange to decompose inflation into money supply, velocity of money, and real output (%Price Level = %Money Supply + %Velocity of Money - %Real Output), it’s pretty easy to see that all of those components have become much more volatile for many countries recently. It’s also easy to see that these variables ultimately become much more volatile when leverage rises.

Putting all that together, what is “different this time, is that developed economies have much higher covariance now than in the past, and high leverage in the developed economies is now pervasive. So whereas before it was isolated to a few emerging economies (Mexico, Argentina, Thailand, Indonesia) whose risk could be mitigated through diversification among countries, it is now everywhere. Since economies are highly exposed to the global economy more than in the past, this high leverage is causing a rise in a global economic uncertainty variable, which can be seen by a rise in volatility in a host of markets, for a diverse set of assets.

No matter how this leverage will ultimately be reduced, what’s clear is that this increased risk in Sovereign bonds isn’t going down anytime soon.

For more on yield curve analysis, read our article on viewing the world's economy through the lens of yield curves.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Will Operation Twist help the housing market?

David Mieczkowski — 2011-10-11T19:37:42Z

The Fed recently launched a third round of quantitative easing, colloquially dubbed Operation Twist. According to the Federal Open Market Committee’s press release: “The Committee intends to purchase, by the end of June 2012, $400 billion of Treasury securities with remaining maturities of 6 years to 30 years and to sell an equal amount of Treasury securities with remaining maturities of 3 years or less.”

The press release highlights two goals of the program: To push long-term interest rates down and to support the mortgage market by reinvesting principal payments on government-held agency MBS and debt.

It’s pretty obvious that the Fed will continue to throw what lifelines it can to rescue the beleaguered U.S. homeowner as long as CPI growth remains in check—and with around 40% of the CPI coming from housing-related costs, it’s a fair bet that it will. To steal a phrase, the question becomes, is the Fed pushing on a string?

To answer this, we need to understand the relationship between the term structure of the yield curve and the effective mortgage rate in the market, as well as the relationship between the effective mortgage rate and house prices.

What exactly is the relationship between the yield curve and mortgage rates?

Below we show the results of a regression performed over the last 250 trading days. It contains the daily difference in the Fannie Mae current rate for 30-day delivery into 30-year fixed charted against the daily differences in yield on the 10-year Treasury as well as the 1x10 swaption implied volatility.

The regression explains just over 60% of the variation in the daily changes in the mortgage rate. Both independent variables are highly significant, with the coefficients being such that a 10 basis point move in the 10- year Treasury would cause a 6 basis point move in the mortgage rate A 10 percent move in the lognormal implied volatility would cause a 9 basis point move in the mortgage rate. Sustained correlation over time is nice, but if the relationship isn’t causative, then Fed action to manipulate the level of the 10-year won’t be effective. Why should we believe the relationship is causative?

A mortgage can be thought of as being long a bond (e.g., a 30 year amortizer) and short a put option (prepay option). For a 30-year fixed mortgage, the long bond position has a modified duration close to that of the 10-year Treasury, so the 10-year makes for a natural benchmark. For the short option position, the highest prepayment risk occurs between one and ten years of seasoning— recent borrowers tend to be somewhat fatigued by the application process, and very seasoned borrowers have typically been exposed to some refinance opportunity in the past that they passed up, implying an indifference or inability to refinance.

So pushing the level of the long end of the curve down or lowering the volatility of interest rates, both things the Fed has done with its two most recent FOMC statements, can push down the mortgage rate. But will pushing down the mortgage rate help the housing market, or the economy at large?

It’s intuitively clear that changes in the mortgage rate should be negatively correlated with house price appreciation. The less that comes out of a fixed monthly payment to pay interest the more can go toward principal, and hence the larger a principal balance that can be borrowed, if all else is equal. The common criticism to the Fed’s focus on pushing down mortgage rates is that falling house prices means collateral is appreciably worse, and lending standards are significantly tighter, locking the bulk of borrowers out of refinance or purchase opportunities, making Fed action irrelevant.

The above table presents the 20 year correlations for house price appreciation, changes in the level of mortgage rates, and the Senior Loan Officer Survey on percent tightening of lending standards in mortgages, as well as the components of the equation of exchange – the money supply (M), the velocity of money (V), and real GDP (Y). Here money supply is measured by the M2 series.

What stands out from this is that, nationally, house prices are most strongly influenced by lending standards, and least strongly influenced by mortgage rates or the money supply. Thus, the data seems to confirm the prevailing wisdom that pushing the mortgage rate down will have at best mild impact on the rate of housing related credit creation and house prices. The Fed clearly believes that mild impact is better than no impact. Let’s just hope that string the Fed is pushing on isn’t actually a rubber band.

If it's not another recession, might it be the apocalypse? (Part 2)

Steve Greiner — 2011-09-30T19:25:43Z

Earlier this week, in part one of this post, I talked about the tough conditions in the market. I used some models to demonstrate the utility of risk analysis and risk modeling in an unpredictable environment.

A real dramatization of rising risks can be understood if we just examine the first 10 days of trading after 7/29/2011. We can plot the increase in risk as measured by 99% VaR along with the return of the S&P 500. The plot below captures this effect and we can see that in just 10 days, we observe a 40% to 45% increase in the VaR. Thus even though the model’s risk measures are ascertained from a 250 day look-back period, the most recent affects in the underlying securities of the S&P 500 offer a dramatic change in the underlying covariance matrix of the index’s constituents, resulting in a significant change in the forecasted Value-at-Risk estimate. This is exactly why calculating risk estimates and paying attention to their forecast can save an investor from a boat-load of losses.

Now we ask how about other countries: How are risk forecasts working in emerging markets for instance? The chart below highlights a VaR timeline over the past year for the MSCI emerging market index. The R² MAC model VaR values are shown as the blue diamonds while the historical VaR is illustrated as red squares, both at 99%. The black dashed line are the 95% CI’s about the historical values. The right-most values are the August-end levels of VaR and are seen turning quickly upward, rightly forecasting the emerging trend of increasing risk.

The increasing risks toward the current market environment are obvious. As are the risks to the EAFE index shown below. We report the VaR in U.S. Dollars, but of course on FactSet you could just as easily reported in EUR, GPB, or YEN currently, with more currencies being added over the next year. These data illustrate convergence of model VaR and historical values within their errors.

To conclude, I just want to harken back to something Jeremy Grantham wrote in his quarterly letter to GMO. Grantham wrote, in a touch of optimism that: “massive bubbles and two equally massive resurrection programs” led to a Fed with no remaining cards to play. Grantham believes another dip will lead us to a more regulated, lower-risk, and “less bubble-prone environment.”

Granted, Grantham wrote this before the U.S. debt credit rating downgrade by S&P and the markets’ recent downward path, but nevertheless it’s encouraging to note that our risk models also demonstrated some foreknowledge of increasing risks and at worst were coincident with “catastrophe,” putting them in league with the talents of one of the great market prognosticators of the current generation! As Jeremy warned of government bond downgrades and further market declines last July, he encourages us to have “optimism” in spite of the children in charge.

Lastly, I encourage you to use risk models for the simple reason that at a very minimum, they encourage conscious recognition of rising risks rather than allowing market decline to act as your alarm clock, for by then it’s usually too late to take defensive action. We leave you with Gavin Cassar and Joseph Gerakos recent results published at a Chicago GARP seminar in their study of how hedge funds manage portfolio risk: “funds that use formal models of portfolio risk did relatively better in the extreme down months of 2008, then those that didn’t.”

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

If it's not another recession, might it be the apocalypse? (Part 1)

Steve Greiner — 2011-10-19T16:54:10Z

The markets are roiling again! The VIX and its shorter-term cousin the VIN are approaching local maxima as seen below. In addition, the GSCI commodity index is falling as are world markets in general. Gold is off from its triumphant rise, but only falling back to its longer term trend (seen second image below). Currencies are falling relative to the dollar, even from safe havens like Switzerland, Canada, and Australia. Even in Indonesia, which has a good balance sheet and an overall bullish sentiment due to its lower reliance on exports to the U.S. and Europe, the rupiah is falling. Other countries in similar supposedly privileged situations are not fairing much better.

What we are experiencing is investors globally dumping whatever they own and running for U.S. Treasuries, German bonds, and the U.S. dollar. Some of it is due to margin calls, some due to liquidity (only the most liquid assets can handle the huge flows), however the nature of the herd is to run when the lions are approaching. These lions (or should we say, bears) aren’t weak, nor few in number: Europe-wide issues, the GIIPS countries debt, the mounting U.S. debt, the U.S. failure to work out fiscal policy, the Fed with no ammunition left. Who can blame prudent investors from chasing the LEAST risky positions? In the new normal, there are no risk-less positions.

Near-term S&P 500 Volatility indicator VIN, show above. GSCI Commodity Index, shown below.

Gold has fallen from the short term spike but has only fallen to its longer term trend line.

The global economy is in bad shape and that’s being kind. There’s even talk of the BRICS countries making loans to the IMF! It used to be the developed world held the creditors. Now the question we might ask, is how have risk models worked to reveal to us this coming collective downside move across assets? Have they revealed nothing as Nassim Taleb consistently chimes to the media? Should we abandon modeling risk? The answer is affirmative for risk forecasting and negative for abandonment of risk models, as we shall now demonstrate.

FactSet is currently integrating within their risk suite, a value-added overlay to our vendor’s risk models, including an overlay for our own multi-asset risk model. While the goal is a full multi-asset class risk assessment, we run tests against indexes of various individual asset classes separately. For, if you cannot forecast risk for a single class, you won’t get it right for a true multi-asset class portfolio either.

One way we measure the efficacy of our process is to plot over the last twelve (12) months, the model VaR along with the historical VaR and 95% confidence intervals (CI) on the VaR itself, for various levels of the VaR (for instance, 95% VaR with 95% CI’s or 99% VaR with 95% CI’s). We plot this and track it through time for a host of both equity and fixed income indexes.

We remind the reader that VaR represents the “minimum” loss that can occur over the forecast horizon with some confidence, not “the loss” which could be far greater. In all examples shown below, we’re using a risk model calibrated on a 250 day (trading year) look-back period and scaled monthly forecasted
horizons. These are the R²Daily Global Equity Model and the Axioma US 2 MH Fundamental model, operating on a large cap U.S. portfolio and the S&P 500. The results are captured in the following graph where we show the faster response R² model (red) and the medium response Axioma model (green).

Though both models are predicated on daily exposures of their estimation universe, the exponential weighting function utilized is designed to make the R² model “faster”. Using a monthly updated model generally will not capture the rising risk appearing at the end of August (not shown) which is why we use daily risk models here. In this graph, the trends are very similar for both models, though the VaR values or levels are different.

These data imply that the monthly Value-at-Risk at 99%, have risen to 12% to 18% currently, up from a July low of between 8% to 11%. This implies that risks have increased by a whopping 40% thereabouts from July end to August end, indicating that any portfolio manager of U.S. domestic large cap stocks should begin reducing risk in their portfolio.

It’s important to recognize that though the level of VaRs may be different for differing models (or other risk statistics like tracking error or total risk variance), the trends usually complement each other at the portfolio level. In this case, one model is a very short and quickly reactive model, while the second one is designed for medium horizon reactivity. Think of their contrast like a 20 day moving average versus a 50 day moving average trend-line, both offer trend information but yield different numbers.

If you compare this chart with those of the VIX and VIN reported earlier, you’ll see that the large volatilities in June of 2010 and today complement those forecasted by the risk models shown here.

A close up of the S&P 500 (normalized to fit on the chart) along with the R-Squared Risk model and S&P VaR from 7/29 until 9/16 is shown below. Notice how the rising risks are captured within this small time frame. A monthly updated risk model would have difficulty capturing this transient but real effect. Within just 5 trading days, from 7/29 until 8/4, the VaR has increased from 10.28 to 11.78 for the large cap portfolio and from 11.39 to 13.0 for the S&P 500, both over a 14% increase in risk.

Tomorrow on the blog, we continue our analysis with a closer look at what happened after 7/29/2011, and an examination of VaR in other countries beyond the U.S.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Beware Germans offering Greek gifts

Sean T. Carr — 2011-10-19T16:53:44Z

As a European (although not a Euro-tied) citizen, I haven’t been able to avoid sitting aghast and watching the ongoing drama that is the discussions going on in the Eurozone these past weeks. The press had a field day as a distressed Mediterranean faced near-certain default but was rescued at the last minute by their heroic Teutonic & Gallic cousins. It has all the necessary pieces for a great bonding story that further cements the ties among the single-currency countries.

Unfortunately, as a story it also has all of the pieces for a great tragedy, and I think that is surely where the tale is going. It is more a matter of when and how the default will come, rather than a question of if it will come. I was looking for information and trying to get a feel for the crisis and to tell the truth none of the signals are good. Let's examine three.

1. The Athens Stock Exchange Index

Above is a 6-month chart that shows a near constant decline with each small spike being a micro-rally following the latest meeting and pledging of austerity (e.g., 7th September: “The Athens stock exchange’s main index closed up 7.98%, cheered by government pledges to speed up privatization and reforms days after EU and IMF representatives cut short a visit to determine whether to grant more bailout aid.”) Indeed the latest set of announcements on September 15 drove the STOXX50 up around 3%. Surely the fact that this regular positive news has no lasting effect has an implication on the quality of it!

2. Government Yield Data

This strange-looking chart is an inverted yield curve—a rare beast indeed that has historically been brought about in a couple of ways: 1) A central bank pushes up short-term interest rates in an effort to combat inflation and the long end of the curve is just slow to catch up (e.g., US Federal Reserve 2004-2006), or 2) People trading that debt are worried they’re not going to get their money back, volumes collapse and prices dive to the floor, thereby pushing the yields up. Which do you think more likely?

3. CDS Levels

We can see that the spread on a 1 Year CDS for Greek government debt is almost 9000 basis points. So in order to insure a sum of €10 million against default the purchaser would have to pay €9 million— not a trade that I can see a lot of people making.

Of course these are not three independent signals, far from it, they are all just the same signal being reflected in different instruments, but it is the extreme nature of these signals that should cause you to take heed. There may well be a short term rally as the markets take the positive words in but I believe it is only a matter of time before we move into the final act— and unfortunately the above data shows I am not on my own.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Similarities and Parallels: Have we just come full circle?

Sean T. Carr — 2011-09-16T18:04:21Z

There is a lot of press at the moment about the similarities between where we are today and where we were 3 years ago today. Back then, we were waking up to see that Lehman Brothers had gone bust, the Credit Crunch had become reality, and we were looking into the financial abyss as far as financial markets were concerned. Today, we have an alarming number of parallels: Greece could be Lehman writ large, we have a bubble in ETFs that is drawing comparison to that in CDOs, and in Kweku Adoboli we have a modern day Jerome Kerviel (or Nick Leeson for us older folks).

Of these three parallels, the one that I find the most interesting, probably because of the personal nature (I was an employee of Barings back when it was still called that) is that of the rogue trader.

The reason that this story resonates is best-highlighted by my summarizing a behavioral finance presentation I saw given by Malcolm Smith of Inalytics at PARM in 2008:

Consider the above image: There are cavemen and they are hunting a mammoth, presumably for food. Can we draw a comparison to today’s markets using such a picture?

Well the risks for taking on a mammoth with a sharp stick are plentiful. We can assume that hunting such animals was done in groups with the expectation that not everyone who headed out on the hunt was going to return. Meanwhile, the reward for bringing down a mammoth would also have been great. Lots of meat for starters, but also benefits in that there would skin for leather, ivory for tools, kudos for hunting prowess, etc. All this led to better-fed family and clan respect: A standard risk/reward scenario where certain individuals see the upside as justification for such large risk.

Now look at it another way. If there is fruit on the trees, sheep in the field, or basically any easier game closer to hand, can the risk of death from hunting a mammoth really be offset by the rewards I mentioned? Look at the picture above from a different light and one can project that there is no fruit on the trees, there are no sheep in the field, there is no easy game to be hunted, and the tribe is facing starvation. If that is the situation then perhaps hunting a mammoth with a sharp stick can justify the risk irrespective of upside.

When Smith gave this presentation he suggested that the proposed systematic legislation to be put in place to limit the risks that individuals were taking were all based on the first scenario. Put another way, it was the potential reward that would encourage such action. But Smith suggested that the situation more akin to the second scenario, where an individual was motivated by survival. In survival mode, legislation won’t work, and the situation is significantly more dangerous. It would seem with the UBS case that that is exactly what has happened.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Empirical durations: A tool to gauge mortgage price sensitivity

Mido Shammaa — 2011-08-31T18:02:59Z

Today's blog is contributed by Mido Shammaa, Product Manager for FactSet's fixed income products. He works on prepayment modeling for FactSet's models and leads the integration of third-party vendor prepayment models.

We’re often asked to explain the results generated by our prepayment models. The question we are asked most often is about effective duration. Yet there isn’t a simple way to answer these concerns because effective durations are driven by two complex pieces of the fixed income calculation engine: the prepayment model and the interest rate process. Both are viewed as black boxes by clients and due to the large number of paths and cashflows, they are very hard to verify.

Empirical duration is a method used to approximate a security’s sensitivity to interest rate changes. Empirical duration uses market prices, so it can be an impartial benchmark that can be useful to compare with effective durations generated using a prepayment model such as FSP.

To calculate empirical duration we ran a regression of daily MBS price changes against yield changes (usually the 10-Year CMT). There are many ways to do that, but we chose the most basic one because its results were the most intuitive.*

We did consider both the method proposed by DeRosa, Goodman, and Zazzarino in 1993 and the one proposed by Steve Manson in 2002, but neither provides realistic results in the current market environment.**

We used daily prices provided by Bank of America/Merrill for generic mortgages from recent vintages (2008 through 2011) with a variety of collateral types. In the following charts we plot empirical durations against effective and coupon curve durations calculated using the FactSet Prepayment Model with the UST as a discount curve as of 7/15/2011.

As we can see the calculated durations track the empirical ones very closely.

The trend invariably brings up the question as to why not use empirical durations for hedging, reporting, or portfolio management. There are many reasons as to why using empirical duration is not practical.

Unlike effective duration, empirical duration is not standardized. Practitioners disagree about what time frame should be observed when conducting the analysis, the appropriate frequency to use, or even what rate to do the regression against. In this case, I used all available prices for the instruments that I covered, but if an instrument has a longer trading history then the time frame becomes an issue.
With effective or coupon curve duration the price sensitivity that is measured is clearly defined as to what rates are being shocked (for effective, it’s the spot rate, while for coupon curve, it is the curve) and by what amount. In the empirical duration calculation, however, it is not clear what is driving the mortgage price changes relative to the 10-Year CMT as mortgage prices depend on a number of factors in addition to the rate level (the spread between the mortgage market and the treasury market, correlations between rates and markets, and market volatility). This inability to isolate the effects detracts from the usefulness of empirical durations in risk reporting and hedging.
Most importantly, while calculating empirical durations is easy to do for actively traded mortgages such as generics, TBAs, and recent vintage agencies; you can get incorrect results if you perform a similar analysis for illiquid or non-conforming securities that do not trade. For newly issued securities, or those without a price history, it would be impossible to perform the regression.

Empirical durations are therefore best used as a sanity check on the results of the prepayment model being used.

* The formula is:

Where and the empirical duration of the security. This method is used by Lakhbir S. Hayre in his 2001 report “Mortgage Durations and Price Moves”

** DeRosa, Goodman, and Zazzarino “Duration Estimates on Mortgage Backed Securities” and Manson in “An Empirical Duration Measure for Mortgage Backed Securities”

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

A retrospective on U.S. debts, and the logic of ceilings (Part 2)

Steve Greiner — 2011-08-22T18:01:10Z

In my last post, I discussed how we should understand the way that our ratio of debt to GDP impacts our economy.

In this segment, let’s look at the increase of our national debt through time

I'd like to earmark which president was in office for the particular increase. The following chart illustrates this perfectly but unfortunately doesn’t take us into 2011. The full impact of debt borrowing during 2011 by President Obama is not accounted for in this chart. (Click to enlarge.)

First, we note the strong rise in debt under President George W. Bush. Clearly, the wars of Iraq and Afghanistan meant we had to raise more funds to fight these wars, among other things. It appears that both Bush presidencies have a steep slope in rising debt. However, I want to call attention to a subtlety missed by most of us. When “W.” took office the national debt was $5.768 trillion and when he left office eight years later it was $10.626 trillion amounting to $607 billion per year of debt increase. However, the debt when Obama took office at $10.626 now stands at $14.071 trillion after just two years. This is a whopping $1.723 trillion per year for Obama.

We can argue all day long about whether he had to do it, was left with a lousy economy by W. or not, but Obama's administration did oversee the largest debt increase in the history of the U.S., amounting to $3.445 trillion in just two years. Another way of looking at this involves observing the debt to GDP through time as opposed to just debt growth. The next plot illustrates this nicely. You can click the image to get a larger view.

Here we show the debt levels as the red bars (axis on the left) and the debt to GDP as the blue line (axis on the right). Now, this chart isn’t up-to-date, it was produced in 2006. I’ve cut the chart off around 2007, so you can see only the accurate data in regards to GDP and debt (which as predicted with 2006 data was much lower than what we actually saw from 2006-present). In 2007, the debt to GDP was only 65% to 66%. Thus one can clearly see that raising the debt ceiling at that time or before that time wasn’t such a significant request from congress—the debt was manageable since it was a smaller percentage of GDP. We as a nation raised enough through taxes to service the debt and still run the government.

Now however, given the data I quoted in the first post, from 2007 to 2011 our debt to GDP has risen to 100% and we’re in danger of not being able to afford the debt payments which makes the request of congress to raise the debt ceiling more controversial. In addition, the economy is stalling, unemployment remains persistently high, and our future is “mortgaged” due to all this debt. Hence the reason many in congress see this persistent trend, started by the second President Bush but exacerbated by Obama as making the U.S. face up to this day of reckoning. The debt to GDP has reached a level which is unsustainable.

The last chart pinpoints where the U.S. is on a world map divided into four quadrants: those with rising debt but a declining deficit, those whose debt and deficit are both in decline, those with with a rising deficit and declining debt, and finally those with both a rising debt and deficit. Countries with rising debt and rising single-year deficit spending are shown in the upper right. This includes the U.S. and Japan. It also represents the worst situation to be in. Countries like Greece, Spain, Portugal and France fare better because while they have huge debt, they are shrinking their spending. Healthy countries like Sweden, Korea and Switzerland are shown below the horizontal like and have balanced budgets and little debt.

In conclusion, the issue about the debt ceiling has everything to do with basic fundamental economics. We in the U.S. cannot afford “everything”. Even we need to curtail spending and stop growing our debt in perpetuity. The solution will involve devaluing our currency, making our debt smaller relative to other currencies, and curtailing spending by cutting welfare and entitlement programs like Medicare, government pension and social security while raising some taxes. There is no other way. Unfortunately this means unemployment will remain stubborn highly for some time and economic growth will be muted also for just as long. Meanwhile, if more regulation and more growth in the size of government occurs then we will have no choice but to take the “Grecian” formula for ourselves.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

A retrospective on U.S. debts, and the logic of ceilings (Part 1)

Steve Greiner — 2011-08-19T18:02:18Z

The debt ceiling has been raised, and a new wave of financial tumult has surfaced regardless of the fact that the U.S. raised its debt ceiling. Still, it is useful to spend some time talking about our national debt, which is still so much a topic on everyone’s lips as a result of the U.S. getting a downgraded S&P rating.

There’s a lot of misrepresentation about the national debt in the media in terms of its size and significance. In addition, many people confuse the budget deficit with the national debt. The budget deficit in lay-terms represents a single year’s difference in income (from taxes) versus government spending, while the national debt represents the amount of accumulated debt from budget deficits over many years. Also, I want to challenge the conception that raising the debt ceiling, simply because it’s been done before in other administrations, was a “good thing” or “no brainer."

I believe what’s missing in these arguments is a historical perspective about the U.S. national debt. It’s not so much a fault of the President of the U.S. per-se (i.e., Bush vs. Obama) as it is about a general failure of our congress, senate, and the sitting president to understand the economic underpinnings of what has been occurring over many years, over many sitting presidents. Moreover, in the past raising the debt ceiling was easy because the debt to GDP ratio was small.

Let's illustrate the concept with an example: Consider if you have income of $50,000 per year and credit card debt of $5000 on a card with $10000. In this situation, you’d have a debt to income ratio of 10% ($5000/$50,000, if your credit card represents your total debts). If you then spend $5000 more on vacation, now your total debt is $10,000. You’ve reached your credit limit and have 20% debt to income ratio. If your job is secure and you’ve had it for many years you can call your bank and ask them to raise your credit limit to $15,000 and they probably would. This can go on until your debt to income level approaches some limit.

The argument about your debt ceiling, your credit card limit, in this example, will get more and more heated until finally the bank says, “no more credit” and stops raising your limit. Now, what is the acceptable credit limit you might ask in percentage of income terms? It is certainly not 100%. If you have $50,000 of credit card debt with 10% interest rate and only a $50,000 income, the credit card debt service will begin to eat away your take-home pay each month.

It’s logical to have a credit limit to protect you from yourself, to protect you from paying too much money in interest on debt, so that you will not have to declare bankruptcy, so you will not have to default on your debt. However, if your income is growing at 10% a year, then next year your income will be $55,000. Your debt to income ratio will fall simply because your income went up, not because your debt decreased.

It’s logical to have a credit limit to protect you from yourself, to protect you from paying too much money in interest on debt, so that you will not have to declare bankruptcy, so you will not have to default on your debt.

For a nation, there are two ways to mimic this trend— the obvious way is to increase your GDP (i.e., grow your economy) and the non-obvious way is to devalue your currency. This analogy makes it all easy to understand. Now, consider: What is the appropriate amount of debt to income, or debt to GDP for a country to not go bust or default?

To put this in a global perspective, the following chart offers a list of countries that most of us recognize their debt as a percentage of GDP as collected by the International Monetary Fund (IMF) and the date of the data. Now, you hear in the media these days about Europe’s woes from the PIIGS countries (Portugal, Ireland, Italy, Greece and Spain, which I rename the GIIPS, as I find PIIGS insulting). I highlight in bright yellow these countries. I also highlight Japan and the U.S. in beige so you can compare the debt to GDP of Japan and the U.S. with the GIIPS countries that are travailing Europe these days.

When you hear people say, “The U.S. is just like Greece”, you can see why they say that. Their debt to GDP is 130% while the U.S. is approaching 100% (it’ll surpass 100% this year). Italy, Ireland, Iceland are all just a little bit ahead of the U.S. and Portugal is just behind us.

Spain is way below the U.S. and France is way up there, as is Belgium along with some smaller countries most of us don’t care too much about. However so is Singapore. Singapore is an example though where their GDP is growing so fast that they’ll stay ahead of their debt, analogous to the 10% income growth of the individual I spoke about. But the other countries all have low single digit GDP growth and some negative growth.

That is where the debt to GDP ratio becomes important. When the debt to GDP ratio becomes large like the U.S., which grew from 40% in 1980 to ~100%, the credit rating agencies begin to lose confidence that the country can make debt payments regularly. In turn, buyers of our debt begin to demand higher interest rates to purchase new debt just like the credit card agencies will raise the interest rate on your credit card. Thus, if the 3% interest rates the U.S. pays on its debt to creditors rises to 5% or 6%, the amount of money paid each year to creditors doubles which leaves less for the government to operate, less for Medicare, less for road construction, and so on. This is why a ratings downgrade, as we saw S&P do, is so important. It keeps the cost of paying debt managable. So we see that Japan has really high debt to GDP and also we know that their economy has stalled.

So the impact of high debt is also that it slows the economy and wreaks havoc with growth of employment, growth of business and lowers the general earnings of everybody. Japan can sustain higher levels of debt simply because most of their owners are the Japanese themselves, while the U.S. has a much higher incidence of foreign buyers of our debt. When Japan pays interest on its debt to its debt holders, it mostly goes to the Japanese people, while when the U.S. pays interest on its debt, it mostly goes overseas, to the Chinese, Indians and resource-rich countries found in the middle east. These are the major buyers of our debt, hence the media talks about the Chinese loaning us money. Indeed they are.

Next week, I’ll continue with a look at our national debt over time.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

ETFs: The next bubble? (Part 2)

Andrew Kovacs — 2011-08-03T13:07:43Z

Last week, on the blog, I talked about the differences I've seen happen in the ETF market in just a little over 10 years. To read my opening post, click here.

Like many, I believe ETFs are a great tool for investors, particularly as an easy way to manage exposures. Nowadays you can use ETFs to pretty much gain or limit any exposure under the sun. Say you want to participate in the growth of Chinese Infrastructure, no problem,Emerging Global Shares China Infrastructure ETF is one of several options available to you.

What’s that? You want to hedge your U.S. large cap risk exposure to momentum? The Russell 2000 High Momentum ETF may be just what you are after. What if you believe gold is where to invest? Then perhaps you should take a look at UBS E-Tracs CMCI Gold Total Return ETN. It really seems like there is something for every situation.

So what’s the problem? My concerns regarding some ETFs and ETNs stems from the old adage “invest in what you know,” to which I strictly adhere.

For many (probably even most) of these funds, the structure is pretty straightforward and they actually own what it is they are trying to track, but for a growing number of funds this in not necessarily the case. Rather than owning some, or all, of the assets directly, ETFs may employ derivatives like total return swaps to replicate or “enhance” a fund. This method of managing ETFs will usually result in even lower fees and in theory should allow the ETF to more closely track the target index or benchmark. These types of funds are often referred to as Synthetic ETFs because they are constructed using derivatives as opposed to the actual physical assets they are supposed to emulate.

"For many (probably even most) of these funds the structure is pretty straightforward and they actually own what it is they are trying to track, but for a growing number of funds this in not necessarily the case. "

Let’s consider one such ETF, the db x-trackers S&P 500 ETF. If you have a little time to spare, then may I suggest perusing this fund’s hefty 912 page prospectus? When I did that I came up with three things that made me a little wary.

I think many investors (I know I was) would be hard pressed to fully understand how this product works.
The fund attempts to replicate index returns via the use of OTC swaps, which of course means we need to consider counterparty risk. Thankfully ETFs tend to have very strict collateral requirements to protect investors from counterparty risk (in fact this ETF is actually over-collateralized), so in theory this type of risk is somewhat mitigated.
There is a section entitled, “Potential Conflicts of Interest”, which basically goes on to say that other “DB Affiliates” (i.e., divisions of the same company that manages the ETF) may act as the counterparty for swap transactions and contracts amongst other things. In fact, Deutsche Bank AG is the counterparty for this particular fund’s swaps.

So why do the above items make me a bit wary? Well, for starters, investing in a product like this is a potential violation of the “invest in what you know” philosophy to which I subscribe. Second, while the db x-trackers may be amazingly collateralized, how can we be certain about the quality of the collateral (e.g., are illiquid or hard to value assets allowed to be used as collateral)? And what happens if something bad befalls Deutsche Bank? Remember,as the counterparty to the swap contracts they are also the ones supplying the collateral. It was just a couple of years ago that some other large, reputable financial institutions became embroiled in circumstances which ultimately led to their default or even dissolution and taught a lot of people that counterparty risk is definitely a real risk. Personally I am not quite ready to assume that a situation like that can never happen again.

Not surprisingly I am not the only one who is concerned about the risks of products like synthetic ETFs. As I was wrapping up this post I noticed a news story with the title “EU watchdog may ban some ETF retail sales," which goes on to say that ESMA (The European Securities and Markets Authority) is considering establishing some controls around the sale of synthetic ETFs to retail investors.

I am not trying to pick on the db x-trackers (or even synthetic ETFs as a group) and I am definitely not suggesting people avoid these types of products. Rather my main goal in writing this is merely to suggest that, while ETFs are a fantastic investment tool, investors need to cast the same critical eye on ETFs that they do on other complicated investments and not be lulled into a false sense of security because of their popularity and easy access. With the seemingly explosive proliferation of Exchange Traded Products coming to market every day I can only imagine that these products will only become more and more complicated and less and less transparent.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

ETFs: The next bubble? (Part 1)

Andrew Kovacs — 2011-07-29T21:12:39Z

Recently I have had discussions with folks inside and outside of FactSet about the seemingly exponential growth of Exchange Traded Funds (ETFs) and their cousins Exchange Traded Notes (ETNs) over the last few years, and whether or not this growth should be cause for concern amongst investors.

When I started at FactSet back in 2000, ETFs were just starting to regularly appear in client portfolios (I know they had been around for quite a while prior to 2000). There were only a few names that came up back then—SPDRs, iShares, and HOLDRS—which I was able to confirm with a quick backtest that showed we had data for about 95 ETFs as of December 2000. I thought as a starting point for my comments I would begin with a quick “then” and “now” comparison of the ETF market and how it has changed since 2000 and today. Please bear in mind that this is purely from my perspective and by no means intended to be a detailed development history of the ETF market.

Then: If a client held an ETF all he wanted to do was make sure we could supply a price and make sure it contributed to his portfolios’ total market value and return.

Now: If client holds an ETF she still wants to be able to represent it as a single physical asset in the portfolio, but she also wants to be able to “look through” the ETF in an attempt to understand the direct and indirect exposures. She may also want to see how the constituents of the underlying basket of securities contribute to the portfolio's overall ex-ante tracking error (which of course requires additional data and functionality beyond a simple price).

"If a client holds an ETF [today] he or she...wants to be able to 'look through the ETF in an attempt to understand the direct and indirect exposures..."

Then: ETFs were primarily set up much like traditional equity index funds in that they actually held the same basket of securities in the same proportions as the equity index or sector they were attempting to replicate.

Now: There are a variety of ETFs now ranging from the traditional index funds, to actively managed ETFs, Commodity ETFs, Bond ETFs, Currency ETFs, and Synthetic ETFs (I’ll come back to some of these in a minute).

Then: In 2000 there were only 3 major ETF sponsors: Barclays (iShares), State Street (SPDRs), and Merrill Lynch (HOLDRS) and about 95 ETFs in the marketplace.*

Now: As of July 2011, there are currently 35 such sponsors in the U.S. and Canada alone (plus a bunch more in handful of other countries around the world) and approximately 1,500 ETFs/ETNs in the marketplace.*

As you can see from the information I've shared above, the appetite for and the number of ETFs available in the market place is rapidly increasing.

In the next post, I will address why I feel investors may want to tread with caution when it comes to the ETF market, particularly synthetic ETFs. I'll analyze some examples and explain some points you may want to consider when investing.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

* The numbers I have provided are based on counts I generated from security screens and data available on FactSet. As a secondary reference I also reviewed some stats collected by the Investment Company Institute (ICI) 2011 Fact Book. This reference provides a great overview of US Registered Investment Companies and provides a chapter on the current state of the ETF market in the U.S. For instance, the ICI Fact Book details how, from 2000 to 200, the number of ETFs in the U.S. grew by 279 and the net asset value increased by about $357 billion USD. From 2006 to 2010 the number of funds increased by an additional 591 and the net asset value climbed another $569 billion.

From News Corp to national crises, don't judge market volatility by the headlines

Bryan Hoefs — 2011-07-18T15:53:25Z

Over the past few days and weeks, the biggest story in the UK has been about News of the World—the tabloid that allegedly committed several crimes in order to publish more and more sensational stories. Journalists appear to be under incredible pressure to come up with headline-grabbing articles.

It's not just the tabloids that are creating sensational headlines. Words like rocked, soars, plummet, panic, and contagion occur very frequently on business headlines these days. Whether it’s the U.S. threatening to go into default if the Democrats and Republicans can’t come to some sort of an agreement by August, Eurozone debt problems, or the potential breakup of the Euro, it all must surely impact the volatility in the markets. After reading recent headlines, it’s understandable if you believed that volatility was currently at an all-time high.

But is volatility high, in fact?

Overall, it’s hard to say it is, at least when looking at the recent past. If you look at the VIX, commonly used to gauge anticipated volatility in the markets, there isn’t evidence for high levels of volatility at present. Yes, it has increased somewhat over the last few weeks, but the VIX is still at its lowest levels since 2007. At the start of the financial crisis at the end of 2008, it was three times higher than today.

The Eurozone debt crisis has caused a lot of concern in the markets, and volatility has increased for the most affected regions. For example, on July 11, the spreads of Italian bonds relative to the German benchmark bond increased by the largest amount since the existence of the Euro, and have hit a record high.

While the European and Italian equity markets have been more volatile than other equity markets, volatility still isn’t at the same level as few years ago. Using the R-Squared Equity Risk model, which is a short term equity risk model available on FactSet, I’m comparing predicted absolute risk of the MSCI Italy index and MSCI Europe. You can see that the levels predicted aren’t nearly as high as a few years ago, even for Italy.

Just to return briefly to News Corp, it’s safe to say that single stock volatility has increased. The stock is down 15% over the last 10 days since story broke. Despite these levels, even this controversial stock is still not at its most volatile point. Moves relative to the S&P 500 were much greater in 2008 and 2009.

So while individual markets and companies have seen increased volatility over the past few weeks, it’s nothing like the systematic increase in risk seen in 2008. What will happen if the Euro breaks up, or the U.S. doesn’t make its interest payments, remains to be seen. This is just another reminder that you can’t judge a book by its cover, and you can’t make assumptions about underlying volatility simply by reading the headlines.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

New Quants: More Reflections on Adapting

Steve Greiner — 2011-07-06T18:01:01Z

There was a lot of smoke blowing immediately after the credit crisis as we looked for sources of blame. One book that got a lot of press was The Quants by Scott Patterson. Mr. Patterson was a Wall Street Journal reporter and clearly has a flare for the imaginative. While The Quants makes what is normally a quite dry subject (like accounting or actuarial science) an easy read and adds adventure to the quant story, there’s much in it that’s inaccurate and hyperbolic. For instance, Patterson presents conversations (one between a waiter and Cliff Asness and one between Peter Muller and Ken Griffin) in language I find hard to believe occurred verbatim.

One of the main themes of the book however is really about how Wall Street whiz kids brought the house down with their impetuous, brilliant, and extremely aggressive nature. It implies that they made their fortunes by robbing less intelligent clients and their investors. It also proposes that, simultaneously, other quants used badly mis-specified models based on the normal curve (specifying default correlation among bonds for instance) to underprice risk and bring about huge losses for the banks. If you took both exaggerations and divided them by 10, you’d probably reach something nearer the truth. The overall pain of losses was spread pretty much across all quants (except John Paulson) and non-quants during the credit crisis.

It's ideas and exaggerations about who specifically was at fault during the crisis that led me to write my book, Ben Graham Was a Quant.

However, one good takeaway from Paterson's book is that it makes you think about whether quants have learned anything about major market turns and whether they’ve adapted their models and modeling techniques to consider the impact of “Obsidian Uncertainties” (i.e., Black Swans) and ELE (extinction level events). The answer to that question is a stark yes.

Have quants modified their models to consider the impact of Black Swans and ELE? The answer is a stark yes.

The first example of it comes with UCITS mandating VaR requirements for EU mutual funds to less than 4 breeches per year at 99% CI. In 250 trading days, 2.5 breeches of the 99% VaR is right on target. UCITS mandate therefore is a signal to quants to “tighten up.”

Another good example comes from the increasing usage of stress testing one’s portfolio against variables that could move your portfolio toward large losses. Thus, it’s no longer sufficient to create an Alpha model based on backtesting through turbulent periods alone. Now quants are examining their quantitatively derived portfolio behaviors by using covariance matrices from the past to forecast the risk from credit crises, LTCM debacles, Asian contagion security dependencies and so forth, all of which involve situations where idiosyncratic risks take a backseat to market risks in a major way.

Crisis events are characterized by factor efficacy falling off considerably while securities increase their correlation as well. When that has happened in the past, quants used to hold firm and wait for the correlative nature of the markets to return to pre-crisis levels. Stock “de-correlation” meant factor efficacy was returning. Now however, quants have learned that these periods may persist for long periods of time and that one way of prepping your portfolio for these events is to examine forecasted risks from short horizon risk models (~1 year of daily values).

Crises are characterized by factor efficacy falling off while correlation among securities increases. When that used to happen, quants held firm and waited for correlations to normalize. Now, quants have learned to expect longer crisis periods and take steps to predict market inflections faster.

The tricky part is, adaptation in one’s overall investment strategy due to market conditions is exactly what Ben Graham taught us not to do. During the technology bubble, for instance, many value investors moved more toward growth only to go out of business when the bubble burst, besides introducing enough style drift that consultants fired them for that reason alone. Ben Graham’s philosophy is about maintaining investment process discipline and not reacting to the whims of Mr. Market. However, for the truly post-modern quant, adaptation is the discipline. From the quant’s perspective, application of the Ben Graham principles to the investment process is about adhering to risk mitigation “come hell or high water." It’s about providing the portfolio with a margin of safety through the methods available in the quantitative art.

The credit crisis has truly allowed quants to leverage their methods and think about risk in new ways. This indeed has raised the IQ of the intelligent quantitative manager.

In this way, inflection points of the market are more quickly spotted than using a long horizon risk model (~60 months) and adjustments to the portfolio can occur by rotating more quickly to factor bets that are more efficacious in the new environment.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Talking risk with model providers: Recap from the FactSet European Symposium

Sean T. Carr — 2011-06-03T18:02:52Z

I was fortunate last week to be able to spend 40 minutes at the extremely popular, standing-room-only Risk Panel session at our European Symposium. In the room were a host of individuals representing all of the risk models currently offered by FactSet:

Sebastian Ceria, President & CEO, Axioma
Dimitris Melas, Executive Director & Head of Research, MSCI BARRA
Anish Shah, Senior Researcher, Northfield
Jason MacQueen, President & Founder, R-Squared
Laurence Wormald, Head of Research, Sungard APT

It was an opportunity for panel chair and fellow blogger Steve Greiner to pose some industry specific questions that he felt relevant to the audience, and also for an exchange of views regarding some of the fundamental differences in the approaches of the different companies.

A transcription could never do the discussions full justice but I did want to highlight a couple of the points raised as worthy of a secondary airing. I have therefore summarized a couple of the questions and selected relevant answers below. I hope that you will also find them interesting:

Chair: What’s the most appropriate way of incorporating commodities into a multi-asset class framework?

Laurence Wormald, Sungard APT: The answer is to build a proper multi-asset class framework–not to select some factors as being suitable for some assets, other factors for different assets. Principal Components is a decent methodology for capturing, on average, the moves across asset classes. It is important to note that modelling of commodities is not just about a spot price, but that it is also important to include the forward curve in the calculation.

Anish Shah, Northfield: Northfield has had Commodities in their model for years, as the factors within their models capture the varying nature of the correlations through regional, industry, and factor exposures. A simple correlation analysis of Gold with S&P 500 shows a swing between positive and negative, so it is important not to get too fixated on an accuracy that cannot be achieved.

Chair: How important is it to use a risk model that has factors that are common to your investment process?

Sebastian Ceria, Axioma : Optimizers find holes in a risk model and ‘loads’ on those holes. One way of getting around this is building your own risk model — the problem here is that if you are wrong then you may not find out until it is too late. Another challenge is understanding how your constraints (imposed or welcomed) affect your alpha capture. Axioma developed a technique called the alpha alignment factor to attempt to solve these ‘risk under-estimation’ problems.

Jason MacQueen, R-Squared: One explanation for the inability of managers with skill to outperform their indices is the unsuitability of their portfolio construction process to reflect where they feel their ‘alpha’ bets should be. The selected risk model used in this process is a major factor in this breakdown. It is therefore crucial to have a risk model that best reflects your own definitions of sector, industry, valuation, etc., to reflect the investment process and maximize ‘alpha’ capture.

Chair: For your particular set of risk models, what is the strongest element of the forecasting methodology, e.g. the risk factors used to build up the systematic portion of risk, or the mathematical techniques used in model development?

Dimitris Melas, MSCI BARRA: By far the most important element of our methodology is to use fundamental factors in our risk models, going back several decades to the work of Barr Rosenborg. Fundamental factors have a number of inherent advantages, aligning very well with the investment philosophy of both institutional fundamental as well as quantitative managers, and also you are able to predict the risk of assets that do not have a long trading history.

Sebastian Ceria, Axioma: Whether you are using fundamental or statistical factors (we use both), it is important to analyze how you build those factors. We use daily data to give better granularity, updating the models daily also, but that in itself brings its own challenges due to the asynchronous nature of global markets. Monthly data updates introduce several problems including “crowding,” as was seen with a lot of quant managers in 2007 and permits people to “front-run” and take advantage of up-coming trades.

Jason MacQueen, R-Squared: I’d also like to add that the mathematical limitations that force a discrete beta allocation to industry, country, etc. is done for convenience and leads to inaccuracy. You cannot assume that all French Banks have a Beta of 1 to France and a Beta of 1 to Finance, allowing these betas to float gives a much better forecast.

Dimitris Melas, MSCI BARRA: The alpha and risk factor misalignment problem does exist, but there are simple and straightforward ways that do not require the creation of a whole new theory and technology to tackle it, one shouldn’t necessarily use a sledgehammer to crack a nut.

---

Finally, I’d like to reiterate my thanks to all of the individuals involved in delivering a most entertaining session, feedback was great from all attendees who relished the opportunity to see the providers in a unique environment (and wonderful setting!)

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

What is the true size of China's monetary might?

Steve Greiner — 2011-05-27T18:03:21Z

I've been thinking about big numbers a lot lately. Specifically, large amounts of money.

No number needs more transparency than the U.S. debt level.

It is always easier to gauge the magnitude of a number when viewed collectively. As with people’s heights, when somebody 5’6” is standing next to somebody 6’5”, it’s easier to grasp their values in when compared.

First for comparison, the U.S. GDP these days runs around ~$15 trillion dollars. That’s $15,000,000,000,000 per year that our economy produces. The U.S. debt is also of similar magnitude, but with a sign change (-$14.7 Trillion) making the debt to GDP ratio about ~100%. It’s 140% for Greece and over that for Japan. The difference is Japan’s debt is 90% owned by its own people, whereas the U.S. debt is half owned by Americans, the other half is owned outside the U.S.

Now, the current budget deficit of the President of the United States is around $1 trillion if there are no cuts (there will be). What does this mean? If this budget is passed by congress, it would raise the U.S. debt by a trillion in a single year, to $15.7 trillion. This amounts to ~$52,333 per U.S. citizen, very roughly.

The U.S. runs a trade deficit every year. In services (consulting, paperwork, general business), we give more than we take in, but by far and away we take in more foreign goods than we give out. This deficit runs to -$668 billion per year or about -4.5% of GDP. China on the other hand has a GDP of about 1/3 of the U.S. of about $5 trillion per year, with a trade surplus of $169 billion, which is about 3.4% of their GDP. However, China has a surplus of foreign exchange reserves of $3 trillion whereas the U.S. has...well, debt. This $3 trillion in reserves is 60% of China’s GDP.

Now, this Chinese surplus is invested in over a trillion dollars of U.S. Treasuries. They on the other hand, have no interest in our dollar falling (devaluing) as it has been, as their investment is losing money when that happens. Nor do the Chinese want the U.S. to default; if it does, they won’t get paid dollar for dollar either. So given the fear of that happening, what might the Chinese invest these proceeds in to diversify away from U.S. Treasuries?

Well, for one the entire amount of commercial mortgages owed in the U.S. collectively is $2.4 Trillion. Meaning the Chinese could pay the entire amount of listed mortgages on commercial real estate in the U.S., take a huge ownership in buildings and land here, and still have $600 billion leftover. The 2008 to 2010 loss in total real estate in this country was $8 trillion, just to put things in perspective.

China could pay off the entire debt of Spain, Ireland, Portugal, and Greece and still have $1.5 trillion leftover, a full half of their surplus. In addition, using this half of their surplus, they could buy all outstanding shares of Apple, Microsoft, IBM, Google, and Exxon.

Or they could spend the whole $3 trillion and buy Exxon, Apple, GE, Microsoft, IBM, Chevron, Berkshire Hathaway, Walmart, AT&T, Proctor & Gamble, Johnson & Johnson, Oracle, JPMorgan, and Google. They’d spend all their money, but considering that on January 1st their surplus was $2.85 trillion and by the end of March it was $3 trillion, after buying all these companies, by the end of next month, they’d have another $15 billion in cash to do something with. If China bought all the companies in the Russell 2000 index of small cap stocks, all 2000 of them, they’d still have $1.4 trillion dollars left over.

All of Manhattan’s taxable real estate amounts to just shy of $300 Billion. China could buy the whole island and have over $2.5 trillion dollars left! We could throw in all the property of Washington D.C. for another $232 Billion and make them overpay and they’d still own Manhattan and D.C. and have $2 trillion leftover!

Now imagine if you will; an alternate universe in which the U.S. had a trade surplus of 60% of our GDP like the Chinese? A whopping $9 trillion dollars instead of -$14.7 trillion in debt! Who of us would be worried about social security, health insurance, and Medicare under that circumstance? Food for thought!

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Is the world running out of commodities?

Steve Greiner — 2011-05-24T18:00:51Z

Recently, I read a very thorough analysis of long term commodity prices from one of my favorite strategists which prompted me to think about natural resources. I wondered: Have we entered a new paradigm when it comes to world natural resource use and distribution? This basic premise has to do with population growth, namely, the rise of China and India and their consumption of resources at a scale the world has never known. All this consumption sits on top of the developed world’s continued use of materials and resources.

For instance, as of 2010 China’s share of global consumption was:

Cement 53.2%

Iron Ore 47.7%

Coal 46.9%

Pork 46.4%

Steel 45.4%

Lead 44.6%

Zinc 41.3%

Aluminum 40.6%

The list of resources that China consumes at great rates will only increase as the world’s second largest economy continues to grow unabated.

Allow me to use oil as a data point: From 1878 until 1971 oil hovered about $16/barrel, a small amount in today’s dollars. However, from 1971 to the present, it rose precipitously and today it is not only at record levels but has moved 6 standard deviation above $16 in today’s dollar terms. Brent Crude closed as high as $117.52 recently. With so many other commodities performing the same way, is this demonstrative of a paradigm shift in global natural resource supply and demand? Another example helps to put a nail in the coffin on this idea: Since 1994, one has to dig up an extra 50% of ore to get the same tonne of copper and this 150% effort has to be done using energy at 2 to 4 times the former price.

I read Jim Roger’s book, Investment Biker in 1997. He had finished a motorcycle ride around the world and much of that book is a mini-summary of global economics. He was the first person I heard talk about the coming commodity boom and his visits to many developing countries during this trip convinced him the world would soon be needing tremendous amounts of raw materials. More than 13 years later, we are seeing this come to pass.

For the U.S., the purchasing power of the dollar continues to fall, especially relative to other currencies. The following chart shows the return of Silver, Gold, and Brent Crude from May of 2009 until now, in various currencies. Hong Kong currency represents the Yuan in this plot since the HKD is pegged to the dollar. Notice however that measured in Swiss Francs, Australian, and Canadian dollars, the appreciation of these three commodities hasn’t been nearly as severe as compared to what U.S. dollar consumers are paying (or earning on these commodity investments). Is the fall of the dollar inviting demand for commodities as a hedge?

This run-up in prices hasn’t been missed, and just before the credit crises of 2008 began, speculators took the media’s blame for the huge price increases even though the CFTC’s Interagency Task Force’s July 2008 Report on Crude Oil said:

The Task Force’s preliminary assessment is that current oil prices and the increase in oil prices between January 2003 and June 2008 are largely due to fundamental supply and demand factors. During this same period, activity on the crude oil futures market—as measured by the number of contracts outstanding, trading activity, and the number of traders—has increased significantly. While these increases broadly coincided with the run-up in crude oil prices, the Task Force’s preliminary analysis to date does not support the proposition that speculative activity has systematically driven changes in oil prices.

There have been other reports offering the same vindication that the fundamentals of supply and demand are changing such that price rises of commodities are due to shortages. Currently corn stockpiles are at decade lows. How can this have any other impact other than that corn futures must rise, especially when growing middle class consumers in China, Indian, Vietnam and Indonesia are hungering for more (historically) western styled foodstuffs?

The next chart documents commodity index price rises in energy, petroleum, industrial and precious metals, agriculture, livestock, and softgoods. Never before has the correlation across varying commodities been so high (i.e., all commodities moving lock-step in one direction, up) other than during WWI and WWII when shortages abounded on a global scale. Fortunately we’re not in a global war, but the cause is likely the same, global shortages of commodities.

Another point of reckoning has to do with the much larger availability of commodity ETF’s and mutual funds which are allowing the retail investor to participate in this asset class. A decade ago one had to buy physicals, with ensuing storage problems, or futures, both difficult for the retail market to handle. In addition, the emergence of pension funds and institutional asset managers which make commodities part of their holdings too gives credibility to the bull market in commodities as well as helps to keep prices higher by creating demand for the asset class.

Nevertheless, regardless of the cause, the demand means there’s more reason for commodities prices to continue to rise until this demand can be met. The question I’m having a hard time answering is, will it?

Getting a picture of your portfolio holdings, from any angle

Chris Ellis — 2011-05-24T17:58:57Z

On May 10, Chris Ellis will present on Security Exposures Analysis at the CFA Annual Conference in Edinburgh, Scotland.

Security Exposures Analysis, which we have addressed on the blog before, helps users understand how trends in the market may impact their holdings. For example, it can provide a manager with a view of which portfolios have a heavy weighting in Japanese banks. Chris describes that you can take the analysis further by focusing on thresholds (e.g., every security representing more than 3% of portfolio weight).

As a preview to Chris's presentation at the CFA Annual Conference, the following 10-minute video helps to introduce the topic. For more in-depth information, listen in on parts 1 and 2 of our podcast on the implementation of Security Exposures Analysis in FactSet, linked from the bottom of this post.

Preview Chris's CFA Annual Presentation "What's Your Risk? Security Exposures Analysis"

See it live, May 10, 2:00 p.m. in Edinburgh, Scotland at the CFA Annual Conference. Details here.

For more information on Security Exposures Analysis, please listen to our two part podcast series.

Part 1: Understanding the basics: Why exposure analysis matters

Part 2: Examining the trends: Expanding exposure analysis to boost manager performance

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Defining "Quant" After the Credit Crisis

Steve Greiner — 2011-05-17T18:10:09Z

Warren Buffett has an undying reputation as perhaps the most quintessential investment manager who has ever lived. However, what is not as well known is that he was a student of Benjamin Graham’s. I was overheard some time ago as describing Ben Graham as being a quant. I brought this up because of the shellacking quants were taking: Being accused as the cause of the credit crisis of 2007-2009. The point I was making was that, in spite of the hysteria of the credit crisis (when quants were loudly called the harbingers of the financial meltdown) Ben Graham’s own words put himself in their camp and nobody would argue about his investing acumen nor discredit his methodologies the way they were disparaging quantitative investment methods.

Way back in 1949 when Graham published The Intelligent Investor, he listed seven criteria that in his opinion defined “the quantitatively tested portfolio.” There cannot be any other interpretation than that of the author himself, who concluded that the application of these criteria builds a quantitatively derived portfolio. Thus begins quantitative asset management, its birth given to us by Benjamin Graham, who, even in the latter years of his retirement while living in La Jolla California, continued to research imaginative ways to “auto-magically” invest one’s assets in purely mechanical ways.

To be a quant puts one in league with Ben Graham, practicing active management.

Ben Graham stated, “I deny emphatically that because the market has all the information it needs to establish a correct price, that the prices it actually registers are in fact correct.”

His sentiments are echoed in the words of several esteemed managers and quants who came after him. These investors include Mohamed El-Erian, the very successful Harvard Endowment CIO who now works at PIMCO; Jeremy Grantham of GMO; and lastly, George Soros, who said, “First I contend that financial markets never reflect the underlying reality accurately, they distort it in some way or another and those distortions find expression in market prices.”

It is an interesting observation to me that many long-time practicing investors with healthy investment records stand on the side of inefficient markets, whereas inexperienced (that is to say, from investment management experience) academics support the opposite view and the Efficient Market Hypothesis.

Thus we begin both a diagnosis of Ben Graham the man and his investing technique and the application of modern quantitative processes to Graham’s methodology. Thinking of all this, I began to write Ben Graham Was a Quant, in part to remedy the miscommunication that led to a disconnect between how the quantitative method ought to be practiced, and how it was interpreted to be practiced by the public during the Credit Crisis. I tried to define in the book what exactly quants are, how they ply their trade, and what the benefits are to quantitative asset and risk management. For instance, the following charts from chapter 6 in the book, typifies the ease with which Ben Graham’s “recipe” can easily be coded within FactSet’s Universal Screening module. Here we show very simple FactSet mnemonics for a simple screen, using Graham’s methodology.

This kind of script can also easily be ported directly to FactSet Alpha Testing, where returns can be calculated simply by sorting of these factors over a long time period as shown in the next table.

Then, the output of the Alpha Test can be imported directly over to Portfolio Attribution module within FactSet for reporting and independent analysis. The ease of building these models; running them through a series of tests; and reporting their returns, characteristics, and statistics has surely made the workday easier and more productive for the practicing quant.

Of course, my book covers much more than what quants do and how they work. Throughout, I cover defining Alpha and risk, then work on using FactSet software in Alpha Testing to run factor studies and statistics. Finally, I attempt to tie the two parts together by discussing where Graham gets his “Alpha” from and how large it really is. There is, of course, much more than I can address here, but I’d encourage anyone interested in really understanding the depths of the quantitative method, as defined by the great Ben Graham, to consider a look inside the cover.

Three minutes on Ben Graham Was a Quant

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

The Value of Updating Risk Models Daily

Dr. Sebastian Ceria — 2011-05-17T18:10:42Z

Contributed by guest blogger Dr. Sebastian Ceria,
CEO at Axioma, Inc.

Long-term Investment Approaches Cannot Ignore
Short-term Volatility

Most institutional investors seek long-term returns. So the idea of using risk models that are updated daily has often been dismissed as overkill by portfolio managers focused on long-term investments.

The argument goes like this: Risk models that are updated daily, even if calibrated for longer investment horizons, will only drive one toward short-term trading and excessive turnover. Such results are, by definition, at odds with a focused long-term strategy.

In theory, that sounds fine. In practice, unfortunately, the argument crumbles as portfolio managers confront the need to respond both to asset-owner angst, compliance constraints within their mandates, and the effects of very real events with global impact. Indeed, when one considers the nature of investment management today, and the contract between investment managers and their clients, it’s simply not realistic to expect asset owners and portfolio managers to blithely ignore unnerving events of the sort we’ve seen lately: Japan, the unrest in the Arab world, Portugal, the Gulf oil spill, and so on.

Long-term investment strategies cannot ignore the short term. And let’s face it: Many portfolio managers rebalance more frequently than they are likely to acknowledge. Why? Because nowadays new information that affects their decision process is constantly available. Or perhaps they are adding funds to their portfolio. Or withdrawing funds. Or adapting in some other ways to changes in the market.

Which brings us to the disaster in Japan, a timely case in point.

Risk models can get hung out to dry for their failure to effectively predict events. The earthquake in Japan was, obviously, a sudden, wholly unpredictable natural disaster. To say that risk models—factor risk models, anyway—failed to predict the tsunami is, of course, absurd. The only course of action in a case like this is to adapt to the situation. The only way to do that is to update your models with new information.

The question is, do risk models that are updated daily react to such events fast enough to provide useful insights? Research papers published by Axioma on March 18, 2011 (Market Aftershocks? The Global Impact of the Japan Earthquake as Seen Through the Lens of Axioma’s Daily Risk Models) and a subsequent paper published on March 28, 2011 (Market Aftershocks? A Seven-Day Update) addressed this very question. Axioma’s risk models are updated daily, so we were able to observe market developments as they occurred. Furthermore, risk models updated on a daily basis also provide access to up-to-date risk factors (B/P, ST/MT momentum, size, etc.). These factors are particularly relevant for value/growth managers who take exposures to the respective factors very seriously. While the covariance matrices evolve slowly, the changes in the risk exposures can be quite dramatic in response to major unexpected events, thereby adding additional value to daily updates.

The Japan earthquake and tsunami occurred on March 11; roughly speaking, the middle of the month. For portfolio managers with a direct Japan exposure, the disaster clearly mandated a rebalancing. If they used a monthly model, there were but two choices: use either the old covariance matrix from Feb 28^th, which they knew was wrong, or wait 20 days for an updated model. The opportunity risk is too big to even contemplate.

Event-driven market surges in volatility and opportunity are likely to be the the rule rather than the exception going forward. The current environment of high correlations, combined with the highly liquid cash flows into and out of ETFs, which have a leverage effect on the market, makes the market especially susceptible to rapid responses to market events.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Is Diversifying in Asia Dead?

Willett Bird — 2011-03-18T18:00:49Z

I recently had an interesting meeting with a senior Portfolio Manager here in Hong Kong at a large multi-national firm which has been working in Asia for over twenty years. After the initial pleasantries, I asked her how her funds were performing this year: “Great, we’ve been bullish on China and Oz and those have both paid off for us well this year.” Yet I was more intrigued by her reply when I enquired about fund inflows. I’m paraphrasing here, but basically her answer was:

Inflows have been fantastic as there continues to be a global perception that investing in Asian equities is a materially diverse investment. This simply is no longer the case. Americans and Europeans think that by shoving money across to Asia they will immediately pick up diversification benefits, and while I am the beneficiary, they are simply wrong.

Upon returning to my desk, I decided to do a little research. On the web, it was easy to find arguments in for both sides, and all of them seemed (or claimed to be) “conclusive. So I decided to do a little research on my own. I went into FactSet’s Chart Center application, selected the FactSet Market Aggregates for the major countries in the Asia Pacific region, and ran a simple rolling 2-year correlation versus the FactSet Global Market Aggregate.

Click image to enlarge

Clearly, there has been a material change in the diversification benefits of investing in Asia over the past 10 years, as you can see the spreads between countries grow tighter and tighter throughout time. My client was correct that over the last 10 years, the correlations versus the world have gradually converged, especially leading into the Global Financial Crisis. During this period, the correlations remained high and steady but it is worth noting that over the last 6 months, the correlations between these countries and the World have started to taper off.

Balanced Risk Explained

Steve Greiner — 2011-03-17T21:45:03Z

In our interview with Dr. Steve Greiner, FactSet's Director of Portfolio Risk Research, we discuss the concept and details of FactSet's Balanced Risk.

Listen to our 15-minute discussion with Steve Greiner using the player below, or read highlights from the interview below to understand more about the unique ways that Balanced Risk can be used to understand a cross asset-class portfolio.

More on Steve Greiner's presentation at GARP's 12th Annual Risk Management Convention, March 9 in New York City.

What does Balanced Risk accomplish?

From a 10,000-foot view, balanced risk sets out to accomplish three things: 1) Allowing the user go back through market time periods to understand risk based on real past circumstances, 2) Determining cross-asset class correlation, 3) Providing added Value at Risk and Stress Testing measures that fit in well with vendor models also available on FactSet.

How does building a risk model for the equity side differ from building one for the fixed income side of the market?

On the equity side, risk models are really built as what we would refer to as a multi-factor risk model. Due to liquidity issues, you cannot create a multi-style risk model for fixed income. You must, instead, understand what common fixed income risk factors exist and how they interact with interest rates.

Why is VaR considered to be only partially accurate and how can its accuracy be improved?

VaR, by definition, cannot be exactly accurate.Whether we use a risk model with one year of daily data or 60 months of monthly data, VaR is still attempting to predict how your portfolio will look at some future point. To use an analogy of the weather, if you look outside and attempt to predict the weather over the next two hours, or if you try to predict the weather over the next two weeks, we know that your first forecast will be more accurate. Therefore, regardless of whether your VaR numbers are built on a longer or shorter look-back period, it is really a matter of whether you're asking your VaR to predict portfolio performance one week, one month, or one year from today that determines accuracy. To summarize, the shorter the forecast horizon, the greater the accuracy of the model.

What is one unique aspect of FactSet's risk product?

We allow the user to go back in time and chose from specific stressed periods of the market and choose the covariance matrix from that time period. The user can then refer to that covariance matrix to calculate the VaR of his or her portfolio today if that situation was to occur. So essentially, you can compare what your VaR would be using the dependence structure of assets in the past versus the structure today.

However, whether you create a VaR for fixed income based on the covariance structure of the past or the one that exists today, the only difference for fixed income will be the period of time of data. For fixed income, the issue comes back to being able to use Monte Carlo scenarios using interest rate shocks, and so forth.

Our Most Anticipated Sessions at GARP's Annual Risk Management Convention

Michael Pappas, Quantitative Specialist & Matthew Cioppa, Senior Consultant — 2011-03-17T21:46:25Z

This March 8-9, FactSet will be attending a two-day conference hosted by GARP (the Global Association of Risk Professionals). We’re very interested in a number of sessions and hope to see some of our loyal blog readers there. For now, our attendees wanted to tell you about a couple of sessions that they're very interested in attending at this year’s GARP Annual Risk Management Convention. Perhaps you’ll see our Quantitative and Risk FactSetters at our booth.

Note: If you’re attending, don’t miss FactSet’s own session on Balanced Risk. Learn more.

We’re most excited to check out the session on “reverse” stress testing: Growing Role of (Reverse) Stress Testing (Tuesday at 1:30 p.m.). As we understand it, “reverse” stress testing includes thinking of the possible negative outcomes or vulnerabilities for the firm, enterprise, or portfolio, and then identifying scenarios that might cause this to occur.

A good stress testing set-up needs to take multiple factors into account to be robust and dynamic, and managers or creators of those factors/stresses should seek to choose a set that includes both events likely (or feared) to occur, as well as events which have specific meaning to the holdings of the portfolio or enterprise in question. For example, the Portfolio Manager who also engages in currency speculation will want to test his portfolio against changes in exchange rates in addition to index and economic shocks.

From the portfolio aspect, there seems to be little or no difference between reverse and regular stress tests, so it will be good to gain some color on how this applies on an enterprise risk level and what tools can be used from our financial risk background to help achieve a stress test on an enterprise level.

Unfortunately the session on reverse stress testing occurs at the same time as another interesting one: Navigating the Regulatory Landscape – Practitioner’s Perspective (Tuesday 1:30 p.m.). It would be interesting to hear some of the public opinions from clients on how these regulations will actually affect them in their day-to-day. We’re also curious as to what extent banks are losing top talent to hedge funds, or what, if any, repercussions have been felt internally for the groups impacted by these regulations. Given the potential overlap of Dodd-Frank and Basel III, there’s bound to be a lot of contention about how the two rule sets interact, and what rule set will be enforced in the case of an interaction or contradiction. Is it just me, or does anyone get the visual of a 1-pane New Yorker cartoon where a kid pits his parents against each other, whining “But Mom said it was OK?!”

Another interesting session is in the track “The Next Crisis.” Akshay Kapoor has a lecture entitled Managing Credit Risk in a Post Crisis World (Tuesday 4:30 p.m.), and we’re eager to hear his take on how the recent credit crisis is affecting trends in the use of credit risk management by institutional investors. The crisis taught us that there is a great deal of instability in the credit market and suggests that credit models that are currently being used might need to be revamped. The emphasis on flexible models and on understanding how risk impacts the credit markets has been a continued interest among FactSet’s risk clients and indeed our own session at the event discusses the importance of modeling risk for the fixed income side of the portfolio.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Capturing Euro-Sovereign Default Correlation through Contingent Claims Analysis

David Mieczkowski — 2011-02-10T19:01:53Z

You’re the managing director of enterprise wide risk management at a major bank. A European balanced mandate portfolio manager in the asset management group is boosting his yield by being overweight the PIIGS and underweight the Euro Stoxx 50. The FX desk is managing a healthy P&L by putting on a USD/EUR carry trade using a weighted average position in Euro-Sovereign treasury bonds. A European CDS trader has sold a first to default swap on a basket of Spanish and Portuguese firms, with a large notional amount, to a European hedge fund and is partially hedging it by taking short positions in Spanish and Portuguese CDS. How do you assess the individual and combined risks to the firm posed by these three activities?

Obviously, a good Euro-Sovereign credit model is necessary, but as the examples point out, a single country model isn’t sufficient. We need a model that is capable of capturing not only the interdependence of the Euro-Sovereign credit risk itself, but one which is also capable of capturing the interdependence between Euro-Sovereign credits, Euro-Corporate credits, and European equity positions.

In a pair of earlier posts (Why default correlation matters Part I & II), I discussed how default correlation is central to analyzing the VaR of a portfolio of corporate credits. I also outlined how firm value models (equivalently, Contingent Claims Analysis) provide a convenient framework to capture correlation between corporate credits, as well as between corporate credits and equity. In this post, I’ll discuss how that framework can naturally be extended to include Euro-Sovereign credits, and provide some evidence of its efficacy.

The Euro-Sovereign Spread Model

The model will use the basic contingent claims framework that underlies the Merton model. To use this balance sheet centric model, we will need to determine suitable inputs for:

The market value of the junior liability claim (equity)
The volatility of the junior liability returns
The notional amount of the senior liability claim (bond) or Debt-to-Equity ratio
The duration on the senior claim
The risk free rate
Recovery rate upon default

With these inputs, we can determine the asset value, the asset volatility, and then the market value of the senior liability claim (bond price). The last input is technically not needed in the Merton framework, but it is convenient to specify a recovery assumption and utilize the approximate relation amongst spread, probability of default, and recovery to determine the spread:

Obtaining these inputs in the corporate model is straightforward. For the Euro-Sovereign model, I use the following as proxies:

FactSet Country Aggregate index value (in Euros)
Implied volatility of the FactSet Country Aggregate index
returns, determined by taking the 60 day trailing historical volatility multiplied by an implied volatility factor. The implied volatility factor is derived by computing the ratio of the VStoxx implied volatility level over the 60 day trailing volatility of the Euro Stoxx 50.
Global Insight’s three-year ahead Forecast Debt-to-GDP, reset on a yearly frequency
Average Life of the scheduled P&I on all outstanding Sovereign debt
Rate on the Euro benchmark with equivalent duration as above
Country dependent recovery that ranges from 35% for Greece to 85% for France

The Results

First we present the time series plots of the actual and model predicted spreads for the major¹ Euro-zone members.

Non-PIIGS:

PIIGS:

Overall, the model predicted spreads track actual spreads fairly well, although there is some divergence for some of the PIIGS in the last eight or so months, where the model predicted spread is too low. One explanation for this is that the implied volatility for the PIIGS is, in fact, higher than the proxy derived using the VStoxx. Below I show the time series of implied volatilities that matches the existing spreads.

Non-PIIGS:

PIIGS:

The calibrated volatilities match the implied volatilities for the Non-PIIGS better than the PIIGS. This is not surprising, since we use the VStoxx to help derive the country implied volalatility, which is clearly going to be a more accurate measure for the countries that have representation in that index. Note, in fact, that of the PIIGS, the two countries that have representation on the Euro Stoxx 50 (Italy and Spain) have calibrated volatilities that are closer to the VStoxx derived volatility than Greece, Ireland or Portugal. What is also of note is that the implied volatilities calibrated from the spread model are within the realm of plausibility, suggesting that the Contingent Claims Analysis approach has merit.

Now that we’ve looked at performance of the model on a country by country basis, let’s turn to measuring the joint predictive power. Let’s start by taking a look at the historical correlations.

Pair-wise spread correlations are fairly high among the 10 major Eurozone countries in general, but the PIIGS and non-PIIGS subgroups have higher correlations in-group than out-of-group.

Each country’s daily spread changes are negatively correlated with their FactSet country aggregate equity index returns, while being positively correlated with daily equity index volatility. In particular, spread/return and spread/vol correlations differ from zero more significantly amongst the PIIGS than the non-PIIGS.

Now the correlations amongst the actual country index returns, volatility changes, and model predicted spreads:

On the whole, the correlation structure is reasonably well preserved. The better the predicted spreads for a country match the actual, the closer the correlations match. What’s worth noting is that significant amount of the correlation structure is preserved even for the countries that are the least accurate. For corporate credits, I showed in prior posts that the VaR of a portfolio of investment grade credits was dominated by the default correlation and that significantly more yield could be achieved at the same risk through careful portfolio construction that gave full consideration to the correlations. Contingent Claims Analysis shows that the same holds for the Euro-Sovereigns.

¹ Germany and the Lilliputians (Cyprus, Estonia, Luxembourg, Malta, Slovakia, Slovenia) are not displayed. Germany effectively IS the Euro benchmark and the others do not have reliably liquid spread or equity index data.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Vlog: What's the equivalent of idiosyncratic risk for fixed income?

Steve Greiner — 2011-01-12T19:01:04Z

Classic equity risk models include idiosyncratic risk. Steve Greiner explains what FactSet uses as an equivalent measure for fixed income.

Risk and Reward: Risk Budgeting and Cross-Asset Correlations

Dr. Laurence Wormald — 2011-01-06T19:00:51Z

Contributed by guest blogger Dr. Laurence Wormald, Head of Research at SunGard APT

It’s a truth universally acknowledged that a fund manager seeking to take effective control of investment risk across asset classes must be using or looking to use risk budgeting. However, it may be a lesser known fact that to implement risk budgeting successfully, the most important element is a proper estimation of cross-asset correlations.

There are those who allocate risk to country/region/asset class desks without managing their interaction. This group may be faced with regular surprises, as these interactions can either lead to risk increasing or risk reducing. There can be no expectation of reward without a proper estimate of the risk incurred in these interactions, and thus the portfolio is probably less efficient than it could be.

On the other side, there are those who budget properly to create a risk reduction via cross-asset-class correlations. This group can take more risk in the alpha generation process – a genuine benefit to the investment manager from diversification.

Being able to monitor actual risk expenditure against budget, via what is sometimes called “covariance accounting,” to break down tracking error or total volatility in terms of the cross-asset-class correlations is a vital investment capability that requires a robust underlying risk model.

The robustness of the correlation estimates in the multi-asset class models is ensured by using the same principal components methodology which SunGard uses for single-asset-class models. Principal components analysis is a powerful technique for separating signals from noise in any dynamic system, and that is what is required to robustly estimate the systematic correlations between assets within different classes (such as sovereign and corporate bonds, equities, and commodities). Historical data is always noisy, but by using principal components techniques, that noise can be effectively filtered out before estimating the systematic correlations.

However, as the last few years have shown, risk management is an art as well as a science. The art comes in choosing what we call the “Estimation Core,” that is the set of assets and macro factors used in estimating the APT components and in choosing the most appropriate number of principal component factors for each multi-asset class model.

Our first criterion for inclusion in the estimation core is that the historical data be fully validated, so that problems associated with stale pricing or missing returns are minimized. For a model based on weekly data, we require 180 weeks of scrubbed returns data for any asset which will be included in the estimation core.

SunGard APT’s research team uses extensive back testing in selecting the best-validated datasets for each asset class, including the most important macro series (selected from a set of approximately 30 macro series such as FX rates, key interest rates and credit spreads, equity, commodity, and volatility Indices) across all assets before extracting the principal components. In selecting, we look for significant explanatory power both in periods of normal market behavior and during market crises, considering crisis scenarios back to the 1990s.

Next, we separately choose estimation cores for each of the major asset classes (consistent with those for our single-asset-class models) before combining them in the multi-asset class estimation. In this way, corporate bonds, for example, inherit exposure to their issuers as well as to rates and credit spreads, while the cross-correlation of equities to commodities is simultaneously estimated with that to FX.

Finally, we test on random matrix theory to check that the principal components are in fact capturing the behavior of the systematic driving terms which drive cross-asset-class correlations. This approach makes the multi-asset class models robust enough for FactSet users to estimate a complete set of cross-asset correlations for use in risk reporting, risk budgeting, and optimized portfolio construction. For more information on optimized portfolio construction, view our recorded webinar on robust optimization.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Taking Risk's Top Five Posts of 2010

2010-12-21T19:01:36Z

Our worldwide experts posted dozens of entries this year sharing their latest research and insights. These posts were the most popular of 2010:

Is it a small world after all?
Do regional models still matter in the global economy? A comparison of global models to specialized regional and local models.
The difference a daily risk model makes
What can daily calculation of factor returns, covariances, exposures and residual risks do for a risk practitioner?
The illusion of stability
Why a less volitile year is bad news for risk management.
How skillful a manager is Paul the Octopus?
Was the late Paul the Octopus really skilled in his World Cup predictions? If so, where does that rank him against some of the greatest fund managers?
Introducing Betamax: A new measure of covariance stationarity
Risk managers are recognizing the importance of correlation, but there have previously existed precious few quantitative tools around to address correlation risk.

Thanks for reading in 2010! Don't miss commentary from FactSet's top risk specialists in 2011. Subscribe by e-mail or RSS to receive the latest posts as they are published, and join in the conversation by posting your own comments and questions.

What area of risk management presents firms with their most acute challenges?

Chris Ellis — 2010-12-16T18:20:41Z

This week, I participated in a Risk and Compliance web seminar hosted by Waters Technology. Our panel attempted to tackle challenges faced by risk managers and compliance officers as they try to get as close to a real-time view of risk exposure as possible.

If you’d like to listen to the full presentation, you can access the archived recording here. During the discussion and later from clients, I was asked some noteworthy questions, and I want to share my views in this forum.

What areas of risk management currently present firms with their most acute challenges?

This question was presented to the Waters Technology audience, which answered “Getting a single, coherent view of exposure across business lines throughout the entire organization.”

With the release of Single Name Security Exposures analysis last week, FactSet is able to help investment managers directly address another key facet of this challenge.

Single Name Security Exposures lets you quantify and analyze your exposure to a given security or group of securities across all or a subset of portfolios, equity or fixed income, at a single point in time or across time. There are two parts to this. First, you define the universe. That universe could be a watch list, securities passing a formula, securities passing a screen (e.g., all Continental European microcap biotech stocks with a bond rating above or below a certain measure), or top N largest and smallest exposures. You can define the universe at a security, issuer, or ultimate issuer basis.

For part two, you define the group of portfolios. This could be the entire enterprise – all the portfolios in the organization. You could group by geographic location, such as everything in your London or Frankfurt or Tokyo office. Or you could choose portfolios of a certain mandate or portfolios managed by a particular manager or team. Really it’s any set of portfolios you want.

Once you’ve defined the universe and group of portfolios, you can think of security exposures as analyzing the intersection of a Venn diagram. Dissect the data in any form you like to examine exposures there.

What are the challenges facing financial institutions in developing a reliable cross-asset, firm-wide view of their risk exposure?

When you look at a cross-asset firm-wide view, the key piece of data that’s particularly challenging is getting the entity data to map up subsidiaries to parent companies and then to ultimate parent companies and then multiple asset classes. That gets complicated quickly when subsidiaries are issuing debt that’s ultimately the responsibility of the parent company and therefore linked to the risk of the parent. We’ve spent an enormous amount of time on this entity data data as part of our ever-growing content collection effort. That data is used in quite a few FactSet applications (particularly Portfolio Analysis), but Single Name Security Exposures really brings it to the surface as a nuanced application option.

What does the parent/child entity data cover?

Our entity mappings include equity, preferred stock, debt, equity options, and credit default swaps. You control on a per-report basis how broadly “parent” is defined. Decide: are you analyzing issues, issuers, or ultimate issuers?

Does the parent/child relationships data also account for ADRs and Ordinary Shares?

Yes. The parent/child relationship does match ADRs and parents. You can choose to match those only, include all share classes, all securities from immediate issuer, or all securities from ultimate parent. As an example, you can link PNC and National City debt or AES and debt issued by Indiana Power and Light Company (IPALCO).

Talking with our clients, we believe this type of cross-asset firm-wide exposures analysis will only become more important as compliance standards tighten. Contact me if you’d like to learn more or to request access to an in-depth web demo of the application. Or leave a comment below if you have a question I didn't cover.

You can also learn more about Single Name Exposures Analysis and try the product for yourself at our 2011 Investment Process Symposium. Register by December 15 to receive the early bird rate of $499.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

What’s your risk? Single name security exposure analysis

Chris Ellis — 2010-12-08T19:39:16Z

This Thursday, I'll be delivering a live webcast on FactSet’s new Single Name Security Exposures tool. Single Name Security Exposures lets you look across all portfolios or a subset of portfolios to quantify your exposure to a security, an issuer, an industry, a country, or a specific set of securities.

Simply put, Single Name Security Exposures tells you how much you own and where you own it. The analysis is straightforward, so it's easy to act upon. This isn’t a predicted standard deviation of excess returns, and there isn’t a theoretical dimension. The results are risk that everyone can easily understand.

Single Name Security Exposures clearly connects to the key news of the day. When a major market event happens, it isn’t clear when that event will be reflected in ex-ante risk numbers. Also, it isn’t necessarily clear what change will result from the event. With Single Name Security Exposures, the “when” and the “what” don’t have any ambiguity. It isn’t ex-ante. It isn’t ex-post (though you can run historical analysis). It focuses on right now.

The ongoing reporting aspect of Single Name Security Exposures revolves around compliance, while the ad hoc value relates to the news of the day or a big market event, including:

Country exposures like financial concerns in Ireland or political concerns in South Korea
Company exposures, such as outstanding results from AAPL
Industry exposures such as disappointing numbers of U.S. existing home sales
New statistics suggesting, for example, that the number of Americans with diabetes will spike in the next 30 years

As a risk manager or CIO, when you formulate your views on macro trends that will guide the market or as you try to act on the ideas from an investment committee meeting, you want to know your exposure to the companies that embody trends or your key insights. Single Name Security Exposures is a compelling complement to ex-ante risk analysis in that research.

Now, when we talk about this type of analysis, it is easy to cite the Greek debt crisis or Enron as great examples of why we need to consider these exposures. Similarly, when we talk about ex-ante risk analysis, it’s easy to focus in on excessive risk or high risk numbers as always bad. This clearly isn’t true. While the most glaring examples of why Single Name Security Exposure analysis is relevant tend to be negative, the analysis lends itself to positive or bullish investment ideas even more than ex-ante risk analysis.

When a company beats estimates or announces a break-through innovation (in technology or in medicine or in a consumer product), you want to understand your exposure. In the last few years, Apple is a great example of this. In the context of Single Name Security Exposures, it is the counter to Enron.

As another example, I mentioned wanting to understand your exposure to a significant increase in diabetes in the U.S. That’s bad news for society, but from an investment perspective, if it occurs, it is going to mean increased importance and profit for a group of companies.

To support this analysis, we combine portfolio holdings data clients store on our systems with comprehensive parent/child entity data. By having clear connections between parent companies, subsidiaries, and all related securities, you can examine exposure to an individual security or all securities related to an issuer, including common equity, preferreds, debt, equity options, and credit default swaps.

I hope you'll join me this Thursday, December 9 for a first look at Single Name Security Exposures. I'll also be taking your questions. The live webcast will take place at 2:00 p.m. EST/11:00 a.m. PST. Register here.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Global risk models that aren’t skewed by asynchronous trading

Sebastian Ceria — 2010-11-22T20:01:33Z

by guest blogger Sebastian Ceria, Ph.D., President and CEO, Axioma, Inc.

In a recent abstract, Professor Bernd Scherer warned against the use of unadjusted daily stock market data to update global risk models, a practice he said results in “spuriously low correlations between stock markets.” As a provider of global, regional and multi-country risk models — models that are all updated daily — we couldn’t agree more.

Prof. Scherer correctly illustrates how the “use of daily accounting data would have underestimated the Value at Risk of an equal-weighted portfolio of G7 equity stock markets almost all the time. The use of unadjusted daily data becomes most troubling in periods of market crisis where underestimated correlations suggest a diversification benefit that is not real.

Prof. Scherer goes on to point out that this problem can only be overcome with “data synchronization models.”

Axioma has been acutely aware of this issue since it first began producing risk models, as all of its risk models are fully updated on a daily basis using daily closing return data. To address this issue, Axioma released in May 2010 a proprietary “returns-timing” adjustment methodology that specifically accounts for asynchronous trading between markets.

Most data synchronization models estimate a vector auto-regressive moving average (VARMA) model of returns. Axioma’s synchronization model uses a simple, first order vector auto-regressive (VAR) model. This model relies on just one day's lagged data (the first day is by far the most significant) and provides a robust estimate of the synchronized returns. To date, Axioma has shied away from using moving averages and higher order models as these are more difficult to estimate, require more modeling assumptions, and, in our experience, provide less robust results.

Axioma’s “returns-timing” model allows a number of important variables to be more accurately estimated in our risk models. First and foremost, it corrects the correlation underestimation between assets that trade at different times. Second, the model corrects the specific returns of ADRs and similar instruments whose underlyings trade at different hours than the ADRs. Third, the model allows returns to be decomposed into local market returns and global market returns. The global market returns can also be easily decomposed into industry and sector returns. This allows users to quantitatively assess whether a section of a market, such as banks, moving on one day in one part of the world moved in that same section of the market in a different part of the world later on the same day or on the next day. Finally, the model ensures that factor returns are synchronized, which improves ex-post attribution analyses and helps portfolio managers to understand the factors underlying performance.

For more, download Axioma’s research paper Returns-Timing: A Solution to Asynchronicity.

Receive new blogs by e-mail and follow @FactSet on Twitter.

Vlog: How does FactSet account for interest rate moves in its Value at Risk forecast?

Steve Greiner — 2010-11-22T19:34:12Z

Notes from the road: FactSet Balanced Risk hits Asia

Andrew Kovacs — 2010-11-22T19:34:12Z

It's been a busy month for us in Asia and Australia as we've been traveling extensively in the region to introduce the short-term balanced risk model FactSet recently developed with our partner R-Squared.

We developed the model this spring as, over the past several years, it has become apparent that there is a need in the market for a more complete risk offering for balanced fund managers. In our seminars in Tokyo, Singapore, Hong Kong, Sydney, and Melbourne, we covered four topics:

Willett Bird and I kicked things off with a brief introduction to FactSet’s fixed income portfolio analytics. In particular our presentation focused on how we generate derived analytics for a wide variety of fixed income instruments: an essential part of the process which ultimately allows us to produce risk analytics for the fixed income portion of a portfolio.
Jason MacQueen from R-Squared then gave an engaging presentation that first summarized the evolution of risk modeling and the various approaches that have arisen over time. Second, he detailed how, in building the short term equity risk model employed in FactSet balanced risk, R-Squared was able to incorporate many of the best elements of risk modeling developed over the years within the industry.
During the third presentation, Steve Greiner of FactSet got to the heart of the matter in a presentation describing the essential elements of FactSet’s solution for balanced risk. Steve provided details about the methodology used to produce the underlying model, how the model is used to compute Monte Carlo VaR and related statistics, and examples to illustrate the stability and robustness of the model.
In the final presentation of the day, Jason provided some real case studies to remind everyone that risk is a vital part of every investment process, if for no other reason than it allows portfolio managers to focus on their strengths and eliminate unintended bets from their portfolio construction process.

The events were a great way for us to hear directly from investment professionals on their needs in this area, answer questions, and share ideas. We're planning several future events that will include details on FactSet’s balanced risk offering, including the Investment Process Symposium in Miami next March. I hope we will see you at one such event.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

The U.S. elections one week out: What's in store for equity and fixed income markets?

Steve Greiner and Bill McCoy — 2010-11-22T19:34:13Z

There’s been sea of change in control in the Congress of the U.S. in this last election. Many in the media attribute it to a disenchantment with the White House’s treatment of Wall Street and corporate America. Other media believe that Obama tread on traditional American values. Still others point to the bank bailouts and egregious quantitative easing, the Fed money printing, and government spending without limits to the takeover in Congress of the Republican majority.

Since November 2008, the VIX, the CBOE Market Volatility Index, has fallen from 70 to today’s benign level near 20. The S&P 500 has rebounded dramatically from its March 2009 lows but still trails by 300 points its early 2007 highs. Gold, as measured by the SPDR Gold Trust (GLD), has traversed a straight line over the last five years in a single direction: up. This last measure of GLD portrays a dangerous direction for the economy in spite of the lower VIX and attests to the fear the world has for the U.S. dollar going forward.

The defacto reserve currency the U.S. dollar has been affords the United States two special interests. First, since the world essentially must buy the dollar -- for most global trade is conducted in dollars -- it allows the U.S. to borrow at cheap rates, relative to other countries, since our debt is in demand. Secondly, the U.S. can print its way out of financial straits, thus it can inflate its way out of debt conveniently if it has too much. Argentina, Greece, Portugal, Latvia, Spain, and Ireland are forced to acquire more stringent budgetary considerations to manufacture a solution to their debt problems, but the U.S. (whose debt is 94% of GDP at this writing) doesn’t have to, all due to the U.S. dollar as reserve currency. Unfortunately, it doesn’t mean our economy is healthier than Greece, it just means we can prolong paying the piper. The citizens of the world recognize this fact and are short the dollar and long gold mostly for this reason, albeit it’s becoming a crowded trade.

Meanwhile, large cap U.S. stocks may have increasing profitability due to their increasing percentages of revenue coming from the advancing emerging markets. In addition, the changeover in government means that Congress will be controlled by the Republicans, while Obama rules from the White House, resulting in gridlock on developing any new government programs. These two situations combine to allow some of the global fear of U.S. future policy to subside. This all portends for small to moderate gains in U.S. markets over the next year, while our burgeoning national debt and large unfunded liabilities (e.g., social security, government pension liabilities, and welfare) continue to grow unabated, leaving the longer term prospects for growth in the U.S. quite dismal.

Steve Greiner

For the U.S. fixed income markets, Tuesday’s election had a modest impact compared to the more significant resumption of Quantitative Easing II (QE2) or the stronger than expected employment numbers. Near term, fiscal policy is expected to be gridlocked, as neither party has a clear majority or mandate to enact legislation. Given that backdrop, the resuscitation of the economy falls on the only player with any hope of maneuver: the Federal Reserve.

However, the Federal Reserve can only do so much. As Chairman Bernanke has indicated, Washington must do its share of the work for fiscal leadership. There are several problems that will require Washington’s attention, and somehow the two parties will have to work together to solve them. Most importantly is the looming crisis in state and local government funding. California, Illinois, and several municipalities are struggling to plug holes in their budgets. Next up are the remaining specifics of financial reform. In addition, Congress must decide on the disposition of Fannie Mae and Freddie Mac. Finally, there is reaction of other countries, as America tries to inflate its way out of the current crisis. Such measures can only be successful as long as everyone is not doing it at the same time.

Bill McCoy

Coming soon: Systematic single-name exposure analysis

Chris Ellis — 2010-11-06T05:35:52Z

I’ve been quiet on the blogging front lately, but I wanted to share with you some of what we’ve been working on in FactSet product development. Over the last two months, I’ve visited 20 of our largest portfolio analytics clients in the U.S. to gather feedback on FactSet’s upcoming single-name exposures application (due to be released before the end of 2010).

This new application brings together four key pieces of FactSet, the first of which is client portfolio holdings data. Over 700 investment managers and plan sponsors use FactSet for performance attribution, characteristics analysis, and ex-ante predictive risk. To facilitate portfolio analytics, those 700 clients are loading over a million portfolios onto our system every night, which means we have a lot of data at hand that is integrated from accounting systems, custodians, and prime brokers. We have more than 50 custodian and prime broker position feeds that come into the system every night.

Second, that portfolio data is combined with the entity data. You can combine ADRs and GDRs with their parents, combine A shares with B shares, equity with corporate debt, or credit default swaps and options. Those are four key security types that get to the heart of issuer exposure.

By having all of that rich entity data about parents and children and securities, FactSet users can look at exposure to an individual security and all related equity securities or all the securities related to the issuer.

The third piece is pulling in screening, which allows you identify the universe of companies using a set number of criteria. When you are doing exposure analysis, sometimes you don’t want to look at an individual security but you want to look at the securities that fit a certain theme. That could be securities that have been downgraded by Moody’s or S&P in the last six months that are in a particular sector and region.

The fourth key piece is that we have wired all of those results into the same reporting and charting tool that we use for attribution analysis, so users can slice and dice the data and turn it into reports and charts that make it easier to digest and act upon the information.

Integrating these pieces has been relatively straightforward, but the devil is in the details. For example, when you define a parent/child relationship between a structured product and the issuer of that structured product, when should the structured product make the security link to the bank that created the security and when is it just an issuer? That gets into the detail of where there are guarantees or links between the issuer and the bond holder. For the next month or so, we’ll work to tie up loose ends and add functionality, such how we integrate and update history of the data for time series historical trend analysis.

In my visits this summer and fall, I was startled by how many very sophisticated clients were performing this analysis in an incredibly manual, non-systematic way. Let me know in the comments: will this single-name exposures application change your approach, or have you perfected this process? What’s your biggest challenge when doing this on your own? What would you want to see in a FactSet application?

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Why Default Correlation Matters (Part 2)

David Mieczkowski — 2010-11-06T02:28:30Z

In part 1 of this series, I highlighted how important default correlation is to credit risk management. In particular, I showed through some simple examples how default correlation dominates the risk calculus. I provided a simple link between default correlation and asset correlation using a simple one factor model. While this might seem highly stylized, it is worth pointing out that this setup is essentially the mechanism that is used to quote CDO tranches in terms of base correlation.

In this post, I’ll continue the discussion by showing how to pull information from the equity markets in order to estimate asset covariance. The solution for corporate bonds is twofold. First, we will apply a firm value or structural model of default (e.g., Merton Model) to link equity return correlation to asset return correlation, and asset return correlation to default correlation. Second, we employ a factor model structure on the equity returns to dimension reduce the problem.

The original and simplest of the firm value models is the Merton model. In the Merton model, firm asset value is modeled as a geometric Brownian motion, and firm equity is modeled as a European call option on the value of the assets, where the barrier (strike) is the notional K on the outstanding debt. Because assets equal debt plus equity, owning the debt is equivalent to being long the assets and short a call option on the same. Although this model is highly simplified in its assumptions about a firm’s capital structure, the appeal of the model is that it provides a direct tractable link, in the form of the Black-Scholes option pricing, between the debt, the assets, and the equity. Mathematically, the Merton model provides us with two non-linear equations which we can attempt to solve for the asset value and the asset volatility.

Where the A is the value of the assets, D the notional on the debt, E the value of the equity, C is the value of the European call option, and the s’s represent the volatility of the assets and equity respectively. Once we solve the system, we have the pieces needed to determine the probability of default P, and then by adding a recovery assumption, we can approximate the bond spread by

Tunring to factor models, we recall that in a factor model, the goal is to express the N individual returns in terms of a linear combination of common factors as,

Where,

The main benefit of this is that the covariance/correlation structure is highly simplified. If the factos are orthogonal (and we can usually perform a little technical magic called "factor rotation" to make this so) then the variance and covariance structure is simply given by, With this more robust framework I can now show you the effect of a correlation shift on credit VaR in a real world setting. To see this framework in action, we just need a portfolio of credits, and an equity factor model. For the sample portfolio, I chose the Bank of America Merrill Lynch U.S. Corporates Large Cap/Industrials (5-10 Y) (MLCIL6) index, which has a weighted average rating of A3. For the equity factor model, I did a simple principal component analysis (PCA) on the daily returns for the prior 250 days and retained the top 10 principal components. I use PCA for the factor model in my example because the principal components are already orthogonal (uncorrelated), hence it gives me a simple way to alter the correlation structure while preserving the individual variances within a Monte Carlo based simulation by using equations (6) and (7). This will let me show the effect on the 99% VaR of a correlation shift for a real life portfolio, while at the same time not altering any of the marginal probabilities of default.

The scree plot for this is shown below. Effectively the first 10 PCs explain roughly 58% of the daily variation in returns, with the first component accounting for more than 40% of the variation alone. It is worth pausing for a moment to compare this to the simple one factor asset value model presented in part 1. The scree plot indicates that such a simple model may not be terribly inaccurate.

To use the details of how to use the above framework to generate a VaR number is as follows:

Given the observable equity value, equity volatility, and outstanding firm debt, calculate the asset value and asset volatility by solving equations (1) and (2).
Calibrate the debt level to force the spread relation (3) to match the observed starting bond spread.
With the equity factor model in hand, run a bunch of simulated returns of the common and idiosyncratic factors to generate a bunch of equity return scenarios. Assume debt is unchanged, and then calculate asset value changes as A(sim) = A(base) + E(sim) – E(base).
Hold the asset volatility constant and use the asset value changes to compute new probabilities of default, and then through the spread relation (3), compute a spread change.
Use the simulated spread changes to compute the portfolio weighted average returns (due solely to spread moves), and calculate VaR.

To see the effect of an equity return correlation change on my real life portfolio of credits, I did two tests. First, I perturbed the issuer equity factor loadings, while preserving the % of variation due to idiosyncratic risk. In general, this spread the systematic risk out across the factors more evenly, and as a result, lowered the average pair-wise equity return correlations. As expected, the VaR is a function of the average correlation. The results of this are below. Recall that the equity factor model is using daily returns, so the analysis effectively shows the impact on 1-day VaR on the BofA Merrill Lynch U.S. Corporates Large Cap / Industrials (5-10 Y) (MLCIL6) index portfolio, in basis points.

Comparing the trend line to the single factor model discussed in part 1, the impact of increasing the correlation from .3 to .4 is roughly 10 bps, which represents an increase of 25% to the VaR. This compares to a 20% increase in VaR when going from a correlation of 0.3 to 0.4, for the simple single factor model.

The second test was to shift variance from the idiosyncratic component to the systematic components in a relatively uniform way. The graph confirms that shifting risk from uncorrelated idiosyncratic components to the systematic factors increases average correlation and the VaR. Slightly different than in the stylized single factor model, the magnitude of the effect declines with rising correlation. This is mostly due to the fact that the idiosyncratic exposures are heterogeneous in our real life portfolio.

Finally, to emphasize the point that correlation dominates, I examined the effect of holding the correlation structure constant, but simply increasing the individual risk neutral probabilities of default (achieved by raising the equity volatilities). The base case weighted average probability of default was roughly 4%. The 95%-VaR level for the base case was 44bps. There are a few really interesting aspects of the graph to note. First, there is a clear pivot point in the risk at about 9%. In a real life portfolio, the linear increase in marginal risk of the stylized example does break down at some point.

Subscribe by e-mail to receive new Taking Risk posts as they are published, and follow @FactSet on Twitter.

Why default correlation matters (Part 1)

David Mieczkowski — 2010-11-06T02:28:30Z

In this post, I will focus on the underappreciated topic of default correlation, how it might be related to equity correlation, and what it means for the VaR of your corporate credit portfolio. Balanced fund managers take note! To begin, let's establish some terminology. Consider obligors A and B, and a time horizon T. Let's denote the probability of default of A before T by p_A and similarly the probability of default of B before T by p_B.

Now for a little pop quiz.

Question: With knowledge of p_A and p_B you can determine which of the following?

P_A|B The conditional probability that A defaults given B has defaulted
P_AB The joint probability that both A and B default
r_AB The linear correlation coefficient between the default indicator events I_A and I_B
None of the above

Unfortunately, knowledge of the marginal probabilities of default is insufficient to determine the conditional probabilities of default, the joint probability, or the linear correlation. So the answer is none of the above. Knowledge of the marginal distributions and the linear correlation of default is, however, enough to determine the conditional default probabilities, and the joint probability, of default. The relationships between these quantities are given by: Before discussing how one might go about estimating the linear correlation of default, let's just take a moment to see just how important this number is for credit risk. Let's suppose that the individual default probability for each obligor is 2%, so that p_A = p_B = 2%. Below we plot the joint probability of default, and the (identical) conditional probabilities of default. It is important to note that the linear default correlation completely dominates both the conditional and joint probabilities of default. The joint probability of default, for example, is roughly 10 times as large under a correlation of 20% as under 0%, and the magnitude of this effect increases as the individual probabilities decrease. This is deferent from our experience with correlation as it applies to equity return variance, where the marginal effect of a lower average correlation is constant on a relative basis. Thus the relative value of proper diversification increases as the as risk decreases for credit sensitive portfolios.

Given that historical analysis by the ratings agencies suggests that the five-year cummulative default rates of the majority of investment grade rated bonds is below 2%, this suggests that for a typical investment grade portfolio of credits, default correlation should dominate the risk calculus. In particular, the use of average credit rating by some bond funds to summarize credit risk is inappropriate, not because it hides a few lower quality bonds, which may skew the loss risk, but because it provides absolutely no information about the dominant risk: the default correlation.

To illustrate the impact of default correlation on the portfolio VaR, consider the case of an equally weighted portfolio of 100 obligors with identical independent individual probabilities of default of p = 5% over a horizon T and zero recovery. Neglecting the interest rate component, the VaR over the horizon of this portfolio is characterized simply by the number of defaults over that horizon. In this example, we can compute the probability of there being k or fewer defaults using the cumulative Binomial distribution functionBelow we display the 99% VaR for this stylized portfolio. As with equities, fully determining the joint probability of default would require us to estimate all the pair wise correlations within a portfolio. Even for a portfolio of 100 obligors, this would require the specification of 4,950 pair wise correlations, which would in turn require obtaining at a minimum 4,950 default events. This is simply not feasible. Even for the equity world, where there is ample return data, a factor structure is employed to make the correlation problem tractable.

The solution for corporate bonds is twofold. First, we employ a factor model structure on the equity returns to dimension reduce the problem. Second, we will apply a firm value or structural model of default (e.g., Merton Model) to link equity return correlation to asset return correlation and asset return correlation to default correlation.

To keep things simple for now, let's consider a one factor model of the firm assets directly (I will incorporate equity data in a later blog post). Specifically, let's assume that all assets are driven by a single, common factor with a standard normal distribution.

Because the idiosyncratic components are assumed to be uncorrelated, the covariance between the assets will be r. If we normalize the asset values first, then we can view r as an asset correlation. For the structural model part, we will assume a simple barrier model. Namely, an obligor defaults if its asset value at the horizon T falls below some critical level K.

If we assume that all obligors in an equally weighted portfolio have the same barrier K, then uncorrelated idiosyncratic components means that conditional on a realization of F, the probability of having m or fewer defaults in a N obligor portfolio is given by the Binomial cumulative distribution function (4) above. Since the common factor has standard normal distribution, the VaR is given by further integrating the Binomial above against the Gaussian density function. Below are the plots of the 99% VaR for a 100 obligor portfolio for different choices of r, where the barrier K is set so that zero correlation corresponds to a marginal probability of default of 2%.

We can immediately see two things from the figure. First, the VaR converges to the independent binomially distributed case as correlation goes to zero (compare to the prior graph). Second, a 2% individual probability of default with a 20% asset correlation has a 99% VaR of 14. We would get the same VaR by having a 7.5% individual probability of default with a 0% asset correlation. To put this in context, this is like saying that an average BBB rated portfolio would bear the same credit risk (as measured by 99% VaR) as a BB+ porftolio. Three guesses on which portfolio would have a higher yield.

Here, I highlighted the importance of default correlation and through a simple factor model, showed how asset correlation dominates the risk equation for a portfolio of investment grade credits. In my next post, I'll go into more detail as to how an equity factor model can be harnassed to estimate the credit risk, and how correlation "stress testing" can be leveraged by looking at some real world examples.

Subscribe by e-mail to receive new Taking Risk posts as they are published and follow @FactSet on Twitter.

Are there really any differences in the risk model providers? Asia-Pacific edition

Willett Bird — 2010-11-06T02:28:30Z

In the eight years I have been discussing risk with our clients in the Asia Pacific region, the dominant question I continue to receive is typically phrased: “Are there really any differences in the risk model providers?” Historically this has led to a discussion on the differences in each of the risk model provider’s methodologies. Last year, my colleague Chris Ellis put together a series of blog entries (part 1, part 2, part 3) answering that exact question, which I have since pointed my clients to for answers to their queries. However, inevitably, a couple of days later, I hear back from them with a rebuttal along the lines of: “For the U.S. that makes sense, but what about Asia?”

So I will build off of my colleague’s previous analysis and see if the same results hold for the Asia Pacific region. My analysis is going to be considerably simpler than what Chris put together given the relative size of the market in the Asia Pacific region. Style indices are few and far between; i.e., there are not many Asia Pacific ex-Japan mid-cap growth funds to speak of. For the purposes of this blog entry, I will focus my attention on the three major markets in the region: Australia, Asia ex-Japan, and Japan, and I will perform the same analysis previously conducted using the Northfield, APT, and Axioma risk models. (At their request, Barra will no longer be included in comparisons of the different vendor risk models.)

Using the FactSet LionShares database, I retrieved fund constituents for the 30 largest funds in each market as of June 30, 2010. My initial focus is tracking error (active risk) using the MSCI All Country Asia, S&P ASX 300, and TOPIX as the benchmarks for the three respective markets (Asia ex-Japan, Australia, and Japan). For the Asia Pacific region, I used the respective global models for each of the three providers, and for the Japan and Australia analysis, I used the country-specific models for the analysis. I will refer to the three models simply as A, B, and C. The purpose of this analysis is not to suggest which model is the “best,” as this analysis is not attempting to answer that question and may mislead the reader to conclusions about the “best” concept. I am trying to assess only whether there are differences between the providers.

Let’s first take a look at the average tracking error:

It’s fairly easy to visually see that Model C is suggesting a lower tracking error in all three markets with Model B portending the higher tracking error for all three markets. Model A seems to sit squarely in the middle, although leaning more towards Model C, especially in Australia. But the key question is, are the differences we see here statistically significant?

I used the same statistical technique used in the previous blog entry: the Welch’s T-test. I will deem any value greater than two to be statistically significant.

This test further bolsters the conclusions we were able to see fairly easily from reviewing the average tracking errors. While there is a discernable difference between Models A and C, we can see the differences are not significant. However, the significance between Models B and C is consistent for all three regions.

This was a very simple exercise; I would most likely want to include more funds and a more robust set of portfolios and benchmarks before making any firm conclusions, but we can observe from this brief analysis that two of the three risk models I used in this exercise do appear similar while the third appears to suggest a consistently higher tracking error. This builds upon my colleague’s previous analysis that there are indeed differences between the different risk model providers.

Please share your thoughts in the comments.

Subscribe by e-mail to receive new Taking Risk posts as they are published and follow @FactSet on Twitter.

How skillful a manager is Paul the Octopus?

Mike Joel and Matthew van der Weide — 2010-11-06T02:28:31Z

by guest bloggers Mike Joel, Quantitative Specialist, London, and Matthew Van Der Weide, Quantitative Specialist, Amsterdam

Now that the craze of the World Cup is over, vuvuzelas have silenced, and Paul the Octopus has retired from making predictions, it seems like the time to evaluate. Could it have been luck, or was Paul the Octopus really skilled, and, if so, where would that rank him against some of the greatest fund managers?

The first argument of sceptics is that there are potentially thousands, if not millions, of animals with assumed predictive skills and only those who predict correctly come to the surface. That does sound like a plausible story; had Paul been wrong in the group phase, surely we would not have heard as much of our squishy oracle. In a way this reminds us of the (e)mail scams that recommend a stock every week, building up a track record week after week, trying to persuade you to invest. Although these stories sound compelling if there are a number of "correct" predictions in a row, it cannot be ignored that these frauds simply stop writing the people for whom they made a wrong prediction. Something that Nassim Taleb has shown all to clear in his book Fooled by Randomness: in the end this is simply a numbers game. Eight out of eight predictions right and given 50/50 odds (we will address both assumptions later), that would mean that there is a 1 in 256 chance for a random animal in a German zoo becomes the next Oracle.

But that is not the complete story, Paul actually has a bit of a track record. Although it is doubtful that it improves his statistics, at least it shows he has been around and we can conduct some analysis. Back in 2008, he made predictions for the Euro 2008 Championship and got four out of six right then. Still assuming 50/50 odds, would you have invested with a manager that has Paul’s track record? A simple coin flip would have given you three out of six (assuming 50/50 odds, mutual exclusivity of events, and the caveat that you might need to quit your job to be able to repeat the experiment enough to obtain stable results).

So when we look at the total track record of the eight-legged oracle, he predicted 12 out of 14 matches correctly (again, of these mutually exclusive events!). Through the use of our friend, the factorial, we can calculate the probability of achieving this:

where x = number of occurrences and y = number of wins

Which ultimately translates to 91/(2^14)=0.0055, or 0.55% probability.

So what about those odds? As we pointed out earlier, they are assumed to be 50/50. Ever wondered why sometimes the bets at a bookmaker do not seem to add up? Effectively for the group phase there is a one in three chance; teams can draw and bookies can make money. Hence, given the extra choice, the odds of making the right prediction on 12 out of 14 matches plummets to 0.0019%, which is still slightly better than getting five out of six numbers (not including the bonus number) correct in the National Lottery (~0.0018%)= £1,500. Although when placing a bet with a bookie, one (usually) has detailed knowledge of team, hence there is more than luck in play and the probabilities are probably biased.

So how does Paul hold up against a good or even average fund manager? Grinold and Kahn define the Information Ratio (IR) as the Information Coefficient (IC) times the Square Root of Opportunity
(). If the IC is our measure of skill, we will assign this the probability of Paul correctly predicting 12 out of the 14 matches (for lack of a better measure). This gives us an IR of 12/14 * Sqrt(14) = 3.2, which is remarkable, although given the light track record, not reliable. To maintain the same IR going forward, Paul does not have to maintain his skill, but merely place more bets!

Now to compare him to a fund manager, how can we describe Paul the Octopus’ investment style? He is not your typical long-only equity investor; his style more resembles that of a hedge fund: long team A, short team B (given that one of the co-authors is Dutch we will not disclose the teams). If one has to put a label on what the World Cup is, "Event Driven" probably fits best. Looking at the information ratio over the last 14 months, for the Lipper Tass Event Driven hedge fund universe, where would that put Paul the Octopus? His performance is not too bad as he ranks comfortably amongst the best names in the top quartile.

This brings us back to the main question: is Paul’s ability the result of skill or luck (putting aside any bias that may influence the numbers, such as shape or colour of the flag, an inclination to favour Germany, etc)? Given Paul’s predictions and where he falls relative to other funds of his style, he is a highly skilled manager. But whether this is true skill, whether there is a correlation between the winning team and characteristics that attract Paul or whether this is pure luck, the jury is still out; there are just not enough data points to quantify skill vs. luck. Perhaps Paul’s situation is a case of survivorship bias to the extreme (also quite literal), but one thing is certain: the next time one peruses a marketing brochure, it may be worth contemplating Paul and his ability. Any number of calculations/statistics will make Paul look like one of best managers, and although the numbers may look good, one must delve into the details behind the numbers. The risks for managers involved in pursuing the stats and making it to a top quartile ranking, though, are minimal compared to being part of our next dish.

Subscribe by e-mail to receive new Taking Risk posts as they are published and follow @FactSet on Twitter.

Whose fault is it when the market crashes. . . and so does your portfolio?

Steve Greiner — 2010-11-06T02:28:31Z

Ah risk. Risk is something that’s fun to talk and write about. Part of this is due to risk being such a vague topic -- no matter what you say, you’re right, as long as you’re talking about the “bad” side of investing. It’s the wicked step-sister to Alpha. For those younger baby-boomers, or older generation-X participants, picture the robot on Lost in Space waving his arms frantically yelling, “Danger, danger Will Robinson!” This conveys what risk is: very alarming and scary, but so worthy of study. Calamity may never occur, but it could occur, and that’s what scares the “bejeebers” out of most of us.

When the credit crisis began in 2007, it carried with it a wave of criticism of quantitative models, especially risk models. Books like Nassim Taleb’s The Black Swan and Scott Patterson’s The Quants: How a New Breed of Math Whizzes Conquered Wall Street and Nearly Destroyed It would lay apart the 50 years of risk modeling predicated on risk being defined as volatility and volatility being measured by the standard deviation of return, all under the “Gaussian Approximation.” While some of what they may have written is true, their main complaint centers around using historical data and models created from it in diagnosing the current environment and for predicting the future. Ben Graham, Warren Buffet’s teacher on the other hand, said:

"It is true that one of the elements that distinguishes economics, finance and security analysis from other practical disciplines is the uncertain validity of past phenomena as a guide to the present and future. Yet we have no right to reflect on lessons of the past until we have at least studied and understood them.”

In addition, Graham said if experience cannot help today’s investor, then we must be logical and conclude that there is no such thing as investment in common stocks and that everyone interested in them should confess himself a speculator. The reason we like exploring the past has to do with the comfort we draw from our ability to use existing constructs and memorable ideas to explain historical developments, make current decisions, and estimate our future directions. That, of course, is why we may be late in recognizing significant shifts in the market in accordance with Nassim’s findings.

To that end, optimism and over-confidence have always accompanied bull markets and pessimism always bear markets. In practice, most of us find three-day weather forecasts useful. We buy flood insurance in areas where floods have historically occurred, and in fact we make decisions every day based on past experience. So the logic of taking the outcomes of the past and counting them up to form a distribution with which to make future decisions does have some precedent. The normal distribution (Gaussian) is not representative of a future distribution but is one for picking members out of a population and has no future values in it. Since the future hasn’t happened yet, there isn’t a full set of outcomes in a normal distribution created from past experience. When we use statistics like this to predict the future, we are assuming the future will look like today or similar to today, all things being equal, and also assuming extreme events don’t happen altering the normal occurrences. This is what quants do. They are clearly aware that extreme events happen, but one doesn’t throw away a useful model just because some random event can happen. I wear seatbelts because they will offer protection in the majority of car accidents, but if an extreme event happens, like a 40-ton tractor-trailer hitting me head on at 60 MPH, the seatbelt won’t offer much safety. Do we not wear seatbelts because of the odd chance of the tractor-trailer collision? Obviously we wear them.

To add color, however, and offer an apologetic for quants, consider that there are more than one “cause in effect” for almost all known observations of any phenomena in the universe! Scientists usually attempt to understand the strongest influencers of an outcome or event, not necessarily all of the influencers of an outcome. So in reality, multiple causes are in effect for every event, even those which are extreme or “Black Swan.” Extreme events have different mechanisms (one or more) that trigger cascades and feedbacks, while everyday normal events (those occurring that aren't extreme) have a separate cause. One only enters into the conundrum of explanation when one tries to link all observations, both from the world of extreme random events with normal events when in reality these are usually from separate causes. In behavioral finance literature, this falls under the subject of multiple-equilibria and has been demonstrated in markets where there is deviation from no-arbitrage opportunities for a considerable amount of time and is of course noticed where structural mispricings occur for longer periods than a single equilibria mechanism would allow. In this context, highly non-linear and chaotic market behavior occurs where small triggers induce cascades and contagion similar to the way very small changes in initial conditions brings out turbulence in fluid flow. The simple analogy is found in the game of "pickup sticks," where the players have to remove a stick from a randomly connected pile without it affecting all the others. Eventually, the interconnectedness of the associations between sticks results in an avalanche. Likewise, so behaves the market.

To those who would criticize the use of a normal curve to examine past data or as criticism for modeling regularly observed events, understanding the statistical interpretation of a time-series of past data is not related to the applicability of history to appear for explanation after a market meltdown. It is not very helpful after the fact to explain chapter and verse why what happened should have happened because it happened before, because while we judge life in reverse, we live it forward. Graham also purported and even had in his lecture notes the example of a stock reaching new highs, that after awhile it goes down to levels below previous highs and that one can take this example as a warrant for purchasing securities based on past or historical histories at lower prices than the current high. Hence, he himself didn’t dismiss the lessons of history but used models as in this example and even admonished the use of empirically determined methodologies when the data was of a regular variety -- that is, of events that happen frequently enough to obviate their inclusion as extreme events or Black Swans. So indeed Graham was an empiricist to some extent; however, it’s also clear that anomalous investing strategies, uncovered from analyzing the past market through empirical analysis, do not constitute a proof of enhanced return availability through these methods, nor does risk prediction based on normal approximations. However, the normal approximation is valid most of the time and predicts normal events quite well. Quants have the tools to apply in extreme event risk prediction, should the cause of extreme events be someday codified. But one cannot blame the quants for their failure to include the unpredictable in their everyday analysis. Modern risk models at FactSet, are predicated on sound economic principles, but if an asteroid hits the New York metropolis, even the famed Metropolis Monte Carlo algorithm wouldn’t have predicted it.

Extreme earth shattering market events happen, but by other mechanisms than standard risk statistics should be expected to predict. This is no error on their part, but an error of expectation of the user. I’ll wear the seatbelt and in the majority of car accidents I have with other cars under 40 MPH, I’ll be safe. But to expect that seatbelt to save me with a semi speeding head-on at 65 MPH, the failure is mine not to have prepared for that event some other way, other than relying totally on my seatbelt! So when the market runs headlong downward at 60 MPH, and the VIX shoots to 90, you better have devised some other strategy other than relying on and then later blaming your risk model or risk modelers.

To receive new posts by e-mail, subscribe to this blog.

Stock-picking set for a strong return, or is BP dirtying the waters for all?

Sean T. Carr — 2010-11-06T02:28:31Z

This blog is aimed at providing a commentary on risk, and while I have highlighted BP (BP-GB) in the title I am not looking to add to the existing plethora of articles already out there on blame, value, or even jingoism. Indeed I could easily have used Apple (AAPL-US) or Google (GOOG-US) to draw the eye instead. These companies are the very much in the news at the moment along with BP; Apple’s releases of the iPad and the latest iPhone incarnation have resulted in consumers queueing around the block, and this in turn has been reflected in the stock price, analysts estimates, etc. Google, on the other hand, continues to fight the good fight in China, although looking at its price movement against that of Baidu (BIDU-US), I can only conclude that it is losing.

I was therefore moved to consider whether this was a signal of an increase in stock specific risk and a move away from the systematic/beta/allocation drivers that we have become used to. Managers would be able to add more value through their stock selection and we could link to this article with confidence that active management was set to step back into the spotlight. To investigate I looked to generate the average pairwise correlation of stocks, hoping to see the correlations of companies within industries/sectors decreasing as the unsystematic effects grew. What I actually saw in the Energy Sector was therefore a little surprising: the steady decline is correlations from the peaks of late 2008 to a bottom in April before a sharp rise again.

This rise implies that the sector is increasingly moving together, potentially understandable from a stock specific point of view where BP is a company with many exposures and many dependencies. While not discounting this, indeed with the final "bill" not yet fixed, as well as reports of BP becoming a potential takeover target, it is a company with a very specific set of risks. I suggest that it is the potential changes to overall regulation regarding drilling, both in the U.S. and other regions, that would affect all competitors as well as BP that makes this a systematic risk, and it is therefore this risk that explains the rise in correlations. This is only one case, and unfortunately negative in nature, but I believe that it does highlight how a specific risk can become systematic through contagion.

To receive new posts by e-mail, subscribe to this blog.

Hedging Against Deflation

David Mieczkowski — 2010-06-29T15:23:20Z

In early June, Thomas Hoenig, President of the Kansas Fed, gave a speech in which he advocated his view that, due to the strength of the recovery, the Fed should begin raising rates and even begin to shrink its balance sheet by beginning to sell off some of its mortgage portfolio. The man who outranks him, Fed Chairman Ben Bernanke, seems intent on making sure that the economy has sufficient momentum going before attempting to “pop the clutch” and see if the growth engine can turn over on its own power. With the upcoming meeting of the Federal Open Market Committee (FOMC), the question is not whether the Fed will stay pat with respect to rates (Fed futures have a 100% implied probability they will), but whether the Fed will lift its “extended period” phraseology. While we can debate whether the threat of inflation really is on the horizon, it makes sense to understand the impact of the more near term risk of deflation.

What the Market Thinks

One way to see what the market thinks about inflation risk is to take a look at the break-even inflation rate needed to equate the real yield on TIPs with the nominal yield on maturity matched non-inflation protected treasuries.

The break-even inflation can be viewed as the extra nominal yield demanded by non-inflation protected investors for bearing inflation risk, and so it is necessarily higher than the actual expected inflation rate. If we assume this extra premium is 50 bps, then the market is implying that it expects inflation to stay below 1% for a little over the next four years (spot CPI will be a little higher than the average period CPI). Thus, the market expects very tame inflation for an extended period.

The Case for Deflation

In a previous commentary, Black Swans and Money Helicopters: Staying Ahead in a Nonlinear World, my colleague Daniel Stachkov laid out the monetarist case for a growing threat of inflation as a result of the dramatic monetary response of the Federal Reserve Bank to the credit crunch. In response, I recently wrote a commentary, Inflation and Credit Money: Why Money Helicopters Don’t Matter, which takes the opposing viewpoint, that deflation is the greater risk, and discusses in detail the mechanics of why.

In a nutshell, I point out that it’s the demand for credit that determines the supply of money, not the Fed, and that it is therefore the total demand for dollar denominated debt that we should be focused on. I show that, even with the extra government borrowing and Fed stimulus, total U.S. dollar denominated debt is in the early stages of a protracted contraction, being more than $3.5 trillion below its trend line. A large portion of this total deleveraging is due to mortgage and housing related distress. I also claim that the shadow inventory of seriously delinquent, but not yet foreclosed homes, in combination with a possibly rising mortgage rate, is likely to lead to a double dip in home prices, on the order of a further 20% decline. I conclude by discussing the parallels between the U.S. and Japan.

Let’s presume the thesis I laid out in my commentary is correct and look at the impact to some asset classes.

Assessing the Risk to Mortgages and Banks

Since housing tipped off the credit crunch, it makes sense to start with looking at the impact on existing mortgage debt due to a further 20% decline in house prices. To do this analysis, I ran a representative portfolio of 2000 bonds taken from non-agency issued CMOs (representing roughly $600 billion in current face value). I used FactSet’s Fixed Income Manager software, and the LPS Applied Analytics prepay and loss model, to run the analytics.

This analysis indicates that investors in the non-agency CMO universe could lose an additional 20% on the existing notional outstanding. Total bank exposure to the mortgage market is roughly $4.9 trillion ($3.3 trillion in first liens held in bank portfolios in non-securitized form, and $1.6 trillion in securitized private-label form). A 20% loss on this would amount to $980 billion. Since we already know how that type of scenario will impact stocks, let's consider the other side of the balance sheet.

Interest Rates and Fixed Income Portfolios

Short term rates are already at zero. To see how low the yield on long term rates could go, we turn to the TIPS market yet again. If we enter a double dip recession, and the market capitulates to the idea of a protracted period of zero inflation, the nominal yield on long term treasuries would approach the nominal yield on comparable TIPs. With the yield on the 10 year TIP yielding 2%, we get a sense for just how low rates on the 10 year Treasury could realistically go (2% is not that far above what the yield on the Japanese 10 year has been trading at, incidentally).

To test the impact on fixed income portfolios, I ran a static analysis on the Barclay’s U.S. Aggregate index by assuming that the nominal yield on Treasuries would approach the current yield on comparable TIPS.

Holding spread constant, a falling interest rate deflationary environment leads to positive price return across all the fixed income classes, with a weighted average return of 10% for the index as a whole. Yields, being inversely related to prices, decline. The point of this quick analysis is to remind us that during a debt deflationary cycle, asset values tend to fall, while existing debt, after hedging default (or spread) risk, will tend to rise. Long-term Treasuries, in particular, seem an appealing deflation hedge.

Read the full Inflation and Credit Money: Why Money Helicopters Don’t Matter ebook.

Is it a small world after all?

Andrew Kovacs — 2010-11-06T02:28:31Z

In the recent post, Risk model providers respond. Part 1, do regional models still matter in the global economy?, we heard from some of our risk partner vendors. The general sentiment seemed to be that regional/local models are still relevant in our ever more interrelated world. The argument used to support this idea was that different investment strategies will look at the risk-return trade-off in different ways depending on the markets and securities they invest in; consequently a variety of models are still necessary to reflect the overall investment philosophy of a strategy.

This post inspired me to look at two different portfolio strategies for two different markets and compare how global models would appear compared with specialized regional/local models.

The first scenario was simple in that I took the Lipper Active Index Lipper European Region to represent a portfolio vs the MSCI Europe Index. Ostensibly the portfolio is relatively similar to the benchmark; however, there are number of small “bets” evident based on sector and country allocations as well as a number of out of benchmark security bets (though nothing drastic). The following charts both show the actual tracking error of the portfolio vs the benchmark along with the ex-ante TEs estimated by both Global and European risk models, respectively, from three of the risk vendors available on FactSet. The period is from December 31, 2007 to May 29, 2009. The historical tracking errors are based on 12 months of returns and are forced to coincide with the 12 month estimation period used by the model provider’s risk models.

The results don’t seem to be dramatically different, but we can see that the spread between the estimated tracking errors and the actual tracking error is definitely smaller on the second chart suggesting that the European models did a better job in estimating overall risk than the global models for this regional portfolio.

I ran a similar test using a hypothetical portfolio of my own design in a single country market. I created a long only Australian Equity Portfolio with a strong mid-cap size bias. I then compared the results, as above, using the same three global models and the corresponding country models. Here are the top line results.

The results are similar again. The spread between the actual and estimated TEs is smaller when using the Australian models when compared to the global models.

So far so good.

I wanted to go one step further and break down the total risk into the security-specific and factor components to see where the risk was coming from according to the global and regional/local models. The results are below:

With the exception of Model #2, the split between Factor and Security Specific Risk in the case of the Global vs European models does not appear to be significantly different.

Here are the Australian results:

Here we can see that all three Australian models ascribe a higher portion of risk to factor risk (overall) than the corresponding global model does, which is pretty intuitive.

These are obviously very simple tests, but I do think that they more or less confirm the comments of the risk model providers in that Local or Regional models do seem to do a better job of estimating risk for these specific scenarios. I think that we can also see that in the case of my European example the global model results were pretty similar to the European model results in terms of the factor/security specific breakdown. In the case of the Australian scenario; however, the global models consistently ascribed a higher portion of risk to the security specific component and therefore did not work as well is describing the actual risk of my portfolio based on my investment philosophy.

Based on my simple analysis I am going to conclude that the risk providers are right; you need different models for different situations. There is no one model fits all solution.

Since this is far from exhaustive analysis I would be interested in hearing about research into this area or even more anecdotal evidence to support the case for specialty risk models. Please share your thoughts in the comments.

To receive new posts by e-mail, subscribe to this blog.

APT, Axioma, Barra, Northfield, and R-Squared respond to six key questions

FactSet Risk — 2010-06-23T18:16:26Z

Throughout May and June, FactSet Risk has been featuring conversations with five major risk model providers. These six conversations highlight key management distinctions and help to define the differences among the companies that help you manage risk.

Find out which providers favor measures of risk such as Monte Carlo VaR, learn how they define their competitive advantages, and much more.

Request the free eBook containing all the responses from Barra, APT, Axioma, Northfield, and R-Squared.

Listen to podcasts from the series:

1: If we've gone global, how should models adapt?
2: What has the market crisis done to risk modeling?
3: Looking to the horizon length
4: Can you do it all with just one model?
5: When measuring extreme losses, which risk measures work?
6: What's your competitive advantage?

To receive new posts by e-mail, subscribe to this blog.

What's more important: Country or Sector decisions?

Bryan Hoefs — 2010-06-23T18:16:26Z

Speaking with clients recently, the topic has come up as to what has been driving performance: country or sector decisions. There are two essentially sides of the coin. One side says the world is getting smaller, and companies of all sizes are acting more and more globally, producing and selling goods and services across the globe. Global factors such as commodity prices affect companies in a given sector in a similar fashion, regardless of where the company is located and conducts its business. The country is then less of a factor than the sector it operates in. If this is the case, than allocating sectors should have a larger impact on relative returns than allocating countries.

The flip side says country is the driving factor. With the problems and risk of sovereign debt default increasing in various countries throughout the world, country-specific factors are becoming more important. Companies operating in one of the PIIGS countries have a very different operating environment than companies operating in countries with out these debt problems, say Germany, Norway, or China.

Both have sound arguments. So which is it? From a risk perspective, the answer is a resounding both.

For this analysis, I’m using APT and the Lipper Active Indices. I’m using APT to narrow down to Country and Sector factors, ignoring other economic and style factors for now.

Looking at the Lipper Active Global portfolio vs the MSCI World, the contribution to TE from Country factors has been decreasing, and the contribution from sector factors has been generally increasing. So the first argument plays out.

Looking at the Lipper Active Europe portfolio vs the MSCI Europe, the opposite is happening. Country factors have been playing a more important role than sector over the last few years.

Are portfolio managers taking advantage of any potential shift? One way to look at this would be to compare the allocation and selection effects in a performance attribution report grouped by country with one grouped by sector. If allocation is higher in, say, the country report vs the sector report, we can assume country decisions are driving performance more than sector decisions.

Using the Lipper Active Indices, there does appear to be a trend that country decisions are becoming more important with European managers. The first chart is showing the annual attribution effects for the Lipper Europe Active Index vs the MSCI Europe, when grouped by Sector. The light blue bar is the allocation effect, which doesn’t appear to have a material impact on the relative performance of the portfolio.

Turning to country, you will see that the allocation effect has been increasing the last few years, indicating that managers are increasing taking advantage of country allocation decisions. The last two years contributing positively to relative performance.

There is less of a consistent story looking at the average global manager. In 2008, sector allocation was a large driver of relative performance, with managers taking advantage of the poorly performing financial sector. However, in 2009 sector allocations didn’t have a big impact. Perhaps the market uncertainty is causing managers to take their foot off the gas in regards to sector bets.

The world will continue to become smaller, with companies operating globally and global factors becoming more and more important. But the sovereign debt problems we’re currently seeing don’t show any signs of letting up either. Whether sector or country allocation decisions are more important depends a lot on the region you’re investing in, (e.g., Global vs European). This quick and dirty analysis reaffirms that both sides of the coin are right.

To receive new posts by e-mail, subscribe to this blog.

Live webcast: What is the true cost of your portfolio analytics service?

Risk News — 2010-06-23T18:16:26Z

Join FactSet on May 26 to uncover the true costs of your portfolio analytics service.

Wednesday, May 26, 2010
4:00 p.m. GMT/11:00 a.m. EDT/8:00 a.m. PDT

Register here.

Whether you’re using internal analytics systems or the “free” portfolio analytics offered by other providers, the cost can be more than you think.

Time: Finding fast answers to ever-changing questions, producing month-end reports, managing a mountain of data, and maintaining homegrown applications keep you from focusing on what counts. FactSet helps you cut through information quickly to get clear and concise answers.
Insight: Without fully interactive attribution, time series, and returns-based analysis, are you getting a complete picture of your relative performance? Only FactSet tailors the analysis and reporting to mirror your investment process, giving you a 360-degree portfolio review.
Accuracy: Cut and paste reporting and pre-defined grouping options can threaten the accuracy of your reports and analysis. FactSet offers the perfect balance of automated reporting and ad-hoc analytics, cohesively built on a single platform with full auditability and control of calculations. Our data is continuously tested and held to a margin of error of 0.005%.
Support: Does your service offer a dedicated support team that understands your processes and is available to answer any question, anytime you need it? FactSet’s dedicated account support, product, and industry experts serve as an extension of your investment team, giving you access to live answers 24-hours a day.

Space is limited and available on a first come, first served basis.

To receive new posts by e-mail, subscribe to this blog.

Risk model providers respond. Part 1, do regional models still matter in the global economy?

Risk News — 2010-06-23T18:16:26Z

The FactSet risk team is often asked to describe the differences among the various risk model providers. In the past, we have drawn our own conclusions in the eBook "Really, how different are the risk model providers?" But this time, we posed the question to key representatives from APT, Axioma, R-Squared, MSCI Barra, and Northfield, asking each provider to define their thoughts on modeling, the market, and risk factors.

In the first of our six-part podcast series, we ask the providers to comment on the usefulness of global models versus regional and country models. Listen now.

Next week, we’ll feature another round with the providers as they tell us how they’ve adapted their models after the market crisis.

Request the related eBook, "Really, how different are the risk model providers?"

Subscribe to receive new blogs by e-mail as they are posted.

The Abacus Deal: Could investors have seen what was coming?

Risk News — 2010-06-23T18:16:27Z

Editor’s Note: As investors know, the complex security called Abacus 2007-AC1 (referred to here as the Abacus deal) was discussed at length during the SEC case against Goldman Sachs. Guest blogger Doug Carey, FactSet Financial Engineer, spoke on the FactSet podcast this week about the underlying collateral of the deal, why investors may have overlooked the poor loan quality, and how analysis can point to diminishing loan health. He also offered additional commentary about the structure of a variety of loan types as they changed throughout the crisis. Below, he discusses what he learned in his analysis of an example loan underlying the Abacus deal.

By now most investors, and even many people who aren’t professional investors, are familiar with the now infamous ABACUS 2007-AC1 deal. This deal is a synthetic collateralized debt obligation (CDO) backed by several tranches from subprime mortgage deals, most of which were issued in 2006. The story ended with several money managers and hedge funds losing over $1 billion as the mortgages backing this deal, and mortgage loans in general, began defaulting at record rates in 2007.

We now know most of the facts about what happened with this CDO deal, who lost money, and who gained. But in terms of risk analysis it can be enlightening to step back in time to February of 2007 when this deal was being actively marketed. Were there signs that the loans ultimately backing this CDO were already in trouble? What data could investors have looked at using FactSet software to warn them that these loans, and therefore the long side of the entire CDO, would run into serious trouble?

The Abacus CDO is backed by multiple tranches, which in turn are backed by thousands of loans each, so it might seem like a daunting task to begin analyzing the credit risk of the entire deal itself. The only way to really begin the risk analysis is to look at each of these underlying tranches in turn and dig into their credit characteristics one by one. We won’t look at each of these deals here, but instead will select one subprime deal, ACE06HE4, from the Abacus CDO that is representative of the typical deal backing Abacus. Let’s look at some credit fields from ACE06HE4 that might have been useful to investors in February of 2007. Using FactSet, I examined the loans backing a deal such as Abacus and the loans' geography.

The loans we are analyzing here were only eight months old, on average, in February, 2007. Yet 30+ day delinquencies are already over 11% and over 5% of the loans have already been foreclosed on. Most subprime mortgage deals were structured to supposedly handle about 15% total foreclosures and this one is 1/3 of the way there after only eight months! These loans were off to a bad start.

Another approach I took when analyzing the credit of each individual deal in the Abacus CDO was to look at the credit-worthiness of the loans as well as the borrowers’ ability and willingness to pay the loan back.

Now let’s look at these metrics compared to ACE03FM1, a deal sold in 2003 by the same issuer.

One of the best predictors of defaults is the Loan to Value (LTV) ratio. If borrowers have less invested in their home, they are more likely to walk away if prices drop. In addition to this, the higher the LTV, the less money the lender will get out of it in a foreclosure, which increases loan losses that much more. The LTV for loans in ACE deals increased by a substantial amount from 2003 to 2006. If you wanted to dig even deeper, you can actually run First Loss CDR on loans like these in FactSet to determine the exact CDR necessary before the bond in question begins to lose money.

In conjunction with the higher LTVs, the average loan balance went up by 60%. This move to higher loan balances shows that issuers were taking more risk on each individual loan while decreasing diversification in each deal overall.

Interest only loans became much more prevalent in 2006 for most mortgage issuers. The number of interest only loans in the ACE 2003 deal was only 1%, but shot up to 22% in the 2006 issuance. In addition, the number of hybrid ARM loans increased from 58% to 85%. For both the interest only loans and the hybrid ARM loans, higher fixed interest rates kicked in just as refinancing became much more difficult for these borrowers. This led to more borrowers defaulting on their loans as they could not afford the higher monthly payments.

All of these credit fields show fairly alarming changes, but increases in risk like this were not isolated to just this issuer. It was happening across the board in the mortgage world and there are multiple examples like this deal that would show the same increases in risk and decreases in loan quality. The information was there to see, but for multiple reasons that have been debated countless times for the past three years, many investors either did not see it or chose to overlook it.

The moral to this story is analysis begins with the loans backing the deal that is under consideration. It is not enough to rely on agency ratings and marketing materials. It is always important to have access to the credit data of the loans that will eventually determine whether an investor will make money or lose money. By combining analysis of credit characteristics of the loans alongside various loss scenarios for the security itself, you can get a picture of the health of the overall security, and hopefully avoid the pitfalls so many fell into.

Hear Doug's podcast on FactSet Insider.

To receive new posts by e-mail, subscribe to this blog.

Reverse Stress Testing: What are the chances.....

Sean T. Carr — 2010-06-23T18:16:27Z

Much has been written in this blog regarding stress testing and how it should be included as part of a complete risk analysis process in parallel with more standard exposure, tracking error, contribution analysis, etc. There has also been an amount of recognition that an increased effort is required in order to spread the understanding, and therefore responsibility for, risk exposure across areas of a firm outside merely the risk and performance teams.

This week's blog therefore aims to shed a little light on an area that sits alongside standard stress testing and potentially services that educational need: the concept of "reverse stress testing." Reverse stress testing is based on exactly the same principles as standard stress testing, but by approaching the analysis from a different, and some might say more intuitive, angle, conclusions can be drawn as to the risk exposure of a portfolio.

In standard stress testing, you consider an extreme event in terms of a shock to a valid market data series (e.g., S&P Financials drop 18%), and then use a model such as the conditional multivariate normal distribution to measure the impact that event would have on an existing portfolio. In reverse stress testing, you instead set limits on those impacts (for reasons such as mandate, internal compliance, etc.), model the implied shocks necessary to break those limits, and then consider the potential for that shock to actually happen. This process can leverage internal expertise in a manner that might not be considered through standard risk calculation.

Consider an example: The manager of the ABC Global Energy and Resources Fund has an imposed absolute monthly loss trigger of -10%, which might lead to a reallocation of mandate. He therefore looks to generate some reverse stress tests modelling impacts to several relevant measures generating the following scenarios:

The table shows three modelled events, the impacts of which be a 10% loss to the portfolio, alongside the standard deviation and implied probability of the event. The event that immediately stands out is the potential fall in the oil price of 20%, and, when we consider that these events are modelled over a one-month period, I believe the current climate allows a rapid disregarding of it. At the time of writing, the impact of the Greece de-rating is only just starting to feed through, and therefore the MSCI World event seems a lot more likely and is therefore the stress test I would allocate resources to in terms of exposure analysis.

This is only a simple example, but I hope that I have demonstrated that although the mechanics of the calculations are the same when it comes to stress testing, it is the analysis of the "reversed" tests that adds some extra colour to the above risk analysis and underlines how it is possible to further leverage indirect risk knowledge and experience and knowledge.

To receive new posts by e-mail, subscribe to this blog.

Client webcast: Five Efficient Ideas for Portfolio Analysis

Risk News — 2010-06-23T18:16:27Z

FactSet clients are invited to join us for this live webcast, where we'll introduce the latest time-saving features of FactSet's portfolio tools and take your questions.

Registration is available for FactSet clients only. Space is limited.

Wednesday, April 28
Europe/Middle East: 12:00 p.m. GMT
North America: 2:00 p.m. EDT/11:00 a.m. PDT

Work even more efficiently in FactSet's portfolio applications with the five fresh ideas presented in this webcast:

Portfolio Spectrum: Incorporate multiple Portfolio Analysis reports and charts into a single view through our new Spectrum functionality. Launch reports and charts into multiple panes within a FactSet workspace. Create a custom view to simultaneously bring in the most relevant items for your analysis.

Report Schemes: Use predefined or customized report schemes to format the colors, fonts, and line separators of the background and data within your reports.

Audit Data and Column and Group Definitions: Increase data transparency and your understanding of summary information through the audit feature. Examine the underlying components of time series, group level, or aggregate data items and columns. Quickly view column calculations and grouping definitions directly within your workspace

Custom Workspaces: Create, organize, and save the most relevant report and chart views for your analysis and reporting. Highlight your most frequently used reports and charts. Save either reports or charts as defaults..

Interactive Charting: Enhance your knowledge of what's driving performance through custom charts.

To receive new posts by e-mail, subscribe to this blog.

The hidden risks of Chinese ETFs

Willett Bird — 2010-06-23T18:16:27Z

If you subscribe to the CFA Institute Financial Newsbrief (as originally recommended in a great post on financial and risk management journals by my colleague last year) and read it every day like I do, you cannot help but notice the Financial Products section included in nearly every newsletter has been ablaze with new ETF products introduced to the marketplace. Of late, hardly a week goes by without a new family of ETFs being launched focusing on China.

While the ETF market as a whole has grown to over U.S. $1 trillion in assets at the end of 2009, up from U.S. $711 billion in December 2008 (Source: Blackrock), the ETF market is still in its fledgling stages in China and has just U.S. $6 billion in assets. There is a very good reason for this.

Most of us are aware of the inherent risks of investing in ETFs; some are obvious, some less so, but I think most investors would agree on at least the following:

Liquidity: Not as much an issue for large equity ETFs (e.g., SPDR is U.S. $80 billion and the bid-ask spread is only a penny), but when the basket of securities held in an ETF are illiquid, differences between the NAV of the ETF and the value of the underlying securities can be significant. The Wall Street Journal last month highlighted this exact risk for bond ETFs.

Leverage: Another post from my colleague on leveraged and inverse ETFs explains that because of the increased volatility of these levered ETFs, the return is not always what you might expect.

Counterparty: Many investors assume that an ETF actually holds the underlying securities, but this isn’t always the case since it might not be practical, economical, or legal. As such, often ETFs will engage in buying and selling derivatives contracts in order to mimic the desired basket of securities. As we painfully discovered with MBS over the past few years, every derivatives contract requires a counterparty to deliver on the contract. Delivery does not always happen.

However, in China, there are added risks. Many investors in China perceive Chinese ETFs to be far less risky than investing in individual stocks because an investor avoids the unique risks of investing in individual Chinese stocks that are far more significant than in the developed world (overwhelming bureaucracy, industry regulators, corporate governance, ownership structure, auditor experience, finance team experience, insider trading). Incidentally, I am not suggesting these issues don’t exist elsewhere, but in China, they are certainly more prevalent and of much bigger concern to investors in individual securities.

With these risks, it would seem that investing in ETFs would be a far less risky proposition for an investor, but consider the following:

Risk Management: No better illustrated than the recent embarrassment suffered by Bank of Communications Schroder, a five-year joint venture, with their Shanghai Composite 180 Corporate Governance ETF. As first reported by Asian Investor and expanded on by ETF Daily News, a simple typo, which could have resulted in a multi-billion Yuan loss, was instead resolved for a far smaller claim. The very fact that the firm’s risk and compliance systems did not catch the problem is troubling.

Currency: For the better part of 10 years, China maintained a peg to the U.S. dollar, but on July 21, 2005, the government changed course, allowing a jump of 2% in the renminbi against the dollar overnight and a gradual strengthening of the renminbi over the following three years. However, due to the economic crisis, China decided to re-enact the peg in mid-2008. As a result of U.S. Treasury Secretary Timothy Geithner’s trip to Hong Kong and Beijing last week, this issue is back front and center. As reported in The New York Times, China seems ready to make a major policy change with regards to its currency in the coming days and the strong possibility that it will again allow its currency to rise versus the dollar.
Based on this information, we would surmise with the expected currency correction in the coming days (3-6% based on industry expectations) that there would be a jump in the WisdomTree Dreyfus Chinese Yuan Fund (largest USD-CNY currency ETF), and yet, despite the huge volume on April 7 (3.4m shares, 10X the 60-day moving average), this has yet to happen. It could be argued that this is a riskless opportunity, since it is highly unlikely that the Chinese government would raise the peg, as the political fallout would be felt from Beijing to Washington, D.C. So why hasn’t this ETF increased on expectations? What do others know that we do not?
Local Insights: Last summer, I was speaking with my father about how absurdly over-valued the Chinese stock market was (in my very humble opinion) after a recent run up and that it was a shame that there was no way to short the Chinese market.
I didn’t think much about my comment, but a couple of days later, my father sent me an e-mail indicating that I was wrong and that you could invest in FXP to short the mainland stock market.

And therein lays the danger of not doing your due diligence and/or not knowing the local landscape of a market. In name alone, FXP (ProShares UltraShort FTSE/Xinhua China 25) would definitely seem like good vehicles for shorting the Chinese stock market.

But my father was unaware of three important things about this ETF and short selling in China:

1. Up until last month, short selling in China was strictly prohibited. There was simply no way of doing it. But as of last month, short selling Chinese stocks is now just very difficult. There are a limited number of brokers who are permitted to do engage in this activity and there are few securities that they are allowed to short sell.

The whole short selling concept is still very new to the Chinese market and the regulators are relaxing the regulations extremely slowly in an attempt to control the introduction of short selling of the market. This relaxation of short selling rules is in large part due to the introduction of index futures to the market place which is scheduled to commence on April 16.

So while it would be incorrect for me to say you cannot short sell Chinese stocks, it is still difficult enough to do that it would be very hard for someone to maintain an ETF that is short any segment of the Chinese market.

2. So then how does FXP short stocks? The FTSE/Xinhua China 25 is a benchmark of Chinese stocks listed in Hong Kong. "Listed" is the key word, as there is a big difference. Historically, mainland companies listed in Hong Kong trade at far reduced valuations than their sister shares trading on the Shanghai or Shenzhen stock exchanges (The current P/E of the Hang Seng is 17; the P/E for the Shanghai A Shares is 29). As such, the returns of the mainland and Hong Kong listed stocks do not always move in lockstep together.

3. As the name indicates, the FTSE/Xinhua China 25 is only 25 stocks. So not only is ETF shorting Hong Kong listed stocks, it’s not shorting very many of them.

I am not suggesting that there is no correlation between the FTSE/Xinhua China 25 and the CSI 300 (most widely used benchmark in China); they are highly correlated. But consider that the FTSE/Xinhua China 25 is up 2% year to date while the CSI 300 is down 5% over the same time period (as of April 15, 2010), so there can be noticeable differences over shorter time frames.

I fired back the above explanation to my father. I think my old man is happy he did not invest in this security:
In conclusion, the question needs to be asked: Are investors being misled into investing in Chinese ETFs under the guise that they are a less risky way to gain access to a market when the reality might be that they are in fact far riskier than owning the individual stocks?

To receive new posts by e-mail, subscribe to this blog.

Follow up: The difference a daily risk model makes

Samuel Choo — 2010-06-23T18:16:27Z

From my last post, a commenter posed the question:

"The benefit for steplike intra month risk increase is nice, but what about a single day peak - getting nervous ... Is it measuring or forecasting risk numbers?"

This is a very valid concern and worth addressing. I took a look at the days prior to and after a jump in stock specific risk for a number of companies. I generated these charts to show what happened to the levels of risk after the change.
The bars in yellow is the first day of the "spike" in risk. As you can see, risk levels maintain the new higher level of risk, although we must acknowledge there is some amount of correction in the days following (take BBBY for example). I went through my entire testing universe and could not find a "single day peak" where there is an increase in risk followed by an similar magnitude decrease in risk.

Thank you for your question. Please leave a comment if you have further questions about this or any other blog entry.

To receive new posts by e-mail, subscribe to this blog.

The difference a daily risk model makes

Samuel Choo — 2010-06-23T18:16:27Z

Axioma recently became the fifth risk model provider on the FactSet platform. You may be thinking, "Sigh, another risk model to choose from?" Although the risk model landscape has been expanding with refinement of factors and model construction, I found that Axioma provided a distinctive look at equity risk modeling.

The feature of these risk models that drew my attention was that they are calibrated daily. Yes, that's right — daily calculation of factor returns, covariances, exposures and residual risks. Much of the model specifics are outside the scope of this blog, but I did want to take a simple look at what a daily risk model can do for a risk practitioner.

Using FactSet's Alpha Testing, I generated a daily backtest of the S&P 500 for the past year, looking for a company that had a drastic change in residual risk. I was able to identify Airgas (ARG) as a company that had a 196% increase in specific risk from February 4-5, 2010. In the chart below, I show the stock specific risk in % annualized volatility (SSR, or residual risk) along with the price of ARG. We clearly see the jump in SSR on February 5.
I include in the chart, the SSR for ARG using each of the Axioma U.S. models: Short Term Fundamental, Short Term Statistical, Medium Term Fundamental and Medium Term Statistical. I also overlay the price to show how the market reacted during this same period. The chart shows the quick reaction to the news by all risk models, with the short term models having the largest spike. Although the jump may be seen as an overreaction, we do notice the levels of risk staying steady over the rest of the month. To address some concerns of a very “jumpy” risk model, Axioma incorporates a Dynamic Volatility Adjustment and appropriate autocorrelation adjustments to produce relatively steady risk figures. (See these research reports for more information.)

ARG provides a great example of the benefits of daily calibrated risk model. The change in this company occurred toward the beginning of the month (February 4). Risk models that are estimated monthly would not capture this increase in risk until the next monthly estimation period. If ARG was part of your risk analysis, you would have been using stale risk data for 15 trading days.

This example shows what could happen to one company in a single day, but what could it mean if several companies in your portfolio experienced the similar scenarios? What implications would this have on the total risk of your portfolio? The implications are even scarier if you were to have a concentrated position in the company. Could you last 15+ trading days not knowing that the risk level for your company has significantly grown?

Any risk practitioner would agree that having current data is better than having stale data. I did not touch upon what having daily factor returns could imply for quant managers that take directional bets, but again we can assume that having daily factor data would have benefits above monthly data The risk landscape is expanding with a plethora of risk models available to the market. With so many to choose from, it comes down to which risk model fits your specific needs. If daily data fits your palette, Axioma may be your risk model of choice.

For more about Axioma models on FactSet, see the press release. Read more about Axioma at their website and access their research papers.

To receive new posts by e-mail, subscribe to this blog.

Introducing Betamax: A new measure of covariance stationarity (Part 1)

David Mieczkowski, PhD — 2010-06-23T18:16:27Z

Setting the Stage

Risk managers are recognizing the importance of correlation, but there are precious few quantitative tools around to address correlation risk. While the adage that "during a sharp market correction, correlations go to one" is widely repeated, this is usually a qualitative statement rarely accompanied by any real quantitative measure. In this post, I will introduce a new quantitative measure of covariance stationarity (also known as weak form stationarity), which is an important topic that is generally misunderstood. In a finance context, covariance stationarity simply means that the covariance (or, alternatively, correlation) structure of asset returns is constant through time. This assumption is made by some equity return factor models, structured credit models, portfolio optimizers, and return attribution models and has implications for tracking error. As such, it is implicitly incorporated into many of the software packages popular among risk professionals. Discussion of this new measure begins with a visit to an old friend: Beta.

Beta is perhaps the oldest and most well known financial risk measure. Defined as the slope of the security market line (SML) in the original Capital Asset Pricing Model of Sharpe and Lintner, Beta is the classical measure of non-diversifiable risk. As far as risk metrics go, Beta had a good run. Born in the '60s, Beta was instantly loved by rational economics and a decade of adoration culminated in its coronation as king in the early-'70s by Fama-MacBeth. As with any monarch, Beta’s reign was almost immediately besieged by challengers (such as the APT) and faced its fair share of assassination attempts (Roll’s critic, Fisher Black’s commentary on market noise). Yet it managed to survive, gain influence, and have a bit of a golden age through the following two decades. Sadly, Beta was rather violently overthrown in the early-'90s by Fama-French, who loudly declared in 1992 that “Beta is dead” and at the same time advanced their new three factor model. The king is dead. Long live the king!

There are many criticisms of the CAPM and, by extension, Beta. The first might be that it is paradoxically on the one hand a Nobel Prize worthy contribution to finance, yet on the other hand has as one of its core tenets, indeed its very mechanism of action, the assumption that all investors know the Beta of each stock fully and are all too willing to engage in statistical arbitrage to return the expected return of all assets back to their security market lines. It assumes that all investors know, and agree upon, the marginal distribution for every asset and, more importantly, the joint distribution of returns for all assets. In particular, it is assumed that investors have full knowledge of the return covariance structure of all assets within the market. That empirical tests have so far shown that if it ever was a valid theory, it worked better prior to its discovery than after two decades of brow beating MBA students and the investing public with it. That is a little something I like to call irony.

Putting all other assumptions and criticisms aside, our aim in this post is to examine one particular assumption that not only applies to the simple two factor model proposed by the CAPM, but of a very common assumption made by a host of statistical factor models. In this post we will examine the assumption of covariance stationarity and will limit our analysis to the U.S. equity market. We do this because equities have the deepest record of return data.

As pointed out by Chan-Lakonishok (Are the Reports of Beta's Death Premature?), in their rebuttal to Fama-French:

“Even if there were no compensation for Beta risk, this does not mean that Betas serve no use for investment decision-making. As long as Beta is a stable measure of exposure to market movements, investors should still consider the 'Beta factor' of a stock.”

Chan and Lakonishok go on to show that, at least for the 10 largest down months, high Beta stocks perform worse than low Beta stocks. Of course, if the general adage holds true, that when the market undergoes sharp corrections correlations go to one, then a similar result would be true of stocks with high variance relative to stocks with low variance. The important assertion of the CAPM is that it is not individual stock variance that investors should demand compensation for, but rather the individual contribution to the variance of the market return that determines expected return. This individual contribution is a stock’s Beta. Why all this discussion of Beta?

In part two, I’ll make some critically important observations about Beta that lead directly to a natural measure of covariance stationarity.

Continue to part two.

Subscribe to receive new posts by e-mail.

Introducing Betamax: A new measure of covariance stationarity (Part 2)

David Mieczkowski — 2010-06-23T18:16:27Z

The Main Course

Continuing from last week's part one: Beta is equal to cov(R_i,R_m)/var(R_m). Therefore, aggregate (that is, market wide) knowledge of beta is equivalent to knowledge of market covariance structure. Unfortunately, estimating the true Beta is difficult. Individual stock return series have a very high signal-to-noise ratio. Cross-sectional studies address this by forming portfolios and estimating Betas for those portfolios. Additionally, when ranking stocks on any estimator, you are also invariably ranking estimator error along with the actual signal of interest. This is known as the errors-in-the-variables problem, and a popular approach to solving it is highlighted in Fama-MacBeth. Essentially, their solution entails forming rankings based on an initial estimate for Beta and then reforming Beta estimates with post-ranking data. Yet, for all the acknowledged difficulty in forming quality estimators of Beta, there seems to be one thing that nobody finds the least bit problematic. For monthly return series, Betas are almost universally calculated using a look-back period of five years (60 months). Have you ever wondered why 60 months?

The answer has to do with convergence of the sample estimator. If the time series of returns is covariance stationary, 60 months will reasonably ensure that the sample estimator for Beta is not overly clouded by error of a small sample (being clouded by the return variance is another story).

To introduce our measure of covariance stationarity, I ask the universally overlooked, yet critically important question: how sensitive is Beta to the look-back period?

The figure below gives an example of how Beta and the regression R-sqr vary as a function of look back period for a typical stock in both a stable market and an unstable market environment. In the top two graphs we see that the regression R-sqr increases steadily as the look back period increases from 10 to 60 months. The example stock Beta is also seen to be reasonably volatile, but is stabilizing with increasing look back period.

For the stock in the unstable market, we also see Beta trending up as the look back period increases. However, the story for the R-sqr is completely opposite from the stable market. Here the R-sqr is clearly trending down as the lookback period increases. A similar pattern emerges for many stocks.

To define the risk measure, there are two things to note. First is that if individual stock returns are truly covariance stationary, then we would expect to see the R-sqr rise with a lengthening look back period, not fall. Second is that in choosing the optimal look back period, R-sqr should be our guide, as it measures goodness of fit. The most appropriate look back period for a stock should then be that look back period that maximizes the R-sqr. The look back period that maximizes the R-sqr within a given range is our new risk measure. He who discovers a thing gets to name it and so with tongue firmly in cheek, I’ll call this measure of covariance stationarity "Betamax."

Definition finally in hand, let’s take a look at how Betamax works to measure the degree of covariance stationarity on aggregate. To test my measure, I began by taking the largest 2,000 stocks in the U.S. market (NYSE, NASDAQ, AMEX) by market capitalization separately for each month from March 2010 to June 1980 (roughly 30 years worth of return data). I then screened the stocks to ensure that they had monthly return data for each of the prior 60 months. I also eliminated from the analysis stocks that had any single one month return in excess of 200%. This left, on average, around 1,500 stocks for each month.

For each stock, for every month, I estimated Beta by regressing the one month market return (proxied by the equal weighted return on the universe of stocks selected above) on the one month individual stock returns for each possible look back period (2-60 months). I assigned a Betamax to each stock by finding the look back period that maximized R-sqr in the range of 10 to 60 months (I also tried using a range of 30 to 60 months, which led to similar results). I determined a stock’s Beta to be the Beta estimator that corresponded to using the look back period equal to the Betamax. For each month, I formed decile portfolios by ranking stocks by their Betas. For each month, for each decile, I also determined a median Betamax. Finally, I calculated the mean of the decile medians. This figure can then be viewed as an aggregated measure of total market covariance stationarity. The image below displays the results of this measure through time.

On an aggregate level, Betamax works well to capture a number of major market shifts. All the major market corrections of last 30 years are captured: Black Monday, the Asian debt crisis and collapse of LTCM, the bursting of the internet bubble (demarcated by a sharp rise in Betamax), and the credit crunch. Even on the aggregate scale, but especially at the level of the individual decile medians, Betamax usually decays rapidly several months before market corrections. Additionally, Betamax decays sharply at several points that do not correspond to market corrections but can actually be traced back to the start of extended bull markets. Betamax falls off steeply from 1985-1986, which roughly began the strong bull market run up that precipitated the ’87 crash. Betamax also falls off steeply from 1995-1996, which roughly marks the beginning of the tech bull market.

The last image shows the median Betamax for the high and low Beta deciles. The most important take away here is that the high decile stocks almost never display signs of prolonged covariance stationarity.

Finally, I compared the market aggregate Betamax to the VIX. Not surprisingly, I found a strong negative correlation of around -50%. Yet, there is significant variation in Betamax not explained by variation in the VIX alone. Since a stock’s Beta is the correlation of the stock return with the market return multiplied by the ratio of the stock’s volatility to the market volatility, this implies that the failure of the market to be covariance stationary is partly due to the time varying nature of volatility, and partly due to a time varying nature of the correlation structure.

As I said at the beginning, this is an important topic that is not well understood by risk professionals. In these two posts, I have introduced a new measure that can be used to quantify covariance and correlation stationarity and established that the market is not stationary for long. Armed with this knowledge, I will use future blog posts to explore the full implications of this failing for popular risk tools.

Subscribe to receive new posts by e-mail.

Hindsight is 20/20 - Or is it?

Sean Carr — 2010-06-23T18:16:27Z

We're now 12 months on from the 10-year recorded low of MSCI World, the "bottom," as it were, following the global financial crisis, the credit crunch, etc. Since those gloomy days we have seen a sharp rally with markets across the world recovering and recording stupendous 12-month returns both in Developed Regions (MSCI The World 59%, S&P500 53%, Euro STOXX 50 59%, Nikkei 225 61%) and Emerging Markets (China Shanghaia Composite 42%, Bovespa 133%, S&P/CNX Nifty 124%). As the indices have rallied, so have the funds that are tracked against them, and while mid-2009 was a period in which we saw investors glad of any kind of positive return, this sustained improvement has brought an increase in focus on the quality of that return and a subsequent interest in the use of attribution in analysing it.

Attribution primarily adds value through the improved understanding of contributions to return achieved through active management. This is all well understood and needs no further explanation here. What I would like to highlight though is the importance of running this attribution as merely a part of the performance analysis (albeit a major one), but not as the sole analysis. I want to build on the complementary analytics discussed here, and underline the importance of partnering risk analysis with attribution, especially now.

To help clarify my argument, let us consider a very simple attribution analysis of a portfolio and benchmark combination, but rather than select any particular fund which might include bias, let's use two indices: the S&P 500 against MSCI World. This is merely to highlight particular biases in the standard analytics.

The 7.23% underperformance of the S&P 500 is split between an allocation effect (-3.23%) and a selection effect (-4.0%), i.e. the active allocation of the "fund" against the benchmark lead to an underperformance as well as the actual stock selection differences between the two. Now, as this is comparison of a U.S. Index against a World Index, I am not going to comment on the differences in constituents and the subsequent selection effect, but I do want to consider the sector allocation differences and the effect that that had.

The Financials sector of MSCI World has returned >100% over the last 12 months. Despite the falls that we saw in the sector in the prior six months, it was still the largest sector in terms of market cap weight 12 months ago by >3% above Healthcare. The largest weight combined with the largest return unsurprisingly leads to the largest contribution from Financials to the performance of MSCI World. Any fund that took a negative active weight to Financials (in this case our "fund" happened to have had a 5% underweight relative to the "benchmark") is going to see a large negative total effect being attributed to its performance (in this case -1.84%, or one quarter of our overall underperformance).

Criticism of the underperformance of a fund (such as our example one) would therefore be expected to focus on the decision to take such a large underweight against a sector that makes up such a large portion of the benchmark; what was the manager thinking? Let's consider that by going back 12 months and consider what the risk profile of MSCI World was at that time.

This chart shows the estimated volatility of MSCI World over the 12 month period from June 2008 until June 2009 as forecast using the R-Squared Daily Global Equity Model. The peak of the chart is the second week of March and the largest contributor to risk was (you'll have guessed by now) the Financial sector. This sector had also recorded returns of <-60% over the previous six months. Armed with this more complete data it is much easier to see what the 'manager' might have been thinking: the market was extremely risky, most of that risk was attached to the Financials sector and that sector had hugely underperformed. Or put it another way, would anyone encouraging an overweight in that sector with this information in front of them have been given any credibility at all? I therefore encourage you to take away a couple of things from this week's blog:

Beware convenient date selections, the last 12 months look great, the last 24 not so much!
Use all the tools available to you rather than letting one specific analysis methodology be the only source you consider.

To receive new posts by e-mail, subscribe to this blog.

The reports of my death are greatly exaggerated, take 2

Bryan Hoefs — 2010-06-23T18:16:27Z

A few months ago I wrote about the death of the dollar, and I thought it was an appropriate time to follow up on the topic. Then, the press was discussing why the dollar’s demise was upon us. From some quick analysis, it didn’t appear that the average fund manager believed the hype. So who has been right so far, the press or the fund manager?

Since mid-October, the dollar index has strengthened 2.9% (as in the other post, this is from the Fed Reserve). The dollar increased almost 10% against the Euro, 6.35% against the Pound, and stayed flat against the Yen. Looks like the dollar is still alive and kicking. Maybe the fund managers had it right, and the death of the dollar was premature.

Lately, talk has moved from the death of the dollar to the death and potential breakup of the Euro. Much has been written about the PIIGS countries (Portugal, Ireland, Italy, Greece, and Spain) and how their massive borrowings may overwhelm the entire Euro community. Is all that bacon clogging the Euro’s arteries, and will there be a heart attack?

The much talked about decline will no doubt be tough on many residents in the Euro area, with unemployment most likely rising. But how will this affect equity markets? Like before, using FactSet’s Stress Testing tools, I’ll examine the effects of a 30% decline in the Euro, from the perspective of a Euro based investor. I’m using a Euro index against a handful of currencies, as published by the European Commission. I’ll use the R-Squared Global Equity model, using Northfield and Barra will produce similar results.

This Stress Test shows the MSCI Euro index increasing 10% from a 30% decline in the Euro. The sector story sounds familiar. The Financials and Materials sector are again the biggest beneficiaries. Exporters are helped by a decline in the currency. Defensive sectors such as Health Care and Consumer Staples are again poor relative performers.

Moving the MSCI World index, the stress test predicts an increase of about 26%. North American and Japanese stocks are the best performers. Euro stocks will have a positive return, but underperform the overall index.

So, what are investment professionals doing? We know many hedge funds are positioned for a fall in the Euro, taking the largest short bets in history recently. How about the average fund manager? Using the Lipper Active indices, we can get an idea of how the average active fund manager in a given strategy is positioning their portfolios. Let’s look at the Lipper Global Region against the MSCI World. In the Euro decline scenario, the average manager would underperform the index by about 1%. The average fund is underweight outperforming North America and Japan, while being overweight Europe sectors. These allocations are the main cause of the predicted underperformance. So it doesn’t appear the average fund manager believes in a major collapse in the Euro.
Will there be a Euro heart attack? The press and many hedge funds think so, hedge funds putting a lot of money behind that belief. The positions the average longer term investor is taking leads me to believe they think the Euro will make a recovery. Time will tell who is proven right. But if recent history is any guide, I can't help but think that a few months from now we will have forgotten about this "Euro death" and move on to the next “dead” currency.

Don't miss a post! Subscribe by e-mail to receive new entries as soon as they are available.

Don't miss FactSet's upcoming Risk events

Risk News — 2010-06-23T18:16:27Z

The next few weeks are the perfect time to see FactSet's risk tools in action. Catch us at these events.

March 10 Seminar
Risk: The challenges of multiple asset classes

Join FactSet and the CFA Toronto for a luncheon on March 10, where Senior Product Manager Bill McCoy will address the multiple challenges of building a risk model that truly meets the needs of multiple asset classes.

March 11 Seminar
Fixed Income Management Essentials: From single bonds to risk analysis

We'll be in Boston on March 11 from 9:00 a.m. to 1:30 p.m. to present a comprehensive and educational event on the challenges of managing a fixed income workflow.

March 11 Luncheon
Risk for Super Funds

In Melbourne, we'll discuss why Super Funds should not only be thinking about risk, but analyzing it in a sophisticated way. FactSet and our risk partners APT, Barra, and Northfield will describe how Super Funds can measure, manage, and understand the risks they are taking or external managers are taking on their behalf.

March 17 Live webcast
Accurately Measuring Risk Across Asset Classes

Most risk models are either equity focused with some fixed income flavor or fixed income models with some rudimentary equity addition. An accurate total risk model must unite, in one framework, descriptors of various equity, currency, and fixed income risks to provide a very granular view of both equity and fixed income markets. Only this type of model can truly report risk across asset classes.

Daniel Satchkov will demonstrate how to accurately measure risk across all asset classes through a collection of risk statistics such as Value at Risk (VaR), Expected Tail Loss, Kurtosis, Skewness, Tracking Error, Stress Testing, and others.

For the latest in FactSet events, follow us on Twitter @FactSet.

To receive new posts by e-mail, subscribe to this blog.

A quick VIX?

Risk News — 2010-06-23T18:16:27Z

by special guest blogger Matthew van der Weide, FactSet Quantitative Specialist, Amsterdam

On this stage, we have discussed the suitability of a risk model in terms of matching its forward looking horizon to a client’s investment process. I’m the first to agree that this indeed is an important dimension and should be considered carefully. What use has a risk prediction if the information is too late to react to? That said, I have noticed a trend among the different risk model providers to provide shorter term horizons in addition to their standard models. Both Barra and Axioma provide shorter and longer term versions of their risk models, and Northfield has produced near term versions of most of its models. The major focus of such models is to pick up increased levels of risks faster. Now, one could philosophise whether these additional horizons have arrived in time and whether they could have provided adequate warning signals for the multi sigma events we encountered in late 2008, but the fact remains that these models are here now and can provide useful insight if deployed in the right manner.

To highlight this responsiveness, consider the observed jump in the VIX over the last couple of weeks (while not all encompassing, the VIX is considered a valid proxy as a broad indication of perceived risk in the markets). As a broad indicator, the recent movement gives as an opportunity to see if and when a short-term model picks up on such an event. To do so, I decided to look at the absolute risk levels of some broad indices, in this case the FTSE All World, S&P 500, and MSCI Europe. I did this using the R-Squared Global Risk model, a daily updated model designed to predict risk on the ultra short horizon and to be very responsive.
Considering the above charts, we clearly see increases in the levels of risk after spikes in the VIX, especially during the last week. The magnitude of the impact in the spike seems related to the amount of exposure to the U.S. Market in our three indices (VIX is a measure of the volatility of the U.S. S&P 500 Index options).

Now consider the monitored impact on a monthly model. It’s going to be both less timely and a more drastic. This because the factor variances will be updated from month end to month end, so a fund may perceive a shock in predicted risk from one day to the other (when the model updates), while the fund composition itself didn’t change.

We’re not suggesting a quick fix here, saying one should only look at short term risk. But we do advocate multiple horizon measurement; longer term risk management may reflect the strategy of a fund, but when markets volatility rises, the information and analysis permitted through a shorter term model carries real value.

Subscribe by e-mail to receive new Taking Risk posts as they are published.

Notes from the IPARM conference in Hong Kong, part 3

Willett Bird — 2010-06-23T18:16:27Z

The second day of the Third Annual Investment Performance Analysis and Risk Management Asia 2010 (IPARM) conference at the Kowloon Shangri-La in Hong Kong, China began with a panel discussion on the lessons learned from the global financial crisis and the roadmap for 2010 and beyond. Jean-Marc Sabatier, Head of Risk Management Asia for Amundi, started off the conversation talking about how 2010 is the year of opportunity for risk management, particularly in Asia. He cited two reasons:

Human resources. Management finally sees the need, nay, the requirement, for a real risk management team in place and are finally willing to pay and support to keep a highly respected team in place.

The balance of power has switched (a little bit). In the past, if the risk manager said “no,” it was acknowledged, but then everyone moved on. Now, risk managers are more easily and readily able to say “no” to portfolio and managers and marketing groups and actually carry some weight.

Jean-Marc also added that the job of a risk manager is no longer only about reporting; the job of a risk manager begins with the risk report. Oliver Bolitho, Managing Director from Goldman Sachs Asset Management, then discussed the concept of regret risk which in Asia is related “to a ‘face’ thing that leads to taking logic off the table." Oliver believes this is one of the bigger issues facing the industry as he has see countless examples of portfolio managers turning off their thinking when faced with an investment decision.

From the audience, the panel was asked who should have the final say on the risk of a portfolio? The portfolio manager? Risk manager? Combination of the two? Someone else?

Oliver jumped in first by describing risk as a culture. He continued by saying that if we tried to codify risk, it will get boxed in and will not be there when we need it most. By way of example, if we codified risk (e.g, you could only buy securities with a certain rating), just think what people would have bought in the past few years. Oliver concluded that he thought the risk manager should have the final say, but it should not be up to a single person.

Also in response to the question, Dr. Lincoln Rathnam, CFA, Global Head of Investment Management for EM Capital Management, mentioned the positive experience he had working for an investment management firm that was a partnership and anyone at the firm could say no. Anecdotally, Lincoln thought that having a corporate structure of a partnership was a prime reason why Brown Brothers Harriman escaped relatively unscathed from the financial crisis; everyone at the firm had a veto and they avoided the toxic assets that others so readily accumulated. But ultimately, Lincoln’s answer was that he thinks there needs to be a balance of power between the portfolio manager and the risk manager, one party always wants to say yes, the other party wants to say no, and a middle ground that must be found.

There was also a brief discussion about the “age factor” of risk managers (also known as the “value of experience” to the older demographic). In Asia, and perhaps globally, many risk management teams are junior, i.e., it is often a junior member of the investment management team and all too often someone who has not gone through many of the historical ups and downs. There were no firm answers on how to address this issue, although Jean-Marc mentioned that Amundi recently announced a new policy in which all Portfolio Managers must spend at least three years serving in a risk management capacity. Lincoln made the analogy to General Electric back in the Jack Welch days when he mandated as part of their executive management program that everyone had to spend some time in internal audit.

Next up was Dr. Stan Uryasev, Editor-in-Chief of the The Journal of Risk, who gave a talk on deviation CVaR (Conditional Value at Risk). While this risk measure has been around for a while, as a co-inventor of the methodology, Stan was able to expand on the methodology both in theory and practice. Of particular note, Stan emphasized that CVaR is most useful for risk management, not risk measurement. I strongly encourage those of you interested in learning more to check out his slides that he has made available on his website here.

After serving on the panel, Lincoln pulled double duty with a presentation on stress testing, although to me, the most interesting part of his talk was when he touched on the topic of crisis. Contrary to most, Lincoln believes that “crises are not rare events. We go from crisis to crisis to crisis. It is our nature.” Not a comment you can ignore coming from a man with over 30 years of investment management experience, and to solidify his point, he made reference to a list of financial panics, scandals, and failures. Far from comprehensive, I am sure, but it did cement his point that the next crisis is never too far away as it goes through crises starting in the 17th century and ominously ends with nothing next to number 210.

The final panel of the day focused on finding a risk model or performance system that is appropriate for your investment process. Dr. Laurence Wormold, Head of Research at SunGard APT, started things off with his three pillars of risk analysis:

Risk measures: The simple stuff, tracking error, VaR, etc.

Attribution: In Laurence’s words, “turning 1 number into 100.”

Stress testing and scenario analysis: In his mind, this is the most often ignored aspect of risk analysis as he firmly believes in building shocked market risk models.

A member of the audience immediately jumped in questioning Laurence’s assertion as every firm that he knew of did some form of stress testing. Lawrence acknowledged this, but added that for most firms, stress testing is a box ticking exercise that is largely ignored throughout the company. The stress testing that most firms do lack imagination and is too simplified (e.g., S&P 500 goes down 20%). For the most part, the stress testing that is in the marketplace today suffers from a herd approach; everyone is testing the exact same thing. Laurence further suggested that the current tests should be anchored in economic plausibility; a firm should start from a historical event and then invite colleagues to take that information and think about other ways to create realistic scenarios.

Overall, IPARM Asia was a well organized conference with a very solid slate of speakers and I was quite happy to hear that the organizers have already announced that the fourth annual conference will take place in Hong Kong again next February.

Subscribe by e-mail to receive new Taking Risk posts as they are published.

The illusion of stability, part 2: What fuels bubbles?

Daniel Satchkov — 2010-06-23T18:16:27Z

Blame It on Taylor

"It wasn't me. It was the one-armed man. Alright, alright, I confess. I did it, you hear? And I'm glad, glad, I tell you! What are they going to do to me, Sarge? What are they going to do?"

"Sorry, son, that's not my department."
- Jim Carrey in the The Mask

As if to respond to my 2009 year in review post, Ben Bernanke came up with a huge speech (please do not smile; I wrote “as if”) outlining why zero real interest rates have had no effect on the housing prices. With this groundbreaking discovery, he opened 2010 by declaring war on common sense and possibly laying claim to the Noble Prize. Yes, you heard it right, zero interest rates do not affect housing prices.

Let's consider some of the Chairman’s arguments:

The Taylor Rule

Bernanke invoked a formula called the Taylor rule which purports to proscribe appropriate monetary policy to strike a balance between economic growth and inflation of asset bubbles. Bernanke suggested that if forecasted inflation rate is used as an input into the Taylor rule then Fed’s policy of zero real interest rates in the years when the housing prices were exhibiting explosive growth then the rule suggests that the policy was appropriate. Taylor himself responded to Bernanke explaining that he misinterpreted and misused the rule. Without getting into any mathematical detail, the first question to ask is, "How valuable can a rule be, if different inputs can be used to provide any result you want?" Imagine a space engineer whose creation crashed and injured the public proclaiming that he used a different gravity constant and therefore was correct. The point is that economics is not physics and a heavy dose of common sense is needed. Even Taylor himself said that mathematical discussions should not obscure the plain fact that interest rates were ZERO.

Smoke and Mirrors

"The fact that our econometric models at the Fed, the best in the world, have been wrong for 14 straight quarters does not mean they will not be right in the 15th quarter." - Alan Greenspan in a testimony before Congress

In the second argument, Chairman Bernanke made a similar attempt to cloud economic discussion with equations carrying a boatload of assumptions which will give most any answer desired. He referred to a set of econometric models developed at the Fed based on the technique called Vector Autoregression. Using past values of housing prices, interest rates and other economic variables, the model can be coaxed into showing what values are reasonable for any one of the variables given the actual values that were observed for the rest (conditional distribution). Using that technique, it can be shown that only about half of the explosive housing growth can be attributed to the Fed’s policy. In addition to eschewing common sense, there are fairly obvious and fatal statistical problems with Bernanke analysis:

The model was calibrated during the period of non-zero interest rates and then linearly extrapolated into the zero interest rate period. It assumes that the relationships between variables are linear and constant, that there is nothing inherently different about zero interest rate environment and that only changes in variables matter. Needless to say, all these assumptions are flawed, and this is a perfect example of the fallacy of linearity of which I have written here.
The range of housing prices conditional on the interest rates was limited to two standard deviations. Even using Bernanke’s own heroic assumptions there is still 32% of the distribution remained outside. When looking at the graphs at the end of the speech, it become fairly obvious why two standard deviations were used. Simply put, three standard deviations would not give the desired answer; the line would be too close to the actual housing line, thus suggesting that even with all its flaws the model rather contradicted the speaker.

Adjustable Rate Mortgages and Rates

The third argument was that the problem stamped from lax regulations not from zero interest rates. Here is one quote from the speech:

“Moreover, less accommodative monetary policy would not have had a substantial effect on ARM payments…Clearly, for lenders and borrowers focused on minimizing the initial payment, the choice of mortgage type was far more important than the level of short-term interest rates.”

Here, the Chairman is partially right, and his call to create a systemic risk agency can only be applauded. However, to argue that low interest rates had nothing to do with the proliferation of creative mortgage lending is to ignore the reason why people took on ARM mortgages in the first place. They did so because housing prices were rising at a tremendous pace for a number of years before. As Bernanke himself shows, ARM mortgages did not become popular until 2006, which was closer to the end of the bubble. Without zero interest rates, we would not be talking about them.

More Smoke and Mirrors

“'You might just as well say that "I see what I eat" is the same thing as "I eat what I see"!'

'You might just as well say,' added the March Hare, 'that "I like what I get" is the same thing as "I get what I like"!'

'You might just as well say,' added the Dormouse, who seemed to be talking in his sleep, 'that "I breathe when I sleep" is the same thing as "I sleep when I breathe"!'

'It is the same thing with you,' said the Hatter …” - Lewis Carroll, Alice in Wonderland

Bernanke's last point confuses the whole matter even more:

“In particular, we need to understand better why some countries drew stronger capital inflows than others. I will only note here that, as more accommodative onetary policies generally reduce capital inflows, this relationship appears to be inconsistent with the existence of a strong link between monetary policy and house rice appreciation.”

Essentially, what this passage is saying is the following: Countries with real estate bubbles exhibited higher capital inflows. Since higher capital inflows usually happens when interest rates are high and not when they are low all this is very confusing. The interest rates must not have been low after all.

This kind of reasoning implicitly assumes that economics can be described by the same functional relationships as Newtonian physics, i.e., that if high interest rates cause capital inflows then high capital inflows will not occur when interest rates are set too low. It assumes a kind two-way relationship that is called one-to-one correspondence in mathematics.

Why did the Chairman have to avoid considering the possibility that capital inflows can be caused by reasons other than high interest rates? Is it not the obvious conclusion that the capital inflows were caused by the very presence of the various bubbles that Alan Greenspan and Ben Bernanke then denied? Who cares about a few percentages in interest rates when values of assets are doubling?

What does all this mean for a risk manager? It means that Ben Bernanke sees no inherent problem in zero interest rates as long as Consumer Price Inflation stays low. This in turn means that we are in for another decade of Black Swans. Buckle up.

Subscribe by e-mail to receive new Taking Risk posts as they are published.

Notes from the IPARM conference in Hong Kong, part 1

Willett Bird — 2010-06-23T18:16:27Z

Greetings from Hong Kong! I will be attending the Third Annual Investment Performance Analysis and Risk Management Asia 2010 (IPARM) conference at the Kowloon Shangri-La later this week. The past two events have had an excellent range of speakers and I am looking forward to sharing some of the more interesting talks with you from the 21 speakers from all over the world who will be headlining the event.

In particular, I am looking forward to hearing from:

Trevor Persaud, the Head of Investment Risk Oversight and Performance at Prudential Asset Management in Singapore (and Chair of the conference) who will discuss redefining the traditional role of risk managers.
Dr. Lincoln Rathnam, CFA, Gloal Head of Investment Management at EM Capital Management who will be giving a talk on stress testing.
Jean-Marc Sabatier, Head of Risk Management Asia for Credit Agricole Asset Management who will be speaking on data management best practices.
Daniel Wallick, Principal in the Investment Strategy Group at Vanguard who will be examining some of the key challenges of risk adjusted performance measurement in the current market.

Watch this blog later in the coming week as I post on some of the interesting topics covered at the IPARM two day conference.

To receive new posts by e-mail, subscribe to this blog.

Notes from the IPARM conference in Hong Kong, part 2

Willett Bird — 2010-06-23T18:16:27Z

A day full of fascinating discussions at the Third Annual Investment Performance Analysis and Risk Management Asia 2010 (IPARM) conference at the Kowloon Shangri-La in Hong Kong, China. Hope you've been enjoying my live tweets from the conference.

Kicking the event off was Trevor Persaud, the Head of Investment Risk Oversight and Performance at Prudential Asset Management in Singapore. Trevor's topic was redefining the roles and responsibilities of performance analysts and risk managers to better support the investment management teams. Trevor started off his presentation by asking the audience whether the portfolio manager at their respective firms was the only person at the firm who can say exactly what is going on in a particular portfolio. While Trevor acknowledged that this was expected since the PM should have the expertise, he questioned whether this lack of challenge and oversight is healthy for the fund. From his experience, he has found that independent expertise of a fund helps moderate the action a PM might otherwise undertake.

Having worked in both Europe and Asia, the first question asked of Trevor by the audience was what were the big differences between the risk management function in the two regions? Trevor cautioned his response by saying this was not a sweeping statement, but he thought that in Asia, up until recently, the role was purely operational, very limited independence and oversight. However, of late, he has noticed that in Asia, there is a thicker layer of senior management (compared to Europe where the Portfolio Manager is king), and management has been more receptive to risk managers looking to take a more active role and becoming more than just serving a pure operational function. In Europe, Trevor thought that risk managers have taken on more responsibility than their Asian counterparts, but that a risk manager's ascension is capped, there comes a point (and I hope he was speaking from personal experience) when the risk manager is at peril of overstepping his or her bounds in the hierarchy of a European investment management firm.

Daniel Wallick, Principal in the Investment Strategy Group at Vanguard, gave an equally interesting talk and got everyone's attention when he equated risk management with the Allegory of the Cave from Plato's The Republic.

For those who haven't read the book since high school, Daniel's analogy was that risk is similar to man looking at the shadows on the wall in the cave; we are not really sure what we are looking at. Daniel's talk expanded into the two reasons why we need risk management (people are imperfect and crises happen) and ended with four interesting fundamentals that are followed at Vanguard:

Risk management is an integral part of the investment decision making process.
An enduring value-added investment program in good ties and (crucially) in bad.
Top-down qualitative judgment/rigor + discipline in quantitative measures.
No substitute for the judgment and experience of the risk management and investment teams.

Daniel's session ended with a question from the audience on his assessment of the current risk environment in the U.S. given his experience and focus there. Daniel replied that he was not specifically concerned near term with inflation (in the U.S.), but what concerned him was how exactly should the Fed back out of what it has been doing and how/when do they do that.

There were a couple of additional speakers today who covered topics that I will take up in a future blog post. As for tomorrow, we have an equally interesting set of speakers and topics coming up in day two of the conference, some that have caught my attention include:

Dr. Stan Uryasev, Editor-in-Chief of The Journal of Risk, who will be seeking alternatives for VAR as risk measure.
A panel discussion on finding the right risk model and measurement system that is appropriate for your investment decision process, particularly interesting since my colleague covered this topic in a blog post last month.
Peter Urbani, CIO at Infiniti Capital who will be revisiting risk management and performance measurement for hedge funds.

Please check back early next week for a recap on these interesting topics.

Don't miss the next post in this series. Receive new blogs by e-mail.

The illusion of stability, part 1

Daniel Satchkov — 2010-06-23T18:16:27Z

“The real trouble with this world of ours is not that it is an unreasonable world, nor even that it is a reasonable one. The commonest kind of trouble is that it is nearly reasonable, but not quite. Life is not an illogicality; yet it is a trap for logicians. It looks just a little more mathematical and regular than it is; its exactitude is obvious, but its inexactitude is hidden; its wildness lies in wait.” - G.K. Chesterton

We have seen this happen many times. A financial crisis erupts and everybody, including the New York Times, remembers risk management. There appear lengthy expositions of the falsity of assuming normally distributed returns, and everyone loudly wonders why the industry was not warned by risk models. However, as soon as the situation stabilizes – or rather appears to stabilize – risk management is again relegated back to the specialized conferences and, incredible as it may seem after the last twenty years, tracking error is again used to completely describe the risk profile of the portfolio.

This is not a conspiracy; rather it is the effect of what John Cassidy the illusion of stability. This illusion is supported by a few pillars, each of which I intent to discuss in this and the next few posts. The first pillar is the economic theorizing that comes from looking at the economy as a physician looks at the elementary particles or an astronomer at the galaxy. This is the logician’s trap taught in every institution of higher learning.

In January 2009, the Basel Committee on Banking Supervision finally attempted to disconnect the feeding tubes and get out of the matrix when it proclaimed that:

“most risk management models, including stress tests, use historical statistical relationships to assess risk. They assume that risk is driven by a known and constant statistical process. Given a long period of stability, backward-looking historical information indicated benign conditions so that these models did not pick up the possibility of severe shocks nor the build up of vulnerabilities within the system.”

I know I have used the above quote before. Nevertheless, I keep using it because I believe that it not only provides in a capsule form many of the key misconceptions about risk management, but gives a glimpse of possible ways to deal with them. I have written before about the problems with the present paradigm and the ways of correcting for them. Now that I have put my logical cart before my imaginary horse, let me go back and find the horse. In other words, I would like to briefly discuss the sources of the misconception as I see them, mainly in the economic theories of equilibrium. Why did much of the financial industry believe that “risk is driven by a known and constant statistical process”? Surely, this is not an obvious observation; it requires a certain mindset, a view of the financial markets as a kind of galaxy that we can observe with the telescope and count that the resulting calculations will not need to be changed from day to day.

To understand the source of this view we need to look no further than Leon Valras, a brilliant French economist who was the first to attempt to create a mathematical equilibrium model of the economy. His friends later remembered that he was very inspired by the book on physics that he had then recently read. It fascinated him so much that he vociferously proclaimed his intention to create a new science of political economy, one that would governed by calculus equations just like physics. “Equilibrium” is derived from physics, and it was only natural to look for the equilibrium in the economic system given his premise. One key feature of his model had far reaching implications and it affects virtually every area of economic and financial thinking including risk management. This idea in French is called “totonemont” and it essentially defines the process of gradual adjustment by which participants in the economy slowly move it toward its equilibrium. The process roughly is as follows:

Sellers and buyers announce the prices at which they are willing to transact.
If the prices match, then the equilibrium is reached according to a set of equations written down by Valras.
If the demand and supply are mismatched, prices are altered in increments until the balance is reached, a gradual process called “totonemont.”

As we can see, this model appears to simply follow common sense, something we observe in our daily lives. However, we need to ask a question that is extremely relevant today for any finance practitioner: what makes this change gradual? Why would we assume that the process is constant and stable? The basic answer to these questions given by quantitative economics and quantitative finance is the assumption that supply and demand are relatively stable.

It is interesting to know that Valras explicitly applied this model to financial markets, despite the fact that Europe saw a number of financial bubbles in the18th and 19th centuries. If we are talking about wheat or corn, it might be reasonable to suppose that neither the consumers’ desire to consume them nor the producers’ ability to produce them will change very quickly. The first is limited by the physiology of the human beings, and the second by the physiology of planet Earth. But the situation is quite different for financial assets. There is no obvious limitation on the amount of financial assets buyers are willing to purchase other than the supply of liquidity and credit in the economy. And as we recently saw, the supply of financial assets may grow very quickly if the demand is there (the MBS and CDS markets are only two recent examples).

More so, as German economist Gustav Schmaller observed more than a hundred years ago, demand can actually increase with price. By now it should be obvious that the situations of demand increasing with price or decreasing with it are so common in finance that they almost constitute the rule. This problem lies waste to any attempt to view financial markets as stable or constant. As risk practitioners, we should be aware that there have been and will be periods when supply and demand change drastically, making the models calibrated in normal times useless. When demand for financial assets falls with prices, it creates a wave of demand for and subsequent shortage of liquidity, which is really what is hiding behind the frequently mentioned “rise in correlations.”

The fact that we do not have a stable process that easily lends itself to modeling should not deter us from quantifying the risk of our portfolios and our exposure to such extreme situations. Our primary answer to this problem has been the development of the Event Weighted method of stress testing. This method assumes that the situations in which demand falls with price have many similarities across time, and therefore we can look for similar periods of liquidation to estimate how our portfolio will respond to future instances. This assumption has shown its validity in empirical tests and certainly does not require the leap of faith involved in assuming that financial markets are as orderly as the Solar galaxy in its motion. The financial system may not be stable and constant, but the reaction of the participants in times of instability can be modeled to supply a risk manager with valuable input for decision making.

Don't miss the next post in this series. Receive new blogs by e-mail.

Considerations when implementing a risk management system, part 3

Andrew Kovacs — 2010-06-23T18:16:27Z

Continuing from my previous posts, I will address the next in our series of questions to consider when implementing a new risk management system. This is the final post of the series. As a reminder, I am summarizing some key points you should consider when selecting a risk system.

Here again are the questions I address:

Part One:

Who are the Stakeholders?
Why do you need risk?

Part Two:

What are your options?
Where do you need to see risk?

Part Three:

How should you implement your solution?
When should this take place?

How should you implement your solution?

Let's first address this point: Goodness of Fit does not apply to models only (Risk model selection). We have taken the time to identify the stakeholders, our analytical requirements, and what we need from a risk model. Now all we need to do is pick a model. Here are some things to keep in mind during the evaluation process.

Do you really understand how the model is constructed and what the output tells you?
As a risk practitioner it is imperative to have a firm grasp of how a particular model is constructed and how that model’s results are to be interpreted. It’s not enough to understand that a model uses pre-specified factors or principal component analysis to estimate risk. The reality is that any risk vendor worth its salt should be able to provide lots of details regarding model construction and interpretation. In the end, if you don’t understand the model how can you effectively communicate the results to your clients or apply them to your investment decision making process?
Is the provider open up about their methodologies?
In this day and age, if a model provider operates like a black box, I would look elsewhere. Certainly there may be elements of the risk model creation that are considered proprietary by a third party vendor, but there is really no excuse for a vendor limiting a client’s access to construction details. In my mind, the better you understand a risk model at a fundamental level, the better you can use it to understand your portfolio's risk.
If you have questions, do they have answers?
Documentation and transparency are important, but there will always be questions unique to your firm and risk provider needs to be able to help you answer and these questions. Ultimately behind every risk model there are people, and this means that during the model selection process you need to evaluate the relationship with the people behind the risk model as much as anything else. If you cannot have an open dialog with your risk provider you are not going to get the most out of the system.

Next up: data (unfortunately there is more to risk than risk models). Risk analysis requires an underlying set of data, which is to say, risk measures are only as good as the data they are based on. So we need to think about what data is actually needed to perform any risk analysis. This is one of the most overlooked points in risk analysis. It boils down to a simple question. Do I want to manage data or manage money? So what are the data sets we should be thinking about?

Risk Model: Why we need this is should be self evident.
Portfolio data, benchmark data, and pricing: As we all know, in order to generate portfolio risk we need portfolio and benchmark weights. That means we need portfolio and benchmark holdings and quantities and pricing to calculate accurate market values and weights. It is important that you are comfortable with the accuracy of the portfolio and benchmark data.
Security-descriptive data: If there are securities (e.g., derivatives, unlisted, futures, trusts, real estate) in your portfolios that are not covered by the risk model(s) you use, you need a way to increase this coverage. One of the most common means is to supply security terms and conditions. Different firms have different levels of access to terms and conditions data. E.g., If you are Plan Sponsor you may not have a good source for this data and may have to rely on your managers to supply it. Where will you get this data if you need it and how will you store it?
Fundamental and economic data: There are three good reasons to think about this type of data:
1) To give a clear picture of portfolio’s current situation in a way that makes sense to people unfamiliar with risk, it is often helpful to include other data in the analysis to illustrate a point. For example, if you have underexposure to something like “size” or “value,” it may help your cause to include market cap or valuation measures along with your risk analysis to help with the interpretation.

2) If you plan on optimizing, you will likely want to have the ability to incorporate market data into your models to tilt your portfolio(s) towards real world factors that are important to you investment process.

3) If you plan on applying any stress tests, you will undoubtedly need market data to create the scenarios you wish to test (e.g., rising oil prices, decreasing interest rates, changes in trading volume).
History and timeliness: Make sure that you have a good handle on what kind of history you need and will have access to as well as how often the data is updated. If you are concerned with historical ex-ante risk analysis or optimization, you will need historical data for the portfolio, benchmark, and risk model before you can move forward.

Choosing a risk provider is a big decision, but it should not be made in isolation, so consider the following. Risk models are no longer linked exclusively to the model providers; they are now available through a variety of platforms, integrated to varying degrees. Because of this you may have an opportunity to not only solve your risk needs, but to potentially also meet other needs or solve other problems at your firm unrelated to risk. This could mean consolidating services, saving money (or at least spreading the cost), and minimizing redundant processes.

Scalability and flexibility are particularly important because it may mean you can use one system for multiple purposes within a risk framework and potentially beyond. If you belong to a Risk Team, you may only care about risk itself, but many financial professionals wear multiple hats these days and are interested in several things (e.g., portfolio management, risk, performance, marketing). In the past, you may have needed more than one platform to meet all of these needs. Now if you can find a platform that is both scalable and flexible and still meets your core risk needs, there is a good chance that you can consolidate services, which in turn can lead to cost savings and distribution. In this environment every cost is being scrutinized so any service that allows you to get good value for cost is in demand.

Of course related to all of this is the data behind the scenes. Most investment firms would rather stay away from the business of managing data and stick to their core competencies. As such, make sure you understand how the platforms you are considering integrate data, from your own portfolio holdings to benchmark data to third-party data and beyond. You might find a system that does much of what you need but still requires you to plug in lots of different data sources to get the job done.

Kick the tires! You wouldn’t buy a car without looking under the hood and taking it for a test drive. Implementing a risk system can be difficult, so take advantage of trials, set some goals, and at the very least make sure you have satisfactory answers to the following:

Is it easy to test out a simple situation?
If you can’t get results for a domestic equity portfolio easily, don’t hold your breath when it comes to large multi-asset class portfolios.
Is the support responsive?
If you don’t get the help you need during a trial, forget about when you are client.
How is the software?
If you can’t use the software, you can’t analyze risk.

When should all of this take place?

There is no perfect timeline for selecting and implementing a risk system, but here is a rough guideline of how it often works:

Investigate the needs and requirements internally before anything else. Do as much leg work internally as you can before casting your net and looking at providers. Meetings and demos will be much more effective if you have a good grasp of what you think you need.
Look at the options available in the marketplace. Do some research about the model types you might be interested in. If possible, attend relevant conferences. Contact vendors and ask for information.
Meet with the providers and have them explain their solutions in the context of your needs. Risk providers have lots of experience; they should be able to do this and this may force you to re-evaluate your questions.
Narrow the field. Based on meetings, demos, and conversations, you should have some comfort at this point about who you think are legitimate options.
Request a trial of the top candidate(s). Keep your goals and objectives in mind. Start thinking about implementation. Perhaps part of the trial can be dedicated to moving forward in this regard.
Don’t lose focus. Circle back to your original requirements. Are you still on target or have you drifted away from your core goals?
Purchase approval. Depending on a firm’s purchase approval process this step can sometimes significantly delay or impede implementation. Costs should be discussed early so that there is no confusion or surprise at this stage.
Full implementation. Fully implementing a risk system may take a while, so create a reasonable time line and start simple or with the key portfolios.

Finally, anything worthwhile tends to be difficult, and I believe that implementing a risk system falls into the category of something that is worthwhile. If you have specific questions, please contact me.

To receive new posts by e-mail, subscribe to this blog.

Considerations when implementing a risk management system, part 2

Andrew Kovacs — 2010-06-23T18:16:27Z

Continuing from my previous post, I will address the next in our series of questions to consider when implementing a new risk management system.

Part One:

Who are the Stakeholders?
Why do you need risk?

Part Two:

What are your options?
Where do you need to see risk?

Coming in Part Three:

How should you implement your solution?
When should this take place?

What are your options?

Now that we are clear about who we are trying to please and the reasons we need risk in the first place, we can tackle finding the right risk model for the job. To do this we need to be able answer three primary questions:

What market(s) do you invest in?

Firms often try to use a single broad model to analyze many smaller markets that they care about. While this may be a more cost effective option then buying many market-specific models and the large model may, in fact, “cover” all of the securities they care about in the smaller markets, these firms are not taking advantage of the research and development performed by the risk vendors to design their models for specific purposes. For example, it seems to be more and more common to use a Global Equity model to analyze equity portfolios that invest only in single countries. While you will certainly be able to calculate some risk numbers, I would argue the value of these numbers is reduced. Most global models are designed to use or capture factors that apply across many diverse markets and therefore are ideal for global investors, while single country models typically use or pick up factors that are unique to a single market. Imagine using a Global model to analyze an Australian equities portfolio. Most global models will likely be dominated by factors that primarily affect a few large countries; is the predicted risk of such a model ideal for this situation?
What asset classes do you care about (equity, fixed income, derivatives, unlisted assets, etc.)?

You can easily extend the rationale from point 1 to multi-asset class models. If you could, would you use an equity model to analyze a REIT portfolio?
What are the right time horizons (Investment Horizon vs Model Horizon)?

What about the often overlooked time horizon? If some of your portfolios are short term by nature, looking at an estimation of risk based on 12-month time horizon would not be particularly useful. What if you have a long-term investment horizon, but you want to understand your short term risk exposures? There are certainly legitimate and very good reasons to use models calibrated for different time horizons; just make sure that you understand the limitations and applications of such models before making a decision about which model is right for you.

In the end, I think we need to keep sight of something we already know, but often push to the background: all risk models are estimates based on some simplifying assumptions. One model is not going to be ideal for all purposes and situations. We need to do our best to align the model assumptions with our view of the world and our reasons for analyzing risk in the first place. If I care about measuring sensitivity to short-term market volatility, then I should not expect a model with a long horizon to provide meaningful insights.

The ultimate goal should be to fit models to our purposes. In many cases this may mean using multiple risk models.

Where do you need to see risk?

At this point, we have identified (from our list of stakeholders) the people/groups that care about risk, but as I have mentioned, the degree to which these stakeholders care can certainly vary significantly across individuals and groups. My goal here is to help you think about communicating what you know in a way that is meaningful to the end consumer of the information.

I have seen numerous investment managers who already have a risk system in place follow a business model where a small team of risk professionals analyze, generate, and communicate risk results for everyone else in the firm. While there is no denying that having a risk team is still a great idea (who better to set up standards, objectively monitor risk, and communicate results then people who specialize in exactly this type of analysis?), there are definitely some good reasons to make this information more accessible throughout a investment firm. For example, technology and software have improved dramatically since risk systems first came into use, which should allow for much broader and more efficient distribution of risk analytics across a firm. It still may fall to a special risk team to monitor and manage risk across portfolios, teams, etc., but there is certainly no reason for others who might be interested in portfolio risk (e.g., PMs, CIOs, Analysts) to have little or no access to the same information on a regular or ad hoc basis. In other words, there is no technological reason for limiting access to risk data to a single person or team.

At the end of the day, there will likely always be a need to have some risk information “pushed” throughout many firms, but the ability for entirely separate, independent groups to dynamically “pull” data throughout an organization should become standard practice as time goes on. If nothing else, it should allow:

Better integration of risk within the investment process (e.g., if fund managers can monitor their own risk, they should be able manage their portfolios in such a way that they can better justify investment decisions from a risk/return perspective)
Improved dialog amongst teams (e.g., if a risk team, board, or CIO is able to monitor risk across portfolios, they can ask meaningful questions to PMs before risk becomes a problem)
More frequent and timely dissemination of risk results internally and externally

The main takeaway from this section is the need to understand the means and options by which you can communicate analytics across the firm when you are working with your risk model vendor. Make sure that you have scalable solution that can grow as your needs grow and change.

I will be wrapping this subject up in the next installment.

Don't miss the next post in this series. Receive new blogs by e-mail.

Considerations when implementing a risk management system, part 1

Andrew Kovacs — 2010-06-23T18:16:27Z

For many people, transitioning from one year to the next is often a time of change. Consequently, I suspect that many firms and risk professionals may at this very moment be contemplating making changes to their current risk systems. Over the years, I have had a fair bit of experience aiding and abetting others who have been tasked with making decisions regarding risk systems, so I want to distill some of this experience into guidelines that may be useful should you find yourself in such a position this New Year.

Disclaimer: As a FactSet employee I am obviously biased in my views of risk providers, but my goal here will be to be present as objective a set of guidelines as possible. I have gone through this process enough times to know that the better prepared an investment management firm is (as it relates to risk systems decisions) the better off everyone involved in the process is as well.

Over the next two or three posts, I will address these questions:

Who are the Stakeholders?
Why do you need risk?
What are your options?
Where do you need to see risk?
How should you implement your solution?
When should this take place?

Who are the Stakeholders?

This is arguably one of the most important questions to answer early since in many ways it will dictate the answers to remaining questions. Common stakeholders include:

CIO
Board
Risk Manager/Risk Team
Portfolio Managers
Performance/Reporting Team
Marketing Team

Obviously the ways stakeholders will use risk data can vary dramatically across users/teams, and a good model and system should be able to provide relevant results to each type of end consumer in a meaningful way. For example, a Risk Team may care about the aggregated risk across a collection of portfolios and how each portfolio contributes to their overall equity risk exposure, whereas a Portfolio Manager within the same firm may want to determine which securities are contributing to his/her portfolio active risk.

I often think about the question of stakeholders from the perspective of how a particular stakeholder may actually think about or look at risk: at a summary level, detail level, or both.

In addition to these three vantage points, there is another dimension to risk that is often overlooked, and that is comprehension. By this I mean that the above stakeholders can usually be grouped into an additional two categories:

Those who understand and care about the interpretation of risk results
Those who do not (but this does not necessarily mean they are not a stakeholder)

Why is it important to understand these distinctions? If you care about details and comprehension, you probably need a different set of tools compared to someone who only cares about summary level information and is not bothered about the meaning of the results. An example of a risk stakeholder who needs results but doesn’t necessarily need to truly understand what the results mean would be reporting and marketing groups, many of whom need to report these numbers but are not required to really understand their meaning in depth.

Why do you need risk?

Choosing a risk model/system can be hard; why are you doing this to yourself?

Less than ideal stand alone reasons:

Need to be able to tick this box RFPs
Board reporting
Seems like everyone else has one

Some good reasons:

Determine if risks are aligned with expectations
Measure/manage risk across portfolios, products, asset classes in a meaningful way
Understand unintended exposures to factors in the market

This is probably obvious and in many ways the "why" should be an easy question to answer, but if you are part of the group looking at implementing a risk system, you need to have a good grasp as to why you are doing this in the first place. If it is simply to tick the box in an RFP or to include tracking error in some large amorphous report with little thought or emphasis on the meaning, then you are doing it for the wrong reasons and it is only a matter of time before you realize you are spending a lot and getting a little.

So my advice is to take the time to really understand why risk is important to your firm and how risk measurement and management can enhance your firm’s core competencies. There is nothing wrong with including the stand alone reasons I listed above as part of the overall answer to "why," so long as they are not the only reasons.

I will pick this up again soon with the next question in my list. Happy New Year!

To be continued. . .

Don't miss the next post in this series. Receive new blogs by e-mail.

My predictions for the risk landscape of 2010

Sean T. Carr — 2010-06-23T18:16:27Z

With the New Year I have no doubt you are also all looking to welcome in some changes, perhaps in the form of resolutions such as stopping smoking, losing a bit of weight, drinking less, spending more time with the family, more exercise, etc. Now we all know that past performance is no indicator of future performance, but on a personal level (and with the weight and exercise resolutions firmly in mind) I can easily see myself slipping back to the old ways all too soon and, therefore, am not going to go down the line of making these sort of pledges.

I decided instead that I might look forward to the year ahead and throw a few thoughts out there about how we may see the risk landscape change. There has been no new regulation as of yet from the SEC, FSA, or others, but I believe it is safe to say that the nature of any coming will not be of the "report less often and in less detail" variety.

Rationale: We have moved far from the old mantra "all risk is bad," but will regulation come in demanding justification any and all risks taken?

There is a level of risk inherent in any financial area, whether it comes from being exposed to a "split-strike conversion" strategy or just investing in a high yielding Icelandic bank account, so it is not the fact that there is risk per se that needs to justified, but that the risks being taken are reflected in the expected returns balanced with a complete picture of what the risks are. I hope that any new regulation in this area is focused on the explanatory angle rather than the justification one.

Responsibility: Will regulators look to firms to appoint individuals to take signatory control over portfolio risk levels in some kind of extension of the compliance departments, or will there be further encouragement towards a group responsibility?

Personally, I believe that we will see a move by firms themselves towards increased education, looking to ensure that everyone who has any interaction with a fund is aware of the risk characteristics of that fund. E.g., analysts providing recommendations for conservative, blue chip strategies should not be encouraging short term, alternative, emerging market exposures; portfolio construction tolerances should be set considering historic and comparative peer limitations; factsheets will report deeper variance contribution than just the top 10 holdings, etc.

Reporting: Will there be a push towards a "new" measure now that the "flaws of VaR" have been exposed?

Portfolio risk cannot be distilled down to a single number. Indeed some people would describe risk as more of a landscape than a point, but will providing a whole multitude of numbers really increase people’s understanding of risks and in what magnitude they exist? I have seen in a few clients already an acceptance that it will be necessary to embrace multiple methodologies that include different horizons (historical periods as well as both short term and long term ex-ante forecasts), different assumptions (normal vs. fat-tail, Stress Testing, Monte Carlo vs parametric, etc.) as well as different risk measures (Tracking Error, VaR, CVar, Expected Tail Loss). Making all of this available in an accessible, timely, and potentially interactive manner is a challenge that they are already looking to surmount.

These are just three areas that I wanted to comment on, and must (for legal reasons no doubt), stress that all comments made here are the views of me alone and do not in any way reflect those of FactSet Research Systems or FactSet Europe Ltd. But I encourage you all to reply directly or through the comments section below laying out your own thoughts to what possible changes to the risk landscape 2010 might bring.

To receive new posts by e-mail, subscribe to this blog.

FactSet's Taking Risk blog

Our most anticipated sessions at GARP's 13th Annual Risk Management Conference

Eurozone risk: Managing risk in uncertain times

Management strategies: Reflections on fundamental vs. quant strategies during and after the credit crisis

Top 5 Blog Posts of 2011

Currency risk forecasting in uncertain times

What Olympus teaches us about corporate governance and risk

Five easy steps to fixing the ratings agencies (Part 2)

II. Accreditation

III. Accountability

IV. Special Rules for Special Vehicles

V. External Awareness

Conclusion

Five easy steps to fixing the ratings agencies (Part 1)

Regulatory Fixes?

I. Consistency and Transparency

Is Greece uniting or dividing the rest of Europe?

* We used a European, daily-updated, short-horizon, statistical factor model.

Financial Blogs We're Reading

Sovereign Yield Risk: This time it's (not) different

Will Operation Twist help the housing market?

If it's not another recession, might it be the apocalypse? (Part 2)

If it's not another recession, might it be the apocalypse? (Part 1)

Beware Germans offering Greek gifts

Similarities and Parallels: Have we just come full circle?

Empirical durations: A tool to gauge mortgage price sensitivity

A retrospective on U.S. debts, and the logic of ceilings (Part 2)

A retrospective on U.S. debts, and the logic of ceilings (Part 1)

ETFs: The next bubble? (Part 2)

ETFs: The next bubble? (Part 1)

From News Corp to national crises, don't judge market volatility by the headlines

New Quants: More Reflections on Adapting

Talking risk with model providers: Recap from the FactSet European Symposium

What is the true size of China's monetary might?

Is the world running out of commodities?

Getting a picture of your portfolio holdings, from any angle

Defining "Quant" After the Credit Crisis

Three minutes on Ben Graham Was a Quant

The Value of Updating Risk Models Daily

Long-term Investment Approaches Cannot Ignore Short-term Volatility

Is Diversifying in Asia Dead?

Balanced Risk Explained

Our Most Anticipated Sessions at GARP's Annual Risk Management Convention

Capturing Euro-Sovereign Default Correlation through Contingent Claims Analysis

Vlog: What's the equivalent of idiosyncratic risk for fixed income?

Risk and Reward: Risk Budgeting and Cross-Asset Correlations

Taking Risk's Top Five Posts of 2010

What area of risk management presents firms with their most acute challenges?

What’s your risk? Single name security exposure analysis

Global risk models that aren’t skewed by asynchronous trading

Vlog: How does FactSet account for interest rate moves in its Value at Risk forecast?

Notes from the road: FactSet Balanced Risk hits Asia

The U.S. elections one week out: What's in store for equity and fixed income markets?

Coming soon: Systematic single-name exposure analysis

Why Default Correlation Matters (Part 2)

Why default correlation matters (Part 1)

Are there really any differences in the risk model providers? Asia-Pacific edition

How skillful a manager is Paul the Octopus?

Whose fault is it when the market crashes. . . and so does your portfolio?

Stock-picking set for a strong return, or is BP dirtying the waters for all?

Hedging Against Deflation

Is it a small world after all?

APT, Axioma, Barra, Northfield, and R-Squared respond to six key questions

What's more important: Country or Sector decisions?

Live webcast: What is the true cost of your portfolio analytics service?

Risk model providers respond. Part 1, do regional models still matter in the global economy?

The Abacus Deal: Could investors have seen what was coming?

Reverse Stress Testing: What are the chances.....

Client webcast: Five Efficient Ideas for Portfolio Analysis

The hidden risks of Chinese ETFs

Follow up: The difference a daily risk model makes

The difference a daily risk model makes

Introducing Betamax: A new measure of covariance stationarity (Part 1)

Introducing Betamax: A new measure of covariance stationarity (Part 2)

Hindsight is 20/20 - Or is it?

The reports of my death are greatly exaggerated, take 2

Don't miss FactSet's upcoming Risk events

A quick VIX?

Notes from the IPARM conference in Hong Kong, part 3

The illusion of stability, part 2: What fuels bubbles?

Long-term Investment Approaches Cannot Ignore
Short-term Volatility