It was not very long ago that investors had become incredibly complacent about market risk. Over recent years, I have written a number of articles that suggested that market volatility was in temporary decline, but that volatility was likely to return. The writing was on the wall for a substantial increase in volatility. This has obviously occurred, and market volatility as measured by VIX is at extremely high levels.
The greatest danger in low volatility environments is that investors tend to become increasingly aggressive because nothing bad has happened for a while. This factor was a substantial component in the surge in investment in emerging markets, even though these markets have historically been incredibly risky. Too many investors assumed that emerging market stocks would remain docile and well-behaved, despite considerable historical evidence of high risks in these markets.
Market volatility tends to go through extended periods of highs and lows. We are currently in a very high volatility state and we don’t know how long it will persist—but we do know that volatility tends to show considerable persistence. The current prices for even long-dated options on the S&P 500 (expiring at the end of 2010) imply volatility in the S&P 500 of 36%, more than twice the long-term average of 15%. With the enormous spike in market volatility over the several months through October, the trailing three-year volatility of the S&P 500 is now at 14.9%. Over the recent three years through much of 2007, the S&P 500 exhibited volatility at about half this level.
One of the most striking features of the markets in September and October of 2008 was the dramatic increase in correlations between asset classes—everything declined at the same time. This is a fairly common phenomenon when investors panic and sell everything. This indiscriminate selling leads to unique investing opportunities. Good companies and assets get sold off disproportionately, leading to substantial discounts that cannot be justified.
The recent market volatility is way above the long-term average, whereas we had volatility way below the long term average for several years prior to 2008. This variability is part of a natural cycle. When volatility was low, investors flocked to the most aggressive asset classes that they could find, like gold and emerging markets. Now that volatility is high, what will happen? What strategies make sense?
There is considerable discussion in the media and among portfolio managers of the need to use various forms of financial derivatives to manage the exposure of a portfolio to rare and extreme events. This is sometimes referred to as “tail insurance”—see for example the discussion by El-Erian in When Markets Collide (p. 279). In a nutshell, tail insurance aims to provide protection against very extreme events such as we have observed in October of 2008. In its simplest form, tail insurance could come in the form of purchasing put options on the major indices (like the S&P 500). Buying a put option on major market indexes is functionally similar to an insurance policy that protects the holder against massive declines in the index—hence the name of the strategy.
There are a couple of problems for investors who would like to avail themselves of this type of strategy. First and foremost, with market volatility far above its historical levels, purchasing long-dated put options today will be very expensive---the prices of options are largely determined by the volatility of the underlying (index, stock, etc.).
I believe that market put options are considerably over-priced at the current time. This sort of variability in the pricing of risk protection is common in insurance. The longer you go without a big natural disaster, the more reluctant people become to pay for insurance. Right after a big disaster, the cost of insurance goes up—simply because people perceive a greater risk when an extreme event is fresh in their minds (this is known as the availability heuristic).
The second problem with simply buying put options comes down to discipline. I believe that very few investors have the stomach to purchase options each year (you would need a rolling strategy) across a prolonged period in which the options have not paid out. Tail insurance, by its very nature, does not pay out often. It is certainly true that investors who purchased long-dated put options on the S&P 500 several years ago have benefited handsomely—but those same options are much more expensive today because of the high degree of fear in the market.
A new PIMCO fund (PGAPX) will use tail insurance practices as part of its broader strategy, with the aim of out-performing a generic 60% MSCI World Index / 40% Lehman Aggregate Bond Index benchmark. This is a fascinating time to launch such a fund. I would imagine that the fund will be quite light on the purchasing of options or other derivative strategies that require purchasing of protection against market volatility (there are many synthetic strategies for accomplishing this that do not require purchase of puts) simply because implied volatility is so high (i.e. these options are so expensive). On the other hand, the launch of this fund has been timed so that any long positions have the benefit of being purchased at a massive discount.
Making Risk Aversion Pay
From my perspective, there is a very simple narrative that suggests that investors can benefit from the current extreme risk aversion. Many investors have fled the market—we have seen massive outflows. When these investors re-enter the market, where will they go? There are two reasons why it is likely that they will tend to purchase low volatility equities. First, with volatility at high levels, it is simply rational for an investor who knows his or her risk tolerance to choose assets that are fairly conservative relative to historical levels. Second, there is the basic fear factor: investors who have been burned by high Beta or high leverage will likely tend to enter the market seeking safety. Investors who buy up these asset classes today exploit factor one and will benefit from factor two.
Many investors are simply highly shell shocked because they leveraged up during a period of low volatility and then were stunned when volatility reverted to its historical mean—and even more when volatility overshot the mean on the high side. Along with this, Betas and correlations have tended to increase.
There is a potential for a “flight to quality”, but quality means something very different than it did before the massive collapse of a range of well-known financial institutions with high ratings from Moody’s and S&P. The quality of earnings matters—a fact that many investors forgot, or perhaps never learned. The failures of the ratings agencies will naturally serve to make investors far more wary about how to determine whether a company’s stock should be considered “investment grade.”
Utilities seem a likely destination for many investors as they re-enter the market, not least because utilities are typically seen as “defensive” stocks—but they also have a range of features that make them attractive. The earnings from utilities are driven by a range of factors that are not common to many other parts of the S&P 500. Despite this, utilities have been sold off just as fast in recent months as the S&P 500—see the chart of IDU vs. SPY:
click to enlarge
This temporary increase in correlation provides the potential for substantial gains as investors become less frightened. The fundamentals of utilities have not changed, so this very high recent correlation in sell-off makes little sense.
Irrational Pricing of Risk
One way to get a sense of the extreme fear in the market is by looking at the implied volatility associated with certain stocks. Implied volatility is the volatility must be assumed in order to reconcile the prices of options and the current stock price (or index price). Using options prices, it is possible to examine the relative implied risk levels in individual stocks and indices.
As of November 3, the annualized implied volatility for options on JNJ expiring in Jan 2010 is 31% (source: ivolatility.com). Given JNJ’s historically low volatility and low Beta, this is a very high implied volatility from the options market. If I adjust the baseline market volatility in Quantext Portfolio Planner so that options on SPY match the long-dated options implied volatility, I get a forward looking volatility for JNJ of around 19%--much lower than where options are trading. If options on SPY have implied volatility of 37% and JNJ has implied volatility of 31% (not much lower than the market as a whole), something is awry.
A similar example can be seen for Southern Company (SO), one of my favorite examples of a low volatility stock. The Jan 2010 options on SO imply annualized volatility of 32%, while QPP projects 16% when we adjust the baseline market volatility to match options on SPY as before. Results like those for SO and JNJ can be found in a wide range of individual stocks and they suggest that investors have become indiscriminately risk averse.
The very high implied volatility on stocks like SO and JNJ is even more striking, given that these two stocks have weathered the market decline remarkably well (see chart above).
These results suggest that investors are too risk averse to buy up these companies, even at very attractive prices. This is a natural part of the market cycle---investors are so fearful that they believe that even stocks which have historically been very low risk not have higher-than-acceptable risk levels. Once investors process the reality that these stocks have weathered the decline very well, it is not unreasonable to think that these stocks will be how they re-enter the equities market.
Index Performance vs. Volatility
Quantext Portfolio Planner (QPP) was forecasting a higher volatility environment well before the volatility shock that we have experienced in 2008, as noted in a range of articles (see Section 1) and cited elsewhere. QPP also showed that the performance of a wide range of sectors and countries were negatively correlated to VIX—i.e. when market volatility goes up, the returns on many broad indices tend to go down, and vice versa. Our experience in 2008 has only strengthened these correlations---as volatility has gone up dramatically, returns on most major indices have been strongly negative. VIX reached extremely high levels in October of 2008—closing as high as 80 late in the month:
Even with substantial declines since its peak, a VIX of 80 is well above the long-term average and even way above its peak during the collapse of the dot-com bubble:
Given the negative correlation between asset class performance and VIX, a declining VIX bodes well for returns---and VIX is at such high levels that a decline is in the cards, though it may take some time. For VIX to be able to decline for a substantial period of time, it must start from a fairly high level—as it is now.
The fact that the high negative correlation between VIX and asset returns was such a good indicator ahead of the current decline has a lesson: plan portfolios with estimates of VIX and sensitivity to VIX in mind. More than two years ago, for example, I used QPP to analyze Berkshire Hathaway’s (BRK.A)top equity holdings. The analysis suggested that Berkshire’s portfolio would actually benefit from higher volatility conditions that QPP was forecasting. This projected benefit was not seen for the S&P 500 or other major indices. Since that analysis was published, Berkshire has indeed demonstrated its ability to cope with the massive surge in volatility:
Perhaps the most productive shift that could occur from the bear market of 2008 would be for investors so spend more time thinking about risk. Risk, as measured by volatility and other standard metrics, is a central concept that investors need to consider in their planning---but many investors have very little understanding of risk management. There are common metrics of risk and these can inform decisions. There are forward-looking models (like QPP) that can suggest whether or not a portfolio is reasonably well positioned to deal with a substantial change in volatility. A great deal of the pain that investors have suffered is simply due to the tendency of investors to use recent years as the benchmark for expectations. Over several very low volatility years (2005-2007), investors assumed that volatility was gone for good and cranked up their portfolio risk levels accordingly. This was an ill-informed decision, supported by no solid research and based on a disregard of history and even market-based risk measures (like implied volatility). If investors do nothing more than use reasonable risk analytics to estimate portfolio risk (rather than recent performance), they are likely to end up with portfolios that avoid unacceptable loss levels.
On a broader perspective, tracking market volatility and portfolio volatility in short-term and long-term horizons (which requires portfolio analysis tools like Quantext Portfolio Planner) provides substantial value. We will certainly hear a great deal more about tail insurance in the aftermath of 2008, but this is only one of the many ways to manage total portfolio risk. Investors need to ask questions about their portfolios that include risk as a consideration. If volatility has been low for some period, it is crucial to make sure that the portfolio remains capable of sustaining a volatility shock (an abrupt shift to higher volatility). If volatility has been very high for some period, investors will want to avoid a situation in which a reactive portfolio allocation is so conservative that it will miss much of the recovery. Along the way, the combination of portfolio analysis tools (like QPP) and market input (such as implied volatility) can provide the basis for improved decision making.