Chapter 34
Volatility and Risk Measures—A Point of View

From an investment point of view, risk is about the potential to lose money. Other risk measures are just ways of trying to analyze the probability of losing money. Changes in the business world should cause people to be careful how they analyze risk because so many risk models are based on using data from the past. The linkages between this older data and the current markets are likely to be different and you need to take some time analyzing the current period to see what risk measures really apply.

While loss is the real risk in an investment, it is quite common to use volatility as a measure of risk. Changes in the value of an asset and the comparison of the value between two assets are often linked with volatility, but volatility is not a measure of loss. Volatility of returns is usually measured utilizing standard deviation of returns over a specific time period typically using daily, weekly or monthly returns. Much of the work on using volatility as a measure of risk was developed as part of modern portfolio and efficient market theory. This work focused on the returns of stocks. The research utilized data on liquid assets in which investors had symmetrical information flow and good record keeping and disclosures of transactions and volumes traded. Part of the key when using volatility as a measure of risk is that you have good data and that means the asset you are examining should be actively traded. Volatility in the valuation of an asset does tell you whether there are big or small changes in the market perception of the value of that asset and how often that happens. Higher levels of volatility does add risk because it increases uncertainty of how the market is valuing the asset. If the volatility is related to a corporate investment, it may signal things about the company’s operational stability, too. When you have shorter investment time frames, the volatility of valuations matters more as you will have fewer entry and exit point options, and if an asset has more volatile pricing, your time horizon probably has a bigger probability of aligning with a valuation that is a bad valuation for you. Volatility in the price of an asset indicates that the market place is less certain of an asset’s prospects, and volatility in a company’s operation metrics can indicate less predictability. However, volatility is not loss. One of the problems of just measuring volatility is that you could easily devise a scenario where you look at two investment assets over a twenty month time period. The first asset we will call asset Alpha. It might be down about 40% from the price you bought it and asset Bravo could be up 400% and not had a single month of negative returns, but asset Alpha could have a lower standard deviation of monthly returns. In this case, the lower volatility does not help to measure the greater level of risk. In theory, this occurs in part because the use of standard deviation assumes that the returns will be normally distributed, and nothing is always normal. It appears that asset Bravo’s returns are not normally distributed. There are many adjustments you could apply to the data by choosing selected time periods, looking at volatility only in periods of price declines or using statistical smoothing techniques, but blindly using volatility to measure risk can be misleading.

Analyzing periods of loss can be a valuable tool for understanding risk. However, measures of loss can be tricky because you need to analyze what caused the loss and how applicable it may be to current risks. If you take an average of periods when investment losses occurred over an extended time period, it might include months that had losses that were triggered by rising interest rates, a global collapse in commodity prices, a war, or trade tensions. One of these periods of “drawdowns” might have been wide-ranging and impacted every asset’s valuation or it might have been a period when drawdowns only impacted a specific commodity, region or industry. Just as you can measure volatility in various ways you can measure drawdowns in various ways. The time frame you choose is important, both the decision over how many years you want to analyze the data and then whether you care about the drawdowns during a day, week, month or year. You must make sure the data is applicable to the current time frame you are analyzing. You also can analyze different types of drawdowns such as the average drawdown during these periods or the maximum or minimum drawdown. It may not be only for periods of drawdowns. You may want to look at any time period where the returns were under a selected threshold. Comparing the frequency and the amplitude of the drawdowns between specific assets or asset classes can be a valuable risk assessment tool, though it is not as widely used as volatility.

Volatility has other uses than just measuring the risk of an investment return. You can look at the volatility of other factors and metrics. For example, when examining the operations of a company, you can look at the volatility of cash flows or the volatility in capital expenditures. A more consistent metric can give you a higher level of confidence in you projection scenarios for that company. Greater volatility could indicate greater business risk. You could also compare the volatility in a cash flow metric of a company to the volatility in the asset valuation of that company and see if they are reasonably correlated and if the business risks are reflected in valuation. There are numerous other possibilities. Similar functions could be constructed using macroeconomic data; you could do a comparison of the moves in volatility in food prices to the demand for out-of-home dining. These types of comparisons can give you some insights, but you must be sure to understand the context of the time periods you are examining relative to the time period that you are investing in and carefully analyze the logic of comparing the data.

It is often important to compare different asset classes and to decide how you wish to allocate your investments across them. For example, consider how you want to be invested in stocks, bonds and real estate. There is an endless body of writing on asset allocation, and the right allocation can shift over time and can vary greatly depending on the investor’s goals. Volatility is often part of the toolset used to compare asset classes. Much of the research that utilizes volatility is concentrated on very liquidly traded assets. This means the market is giving investors regular data that can be used to calculate gains and losses at least every day and frequently every second. These highly liquid assets are typically stocks and currencies in developed country markets and sovereign government bonds in the strongest credit quality countries. There are many assets that do not have nearly the same amount of trading volume or liquidity as equities, government bonds and currencies. These might include corporate bonds, real estate and tranches of securitized debt. Sometimes these markets are very inefficient and illiquid. An asset that is illiquid may have very low volatility, but that may be because there are no transactions for several days so that prices have not moved. A lack of transactions does not mean that there is less risk in the investment. As a matter of fact, you could make quite a case that if something rarely trades, then there are fewer valuation data points and possibly fewer buyers so it has greater risk even though it may have mathematically lower volatility. There are also structural issues that have to be considered when comparing various asset classes that can distort relative volatility. For example, much of the return in stocks price has historically come from appreciation or depreciation of the value of the stock and much less from the stream of income from dividends, whereas in corporate bonds the bulk of the return comes from the interest income stream. The structural issues are real and given these structural differences it is probably of limited value to compare volatilities of these types of asset classes. Differences in liquidity and trading volumes make comparing volatility across asset classes difficult to do consistently. It does not mean that this type of comparison does not add some value, but it should not be the sole tool used in comparing the risk and the risk reward ratios of various asset classes.

Risk and returns are intertwined; comparing the apparent risk you are taking to the potential reward is a common and valuable process. However, not all tools that are available to analyze return and risk fit all situations. These tools can be used to analyze an individual investment, but are more commonly used to analyze an entire portfolio or to compare asset classes in order to make decisions on allocations. The most popular measure to compare return versus risk for investments is, refreshingly one of the simplest, the Sharpe ratio. This ratio takes the return of a portfolio and subtracts the theoretical “risk-free” rate of return (which is usually assumed to be a government bond), and then divides it by the volatility of returns of the portfolio. The result is a simple return per unit of risk measure, assuming you accept volatility as a measure of risk. There are at least two important variations on this formula. The Treynor ratio changes the measure of risk, in the denominator. It uses beta as a measure of risk rather than overall volatility of returns. Beta measures how much of an investment return is due to changes in the market and how much is specific to that investment. A beta of 1 means the investment moves in line with the market; if it is 1.5, then it means it moves more. Making sure you utilize the right market from which to measure beta is important. The Sortino ratio adds loss into the equation and that is very valuable. In the Sortino ratio, the denominator utilizes the standard deviation of periods when the asset had negative returns, which is sometimes referred to as downside deviation. This tries to separate out good volatility from bad. It could be adjusted. Rather than just including periods of negative returns, it could utilize the volatility during any period in which a certain threshold return is not met. As an example, for a bond portfolio, this could be counting any period in which returns fell below the average coupon rate. The Sortino ratio may help when the data on the volatility of returns does not follow a normally distributed pattern. Which ratio you use will also depend on what exactly you are trying to measure. They can all be helpful, but all have flaws.

For any investor or decision maker, risk versus reward is a constant exercise. For security, investing the simplicity of the Sharpe ratio makes it very attractive to use, its simplicity has led to over use. Analysis should include an understanding of any idiosyncratic events that might have occurred during the period you are examining and how the constituents in the universe might have changed relative to the current time period. The distribution of returns should also be analyzed.

Risk and return are intertwined. Valuation changes are generally what drive returns, and valuation is driven by the perception of potential upside versus potential downside. Both of these perceptions can move significantly due to change in uncertainty.