Over the next few weeks, I’m going to write about robustness. This is an idea that addresses risk in any type of planning. I have found that risk is particularly poorly addressed in market-related thinking. Investment plans discuss risk, but fail to control it. Particularly when it comes to asset allocation, the financial industry seems to have missed the boat on addressing the risks that are understood by common sense.
The subject of risk has been addressed in numerous ways in the past. For a full treatment of the history of risk-related thinking as it relates to investment markets, I refer you to Against the Gods, by Peter Bernstein. Risk has been equated with the likelihood of a desired outcome. If the probability of “winning” is lower, the risk is higher. This led Harry Markowitz to suggest that volatility could define, or at least stand as a proxy for, risk. When the volatility and correlation of two assets are known, volatility of a portfolio can be controlled.
Because the idea of controlling risk is like the Holy Grail of the investment industry, people seem to have latched onto this idea despite its failings. Modern Portfolio Theory was developed with optimised asset allocation. This was supposed to provide the maximal return for a given level of risk, or minimise risk for a given level of return. The problem lies in the fact that volatility does not equate with risk. First, investors want up-side volatility. Second, the fact that the exact value of an investment in the future cannot be known is not equivalent to the intuitive definition of risk: to possibility of loss. Third, the math that’s used to describe volatility, based on Gaussian (normal) statistics, does not accurately reflect the type of uncertainty found in financial markets, as can be seen when measures of volatility of a single asset vary over time.
As early as 1921, Frank Knight drew a distinction between uncertainty and risk. Uncertainty is a state where the probability of an outcome is unknowable. His thinking about uncertainty has not been widely adopted in academic theorising about financial markets, probably because it does not lend itself to neat mathematical solutions. However, it is a much nearer representation of reality. Using an inappropriate model has led the financial industry to underestimate risk, to make poor choices and to threaten the financial system of the western world in 1998 (LTCM) and 2008 (credit crisis).
If prevailing academic theories of risk and risk management are flawed, what is a better way to think about risk? Referring to the causes of the recent credit crisis of 2008, the cause was mainly due to leverage and inappropriate lending and borrowing. In the near collapse of the bond market in 1998, the cause was the over-use of leverage in the investments of Long-Term Capital Management. Their bet on the market eventually became profitable, but well after their investment-related debts became untenable. Volatility by itself doesn’t induce loss, but the use of borrowed capital makes it much more likely that a position will become untenable and force a sale at a loss.
If volatility isn’t a useful proxy measure for risk, could leverage be? Imagine a profitably run business. It will face varying economic environments in future. Some quarters will be more profitable than others, and some quarters it will experience negative profits (a loss). This situation is to be expected, but isn’t a problem as long as no one has a claim against the productive assets of the business. If the business continues to be unprofitable, management can sell assets and invest in more profitable areas or shift the focus of the business. But if their is a claim against the assets of the business, with a mortgage, a bond issue, or other lending, management may be forced to sell assets at an inopportune time, incurring a permanent loss.
Volatility is a poor proxy for financial risk. Financial markets are uncertain in a way that is not measurable. The Gaussian tools that we have to measure and compensate for normal variation does not apply in the financial realn, and behaving otherwise has lead to market crises in the past. Leverage is more closely related to the possibility of permanent loss suffered in a business context. The way to compensate for unmeasurable uncertainty is by increasing robustness.