Some Thoughts on Risk (May 2011)

Risk of Overconfidence

Overconfidence can be another source of risk.  Successful investing requires humility. Whenever you buy a stock, you are willing to pay more than any other potential buyer who is looking at that stock.  This phenomenon is known as winner’s curse.  When there are many sophisticated potential buyers looking at a security, their reluctance to buy triggers warning signals. Although you may be hard-working and skilled, other competitors presumably have similar confidence in their own abilities.  Either you have to be significantly better in analyzing securities than all other potential buyers, or you need to be trading in securities that have not attracted the attention of other sophisticated traders.  A desire to avoid the winner’s curse might lead you to find comfort trading in obscure stocks in obscure markets.

An often overlooked risk is excessive confidence in forecasts that are based solely on statistical analysis.  If you torture data long enough, it will confess to anything.  Therefore, it is important to base your analysis of data on plausible stories.  Though these often take the form of mathematical models, they are still only stories.  Their plausibility depends, in part, on their logical consistency and based on realistic assumptions about human behavior.  you gain more confidence in the story if the story fits the data.  However, even if you are very confident in the ability of your models (stories) to explain the past, there is no guarantee that the same models will accurately predict the future.  The danger of financial firms hiring analysts with strong math skills is that they may confuse economics with physics.  Our understanding of financial markets is not nearly as good as our understanding of physical phenomena.  In quantum mechanics, probabilities are real; in financial markets, they are a measure of our ignorance.  Evidence is suggestive, never conclusive.  Data and anecdotes imperfectly describe a past reality.

Despite the common cliché, history never repeats itself.  Economic conditions are always changing.  In 2006 self-described experts on real estate markets were still predicting continued price increases in real estate, and were scoffing at the notion of a national decline in prices. They were basing their analysis on the observation that we had not experienced a national decline in real estate prices since the Great Depression.  However, during most of that time period, mortgages had 30 year durations with fixed interest payments and 20% down payments.  There was no interstate banking or securitization of mortgages and other loans, and throughout this period personal debt as a fraction of median per capita income was far lower than its value in 2007.  All in all, the historical evidence was of little relevance in understanding the conditions prevailing in 2007.

Even if circumstances arise that seem similar to those of previous periods, people would be aware of the outcomes when those circumstances prevailed and will alter their behavior accordingly.  Indeed, they will base their behavior not only on their knowledge of the previous outcome, but also on how they think everyone else will be changing their behavior.  None of this is to suggest that you should ignore data or the implications of rigorous models, just that you should use statistical analysis cautiously.  The alternative of “faith based” investing is hardly an improvement on data driven investing.  What I advocate is to think carefully about the fundamental value of securities and to think about how that value will be realized.  You can use historical data to guide your thinking and as a check on your analytic models.  This may not be the most exciting way to invest but it avoids the Scylla and Charybdis of either investing based on opaque statistical models or on gut instinct.

Macro-economic Risks

The other source of unknown risks is unanticipated macroeconomic events.  For instance, I didn’t anticipate the rebellion in Tunisia, the dissolution of the Soviet Union, or the collapse of AIG.  However, I did expect that unexpected events would happen.

Why I’m not as pessimistic as in 2006

One source of optimism is that the public’s perception of the problems facing the developed economies is getting closer to the reality of the magnitude of those problems.  Clearly, developed economies are facing serious fiscal problems and states and municipalities have huge
unfunded pension and health care costs.  I am pleased that these problems are finally receiving attention.  While it would be even better if the problems were being addressed, it is surely a good first step that they are noticed. 

I do not share the hysteria regarding quantitative easing (QE2).  I do not understand why the Fed’s purchase of longer dated government securities is decried as the end of civilization, but its routine purchase of short dated government securities is regarded as a non-event.  The distinction should not be the maturity date of the securities, but rather the size of the purchase given the circumstances.  For many years, the Japanese central bank has been expanding its money supply and entities of the Japanese government have been buying Japanese bonds – the result has been deflation.  (The Japanese also have a net debt to GDP ratio of somewhere north of 100%.  Interestingly, the amount above 100% depends on how one counts assets and liabilities held by entities controlled by the Japanese government and other quasi-governmental entities.)  Though I personally disagree with those who argue that high inflation is imminent and that the dollar is about to crash, I am fully cognizant of the fact that I do not possess a crystal ball.