On Fat Tails

I’m reading an investment book that is arguing for market timing.  I’m not impressed with the line of argumentation so far.  I just finished a chapter where the authors pointed out that security price movements are more volatile that the normal distribution would admit.

This is a well known result, or at least it should be well-known.  What I hope to contribute to the discussion is why the tails are fat, and skewed negatively.  There is a famous saying in investments:

Cut your losses, and let your winners run

I regard this saying as vapid, because I have had so many investments where the price action was bad initially, but ended up being incredible investments.  I have also had companies stumble after prior gains, and persevere for greater gains.  Intelligent asset management does not react to the past, but analyzes future prospects, and looks at current margin of safety.

But imagine a situation where many parties have their plans, and they are all similar.  I’ll give a few examples:

  • Institutional investors decide in 1986 to follow the momentum, but be ready to sell if the momentum breaks.  They want upside, but want to protect the downside.
  • Japan was a total momentum market up through 1989, and the reverse thereafter.  Loose monetary policy was an aspect of that, as was a loss of fear, warrant speculation, etc.
  • Those investing in hot emerging markets in the mid-90s did not recognize valuations getting stretched, and the inability of the countries to maintain stimulative policies amid falling currencies.
  • The guys at LTCM were geniuses until they weren’t.  They had no idea of the risks they were taking.  They did not have an ecological view of investing.  Essentially, they thought liquidity was free, until the jaws of the trap snapped shut, and they died.  Taking a concentrated position is a risk, because the investing typically pushes up the price.  When you are so big in a position that you are affecting the market price, that is a bad place to be for two reasons: 1) if you sell, you drive down the price for future sales, and 2) you no longer know what the fair price would be if you weren’t there.
  • Aside from that with LTCM, their brokers mimicked their trades, accentuating the boom-bust, but the brokers had risk control desks that forced them to sell out losing trades, which further hurt LTCM.
  • Think about residential mortgage bonds in 1994.  So many players thought that they had mastered the modeling of prepayment risk only to find amid a Fed tightening cycle that many wanted to limit their interest rate risk as rates skyrocketed, fueling a self-reinforcing panic.
  • Consider tech stocks 1998-2000.  Momentum ran until the sheer weight of valuations, together with insolvencies, crushed the market as a whole, and tech stocks more.  Think of European financial institutions getting forced by regulators to kick out US stocks in September 2002, putting in the bottom.  Regulators almost always act too late, and exacerbate crises, but they should do that, because worse things would happen if they didn’t.  (Later = bigger crisis, Earlier = Some Type II errors, regulating where it was not needed).
  • Finally, consider the housing/banking crisis in the US 2005-2009.  People bought homes with a lot of debt financing, and short-dated debt financing.  Banks levered up to provide the financing.  Shallow credit analysis allowed banks to take on far more risk than they imagined.  It all ended in a trail of tears, with many personal, and not enough corporate bankruptcies, with the taxpayers footing the bill.

In each of these cases, you have correlated human behavior.  The greed of investors gives way to fear.

Now if you are thinking about Modern Portfolio Theory, where market players have perfect knowledge, this doesn’t make sense.  These crises should not happen.  But they happen all too regularly, and I will explain why.

Men are not greedy as much as they are envious.  This leads to mimicking behavior when things are going well.  Those not currently playing want a piece of the action, and so they imitate.

Modern Portfolio Theory implicitly assumes that market players don’t react to the actions of other market players, but that is false.  Most market players don’t think; they mimic.

That is what leads to fat tails, because when people move as a herd, you get dramatic price moves.  Because fear is a greater motivator than envy, that is why the big downward moves are almost always greater than the big upward moves.

Add into that the credit cycle, because gains on credit-sensitive bonds are small, but losses are huge when they occur.  The distribution of outcomes has a long left tail.

The main point here is that price movements are non-normal because market players act as a group.  Their behavior is correlated  on the downside, and to a lesser extent on the upside.

Among other things, this means Modern Portfolio Theory is wrong, and needs to be severely modified, or abandoned.  It also means that we need to watch the credit cycle, and speculative activity to get a sense of how committed the hot money is to risk assets.  Hot money follows trends.  Cold money estimates likely returns over a market cycle, and invests in the best ideas when they are out of favor.

I don’t think timing the market is easy.  I do think that fundamental investors have to look at whether they have a lot of opportunities, or few, and vary their safe assets opposite to opportunities.

So beware the fat tails — we haven’t had a lot of volatility recently.  Maybe we are due.