The following is an excerpt from the research article "Fighting Greek Fire with Fire: Correlation, Volatility, and Truth" from Artemis Capital Management LLC.
40 years of Mean Reversion
click to enlarge
The degree of up and down days in the DJIA is at the most extreme level in recorded history, representing a pinnacle in an era of daily mean reversion. It only takes a casual observer of markets to see that the propensity for large-up days followed by large-down days seems particularly vicious in today's cycle. The excessive intra-day and day-to-day volatility is nauseating to professional and retail investors alike and multi-100 point swings in the DJIA are all too common.
On a macro-level, changes in the size and volatility of the money supply may be connected to a phenomenon called serial correlation drift. Modern derivative pricing theory is based on a conceptual idea that knowledge of past prices has no bearing on future returns (martingales). Despite this fact there is evidence that asset returns show signs of "serial correlation" whereby past returns are correlated (to some level) with future results. There are two forms of serial correlation 1) Negative serial correlation measures the propensity for today's return to be the opposite of yesterday's and rewards reversion to the mean strategies. For example, an asset that alternates between being +1% and -1% every day demonstrates perfect negative serial correlation. 2) Positive serial correlation is associated with consecutive days of asset price movement in the same direction and rewards trend following models. Both forms of serial correlation can occur in up or down trending markets.
We are going nowhere at the fastest pace in market history. The rolling one year serial correlation of daily lagged logarithmic returns in the DJIA reached a generational peak on May 25th, 1971. Less than three months later on August 15th, 1971 President Nixon surprised the international monetary system by cancelling the direct convertibility of the United States dollar to gold. After the "Nixon Shock", positive serial correlation in DJIA daily returns began a four decade decline. On August 11th, 2011 we reached the lowest levels of serial correlation in the 82 year history of the DJIA almost exactly 40 years to the day that Nixon abandoned the gold standard.
Is this a statistical coincidence? A random coin flip, whereby heads represent a +1% day on the market and tails a -1%, will also occasionally exhibit serial correlation extremes on a rolling one year basis. Despite this fact, this coin flip test, run over 100 years through 10,000 simulations shows nowhere near the serial correlation drift seen in the DJIA results and has much lower positive and negative extremes (see one sample simulation to the right).
If DJIA serial correlation drift is real, is it possible that monetary expansion has artificially rewarded stock market mean reversion strategies (such as value investing and buying on dips) for the past 40 years? If this is true does today represent the beginning of the new era of trend following and volatility?*
The hero of the last 40 years is Warren Buffett, who personifies mean reverting buy-and-hold value investing. However, if this data bears any truth, it may be possible that in the next 30 years a new generation will be worshipping his shadow in the form of some yet unknown trend-following commodities or bond trader.
Keep an open mind.
*A return to trending or momentum markets means the best way to make money is to buy asset classes that have performed the best recently. At least in the near-term (past 6 months), that means high quality bonds (TLT, LQD), commodities (GLD, SLV), technology (QQQ, XLK), and volatility itself (VXZ, VXX).
Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
Additional disclosure: Artemis Capital maintains a variety of long or short volatility positions in various indices (including potentially some mentioned in this article).
