Archive for February 2009

Can investors successfully time the market?

In these uncertain and volatile times, investors often wonder if the stock market can be timed, such that low returns can be avoided and high returns can be realized. There is much evidence that markets indeed can be timed.  Various signals help predict market downturns so that investments can be taken out of the stock market and safely parked in cash or treasury bills.  When the signals indicate high future market returns, investments can be moved back into the stock market.  

There is a variety of signals that were shown to predict stock market movements.  For example, signals based on financial ratios, such as the price-to-earnings ratio (P/E), the book-to-market ratio (B/M), and the dividend yield (D/P); signals based on interest rates (both short- and long-term rates); signals based on the maturity spread (defined as the yield spread between long- and short-term treasury bonds/bills); signals based on the credit spread (the yield difference between BAA- and AAA-rated corporate bonds); signals based on the implied volatility index (VIX); and so on.  The intuition, for example, for the P/E-based signal is the following: when stock prices are high relative to earnings (the P/E ratio is high) stocks are considered to be overvalued and thus expected to earn low returns.  And vice versa: when price levels are low relative to earnings (the P/E ratio is low), stocks are cheap and are expected to appreciate and earn high returns.    

There has been plenty of empirical evidence presented that the above indicators work in predicting stock market movements and therefore can be used to successfully time the market. However, a recent paper by Andreas Neuhierl and Bernd Schlusche ( “Data Snooping and Market-Timing Rule Performance”) disputes these findings.   The authors convincingly show that much of the documented success of market timing strategies can be attributed to data snooping.  

In this context, the term “data snooping” simply means that previously-reported successful strategies were discovered by searching through many different parameters for the market timing indicators until an indicator that works has been found.   The “successful” indicator was thus found by pure luck (or by intensive and deliberate search) and is not guaranteed to work in the future.

After the authors correct for data-snooping biases, none of the market-timing strategies beat a buy-and-hold strategy.  A small exception is a strategy based on signals from several indicators at once, which only works for the period of 1981-1994.

And, of course, why would market-timing strategies work if any investor can take advantage of them?  By paying attention to market timing signals, investors bring prices in line with the fundamentals, and the indicators cease to be useful in predicting market movements.  It is entirely possible that there might be a strategy out there that has never been reported and it works precisely because only a select few know about it!