David, re. <<why do all the experts and pundits I hear on
CNBC and other places almost all agree that market timing is too difficult
and, when attempted, generally lowers overall returns>>
I would think that they are making an empirical observation that is mostly true. I have read about studies showing that market timing newsletters, for example, are ineffective in predicting the direction of the market. Aside from these empirical findings, there are probably good fundamental reasons to dismiss most schemes, for example those that rely on guessing factors that are unguessable (which direction interest rates or oil prices will go) and the chartist schemes that assume that historical price data (head & shoulders etc.) can be used to predict future market directions.
My effort to use investor sentiment has at least a little fundamental basis (sentiment governs the rate of money flow into and out of the market) and it is something that can be back-tested empirically using historical data. Such a backtest shows that, during a tiny percentage of the time when the market sentiment is extremely bullish, you can safely get out of the market without losing any significant upside yet have a high likelihood of avoiding a severe downturn. One's compound return over 10 years is increased a significant percentage by moving into cash when the 2-week average of the AAII survey exceeds 55% bullishness. That's a finding based on backtesting, i.e., it is an empirical finding. Since the empirical test is based on only 10 years of data, including only so many corrections and rallies, it is certainly possible that it isn't good enough to rely on. Maybe in the long run using sentiment data will also be proven not to work. But so far, based on 10 years of data, it does improve one's overall return. |
