This column originally appeared on Real Money Pro at 8:33 a.m. EST on March 6.
NEW YORK (Real Money
) -- Is there a secret sauce to predict the market's return?
Naturally, the answer to this question by all individuals and professionals alike should be a resounding, "No!"
That doesn't stop nearly every brokerage firm, economist, strategist and avid market participant from attempting to mix one, though -- and that includes me in my fair market value
The truth is: You would probably have just as much accuracy as a psychic with only very basic understandings of the stock market.
The holy grail of any market participant is a model, a secret sauce that aids the decision-making process in a way that allows for the prevention of losses and exposes opportunities for gains.
As a fun, exhaustive and torturing exercise, we have attempted to create this holy grail model.
Out of 40 different indicators that we considered for inclusion in our model, we narrowed down the list to just seven. In an effort to predict monthly S&P 500
returns, we used the change in the following indices:
1. Leading Economic Indicator -- Total;
2. CRB Index (inverse);
3. Seabreeze proprietary Durable Goods Index (with one-month lag);
4. Trade-Weighted Dollar Index (inverse);
5. University of Michigan Consumer Sentiment;
6. U.S. Exports -- Monthly (including a one- and two-month lag); and
U.S. Investor Sentiment Index (with monthly lags through three months).
What might be surprising to most individuals in this industry is that it is not economic data that matters most; it is sentiment. The old saw "the data don't matter, until they do" hits the nail on the head.
In an effort to provide a little statistical evidence of the fact that monthly S&P 500 returns are immensely difficult to predict (if not impossible), with the help of Seabreeze analysts Nick Pollari and Kelley Hopkins, we have compiled the following information.
- Time frame: February 1995 through December 2012 -- monthly data, 215 observations.
- Total independent variables: 12 -- statistically significant variables at the 90% significance level: 11 (one-month lag of monthly U.S. exports is not statistically significant).
- All independent variables have positive coefficients with the exception of the CRB Index (inverse), providing for a model that has economic sense
- R-squared = 0.3966
- Adjusted r-squared = 0.3607
The bottom line: With the r-squared and adjusted r-squared of less than 0.5, a linear regression does not help us accurately predict the returns on the S&P 500.
Stated simply, as hard as professional and unprofessional investors might try, there is no secret sauce or model that can assist us in answering (with precision) the question of "How Now, Dow Jones?
EXCLUSIVE OFFER: Jim Cramer's Protégé, Dave Peltier, only buys Stocks Under $10 that he thinks could potentially double. See what he's trading today with a
14-day FREE pass.