Using the Past to Predict the Future
Friday, September 10, 2010 at 01:51PM How many times have you looked at a chart and agonized over which way the index or stock was going to go next? Well, many traders are doing that right now with the S&P 500. As shown in the chart below, after a sharp decline in late April and early May, the S&P 500 has essentially been range bound for the last three months. The $64K question now is, what's next?
Since most technical analysts believe that chart patterns tend to repeat themselves, what if there was a period where the S&P 500 exhibited a chart pattern nearly identical to the current one. Would the market's performance following that period provide a guide as to what to expect in the future? Using some quantitative analysis, we went back and compared the last six months in the S&P 500 to every other six month period since 1928. After comparing all of the other periods, we found that the one stretch of time that most closely resembles the last six months is the period from November 1959 through May 1960. As shown in the chart, the two periods have very similar patterns.
So if history does continue to repeat itself, what does the future hold? Looking at the year following the 1959/60 example, the S&P 500 ultimately traded more than 20% higher in the 12 months from May 1960 through May 1961. However, this was only after the index drifted lower for an additional six percent over six months before reaching its ultimate bottom in the Fall of 1960. Now, this isn't meant to imply that the S&P 500 will continue to follow the pattern from the 1959 to 1960 period, but it is interesting to note some similarities. Back then the S&P 500 had just emerged from a recession less than two years earlier (April 1958) and wasn't far from entering another one in April 1960. Today, we are just over one year removed from what most investors consider the end of the Great Recession (June 2009) and in what is at least a slowdown in economic activity. While the majority of investors still believe we will avoid the double dip recession this time around, the 1959 to 1960 example suggests that even if we do go back into a recession (as we did then), a new bear market is not necessarily a sure thing.
Want actionable advice from Bespoke? Subscribe to Bespoke Premium today.
Reader Comments (6)
An interesting one - Given the midterm elections and the price action in the midcaps, the analog that looks most similar to me is 2006 - check out MDY then and now - not perfect, but quite close
Another approach that would be worth looking at is finding the best match in price action without deviating from the Mar-Sep period. In other words, deemphasize the importance of a near perfect match in price action and instead find something that is reasonably close. Without accounting for the seasonality issues of the stock market, the relevance of the above comparison seems a bit lacking (Nov-May has a way different profile than Mar-Sep). Anything fairly close to the same price action from Mar-Sep in prior years...now that's something worth considering. Even better, finding similar price action from prior years which also happen to be election years. Again, this is just another consideration and is not intended to say focusing on strictly price action is not interesting. Thank you for the effort. Love the site. Keep up the good work.
It is interesting to note the uptrend started in October.
This is like trying to match patterns on a roulette wheel. The same patterns are bound to emerge in different time frames but subsequent spins are still roughly independent. Yes, there is autocorrelation among stock prices, but there could also be autocorrelation within a dealer's spins. The problem is that if the dealer switches unexpectedly (say Greece Spain and Austria default on their debt) then historical comparisons are useless.
Did you "backtest" this analysis method and try it for periods prior to current so you could put a confidence interval around the results?
ETF, INDONESIA.
8-10-10 BUY (EIDO) ON THE FIRST CLOSE ABOVE 26.30
CHECK THE BLOG - http://etf.typepad.com/blog/
It is important to note that although the stock market showed a rise during the time you outline here, as well as a flat to slight rise throughout the entire 60's, that people actually lost money due to loss of purchasing power of the dollar. So, even though it would look as if a person would have made 21% return, it may not actually be the case. Just remember, in a system that is manipulated (i.e. not a free market) the numbers aren't always what they seem.