Sunday, July 27, 2008

Our Value Investing Backtest: Part 1

Saj and I believe in value investing. First, its hard to ignore the results of great value investors such as Warren Buffett, Walter Schloss, Francis Chou and others. Additionally, we found the academic paper by Fama and French compelling. Fama and French found that small cap stocks and stocks with a higher book value to price (low P/B) have better returns than the market as a whole. We performed our own independent research with North American securities that reproduced the small cap and low P/B effects but also demonstrated a low P/E effect. In our study, small cap, low P/B and low P/E quartiles of NA stocks outperformed the market on average during time period between 1995 and 2006.

What we really wanted to know was whether we ourselves could best the performance of the low P/E quartile stocks from our previous study by using primarily value investing techniques we learned from George Athanassakos at the Richard Ivey school of business (tweaked a bit). To answer this question, we did full valuations on Toronto stock exchange low P/E stocks on the randomly selected years 1998-2000, 2003 and 2006. We randomly selected these years using an excel rand() function for years between 1995-2006 . Our sample dataset had been pruned down using various constraints in order to keep the exercise manageable but yet relevant! Since the purpose in this post is to share our results, I will defer a full explanation of how we filtered down to the working set in another post.

Our approach involved doing a complete intrinsic business valuation for each of the 55 stocks in our dataset. Saj and I did our own independent valuations using the same framework and then would then discuss (mostly in a civil way), each valuation to ensure we had similar valuations. After completing all the valuations we would find the price data and compare to our entry prices. If our entry price on a stock exceeded the current stock price, we would record it as a buy, otherwise it was not a buy for us but remained in the low P/E quartile basket.

Assuming a holding period of only 1 year and accounting for dividend payments, our selective buys yielded an average compound return of 29.1%! The low P/E basket of stocks averaged results of 20.6% during the same period. We found these results very encouraging and are currently valuing securities in (primarily) North America for the purposes of establishing an investment fund this year.

No comments: