We repeated the same study for Canada using the top 50 companies within the S&P/TSX Composite index. Our focus on large companies is due to better data for this market segment, particularly earnings as there is greater analyst coverage for consensus earnings forecasts. It also helps reduce liquidity constraints that can quickly emerge for smaller companies in the TSX index. The test period for this study was December 31, 1997 to December 31, 2003. As with our US study, all returns are capital gains only with no transaction or management fees assessed.

Our study divided the top 50 companies of the TSX into quintile portfolios with 10 stocks in each. The quintile 1 portfolio contained the 10 companies with the greatest upside to model price chosen from the top 50 companies in the TSX, while quintile 5 contained the 10 companies with the least upside (greatest downside) to model price. The results show that the top quintile portfolios outperformed the lower quintile portfolios as well as the S&P/TSX Composite index.

When we measure the volatility of the quintile portfolios we notice that the fifth quintile has by far greatest volatility. Much of this is due to the relatively short time period covered by the study which also encompassed the rise and fall of companies in the technology sector, which found themselves in the lower quintile portfolios. The U-shaped relationship we found in our S&P study is less visible in this case although evidence of it still appears.

When we look at the return / risk profile of the portfolios, we find that the top quintile portfolios outperformed the lower quintile portfolios and the S&P/TSX Composite.