George Athanassakos is a professor of finance and holds the Ben Graham Chair in Value Investing at the Ivey Business School, Western University. His most recent book is Value Investing: From Theory to Practice.
Cheaply priced stocks, based on metrics such as low price-to-book or price-to-earnings, tend to outperform more expensively priced stocks in the long run.
There are short periods, however, in which these so-called value stocks do underperform the pricier growth stocks.
That includes between 2012 and 2020. In Canada, for example, cheap stocks on average underperformed expensive stocks by 1.5 percentage points over that period. And in the United States, the underperformance of cheap stocks was much more pronounced.
This has led many academics to conclude that “value investing” is dead. It is important, however, to note that value investing as defined by academics is not what value investors do. Value investors do not buy all cheap stocks, as academics assume; they buy the best cheap stocks after careful due diligence and valuation.
Value stocks, relative to growth stocks, did not do as poorly in Canada as they did in the U.S. over the 2012-20 period because the Canadian market is dominated by firms in the financial industry as well as in the materials and energy sectors. Canada does not have the exponentially growing pharmaceutical and high-tech companies as exemplified by American firms such as Apple, Microsoft, Nvidia, Google and Amazon.
However, looking at aggregate statistics masks interesting findings about the performance of cheap versus expensive stocks. As always, the devil is in the details.
Here’s one example: In Canada, the lowest-priced value stocks still outperformed the lowest-priced growth stocks. That’s even though growth stocks, on aggregate, marginally outperformed value stocks.
In the U.S., cheap (value) stocks within the technology and non-technology industry still outperformed expensive (growth) stocks in their respective industries, even though growth stocks, on aggregate, severely outperformed value stocks.
To provide disaggregated evidence on the performance of Canadian cheap stocks, I obtained a sample of Canadian companies from the Compustat database for 2012 through 2020.
My sample excluded interlisted stocks. The goal was to provide insight into the Canadian experience, and Canadian stocks traded in the U.S. may show patterns closer to those of American firms. For each year of the sample, I ranked firms using price-to-book (P/B) from low to high and then divided the data into quartiles. Quartile 1 (Q1) included stocks with the lowest P/B firms (the value stocks) and Quartile 4 (Q4) included stocks with the highest P/B firms (the growth stocks). In addition, I computed a time series of non-overlapping returns for each stock within each quartile for the period in question.
I found that, on aggregate, cheap stocks underperformed the expensive stocks on average by 1.5 percentage points. However, disaggregated data by price shows more encouraging results. Low-priced cheap stocks outperformed low-priced expensive stocks on average by 4.5 percentage points. Low-priced stocks tend to be more obscure, are followed by fewer analysts, and are typically less liquid than high-priced stocks. The aggregate results were affected by the poor performance of high-priced cheap stocks, which underperformed high-priced expensive stocks on average by 2.5 percentage points.
In other words, active portfolio management focusing on cheap stocks can still be profitable in Canada for low-priced cheap stocks, irrespective of the overall performance of value versus growth stocks.
Similarly, recent research titled Resurrecting the Value Effect: The Role of Technology Stocks, by Ryan Lee of Baylor University, also concludes that despite the poor performance of cheap stocks, on aggregate, in the U.S., industry-specific cheap stocks still outperformed.
The author proposes separating tech versus non-tech when comparing cheap (low P/B) versus expensive (high P/B) stocks. He argues that the conventional approach of separating cheap from expensive stocks overlooks the unique characteristics of the technology sector and fails to account for how company-specific characteristics influence valuations. This way, the “value” effect can be better captured within each industry.
In the tech subsample, the author finds that on average, tech value stocks outperformed tech growth stocks by about 11 percentage points between 1991 and 2019. In the non-tech stocks subsample, value outperformed growth by about six percentage points. When looking at the 2001-19 period, when value severely underperformed growth in the aggregate sample, value beat growth by about eight percentage points in the tech subsample and by three percentage points in the non-tech subsample. Both excess returns were statistically different from zero.
As a result, systematically misclassifying all tech stocks into growth stocks because of their high P/B biases the value-versus-growth effect. Similarly, misclassifying all low-priced stocks as value because of their obscurity characteristics also biases the value-versus-growth effect.