Why value investing works in the long run

By George Athanassakos *  >>>  

Investors widely use the terms value stocks and growth stocks, but many do not know what they mean. Academic researchers sort stocks by price-to-earnings (PE), price-to-book (PB) or other valuation metrics from low to high and form a number of portfolios from the sorted stocks. They call the lowest PE stocks ‘value’ stocks and the highest PE stocks ‘growth’ stocks. While academics do not know which stocks from the value group value investors will eventually buy, they do know that value investors mostly choose stocks from the low PE group, the so-called value stocks, and avoid stocks from the high PE group, the so-called growth stocks. This is what I call the naive value investing approach.

But let’s look more closely at this naive definition of value investing and examine why naive value investing tends to focus on low PE or PB stocks.

These ratios are a function of the growth rate of earnings going forward. This relationship can be found in a mathematical formula derived from the equity valuation model taught at universities around the globe. Companies have low or high multiples because markets expect low or high earnings growth. However, the way growth comes into the mathematical formula implies that high multiple firms are expected to sustain high growth forever; vice versa for low multiple firms. That is the markets tend to be over-optimistic about growth for high multiple firms and over-pessimistic about growth for low multiple firms. Some of the companies for which investors are over-pessimistic indeed go bankrupt, but most do not and those that they do not tend to do much better in the future as the markets were excessively pessimistic.

The best way to see this in action is to examine what happens to the value premium (i.e., value minus growth stock returns) when companies announce positive vs. negative earnings surprises. A positive surprise is when the company reports earnings that exceed expectations by at least 10%. A negative surprise is when the company announces earnings that fall short of expectations by at least 10%. When looking at results for U.S. stocks for the period 1983-2018, we see that when companies announce positive earnings surprises the value premium is 2.34% three months following the earnings announcement and 6.69% over the following year. When companies announce negative surprises the value premium is 6.04% three months following the earnings announcement and 21.51% over the following year.

What could be the reason for this? The PE or PB ratios are a function of the company’s earnings going forward, with high PEs implying over optimism and low PEs implying over pessimism about the company’s earnings growth. If we are overoptimistic about growth and we get good (or better) numbers, we sort of expected it. But if we are over-optimistic about growth and we get bad numbers, we are very disappointed and react negatively. Similarly, if we are over-pessimistic about earnings growth and we get bad (or worse) numbers we sort of expected it. But if we get good numbers, we are exuberant and react positively.

The value premium is positive irrespective of whether the earnings surprise is positive or negative. However, the value premium is much larger when the earnings surprise is negative. Growth stocks react much more negatively vis-ˆ-vis value stocks when news is bad. For growth stocks investors expect good numbers. The good numbers are baked into stock prices and so there is little reaction to good (or better) news. But when numbers are bad there is a  downward  overreaction  leading to the sharp increase in the value premium as growth stocks react much more negatively to negative news than value stocks.

It is interesting to examine similar statistics from earlier periods produced by other studies, such as from the 70s and 80s. The value premium is mostly similar when news is good and bad, with the exception when news is bad for growth stocks in terms of returns over the year following the earnings announcement, in which case the value premium is about 10% vs. 21.51% in the 1983-2018 period. It seems that the sharp rise of fast-growing firms with fat growth expectations in recent years has exposed such stocks to sharp negative reactions when news is not as good as expected.

At the same time, these findings offer further support to the argument that the value premium is not driven by risk, as if risk was driving the results of na•ve value investing, we would expect value stocks to have been decimated during bad earnings numbers, but they do not. It is rather the growth stocks that are.

 * George Athanassakos is a professor of finance and holds the Ben Graham Chair in Value Investing at the Richard Ivey School of Business, University of Western Ontario.