Some simple but powerful investing models suggest owning European stocks.
Investors who "know" that European stocks are doomed are probably fooling themselves. University of Pennsylvania psychologist Philip Tetlock conducted a mammoth 20-year study on expert predictions. He found their performance modest, handily beaten by naive statistical models that used historical averages or simply extrapolated present conditions to the future. His findings add to an already considerable body of research showing the superiority of simple mechanical models over human judgment in prediction tasks.
Tetlock classified experts as either "hedgehogs" or "foxes" and compared their records. Hedgehogs are grand thinkers who wield a universal model of how the world works; foxes are ad-hoc and eclectic thinkers who see the world through many different lenses, never certain that any single one apprehends the truth. Hedgehogs scored the most impressive and memorable predictions owing to their high confidences in extreme predictions. But the diffident foxes were better overall forecasters. Unfortunately, the "experts" who appeared most frequently in the media tended to be extreme hedgehogs, confident without good reason.
These findings can be applied to financial markets, which are really prediction markets. How do we become more foxlike, or better yet, model-like in our investing behavior? Don't let confident-sounding pundits reassure you or scare you silly. Prefer robust, intuitive models to elaborate and vivid stories. And acknowledge that you probably don't know as much as you think you do. To make these lessons concrete, let's walk through three simple models that I believe say important and true things about the markets.
The Gordon Model
The Gordon dividend discount model answers the most important investing question: What kind of return can I reasonably expect from an investment? It can be elegantly stated as k = d + g, where k is return, d is yield, and g is the growth rate of the dividend. In other words, your expected return on an asset is its current yield plus the estimated growth in per-share dividends.
The model has strong theoretical and historical foundations. It acknowledges that an asset's intrinsic value is determined by its future cash flows. And it's done a good job in forecasting long-run stock market returns. It works because, in the long run, corporate earnings can grow only as fast as the economy does; otherwise, earnings would eventually make up the entire pie. A large economy like the United States seldom experiences big fluctuations, and has strong mean-reverting tendencies, so the stock market's per-share dividend growth over decades-long spans is bounded by a narrow range. Over 30-year rolling windows, the S&P 500's real dividend per share growth has averaged about 1% from 1900 to 2011, with a standard deviation of 1%.
Note that per-share dividend growth lagged real gross domestic product growth by 2% annualized. The gap reflects capitalism's creative destruction: Current shareholders are diluted away by new companies springing into being and established ones issuing new shares. (William Bernstein's "The Two-Percent Dilution" provides a longer treatment of this issue.)
Plugging in the historical per-share real dividend growth rate (1.0%) and adding an extra term for net share buybacks (0.5%), the model suggests that Europe's long-run after-inflation expected return is 5.5% annualized (4% yield plus 1.5% real per-share dividend growth). This beats out the U.S.' 3.5% expected return (2.0% yield plus 1.5% real per-share dividend growth). Sensibly, investors demand a higher expected return to invest in sclerotic, about-to-fly-apart Europe.
Most markets display mean reversion, the tendency for beaten-down assets to outperform over the long run. Markets tend to overreact to bad news, sending assets below fair value. Value strategies exploit this tendency by buying low and selling high. The most-studied value strategy is price/book sorting, popularized by Eugene Fama and Kenneth French. Each year it buys the stocks with the lowest price/book ratios and either short-sells or avoids those with the highest ratios. It has worked in nearly every equity market studied. In fact, you don’t even have to sort on price/book, so long as the measure is plausibly tied to a fundamental measure of value.