GMO, a big asset allocator in the U.S., has said for some time that U.S. quality stocks are relatively cheap. As of July-end, they're projecting 2.4% annualized real returns for U.S. quality stocks and negative 1.4% returns for U.S. large cap stocks over the next seven years. We can test their claim by using the simple but powerful Gordon growth model:
r = y + g,
where r is total real return, y is yield, and g is real per-share dividend growth in perpetuity. This model tells us that returns are a function of present yield (technically, present yield using next year's dividend) and the expected per-share growth of that dividend. Unlike GMO's model, it assumes that yield, or valuation, remains constant. The model is useless for short-run forecasting. Valuation changes can exert enormous influence on total return, even over decade-long spans. A doubling of yield (implying a halving of price) can drag down 10-year annualized total return by almost 7%. This model is ill-suited for most individual equities, which have distinct growth profiles and finite lives, but it's reasonable for aggregate markets, which earn a relatively constant share of national output, and perhaps ultra-safe, predictable, slow-growth stocks.
The trick with the model is coming up with reasonable forecasts for per-share dividend growth in perpetuity. The market's per-share dividend growth has averaged 1%-2% real since the 1880s but has averaged 2.8% real since 1976-end thanks to increasing use of share buybacks in lieu of dividends, rising margins, falling tax burdens, the declining power of labor, and perhaps better management. Quality stocks, as defined by the MSCI USA Quality Index, have averaged 4.6% real dividend growth over this period, almost 2 percentage points more than the broad market. (As an aside, I'm using the personal consumption expenditures deflator as my price index, which reduces inflation relative to the consumer price index by 0.56 percentage points annualized over this period.) MSCI defines quality stocks the same way GMO does: low earnings variability, low debt/equity ratio, and high return on equity.
.png)
A naive approach is to use the historical dividend growth of the quality index to project future growth. But we know from the longer U.S. sample that dividends and earnings grew abnormally fast over this period. Moreover, it would imply investors should be willing to pay almost any price for quality stocks today--an absurdity. As of September-end, the MSCI Quality Index yields 1.7% and the MSCI USA Index yields 2%. Plugging in our historical growth estimates, we get 1.7% + 4.6% = 6.3% expected real return for quality and 2% + 2.8% = 4.8% expected return for the market. Assuming quality stocks aren't riskier, investors should be willing to bid up quality stocks to the point where their expected return equals the market's. This would mean a 0.2% yield on quality stocks, or a 750% increase from today's prices.
.png)
Something is wrong here: our model, our assumptions, or both. One assumption that's clearly wrong is using historical per-share dividend growth of quality stocks to project future growth in perpetuity. Almost by definition, a back-tested equity strategy that shows outperformance will show superior dividend growth. No unsuccessful back-test ends up as an investment product, so the outcomes we see are biased to favorable outcomes and not representative of future outcomes. I find it somewhat suspicious the index begins around the nadir of quality stock valuations, right after the Nifty Fifty crash. The financial statement data needed to back-test the strategy begin in the 1960s at the latest. Unfortunately, the degree of bias is impossible to determine with precision without knowledge of the number of different back-tests attempted. And even if a back-test isn't too biased, quality stocks are now believed by increasing numbers of investors to outperform, surely lowering the strategy's expected returns.
Another reason past dividend growth might be overstated is the tendency for quality stocks to more aggressively repurchase shares than the typical firm. Over time, valuations of all stocks have risen, reducing the growth boost afforded by share buybacks.
Despite reasons for skepticism, I think the quality effect is real. Various quality measures have been found to produce superior returns worldwide and over long periods. I just don't believe that we can expect almost 5% real dividend growth from them forevermore.
Say, grow per-share dividends by 1%-2% and quality stocks by 2%-3% over a full business cycle and after inflation. Plugging these values into the growth model suggests U.S. stocks overall are priced to return around 3%-4% and quality stocks around 4%-5% real. Keep in mind that even over 10-year periods, valuation changes can account for the majority of the stock market's returns, so there are huge bands of uncertainty around these forecasts. Another way to look at relative valuations is to look at the yield difference between the market and quality stocks. At 0.3%, it's near the bottom of its historical range.
.png)
Quality's relative cheapness has three plausible, non-mutually exclusive explanations.
1) Investors are forecasting slower per-share dividend growth for quality stocks relative to the market. Since the end of the financial crisis, lower-quality firms have been quickly raising their dividends from extremely low levels. Financial institutions have had their dividends suppressed by regulators, and this may change.
2) Investors are expecting higher returns from quality stocks to compensate for their greater risk. Because they are faster-growing, quality stocks derive a greater portion of their present value from future cash flows than junk stocks, meaning quality stocks should be more sensitive to interest-rate changes. With ultralow rates, the market could be pricing in this risk. However, quality stocks have historically exhibited lower drawdowns, their dividends steadily grew right through the financial crisis, and they don't seem to be terribly interest-rate-sensitive.
3) Investors are mispricing quality stocks, not fully appreciating their strong and persistent pricing power. I find this story compelling. Most investors focus on how earnings surprises will evolve over the next few quarters or years, whereas quality stocks often realize their advantages over longer horizons.
