Do investors treat stock dividends as part of total returns, or do they view them as a separate income stream? In their December 2016 paper entitled “The Dividend Disconnect”, Samuel Hartzmark and David Solomon investigate whether trading and pricing of stocks exhibit a “free dividend” fallacy (disregard for the fact that dividends directly debit stock price as paid). Specifically, they test whether investors: (1) consider both dividends and capital gains when evaluating stock performance; (2) view dividend stocks differently based on market conditions/competing sources of return; and, (3) reinvest dividends and capital gains differently. Using daily individual trader data during January 1991 through November 1996, quarterly institutional and mutual fund holdings data (SEC filings) during 1980 through 2015 and contemporaneous daily stock and stock index prices, return and dividend data, they find that:
- Correlation tests suggest that noisy prices make it difficult for investors who only occasionally check their portfolios to see the error of the dividend fallacy.
- Overall, when trading based on past performance, investors tend to focus on capital gains and underweight dividends. However, they tend to hold dividend stocks longer than non-dividend stocks, and the likelihood of selling dividend stocks depends less on past capital gains.
- Investor demand for dividend-paying stocks is higher when interest rates and recent market returns are lower, consistent with investors viewing dividends as a distinct income stream. In addition, demand is higher for stocks with stable and recently increased dividends. Investors buying dividend-paying stocks when demand is high sacrifice 2%-4% per year in expected returns.
- Institutions and mutual funds generally do not reinvest dividends into the stocks paying them. They instead purchase other stocks, leading to predictable market price pressure on days with large aggregate dividend payments.
- Specifically, a one standard deviation increase in aggregate daily dividend payments indicates a 0.2% boost in daily market return (compared to a 0.04% average daily market return).
- The upward price pressure is greater for non-dividend stocks than dividend stocks.
- In other words, it appears that most portfolio reformation rules do not use total returns.
In summary, evidence indicates that investors tend to overweight or underweight dividends when making particular portfolio decisions, at odds with modeling/tests that use total returns (implicitly assuming stock-by-stock reinvestment of dividends).
Cautions regarding findings include:
- Stock trading data for individuals are very stale (20-25 years old).
- The study does not address whether investors can meaningfully exploit findings.
