The Dividend-Clientele Controversy and the Tax Reform Act of 1986
Hearth, Douglas, Rimbey, James N., Quarterly Journal of Business and Economics
The Tax Reform Act of 1986 (P.L. 99-514) is categorized as the most extensive overhaul of the United States tax code in 30 years as well as the most fundamental reform of the federal tax structure in history (Joint Committee on Taxation, 1987). One of the major changes involves the taxation of capital gains relative to dividend income for individual investors. Prior to the Tax Reform Act, the top individual tax rate for all taxable ordinary income, including dividends, was 50 percent. On the other hand, because there was a 60 percent deduction for capital gains, the top effective individual tax rate for capital gains was only 20 percent. The Tax Reform Act eliminated the 60 percent deduction for capital gains; beginning in 1987, all taxable investment income (whether capital gains or dividends) was taxed at the same rate, up to a top rate of 28 percent.(1) Effectively, the tax rate on capital gains rose while the tax rate on dividends fell.
The relationship between dividends and taxes is the subject of much study and debate in the financial economics literature. Campbell and Beranek (1955) were among the first to observe that because the ex-dividend day stock price decline is less than the amount of the dividend on average, tax-induced dividend clienteles may exist. Elton and Gruber (1970) use dividends and price changes on ex-dividend days to document the tax clientele effect more rigorously, as well as to compute marginal tax rates.(2) They find that as the dividend yield increases, the ex-dividend day price drop increases relative to the dividend.
Miller and Scholes (1978) argue that the Elton and Gruber analysis does not take into account such factors as short-term trading activity and the impact of tax-exempt investors (such as pension funds). They suggest that the positive relationship between dividend yield and ex-dividend day relative price changes can be explained and demonstrate that individual investors need not pay more than the capital gains tax rate on cash dividend receipts.
Other researchers have attempted to replicate Elton and Gruber's findings with mixed results. Kalay (1982) removes two potential biases from the Elton and Gruber methodology and obtains a similar outcome.(3) Barclay (1987) examines the ex-dividend behavior of U.S. stocks before and following the implementation of the U.S. federal income tax in 1913. Using a variation of the Elton and Gruber methodology, his findings generally support the tax clientele hypothesis.(4)
By contrast, Lakonishok and Vermaelen (1983) in an investigation of the ex-dividend behavior of stocks listed on the Toronto stock exchange before and after significant changes in Canadian tax law find evidence that contradicts the tax clientele hypothesis. Skinner and Gilster (1990) further argue that the tax clientele effect identified by prior studies may be an industry yield effect. Using dividend and price data from 1980 through 1985 and the basic Elton and Gruber methodology, they find a strong positive relationship between EG statistics and dividend yield. They find that once utility companies (with their generally high dividend yields) are separated from nonutilities, the positive relationship between dividend yield and the EG statistic disappears; the relationships in both subsamples appear to be almost random.(5)
Two other recent studies examine the tax clientele hypothesis in light of the 1986 Tax Reform Act and reach opposite conclusions. Robin (1991) examines the impact of TRA on ex-dividend day rates of return, arguing that the tax clientele hypothesis would predict a drop in ex-dividend day rates of returns following implementation of the 1986 TRA. His findings generally support this argument. By contrast, Michaely (1991), using a variation of the Elton and Gruber methodology, finds that TRA had little impact on ex-dividend day behavior. This dichotomy of empirical findings leaves the issue of tax clienteles unresolved. …