CAPITAL GAINS TAXES: Empirical Analysis 4

Rough approximations using the regression coefficient estimates provide some confidence that coefficients are reasonable. The coefficient on the dividend status variable can be used to estimate the market’s expectation of the eventual capital gains tax rate reduction. Let 8d equal the percentage of a dividend-paying stock’s value expected to be taxed as dividends, and 8n equal the percentage of non-dividend-paying stock’s value expected to be taxed as dividends. Let Ao equal the change in expected ordinary income tax rates, and Ac equal the change in the expected capital gains tax rate. The reduction in firm value for dividend-paying firms, relative to non-dividend-paying firms can be represented as: (8d – 8n) (A0 – AcAa), where Да is the change in the likelihood of a capital gains tax rate cut.

Assume the budget agreement did not affect expectations about future ordinary income tax rates (Ao = 0) and that the profits of the dividend-paying firms are taxed as dividends (8d = 1) and the profits of the non-dividend-paying firms are taxed as capital gains (8n = 0). Also assume that the budget agreement shifted expectations about the capital gains tax cut from impossible to certain (Да = 1). Using these parameters, the regression coefficient implies that at the time of the accord, the market anticipated future capital gains tax rates would fall by 6.24 percentage points. With a capital gains tax rate of 28 percent when the agreement was reached, the coefficient suggests that the anticipated future capital gains tax rate was approximately 22 percent.

If some profits of dividend firms are expected to face capital gains taxation, the regression coefficient implies a lower future capital gains tax rate. For example, if 8d is 0.5, the implied future capital gains tax rate is about 16 percent. These simple estimates seem reasonable. Recall immediate speculation in the business press following the accord indicated an eventual capital gains tax rate between 15 and 20 percent. TRA 97 reduced the capital gains rate to 20 percent.
Whoever thinks getting an affordable instant loan online is a utopia is simply wrong. We, as a lender, will be glad to prove you wrong if you believe the same, by offering you most reasonably priced loans that that will get you out of one financial trouble without getting you into another one. Apply with us now at
An alternative approach is to assume the budget agreement only shifted expectations about passage of the bill. That is, the market anticipated a 20 percent capital gains tax rate (Ac = 8) if legislation passed. Solving for Aa, the regression coefficient of -6.24 suggests that agreement increased the probability of a tax cut by 78 percent.

Control Variables

While suggestive, the average returns in the simple regression do not control for the possibility that other factors may have caused these differences in returns. To ensure that other factors are not driving the results, the regression is modified to include several control variables. The choice of control variables is somewhat ad hoc because theory is insufficiently rich to proscribe other effects of the tax rate change. The controls are:

Size = log of market value as of April 28, 1997;

Debt/Assets = total liabilities divided by total assets at year-end 1996;

Return on Assets = 1996 net income divided by year-end total assets;

Book/Market Ratio = book value divided by market value of equity at year-end 1996.

Size is designed to ensure that the results are not caused by smaller, non-dividend-paying firms outperforming larger, dividend-paying firms. Debt/Assets is included to control for differences in leverage across firms. Return on assets is designed to capture differences in profitability across firms. Book/Market ratio is included to control for differences in unrecognized assets, growth prospects and risk across firms.