The Bush Capital Gains Tax Cut after Four Years: More Growth, More Investment, More Revenues

Policy Reports | Taxes

No. 307
Thursday, January 10, 2008
by Stephen Moore and Tyler Grimm

The Lock-in Effect of Capital Gains Taxes

The major explanation for the lower tax collections at higher tax rates is the lock-in effect described earlier.16 Even opponents of a capital gains tax cut generally concede that there is a lock-in effect.  The Congressional Budget Office recently stated, “There is strong evidence that realizations of capital gains decline when tax rates on gains are increased.”17 Figure V shows the inverse relationship between capital gains realizations and capital gains tax rates from 1954 to 2005.

“Capital gains ‘locked in’ due to high tax rates were realized due to tax rate cuts.”

Figure VI illustrates the lock-in effect differently.  It shows, for 1954 through 1990, the ratio of total capital gains earned each year to the amount of those gains that was realized.18 The figure shows that when the capital gains tax rate is low, the ratio of unrealized capital gains falls (that is, investors are more likely to sell their assets).  After the increases in the capital gains taxes in the late 1960s and early 1970s, the unrealized capital gains ratio rose from about 35 percent to almost 60 percent.  After the 1978 and 1981 tax cuts, the 60 percent ratio tumbled to a 40-year low in 1986 of 20 percent.  By 2000, the ratio was back up to 45 percent.19

“More ‘locked in’ assets were sold after rate cuts.”

So again we ask:  Why is it that revenue estimators continue to overstate revenue losses and underestimate the unlocking effect from a lower capital gains tax rate?  Alan Reynolds, chief economist with the Cato Institute, explains the consistently misguided forecasts.  Reynolds notes,

All of the many studies of the revenue effect of the capital gains tax simply assume that there is no effect at all on the price of stocks and bonds, and also no effect on business investment or real GNP.  That is, the revenue estimators ignore the most important effect.20

Ratio of Available Capital Gains Realizations, and Capital Gains Tax Rate, 1954-1994


If the government revenue forecasters were to incorporate even modest economic growth responses into their static models and appropriately estimate the lock-in effect, the computers would spew out substantially different conclusions.  That was the finding of a recent analysis by economist Martin Feldstein of the National Bureau of Economic Research on the sensitivity of government models to changes in economic growth.  Feldstein concludes that if:

the improved incentives for saving, investment and entrepreneurship were to increase the annual growth rate of GNP [over five years] by even a microscopically small four one-hundredths of 1 percent — for example, from the CBO’s estimate of an average 2.44 percent real GNP growth per year to 2.48 percent — the additional tax revenue would be about $5 billion a year and would turn CBO’s estimated revenue loss into a revenue gain.  In short, the potential economic gains from a capital gains [tax] reduction are substantial and the potential revenue loss is doubtful at best...The slightest improvement in economic performance would be enough to turn [CBO’s] revenue loss into a revenue gain.21

“Government revenue estimates ignore the incentive effects of tax cuts.”

In sum, if one accepts the notion that a capital gains tax cut promotes economic growth (as, once again, the evidence suggests was the case after the 2003 rate cut), then revenue losses from lower capital gains taxes will be much lower than anticipated.  In some cases the lower capital gains tax generates more revenues.

Read Article as PDF