Why the Capital Gains Tax Rate Should Be Zero

Studies | Taxes

No. 245
Friday, August 31, 2001
by Bruce Bartlett


Is a Tax on Capital Gains a Double Taxation of Capital?

Figure I - Taxable Capital Gains versus Real Gains for Selected Years

Since the 1920s, theoretical arguments about capital gains taxation have principally revolved around efforts by self-styled tax reformers to get rid of the capital gains preference. In general, the reformers have relied on a definition of income developed by Robert M. Haig and Henry Simons to press for full taxation of capital gains as ordinary income. Known as the Haig-Simons definition of income, it consists of all consumption during the course of a year plus the change in net worth.35 Thus, under Haig-Simons, even unrealized capital gains would be subject to taxation (and theoretically, unrealized losses would be deductible).36

"Under the Haig-Simons definition of income, an increase in the value of an asset would be subject to taxation even if it's not sold."

It is hard to say why the Haig-Simons definition of income achieved its status as the "official" definition universally used by economists and tax lawyers. Part of it has to do with its simplicity and part to the vigor with which some of its adherents have promoted it as the only theoretically pure definition of income. The latter would include especially economists associated with the Brookings Institution, such as Joseph A. Pechman and Richard Goode.37

The Distinction between Income and Capital. The central problem with the Haig-Simons definition of income as it relates to capital gains is that it implicitly endorses the double taxation of capital. The fact is that capital gains only arise in the case of an income-producing asset (except, obviously, in the case of collectibles). The value of the asset is simply the discounted present value of the future flow of income associated with that asset (e.g., rent in the case of real estate, interest in the case of bonds and corporate profits in the case of stocks). Thus, if the income stream (rent, interest profits/dividends) is taxed, then any additional tax on the underlying asset (real estate, bonds, stocks) must necessarily constitute a double tax on the same income.

"Decreases in the value of an unsold asset would be deductible."

Of course, asset values rise and fall all the time with no change in income. But permanent changes only come about because of a permanent increase in income flows. As the great economist E.R.A. Seligman put it, "Capital is a capitalization not simply of present or actual income but of the present worth of all future anticipated incomes. There can be no permanent change in the value of the capital unless there is at least an anticipated change in future income."38 More recent analyses confirm Seligman's characterization. As Richard Kopcke of the Federal Reserve Bank of Boston recently wrote in the bank's economic journal:

According to common financial theories, an asset's price depends on its prospective income. In applying this description of asset prices, this article concludes that capital gains are not necessarily a proper element of taxable income separate from ordinary income, because these gains (or losses) reflect changes in the prospective income offered by assets. When this income is itself taxed fully, the asset's price, and any changes in its price, tend to reflect the tax burden on this income. In these circumstances, a capital gains tax imposes an additional levy beyond that on ordinary income, which effectively imposes a relatively high tax rate on any changes in an asset's earnings. Consequently, a levy on capital gains is a poor substitute for an income tax. Only changes in an asset's earnings generate taxable capital gains; the initial earnings would remain untaxed.39

"If the income stream of an asset is taxed, any additional tax on the asset is a double tax."

Double Taxation of Bonds. Perhaps the clearest case of the capital gains tax being a double tax relates to bonds. Bonds are issued at par with a fixed income flow that is determined by the market rate of interest at the time of issue. However, the price of the bond may rise and fall as market interest rates change over the life of the bond. Since the nominal income flow is by definition unchanged, it is hard to see how selling a bond at a profit ever creates income in the aggregate - any gain to the seller in terms of command over resources is exactly offset by a loss to the buyer. As Martin J. Bailey explains:

Additions to wealth that result from a fall in the real interest rate are not income. Permanent income does include accruals to wealth due to all other causes, because such accruals could be consumed without impairing the ability to go on consuming at the same level forever. In contrast, consuming additions to wealth that result from a fall in the interest rate would impair the ability to go on consuming at the same level. A change in wealth due to changes in the interest rate reflects a price change, not a tangible change in goods available.40

"Capital gains and losses are not counted in the calculation of GDP."

Effect of Capital Gains on National Income. Confirming the broad acceptance of the principle that capital gains do not constitute income is the fact that the National Income and Product Accounts of the United States, from which the gross domestic product is calculated, have never included capital gains as part of the nation's economic income. Nobel Prize-winning economist Simon Kuznets, father of the national income accounts, explains, "Capital gains and losses are not increments to or drafts upon the heap of goods produced by the economic system for consumption or for stock destined for future use, and hence they should be excluded."41 In 1998, $414.2 billion of capital gains included in adjusted gross income for tax purposes were excluded from personal income in the national income accounts.42

Neither Haig nor Simons ever addressed this argument. In truth, their rationalization for taxation of capital gains rested more on ideological grounds than scientific analysis. As Haig once put it, "an income tax which would allow capital gains to escape unscathed would, in this country at least, be an ethical monstrosity."43 Simons took the same basic view. "The main and decisive case for inclusion of capital gains rests on the fact that equity among individuals is impossible under an income tax which disregards such items of gain and loss," he wrote.44

Figure II - Taxable Capital Gains versus Real Gains

Thus, for all its presumed scientific precision, it turns out that the Haig-Simons definition of income is nothing more than an opinion based on ideological preferences for a more egalitarian society. In any event, the U.S. tax system does not remotely correspond to a pure Haig-Simons tax base. If it did, there would, for example, be no corporate income tax, which even Simons and other Haig-Simons adherents concede is a pure double tax when dividends are also taxed.45

Thus, one can argue that while a pure Haig-Simons approach might be an improvement - it would at least eliminate the lock-in effect, since the tax would no longer depend on whether an asset was held or sold - it makes no sense to argue that Haig-Simons demands full taxation of capital gains while ignoring all the other exceptions to a comprehensive tax base in our current tax system.46 If supporters of eliminating the preference for capital gains wish to cite Haig-Simons in support of their position, they must also be willing to accept all the rest of the definition as well, which would, among other things, require taxation of unrealized gains, taxation of imputed income (such as the rent one receives for living in one's own home) and abolition of the corporate income tax. One cannot pick and choose.

"Both Haig and Simons agreed all gains should be adjusted for inflation before being taxed...[and] Simons conceeded that most capital gains are 'largely fictitious'."

Adjusting Capital Gains for Inflation. Finally, regarding capital gains, it is important to know that Haig and Simons both accepted the principle that all gains should be adjusted for inflation before being taxed. As Haig wrote in 1921:

If income is defined as the total accretion in one's economic strength between two points of time, as valued in terms of money, it is clear that his income will reflect every change in the value of money between those two points of time in so far as the items entered on the balance sheets at those times affect the computation. If the level of prices goes up 10 per cent the money value of my assets will ordinarily follow at a like rate. That particular increase in value does not really indicate an increase in my economic strength. My power to command economic goods and services has not increased, for the money-value of these goods and services has likewise increased.... If it were possible to modify the concept of taxable income so as to eliminate this variation it would certainly be desirable to do so.47

Simons conceded that most capital gains are "largely fictitious" once inflation is taken into account. In principle, he said, tax law should adjust gains and losses for changes in the price level. "Considerations of justice demand that changes in monetary conditions be taken into account in the measurement of gain and loss," Simons wrote in 1938.48 Numerous studies have documented that failure to index capital gains for inflation has subjected taxpayers to significantly higher taxes. [See Figure I].
In 1973, taxpayers realized nominal gains of $4.5 billion, but a real loss of $1 billion.49

  • In 1977, taxpayers paid taxes on $5.7 billion of nominal gains that actually represented a real loss of $3.5 billion.50
  • In 1981, taxpayers paid taxes on $17.7 billion of nominal gains that represented a real loss of $5 billion.51
  • In 1985, nominal gains amounted to $78.8 billion, but real gains came to just $63.5 billion.52
  • In 1993, nominal gains of $81.4 billion fell to $39.5 billion when adjusted for inflation. (Disregarding current limits on the deductibility of losses would produce a nominal gain of $71.9 billion and a real loss of $19.4 billion.)53
  • Over the entire period from 1946 to 1977, economist Robert Eisner found that there were no real capital gains whatsoever.
  • On balance, he found that households suffered a real loss of $231 billion on nominal gains of almost $3 trillion.54 [See Figure II.]

Read Article as PDF