In Sunday's New York Times, as part of its Sunday Dialogue, there was a discussion of whether capital gains should be preferentially taxed. Unfortunately, I don't think the discussion really identified the important issues. And I'm always frustrated by the public discourse about capital gains taxes. So, here's a summary of the best arguments pro and con in 500 or so words.
The problems with taxing capital gains preferentially are pretty easy to identify. The capital gains preference means that unearned income is taxed at lower rates than earned income, a result that many believe is unfair. The preference also requires the IRS to police the line between capital gains and ordinary income, and it encourages taxpayers to in engage in transactions that turn ordinary income into capitals (or capital losses into ordinary losses). Finally, an administrable line between capital gain and ordinary income can lead to bizarre results where items that should be ordinary income are mis-characterized as capital gains (see, e.g., carried interest).
So, what are the best arguments in favor of the preference? The preference is a result of the realization rule, which is the Achilles heel of the income tax. Investment gains are taxed (and investment losses deductible) only upon a realization event (most commonly a sale). Until then, the built-in gains and losses are ignored for tax purposes. Such a realization-based system is not a very good way to track increases and decreases in wealth (consider Mark Zuckerberg), but the other option (mark-to-market) is thought to be too difficult and costly to administer. Because of the realization rule, taxpayers have a tax incentive to hold onto their winners while selling their losers, thereby deferring tax. The incentive to hold onto appreciated investments is a tax-created friction that impairs the ability of capital to move to its most productive uses. This lock-in effect means that a taxpayer may hold onto Investment A even though Investment B is better because the tax toll charge for selling A is too high. The capital gain preference reduces this toll charge.
The realization rule also requires that limitations on investment losses must be curtailed. Otherwise, taxpayers with large, diversified portfolios would be able to sell their losers (while holding onto their winners) and avoid tax on their non-investment income like salary. To prevent this, the tax code allows deductions for capital losses only to offset the taxpayer's capital gains (plus a measly $3,000 against ordinary income). But this capital loss limitation creates its own problems. If a taxpayer is not very diversified, or if the overall market declines substantially, it's quite possible that a taxpayer will actually have overall net losses. In such a case, the capital loss limitation is unnecessary and merely adds injury (no deduction) to insult (a real overall investment loss). The prospect of this can discourage appropriate investment risk-taking because the government could, like a bad betting partner, participate in your gains if you win but not your losses if you lose. The capital gains preference reduces the impact of this assymetry.
Other common arguments in favor of the capital gains preference, such as that it offsets inflation or the problem of bunching multiple years worth of income into one year, are not as persuasive as the ones above. Inflation rates are not very high and tax rates are pretty flat, making inflation and bunching concerns less significant than they have been in the past. Even if they were significant, a more narrowly tailored solution (e.g., indexing basis or income averaging) could easily be devised.