No serious economist thinks higher dividend and cap gains taxes are efficient ways to raise revenue. Why not limit deductions for high earners instead?

 Friday's weak employment report reminds us anew of the flagging U.S. economic recovery. While the Obama administration discusses additional stimulus packages, Treasury Secretary Tim Geithner is arguing that we should roll back key elements of the Bush tax cuts passed in 2001 and 2003. The administration is particularly skeptical about the benefits of today's lower rates on dividends and capital gains.

The tax on dividends, for example, is currently 15%, but it could increase to as high as 39.6% if the 2001 and 2003 tax cuts expire. On top of this, a new 3.8% tax on investment incomes for high-income earners begins in 2013 to help pay for ObamaCare. The administration's arguments for higher taxes on capital center on fairness and the need for deficit reduction.

These arguments are seriously mistaken. The relationship between investment, capital and wages is such that workers are better off if capital is not taxed at all.

Think of the economy as a pie split among workers, savers and the government, with the government's slice fixed. The savers' slice will equal the after-tax return on each unit of the capital stock, and what's left goes to workers as after-tax wages. The fairness advocates in effect claim that low tax rates on dividends and capital gains increase the share of the pie that goes to high-income savers. But the low tax rates increase the absolute size of the workers' slice by making the entire pie bigger. That's because low tax rates encourage capital accumulation, productivity and wage growth.

To understand how this works, it's worthwhile to return to President Bush's proposal, in January 2003, to substantially reduce the double taxation of corporate income by eliminating investor-level taxes on dividends. (The actual law reduced tax rates on dividends and capital gains to 15%.)

There are at least four channels through which Mr. Bush's tax reform (proposed and passed) raised the long-run productive capacity of the economy—that is, increased the size of the pie. First, since lower taxes mean higher returns to investors, those investors allocate more funds to corporate capital. Corporations can raise capital for investment more cheaply. As a result, the nation's capital stock and output increase.

Second, reducing or eliminating the differential tax treatment between corporate and noncorporate investments means that investment flows are not channeled artificially by tax considerations and the overall productivity of the economy increases.

Third, lowering or eliminating taxes on capital mitigates distortions in our financial structure. Prior to 2003, equity financing was disadvantaged relative to debt financing, with taxes levied twice, at the corporate level and again at the investor level. Because interest payments to debt holders are deductible at the corporate level, debt financing was taxed only once, at the investor level. This system contributed to over-reliance on debt financing. The 2003 tax cuts reduced this bias substantially. Nonfinancial companies went into the recent crisis with lower leverage as a result, a very good thing.

Fourth, low taxes on dividends encourage firms with few growth opportunities to distribute the funds to shareholders. Those shareholders could then reinvest funds in other, more innovative and productive ventures—another very good thing for economic efficiency and growth.

Putting together the effects of greater capital accumulation and improved capital allocation, I estimated at the Council of Economic Advisers that, despite slightly higher interest rates caused by an increase in government debt, the president's 2003 proposal would raise real GDP permanently by about $75 billion annually.

How do my predictions look in retrospect?

Recent research has highlighted the predicted effects of the dividend tax cut on the economy's efficiency in allocating capital. Raj Chetty of Harvard and Emmanuel Saez of Berkeley concluded in the American Economic Review that the 2003 tax cuts led to a large increase in dividend payouts, particularly for slow-growth firms, improving the efficiency of capital allocation and investment.

One can, of course, make the argument that we should put the whole tax system on the table to debate fundamental tax reform. I am all for that. But I know of no proposals by economists for fundamental income or consumption tax reform that do not include low or zero tax rates on dividends and capital gains. Indeed, "tax reform" centers on permanently lowering marginal tax rates while broadening the tax base to secure adequate revenues. Any reform plan that does not maintain or reduce marginal tax rates—permanently—is moving in the wrong direction.

If the Obama administration's goal were truly fairness, it could propose an increase in the average tax rate on higher-income earners without raising marginal rates—for example, by limiting deductions. Does the Treasury really believe that raising dividend and capital gains taxes addresses its fairness concerns at the lowest cost in terms of reduced economic activity?

Deficit reduction is a legitimate object of concern. But if this concern is the dominant one, I am aware of no serious analysis that would claim smaller costs to the economy—in lost output and foregone economic growth—of raising capital income taxes as opposed to increasing other taxes or limiting deductions or reducing federal spending.

If President Obama is interested in promoting growth now and in the future, he should commit to retaining the low tax rates Congress passed in 2003.