- One of the main arguments against raising capital income tax rates is that doing so discourages savings and investment and hinders economic growth.
- However, academic research on taxes and growth suggests that this argument has no real basis.
- And the primary alternatives to capital income taxation—labor income taxes and increased government borrowing—carry their own potentially adverse economic effects.
According to this view, raising taxes on capital income discourages savings and investment. A diminished flow of savings and investment causes, in turn, slower growth in the economy’s stock of capital goods, such as plant and equipment. And when the capital stock grows more slowly, so too do output, employment, productivity, and wages.
On the surface, the growth argument against capital income taxes seems clear and compelling. And many policymakers and pundits—on both sides of the aisle—appear to regard it as common sense.
A very different picture emerges, however, from the academic research on taxes and growth. Scholarly evidence on the growth argument against capital income taxation is mixed at best. Indeed, it would not be unreasonable to conclude, based on the best available theory and data, that the growth argument has no real basis.
The object of this policy brief is to lay out the relevant research on capital income taxes and growth. The first part of the brief discusses the growth effects of capital income taxes considered in isolation from other policies. Both domestic and international dimensions of the issue are reviewed. The second part of the brief concerns the growth effects of alternative policy measures that would be likely to go hand in hand with any reduction in capital income taxes—namely increased labor income taxation and increased government borrowing.
Domestic Savings and Investment
A key premise of the growth argument is that taxing capital income decreases savings and investment. The research on capital income taxes, however, casts doubt on this premise. Consider first the theory and then the data.
Even the simplest models of savings yield no clear prediction regarding the impact of capital income taxes on savings.
Imagine an individual deciding how much to save for retirement. When the government takes a portion of what is earned on savings, every dollar of planned future consumption requires more dollars of present savings. Thus, taxing the return to savings effectively raises the price of retirement consumption in terms of forgone current consumption.
It is fair to assume that this price increase reduces the individual’s planned level of retirement consumption. It is also fair to assume that, all else the same, it makes the individual worse off. But does it also reduce the individual’s retirement savings?
Savings is the not the same as desired future spending: desired future spending is what one ends up with; savings is what one puts away. The distinction matters because, while capital income taxes may decrease the amount that one desires to end up with, they increase the amount that one must put away to end up with any given amount.
Suppose, for example, that before a tax is imposed on interest earnings an individual desires to consume $110x in retirement. If the relevant interest rate is 10% he or she saves $100x for the purpose. If a 50% tax is imposed on such interest, the individual may decide to consume only $105x in retirement. Yet the individual will still save the same amount: $105x in retirement consumption requires $100 in savings when the after-tax interest rate is only 5%.
There are many ways to extend this very simple framework. One could account for other reasons to save besides retirement consumption, including bequests to heirs or self-insurance against injury or property loss. One could account for the borrowing (that is, the negative savings) of early-career taxpayers who expect to earn more, rather than less, in future years. One could also take into consideration the tendency for individuals to make decisions based on rules of thumb, rather than rational calculation.
But none of these complications sharpen the predictions of the theory. Indeed, most only serve to introduce additional sources of ambiguity that make it even harder to argue that higher taxes necessarily reduce individual savings.
Figure 1: Personal Savings Rate and Maximum Effective Rate on Long-Term Capital Gains
Drawing conclusions from the empirical research on capital income taxes is similarly challenging. This is partly a reflection of the theoretical ambiguities described above. It is also due to data limitations that make it difficult to isolate the impact of capital income taxes from the potential impact of myriad other factors.
Figure 1 shows the sixty-year time pattern of the personal savings rate (the fraction of after-tax personal income not consumed) and the highest marginal effective tax rate
on capital gains (accounting for deduction phase-outs, etc.). The graph reveals little discernible relationship between such tax rates and savings. Notice in particular that the personal savings rate appears to move with, not against, the capital gains rate until the mid 1980’s and from the mid-1990’s until the mid-2000’s.
Yet one must be very cautious about drawing conclusions merely from the relative ups and downs of these two lines. Other determinants of savings may have been changing at the same time as taxes. The effects of these other changes might mask a clear negative relationship between saving and taxation. It is possible, for instance, that the personal savings rate would have fallen even faster than it did in the early 1980’s if not for the simultaneous reduction in capital gains tax rates.
Several studies have attempted to control for the non-tax determinants of savings, so as to capture the true impact of taxation. Estimates in this literature vary substantially and are difficult to compare. But most academic commentators who have assessed this literature suggest that the takeaway is that taxes on capital income have essentially no impact on savings.
Some researchers have argued that the preceding group of studies provides misleading predictions regarding the impact of tax policy changes. The concern is that the measured responsiveness of savings to taxes is not a fixed fact relative to tax policy, but is rather determined in part by the tax regime in place at the time of measurement. According to this research, such measurements therefore cannot be used to predict the impact of new policies.
In response, a second set of studies attempts to estimate the underlying, policy-constant parameters of individual choice—from which the effect of policy changes may then be deduced. Most such research concerns one particular parameter, the “intertemporal elasticity of substitution.” This parameter measures how quickly an individual’s perceived value of additional future consumption increases as the individual’s quantity of future consumption decreases. The studies in this literature indicate that the increase is relatively rapid. Each percentage point decrease in the ratio of future to current consumption appears to increase the value of additional future consumption (in terms of current consumption) by 1 percentage point or more—significantly more in some studies.
Applied to the retirement savings example above, such estimates imply that individuals quickly put the brakes on any tax-induced reduction in planned future consumption. This, in turn, points toward the possibility that the reduction in desired future consumption is overshadowed by the need to save more per dollar of future consumption still desired. Such results thus indicate that savings does not decrease, and may actually increase, in the face of greater taxation of capital income.
International Investment Flows
Measures of global financial integration—such as the flows and stocks, gross and net, of foreign assets and liabilities—have increased several-fold since the 1980’s. It is thus increasingly difficult to adequately describe the growth impact of capital income taxation without assessing the international dimensions of the issue.
The problem with capital income taxes, one might assert, is not that over-taxed U.S. taxpayers will invest less, but that they will invest less in the U.S. This, it is argued, will have the same harmful effect on U.S. capital accumulation as an overall reduction in savings.
As a preliminary matter, it is important to keep in mind that investing abroad does not, in the first instance, allow a U.S. taxpayer to avoid U.S. tax on capital income. The U.S. generally taxes the income of U.S. citizens and residents from whatever source derived.
On the other hand, there are currently opportunities to effectively defer U.S. taxes by investing in foreign corporations that are located and operated so as to trigger little or no corporate level tax—U.S. or otherwise. In this case, the only tax paid, directly or indirectly, by the U.S. investor is the tax paid when and if the investor receives dividends or sells shares.
But while this feature of the current U.S. international tax regime is important and controversial, it has ambiguous policy implications with respect to setting the proper level of capital income taxation. The problem caused by this feature is that the taxation of (certain) capital income from foreign investment is preferred relative to the taxation of capital income from domestic investment. The solution, presumably, is to equalize the two tax treatments and thus eliminate the tax preference for foreign investments. Yet a reason to equate two tax treatments has no implication for the level at which the treatments are equated. Lowering the taxation of domestic investment income to meet the taxation of foreign investment income is but one way to generate parity. An alternative would be to raise the taxation of foreign investment income to meet the taxation of domestic investment income. The point is this: the argument that “U.S. investors will just go abroad” if U.S. capital income taxes are raised ignores the fact that tax rates on foreign investments can be changed too.
If outflows of U.S. investment dollars are part of the story, so too are inflows of foreign investment dollars.
In the simple version of the growth argument described in the introduction to this brief, all capital accumulation was assumed to come from domestic savings, or more to the point, savings whose return is reduced by U.S. capital income taxation. There is, of course, another source of funds for U.S. capital accumulation. This is lending and (non-active business) investment from abroad by those who are not subject to U.S. taxes, and whose behavior would not therefore be directly affected by any change in the U.S. taxation of capital income. For example, 20% of U.S. corporate bonds (by value) are held by foreigners, up from 16% in 2004.
Thus, even if it were true that raising capital income taxes reduces domestic savings, this does not imply a reduction in the stock of productive capital in the U.S. economy. Potentially, as domestic savings ebb, foreign savings flow.
Whether foreign investment is an adequate substitute for domestic savings depends on several factors.
The first factor concerns the extent to which non-U.S. taxed foreign investment seamlessly replaces domestic savings. One possibility is that foreign investment must be coaxed into any space left by receding domestic savings with the promise of higher returns. The necessity of paying higher returns would in turn choke off marginal, but formerly viable, investment projects.
Another possibility is that, capital being extremely mobile, a large pool of foreign capital rests in precarious balance on the precise level of the presently-offered U.S. return, and is poised to rush in and capture the benefit of any upward pressure on such returns caused by receding domestic savings. The result of this would be that the return would never have a chance to increase, and investment within the U.S. would remain at essentially the same level.
Many commentators believe that the latter story, though idealized, is a closer depiction of reality. This is partly supported by the fact that, as noted, measures of global financial integration have increased several-fold since the 1980’s. Net borrowing by the U.S. has increased even more dramatically.
A second factor to consider in assessing whether foreign investment can adequately substitute for domestic savings, centers around the fact that when capital is foreign-owned, the return to that capital flows abroad. The implication is that some portion of the benefits of U.S. growth accrues to foreigners. This raises two questions: How much actually flows abroad? And how much is too much?
At least in recent years, the answer to the “how much” question appears to be “very little.” While the foreign contribution to domestic growth has increased, the reward for that contribution has not. The yield on ten year Treasury bonds has fallen steadily over the last two decades and is now below the rate of inflation in many countries. In the view of the Federal Reserve Chairman the U.S. finds itself amid a “global savings glut.”
How much outflow of capital income is too much depends in part upon one’s underlying objective in promoting growth. In particular, those whose chief concern is jobs and wages are less likely to be concerned with capital owners’ nationality. Domestic capital accumulation increases employment, productivity, and worker compensation without regard to who owns the capital.
If one holds fixed the level of government spending, the chief alternatives to capital income taxation are two: labor income taxes and government borrowing.
Both of these alternatives figure large in the recent fiscal cliff deal. In addition to raising capital income taxes on high income taxpayers, that deal also raises the top rate on labor income from 35% to 39.6%. It also allows a temporary 2-percentage point reduction in the payroll tax rate to expire. Moreover, a .9 percentage point increase in the Medicare payroll tax for high income taxpayers, passed in 2010 as part of health care reform, comes online in 2013. Furthermore, the cliff deal increases the deficit by $4 trillion, according to CBO (compared to what spending and tax policy would have been in the absence of an agreement). Growth-based objections to the cliff deal’s capital income tax increases must also address the growth impact of alternatively steepening the increase in either labor income taxes or government borrowing.
Labor Income Taxes
There are stories to be told about the negative impact of labor income taxes on growth that seem no less theoretically compelling and are no less empirically grounded than the corresponding stories for capital income taxation.
Consider again the retirement savings example discussed above. Recall that capital income taxes reduced desired future consumption, but increased the amount that had to be saved for each dollar of future consumption still desired.
Like taxing capital income, taxing labor income plausibly reduces desired future consumption. In fact, with labor income taxation, there is an additional downward force. As with capital income taxes, labor income taxes raise the implicit price of future consumption—in terms of work hours, rather than in terms of forgone current consumption. Unlike capital income taxes, however, labor income taxes also reduce the after-tax value of individuals’ lifetime earning potential, leaving them less to allocate to all uses, including future consumption.
Furthermore, with labor income taxes, any given reduction in future consumption translates more readily into a reduction in current savings. When the tax is on labor income rather than capital income, there is no counteracting increase in the amount that must be saved for any given amount of future consumption. Adapting the numerical example presented above, suppose that it is an increase in wage tax, rather than interest tax, that causes desired retirement consumption to decrease from $110x to $105x. Increasing the wage tax leaves unchanged the amount that must be saved per dollar of future consumption. Consequently, less future spending always means less current savings. Here the individual saves only $95.45x rather than $100x.
It should also be noted that taxing labor income may also discourage education and training. In terms of economic principles, education and training may be regarded as another form of saving and investment, generating another stock of productive (human) capital that increases output and productivity. Furthermore, to tax labor income is, in part, to tax the return to human capital investment. Thus, whatever one believes to be the mechanism by which capital income taxes hamper the accumulation physical capital, it is worth asking whether this same mechanism applies to labor income taxes and the accumulation of human capital.
The government might also finance lower capital income taxes with additional government borrowing. Through a combination of pre-existing fiscal policies, policies designed to reverse the recent downturn, and the downturn itself, government debt held by the public now stands at over 70% of GDP, a magnitude not seen since the aftermath of WWII.
The impact of government borrowing on capital accumulation and growth is uncertain, but the consensus among academics appears to be that government borrowing is a hindrance.
This consensus rests largely on the view that government borrowing crowds out private investment. According to this view, when the government increases its borrowing, it adds to the demand for savings. This increased demand puts upward pressure on the return to investors that is paid not only by the government but also by private investment projects, which compete with the government for savings. The overall supply of savings—from domestic and international investors—may expand in response to higher rates of return. Yet it is possible that this increase, and then some, will be siphoned off by new government borrowing, so that investment in private projects decreases.
For reasons discussed above, the current danger of crowding out is reduced to the extent that there is presently a “global savings glut”.
Moreover, crowding out will also be less severe to the extent that individuals anticipate and incorporate into their decisions the long-term effects of government borrowing. Assuming that government borrowing must at some point be repaid, lower present taxes must be offset with higher future taxes. If individuals anticipate this, and care about the impact even as it befalls future generations, they may increase current savings in response to government borrowing in order to enable themselves or their heirs to pay higher future taxes. Such compensatory savings would limit the extent to which government borrowing crowds out private savings.
Taking account of these competing factors, the Congressional Budget Office (2012) runs through three “crowding out scenarios” in projecting the economic effects of budget policy. All three reflect CBO’s view (which it takes to be the consensus) that government borrowing reduces private investment. Specifically, the CBO posits that each dollar of additional government borrowing reduces domestic investment by 10, 30 or 50 cents.
On balance, academic research supports a skeptical view of the assertion that capital income taxation hinders economic growth. Even if offsetting changes in other policies are ignored, the growth effects of capital income tax preferences are uncertain. Compelling intuitions tend to melt away on close inspection, and the data tell no consistent story. When the negative growth effects of offsetting increases in labor income taxes or government borrowing are also taken into account, uncertainty begins to shade into doubt. Attempting to spur economic growth with tax preferences for capital income may be like trying to repair one side of the roof with shingles from the other.
About the Author
Chris William Sanchirico, PhD
Professor of Law, Business and Public Policy Penn Law School and the Wharton School
Chris William Sanchirico is the Samuel A. Blank Professor of Law, Business, and Public Policy at the University of Pennsylvania, and Co-Director of the University’s Center for Tax Law and Policy. Sanchirico, whose primary appointment is at Penn Law School, has a secondary appointment in Wharton’s Business Economics and Public Policy Department. This year, while on sabbatical leave, he is Visiting Scholar at the Urban-Brookings Tax Policy Center in Washington DC.
Sanchirico graduated from Princeton University, where he majored in the Woodrow Wilson School of Public and International Affairs. He earned his law degree and his PhD in economics from Yale University.
Sanchirico’s work on taxation employs a range of methodologies reflecting his background in both economics and law. His recent research on the tax treatment of private equity funds, for instance, analyzes subtle aspects of tax law doctrine. His study of work patterns and income tax progressivity applies modern statistical methods to complex datasets. And his contributions to the theory of optimal taxation use mathematical modeling to help refine policy discourse. Much of Sanchirico’s research is freely available at www.ssrn.com.
1 The growth argument, though salient and influential, is not the only serious argument against capital income taxation. Other important issues include: the potential for “lock-in,” whereby taxpayers unnaturally delay selling appreciated investments in order to defer taxation; the fact that the calculation of gain is not adjusted for inflation; and the fact that corporate investors are already taxed under the corporate income tax. Though a full analysis of capital income taxation would necessarily address these arguments, to keep this “brief” true to type, I do not consider them here. Suffice it to say that these alternative arguments—though not without merit—are not so strong as to render the growth argument superfluous.
3 Sandmo, Agnar, “The Effects of Taxation on Savings and Risk Taking”, in Handbook of Public Economics, Vol. 1, Elsevier Science B.V., (A.J. Auerbach and M. Feldstein, eds.) (1985); Bernheim, B. Douglas, “Taxation and Savings,” in Handbook of Public Economics, Vol. 3, Elsevier Science B.V. (A.J. Auerbach and M. Feldstein, eds.) (2002).
4 Lucas, Robert E., Jr., “Econometric Policy Evaluation: A Critique,” in The Phillips Curve and Labor Markets, Vol. 1, Carnegie-Rochester Conferences on Public Policy, Journal of Monetary Economics, Suppl., pp. 19-46 (1976).
5 That is, the intertemporal elasticity of substitution, which is the reciprocal of the ratio described in the text, is generally estimated to fall between zero and one. See Bernheim, “Taxation and Savings,” in Handbook of Public Economics, Vol. 3 (2002).
6 Kose, M. Ayhan, Eswar Prasad, Kenneth Rogoff, and Shang-Jin Wie, “Financial Globalization: A Reappraisal,” IMF Staff Papers, Vol. 56 (2009); Lane, Philip R. & Gian Maria Milesi-Ferretti ‘‘The External Wealth of Nations Mark II: Revised and Extended Estimates of Foreign Assets and Liabilities, 1970–2004,’’ IMF Working Paper (2006).
7 Darby, Joseph B. III & Kelsey Lemaster, “Double Irish More than Doubles the Tax Saving: Hybrid Structure Reduces Irish, U.S. and Worldwide Taxation,” Practical U.S./International Tax Strategies, Vol. 11, No. 9. Pp. 2-16 (2007).
8 Foreigners’ active business income from U.S. sources is generally taxed by the U.S. as if it were the income of a U.S. citizen. Moreover, dividends paid to foreign shareholders are generally subject to a 30% “withholding” tax. Bilateral tax treaties often lessen these tax burdens.
10 Kose, M. Ayhan, Eswar Prasad, Kenneth Rogoff, and Shang-Jin Wie, “Financial Globalization: A Reappraisal,” IMF Staff Papers, Vol. 56 (2009); Lane, Philip R. & Gian Maria Milesi-Ferretti ‘‘The External Wealth of Nations Mark II: Revised and Extended Estimates of Foreign Assets and Liabilities, 1970–2004,’’ IMF Working Paper (2006).
12 Some might argue that capital income tax reductions should be offset by cuts in government spending. But suggesting spending cuts as the means of financing capital income tax reductions begs the question of why such spending cuts are not being to used finance reductions in labor income taxes or government borrowing. The analysis in this brief considers alternative means of financing any given amount of government spending.