Advocates of higher taxes traditionally justify their policies by the spending that they support, either directly by linking taxes to spending or indirectly by linking taxes to deficit reduction. In either case, when pressed they would begrudgingly acknowledge the harmful economic effects of higher taxes but argue that these effects are an acceptable price to pay for the benefits associated with the spending.

In recent years, some advocates have shifted to arguing that higher taxes are benign with respect to the economy and, in some circumstances, can actu­ally enhance economic performance. This raises a fundamental question: Do higher taxes lead to a stronger economy?

Beyond the year-in, year-out wrestling over public finances and spending, the pending calamity in Social Security and Medicare finances has made the eco­nomic effects of taxes a topic of great concern. These entitlement programs are vital to America's seniors but unaffordable and unsustainable in their current forms. The options are limited: Congress must pare benefit growth, raise taxes, or enact some combina­tion of the two--all in large amounts.

Not surprisingly, one typical liberal response to sustain Social Security and Medicare is to raise taxes. This is clear from today's policy and political debates, but this response tends to fall flat if higher taxes hurt the economy. If the traditional view that higher taxes are bad for the economy is correct, then this liberal response would effectively ask American taxpayers to pay twice to close the funding gap created by exces­sive entitlement program promises--once in higher taxes and again through lower wages from a weaker economy.

With both near-term and long-term pressures to raise taxes building, what does the evidence say regarding the effects of taxes on the economy? According to most economic traditions, higher taxes result in lower incomes and less economic output. In the Keynesian tradition, higher levels of taxes depress aggregate demand. In standard public finance, higher taxes permit more government spending, which means that fewer resources are available to the generally more productive private sector. In the microeconomic and supply-side tradi­tions, higher taxes lead to greater distortions in eco­nomic decision making, which generally results in poorer economic outcomes.

In contrast, the budding liberal tradition is that higher taxes are benign or possibly even beneficial. So which is it? Are higher taxes good or bad for the economy?

As discussed in this paper, the evidence very much supports the traditional view that higher taxes are bad for the economy. Taxes affect eco­nomic activity through many channels, however, and the evidence suggests one channel through which higher taxes can strengthen the economy, albeit very modestly. This channel relates higher taxes to lower interest rates and ultimately higher investment levels. Advocates of higher taxes some­times emphasize this "silver lining" channel but ignore the many other channels through which high tax levels have deleterious effects on the economy. The evidence strongly suggests that while the silver lining theory has some substance, it is ultimately threadbare, overwhelmed by the many dark eco­nomic clouds that arise from higher tax levels.

Evidence and Arguments on Aggregate Tax Levels

There are many ways to examine evidence as to whether higher taxes result in more or less eco­nomic activity. The following discussion considers the question from three perspectives: the historical record, the economic and revenue feedback effect of tax changes, and the validity of the silver lining the­ory of high tax levels.

The Post-World War II Record. In a recent study, Christina Romer and David Romer, profes­sors of economics at the University of California at Berkeley, examined significant tax changes and the ensuing economic performances during the post­war period.[1]

The level of taxation can change for many rea­sons. Reliably estimating the historical relationship between changes in tax policy and changes in eco­nomic performance requires filtering out episodes that may distort the results. For example, it is important to distinguish between changes in tax levels caused by changes in the economy and those caused by changes in the tax law. In other words, the direction of causality matters.

Rapid economic growth may cause tax receipts to soar, as happened at the end of the 1990s when tax receipts tied the all-time record of 20.9 percent of gross domestic product (GDP). Strong economic growth generating high tax levels can then lead to the spurious observation that high taxes caused the strong economy. The converse is also possible. As seen at the beginning of this decade, a recession will slow the growth in tax receipts or even cause overall tax levels to plummet. This could suggest that a lower level of taxes is associated in some causal sense with weaker economic performance. Includ­ing such episodes in the study would distort the results. Hence, it is important to exclude episodes when significant changes in tax levels were driven by economic changes.

Similarly, it is important to identify the motiva­tions for tax legislation and to exclude those intended as countercyclical policy. Well-timed tax cuts may prevent an economic downturn or reduce its depth and duration. If the tax cuts are enacted with sufficient timeliness and strength to short-cir­cuit a recession, then the national data will show no real signs of the recession that was averted. In such a case, effective pro-growth, countercyclical tax cuts will seem to have had no economic effect. Similarly, if tax changes were enacted during a recession, it would be difficult to determine whether normal economic growth resumed sooner or later than if there had been no tax changes.

The best test of how tax policy affects the econ­omy excludes those episodes when tax levels changed due to changes in economic performance or when tax policy was changed to restore or main­tain normal economic growth. This leaves two types of tax policy changes to include in the analysis. The first type arises when Congress changes tax policy and thus the level of taxation for reasons other than economic growth, such as to improve the perceived fairness of the tax system. Even though not moti­vated by economic considerations, such legislative tax changes nevertheless offer good tests of the gen­eral relationship between the level of tax collections and the economy.

The second type includes tax changes that were enacted with the expressed intent of improving eco­nomic performance beyond the near term. For example, Congress has sometimes cut taxes to strengthen the economy on a sustained basis, such as the 1981 Reagan tax cut and the 2001 and 2003 Bush tax cuts. Of course, such tax cuts are some­times also intended to address near-term economic weakness, but the authors concluded that there was a longer-term consideration that was sufficient to warrant inclusion in the study. Conversely, Con­gress has sometimes announced that it is raising taxes to strengthen the economy, as with the 1991 tax hike signed by President George H. W. Bush and the 1993 tax hike ushered through Congress by President Bill Clinton.

Romer and Romer used the official narrative record to distinguish legislated tax changes from changes in tax levels attributable to other causes and to identify the motivations behind the legislation. For the executive branch, the narrative record includes the Economic Report of the President, which typically discusses the major tax proposals in some detail; the President's annual budget submis­sion; and the State of the Union address. For the congressional branch narrative, the authors relied on committee reports, conference reports, and sum­maries provided by the Joint Tax Committee and the Congressional Budget Office.

Using these criteria, the authors identified 49 rel­evant tax law changes and administrative actions between 1947 and 2006. Of these, just under two-thirds produced tax changes in the ensuing quarters that were appropriate for study. Using these tax law changes, the authors were then able to assess the central question: Were tax hikes associated with lower economic activity?

The Romer and Romer study presents strong evi­dence that higher taxes tend to diminish economic activity. It found that a tax increase of 1 percent of GDP initially has a modest downward effect on out­put and that the effect grows rapidly before leveling off after 10 quarters for a maximum effect of lower­ing GDP by 3 percentage points. Applied to 2007, the study suggests that after two to three years, a tax increase of about 1 percent of GDP (about $135 bil­lion) would reduce output by about $400 billion annually.

When expressed as tax reductions rather than as tax increases, the study found that "tax cuts have very large and persistent positive output effects."[2] More­over, the authors emphasized that these results were "strongly significant, highly robust, and much larger than those obtained using broader measures of tax changes."[3] In other words, the modern historical record strongly suggests a clear and robust relation­ship between lower taxes and higher economic output.

A View on the Economic and Revenue Feed­back Effect. Taxes affect the economy in many ways. Taxes reduce the resources available to the economy by distorting economic incentives, so less labor is provided, less capital in the form of new saving is available, and less investment in education and new plant and equipment occurs. In addition, the distortions mean that resources employed in the economy are often not put to their most productive uses, further reducing national output and incomes. These are some of the avenues through which higher taxes reduce economic output, as found in the Romer and Romer study.

A recent National Bureau of Economic Research Working Paper by Greg Mankiw and Matthew Wein­zierl, both of Harvard University, sheds additional light on the magnitude of the effects of taxes on the economy.[4] In contrast to Romer and Romer, Mankiw and Weinzierl were interested primarily in identify­ing the feedback effects of tax changes. For example, if Congress reduced the tax on capital income with­out any consequent change in the level of economic activity or taxpayer behavior, then the tax policy change would obviously reduce tax receipts. This is sometimes referred to as the static revenue effect. However, such a tax reduction would increase the level of investment and therefore the level of output and income. The tax revenue gained from this increase in income is the revenue feedback effect, sometimes called the dynamic effect.

To measure the feedback effect, Mankiw and Weinzierl started with a standard neoclassical growth model of the economy, or the Ramsey model. A neoclassical growth model suggests the general nature of the relationships among capital, labor, and output. To make the model operational, the authors applied conventional parameter values for key variables, such as the responsiveness of labor supply to changes in the after-tax wage rate.

Every economic model rests on certain assump­tions, so it is important to examine how alternative assumptions affect the robustness of the model's results. For example, the standard Ramsey model assumes an infinite-horizon consumer, which is a simple way of ensuring that the participants in the economy, such as workers and investors, make wise decisions in terms of both current and future conse­quences. However, it is more intuitive to assume that individuals operate with finite horizons, so Mankiw and Weinzierl introduced finite-horizon consumers into the model. Similarly, the authors considered the effect of assuming imperfect compe­tition in the market for goods in lieu of the typical perfect competition assumption.

In all the variations they tested, Mankiw and Weinzierl found that changing the tax on capital or on labor produced a significant revenue feedback effect and thus implied a significant economic response. Specifically, the authors found that reduc­ing the tax on capital produced a dynamic effect equal to about one-half of the static effect. This dynamic revenue effect can then be used to infer the extent of the change in the overall economy.

For example, if the initial effective tax rates on capital and labor are 25 percent and a 1 percentage point increase in the tax on capital is expected to raise $20 billion per year on a static basis, then the resulting reduction in economic activity from the tax increase will reduce the revenue gain to $10 bil­lion. At a 25 percent effective tax rate, this result implies that the tax hike would permanently reduce annual economic activity by about $40 billion.

Changing the tax rate on labor supply also pro­duces a significant, although smaller, dynamic effect of about 17 percent. This means that a tax increase on labor that was intended to raise $20 billion by static scoring would increase actual receipts by only $16.6 billion because it would reduce economic activity by $13.4 billion.

These results correspond well with another result from the Romer and Romer study. Romer and Romer found that a tax increase equivalent to about 1 percentage point of GDP lowers real GDP by 3 percentage points, and it does so predominantly by reducing the level of investment. The overall decline in business investment according to this analysis is 12.6 percent, but the authors noted that this figure appears to be heavily influenced by inventory investment responding to a cyclical effect from the tax hike. The authors found that non-inventory business investment (i.e., investment in business plants and equipment) declines by a still very signif­icant 6.2 percent per percentage point increase in the tax on capital.[5]

Of course, Congress could raise taxes in a way that does not change the effective marginal tax rates on capital and labor. Many provisions in the tax code, especially in the individual income tax, could be eliminated to raise the average tax rate without affecting economically sensitive marginal tax rates. Examples include the child tax credit, the standard deduction, and personal exemptions. However, if recent debate is any guide, those who support tax increases are most interested in the kinds of tax pol­icies that would raise the effective marginal tax rates on labor and capital, such as higher top income tax rates, higher corporate income tax rates, higher lev­ies on capital gains and dividends, raising or elimi­nating the income cap on payroll taxes, restoring the estate and gift tax, and a new surtax on upper-income earners.[6]

The Romer and Romer study and the Mankiw and Weinzierl study confirm the conventional wis­dom that higher taxes diminish economic vitality. Their importance is heightened when they are con­sidered together because the two studies took very different approaches to reach their common conclu­sion. Nevertheless, two academic studies, regardless of how careful and innovative they are and how well-respected their authors are, cannot settle the question of whether higher taxes are economically harmful, benign, or helpful, and they certainly can­not settle the question before giving due consider­ation to the opposing argument from the advocates of tax hikes.

The Silver Lining Theory of Higher Taxes. Some advocates of higher taxes argue that higher taxes can strengthen the economy. The theory is that higher taxes lead to an increase in national sav­ing, which in turn puts downward pressure on real interest rates and leads to higher levels of national investment and output. This is a narrow yet plausi­ble argument resting on a chain of testable eco­nomic relationships. If each link in the chain proves to be valid and robust, then the argument has merit.

The strength of the tax hike argument is that all but one of the economic links are noncontroversial. A weakness is that all of the connections must be valid for the narrow argument to hold, and the con­nections must be robust for the theory to be relevant. A greater weakness is that even if the narrow theory is valid, the narrow positive effect from higher taxes must be weighed against the broader range of nega­tive effects that taxes have on investment specifically and on the economy in general. This theory might be called the "silver lining" theory of tax policy: Within the dark clouds of negative economic effects from higher taxes, advocates ignore the billowing clouds to focus on a possible silver lining.

The argument begins with the observation that an increase in taxes, holding all else constant, will reduce the budget deficit and thus raise public sav­ing. This is not generally a controversial proposition as long as Congress does not increase spending in response to the higher revenues. However, Con­gress has a long history of spending virtually every dollar it can, so an increase in tax revenues would likely soon be matched with a like increase in spending. Yet the issue here is the possibility of a positive economic outcome, so the assumption that Congress holds spending constant is sustained.

The next link in the chain is that an increase in public saving results in an increase in national sav­ing. This would seem to be an obvious connection, but a respected theory suggests that private saving may adjust in the opposite direction of public sav­ing.[7] Nevertheless, the argument that national sav­ing would increase is sufficiently intuitive and straightforward to permit acceptance here.

The silver lining theory ultimately hangs on the next link in the chain: that a significant increase in national saving will produce a material reduction in interest rates. Those who make this argument are generally referring to lower long-term real (infla­tion-adjusted) interest rates.

On a purely theoretical basis, this is a highly debated proposition for a modern, open economy. The United States is typically a net importer of sav­ing from abroad as U.S. net imports of foreign sav­ing match U.S. net imports of goods and services in the balance of payments. Under these circum­stances, an increase in public saving due to a tax increase could simply supplant a like amount of saving imported from abroad. If this occurs, then a change in tax levels would not appreciably affect real long-term interest rates.

However, the relationship between government deficits and debt on the one hand and real interest rates on the other is empirically testable. Findings from two recent studies that speak to this proposi­tion are presented in the next section.

The next link in the chain is that reducing the real long-term interest rate increases business investment. There is no real dispute as to the direc­tion of the effect: A reduction in real interest rates should increase the level of capital employed. How­ever, theory provides little guidance as to how much additional capital will be employed for a given reduction in the real interest rate. Furthermore, apparently, no modern research exists seeking to quantify this relationship.

The last link in the chain relates changes in the level of capital used in the economy to changes in the long-run level of overall economic activity. In the long run--that is after the economy has adjusted to these changes--an increase in the amount of capital employed raises long-run eco­nomic activity by raising the level of labor produc­tivity. Simply put, more capital per worker means that workers are more productive. This relation­ship is robust and generally uncontested.

In summary, the total chain of events that comprises the silver lining theory can be shown as follows:

Of these links, only the link between increased national saving through budget deficit reduction and lower real interest rates is seriously controversial.

Real Interest Rates and Federal Borrowing

The debate over the effects of government debt and deficits on real interest rates has a long history and is ongoing. Recently, a pair of studies have gar­nered broad general acceptance in terms of method­ology and results, and their results are largely consistent with one another.

Current Fiscal Policy and Real Interest Rates. Eric Engin of the National Bureau of Eco­nomic Research and Glenn Hubbard, Dean of the Columbia Business School, examined the histori­cal record of government debt and interest rates.[8] They started their analysis by positing a simple, intuitive theoretical relationship in which the value of the additional output from an additional unit of capital determines the real interest rate as shown in Chart 1. The downward-sloping line shows the relationship between the level of capital employed and the resulting real interest rate. If the initial level of capital employed is k₀, then the real interest rate is r₀.

If for some reason the level of capital employed in the economy declines to k₁, the economy will then move back up the sloping line until the interest rate settles at r₁. Thus, a lower amount of capital employed will mean that the efficiency of the last unit of capital employed is higher, so the real inter­est rate is higher.

Engin and Hubbard then make the critical assumption that the level of national saving at any point in time is fixed, so a higher level of govern­ment debt means less private saving is available for investment. Thus, issuing an additional dollar of government debt reduces or "crowds out" private investment by a dollar. With less capital employed, the productive efficiency of the last unit of capital is higher, as is the real interest rate.

This relationship between government debt and interest rates seems straightforward, and the issue would appear to be one of estimating the slope of the line in Chart 1. A steep slope would indicate a strong relationship between government debt and interest rates, whereas a nearly flat line would indi­cate a weak relationship. However, there is one fur­ther complication.

It is certainly true, as the authors assume, that the supply of saving from domestic sources avail­able for domestic investment changes slowly over time. However, the United States imports large amounts of saving from abroad and exports large amounts of saving to foreign investors. In any given period, the net of these flows may be quite large. Thus, the total supply of saving--domestic saving plus net imported saving--may vary significantly over even short periods. In this case, the proposed relationship between government debt and interest rates may be very weak. In effect, if flows of foreign saving are significant, then the slope of the line in Chart 1 will tend toward the horizontal.

Using this theoretical framework relating gov­ernment debt, investment levels, and real interest rates, Engin and Hubbard found a statistically sig­nificant relationship between federal debt and inter­est rates, but they also found that the effect is generally very small. Specifically, the authors found that an increase in the level of federal government debt of 1 percent of GDP would increase the long-term real interest rate by about 3 basis points, or 0.03 percentage point. In terms of Chart 1, the line is nearly but not quite completely flat.

These results have interesting implications for recent experience. Federal government budget def­icits have raised the debt-to-GDP ratio from 35.1 percent in 2000 to 37 percent in 2007. According to the Engin and Hubbard analysis, fiscal policy tended to raise the real interest rate over this period by about 6 basis points compared to what it other­wise would have been. The 10-year Treasury bond rate averaged 4.63 percent in 2007, so this analysis suggests that the rate would have been 4.57 percent if the debt-to-GDP ratio had returned to its 2000 level by 2007. On the other hand, the debt-to-GDP ratio actually fell slightly from 2006 to 2007, sug­gesting that the improvement in the federal fiscal picture likely exerted some slight downward pres­sure on the long-term bond rate in 2007.

The slight effect on real interest rates that is attributable to recent changes in the level of federal debt is shown in Chart 2. One line shows the actual 10-year Treasury bond rate for the period. The sec­ond line shows what the 10-year Treasury bond rate would have been if the debt-to-GDP ratio had been held constant. The two lines are almost identical, suggesting that federal fiscal policy has not signifi­cantly affected real interest rates over this period rel­ative to an interest rate-neutral fiscal policy.

Looking further back into the historical record shows a similarly minuscule and inconsistent effect. In 20 of the years since 1970--just over half of the time--changes in federal debt exerted a slight upward pressure (averaging about 5 basis points) on long-term real interest rates. However, for almost half the time, changes in federal debt applied a slight downward pressure that averaged about 4 basis points. The strong implication is that, despite all of the fiscal policy changes since 1970, the effects of deficit finance on real interest rates and the econ­omy were minor, verging on inconsequential.

Projections of Fiscal Policy and Future Inter­est Rates. A second study, performed by Thomas Laubach at the Federal Reserve, found a similar result using a different framework.[9] Laubach con­sidered the effects of projected fiscal policies as opposed to current policies, and he looked at longer-horizon interest rates rather than current lev­els of long-term interest rates.

The advantage of the Laubach framework is that many factors affect interest rates, especially in the short run, so isolating the near-term effects of fiscal policy can be difficult. Often, however, these effects are transitory, as when automatic fiscal policy stabi­lizers operate during a recession or when deficits rise appreciably due to a temporary surge in spend­ing, such as during the response to Hurricane Kat­rina. Levels of government debt expected to prevail several years into the future are unlikely to be affected materially by the current business cycle or near-term transitory spending patterns and thus are more likely to indicate accurately the influence of government debt on future real interest rates.

Laubach used the five-year forecasts for federal deficits and debt as published by the Congressional Budget Office (CBO) and the Office of Management and Budget (OMB). Budget deficit forecasts are notoriously inaccurate, and their ranges of error grow as the forecast horizon extends further into the future.[10] Nevertheless, the CBO and OMB forecasts constitute the best publicly available information about future federal fiscal policy at the time of their publication. Therefore, they offer a reasonable basis on which to assess fiscal policy effects, as under­stood by the market, on future interest rates as set by the market.

To measure market expectations of future nomi­nal interest rates, Laubach takes advantage of the information embedded in the interest rate yield curve. The yield curve is the relation between inter­est rates and the terms to maturity of various debt instruments of similar risk characteristics such as the 91-day Treasury bill and the 10-year Treasury bond. Chart 3 shows a typical yield curve. Short-term interest rates tend to be lower than long-term interest rates because the funds are committed for shorter periods.[11]

Laubach used this information on government forecasts of future debt levels relative to GDP to determine whether there was a consistent rela­tionship to market expectations about future interest rates. The analysis revealed a statistically significant rela­tionship in that a 1 percentage point increase in the projected debt-to-GDP ratio would be expected to raise future interest rates by about 4 to 5 basis points. This result is both quantita­tively small and remarkably close to the 3-basis-point effect found by Engin and Hubbard.

These results have implications for current level interest rates. At the end of 2007, the debt-to-GDP ratio stood at 36.8 percent. The latest OMB fore­casts show budget deficits through 2013, but they are sufficiently small relative to the size of the economy that the debt-to-GDP ratio falls to 33.4 per­cent by 2013.[12] Similarly, the CBO forecasts a 33.2 percent debt-to-GDP ratio in 2013.[13] The average of the two forecasts for 2013 is 31.6 percent, or 4.3 per­centage points below the 2007 level. The Laubach study suggests, therefore, that the expected progress on reducing the debt-to-GDP ratio over the next few years is putting downward pressure on long-term interest rates equivalent to 15 to 20 basis points.

Together, the Engin and Hubbard study and the Laubach study suggest a tentative, developing consensus about the general magnitude of the effects of deficit financing on real interest rates. The studies appear to suggest that, for deficits and debt levels in the ranges seen in recent years and projected in the medium term, the effects on real interest rates are in the expected direction, consis­tent across episodes and across estimating meth­odologies, and very slight--measured in terms of a handful of basis points.

Recap of the Silver Lining Theory

The silver lining theory argues that higher taxes could lead to a stronger economy through a chain of connected effects, including a real interest rate effect. There is little dispute that a generic increase in taxes will increase national saving. Nor is there much dispute about the direction of change in the economy resulting from a change in real interest rates: Lower real interest rates would increase the amount of capital employed in the economy, raising output by raising productivity. The silver lining theory's validity and relevance stands or falls on whether or not government deficits significantly affect real interest rates.

The consensus seems to be that there is a rela­tionship between deficits and debt on the one hand and real interest rates on the other, but that it is very, very weak, so the resulting increase in investment and output as posited by the silver lining theory would be commensurately weak. With respect to this specific set of relationships emphasizing the importance of changes in the real interest rate, the answer seems to be that tax changes of the usual magnitude would have little appreciable effect on the level of economic activity. The silver lining the­ory turns out to be rather threadbare.

Taxes, Real Interest Rates, and Investment

Tax changes affect the economy through many channels. Because the silver lining theory emphasizes investment, it is worthwhile to explore the relation­ship between taxes and investment a bit further.

As described above, a tax increase that raises public saving would be expected to exert very slight downward pressure on real interest rates, which would tend to raise the level of investment by some undetermined amount. However, if the tax increase was on capital, then the offsetting effects on invest­ment must be considered. When looking at silver lin­ings, one cannot ignore the darkness of the clouds.

Every investment must earn enough to compen­sate the investor on an after-tax basis for the time value of money (i.e., the real interest rate) and the risks associated with the investment.[14] In addition, the investment must earn enough to cover the taxes that will be subtracted from its gross earnings. The combination of these factors is variously referred to as the cost of capital, the service price of capital, or the hurdle rate on investment. The important point is that no investment is undertaken if it is not expected to generate a return in excess of the hurdle rate.

For example, if the real interest rate were 3 per­cent per year and compensating investors for risk required an additional 1 percent per year, the required real after-tax return would total 4 percent. If the tax rate on the returns to this investment were 20 percent, then the (before tax) hurdle rate on this investment would be 5 percent.

If the tax on capital income were raised to 30 percent and the proceeds were used entirely for def­icit reduction, then the real interest rate would decline to 2.95 percent and the required after-tax return would fall to 3.95 percent, as tax increase advocates suggest. However, the increased tax rate on capital income would raise the hurdle rate to 5.64 percent--an increase of 12 percent. Despite the suggested real interest rate effect, the net conse­quence of the tax increase would be a significant reduction in investment and economic activity.

This is just one example of many that could explain why the modest positive effect of a tax increase through real interest rates and investment is overwhelmed in practice to produce the traditional result that higher taxes result in less prosperity. Alternatively, if the tax increase falls wholly on labor, the reduction in labor supply must be balanced against the improvement in investment incentives caused by a minute fall in real interest rates. Again, the net result would be a smaller economy.

The implication is that tax increases do indeed harm the processes of economic growth and that arguments to the contrary are not well supported. Federal policymakers should therefore look for every opportunity to reduce taxes, especially those that are most harmful to economic growth such as marginal tax rates and taxes on saving and investment.

Conclusion

On balance, clear and compelling evidence shows that higher taxes reduce economic output. The silver lining argument appears to be valid within a narrow scope of influence on the economy, but it is overwhelmed by the dark clouds of weaker economic performance caused by other effects of higher taxes.

The argument that tax increases are economi­cally beneficial must establish a positive net effect from two opposing elements: The gain due to increased capital investment from a lower real inter­est rate must exceed the losses in output due to the increased economic distortions from the tax hike. Even granting that lower deficits bring downward pressure on real interest rates, the effect appears to be minimal--in the hundredths of a percentage point. Yet even granting that such a real-interest-rate effect exists, proponents of tax hikes as being pro-growth must then establish that the resulting encouragement for business investment exceeds the downward pressures on business investment and labor supply and the misallocation of resources from higher taxes.

Given the modest potential gains in business investment from deficit reduction and the poten­tially significant losses in both business investment and labor supply from a tax hike, it seems entirely implausible that most hikes under consideration would not harm the economy. The silver lining the­ory is superficially appealing and has the rhetorical merit of being relatively easy to explain. However, its strengths end there.

In short, the evidence supports the view that tax increases harm economic performance. As a first priority, federal, state, and local policymakers should eschew tax increases. As the tax burden in the United States continues to rise, policymakers at all levels of government should pursue tax relief to preserve and enhance a strong economy.

J. D. Foster, Ph.D., is Norman B. Ture Senior Fel­low in the Economics of Fiscal Policy in the Thomas A. Roe Institute for Economic Policy Studies at The Heritage Foundation.