Do Deficits Matter?


"[An] excellent, comprehensive, and illuminating book. Its analysis, deftly integrating considerations of economics, law, politics, and philosophy, brings the issues of 'balanced budgets,' national saving, and intergenerational equity out of the area of religious crusades and into an arena of reason.…A magnificent, judicious, and balanced treatment. It should be read and studied not just by specialists in fiscal policy but by all those in the economic and political community."—Robert Eisner, Journal of Economic Literature

"By conceptualizing budget deficits as 'tax lag'—the spending of money by government before it has been acquired through taxation—Shaviro directs us toward issues of generational burden-shifting, macroeconomic performance, and the size of government.…Shaviro is unusually honest about the difficulties that have left these conflicts unresolved. …A sobering analysis."—Kirkus Reviews

"The author's background as a legislation attorney shows in his insightful analysis of political gamesmanship and the unintended consequences of legislation. Shaviro's history, economics, and political analysis are right on the mark. For all readers."—Library Journal

The introduction to
Do Deficits Matter?
Daniel Shaviro

Few topics in American politics are more discussed and less understood than the federal budget deficit. We frequently hear that deficit reduction is vital to our prosperity, but we rarely hear why this might be so. The deficit is blamed for all manner of economic ills, ranging from high interest rates to unemployment to the trade deficit to the low rate of national saving to low productivity growth—whichever seems most crucial at the moment—but little attention is paid to why it might have any of these effects.

The near unanimity in public discourse about the evil of deficits might seem to suggest that economists are similarly unanimous. In fact, however, they disagree fundamentally about whether deficits matter, and, if so, then why. They have been debating these issues for more than two centuries, with consensus occasionally emerging but not persisting. Over the past twenty-five years, the deficit debate among economists has grown increasingly discordant, reflecting the issue's increased prominence, the growing size of reported deficits, and the collapse of 1960s Keynesianism.

Despite the scope and intensity of this debate, no one has ever published a serious, comprehensive study of budget deficits' economic and political significance. Economists generally do not focus on more than one or two of the main issues that we will see deficits pose, nor do they explore the relationships between the issues. They also mainly ignore the question of how deficits affect the level of federal spending, although this is a key part of the national political debate. Moreover, their work is often marred by partisanship of the Left or the Right, or by the impulse to stake out a public stance like a campaigning politician—often as the foundation of one's academic career—and then to stick to it at all costs, ignoring or disparaging all contrary arguments and evidence.

For the first time in two centuries of deficit debate, this book unifies the varied strands of the economic and political science literature, and draws definite, balanced, and inter-related conclusions regarding all of the main issues that budget deficits present. The journey has many stages, and will take us in a number of different directions. I therefore offer the following summary of my analysis and conclusions.

1) Purely as a descriptive matter, why do Americans care so much about budget deficits? The United States has a history of unusual concern about federal (although not state) budget deficits, going back to the earliest days after adoption of the Constitution. Other industrialized nations generally have not shown similar levels of concern, even when they have had comparable or greater budget deficits (or outstanding national debt) relative to their economies.

Our deficit fixation results in part from equating government debt with an individual's or household's debts, which generally could not grow for decades at a time without raising a serious prospect of default. In addition, deficits are a deeply rooted symbol in American history, the meaning of which has changed over time, but that continually relates to distrust of the national government. The historical lineage from Thomas Jefferson's denunciations of the deficit to those of H. Ross Perot should alert us to the cultural conditioning that needs to be left to one side for purposes of clear analysis.

2) Are the national government's debts properly analogized to those of an individual or household? People often exaggerate the analogy, although it is not wholly misplaced. The main distinction is that the federal government, with its power to raise taxes and print money, faces less default risk than would a household that regularly spent more than it took in. In addition, our debt is mainly internal: owed by American taxpayers to American bondholders (two groups that overlap).

While the prospect of involuntary default is remote in the United States at present, we face a problem that is a lesser version of it: that of policy sustainability, or our ability to meet current commitments that are not quite as definite as the pledge to honor government bonds. Current fiscal policy is likely to prove hard to sustain—not only because of the growing national debt, but because of the expected long-term insolvency of Social Security and Medicare. Fiscal policy changes may result in disappointing expectations that our present fiscal policy encourages people to hold. Yet more and more people realize that our fiscal policy must and will change. The main problem posed by policy unsustainability is the shock and dislocation that result when expectations must change too fast.

3) What issues do federal budget deficits raise, apart from concern about default and policy sustainability? The first issue is generational equity, or concern about unduly benefiting current generations at the expense of future generations. The second issue is what macroeconomic effects deficits generally have. At one time, they were lauded by Keynesians as the cure for recession or even significant unemployment, based on the claim that they increased current consumer spending and thereby stimulated the economy. Today, the same causal claim leads many to condemn deficits as a cause of the low rate of national saving. The third issue concerns their effects on the size of the national government. Some supporters of limited government identify deficit spending as a major cause of undesirable government growth, and therefore advocate the adoption of a balanced budget amendment to the U.S. Constitution.

Each concern rests on a common causal claim: that deficit spending reduces the perceived (whether or not the actual) cost of government spending to current consumers and voters, thus inducing them to feel wealthier. They therefore consume more, leave less for subsequent generations, and accept a higher level of government spending than they would have otherwise. Before discussing the accuracy and significance of this causal claim, we must look more carefully at how deficits are defined and measured.

4) Is the budget deficit a meaningful economic measure? The budget deficit would seem most likely to have these claimed effects if it were a meaningful economic measure. Unfortunately, it is not—although there is a meaningful underlying phenomenon, which it mismeasures, about which the above claims can be made.

The budget deficit's main shortcoming is that it is calculated on the basis of cash flow, rather than economic accrual. Suppose that an individual kept his books under the rules that the federal government uses to measure the budget deficit. If he managed to buy a million dollar home for only a thousand dollars, he would seem to have acted imprudently, since, in the year of purchase, he would have increased his deficit. This transaction would be treated identically to losing a thousand dollars at the racetrack. If he instead sold a million dollar home for a thousand dollars, or agreed to pay someone a million dollars next year in exchange for a thousand dollars today (assuming that this was not classified as a loan), he would have accomplished deficit reduction, at least under some versions of the measure, and thus would seem to have acted prudently.

While inaccuracies of this kind may have mattered little at one time in American history, that time has passed. Today, a cash flow measure creates systematic, not just random, mismeasurement, for three main reasons. First, it encourages ignoring the approach of unfunded future spending commitments, as under Social Security and Medicare. Second, it encourages legislative responses that, over the long term, are meaningless or even make the government's fiscal posture worse. "Smoke and mirrors" policy changes, or those that reduce deficits in the short term while increasing them for "out years" beyond the estimating window, have been a feature of all major deficit reduction initiatives in Washington. Third, to the extent that fiscal policy affects economic behavior via its impact on perceived wealth, the long-term elements of such policy that the deficit ignores may matter. Unless people are highly myopic, their expectations regarding how much the government will pay or take from them in the future should affect their current behavior.

5) How could we better describe the underlying fiscal policy that deficits mismeasure? Given the deficit's economic inaccuracy as a cash flow measure, we need a new vocabulary to describe the underlying phenomenon of having spending accrue before taxes. I will use the term tax lag to describe a fiscal policy (such as our present one) in which, over the long term, (a) tax revenues will be inadequate to pay for government spending absent a policy change and/or (b) younger generations and future generations will end up paying for government spending on behalf of older individuals and current generations. The word lag is appropriate because, over the long term, no government spending is free; it all must and will be paid for by someone. Even if the national debt is never repaid, taxpayers bear it economically over time by perpetually paying interest on the debt. Nor would defaulting on our debt obligations eliminate the cost of paying for government spending. Default would merely shift the cost from taxpayers to bondholders—functioning, in effect, as a one-time tax on the latter.

When I use the term tax lag, the reader should keep in mind that I am describing the relative timing of taxes and spending, rather than anything absolute about taxation. I could just as easily refer to spending acceleration. Tax lag can be reduced through policy changes either on the tax side or on the spending side of the federal budget. On the tax side, one can increase the extent to which future necessary taxes are specified and/or increase taxes on current generations and older individuals. On the spending side, one can reduce planned present or future spending, in particular that on behalf of current generations and older individuals.

The most intuitively obvious way to measure changes in tax lag would be through what I call the economic accrual budget deficit (as distinct from the cash flow budget deficit that we currently use). This measure—which I mean only as a thought experiment—would focus on economic accrual over time, rather than cash flow, by taking account of expected future revenues and outlays at their interest-adjusted present value. Suppose that at the beginning of the fiscal year, the national debt stood at $2 trillion, and the present value of all expected future budget deficits stood at $3 trillion. Tax lag would therefore total $5 trillion in present value terms at the start of the year. (This would be the "economic accrual national debt.") Next, suppose that during the fiscal year, there was a $100 billion cash-flow budget deficit, and the present value of all expected future deficits rose by $150 billion. The economic accrual budget deficit would equal the sum of these two amounts, or $250 billion, as the present value of our tax lag would have increased by that amount to $5.25 trillion in the course of the fiscal year.

The measure, as I have described it thus far, would still be too cash flow-oriented in a critical respect. It would fail to distinguish between expenditures that create durable government assets, and those that are immediately consumed. Recall my earlier point that spending a thousand dollars to buy a house is quite different from spending that amount at the racetrack. To the extent feasible, one would want to adjust the economic accrual budget deficit to use standard principles of accrual accounting for government expenditures. At a minimum, expenditures that created lasting government assets would be deducted over their estimated useful lives, rather than in the year of the expenditure. One might also want to consider adjusting for fluctuations in the value of government assets (at least those plausibly held for sale, if not, say, the Lincoln Memorial), and ignoring government asset sales that merely convert property to cash.

Yet, even if one could make these adjustments, another problem would arise. Suppose that Congress in 1998 enacted a head tax of $50,000 per year, to apply starting in 2050 to each adult American, and remain in force as long as needed to pay off all public debt and eliminate all funding shortfalls in Social Security and Medicare. This enactment probably would not alter one's view of our current fiscal policy, whether because it seemed frivolous or because the bottom line, the fact that taxes ultimately will pay for today's spending, was already implicit. This suggests two problems with the economic accrual budget deficit. The more trivial one is that the set of tax and payment rules currently on the books is less important, for some purposes, than the set of rules that actually, credibly constitute our current policy. The more fundamental problem is that what really matters about taxes and spending, at least distributionally, is who pays for and receives them. A tax on younger generations in fifty years may be quite different than a tax on us today, even if the taxes have the same present value.

Responding to these problems in measuring tax lag, the economist Laurence Kotlikoff has proposed a new measurement system to replace the cash flow budget deficit that, while less intuitive than the economic accrual budget deficit, provides more meaningful information. He calls it generational accounting. Its most important innovation, beyond employing principles of economic accrual, is that it compares expected taxes to expected outlays by age group, rather than providing a single overall measure of tax lag. Generational accounting involves computing the estimated lifetime net tax payment and lifetime net tax rate for the average member of an age group—those born, say, in 1930, 1960, 1990, or the future—assuming the continuation of current policy (except that future generations are deemed to make up all the long-term revenue shortfalls). The lifetime net tax payment is the excess of taxes paid over transfers received, computed on a lifetime basis in present value terms from birth. The lifetime net tax rate is the lifetime net tax payment, divided by estimated lifetime income.

Generational accounting thus directly addresses which age groups win and lose under fiscal policy, and sheds light on such policy's likely sustainability. According to Kotlikoff, lifetime net tax rates have been rising throughout the twentieth century, and now stand at astronomical levels for future generations—more than 84 percent under current policy, according to his most recent data.

Unfortunately, any long-term economic accrual measure involves conceptual and computational difficulty. In calculating future years' tax and spending levels, what set of policies should we assume will be followed? How confident can we be in any long-term economic and demographic forecasts? Should mere expectations of benefiting in future years from government spending programs be distinguished from explicit public debt? Can we really determine the incidence, within a multi-generational household, of government taxes and transfers?

While these problems reduce generational accounting's practical value—especially given estimating games in the real world of partisan politics—it remains conceptually superior to the cash flow budget deficit. Generational accounting requires extensive assumptions, but at least it sets forth a meaningful economic concept. Moreover, it addresses an important gap in current understanding. While political debate has moved in the direction of focusing on long-term economic accrual (as through multi-year deficit forecasts and debate about the long-term solvency of Social Security and Medicare), it has tended to ignore the generational implications of alternative policies.

Still, despite the importance of focusing on long-term accrual and on who pays what, discussion of the cash-flow budget deficit is not wholly without value, as long as we keep its limitations in mind. Major deficit reduction initiatives in Congress often would reduce tax lag, even if they also include smoke-and-mirrors elements. One could even argue that the deficit has particular advantages as a guide for public political discourse, despite its conceptual flaws, because of its greater salience and symbolic heft.

6) What aspects of government policy do even long-term fiscal policy measures ignore? So far, the discussion has been limited to fiscal policy, which concerns cash flows to and from the government. This focus is incomplete, since cash transactions represent only a part of total government activity. Both in-kind benefits from government spending, and the various in-kind benefits and burdens resulting from government regulation are similar in principle to cash subsidies and tax levies. Yet they generally are ignored in discussions of fiscal policy, for no better reason than that they are hard to measure and allocate to specific individuals or groups.

One could argue that just as the cash flow budget deficit is not a true economic measure because it ignores long-term accrual, so generational accounting is not a true economic measure because it ignores the value of in-kind benefits and burdens imposed on different age groups by government policy. Consider Kotlikoff's finding that lifetime net tax rates have been rising over time. If this were wholly an artifact of government growth that could increase the value of in-kind public goods and services, it might have little significance. The claim that one can meaningfully evaluate fiscal policy separately from all other government policy is essentially an "all else equal" claim (or hope)—although perhaps not an unreasonable one.

7) All else equal, does tax lag actually have the generational, macroeconomic, and size-of-government effects that have been attributed to budget deficits? This is an empirical question, which turns on whether tax lag actually causes people to feel wealthier and regard government spending as less costly than they would if taxes were accruing at the same rate as spending. The argument against tax lag's having this effect begins from the observation that rational private borrowers recognize that borrowing money generally does not make them wealthier, but merely gives them current cash against a future, interest-bearing charge. If people generally took this view of public debt—considering it, in the aggregate, a charge on their current wealth since it implies higher future taxes, including interest on the deferral—then tax lag would fail to increase their perceived or actual wealth.

The economist Robert Barro asserts that people actually view tax lag this way. Thus, he would expect a $100 billion current year tax reduction (holding government spending constant) not to affect current consumption behavior or have any systematic impact. This claim—called Ricardian equivalence for economist David Ricardo who first described it—suggests that tax lag is largely irrelevant because people make offsetting adjustments. In effect, they put the amount of any deferred tax in a dedicated bank account to pay the tax when it comes due, rather than regarding it as giving them extra money to spend—even if they do not expect it to come due during their lifetimes. Barro derives this claim from a "rational expectations" framework in which people, acting without systematic error, pursue fixed goals regarding how much wealth to transfer to their children net of deferred taxes, and thus achieve whatever end result they prefer without regard to any government policies that fail to alter the "opportunity set."

If Barro were correct, then politicians could, with relative impunity, propose massive current tax increases to eliminate tax lag, leaving the question of spending levels to be debated separately. Yet even the barest familiarity with contemporary politics shows that this is false. Voters generally punish politicians who propose current tax increases, evidently not agreeing with the Ricardian argument that the timing of taxation has no effect on its perceived current level. We also fail to observe other apparent implications of Ricardianism—for example, equal frequency of budget deficits and surpluses, or any tendency of large families to prefer current tax increases based on the view that their share of the tax burden will grow over time as they split into multiple households. All this suggests that increasing tax lag does tend to increase perceived wealth, leading to the claimed generational, macroeconomic, and size of government effects.

8) What should we think of tax lag's generational effects? Many argue that shifting our tax burdens to future generations is immoral. The question is not really one of wealth transfer, which is pervasive and inevitable between generations in any event (as whenever parents spend money on their children or leave bequests). Rather, the issue is better put in terms of rising lifetime net tax rates. Laurence Kotlikoff argues that a norm of "generational balance" requires avoiding any increase in the expected lifetime net tax rates for future generations relative to those on current generations.

I find this norm unpersuasive. The real issue is the overall distribution of lifetime consumption between succeeding generations. This, in turn, depends less on fiscal policy than on present generations' overall rate of saving and productivity of investment, along with decisions within the household concerning such matters as child care, educational investment, and the rate of divorce. There is no apparent reason why government fiscal policy, which is merely one component of everything we do that affects our descendants, should be generationally "balanced." Even the narrower claim that reducing tax lag, by increasing national saving, would shift lifetime consumption in the right direction, may not be correct. For example, if technological advances cause people fifty years from now to be wealthier than we are—just as we are wealthier than people fifty years ago, and they are wealthier than people fifty years earlier still—then changing fiscal policy to benefit future generations would amount to playing Robin Hood in reverse. While per capita societal wealth is not certain to continue increasing, our inability to predict the future makes it hard to know what generational policy would be best.

One might ask: even if some increase in lifetime net tax rates is justifiable, isn't a projected 84 percent rate for future generations far too high? Here the answer might be yes if such a rate were actually possible. Even Kotlikoff agrees, however, that it is not. Rather, this projected rate reflects a computational convention, under which he estimates the implications of current policy (although it may be certain to change) by assigning the entire unprovided-for net tax burden to future generations. He does this both because some arbitrary assumption is needed (since current policy does not tell us how tax lag will be addressed), and as a rough indicator of where our current fiscal policy is headed. A staggering lifetime net tax rate for future generations mainly indicates a sustainability problem for current policy, suggesting that some of the as yet unprovided-for net tax burden will have to be borne, in the end, by members of present generations.

9) What should we think of tax lag's macroeconomic effects? Again, tax lag tends to increase current consumption relative to saving. While many condemn this effect on the ground that the current rate of national saving is too low, it is hard to be sure. What is the optimal rate of saving? No one knows. The whole point of saving, rather than consuming, is to make possible greater future consumption. Yet the choice between present and future consumption presents a difficult tradeoff, and neither is to be preferred automatically. In any case, while saving helps to promote long-term economic growth, such growth probably depends more, in the end, on technological developments than on the precise level of capital accumulation.

Tax lag's effect on the choice between current consumption and saving also matters in Keynesian theory, where one can moderate recessions by making people feel wealthier, thus offsetting fear-induced reductions in consumer spending. However, although Keynesianism has rebounded from its apparent intellectual collapse in the 1970s, even most Keynesians now recognize that the case for an actively countercyclical fiscal policy is extremely weak. The main problems are systematic misuse by politicians pursuing short-term political goals, and undue lag in implementing fiscal policy changes. Keynesian fiscal policy should be purely automatic rather than discretionary, as when unemployment benefits rise and income tax revenues decline during a recession.

10) What should we think of tax lag's size-of-government effects? Again, tax lag tends to increase government spending by reducing its perceived cost. While this does not prove that, on balance, the federal government spends too much, there is much to be said for such a view. Political and regulatory processes often lead to bad and wasteful policy due to their structural problems, such as the power of special interest groups, politicians' and bureaucrats inappropriate incentives, and voters' low information.

Nonetheless, even if by reducing tax lag, one could reduce the number of dollars the government spends, a smaller and better government might not result. For one thing, the size of government and the level of harm from bad policies, correlate only very roughly, if at all, with the number of dollars that the government spends. For another, political pressures might mainly lead to a reduction in good rather than bad spending. For a third, reducing tax lag might simply yield changes in the form of government activity, with regulatory mandates replacing explicit taxes and spending. This has already happened to some extent in recent years. Therefore it is not clear that reducing tax lag would lead either to a genuinely smaller government, or to one whose policies, when misguided, did less harm.

The foregoing discussion suggests that, while tax lag is important and needs to be better understood, the appropriate response to it is unclear. Only on size of government grounds do its effects seem clearly undesirable, and even there reducing it might not accomplish much. This suggests skepticism about enacting a balanced budget amendment (BBA)—the great political issue concerning tax lag in recent years—even disregarding the budget deficit's shortcomings as a measure.

Such skepticism is enhanced when one examines the BBA's likely form, as in the version nearly adopted by Congress in 1995. This version of the BBA would probably be grossly ineffective, even on its own terms, for two main reasons. First, it allows a 60 percent supermajority in both houses of Congress to authorize deficits of any size, and for any reason. While this flexibility has some virtues, in practice it might lead to increased budget deficits! Supermajority requirements may increase the amount of logrolling that is necessary to assemble a winning coalition. Second, the proposed BBA has no built-in enforcement mechanism. It merely states that unauthorized deficits should not occur. Any enforcement legislation that was adopted to fill this gap would lack the constitutional status of the BBA itself, and thus likely fail to provide a binding constraint.

In the end, the most pressing concern raised by current tax lag relates, not to the three main long-term issues that I have identified, but to the relatively short-term, prudential issue of policy sustainability. If, as seems likely, we cannot long pay for everything that current fiscal policy seems to promise—above all, due to the likely long-term insolvency of Social Security and Medicare—then our policy must change, and eventually will change. Along the way, however, there may be disruption and severely disappointed expectations unless the shift to a sustainable policy is made gradually, with broadly distributed impact, and in a reasonable, well-explained fashion. The sooner the shift to a sustainable policy is made, the less disruptive it is likely to prove.


Copyright notice: Excerpt from pages 1-12 of Do Deficits Matter? by Daniel Shaviro, published by the University of Chicago Press. ©1997 by Daniel Shaviro. All rights reserved. This text may be used and shared in accordance with the fair-use provisions of U.S. copyright law, and it may be archived and redistributed in electronic form, provided that this entire notice, including copyright information, is carried and provided that the University of Chicago Press is notified and no fee is charged for access. Archiving, redistribution, or republication of this text on other terms, in any medium, requires the consent of the University of Chicago Press.

Daniel Shaviro
Do Deficits Matter?
©1997, 344 pages, 3 tables
Cloth $29.95 ISBN: 978-0-226-75112-2
Paper $22.00 ISBN: 978-0-226-75113-9

For information on purchasing the book—from bookstores or here online—please go to the webpage for Do Deficits Matter?.

See also: