Macroeconomics was born out of the idea that fiscal policy could be used to stimulate the economy. In this chapter, we consider how decisions on government taxes and spending affect the decisions of individuals to consume and save.
In 2016, the current income of federal, state, and local governments in the United States came to $5.313 trillion. Of this amount, $4.980 trillion was raised as tax revenue—the rest from other sources. The government distributed $2.786 trillion in transfer payments, leaving it with $2.527 trillion to pay for $2.658 trillion in consumption expenditures, $672 billion in interest on government debt, and $62 billion in subsidies. That left a current account deficit of $865 billion, most of which was federal in origin.
As shown in Volume I, Chapter 7, the government collects tax revenues by taxing the economic choices of individual residents, in particular the choice of how much to make in wages by offering labor services to employers and how much to make in asset income by offering financial capital to investors. Federal and state income taxes affect those choices of individuals by taking away a part of their wages and asset income.
Governments also collect revenues by taxing consumption, such as through the federal excise tax on alcohol and state sales taxes. A federal tax on consumption would, as shown, untax net investment but tax wages. But consumption taxes impose a distortion of their own by making work less attractive relative to leisure.
Likewise, means-tested transfer payments affect decisions to work and save, insofar as the individual’s eligibility to receive those payments diminishes with the amount of income he receives. The fact that a tax filer can lose a portion of his earned income tax credit by earning more money, for example, means that the availability of that credit affects his work–leisure calculus.
In Volume I, Chapter 7, we distinguished between the income and substitution effects of a change in the wage rate brought about by a tax change. If a worker (“Adam”) experiences a rise in his after-tax wage rate because the tax on his wages falls, he will want to substitute work for leisure as the cost of leisure rises (the substitution effect). But he will also want to expand leisure because his after-tax income rises (the income effect). Conversely, if he experiences a decrease in his after-tax wage rate because the tax on his wages rises, he will want to substitute leisure for work as the cost of leisure falls, and he will want to contract leisure because his after-tax income falls.
There are likewise income and substitution effects associated with taxes on asset income. If Eve has to pay a tax on interest or dividend income that she receives by saving, then she experiences a reduction in the reward for forgoing current consumption and will, on that account, want to consume more and save less (the substitution effect). At the same time, because the tax makes her poorer, she will want to consume less (the income effect).
This reasoning relates to choices about current consumption and saving as affected by current taxes and spending. But government tax and spending decisions affect individual expectations of future income as well as current income and, thus also, individual choices over future as well as current consumption and saving. When the government spends more and, inevitably, taxes more, the individual experiences a reduction in the present value of funds available to finance his current consumption. This exerts an income effect of its own, which has to be considered in assessing the effects of the new spending on individual economic behavior.
Income Effects of Fiscal Policy Changes
To examine this income effect, we adopt a convention frequently used—and misused—to measure the effects of changes in fiscal policy on government revenues. This convention assumes that changes in tax burdens and benefit distributions impinge only on people’s disposable income, which is the money left with them after taxes and transfer payments, but that they don’t impinge at all on the relative cost of leisure or consumption. It is as if the government taxes people by sending them bills in the mail without regard to their economic choices, in particular, their choices about earning income or saving for future consumption. The taxpayer gets a bill from government in exactly the same manner as he gets a bill from his credit card company, except that his tax bill bears no relation to how much he spent on consumption. Likewise, government distributes transfer payments by sending out checks unrelated to individual economic choices.
This procedure is what a “static” analysis of tax policy implies. In effect, it is assumed that an X-percent increase in the tax rate imposed on some activity will cause an X-percent increase in the revenue collected from taxing that activity and, in the process, have no effect on the individual except to make him poorer. As naïve as it is, that convention is acceptable to the extent that people adjust their consumption and saving choices to the burden that they expect taxes to impose on their current and future incomes. People take into account the current and future taxes they can expect to pay when making their current consumption and saving decisions.
Fiscal policy has to do with the effects of government purchase, G, and therefore net taxes, T, on those decisions. When economists use the expression “fiscal policy,” they usually do so with some variation of the Keynesian economic model in mind—a model we have yet to consider but that relies on the assumption that neither tax law nor government spending impinges on the work–leisure or consumption–saving calculus, as worked out in Volume I, Chapters 3 and 4.
Keynes centered his analysis on the effects of government tax and spending changes on disposable or after-tax income and on the “propensity” of individuals to consume a certain fraction of their disposable income. This propensity is hard-wired into people’s brains and rules out any notion that it might be in a person’s interest to adjust the fraction of his income allocated to consumption or saving according to his preferences, personal discount rate, or intertemporal elasticity of substitution. In the Keynesian model, a rise in G or a fall in T will increase disposable income and cause consumption to rise as determined by people’s built-in propensity to consume.
The assumption about a built-in propensity to consume is at the heart of Keynes’s argument that an increase in government spending will expand the economy. Government, according to Keynesian thinking, uses its powers to tax and spend in order to influence aggregate economic activity, particularly, when adjustments in G or T influence consumption through the effects of those changes on disposable income.
A matter of particular interest is whether government deficits—budgets in which government spends more than it collects in revenue—positively or negatively affect economic activity. A major concern, as relayed by Alan Blinder several years ago in the Wall Street Journal and as noted in Volume I, Chapter 1, is that government deficits “crowd out” private saving and investment and, by doing so, impose a burden on future generations by leaving them with a shrunken capital stock and a reduced capacity to produce goods for personal consumption. Blinder has raised the same concern more recently in a commentary on the Tax Cuts and Jobs Act, signed into law by President Trump on December 22, 2017 (Blinder 2017).
The argument is that, while running a deficit will increase consumption, it might also crowd out investment by pushing up interest rates and by driving private saving away from investment and toward the purchase of bonds that the government issues in order to finance the deficit. Here we will show that increases in G can have this very effect but that any deficit run-up in the process has no bearing on the resulting crowding out. The idea that deficits crowd out investment stems from a variation on the Keynesian model, which is recognized later.
The Effect of Government Taxes on Consumption and Saving
Recall the formulas for the present value of consumption and income over the full length of the planning period (which we wrote down for Eve in Chapter 4):
Equations (2.1) to (2.4) assume that decisions to consume or save are made on the basis of the individual’s optimization calculus. It is as if the individual decides, first, how much income he wants to make going out to the future and then, second, how much of the present value of that income to allocate to consumption and how much to saving, given r and his personal IES and r. Therefore, the first decision is to decide how much income to receive. The second is to decide how much to consume and save, given the income that the individual will receive.
Now we bring taxes into the individual’s optimization calculus. In Volume I, Chapters 3 and 4, there were no taxes or government transfer payments to concern ourselves with, but all that changes now. The individual now pays a tax less any transfer payment from government—a “net tax”—each year over this planning horizon. Let’s designate that net tax as taxt, where the subscript t refers to any year from the current year 1 to the end year n − 1. We then rewrite the individual’s lifetime budget constraint as
or as
where PVyat is the present value of future income after net taxes.
Thus the individual’s optimization problem is to maximize the present value of lifetime utility subject to
The individual sets current consumption equal to the fraction, v, of the present value of his after-tax income that satisfies his optimization calculus, as worked out in Chapter 4:
and the saving rate is
Because we want to think only in terms of income effects, it is useful to suppose that government fixes the size of tax for every individual by using some method that ignores how much income the individual chooses to receive. We could pretend that government decides by lottery how much to tax Adam or how big a transfer to provide Eve. It then sends out bills or checks depending on how the lottery turns out for each person. To further simplify the analysis, we could assume that, once the government decides on how to apportion the current year’s T among individual taxpayers, it apportions future taxes and transfers in a similar fashion. If Adam owes the government X% of a given year’s tax revenue or receives Y% of its transfer payments, he will pay or receive the same percentage under different assumptions about the size of T for every future year.
Under these assumptions, the total amount of money collected in taxes minus the amount paid out in transfer payments, T, in period t is
summed across all individuals for each year t.
Now it’s time to reintroduce government purchases G. We denote government purchases for a given year, t, as Gt. Then the present value of government purchases is
and the present value of net taxes is
Government purchases need not match net taxes in any given year, but, we assume, the present value of government purchases must equal the present value of net taxes over the period 1 through n−1, which is to say that
This means that, in the long run, government balances its budget or, we could say that, sooner or later, government pays for all its spending through current or future tax levies.
Given equations (2.12) and (2.13), we can see that any change in government purchases, ΔG, will cause the present value of government purchases to change and simultaneously require the present value of net taxes to change by the same amount as the change in the present value of government purchases. Let’s denote a change in government purchases in a given year as ΔGt and a change in net taxes as ΔTt.
The change in the present value of government purchases is
The corresponding change in the present value of net taxes is then
By necessity,
With all this established, we have to think about what we mean by fiscal policy. The usual meaning is policy regarding government purchases and net taxes in the near term, which, for our purposes, is the current year. Here, we take a longer term view and consider also how government might choose to make a permanent adjustments in its purchases of goods and services.
It is useful to imagine a rule that calls for maintenance of the status quo, which is to assume no change in government purchases, unless there would be gains to the economy from making some adjustment, temporary or permanent, in those purchases. The question, then, is how a change from that status quo would affect the economy.
But what is the “status quo”? The answer depends on what government can and can’t control. We can think of government purchases as an “exogenous” variable affecting the economy: Government expenditures on defense, schools, roads, and so forth depend on what government chooses to provide in the way of services related to those goods. The state of the economy, then, has no effect on G. If government uses the lottery system hypothesized here, then T is exogenous. Government has complete control over both G and T.
In reality, both G and T are partly “endogenous”: The size of the actual tax liabilities and actual transfer payments at the individual level, and therefore the size of T at the aggregate level, depend in part on the state of the economy and on how individuals react to changes in the laws relating to the imposition of burdens through taxes and the conferral of benefits through transfer payments. The same goes for G, though to a lesser extent.
We ignore this complexity here by adopting the assumption that, over the long run, government simply matches its purchases with sufficient net revenue (revenue left over after transfer payments) to pay for what it spends on goods and services and that it does so without creating any substitution effects on the work–leisure and consumption–saving decisions. So let’s begin by assuming that government has set current period G at some baseline level and then considers making changes in G.
A Permanent Change in Government Purchases
Suppose then that government makes a permanent change in its purchases, ΔG. As n gets very large, equation (2.15) for ΔPVG becomes
How do individual taxpayers adjust to this change? We know that, because the present value of government purchases must equal the present value of net taxes, the change in the present value of government purchases must be matched by an equal change in the present value of net taxes, so that
at the aggregate level, and
at the individual level.
We assume that the individual pays taxes equal to some fixed fraction of T. Let’s call that fraction f. Then
The government must set the change in each taxpayer’s net payment,
tax, for each year over the future so as to satisfy equation (2.19). The individual taxpayer will then adjust his consumption in line with the change in the present value of his after-tax income.
Thus,
and
The individual adjusts his consumption to match the change in the present value of his after-tax income, as determined by his lifetime change in taxes, as necessitated, in turn, by the government’s budget constraint. What happens to saving in a given year depends in part on how much the government decides to tax the individual in that year.
Suppose, then, that the individual reduces consumption by $100 (which is just his share of the aggregate tax bill for the increased G) in response to a rise in G, so that Δct = –$100. If the government doesn’t raise the individual’s taxes at all, his saving will rise by $100. Or if the government raises his taxes by $100, his saving will remain constant, as the individual will use money previously spent on personal consumption to pay the new taxes. Note that it doesn’t matter to the individual if the new spending is allocated for a purpose that he can substitute for his consumption or if the new spending is going to be entirely wasteful.
The change in consumption for the economy will be
where ΔPVyat is the aggregate change in the present value of after-tax incomes that results from the change in government purchases and therefore government taxes over the planning period:
Now let Yt before the change in G, be
Or, given our assumption about the equality of net exports and net foreign investment,
From equation (2.28) we know that the change in G does not cause a change in Yt. Production that previously went into personal consumption is now allocated to government purchases. It is possible that Yt will change as a result of this reallocation, which we take up later. But for now, let’s assume that
In this case, there must be a change in (St + Tt) that just matches the change in Ct. Say, for example, that Ct falls by $1,000. Then (St + Tt) must rise by $1,000. If the government keeps taxes constant, so that it increases its deficit by $1,000, saving must rise by $1,000. If it raises taxes by $1,000, saving need not change at all. Or if it raises taxes by $500, saving must rise by $500. The only condition is that the sum of the two changes equal the change in consumption.
A Temporary Change in Government Purchases
Now suppose that government purchases change only in year t. For each taxpayer,
Now
and
What happens at the aggregate level depends on the size of ΔG.
Because
Then
and
Thus also, if ΔYt = 0,
The sum of ΔIt and ΔNFIt must equal the negative of ΔGt(1 − v). (It + NXt) must fall if Gt rises and rise if Gt falls. Thus, if r = 5% and if Gt rises by $1,000, It + NFIt must fall by $952.38. If Gt falls by $1,000, It + NFIt must rise by $952.38. If Gt rises, there is crowding out of investment. If it falls, there is crowding in of investment. Note that this result is independent of the extent to which the change in Gt is accompanied by a rise in the deficit or surplus.
What happens to saving? Referring to equation (2.38), suppose that ΔGt = $1,000 and that the government finances the increase in government purchases by imposing $1,000 more in taxes in period t. In this case, the existing deficit does not change. If r = 5%, taxpayers will reduce their saving by $952.38 in order to free up the revenue to pay the new taxes. They will pay the remaining $47.62 in taxes out of funds diverted from current consumption. They will reduce their consumption by $47.62 in year t and in every year in the future, rather than endure paying the $1,000 in new taxes all at once in period t.
Now suppose the government does not raise taxes at all, thus adding $1,000 to the deficit. Saving will rise by $47.62. The government will raise the $1,000 it needs to pay for the increase in Gt by selling $1,000 in bonds. Taxpayers will divert $952.38 in current saving from the purchase of bonds and other financial instruments issued by private sector entities to the purchase of the new government bonds. They will combine that $952.38 with the $47.62 that they free up by reducing consumption to buy the $1,000 in bonds the government will issue to finance the deficit. As before, consumption falls by only $47.62 a year as consumers spread the impact of the new government spending over their future.
What this proves is that it doesn’t matter if government adjusts net taxes to match a temporary increase in government purchases. There will be a crowding out of investment but only because government increases purchases temporarily, not permanently, thus inducing taxpayers to spread the resulting hit to their income over their future. The interest rate r will rise in the process. If the government raises taxes to match a temporary increase in purchases, r will rise as aggregate saving falls. Likewise, if it engages in deficit financing, r will rise as savers replace existing saving instruments with the purchase of the bonds issued by government to pay for the new spending.
Figure 2.1 charts government purchases as a fraction of GDP against gross investment as a fraction of GDP from 1935 to 2016. It shows a substantial crowding out of gross investment during World War II and during the fiscal stimulus of 2009–10. Figure 2.2 shows that the temporary spending was financed largely by government borrowing.
Figure 2.1 Crowding out of gross investment
Source: U.S. Bureau of Economic Analysis.
Figure 2.2 U.S. deficits and gross investment
Source: U.S. Bureau of Economic Analysis and Office of Management and Budget.
A Change in Taxes
Now suppose that the government decides to change aggregate taxes by the amount ΔTt. If ΔTt is negative, which means the government cuts taxes (and/or increases transfer payments), it must sell bonds matching the amount ΔTt and must service the bonds over periods 2 through n, raising taxes in increments as ΔTt+1, ΔTt+3, …, ΔTn-1, in order to satisfy equation (2.17).
By assumption,
and, at the taxpayer level,
Thus,
and
Because
Hence, also,
and
Saving simply adjusts to the change in taxes.
Setting Tax Rates
Here we tie in the analysis from previous chapters where we recognize that actual taxes are imposed on choices about economic activity—choices about working, consuming, and saving. Government does not send out bills as credit card companies do but collects revenue by taxing certain activities (e.g., earning income) or certain asset holdings (e.g., holding property). Nor does it typically send out checks to individuals as bonuses for being good citizens. Rather, it sends out checks to individuals that operate mainly as rewards for not earning income. The question is how do we amend the foregoing analysis to consider this reality?
In reality, when the government decides to spend more on purchases or transfer payments, it must eventually recalibrate existing tax rules to raise the amount of revenue needed to finance the new spending (recognizing that it might spread out the process of raising this revenue over a long period of time, so that interim deficits may result). That process must take into account also how the expenditure of the taxes raised will affect individual choices, depending in part on whether taxpayers consider the expenditure to be “useful” or whether eligibility for a transfer payment depends on those choices. We have already worked through this consideration in Volume I, Chapter 7.
If government just randomly sent out bills to pay for what it spent (as assumed in the preceding analysis), none of this would matter. People would simply adjust their consumption accordingly, as described previously. But because it collects revenue by taxing choices, it will, as we have seen, affect those choices. Suppose, for example, that government commits to a stream of new spending on facilities like the Minute Clinics mentioned in Volume I, Chapter 8 and that it decides to raise the revenue by enacting a permanent surcharge on labor income.
The surcharge will have to be set in a way that recognizes the opposing consequences of the income and substitution effects. The strength of the income effect will vary inversely with the value of the new government clinics to consumers. Consumers will not feel that the tax made them poorer if the new clinics provide services comparable to or better than the CVS clinics that the new clinics displace. The tax rate will have to be set high enough to compensate for the negative effect on revenue of the taxpayer’s inclination to substitute leisure for work but there will be little or no sacrifice of leisure to compensate for the burden of the new tax on income. In effect, consumers will simply pay government for a service for which it previously paid a private provider and, to that extent, suffer no loss in utility.
In reality, of course, it wouldn’t simply be a matter of the quality of the government clinic’s services versus those provided by CVS. Because the use of private sector clinics requires co-payments and becomes reflected in insurance rates, there exists a rationing mechanism for their use. If the government just opens the doors of its new clinics and treats people free of charge, the clinics are likely to be overused and, ultimately, become costlier. On the other hand, as some would argue, the government clinics could operate more cheaply by eliminating the need for an insurance company and, perhaps, by extracting other cost savings.
But let’s not get buried in a discussion of socialized versus free-market health care. Let’s just assume that people who previously went to CVS for care now just go to the new “US Care” clinics with no loss in services. They simply, in effect, send money to government that they previously sent to CVS and to their insurance providers, but experience no income effect or on, that account, an incentive to work more. There will remain only a substitution effect and the resulting incentive to work less, along with the resulting necessity of setting the tax rate high enough to overcome the loss in revenue owed to the reduction in work effort. The result is a strong negative effect on work effort, labor income, and GDP. Whether the new spending ends up being good or bad for the economy would depend on the trade-off between whatever efficiencies the government clinics could provide and the negative effect on GDP of the new tax.
We can suppose conversely that the new spending is entirely wasteful. Then taxpayers send money to the government that would have gone to personal consumption and do so entirely as a sacrifice. The income effect at least partly offsets the substitution effect, reducing the tax rate increase needed to finance the new spending. The effect on labor income and GDP is small inasmuch as the reduction in work effort is small, but consumers suffer a reduction in utility as they sacrifice personal consumption for wasteful spending.
However, as these effects work themselves out, individuals will want to reduce their consumption permanently by the amount of the new spending. Insofar as the new spending provides a perfect substitute for consumption on some good (clinics), consumers get “free” what they used to pay for and reduce their personal expenditures on private clinics accordingly. Insofar as the new spending is wasteful, they will want to spread the burden of providing tax dollars to fund a wasteful expenditure evenly over their lifetimes.
Next, suppose again that government undertakes a once-and-for-all increase in purchases. Because individuals will want to spread the tax burden necessitated by this increase over their lifetimes, they will either reduce their saving (if the spending is financed by new taxes) or reallocate their existing saving to the purchases of government bonds (if the spending is financed by borrowing). Either way, r will rise and, with it the cost of capital, causing investment to fall. A further collateral effect is that the supply of labor will rise in response to the rise in r. Correspondingly, wage rates will fall, and GDP will rise. Ultimately, also, the increase in G, though temporary, will give rise to higher taxes, with the same substitution and income effects that are associated with higher taxes, which are made necessary by a permanent rise in G.
Finally, government can choose to reduce T, by either reducing tax rates or by loosening the eligibility requirements for transfer payments, or both. A reduction in tax rates encourages people to substitute work for leisure, because of the substitution effect, and to substitute leisure for work, because of the income effect. A reduction in T ultimately means a reduction in G. If the reduction in G means the elimination of government programs that provide useful alternatives to private consumption, the income effect will be weak and the reduction in tax revenue brought about by the reduction in tax rates will be greater than if the programs were wasteful. The effect of loosening the eligibility requirements for transfer payments on revenues depends on the new benefits provided and on any increase in tax rates (with all the corresponding collateral effects) that is necessitated by the loosening of requirements.
As it turns out, government consumption expenditures amount to about 50 percent of total government revenues. This means that about 50 percent of tax revenues go to providing services that taxpayers would otherwise have to obtain from the private sector. A disproportionate burden of federal income taxes is borne by only about half of taxpayers. The top 50 percent of income earners pay 93 percent of all federal income taxes.
The income effect of federal tax rates must therefore be high and the tax rates needed to raise the requisite revenue correspondingly low. People continue to work in order to provide for personal consumption as taxes imposed to pay for defense and the social safety net rise, blunting the inclination to increase leisure as the price of leisure falls. Another way of putting it is that government takes advantage of people’s demand for personal consumption in order to finance defense and social benefits.
Fiscal Policy as a Stimulus Tool
Chapter 3 of this volume shows how government can manipulate fiscal policy to restore the economy to full employment. This role for fiscal policy presents itself when the economy suffers long-lasting excess supply or excess demand. Excess supply manifests itself as a condition brought about by a failure of prices and nominal wages to fall in tandem with a fall in demand. Excess demand manifests itself as a condition brought about by a failure of prices and wages to rise in tandem with a rise in demand.
The required response by government to a condition involving excess supply is to increase government purchases and transfer payments and to reduce tax collections. The required response to a condition involving excess demand is to reduce government purchases and transfer payments and to increase tax collections. We go into detail in Chapter 3 of this volume.
The Budgetary Baseline
Now we need to refine what we mean when we refer to changes in government revenue and expenditures. There exists some baseline expectation of government spending and tax revenue, going forward. This expectation reflects both the assumed state of the economy and the statutes governing expenditures and taxes. The statutes governing expenditures and taxes have the effect of causing T to rise when the economy is expanding and to fall when the economy is contracting. T rises in an expanding economy as tax revenues rise and transfer payments fall. It falls in a contacting economy as tax revenues fall and transfer payments rise.
The baseline projection for federal government revenue and expenditures for 2018 to 2027, as offered by the Congressional Budget Office, is provided in Table 2.1 (Congressional Budget Office 2017, p. 13).
Projected revenues and expenditures vary with assumed changes in the economy. For its report, the CBO assumed that real GDP would grow by 2.0 percent in 2018, by 1.5 percent in 2019 to 2020, and by 1.9 percent in 2021 to 2027. In the foregoing discussion, an assumed ΔG or ΔT is intended to reflect a change in statutory provisions that determine spending and revenue, given existing assumptions about the state of the economy. Thus, any assumed change in T or G over the future would mean a change in the baseline assumptions behind Table 2.1. Because the government cannot simply decree a change in taxes or expenditures without considering the effects of a statutory change on the economy, it would be necessary to consider how any proposed change in tax or spending legislation would affect the baseline assumptions about GDP growth.
When federal officials “score” new spending legislation, they do not account for the future tax changes that will be eventually needed because of that legislation, unless the tax changes are included in the legislation. Table 2.1 therefore reflects CBO’s assumptions about current (June 2017) law and future economic conditions. It makes no provision for changes in tax law needed to pay for the projected expenditures.
While, under our assumptions, the government need not raise taxes enough to set the present value of revenue equal to the present value of expenditures in a 12-year frame, it must plan to do so within some time frame (50 years? 100 years?) in order to assure people buying government bonds that they will be paid. Thus, the current scoring mechanism is misleading in that it recognizes no limit to government’s ability to borrow. If the government’s ability to borrow were unlimited, no one would expect the government to pay its bills and no one would lend money to the government to finance its deficits.
Policymakers, pundits, and politicians often claim, in effect, that the real gauge of the ability of government to pay its bills is the fraction of GDP accounted for by the deficit. Table 2.1 provides the projected deficits as a fraction of GDP.
This table should be comforting, in that it has the deficit as percentage of GDP only slightly higher for 2027 than for 2017. What that means is that, given the assumptions relating to the growth of GDP over that period, the capacity of the economy to absorb the deficits implicit in budgeted spending plans will be about the same in nine years as it is now (2018). It does not, however, eliminate the eventual necessity of those tax increases or, therefore, of future government surpluses needed to pay the federal debt.
What are the chances that government will, in fact, impose the needed new taxes, given that it chooses to spend more? Figure 2.3 traces the history of the federal deficit as a fraction of GDP since 1929. There we see that the deficit exceeded 25 percent of GDP during World War II and came to about 10 percent during the Great Recession that marked the Obama presidency. The CBO, as we have seen, projects deficits in the 3 to 5 percent range over the next several years.
Table 2.1 CBO baseline projections of federal revenue and expenditures
Fiscal Years 2018 to 2027 ($ billions) |
||||||||||
Year |
2018 |
2019 |
2020 |
2021 |
2022 |
2023 |
2024 |
2025 |
2026 |
2027 |
Total Revenue |
3,531 |
3,687 |
3,853 |
4,011 |
4,178 |
4,361 |
4,545 |
4,742 |
4,948 |
5,158 |
Total Expenditures |
4,194 |
4,475 |
4,628 |
4,891 |
5,205 |
5,419 |
5,628 |
5,967 |
6,300 |
6,621 |
Deficit |
663 |
789 |
775 |
879 |
1,027 |
1,057 |
1,083 |
1,225 |
1,352 |
1,463 |
Deficit as a % of GDP |
2.8 |
3.3 |
3.6 |
4.0 |
4.5 |
4.4 |
4.3 |
4.7 |
5.0 |
5.2 |
What about federal debt? The total amount of debt owed by the U.S. government at the end of 2016 was $19.539 trillion. This figure is not meaningful, however, insofar as a substantial portion—$5.372 trillion—is owed by the government to itself, i.e., to government trust funds like Social Security. Social Security trust fund administrators use taxes (“contributions”) paid into the fund to pay benefits and to “buy” government bonds in amounts equal to the excess of Social Security “contributions” over payments. In this process, the trust fund has acquired over $5 trillion in what it reports as assets. As Social Security payouts rise, it will begin to spend down these assets in order to cover the difference between payouts and revenues from Social Security taxes.
To say that assets held by government trust funds represent government debt in any meaningful sense is, however, nonsense. Suppose that you have a mortgage on your house and that you keep a cookie jar where you put your spare change to pay for a night out for dinner. The mortgage you owe on your house is debt. So is the money you take out of the cookie jar to meet an occasional expense debt too? Maybe you “borrow” $25 from the cookie jar to pay your baby sitter. You can leave a note for $25 in the cookie jar as a reminder to repay your dinner account, but that $25 is not a debt. It is money you owe yourself, unlike the money you owe your mortgage holder.
Figure 2.3 Federal surplus or deficit (−) as a percent of GDP
Source: Federal Reserve Bank of St. Louis.
That is why, in talking about the burden of the debt, it is necessary to consider only that portion of federal debt that is owed the public and not government itself. That portion of the debt came to $14.167 billion at the end of 2016. Figure 2.4 traces federal debt owned by the public from 1939 to 2016 as a fraction of GDP. The debt as a fraction of GDP was an all-time high of about 105 percent during World War II, fell to a low of slightly more than 20 percent in the 1970s and then shot up again, largely as a result of the Great Contraction. Figure 2.5 tracks interest on the federal debt as a fraction of GDP. Both figures reflect the temporary run-up in government expenditures during the Great Contraction.
Figure 2.4 U.S. debt as a percent of GDP 1939 to 2016
Source: Federal Reserve Bank of St. Louis.
Figure 2.5 Interest on U.S. debt as a percent of GDP 1939 to 2016
Source: Federal Reserve Bank of St. Louis.
Multiplying both sides of the equation by 
we get
Subtracting, we get
Because
is very small, we can ignore that term. Solving,
