Report | Budget, Taxes, and Public Investment

Thinking seriously about what ‘fiscal responsibility’ should mean: Full employment and reduced inequality are the most important targets of fiscal policy

Download PDF

Press release

This report is one in a series1 exploring the economic theory and research underlying the model federal tax and budget plan we submitted as part of the “Solutions Initiative” project sponsored by the Peter G. Peterson Foundation.

Normally in Washington policymaking discussions, model tax and budget plans are constructed almost solely for the purpose of showing how a mix of tax increases and spending cuts can lead to lower budget deficits. In this view of the world, “fiscal responsibility” can be assessed entirely by examining the trajectory of the nation’s debt.

A new concept of “fiscal responsibility” needs to be embraced—one in which full employment and equitable distribution in the economy are primary goals of fiscal policy.

Our “Budget for Shared Prosperity” aims to radically change how “fiscal responsibility” should be evaluated. The overarching mandate we set for our budget is to solve the pressing economic problems facing low- and middle-income families. Bending revenue closer to spending in a spreadsheet is a fairly trivial exercise, whereas the real-world effects of changes in taxes and spending are often not trivial at all. For example, the poverty rate of elderly households fell extraordinarily rapidly as Social Security expenditures rose in the mid-20th century.2 Ignoring this tremendous progressive achievement and instead seeing Social Security spending as simply a budget line item that needs to be trimmed in order to move expenditures and revenue closer together would be an extraordinarily myopic way to think about economic policy.

The main threats to typical families’ economic success are (1) chronic downward pressure on aggregate demand (spending by households, businesses, and governments) that drags on economic growth and (2) ongoing high levels of inequality between those at the top of the wage distribution and those at the bottom and middle. Our plan aims to greatly ameliorate the dangers posed by both of these threats. Our budget plan is also able to put the economy’s debt-to-GDP ratio (“debt ratio”) on a downward path over time—but we emphasize that this is worth doing only if it simultaneously solves the more pressing problems facing low- and middle-income families.

Our plan achieves this set of goals by financing most of the federal government’s spending—even highly ambitious new spending programs—with tax revenues instead of debt. As long as the tax system is highly progressive, this pairing of new spending and higher taxes will boost the pace of growth in aggregate demand.

It is important to note, however, that not all spending needs to be financed by taxes instead of debt. During recessions, measures aimed at boosting recovery are more effective when debt-financed, and many public investments that generate large economic and social returns are also fine to finance with debt, even in nonrecessionary periods.

Our view is that the debt ratio is optimally reduced quite gradually over time. In our model budget, in fact, the debt ratio falls more quickly than we’d prefer—as we were given limited opportunities to iterate our revenue scores to hit a more precise path. If this ever became a real-world policy proposal, we would iterate our tax and spending levels a bit more to achieve more gradual debt reduction.

If policymakers are excessively scared of deficits, they will be constrained in responding to recessions and income inequality.

This report explains why we think a new concept of “fiscal responsibility” needs to be embraced—one in which full employment and equitable distribution in the economy are primary goals of fiscal policy. Too often the economics of deficits and debt have been characterized as indicating that rising debt ratios will quickly become economically ruinous, and that lowering the trajectory of debt justifies large economic sacrifices from low- and middle-income households. We review this economics (both theory and evidence) and highlight why this characterization is deeply mistaken. This is not nitpicking—if policymakers are excessively scared of deficits and overstate the damages they can cause, they will be highly constrained in responding to recessions and in constructing policies that can address income inequality. Given how important full employment and a more equitable distribution of income is for the economic health of most U.S. families, we should be careful about sacrificing progress in meeting these challenges in the service of “debt reduction by whatever means necessary.”

Fiscal priorities of the ‘Budget for Shared Prosperity’

We begin with a brief overview of the priorities reflected in the “Budget for Shared Prosperity.” In constructing our budget, we identified the main economic challenges facing ordinary families and explored how radical changes to federal tax and spending policies could help alleviate those challenges. These challenges and our policy responses (as reflected in our budget) are described below, in descending order of importance.

Fiscal priority one: Ensure genuine full employment

Economic challenge facing working families: Full employment is too rarely achieved, and is getting harder for policymakers to reliably achieve.

Policy response: The primary goal of fiscal policymakers should be to ensure genuine full employment. Given other economic trends in recent years (particularly a chronic shortfall of aggregate demand), this will likely require more expansionary fiscal policy than has characterized most of the post–World War II history of the U.S. economy.

The U.S. economy has suffered from a chronic shortfall of aggregate demand (spending by households, businesses, and governments) for most of the past two decades, even as interest rates have been kept extraordinarily low in an effort to spur this spending. This shortfall of demand is why unemployment has remained too high far too often in recent decades, and it is partly why recovery from the Great Recession was so slow. Excess unemployment and delayed recovery have badly hamstrung typical workers’ ability to leverage pay increases from employers and can hence largely explain why wage growth has been so sluggish in recent decades. Given this damage inflicted on households by demand shortfalls, the primary goal of fiscal policymakers must be to ensure that genuine full employment is maintained consistently and that fiscal policy is used to restore full employment as rapidly as possible when recessions hit.

Cumulatively between 2008 and 2017, the U.S. economy lost roughly $5 trillion in potential income (even by quite conservative estimates) and roughly 32 million job-years of potential employment, simply because aggregate demand did not keep pace with the economy’s underlying productive capacity.3 This aggregate demand shortfall is the source of the large spike and much shallower downward reversion of unemployment as potential workers—the most important part of the economy’s productive capacity—were idled simply because there was not enough spending in the economy.4

These losses are calculated simply by looking at the gap between actual GDP and potential GDP estimated by the Congressional Budget Office in 2019. But the demand shortfall also caused a long-term erosion of the economy’s underlying productive capacity. As slack demand and weak labor markets sapped workers’ bargaining power, firms had little incentive to continue chasing productivity-enhancing investment and processes to keep labor costs in check. The result was a sharp reduction in productivity growth. CBO’s estimate of potential output over the post–Great Recession period made in 2019 was far smaller than its own estimate made in 2008. If even half of the decline in the economy’s estimated potential capacity after 2008 was itself driven by the prolonged demand shortfall, then the cost of that shortfall rises to roughly $10 trillion of potential income.5 Figure A shows the devastating effect of this prolonged demand slump. It charts actual gross domestic product (GDP, a proxy for overall national income) as well as potential GDP as estimated before the Great Recession (in 2008) and potential GDP as estimated after the effects of the Great Recession took hold (calculated in 2019). The gap between actual GDP and both estimates of potential GDP is obvious, and this gap is the demand shortfall that led to the joblessness of the recession. The sharp downward revision in the estimates of potential GDP are (at least partly) the effect of slack demand in eroding the economy’s productive capacity.

Figure A

Demand shortfall causes actual GDP to lag potential---and then begins to erode this potential: Actual GDP and estimates of potential GDP made in 2008 and 2019

Actual Potential (2008 estimate) Potential (2019 estimate)
2000Q1 81.9954 80.1116 81.0617
2000Q2 83.4961 80.8493 81.9061
2000Q3 83.6088 81.5878 82.7390
2000Q4 84.1294 82.3213 83.5521
2001Q1 83.8901 83.0454 84.3283
2001Q2 84.3805 83.7475 85.0656
2001Q3 84.0303 84.4301 85.7749
2001Q4 84.2596 85.0882 86.4562
2002Q1 84.9964 85.7183 87.1036
2002Q2 85.5118 86.3272 87.7230
2002Q3 85.8913 86.9192 88.3253
2002Q4 86.0234 87.5002 88.9154
2003Q1 86.5015 88.0778 89.5099
2003Q2 87.2454 88.6520 90.0969
2003Q3 88.7268 89.2262 90.6832
2003Q4 89.7458 89.8030 91.2733
2004Q1 90.2251 90.3840 91.8742
2004Q2 90.9120 90.9726 92.4987
2004Q3 91.7719 91.5697 93.1264
2004Q4 92.6916 92.1744 93.7529
2005Q1 93.7175 92.7893 94.3761
2005Q2 94.1500 93.4135 94.9802
2005Q3 94.9893 94.0461 95.5678
2005Q4 95.5885 94.6864 96.1350
2006Q1 96.8596 95.3326 96.6633
2006Q2 97.0852 95.9847 97.1635
2006Q3 97.2360 96.6419 97.6472
2006Q4 98.0637 97.3051 98.1194
2007Q1 98.2946 97.9716 98.5910
2007Q2 98.8583 98.6441 99.0664
2007Q3 99.3944 99.3199 99.5361
2007Q4 100.0000 100.0000 100.0000
2008Q1 99.4259 100.6826 100.4620
2008Q2 99.9392 101.3661 100.9209
2008Q3 99.3973 102.0504 101.3631
2008Q4 97.2469 102.7356 101.7824
2009Q1 96.1551 103.4199 102.1642
2009Q2 96.0169 104.1042 102.4911
2009Q3 96.3662 104.7894 102.7919
2009Q4 97.4260 105.4754 103.0729
2010Q1 97.7991 106.1622 103.3316
2010Q2 98.7024 106.8491 103.5897
2010Q3 99.4294 107.5385 103.8504
2010Q4 99.9278 108.2304 104.1218
2011Q1 99.6880 108.9275 104.4354
2011Q2 100.4016 109.6312 104.7655
2011Q3 100.3728 110.3410 105.1122
2011Q4 101.5372 111.0575 105.4760
2012Q1 102.3318 111.7807 105.8527
2012Q2 102.7722 112.5091 106.2535
2012Q3 102.9107 113.2417 106.6690
2012Q4 103.0270 113.9768 107.0979
2013Q1 103.9402 114.7120 107.5408
2013Q2 104.0687 115.4479 107.9907
2013Q3 104.8843 116.1831 108.4483
2013Q4 105.7210 116.9182 108.9109
2014Q1 105.4550 117.6533 109.3717
2014Q2 106.7774 118.3868 109.8381
2014Q3 108.0681 119.1202 110.3084
2014Q4 108.5780 119.8528 110.7838
2015Q1 109.4705 120.5862 111.2650
2015Q2 110.3737 121.3188 111.7544
2015Q3 110.6393 122.0531 112.2451
2015Q4 110.7491 122.7891 112.7345
2016Q1 111.1759 123.5276 113.2195
2016Q2 111.8046 124.2687 113.6891
2016Q3 112.3385 125.0131 114.1569
2016Q4 112.8308 125.7618 114.6246
2017Q1 113.3300 126.5147 115.0911
2017Q2 114.1687 127.2710 115.5608
2017Q3 114.9653 128.0324 116.0432
2017Q4 115.6188 128.7989 116.5447
2018Q1 116.2549 129.5696 117.0813
2018Q2 117.4436 130.3471 117.6568
2018Q3 118.4595 131.1279 118.2558
2018Q4 119.2023 131.9130 118.8733
ChartData Download data

The data below can be saved or copied directly into Excel.

Note: Shaded areas denote recessions.

Source: Author’s analysis of data from the Bureau of Economic Analysis, “Table 1.1.6. Real Gross Domestic Product, Chained Dollars,” National Income and Product Accounts, accessed June 2019, and data from the Congressional Budget Office (on potential GDP) in The Budget and Economic Outlook: Fiscal Years 2008 to 2018, January 2008, and The Budget and Economic Outlook: 2019 to 2029, January 2019


Copy the code below to embed this chart on your website.

A key lesson from this prolonged episode of slack demand is that the Federal Reserve cannot simply jump-start a rapid recovery by lowering interest rates—short-term rates were held at essentially zero from the end of 2008 until the end of 2015, and the Fed also undertook measures to hold down longer-term rates. Low interest rates are especially weak tea for spurring recovery from deep recessions in the face of fiscal austerity—like that which characterized much of the post-2011 recovery. Since monetary policy has such limited scope to quickly end recessions and spur recovery, fiscal policy must remain expansionary until genuine full employment is reached. This type of anti-recession spending should be debt-financed when the economy is in outright recession.6 To this effect, the U.S. would benefit from much larger “automatic stabilizers”—spending increases or tax cuts that disproportionately benefit low- and moderate-income households and that kick in automatically, without the need for new legislation, when the economy slows or enters recession.7

However, keeping fiscal policy more expansionary than it has been in the recent past during nonrecessionary periods does not require large increases in budget deficits. Instead, greater public spending can be paired with progressive tax increases. Given that federal spending increases spur far more economic activity than is restrained by an equivalent dollar amount of tax increases aimed at rich households or corporations,8 fiscal policy that increases spending—and funds it with progressive taxation—can spur aggregate demand through “balanced budget multiplier” effects.

Fiscal priority two: Provide a bulwark against rising inequality by spending more and taxing economic ‘bads’

Economic challenge facing working families: In recent decades, income gains have been concentrated at the top and have largely bypassed typical households.

Policy response: The next most important goal of fiscal policy should be to significantly reduce or reverse the rise of inequality that has characterized recent decades, ensuring that income growth for the vast majority of U.S. households matches economywide rates of growth. As part of this effort, tax policy should be used aggressively to reduce economic “bads”—for example, greenhouse gas emissions, excess market power in the finance sector, and inequality more generally.

The U.S. economy has seen an enormous increase in inequality of market incomes over the past generation of economic life, as shown in Figure B. Market income essentially measures income with government transfer payments (like Social Security or Medicare or food stamps) excluded. Inequality in market incomes can be seen as what is generated in the private sector before the effects of the tax-and-transfer system are accounted for.9 For working-age households, the rise in inequality had by 2015 imposed a 43% “inequality tax” on their market incomes, reducing those incomes by roughly $24,000 relative to a scenario in which inequality had not grown since the 1970s. This increase in inequality has been driven by a redistribution of bargaining power and leverage, particularly in labor markets.

Figure B

Unequal growth of market-based incomes has harmed middle-class families: Actual growth in market incomes for the middle fifth of U.S. households, and growth that would have prevailed if these incomes had grown along with economywide average, 1979–2015

Middle-fifth market incomes, actual Middle-fifth market incomes, if they had grown at overall average rate
1979 $54,886 $54,886
1980 $52,301 $52,874
1981 $52,020 $52,779
1982 $50,721 $52,344
1983 $49,824 $52,824
1984 $53,027 $56,196
1985 $53,361 $56,947
1986 $54,231 $60,945
1987 $53,937 $59,453
1988 $54,931 $62,038
1989 $55,346 $62,396
1990 $55,580 $61,977
1991 $53,898 $60,317
1992 $54,167 $61,759
1993 $54,287 $61,905
1994 $54,822 $62,829
1995 $56,871 $65,421
1996 $57,459 $67,580
1997 $58,800 $70,782
1998 $61,058 $74,928
1999 $63,202 $79,037
2000 $63,589 $81,425
2001 $62,896 $76,611
2002 $60,772 $73,085
2003 $60,839 $74,383
2004 $62,779 $78,666
2005 $63,755 $82,746
2006 $64,155 $85,763
2007 $66,630 $87,949
2008 $63,800 $81,749
2009 $60,863 $75,558
2010 $60,579 $77,367
2011 $59,710 $76,887
2012 $59,746 $80,577
2013 $61,300 $79,737
2014 $62,694 $82,812
2015 $64,565 $85,526
ChartData Download data

The data below can be saved or copied directly into Excel.

Notes: Market incomes are incomes before government taxes and transfers.

Source: Congressional Budget Office, The Distribution of Household Income, 2015, November 2018

Copy the code below to embed this chart on your website.

Fiscal policy changes can help ameliorate these trends in market income inequality through many different channels.

First, increases in federal spending can direct resources toward low- and moderate-income households that have not shared proportionately in the fruits of overall growth. This spending provides the largest welfare gains if it is focused in areas where public spending can be more efficient than private spending. Key examples of areas where public spending provides these clear efficiency gains are infrastructure investments, investments in early childhood and education, and social insurance (e.g., financing health care and retirement income). However, even pure cash transfers (like refundable tax credits) to low-income families have been found to have strong investment-like qualities, with social benefits stemming from this spending persisting for years. For example, children in low-income families that receive larger cash transfers see greater scholastic achievement and have higher incomes when they are adults than children in low-income families that receive smaller or no cash transfers. In short, after ensuring full employment, the primary goal of fiscal policymakers should be to figure out what spending is needed to ensure meaningful economic security and opportunity for all. It is important to note that while current federal spending is quite well-targeted to fight inequality, as the lion’s share of benefits goes to families with incomes below the median, the overall level of this spending is quite stingy when compared with that of other rich countries around the world.10

Second, taxes can be deployed strategically to combat economic “bads” that place excess burdens on nonrich families. We see the key economic bads to be confronted as global climate change, the excess power of finance, and inequality more broadly. To combat these, our budget includes a substantial carbon tax (that is rebated in lump-sum fashion to either raise household incomes or at least keep them constant even as households’ incentive to consume carbon-intensive goods is blunted), a small financial transactions tax to displace low-value trades, and substantially higher top tax rates (particularly on capital-based incomes) that can blunt the incentive of privileged economic actors (think CEOs of large companies) to bargain as hard as possible against their own workers and other stakeholders when dividing up the income generated by an enterprise.11 The federal tax system is currently progressive, but combined with far less progressive state and local tax systems, the overall U.S. tax system does little to combat inequality. Boosting the progressivity of federal taxes is the shortest route to using taxes more aggressively to keep inequality in check.

Fiscal priority three: After meeting other objectives, minimize risks stemming from higher debt levels

Economic challenge facing working families: When the economy is at genuine full employment (but only then), larger deficits can push up interest rates and “crowd out” productive investment. Further—while the probability of interest rates rising looks quite small currently—an unexpected future rise in interest rates would present greater challenges if debt ratios were high than if they were low.

Policy response: During normal economic times—and particularly when the economy is close to full employment—most of the federal government’s spending should be financed with taxes instead of debt. However, the chronic shortfall of demand that has characterized the past two decades argues that larger budget deficits should be tolerated in the medium term. For this reason, we do not recommend attempting to reduce debt ratios, and would particularly caution against spending cuts. Instead, we recommend a more gradual approach that focuses on raising revenue progressively rather than cutting spending. In the long run, the overall federal budget does not need to be balanced or run surpluses, but can instead run modestly sized deficits and yet keep debt ratios stable or even declining.

Finally, subject to meeting these other objectives—ensuring full employment, providing sufficient spending to ensure that the fruits of economic growth are broadly shared, and using taxes to combat economic “bads”—the overall level of taxation should be set high enough to keep the risks of high debt ratios (federal debt divided by the nation’s GDP) manageable. We think that these debt risks are often significantly overstated, and we certainly do not think that the nation needs to run persistent budget surpluses or that there is any magical debt threshold above which growth predictably suffers.

But—particularly when progressive taxes and high spending levels combine to make fiscal policy expansionary—it is useful to keep deficits small enough to allow the ratio of overall federal debt to GDP to flatten or even decline during times of full employment. Recently, debates among economists about the proper level of federal budget deficits and debt have become much more nuanced than they have been in the past.12 However, debates among too many policymakers remain mired in thinking that running deficits, except when the economy is in outright recession, is always and everywhere bad.13 This kind of thinking is bad economics, and the rest of this report sketches out what we think is a smart approach to deficits and debt.

Putting deficit reduction in its place: On the list of economic concerns, but not at the top

Next we turn to a deeper exploration of why we believe that achieving full employment and lower inequality are necessary prerequisites to pursuing debt reduction. We provide an overview of the economics of deficits and debt that justifies where we place debt reduction in our list of fiscal priorities.

This is not a call to completely ignore the potential challenges imposed by deficits and debt, but it is a call for better understanding the underlying economics and the trends in real-world data surrounding deficits and debt—so that we can more meaningfully infer just how worried we should be about debt.

What do deficits do?

Federal budget deficits are an excess of federal spending over revenue. They are generally financed by borrowing: Bonds are issued to cover the deficit, and these bonds are bought by households or firms in the U.S. private sector or by foreign entities (households, businesses, or governments).

Deficits stimulate aggregate demand—which is why they are recommended during a recession

All else equal, an increase in the federal budget deficit from one period to the next stimulates aggregate demand in the economy. The spending done by the federal government either stimulates demand directly (say, when the money is spent to build a road) or indirectly by transferring resources to households who can then undertake more consumption spending (say, by sending out Social Security checks). Taxes, all else equal, reduce demand by lowering disposable income. If federal spending rises faster than taxes and deficits increase, economywide aggregate demand gets a boost. Higher demand in turn induces firms throughout the economy to hire more people and this leads in turn to lower unemployment. As roads are built or as more people receive transfers from the federal government, this leads to greater demand for concrete, steel, capital equipment, restaurant meals, and so on. This in turn leads to increased demand for workers to produce these goods and services. This boost to demand stemming from larger budget deficits is precisely why it is almost universally recommended to run larger deficits when the economy enters recession and is starved of demand.

Deficit spending can be more risky when the economy is at full employment

This stimulus to aggregate demand, however, is also the reason many worry about potential malign effects of deficits. If deficits rise when the economy is already characterized by full employment—meaning economywide resources like labor and capital are already fully utilized—then the boost to aggregate demand created by these larger deficits can be accommodated only by crowding out already-occurring spending. In the current institutional framework of the U.S. economy, the way this crowding out is effected is through higher interest rates that are engineered by the Federal Reserve.

The Fed raises rates in the face of higher budget deficits run during times of full employment because they worry that the boost to aggregate demand caused by larger deficits could stoke inflationary pressures. As an example, imagine that the federal government sent $1,000 debit cards to all households, financed by debt and expiring in one month. If this happened when the economy was already at full employment, there would be no workers or equipment available to produce the extra goods and services needed to satisfy the new demand as households went out and tried to spend these debit cards. In turn, greater demand meeting unchanging supply would lead to general increases in prices, or inflation.

One of the Fed’s two institutional mandates is to keep inflation at roughly 2% annually.14 If deficit-induced increases in demand threatened to push inflation too far above this target, they would raise interest rates. Higher interest rates would increase the costs for durable goods purchases financed with debt (think autos and houses and washing machines), thereby lowering demand for these goods. Higher rates would also make it harder for firms to borrow money to invest in tangible plants and equipment (think of a restaurant owner deciding whether or not to take out a loan to replace the furniture in his dining room). Finally, higher interest rates increase foreign demand for U.S. assets and thereby increase demand for U.S. dollars in global markets. The resulting stronger dollar makes U.S. exports expensive in global markets and foreign imports cheap for U.S households, which increases the trade deficit and reduces demand for U.S.-produced output.

In short, larger budget deficits run when the economy is already at full employment can threaten to crowd out investment in tangible capital by businesses, and can lead to higher foreign ownership of U.S. assets. This in turn can potentially lead to slower productivity growth (as the investment slowdown deprives U.S. workers of more up-to-date equipment), and it also means that more of the income that is generated within the U.S. is leaving the country to pay foreign owners of U.S.-based assets. Both of these channels can make future generations poorer than they would have been absent the increase in the deficit.

The data signature of deficits that are too high is far more likely to be high interest rates than high inflation

Occasionally the claim is made that the sole indicator that budget deficits have become too large is accelerating inflation. This claim is often made by those identifying with the Modern Monetary Theory (MMT) school. For example, in a February 2019 episode of the Ezra Klein Show (Klein 2019), leading MMT proponent Stephanie Kelton argued:

If there was a long-term debt problem, the way that you would know that is that there would be a long-term inflation problem. You would have some credible long-term inflation forecast, coming from the Fed or TIPS [Treasury Inflation-Protected Securities] or wherever…there would be evidence that there’s a perception that there’s a heightened risk of long-term inflation, and you just don’t see it anywhere and absent that, for me, there is no evidence of a long-term debt problem.

This seems clearly wrong and seemingly rests on an assumption that the Federal Reserve would not raise interest rates in response to a (real or perceived) excess of aggregate demand over productive capacity caused by larger deficits. For example, the Congressional Budget Office (CBO) annually produces its Long-Term Budget Outlook report (e.g., CBO 2019). This outlook often projects large increases in debt-to-GDP in the coming decades, and it projects that if these increases occur when the economy has no productive slack, then the excess of aggregate demand growth over growth in the economy’s productive capacity will result in interest rates that rise and crowd out private investment, hence slowing long-term growth in the economy’s potential. This outlook does not project rising inflation in the face of the excess growth of aggregate demand over productive capacity even though this excess growth is clearly being forecast. The reason CBO does not project accelerating inflation in the face of excess growth of aggregate demand caused by deficits is obvious: They assume the Fed will still exist in the future and that it will be successful in meeting its inflation mandate, and this success will rest on higher interest rates.

Often proponents of MMT express a preference for taking the primary mandate for controlling inflation away from the Fed; these proponents even argue that the proper stance of Fed interest rate policy is to hold short-term rates at zero, always and everywhere. Regardless of what one thinks about taking the primary mandate for inflation control away from the Fed and handing it to fiscal policymakers, we should be clear that this would be a big change from the institutional structures that currently exist in the American economy. Given the existence of these institutions, economic predictions based on the idea that the Fed would not react to a (real or perceived) excess of aggregate demand over productive capacity by raising interest rates would be wrong.

When there is productive slack in the economy—and it is recognized by the Fed—deficits don’t hurt

None of the malign effects of deficits happen if they are incurred when the economy has excess productive slack (jobless potential workers or excess capital capacity). Imagine if the $1,000 time-limited debit cards (financed by federal debt) were mailed to U.S. households when unemployment was high and businesses had lots of excess capacity. In this case, as households rushed out to spend this extra money, there would be idle resources that could be mobilized to produce the extra goods and services that are newly demanded. This implies that it is output (GDP), not prices, that will be pushed up by the boost to desired spending. Since this does not threaten to push up inflation so long as the economy retains productive slack, there is no need for the Fed to raise interest rates, and the channels of crowding out do not operate.

In fact, because larger deficits ensure a steadier stream of customers, these deficits can induce firms to make sure their investment in tangible plants and equipment does not fall off the way it would in an economic downturn. In short, these larger deficits can “crowd in” private investment.

There is two-way causality between deficits and larger economic trends

So far we have been discussing how a change in the size of the budget deficit affects aggregate demand. However, the relationship between deficits and aggregate demand (and hence the pace of overall growth) is decidedly two-way. If a negative shock to private spending causes the economy to slow or enter recession, then this will mechanically feed back to larger federal budget deficits. As private-sector incomes fall, tax collections fall. And as more people become eligible for safety net programs as the private sector contracts, federal spending rises.

This increase in the deficit that results from slowing economic conditions is unambiguously a good thing. Falling tax burdens and rising safety net spending constitute “automatic stabilizers” that support aggregate demand when private-sector trends undermine economywide spending. A key issue that fiscal policymakers should strive to address going forward is making these automatic stabilizers larger and longer-lasting as the economy slows. While automatic stabilizers and discretionary fiscal stimulus both worked to support the economy during the 18 months of the Great Recession, fiscal policy turned sharply contractionary far too early in the subsequent recovery.15

Public investments should often be deficit-financed even during periods of full employment

As we have discussed above, when there is productive slack in the economy, any increase of public spending that is not financed by higher taxes will boost spending and create jobs. If some of this increased public spending takes the form of investment—spending that yields an asset that will make society richer in the future—then so much the better for sustaining economic growth. Obvious candidates for public investment include core infrastructure and early child care and education. Further, growing research indicates that much federal spending that is not technically classified as “investment” actually does have strong characteristics of making future generations more productive. For example, the Supplemental Nutrition Assistance Program (SNAP, sometimes referred to as food stamps) and Medicaid (a program that provides health insurance to poor families) have both been shown to significantly increase future health and labor market outcomes for children who benefited from these programs.16

Furthermore, even when the economy is at full employment, public investments that are debt-financed likely do not damage future economic growth prospects and may well even boost them. Remember, the key reason why larger budget deficits run during times of full employment potentially damage future growth is because they can push up interest rates and crowd out private-sector investments in productive plants and equipment. But if the deficits that crowd out private investment are themselves financing productive public investments, then future growth may well not be slower. If the debt-financed public investments are as productive as the private-sector investments they might displace, then they will not drag on growth. Given that the last generation has seen a sharp slowdown in the accumulation of public capital, public investments may well have a higher social return than private investments.17

The intuition that it might be fine to finance public investments with debt even when there is no economywide demand slack is most clear when the public investments are aimed at slowing greenhouse gas emissions. There is a strong case to be made that the coming decades will need to see significant growth in green public investments if we are to forestall climate change (think building retrofits for energy efficiency, “smart grid” investments, and public transit expansion, among others). If the bulk of these investments are relatively time-limited—e.g., a one-off “big push” to replace dirty with clean energy sources and transportation options—then financing these investments with debt seems an obvious choice. By all economic forecasts, future generations will on average be far richer than the current generation but for the effects of climate change. If debt-financing green investments today lets us borrow modestly from future generations but spare them from inheriting a world that has not addressed climate change, then this is a staggeringly good deal for these future generations, and one that allows the current generation to make green investments without sacrificing current living standards.

If deficits run during bad times don’t hurt the economy, can their legacy of higher debt be a problem?

The textbook economics of running budget deficits when the economy has productive slack is clear: Budget deficits push the economy closer to full employment and are actively beneficial. However, even if deficits run when the economy is suffering from less-than-full employment don’t cause harm in real time, can the accumulated stock of debt that these deficits leave behind cause problems in the future?

In theory, this accumulated debt could cause future problems. In practice, the key thing to remember is that a government’s budget is not like a household’s budget: A government has no finite life span over which debt must be retired. Hence, governments can always roll over old debts that have been accumulated and are coming due by borrowing more. For a household, this constant rollover of debt would eventually be unsustainable, but this is not the case for a long-lived government. So long as the rate of GDP growth exceeds the interest rate charged on the debt, then the ratio of accumulated debt to GDP will fall over time, steadily eroding the share of national income that must go to interest payments on this debt. Over most of U.S. history, GDP growth rates have indeed exceeded the interest rates on public debt. Interest rates exceeded growth rates for most of the 1980s and 1990s and early 2000s, but fell back below growth rates in the aftermath of the Great Recession. Figure C shows the behavior of both of these variables over the long run.

Figure C

Assumptions that interest rates will exceed growth rates are usually wrong: Five-year average growth rate of GDP (G) and effective interest rates on federal debt (R), 1913–2018

G minus R Growth rate (G) Interest rate (R)
1913 1.79% 3.70% 1.90%
1914 2.06% 3.97% 1.91%
1915 0.86% 2.79% 1.93%
1916 3.71% 5.65% 1.94%
1917 7.46% 9.24% 1.78%
1918 10.80% 12.70% 1.89%
1919 13.51% 15.68% 2.17%
1920 14.82% 17.43% 2.62%
1921 10.28% 13.36% 3.08%
1922 3.05% 6.76% 3.71%
1923 -0.59% 3.57% 4.16%
1924 -1.86% 2.52% 4.38%
1925 -2.91% 1.49% 4.41%
1926 -1.22% 3.19% 4.41%
1927 1.18% 5.56% 4.39%
1928 -0.26% 4.00% 4.26%
1929 -1.05% 3.13% 4.18%
1930 -2.20% 1.93% 4.13%
1931 -6.00% -1.96% 4.04%
1932 -10.21% -6.34% 3.87%
1933 -13.05% -9.34% 3.71%
1934 -12.45% -8.92% 3.53%
1935 -8.87% -5.56% 3.31%
1936 -3.27% -0.23% 3.04%
1937 3.10% 5.97% 2.87%
1938 6.49% 9.20% 2.71%
1939 5.43% 8.01% 2.58%
1940 4.47% 6.96% 2.49%
1941 5.62% 8.07% 2.44%
1942 8.82% 11.14% 2.32%
1943 13.71% 15.85% 2.14%
1944 17.01% 18.96% 1.95%
1945 16.67% 18.43% 1.75%
1946 13.22% 14.90% 1.68%
1947 8.76% 10.44% 1.68%
1948 5.64% 7.41% 1.76%
1949 3.43% 5.34% 1.91%
1950 2.26% 4.32% 2.06%
1951 5.49% 7.65% 2.15%
1952 6.29% 8.53% 2.24%
1953 5.64% 7.96% 2.32%
1954 4.80% 7.14% 2.34%
1955 5.65% 7.97% 2.32%
1956 3.73% 6.08% 2.35%
1957 2.99% 5.39% 2.40%
1958 2.00% 4.42% 2.43%
1959 2.97% 5.47% 2.50%
1960 2.98% 5.64% 2.67%
1961 1.75% 4.54% 2.79%
1962 1.94% 4.82% 2.89%
1963 2.49% 5.49% 3.00%
1964 2.42% 5.58% 3.16%
1965 2.60% 5.86% 3.25%
1966 3.99% 7.39% 3.40%
1967 3.79% 7.40% 3.61%
1968 3.96% 7.75% 3.79%
1969 4.16% 8.19% 4.03%
1970 3.73% 8.11% 4.38%
1971 2.73% 7.44% 4.71%
1972 2.85% 7.79% 4.95%
1973 3.36% 8.57% 5.22%
1974 3.05% 8.65% 5.60%
1975 3.10% 8.96% 5.86%
1976 3.76% 9.85% 6.10%
1977 3.86% 10.18% 6.32%
1978 3.82% 10.40% 6.58%
1979 4.18% 11.04% 6.86%
1980 3.76% 11.14% 7.38%
1981 3.11% 11.35% 8.24%
1982 1.00% 10.35% 9.35%
1983 -0.89% 9.26% 10.15%
1984 -1.74% 9.02% 10.76%
1985 -2.23% 8.86% 11.09%
1986 -3.20% 7.67% 10.86%
1987 -2.61% 7.63% 10.24%
1988 -1.98% 7.82% 9.81%
1989 -2.28% 7.11% 9.38%
1990 -2.21% 6.74% 8.95%
1991 -2.51% 6.15% 8.66%
1992 -2.26% 6.13% 8.40%
1993 -2.29% 5.69% 7.98%
1994 -2.18% 5.27% 7.45%
1995 -1.94% 5.10% 7.04%
1996 -1.21% 5.48% 6.69%
1997 -0.84% 5.68% 6.52%
1998 -0.83% 5.68% 6.51%
1999 -0.79% 5.73% 6.53%
2000 -0.49% 6.00% 6.49%
2001 -0.64% 5.78% 6.42%
2002 -1.02% 5.11% 6.13%
2003 -0.83% 4.83% 5.66%
2004 -0.27% 4.93% 5.20%
2005 0.17% 4.91% 4.74%
2006 0.88% 5.32% 4.44%
2007 1.24% 5.66% 4.42%
2008 0.93% 5.43% 4.49%
2009 -0.65% 3.63% 4.28%
2010 -1.13% 2.84% 3.97%
2011 -1.15% 2.39% 3.54%
2012 -0.71% 2.30% 3.01%
2013 -0.13% 2.39% 2.53%
2014 1.28% 3.65% 2.37%
2015 1.78% 4.01% 2.23%
2016 1.99% 4.13% 2.14%
2017 2.04% 4.22% 2.18%
2018 2.05% 4.39% 2.34%
ChartData Download data

The data below can be saved or copied directly into Excel.

Copy the code below to embed this chart on your website.

Of course, one cannot make ironclad guarantees that GDP growth rates will always exceed interest rates. But GDP growth rates have far exceeded interest rates on public debt in most of the developed world over most of the past two decades even as public debt has risen substantially. The sustained downward pressure on interest rates over that time is a symptom of the chronic shortfall of aggregate demand we referenced previously (a shortfall sometimes called “secular stagnation”). There is very little evidence that secular stagnation is poised to reverse itself anytime soon.18

In our “Budget for Shared Prosperity” developed for the Solutions Initiative project, we aimed to put the debt-to-GDP ratio on a shallow downward path over the next 30 years during times of full employment, even though we think assumptions about how rapidly interest rates will rise in coming decades are unrealistic.19 We are not particularly concerned that interest rates will rise far above GDP growth and push up debt ratios, but we decided to take a conservative approach, hedging a bit against this risk by keeping deficits small.

If interest rates were to begin rising faster than GDP growth rates, this would hardly presage an economic disaster, but it would begin to impose some costs. More of the nation’s income would have to go to pay relatively well-off bondholders, and taxes would have to be raised to keep noninterest spending the same in the face of this higher interest burden. Preemptively raising taxes (particularly through steeply progressive taxes, like those our budget plan proposes) seems a desirable hedge against the admittedly unlikely event that interest rates begin marching sharply upward again. (Of course, politics would need to allow for these types of tax increases.) Given that both deficits and debt will rise steadily over time if they continue on their current path, it seems correct to us that proposals to significantly expand federal noninvestment spending should be matched with corresponding increases in revenue in order to stem the rise of deficit spending. Part of the reason we think it makes sense to slow down deficit spending is that the gap between GDP growth rates and interest rates—while still positive—has shrunk relative to earlier historical periods. This can be seen by looking at the most recent years in Figure C: The bars represent the gap between growth and interest rates, and it’s easy to see that these bars have shrunk in the most recent years.

To be concrete about deficit targets that would stabilize overall debt-to-GDP ratios, we can take the gap between the five-year average GDP growth rate and the five-year average interest rate as of 2018 (representing the five years from 2013 to 2018), shown in Figure C). This gap was roughly 2 percentage points. With GDP growth rates exceeding interest rates by this amount, the primary deficit (excluding interest) that would stabilize the debt-to-GDP ratio roughly at today’s level (of 80%) is 1.6% of GDP. Adding in interest costs—based on long-term interest rate projections—at this level of primary deficit would yield a sustainable deficit of well over 3% of GDP.

This may be too optimistic, but even the latest CBO long-term budget outlook projects that interest rates will essentially equal growth rates on average over the next 30 years, with interest rates exceeding growth rates by roughly 0.4% at the end of this period (CBO 2019). With this smaller gap between GDP growth rates and interest rates, a primary balance with an overall deficit (including debt service) of roughly 3% of GDP would stabilize the debt ratio at its 2019 level (and there’s no reason why stabilizing it at higher levels would be a particular problem).

Finally, any discussion of the U.S. fiscal position should note its utterly enormous potential fiscal capacity stemming from its low tax burden relative to all other rich nations. Among the 20 richest nations in the OECD, the U.S. has the second-lowest overall revenue as a share of GDP (including federal, state, and local taxes), trailing only Ireland—as seen in Figure D. The gap between what the U.S. collects today and what the globe’s highest-taxing countries collect is roughly 20 percentage points of GDP. These high-taxing countries tend to be rich and happy relative to the rest of the world. There certainly seems to be no obvious economic reason why the U.S. could not substantially increase revenue in the future.

Figure D

The U.S. remains an extraordinarily lightly taxed rich country: Total government revenue as a percent of GDP across the 20 richest OECD nations

Country Total government revenue as % of GDP
France 53.68729%
Finland 53.37975
Denmark 52.60747
Belgium 51.30097
Sweden 50.8558
Austria 48.41568
Italy 46.52186
Germany 44.97083
Luxembourg 44.40726
Iceland 43.81241
Netherlands 43.71818
Slovenia 43.18551
Non-U.S. average 42.35749
Canada 40.68857
UK 39.11596
Spain 37.91613
Australia 35.87092
Japan 35.78119
Switzerland 35.42836
U.S. 33.80197
Ireland 26.022
ChartData Download data

The data below can be saved or copied directly into Excel.

Source: Organisation for Economic Co-operation and Development, database, “Public Sector, Taxation and Regulation” tab, accessed July 2019

Copy the code below to embed this chart on your website.


Thinking intelligently about deficits and debt is never “just about arithmetic.” A dollar of generic spending cuts has the same effect on deficits as a dollar in generic tax increases, but the effects of these on human welfare might be radically different. Fiscal policy should be made with the primary aim of improving peoples’ lives. Sometimes this necessarily involves reducing the deficit and debt, and sometimes it doesn’t.

This report has laid out the primary economic challenges facing low- and middle-income households in the United States and sketched out how fiscal policy can help meet these challenges. Minimizing the future risks inherent in reducing the nation’s debt-to-GDP ratio is on the list of priorities that fiscal policymakers should aim to meet—but current state-of-the-art economic research indicates strongly that it should not shoulder its way to the top of this list.


1. The first report in the series describes the key principles for tax reform underlying our plan. See Blair 2019.

2. See Engelhardt and Gruber 2004 for trends in elderly poverty and their connection with Social Security.

3. The $5 trillion figure is obtained by summing the area between the lines representing actual GDP growth and potential GDP (as estimated in 2019) in Figure A, for the years 2008 to 2017, and then multiplying this amount by actual GDP in 2019. To estimate the job-years number, we take the difference each year between the actual rate of unemployment and the estimated “natural rate” used by CBO (2019) in their calculations of potential GDP; we then multiply this difference for each year by the labor force in that year; finally, we sum these totals for all years from 2008 to 2017.

4. The Bureau of Labor Statistics (2019) reports that the annual unemployment rate rose from 4.6% in 2006 and 2007 to 9.3% in 2009. The pre-recession unemployment rate was only regained in 2017 (when it averaged 4.4%).

5. This $10 trillion figure is obtained by summing the area between the line in Figure A representing potential GDP from 2008 to 2017 as estimated in 2019 and the line in Figure A representing potential GDP over that same period as estimated in 2008, multiplying by actual GDP in 2019, taking half of this total amount, and then adding it to the $5 trillion estimate of forgone output estimated earlier.

6. Anti-recessionary spending could also be monetized rather than financed with debt, but unless policymakers could commit to making this monetization permanent, this would have no real advantage over debt finance in spurring recovery.

7. Two obvious examples of automatic stabilizers are unemployment insurance and progressive income taxes. Both of these ensure that households are given more resources when their incomes fall as recessions hit (either transferred directly to them in the case of unemployment insurance or conveyed indirectly as implicit tax cuts in the case of progressive taxation).

8. The logic of this is simple: Rich households’ and corporations’ current spending is not constrained by current income, as they save large amounts of current income. This means they will cut back spending less than cash-constrained households will when their incomes fall, and they will increase spending less when their incomes rise. Bivens and Fieldhouse (2012) present a range of estimated fiscal multipliers from the research literature, confirming the fact that multipliers are much smaller on tax changes for rich households and corporations than they are for lower-income households.

9. Importantly, however, government policy can strongly affect market incomes. Laws that set minimum wages or that make it easier for workers to organize unions can affect the distribution of wage income, for example. And laws that set the terms of intellectual property protection, or regulations that check monopolies, can also affect the profitability of corporations.

10. Indirect evidence of this is presented in Figure D, which shows how much less the U.S. collects in taxes than its rich peers. When the same comparison is made directly with spending data from the same source (OECD 2019), the U.S. again ranks near-to-last in international comparisons.

11. For more details, see Blair 2019.

12. Furman and Summers 2019 and Blanchard 2019 are particularly notable—and well-pedigreed—examples of this useful rethinking going on among economists.

13. Bivens (2018) argues that evidence of this can be seen in the decision of the newly elected House Democratic majority to impose PAYGO budgeting rules in their upcoming term. (PAYGO, or “pay as you go,” rules require that new expenditures or new tax cuts be offset—with budget cuts and/or increased tax revenues elsewhere—so that they do not increase the budget deficit.)

14. This is a positive, not a normative, description. One can argue that the Fed’s inflation target is suboptimally low (as is argued in Bivens 2017b), but it’s a simple fact that 2% is the target.

15. See Bivens 2016 for an overview of these trends toward premature fiscal austerity.

16. See Almond, Hoynes, and Schanzenbach 2011 and Hoynes, Page, and Stevens 2011 for this evidence.

17. See Bivens 2017c for evidence that public investment likely has very high rates of return.

18. See Bivens 2017a on secular stagnation and its causes.

19. As we note in Bivens 2019, and in the introduction to this report, we actually ended up with a slightly steeper downward debt trajectory than we had hoped for, as we were given limited opportunities to iterate our revenue scores to hit a more precise path. But given that our budget would go a long way to solving one of the reasons for chronically low interest rates (the rise of income inequality), and given that in our budget we had control over both taxes and spending and hence could construct them both to reduce debt without putting an undue drag on aggregate demand, we’re not particularly worried about this too-steep path of debt reduction.


Almond, Douglas, Hilary W. Hoynes, and Diane Whitmore Schanzenbach. 2011. “Inside the War on Poverty: The Impact of Food Stamps on Birth Outcomes.” Review of Economics and Statistics 93, no. 2: 387–403.

Bivens, Josh. 2016. Why Is Recovery Taking So Long—and Who’s to Blame? Economic Policy Institute, August 2016.

Bivens, Josh. 2017a. Inequality Is Slowing U.S. Economic Growth: Faster Wage Growth for Low- and Middle-Wage Workers Is the Solution. Economic Policy Institute, December 2017.

Bivens, Josh. 2017b. Is 2 Percent Too Low? Rethinking the Fed’s Arbitrary Inflation Target to Avoid Another Great Recession. Economic Policy Institute, June 2017.

Bivens, Josh. 2017c. The Potential Macroeconomic Benefits of Infrastructure Investment. Economic Policy Institute, July 2017.

Bivens, Josh. 2018. “The Bad Economics of PAYGO Swamp Any Strategic Gain from Adopting It.” Working Economics Blog (Economic Policy Institute), December 18, 2018.

Bivens, Josh. 2019. “Presenting EPI’s ‘Budget for Shared Prosperity.’Working Economics Blog (Economic Policy Institute), June 11, 2019.

Bivens, Josh, and Andrew Fieldhouse. 2012. A Fiscal Obstacle Course, Not a Cliff: Economic Impacts of Expiring Tax Cuts and Impending Spending Cuts, and Policy Recommendations. Economic Policy Institute Issue Brief no. 138, September 2012.

Blair, Hunter. 2019. EPI’s Model Federal Budget and Tax Plan: How We Can Raise the Revenue Needed to Provide University Health Care, Strengthen Safety Nets, and Shore Up Public Investment. Economic Policy Institute, July 18, 2019.

Blanchard, Olivier. 2019. “Public Debt and Low Interest Rates.” American Economic Association Presidential Lecture, January 2019.

Bureau of Economic Analysis (BEA). 2019. “Table 1.1.6. Real Gross Domestic Product, Chained Dollars.” National Income and Product Accounts. Accessed June 2019.

Bureau of Labor Statistics (BLS). 2019. Civilian Unemployment Rate [UNRATE] (data series). Retrieved from FRED, Federal Reserve Bank of St. Louis, July 23, 2019.

Congressional Budget Office (CBO). 2008. The Budget and Economic Outlook: Fiscal Years 2008 to 2018. January 2008.

Congressional Budget Office (CBO). 2018. The Distribution of Household Income, 2015. November 2018.

Congressional Budget Office (CBO). 2019. The Budget and Economic Outlook: 2019 to 2029. January 2019.

Engelhardt, Gary, and Jonathan Gruber. 2004. “Social Security and the Evolution of Elderly Poverty.” National Bureau of Economic Research Working Paper no. 10466, May 2004.

Furman, Jason, and Lawrence Summers. 2019. “Who’s Afraid of Budget Deficits? How Washington Should End Its Debt Obsession.” Foreign Affairs, March/April 2019.

Hoynes, Hilary, Marianne Page, and Ann Huff Stevens. 2011. “Can Targeted Transfers Improve Birth Outcomes? Evidence from the Introduction of the WIC Program.” Journal of Public Economics 95, no. 7: 813–827.

Klein, Ezra. 2019. “Why Should We Care About Deficits?The Ezra Klein Show (Vox podcast), February 21, 2019.

Kogan, Richard, Chad Stone, Bryann Dasilva, and Jan Rejeski. 2015. Difference Between Economic Growth Rates and Treasury Interest Rates Significantly Affects Long-Term Budget Outlook. Center on Budget and Policy Priorities, February 2015. Accompanying data file downloadable at

Organisation for Economic Co-operation and Development (OECD). 2019. database, “Public Sector, Taxation and Regulation” tab. Accessed July 2019.