Report | Budget, Taxes, and Public Investment

Taking ‘Middle-Out Economics’ Seriously in this Fall’s Fiscal Debates

Briefing Paper #366

Download PDF

Press release

For the fourth year in a row, fiscal policy debates will loom large in late summer and early fall in American politics. This time around, the debates will center on the continuing resolution to fund the federal government in fiscal 2014 (starting Oct. 1) and the need to raise the statutory debt ceiling—the legislated cap on outstanding federal debt. It is easy to go straight into the wonky weeds on these issues, and they do deserve detailed discussion. But it is also vitally important to put these debates into a larger context—specifically, how these debates affect the majority of Americans.

Recently, a number of policymakers, including President Obama, have embraced the term “middle-out economics,” arguing that economic policy should aim first and foremost to boost middle-class living standards. This paper lays out what a real “middle-out” approach to fiscal policy issues would look like, by using actual data to diagnose the real problems affecting—and potential solutions for—middle-class living standards.

The clearest economic challenge for middle-class households in the near term is that the U.S. labor market remains deeply depressed and has never come close to a full recovery from the Great Recession. This near-term recovery can really only be achieved by a sharp shift in fiscal policy away from the march toward austerity that has characterized recent years.

But solving the immediate problem of a depressed labor market will not be sufficient by itself to reliably lift middle-class living standards. Over most of the past generation (at least since the late 1970s), middle-class living standards have been held back by the sharp rise in wage and income inequality that has resulted in vastly disproportionate shares of income growth accruing to the very richest households. Ensuring acceptable growth of middle-class living standards means braking, or even reversing, this rise in inequality. Fiscal policy is not the only tool that can do this, but it is one powerful measure that has not been used to its fullest potential.

In the near term, fiscal policy should address the most pressing economic challenge confronting the country by supporting job growth and finally achieving a full recovery from the Great Recession.

The U.S. economy is forfeiting $900 billion in annualized income simply because there remains deficient aggregate demand for goods and services. Filling in this demand shortfall requires reversing the current march toward austerity and using federal spending to boost, and not drag on, economic growth. Therefore, fiscal policymakers should:

  • Reverse the “sequester.” The importance of fiscal policy can be seen in the different outcomes under discretionary spending levels likely to be proposed by GOP and Democratic budget plans. The $91 billion higher discretionary spending in Democratic plans would turn off the automatic spending cuts called for by the sequester and lead to nearly an extra million jobs in the U.S. economy by the end of fiscal 2014.
  • At a minimum, follow historical precedent and provide support to the economy as it tries to fully recover. Of course, simply reversing the sequester would only be a small step toward using fiscal policy to boost growth. Inflation-adjusted government spending in the current recovery is roughly 15 percent lower than during average post–World War II recoveries. If public spending had simply matched historical averages during the current recovery, it would be roughly 15 percent higher today, leading to roughly 5 million jobs—most of which would be private-sector jobs.

In the longer term, we need fiscal policies that boost middle-class living standards by ensuring the middle class is able to claim a larger share of overall economic growth.

By 2007 (the last year before the Great Recession), decades of rising inequality had essentially imposed a 24 percent “inequality tax” on middle-income households, and this rise in inequality has been the primary impediment to middle-class income growth for a generation. Using fiscal policy to check rising inequality means fiscal policymakers should:

  • Preserve the few key economic institutions that are managing to boost middle-class incomes. Social Security, Medicare, and Medicaid have been extraordinarily important in raising middle-class incomes over the past generation, accounting for half of total income growth for the broad middle class between 1979 and 2007. “Middle-out economics” should focus on preserving key areas of strength for middle-class living standards, and these social insurance programs are one such area.
  • Raise top tax rates. While the highest-income households have claimed vastly disproportionate shares of total income growth, top income-tax rates have fallen substantially in recent decades, with the increase beginning January 1 of this year only slightly pushing back against this trend. If federal tax rates on the top 1 percent had remained at 1979 levels, this would have led (all else equal) to more than $150 billion in additional annual tax revenue in 2007. Besides raising revenue, higher tax rates would blunt the drive toward greater inequality. While most of the extraordinary rise in income inequality has been driven by pre–tax-and-transfer trends, recent research demonstrates clearly that this pretax income is itself heavily influenced by marginal tax rates on the highest-income households. Put simply, lowering top marginal tax rates increases the returns to economic activity that doesn’t generate new income, but that does shift income toward the top 1 percent.
  • Enact policies that ensure the tight labor markets needed to help provide across-the-board wage and salary increases. Growth in wages and salaries contributed only 10.1 percent to overall income growth of middle-class families between 1979 and 2007, even as hours worked by middle-class households rose significantly. Even more striking, cumulative growth in wage and salary income declined when excluding the years between 1995 and 2000, when tight labor markets (unemployment fell below 4 percent in 2000) accompanied by an increase in the minimum wage drove across-the-board wage growth. The lesson here is clear—all levers of macroeconomic policy, and fiscal policy most importantly, should be used to push economic growth to levels that will again foster tight labor markets and across-the-board wage growth.
  • Shift the primary strategy for addressing middle-class living standards away from tax cuts. Total federal tax rates on middle-class households have fallen in recent decades, slightly boosting middle-class living standards over that time. But the entire federal tax burden faced by the middle class is dwarfed by the income losses they have suffered due to rising inequality (the “inequality tax”) and the failure to generate a full recovery from the Great Recession. While progressive revenue sources should be the first ones tapped to deal with projected long-run budget deficits, it should be recognized that taxes have not been a drag on the growth of middle-class living standards in recent decades, and that Americans are willing to pay more to gain access to valuable government programs—particularly Social Security and Medicare. Even moving middle-class tax rates just halfway back to what prevailed in 1979 would yield more than $100 billion in additional tax revenue to fund public investments and social insurance programs.

What ‘middle-out economics’ should mean: Solving the jobs crisis in the near term and then reversing inequality

Putting middle-class living standards front-and-center in U.S. economic policy debates would be a clear improvement over focusing only on top-line measures like growth in overall income (which in recent decades has been boosted greatly by the enormous income growth at the very top). Over the longer run, the definition of a healthy economy should be one that delivers decent rates of income and living standards growth to the broad middle class.

In the near term, there actually is no real difference between the key priority of “middle-out economics” and that of maximizing overall growth: Both aim to finally ensure a full recovery from the Great Recession. This failure to achieve a full recovery is by far the most costly ongoing economic failure, and one that benefits nobody.

In the longer run, the most plausible path to boosting middle-class living standards is to ensure that middle-class income growth matches or exceeds growth in overall income. In practice, this simply means braking (and maybe even reversing) the extraordinary rise in wage and income inequality of recent decades.

This paper will mostly focus on the role of fiscal policy in reaching these goals, since fiscal policy debates dominate most of the attention in Washington, particularly in recent years. However, the problem of rising inequality should be addressed by all branches of economic policymaking—macroeconomic, tax, regulatory, labor, and international economic policy, just to name a few.

Near-term fiscal policy goal: Achieving a full recovery from the Great Recession

The U.S. economy remains far from fully recovered from the Great Recession. The “output gap” between actual gross domestic product (GDP) and potential GDP—how much could have been produced had unemployment and capacity utilization not been depressed due to insufficient aggregate demand—stood at 5.9 percent of potential GDP, or roughly $900 billion, at the end of 2012. Similarly, as of the middle of 2013, the share of “prime-age workers”—those between 25 and 54 years old—who are employed (their employment-to-population ratio) remains 4.3 percentage points below its 2007 peak (not to mention 5.6 percentage points beneath the peak value reached in the late 1990s). Furthermore, it was no higher in June 2013 than in June 2009, the official beginning of the recovery from the Great Recession.

Figure A shows the ratio of actual to potential GDP (the “output gap” simply is the difference between this ratio and 1) and the employment-to-population ratio for prime-age workers. Both measures of economic slack have followed a similar trajectory, as they have fallen far below their prerecession peaks and even further below where they were at the beginning of the 2000s.1

Figure A

Ratio of actual to potential GDP, and the employment-to-population ratio for prime-age workers, 2000–2012

Date Employment-to-population ratio Actual-to-potential GDP ratio
 2000-QI 0.818 1.0252763
 2000-II 0.817 1.0358216
 2000-III 0.812 1.0273167
 2000-IV 0.813 1.0240994
 2001-QI 0.813 1.0116242
 2001-II 0.808 1.0093345
 2001-III 0.803 0.9979305
 2001-IV 0.798 0.9930638
 2002-QI 0.797 0.9935851
 2002-II 0.794 0.9911093
 2002-III 0.792 0.9886535
 2002-IV 0.79 0.9818268
 2003-QI 0.789 0.9790202
 2003-II 0.79 0.9808266
 2003-III 0.787 0.9906853
 2003-IV 0.787 0.9935785
 2004-QI 0.788 0.9943256
 2004-II 0.79 0.9950911
 2004-III 0.791 0.9969516
 2004-IV 0.79 0.9995841
 2005-QI 0.792 1.0042099
 2005-II 0.793 1.003003
 2005-III 0.795 1.0052454
 2005-IV 0.793 1.0046486
 2006-QI 0.797 1.0114622
 2006-II 0.797 1.0095835
 2006-III 0.798 1.0037212
 2006-IV 0.801 1.0045079
 2007-QI 0.802 0.999825
 2007-II 0.8 1.002702
 2007-III 0.798 1.0039441
 2007-IV 0.797 1.0022191
 2008-QI 0.799 0.9920376
2008-II 0.795 0.9896675
2008-III 0.789 0.9751258
2008-IV 0.78 0.9477344
2009-QI 0.766 0.9305189
2009-II 0.76 0.9257817
2009-III 0.755 0.9253653
2009-IV 0.75 0.9309783
2010-QI 0.751 0.933023
2010-II 0.753 0.9349312
2010-III 0.75 0.9376759
2010-IV 0.749 0.9398876
2011-QI 0.752 0.9364209
2011-II 0.751 0.9383437
2011-III 0.75 0.9374095
2011-IV 0.752 0.9428356
2012-QI 0.757 0.9433546
2012-II 0.757 0.9422254
2012-III 0.757 0.9453382
2012-IV 0.758 0.9408653
ChartData Download data

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

Note: Prime-age workers are workers ages 25–54.

Source: Authors' analysis of Bureau of Economic Analysis National Income and Product Accounts (Table 1.1.6), Congressional Budget Office (2012), and Current Population Survey public data series

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

The roots of this slow recovery are far from mysterious: The large negative shock to aggregate demand from the bursting housing bubble (starting in 2007) has never been fully neutralized by countervailing policy measures to boost demand. To engineer the demand boost needed for a full recovery, there are two main levers that macroeconomic policymakers have at their disposal. One of these levers—monetary policy—continues to be supportive of recovery, with short-term policy interest rates near zero and the Federal Reserve even attempting to directly push down long-term rates by purchasing Treasury bonds and mortgage-backed securities. However, because economic weakness persists even in the face of this monetary support, and because short-term rates cannot be pushed below zero (a state of the world often referred to as a “liquidity trap”), policymakers should lean more heavily on the other main tool to boost recovery—expansionary fiscal policy.

What this means in practice is simply that the federal government should be injecting purchasing power into the economy, either through tax cuts or increased spending. Overwhelming empirical evidence demonstrates that spending increases are a more efficient stimulus, boosting economic activity more per dollar than tax cuts. For maximum efficiency, increased spending should be deficit-financed. If, however, politics demands that this increased spending be “paid-for” (offset by cuts elsewhere or increased revenues), the economic evidence is clear that financing it with very progressive revenue increases provides the smallest countervailing drag on growth.

During the worst parts of the Great Recession, when job losses reached a staggering 750,000 per month, federal spending was indeed boosted significantly with the passage of the American Recovery and Reinvestment Act (ARRA) in early 2009. One can see this clearly when measuring growth in real federal government expenditures across postwar recessions, as is shown in Figure B. The effect of ARRA shows up in the spike in spending in the fifth quarter following the beginning of the Great Recession.

Figure B

Real federal government expenditures in recessions since 1948

Average of pre-1980s recessions Early 1980s recession 1990 recession 2001 recession 2007 recession (Great Recession)
-4 0.963140021 0.9528188 0.97044003 0.964457854 0.95391793
-3 0.967434779 0.9504226 0.97713043 0.981044283 0.97764432
-2 0.981021883 0.96332 0.99306559 0.980275557 0.98991995
-1 0.986567148 0.9794026 1.00056244 0.986853195 0.99430967
0 1 1 1 1 1
1 1.014804396 1.0135289 1.01055219 1.003944169 1.01743743
2 1.034703311 1.0445011 0.97448043 1.016235026 1.06263883
3 1.038777538 1.0524066 1.01485394 1.016411474 1.04800646
4 1.04026367 1.0598874 1.03319154 1.039515107 1.04086574
5 1.077142838 1.0611965 1.04867324 1.054927847 1.07108068
6 1.063468315 1.0787196 1.07983784 1.055398556 1.16936206
7 1.048189531 1.0891731 1.08865213 1.070381663 1.17200544
8 1.066861121 1.0938261 1.0981944 1.081121333 1.17153832
9 1.092854086 1.0990875 1.1060658 1.121185825 1.19957899
10 1.108703145 1.1212571 1.10075988 1.11483512 1.20792639
ChartData Download data

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

Note: The average of pre-1980s recessions is the average of government expenditures for the six recessions and subsequent recoveries between 1948 and 1980. The early 1980s recession begins in 1980Q1 and spans through 1982Q4 to cover the recession from 1980Q1 through 1980Q3 and the subsequent recession from 1981Q3 through 1982Q4. The start of the recovery begins in 1982Q4.

Source: Authors' analysis of Federal Reserve Board, Federal Reserve Economic Data public data series

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

ARRA broke the economy’s free fall, and even managed to boost growth rates to levels that reliably lowered unemployment through 2010. However, the fiscal boost provided by ARRA was both temporary and left the economy well short of full employment when it ran out in the beginning of 2011. The Obama administration made some admirable efforts to keep fiscal policy from whipsawing sharply negative in 2011 and 2012, even to the extent of delaying long-standing priorities such as rolling back the Bush-era tax cuts on high incomes in exchange for some measures of fiscal support (notably, the 2 percent temporary payroll tax cut in 2011 and 2012).

But even with these efforts, fiscal policy since the official end of the Great Recession (in June 2009) has been sharply contractionary when compared with historical averages, particularly when state and local expenditures are included. The most striking comparison is with the recovery following the steep recession of the early 1980s. The output gap at the trough of the early 1980s recession was actually larger than that at the trough of the Great Recession (and the unemployment rate peaked at nearly a full percentage point higher in the 1980s than the late 2009 peak), yet two years following the 1982 trough, 80 percent of the output gap had been erased. In contrast, four years following the trough of the Great Recession, less than 20 percent of the output gap has been erased. Even more striking, the scope for monetary policy boosting recovery in the wake of the early 1980s recession was much larger—the federal funds rate was dropped by almost 10 percentage points following the onset of recession.

Given the similar size of output gaps at the trough of these recessions, and given as well that subsequent recovery was aided much more by monetary policy going forward from 1982, it seems axiomatic that an atypically large fiscal expansion was needed after the end of the Great Recession to spur full recovery. Figure C shows that in fact fiscal policy was instead atypically contractionary in recent years. It extends the measures of real government spending (here federal, state, and local expenditures) past the official end of recessions and deep into the subsequent recoveries. Real government spending four years into the current recovery is essentially 15 percent lower than during average post–World War II recoveries. Had the historical average of public spending been replicated in the current recovery, roughly 90 percent of today’s output gap would be closed, and the U.S. economy would have roughly 5 million additional jobs.2

Figure C

Real government expenditures in recessions and subsequent recoveries since 1948

Quarter Average of pre-1980s recessions Early 1980s recession 1990 recession 2001 recession 2007 recession
-9 0.903841
-8 0.907022
-7 0.91597
-6 0.919612 0.896781
-5 0.9336863 0.933643 0.908194
-4 0.96147084 0.943037 0.938287
-3 0.9660575 0.949903 0.973082 0.935461
-2 0.98664068 0.956289 1.011474 0.984862 0.929131
-1 0.99031804 0.973453 1.025372 0.991775 0.945731
0 1 1 1 1 1
1 1.03739367 1.003406 1.031986 1.016599 1.010355
2 1.02587153 1.016527 1.047023 1.027015 1.007506
3 1.01607616 1.020481 1.065995 1.033033 1.019912
4 1.03130283 1.01395 1.088692 1.045771 1.024077
5 1.04893481 1.022835 1.101646 1.059014 1.026123
6 1.06428821 1.037549 1.112303 1.077037 1.025589
7 1.07942298 1.049091 1.117083 1.076552 1.022811
8 1.10653676 1.066847 1.114949 1.08074 1.030403
9 1.10616078 1.075917 1.115387 1.099058 1.014113
10 1.133538 1.087306 1.122778 1.099532 1.011238
11 1.14401922 1.103156 1.1257 1.107016 1.014988
12 1.15512171 1.114143 1.118821 1.113766 1.022807
13 1.16543079 1.122944 1.125041 1.133755 1.01599
14 1.19325886 1.141506 1.135722 1.14007 1.016899
15 1.14414684 1.157107 1.149697 1.148254 1.008599
16 1.1566797 1.157997 1.159181 1.157468 1.015667
ChartData Download data

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

Note: The average of all pre-1980s recessions is the average of government expenditures for the six recessions and subsequent recoveries between 1948 and 1980. The early 1980s recession begins in 1980Q1 and spans through 1982Q4 to cover the recession from 1980Q1 through 1980Q3 and the subsequent recession from 1981Q3 through 1982Q4. The start of the recovery begins in 1982Q4.

Source: Authors' analysis of Bureau of Economic Analysis National Income and Product Accounts (Table 3.1)

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

This is an important lesson. Calls to address the jobs crisis with a fiscal boost commensurate to the scale of the problem are often greeted by implicit claims that this would constitute a wild and historically unprecedented degree of public spending. It is not so—we’ve had this amount of fiscal support for recoveries before, in the not-so-recent past, as Figure C shows. There is nothing either economically or historically “unrealistic” about the prospects of ending the jobs crisis by ending austerity.

At the very least, fiscal policymakers should repeal budget “sequestration”—automatic spending cuts negotiated in 2011 that nearly everybody now realizes are unnecessarily dragging on growth. Previous research has shown that repealing the $91 billion in scheduled cuts necessitated by sequestration for 2014 would generate between 900,000 and 995,000 jobs (CBO 2013; Bivens and Thiess 2013).

Fiscal policy goal for the long run: Braking the rise of inequality

A full recovery from the Great Recession would be a welcome first step in restoring middle-class health, but would be far from sufficient. The three decades before the Great Recession saw anemic growth in middle-class living standards compared with both historic trends and the potential for increases given overall growth. This potential was blocked because a vastly disproportionate share of total income gains went to the highest-income households. Accordingly, a truly “middle-out” approach to fiscal policy would try to steer a larger share of the benefits of overall economic growth in the long term to the broad middle class.

Just how important was inequality in damaging the broad middle class?

Figure D shows average annual income growth using two datasets: the Piketty and Saez (2012 updated) data set of cash, market-based incomes of tax units in two periods, 1947–1979 and 1979–2007, and Congressional Budget Office (CBO) data on household comprehensive incomes for the latter period.

Figure D

Average annual income growth by percentile, time period, and data set

Piketty and Saez, 1947–1979 Piketty and Saez, 1979–2005/07 CBO, 1979–2005/07
Average 2.1 1.1 1.8
0–90 2.1 0.2 1.1
91–95 2.6 1.1 1.8
96–99 2.1 1.7 1.7
99–99.5* 1.6 2.5 2.6
99.6–99.9* 1.3 3.4 3.5
99.9–99.99* 1.3 5.1 5.2
Top .01* 2.2 6.2 6.0
ChartData Download data

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

* Data go from 1979 to 2005.

Note: Piketty and Saez data are for tax units; the Congressional Budget Office (CBO) data are for households.

Source: Authors' analysis of data from Piketty and Saez (2012) and Congressional Budget Office (CBO 2012)

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

The data indicate strongly that the rise in inequality is a feature unique to this past generation. Income growth rates are much more uniform in the earlier period (when they range from 1.3 percent to 2.6 percent) than in the later period (when they range from 0.2 percent to 6.2 percent).

To show how important the rise in inequality has been for living standards growth of the broad middle class, we compare growth in average incomes (growth buoyed by rapid growth at the top) with growth in actual middle-class incomes from 1979 to 2007. The wedge between average growth and growth at the middle is essentially the “inequality tax” on middle-class incomes.

Figure E shows that the inequality tax imposed by this wedge was $17,865 in 2007, which is equal to 24 percent of the average comprehensive incomes of households in the broad middle of the income distribution—those households between the 20th and 80th percentiles. Importantly, this rise in inequality was much more damaging to income growth for this population than the decline in overall income growth rates after 1979.

Figure E

Household income for the broad middle class, actual and projected assuming it grew at overall average rate, 1979–2007

 Year Actual Projected
1979 $60714 $60714
1980 $58732.64 $58855.12
1981 $58384.47 $58638.72
1982 $57315.84 $58452.08
1983 $56369.14 $58858.63
1984 $59640.16 $61965.16
1985 $59771.71 $62928.76
1986 $61559.08 $67679.99
1987 $60628.18 $65471.77
1988 $61439.69 $68299.36
1989 $62293.88 $68904.05
1990 $62520.53 $68244.5
1991 $61589.53 $66612.67
1992 $61979.35 $68436.4
1993 $62702.69 $68834.39
1994 $63091.57 $69660.42
1995 $65054.18 $72338.97
1996 $65769.78 $74795.34
1997 $66911.57 $77662.68
1998 $69113.65 $81250.76
1999 $70948.66 $84688.06
2000 $70731.34 $86063.02
2001 $70771.07 $81558.64
2002 $69163 $78177.02
2003 $69349.78 $79418.73
2004 $71651.95 $84002.48
2005 $72760.86 $88145.3
2006 $73381.6 $90812.04
2007 $75411.56 $93276.63
ChartData Download data

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

Note: Data are for comprehensive income, and cover between the 20th and 80th percentiles of the income distribution.

Source: Authors' analysis of Congressional Budget Office (2012)

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

Figure F further illustrates the disproportionately negative effect of rising inequality on the broad middle class by comparing it to how much the slowdown in overall income growth contributed to slower living standards growth at the middle. Economic policy priorities should reflect this.

Figure F

Average middle-class income growth, actual and under two hypothetical scenarios, 1979–2007

Actual middle-class income Middle-class incomes rise at (faster) pre-1979 average rate, but “inequality tax” still hits Middle-class incomes rise at overall average rate (i.e., no “inequality tax”)
1979 $59,763 $59,763 $59,763
1980 $57,812 $58,086 $57,933
1981 $57,470 $57,710 $57,720
1982 $56,418 $56,627 $57,536
1983 $55,486 $55,749 $57,937
1984 $58,706 $59,428 $60,994
1985 $58,835 $59,696 $61,943
1986 $60,595 $62,128 $66,620
1987 $59,678 $61,498 $64,446
1988 $60,477 $62,708 $67,229
1989 $61,318 $63,664 $67,825
1990 $61,541 $63,988 $67,175
1991 $60,625 $63,267 $65,569
1992 $61,008 $63,922 $67,364
1993 $61,720 $64,724 $67,756
1994 $62,103 $65,241 $68,569
1995 $64,035 $67,644 $71,206
1996 $64,739 $68,725 $73,624
1997 $65,863 $70,305 $76,446
1998 $68,031 $73,100 $79,978
1999 $69,837 $75,498 $83,361
2000 $69,623 $75,447 $84,715
2001 $69,662 $76,115 $80,281
2002 $68,079 $74,868 $76,952
2003 $68,263 $75,246 $78,174
2004 $70,529 $78,381 $82,686
2005 $71,621 $80,155 $86,764
2006 $72,232 $81,195 $89,389
2007 $74,230 $83,771 $91,815
ChartData Download data

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

Note: This figure pertains to the broad middle class, defined here as households in the 20–80th percentiles of the income distribution. Income data are the average across this group and deflated using the CPI-U-RS.

Source: Authors' analysis of Congressional Budget Office (2012), as described in text

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

In the figure, actual comprehensive income growth of households between the 20th and 80th percentiles between 1979 and 2007 is compared with what growth for this broad middle class would have been under two scenarios. In the first scenario, we essentially calculate what middle-class incomes would have been if overall income had risen at the rate that prevailed between 1947 and 1979 but the 24 percent “wedge” between middle-class and average growth from 1979 to 2007 remained in place (i.e., the trajectory under pre-1979 overall income growth rates but post-1979 distribution).3 In the second scenario, we simply allow middle-class income to rise at the overall average rate between 1979 and 2007, which can also be thought of as the post-1979 growth and pre-1979 distribution scenario (i.e., essentially holding income growth across percentiles uniform).

The pre-1979 average combined with post-1979 distribution does indeed boost middle-class incomes, from growth of 24.2 percent over the entire period to 40.2 percent growth. However, post-1979 (slower) average growth combined with more equitable pre-1979 income distribution boosts middle-class income growth from 24.2 percent over the period to 53.6 percent.

In short, while income growth of middle-class households clearly slowed between 1979 and 2007 due to both slower average growth and rising inequality, inequality has had nearly twice the effect in slowing income growth for this group. Lastly, nearly 60 percent of the gap between income growth of the middle class and overall average income growth between 1979 and 2007 can be accounted for solely by growth of the top 1 percent (not shown in Figure F).4

Why do the Piketty and Saez and CBO data show different income trends for the bottom 90 percent?

A question that often arises when looking at the income growth trends (as measured in Figure D, earlier) is, “Why do income growth rates for the bottom 90 percent diverge so markedly between datasets?” While the two datasets largely agree on growth rates for the highest-income households, the Piketty and Saez data represented in Figure D show average annual growth of just 0.2 percent for the bottom 90 percent between 1979 and 2005, while the CBO data show bottom-90-percent incomes rising 1.1 percent annually in that period.

The simplest answer to this question is simply that the Piketty and Saez and CBO data measure different things. The Piketty and Saez data measure cash, market-based incomes (for tax units), while the CBO data include these incomes as well as government transfers and noncash employee benefits, such as health premiums (for households). Table 1 provides some evidence on why these two income measures differ so significantly in the past generation by showing the sources of income growth for the broad middle class over the last 30 years, where, again, the “broad middle class” encompasses households in the 20th to 80th percentiles.

Table 1

Change in source of comprehensive income for households in the 20–80th percentile, selected years, 1979–2007 (2012 dollars)

Total income Wages and salaries Capital and business income Pensions* Cash transfers Employer-sponsored insurance Non-cash transfers**
1979 $60,714 $44,823 $5,772 $2,320 $4,423 $2,110 $1,263
1989 $62,294 $41,730 $6,359 $4,172 $5,317 $2,343 $2,374
1995 $65,054 $41,704 $5,147 $5,214 $6,069 $3,017 $3,903
2000 $70,731 $45,849 $5,590 $6,647 $5,980 $2,702 $3,963
2007 $75,412 $46,314 $5,246 $7,238 $6,913 $3,570 $6,132
Total change
19792007 $14,698 $1,491 -$526 $4,918 $2,490 $1,459 $4,868
1995–2000 $5,677 $4,145 $443 $1,433 -$89 -$315 $60
Average annual change
1979–2007 0.8% 0.1% -0.3% 4.1% 1.6% 1.9% 5.8%
Share of total change
1979–2007 100.0% 10.1% -3.6% 33.5% 16.9% 9.9% 33.1%

* Includes "other income" as defined by the Congressional Budget Office

** This is overwhelmingly Medicare and Medicaid, though it also includes other non-cash transfers such as SNAP (food stamps) and housing assistance.

Note: Data inflated using the CPI-U-RS

Source: Authors' analysis of Congressional Budget Office (2012)

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

Between 1979 and 2007, half of the income growth for these households was driven by government transfers (dominated by Social Security and Medicare). And a third of this income growth (33.5 percent) was driven by pension income received for past labor market service (and the contribution from this legacy of now-disappearing defined-benefit pensions faded rapidly throughout the period in question), while another 9.9 percent was accounted for by employer-sponsored insurance—a noncash market income not accounted for in the Piketty and Saez dataset. Cash wages and salaries, conversely, accounted for just 10.1 percent of overall income growth for these households. Overall, market-based incomes (wages and salaries, capital and business income, and employer-sponsored insurance) contributed just 16.5 percent to middle-class income growth between 1979 and 2007.

Some have seized on the fact that the Piketty and Saez data do not include income sources that constituted the majority of income growth for the bottom 90 percent to conclude that these data should be ignored or devalued in policy debates. This is wrong, for a couple of reasons.

First, movements in the cash, market-based incomes that Piketty and Saez do track overwhelmingly drive overall trends in inequality even in the comprehensive income dataset tracked by the CBO. This should hardly be a shock, as cash, market-based incomes account for roughly 80 percent of all incomes even in the CBO data. Further, for the highest-income households, the share of incomes accounted for by cash, market-based incomes is much higher than average, and given that most of the changes in the income distribution are driven by these very high-income households, it makes sense that they will be tracked well by the Piketty and Saez data.

Second, it is extraordinarily important to know just how poorly cash, market-based incomes have provided for rising living standards for the middle class. One way to view the story told by the Piketty and Saez and CBO data is that while the market-based economy has been performing terribly in delivering rising living standards to the middle class, government social insurance programs—Social Security, Medicare, and Medicaid—have been reliably delivering growth.

Concrete recommendations for how fiscal policy should become ‘middle-out’

The information regarding the failure of cash, market-based incomes to deliver rising living standards leads to straightforward recommendations for fiscal policymakers: Preserve (or even expand!) the key social insurance programs, think hard and smart about what “controlling government health care expenditures” means, make boosting wage growth for the broad middle class a priority, and stop leaning on tax rates as a lever in trying to boost middle-class incomes.

Preserve (or even expand) key social insurance programs

We noted before that transfers accounted for half of all income growth for the broad middle class between 1979 and 2007. This makes it clear that those advocating a “middle-out” approach to economic policy should reject any attempt to cut the income growth provided by these programs. These transfer payments—particularly Social Security and Medicare—are disproportionately directed to older households. Since 2007, the importance of these programs has only grown, as household wealth and retirement readiness was battered by the burst housing bubble. Further, the share of income growth accounted for by private pensions has fallen rapidly from 1979 to 2007, further indicating that the “401(k) revolution”—the replacement of traditional defined-benefit pension plans with defined-contribution plans—has greatly damaged retirement security for coming generations. Given this context, the vital importance of preserving Social Security, Medicare, and Medicaid (as well as the still nascent Affordable Care Act) seems clear.

Think hard and smart about controlling health care costs

Medicare, Medicaid, and employer-provided health insurance premiums make an extraordinarily large contribution to middle-class income growth; altogether they accounted for nearly 40 percent of all income growth for the middle class in the generation before 2007 (see Mishel et al. 2012). A number of analysts have sharply criticized income measures (like the Piketty and Saez dataset) that do not include these health expenditures as conveying too-pessimistic news about living standards growth of the middle class, and about how the U.S. economy is working for the middle class generally. However, although these large increases in health-related income flows account for an ever-larger share of costs that taxpayers and employers face, they are not clear evidence that the broader economy is working well for middle-income families. This is simply because the large increase in nominal dollars transferred to these families for the purchase of health care goods and services largely just holds these households harmless against the rapidly rising price of health care. And the failure to rein in price growth in the U.S. health care sector—a failure magnified by the much greater success of peer countries in controlling this price growth—hardly serves as compelling evidence that the U.S. economy is working well for low- and moderate-income households (or for employers or governments or even high-income households, for that matter).5

The influence of rising health care costs is especially important given that they drive essentially the entire long-run increase in projected budget deficits. Just between 2010 and 2013, for example, a projected slowdown in health care costs over the coming decade reduced projected budget deficits by a cumulative $200 billion (Linden 2013). Figure G shows the role of rising health care prices in long-run deficit projections, displaying total federal spending as a share of GDP under three scenarios: the relatively alarmist “alternative fiscal scenario” from the Congressional Budget Office, in which the degree to which per beneficiary health care costs outpace economy-wide growth (a gap often referred to as “excess cost growth”) is 1.6 percentage points; spending if health care costs instead just rose exactly in line with economy-wide growth (“no excess cost growth”); and spending if excess cost growth in the CBO alternative fiscal scenario were cut in half. The results are dramatic. By the end of the (admittedly deeply speculative) 75-year long-term budget window, CBO projects that if excess cost growth continues, total federal spending will rise from 19.8 percent of GDP in 2015 to roughly 29 percent of GDP in 2087, with health care spending accounting for more than 100 percent of this increase (not shown in the figure). With no excess cost growth after 2015, total federal spending as a share of GDP would rise by just over 1 percentage point, and with excess cost growth cut in half, the total rise in federal spending as a share of GDP would be 4.2 percentage points in the years between 2015 and 2087.

Figure G

Total federal spending under three scenarios of excess health care cost growth, 2015–2087

Year CBO alternative fiscal scenario (excess cost growth continues) No excess cost growth Excess cost growth cut in half
2015 19.8% 19.7% 19.7%
2016 19.7% 19.5% 19.6%
2017 19.2% 18.9% 19.0%
2018 18.9% 18.5% 18.7%
2019 18.9% 18.5% 18.7%
2020 18.9% 18.4% 18.6%
2021 18.9% 18.3% 18.6%
2022 19.1% 18.4% 18.7%
2023 19.9% 19.1% 19.5%
2024 20.4% 19.5% 19.9%
2025 21.2% 20.2% 20.7%
2026 21.7% 20.6% 21.1%
2027 22.4% 21.2% 21.8%
2028 22.7% 21.4% 22.0%
2029 22.8% 21.4% 22.0%
2030 22.9% 21.4% 22.1%
2031 23.2% 21.6% 22.3%
2032 23.3% 21.6% 22.4%
2033 23.5% 21.7% 22.5%
2034 23.7% 21.8% 22.6%
2035 23.7% 21.7% 22.6%
2036 23.9% 21.8% 22.7%
2037 23.9% 21.7% 22.7%
2038 24.2% 21.8% 22.9%
2039 24.2% 21.7% 22.8%
2040 24.3% 21.7% 22.8%
2041 24.2% 21.5% 22.7%
2042 24.4% 21.6% 22.8%
2043 24.5% 21.6% 22.9%
2044 24.6% 21.6% 22.9%
2045 24.7% 21.6% 22.9%
2046 24.8% 21.6% 23.0%
2047 25.0% 21.7% 23.1%
2048 25.0% 21.6% 23.0%
2049 25.1% 21.6% 23.1%
2050 25.3% 21.6% 23.2%
2051 25.4% 21.6% 23.2%
2052 25.5% 21.6% 23.3%
2053 25.6% 21.6% 23.3%
2054 25.6% 21.5% 23.2%
2055 25.8% 21.6% 23.4%
2056 26.0% 21.7% 23.5%
2057 26.0% 21.6% 23.4%
2058 26.1% 21.6% 23.5%
2059 26.3% 21.6% 23.6%
2060 26.4% 21.6% 23.6%
2061 26.5% 21.6% 23.6%
2062 26.6% 21.6% 23.7%
2063 26.7% 21.6% 23.7%
2064 26.8% 21.6% 23.7%
2065 26.9% 21.6% 23.8%
2066 27.0% 21.6% 23.8%
2067 27.2% 21.6% 23.9%
2068 27.1% 21.4% 23.7%
2069 27.3% 21.5% 23.9%
2070 27.4% 21.5% 23.9%
2071 27.5% 21.5% 23.9%
2072 27.6% 21.5% 23.9%
2073 27.6% 21.4% 23.9%
2074 27.7% 21.3% 23.9%
2075 27.9% 21.4% 24.0%
2076 27.9% 21.3% 23.9%
2077 28.0% 21.3% 24.0%
2078 28.1% 21.3% 24.0%
2079 28.2% 21.3% 24.0%
2080 28.3% 21.2% 24.0%
2081 28.4% 21.2% 24.0%
2082 28.4% 21.1% 24.0%
2083 28.6% 21.2% 24.1%
2084 28.6% 21.1% 24.0%
2085 28.7% 21.1% 24.0%
2086 28.6% 20.9% 23.9%
2087 28.7% 20.9% 23.9%
ChartData Download data

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

Note: Health care cost levels reflect trends in per enrollee health care costs.

Source: Authors' analysis of Congressional Budget Office (2012)

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

In short, the future trajectory of federal spending growth is entirely driven by trends in per enrollee health care costs. All of this makes bringing health care price growth closer in line with overall price growth job number one for those truly concerned with the long-run deficit outlook. Moreover, simply offloading the cost of health programs onto private households is not only useless in this regard, it is actively perverse from the perspective of “middle-out economics.” Since 1970, Medicare and Medicaid have done a better job of containing cost growth than private insurers, so a cost shift will not only change who pays for health care, it will also simply make it more expensive.

Make boosting middle-class wages a top priority for fiscal policymakers

As noted above, cash wages and salaries contributed just 10.1 percent, and overall market-based incomes just 16.5 percent, to middle-class income growth between 1979 and 2007. This performance is especially poor given that hours worked by middle-class households increased over this period.6 Further, the cumulative growth of wage and salary incomes in all years besides 1995 to 2000 was outright negative.

This period between 1995 and 2000 was characterized by two key economic policy changes: Unemployment was allowed to reach 30-year lows—to levels far below what policymakers had considered dangerously inflationary as recently as the mid-1990s—and the minimum wage was increased for just the second time since 1980.

The lesson of this first change—allowing unemployment to reach very low levels—is (as we noted earlier) directly relevant to today’s fiscal policy debate. The number one problem the U.S. economy faces is joblessness, and the only reliable tool for solving it in the near term is increased government spending, financed either by debt (preferable) or through progressive tax increases. Mishel and Shierholz (2013) have noted the extremely poor wage growth in recent years, and prospects for wages growing more rapidly in coming years are essentially nonexistent unless unemployment rates return quickly to pre–Great Recession levels.

Raise top tax rates to fund government and to keep the outsized income gains from going to the highest-income households

Even as the share of total income claimed by the highest-income households (those in the top 1 percent) more than doubled between 1979 and 2007, the federal tax rate paid by this group declined by nearly 7 percentage points. Had these rates remained constant from 1979 to 2007, there would have been an additional $150 billion in revenue in 2007—an amount nearly as large as the entire federal budget deficit ($163 billion) in that year.7 Further, these tax cuts for the highest-income households have been truly enormous if measured since 1960 or 1970 (as shown in Figure H).

Figure H

Federal individual income tax rates for subgroups of the income distribution, 1970–2012

Year Top marginal income tax rate Effective tax rate for incomes over $1 million Effective tax rate for the top 400 tax returns Average federal tax rate for the top 1 percent
1970 71.8%
1971 70.0%
1972 70.0%
1973 70.0%
1974 70.0%
1975 70.0%
1976 70.0%
1977 70.0%
1978 70.0%
1979 70.0% 35.1%
1980 70.0% 33.1%
1981 69.1% 30.4%
1982 50.0% 26.7%
1983 50.0% 26.7%
1984 50.0% 27.0%
1985 50.0% 26.1%
1986 50.0% 24.6%
1987 38.5% 30.3%
1988 28.0% 29%
1989 28.0% 28.3%
1990 28.0% 28.1%
1991 31.0% 29.1%
1992 31.0% 26.4% 30%
1993 39.6% 29.4% 33.5%
1994 39.6% 28.6% 34.8%
1995 39.6% 29.9% 35.3%
1996 39.6% 30.8% 27.8% 35.2%
1997 39.6% 28.8% 24.2% 34.1%
1998 39.6% 27.5% 22.0% 32.6%
1999 39.6% 27.9% 22.2% 32.8%
2000 39.6% 27.7% 22.3% 32.4%
2001 39.1% 28.3% 22.9% 32.1%
2002 38.6% 28.5% 22.9% 32%
2003 35.0% 24.8% 19.5% 30.4%
2004 35.0% 23.4% 18.2% 30.1%
2005 35.0% 23.0% 18.2% 30.4%
2006 35.0% 22.5% 17.2% 30%
2007 35.0%  22.1%  16.6%  28.3% 
2008 35.0% 18.1% 28.1%
2009 35.0% 19.9% 28.9%
2010 35.0%
2011 35.0%
2012 35.0%
ChartData Download data

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

Source: Authors' analysis of Tax Policy Center Historical Individual Income Tax Parameters (2012), Sullivan (2013), Congressional Budget Office (2012), and IRS (2009)

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

While the direct arithmetic influence of tax rates on inequality is modest since the lion’s share of the rise in inequality occurred in pretax measures, a growing body of research (see Piketty, Saez, and Stantcheva 2012 and Fieldhouse 2013, for example) indicates that tax rates can influence the pretax distribution of incomes through what has been called the “rent-seeking” channel.

Essentially, the argument is that well-placed economic actors (corporate executives, for example) exercise substantial market power over their pay rates. As long as these well-placed economic actors weigh the costs and benefits of exerting more influence to boost their own incomes,8 then anything that increases the benefits of exerting this influence will lead them to do so, leading to a rise in their pretax incomes. A falling top marginal tax rate does indeed increase the benefit to these actors of exerting this influence by boosting their take-home pay. In essence, a falling top marginal tax rate increases the incentive for powerful economic actors to boost the share of overall economic growth they claim. This relationship between top marginal tax rates and the share of income accruing to the top (for example, the top 1 percent) holds in both international evidence as well as time-series evidence in the United States (see a discussion of this in Bivens and Mishel 2013). Crucially, this effect does not seem to be driven by falling top marginal rates significantly increasing labor supply or savings rates; instead, it seems to reflect pure redistribution of economic gains.

Given this, the scope and utility for fiscal policymakers to brake or even reverse trends in rising income inequality by raising top marginal rates is much larger than previously thought, while the economic downsides, such as a reduced labor supply or savings rates, seem very small. Fiscal policymakers should heed the evidence on the non-responsiveness of economic growth to top marginal rates (Hungerford 2012) and concentrate on progressive revenue sources for long-run deficit reduction. Further, because progressive revenue increases have such small effects on short-run economic activity, any near-term fiscal support that politics demands be “paid for” in the long run should certainly be done so by raising taxes on the highest-income households.

“Tax relief” is not a pressing need for middle-class households

One of the great rhetorical maneuvers in recent decades has been the relabeling of tax cuts as tax “relief.” Politicians from both sides of the aisle often place great emphasis on ensuring that middle-class Americans secure “relief” from taxes. Given this rhetoric, one would imagine that federal taxes have eaten up a growing share of middle-class incomes and constitute a growing barrier to middle-class living standards growth.

This is clearly not the case: Average federal tax rates on middle-class households actually fell (and income tax rates—never high to start with—fell almost in half) between 1979 and 2007, as shown in Figure I. The figure shows that average income taxes fell for all groups, with the middle (40th–60th percentile) seeing the largest decline, of 5.3 percentage points. The payroll tax rate, on the other hand, hardly budged, except to increase slightly for households in the 20th–40th percentiles. Tax increases that went into effect at the beginning of 2013 do not materially affect the long-run trajectory of middle-class households’ tax rates.9

Figure I

Tax rates on the broad middle class, 1979 and 2007

Average income taxes Payroll taxes
20-40th percentile 1979 6.90% 7.60%
20-40th percentile 2007 1.70% 8.60%
40-60th percentile 1979 10.40% 8.50%
40-60th percentile 2007 5.10% 8.90%
60-80th percentile 1979 13.00% 8.50%
60-80th percentile 2007 8.20% 9.30%
ChartData Download data

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

Note: Average income taxes include corporate and excise taxes.

Source: Authors' analysis of Congressional Budget Office (2012)

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

While progressive revenue increases should be the first priority for those fiscal policymakers looking to reduce deficits (for all of the reasons cited above), these data on middle-class tax rates clearly signal that there is much room to raise taxes even on middle-class households to fund valuable public investments and social insurance. For example, just raising federal income tax rates on middle-class families halfway back to what they were in 1979 would raise more than $100 billion annually.

Politically, tax increases on families outside the highest income strata are often treated as near-impossible and hopelessly unpopular. While it is true that tax increases are not an easy lift politically, it should be pointed out that many other proposed solutions to projected long-run budget deficits—particularly cuts to Social Security, Medicare, and Medicaid—are also extremely unpopular. In fact, much polling suggests that Americans are willing to see taxes raised to cover the valuable social insurance these programs provide.


“Middle-out economics” is likely good messaging—Americans increasingly understand that something is getting in the way of aggregate economic growth translating into growth in their own living standards, and want to start holding policymakers accountable for this. Whether “middle-out economics” become a guide to substantive policy remains to be seen. This paper has tried to demonstrate what it might look like to take “middle-out economics” seriously for this fall’s fiscal policy debates.

As important as tax and budget policy is, however, we should remember that the growing gap between middle-class incomes and aggregate economic indicators is driven by a range of policy decisions across many arenas. Reconnecting middle-class incomes and overall growth will take more than fiscal policy. But fiscal policy clearly has a role to play, and this report demonstrates how a genuine commitment to “middle-out economics” would shape fiscal policy decisions.

About the authors

Josh Bivens joined the Economic Policy Institute in 2002 and is currently the director of research and policy. His primary areas of research include mac­roeconomics, social insurance, and globalization. He has authored or co-authored three books (including The State of Working America, 12th Edition) while working at EPI, edited another, and has written numerous research papers, including for academic journals. He appears often in media outlets to offer eco­nomic commentary and has testified several times before the U.S. Congress. He earned his Ph.D. from The New School for Social Research.

Hilary Wething is a senior research assistant at the Economic Policy Institute. She provides support to EPI’s economists and policy analysts for a variety of topics, including health insurance, labor markets, inequality, living standards, and macroeconomics. She has a B.S.B.A. in mathematics and economics from Creighton University.

Appendix: Rising prices for health care and how to adjust living standards data for inflation

Table 1, derived from the CBO data on comprehensive income measures, indicates the extraordinarily large role that health-related government transfers and employee benefits have played in generating income growth since 1979. Fully 40 percent of the rise in middle-class incomes can be attributed to increased spending on Medicare, Medicaid, and employer-provided health insurance.

However, earlier versions of the CBO income data deflated these health-related transfers and employee benefits with the same overall price deflator applied to other income sources. But because the inflation measure—the consumer price index for urban consumers, research series (CPI-U-RS)—only includes out-of-pocket medical costs (a small share of all medical spending), it does not adequately reflect overall increases in health care costs. Because of this, and because prices for health care goods and services have risen much faster than overall prices, it is likely that these earlier CBO data overstated living standards growth by using the CPI-U-RS to deflate even Medicare, Medicaid, and employer-sponsored insurance premiums.

When Mishel et al. (2012) corrected for this (by deflating income flows related to health care goods and services with a medical-care-specific price deflator), the inflation-adjusted value of both these health-care-related income flows as well as overall income trends fell considerably. For example, deflating Medicare, Medicaid, and employer-sponsored insurance premiums with a medical-care-specific deflator reduced overall income growth for middle-class families by nearly a third.

In its latest version of the data, the CBO explicitly mentions this criticism of how health care is deflated as one justification for changing the deflator it uses to the price index for personal consumption expenditures (the PCE deflator). However, this is an inappropriate fix.

While the adoption of the PCE deflator does lead to a more accurate weight of health care expenditures in total spending, it also changes other weights in the overall consumption basket. The most consequential changes are the much-reduced weight for housing and energy in the PCE deflator relative to the CPI-U-RS. One reason that these weights are inappropriate for measuring household living standards is that the PCE deflator is based on spending by private households and by nonprofit organizations. This helps lead to the much-reduced housing weight in the PCE deflator. McCully, Moyer, and Stewart (2007) reconciled differences in CPI-U-RS and PCE deflator measures of inflation and found that if the CPI-U-RS accounted for rising health care costs with the same weight as the PCE deflator, then overall inflation would have been 0.5 percentage points larger between 2002 and 2007 than what was actually measured. So, if one simply wanted to use information from the PCE deflator to correct the problem of underweighting health care, one would actually find a significant decline in inflation-adjusted living standards, much as Mishel et al. (2012) found.

However, this 0.5 percentage-point increase in inflation from adopting the (higher) PCE weights for health care is outweighed by the 0.8 percentage-point decline in measured inflation stemming from the (lower) PCE weights for housing and fuel. Further, methodological differences between the construction of the PCE deflator and the CPI-U-RS further reduce measured inflation by roughly 0.2 percentage points per year.

To sum up, while the new CBO dataset, based on the PCE deflator, does indeed solve the problem of underweighting health care expenditures by households, it introduces other weighting problems that counterbalance these, and makes a methodological change that further counterbalances the health care correction. The net effect of this deflator switch is to actually raise measured income growth over the time period, even as the relative weight of (high inflation) health care expenditures is increased. For our purposes, the CBO cure for the underweighting of health expenditures in its old deflator (the CPI-U-RS) is actually a bit worse than the original problem. For this report, we have continued to use the CPI-U-RS as our preferred deflator.


1.  The figure does not extend through 2013 because when this report was published, the Congressional Budget Office had not yet updated its data on potential GDP.

2. See Bivens and Shierholz (2013) for background in converting these differences in public spending levels into employment.

3. While the comprehensive income measure does not go back beyond 1979, we know that overall personal income growth per capita (measured from the U.S. Bureau of Economic Analysis’s National Income and Product Accounts, or NIPA, tables) was slower between 1979 and 2007 compared with the 1947 to 1979 period—1.7 percent in the later period compared with 2.2 percent in the earlier period. We also know that government transfer payments and nonwage market incomes did not grow faster in the later period overall, so unless these income sources have become much more directed toward the middle three-fifths and less directed toward the top reaches of the income distribution, it is unlikely that these substantially offset the much slower growth rate of money income in the later period. Given this, it seems possible to get a sense of what the impact of rising inequality between 1979 and 2007 has been on middle-class income growth, and to compare it with the likely impact of the slowdown in overall growth relative to the 1947–1979 period.

4.  Of course, this exercise implicitly presupposes that one can assume that redistribution away from the top could have been (or could be) accomplished without damaging overall economic growth. Is this a safe assumption? We think the data bear it out. Besides the evidence assembled above indicating that the growth of these incomes is largely due to rents, a number of recent studies have looked directly at the issue of shifting top shares on overall economic growth. Piketty, Saez, and Stantcheva (2013) and Andrews, Jencks, and Leigh (2011) use international evidence to see if there is stark evidence that top shares affect overall growth.

5. See the appendix for a discussion of how health care inflation should shape our interpretation of living standards growth for middle-income families.

6.  Data on the role of increasing work hours in middle-class income growth come from Mishel et al. (2012) and concern households in the middle fifth of the income distribution.

7. This estimate of additional revenue does not take into account the impact of higher rates on labor supply or pretax incomes of this group.

8. See the balance of increased CEO pay versus the “outrage constraint” in the work of Bebchuk and Fried (2004), for example.

9. The temporary reduction in payroll tax rates expired at the beginning of 2013, but it was always assumed that this reduction would not become a permanent feature of the tax code.


Andrews, Dan, Christopher Jencks, and Andrew Leigh. 2011. “Do Rising Top Incomes Lift All Boats?” B.E. Journal of Economic Analysis & Policy, vol. 11, no. 1, 1–45.

Bebchuk, Lucian Arye, and Jesse M. Fried. 2004. Pay without Performance, The Unfulfilled Promise of Executive Compensation, Part II: Power and Pay. Cambridge, Mass.: Harvard University Press.

Bivens, Josh, and Lawrence Mishel. 2013. “The Pay of Corporate Executives and Financial Professionals as Evidence of Rents in Top 1 Percent Incomes.” Journal of Economic Perspectives, vol. 27, no. 3, 57–78.

Bivens, Josh, and Heidi Shierholz. 2013. “Four Years Into Recovery, Austerity’s Toll is At Least 3 Million Jobs.” Working Economics (Economic Policy Institute blog), July 3.

Bivens, Josh, and Rebecca Thiess. 2013. “Budget Debates Matter: The Difference in GOP and Democratic Levels of Discretionary Spending for 2014 Translates into Nearly 1 Million Jobs.” Working Economics (Economic Policy Institute blog), September 4.

Bureau of Economic Analysis (U.S. Department of Commerce) National Income and Product Accounts. Various years. National Income and Product Accounts Tables [data tables].

Congressional Budget Office (CBO). 2012. The Distribution of Household Income and Federal Taxes, 2008 and 2009. “Supplemental Data” [online data tables].

Congressional Budget Office (CBO). 2013. How Eliminating the Automatic Spending Reductions Specified by the Budget Control Act would Affect the U.S. Economy in 2014. A Letter to the Honorable Chris Van Hollen.

Current Population Survey public data series. Various years. Aggregate data from basic monthly CPS microdata are available from the Bureau of Labor Statistics through three primary channels: as Historical ‘A’ Tables released with the BLS Employment Situation Summary (, through the Labor Force Statistics Including the National Unemployment Rate database (, and through series reports (

Federal Reserve Economic Data (FRED). Various years. Database maintained by the Federal Reserve Bank of St. Louis.

Fieldhouse, Andrew. 2013. A Review of the Economic Research on the Effects of Raising Ordinary Income Tax Rates: Higher Revenue, Unchanged Growth, and Uncertain but Potentially Large Reductions in the Growth of Inequality. Economic Policy Institute and The Century Foundation, Issue Brief #353.

Hungerford, Thomas. 2012. Taxes and the Economy: An Economic Analysis of the Top Tax Rates Since 1945. Congressional Research Service, Report for Congress.

Internal Revenue Service (IRS). 2009. “Data on the 400 Individual Income Tax Returns Reporting the Largest Adjusted Gross Incomes.” SOI Tax Stats.

Linden, Michael. 2013. It’s Time to Hit the Reset Button on the Fiscal Debate. Center for American Progress.

McCully, Clinton P., Brian C. Moyer, and Kenneth J. Stewart. 2007. “A Reconciliation between the Consumer Price Index and the Personal Consumption Expenditures Price Index.” Bureau of Economic Analysis and Bureau of Labor Statistics.

Mishel, Lawrence, and Heidi Shierholz. 2013. A Decade of Flat Wages: The Key Barrier to Shared Prosperity and a Rising Middle Class. Economic Policy Institute, Briefing Paper #365.

Mishel, Lawrence, Josh Bivens, Elise Gould, and Heidi Shierholz. 2012. The State of Working America. An Economic Policy Institute Book. Ithaca, N.Y.: Cornell University Press.

Piketty, Thomas, and Emmanuel Saez. 2012. Tables and figures from “Income Inequality in the United States, 1913–1998.” Quarterly Journal of Economics, vol. 118, no. 1, 1–39, updated to 2010 in Excel format, March 2012.

Piketty, Thomas, Emmanual Saez, and Stefanie Stantcheva. 2013. “Optimal Taxation of Top Labor Incomes: A Tale of Three Elasticities.” American Economic Journal: Economic Policy. (forthcoming)

Sullivan, Martin A. 2013. “Economic Analysis: Is the Income Tax Really Progressive?” Tax Analysts, February 25.

Tax Policy Center. 2012. “Historical Individual Tax Parameters: U.S. Individual Income Tax: Personal Exemptions and Lowest and Highest Tax Bracket Tax Rates and Tax Base for Regular Tax, Tax Years, 1913–2013.” Tax Facts. Tax Policy Center. Urban Institute and Brookings Institution.


See related work on Budget, Taxes, and Public Investment

See more work by Hilary Wething and Josh Bivens