Distributional National Accounts: Methods and Estimates for the United States Thomas Piketty (PSE) Emmanuel Saez (UC Berkeley) Gabriel Zucman (UC Berkeley) November 2016
There is a large disconnect today between the study of inequality and macro Macro: use national accounts, with no distributional information Inequality: use survey & tax data, inconsistent with macro totals This gap makes it hard to know growth is distributed and to analyze the causes of the rise in inequality How does growth of bottom 50%, middle 40%, top 10% adults compare to total growth? What part of rise in inequality owes to change in factor shares vs. changes in the concentration of labor and capital? How do taxes and gov spending affect the distribution of growth?
This paper: attempt at constructing Distributional National Accounts (DINA) We construct a micro database of income, wealth, taxes and transfers consistent with national accounts totals in the US: First income inequality series covering 100% of national income First growth statistics by quantile consistent with macro growth First assessment of total redistributive effects of gov. intervention Getting back to Kuznets original intent to study growth & inequality jointly, but with more and better data Ultimate goal: being able to better compare ineq. across countries
Our objective: Distribute national income
National income = labor income + capital income 100% The share of capital and labor in national income 90% Capital income 80% % of national income 70% 60% 50% 40% 30% 20% 10% 0% 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013 Labor income
Reconciling national labor income and labor income reported on tax returns 80% From taxable to total labor income 70% Tax evasion & other 60% Employer fringe benefits & payroll taxes 50% 40% 30% 20% 10% 0% 1916 1920 1924 1928 1932 1936 1940 1944 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 % of national income 2012 Non-filers Wages and self-employment income on tax returns Source: Appendix Table I-S.A8b.
A growing fraction of labor income is missed by tax data
Reconciling national capital income and capital income reported on tax returns 30% From taxable to total capital income % of national income 25% 20% 15% 10% Non-filers & other Retained earnings Corporate income tax Imputed rents + property tax Income paid to pensions & insurance 5% 0% 1916 1920 1924 1928 1932 1936 1940 1944 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012 Didivends, interest, rents & profits reported on tax returns Source: Appendix Table I-S.A8.
Most capital income is missed by tax data
Methodology: How we distribute national income
How we move from fiscal income to total national income 1. Start with public-use samples of tax returns (1962-2010) High quality, oversamples top 2. Supplement public-use files using additional IRS data Age and gender information since 1979 Labor split for couples & fringe benefits on W2 since 1999 3. Impute missing income using SCF and CPS Non-taxable capital income (pension funds, imputed rents): SCF Monetary transfers: imputed based on CPS distribution by family income deciles and basic demographics 4. Distributionally neutral assumptions for public goods
We consider three main concepts of income matching national income Factor national income Sum of all labor income and capital income Pre-tax national income Subtracts contributions for pensions and social insurance, adds corresponding benefits (pensions, disability, unemployment) Post-tax national income Subtracts all other taxes Adds back all other forms of government spending (individualized transfers + public goods) Same national income total for factor, pre-tax and post-tax gives the broadest view of the redistributive effects of the government
Final product: a new tool for the study of inequality & growth Annual micro-data set representative of US pop. (1962-present): Detailed income & wealth variables matching national accounts aggregates Demographic information: age, gender, marital status, children Can be used to compute wide array of growth & inequality stats, and simulate tax and transfer reforms For 1913-1961: we rely on tabulated tax statistics produce series for specific income fractiles (top 10% and above)
The distribution of US national income and growth: Pre-tax vs. post-tax
DINA confirm the rise of income inequality, but post-tax inequality less 50% Top 10% national income share: pre-tax vs. post-tax 45% Pre-tax 40% 35% 30% 25% 1917 1922 1927 1932 1937 1942 1947 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007 2012 2017 % of national income Post-tax Source: Appendix Tables II-B1 and II-C1
Bottom 50% share has collapsed Bottom 50% national income share: pre-tax vs. post-tax 25% % of national income 20% 15% Post-tax 10% 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Pre-tax Source: Appendix Tables II-B1 and II-C1
Post-tax growth for bottom 50% has been anemic, and eaten up by health spending Average income in constant 2014 $ 25,000 20,000 15,000 10,000 5,000 Post-tax Real income of bottom 50%: pre-tax vs. post-tax Pre-tax Post-tax, excluding health benefits Medicare + medicaid 0 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Source: Appendix Tables II-B7, II-C7 and II-C3c.
The bottom 50% does not benefit on net from cash redistribution Average income in constant 2014 $ 25,000 20,000 15,000 10,000 5,000 Post-tax Real income of bottom 50%: pre-tax vs. post-tax Post-tax private Post-tax, excluding health benefits Medicare + medicaid Pre-tax 0 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Source: Appendix Tables II-B7, II-C7 and II-C3c.
For bottom 50% working-age adults, pre-tax income has collapsed since 1979 Real pre-tax income of bottom 50%, by age group Average income in constant 2014 $ 25,000 20,000 15,000 10,000 5,000 All age 45-65 years old 0 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 20-45 years old Real values are obtained by using the national income deflator and expressed in $2014. Income is divided equally among spouses.
At the bottom, only retirees pre-tax income is growing Real pre-tax income of bottom 50%, by age group 25,000 45-65 years old Average income in constant 2014 $ 20,000 15,000 10,000 5,000 >65 years old All age 0 1979 1983 1987 1991 1995 1999 2003 2007 2011 2015 20-45 years old Source: Appendix Tables II-B7 and II-B7b.
Even after transfers, 0 growth in working-age bottom 50% income since 79 Real post-tax income of bottom 50%, by age group 35,000 65+ years old Average income in constant 2014 $ 30,000 25,000 20,000 15,000 10,000 5,000 45-65 years old All 0 1979 1983 1987 1991 1995 1999 2003 2007 2011 2015 20-45 years old Source: Appendix Tables II-C7 and II-C7b.
The fall of the bottom 50% mirrors the rise of the top 1% 22% Pre-tax national income share: top 1% vs. bottom 50% % of national income 20% 18% 16% 14% 12% Top 1% 10% 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Bottom 50% Source: Appendix Table II-B1
1980: Top 1% income = 27 bottom 50 2014: Top 1% income = 81 bottom 50 Real average pre-tax income of bottom 50% and top 1% adults Top 1% real average pre-tax income (2014$) 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 200,000 0 1962 1966 1980: Top 1% = $428,000 1980: Bottom 50% = $16,000 1970 1974 1978 1982 1986 1990 2014:Top 1% = $1,305,000 1994 2014: Bottom 50% = $16,200 1998 2002 2006 2010 2014 52,500 45,000 37,500 30,000 22,500 15,000 7,500 0 Bottom 50% real average pre-tax income (2014$) Source: Appendix Tables II-B7 and II-B10
Changes in standards of living in the United States since 1946 Pre-tax income growth Post-tax income growth Income group 1980-2014 1946-1980 1980-2014 1946-1980 Full Population 61% 95% 61% 95% Bottom 50% 1% 102% 21% 130% Middle 40% 42% 105% 49% 98% Top 10% 121% 79% 113% 69% Top 1% 205% 47% 194% 58% Top 0.1% 321% 54% 299% 104% Top 0.01% 454% 75% 424% 201% Top 0.001% 636% 57% 617% 163%
Gender inequality and its effect on individual-level inequality
When assigning each spouse her own labor income, inequality has increased less 55% Top 10% pre-tax income share: equal-split vs. individuals 50% % of national income 45% 40% Pre-tax income per adult (individuals) 35% 30% 1917 1922 1927 1932 1937 1942 1947 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007 2012 2017 Pre-tax income per adult (equal split) Source: Appendix Table II-B9.
This is due to the decline in the inequality of labor income between genders 400% Average pre-tax labor income of men aged 20-64 / women aged 20-64 350% 300% 250% 200% 150% 100% 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Source: Appendix Table II-F1.
Part of the decline in gender inequality owes to rising female labor force particip. 50% Share of women in the employed population 40% 30% 20% 10% 0% 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Source: Appendix Table II-F1.
Men still make 85% of the top 1% of the labor income distribution 50% 45% 40% Share of women in the employed population, by fractile of labor income 35% 30% 25% All Top 10% 20% 15% 10% 5% Top 1% Top 0.1% 0% 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Source: Appendix Table II-F1.
At the median, no growth for working-age men over half a century 45,000 Median pre-tax labor income: working-age men vs. working-age women Real median pre-tax income ($2014) 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 Working-age men Working age adults 0 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Working-age women Source: Appendix Table II-B13.
Decomposing inequality: Labor vs. capital
Capital is making a comeback in the US 100% 90% 80% 70% 60% 50% 40% 30% The share of capital in pre-tax income 20% 10% Macro capital share in national income 0% 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013
Capital is making a comeback at the top The share of capital in pre-tax income 100% 90% Top 0.1% 80% 70% 60% 50% 40% Top 1% Top 10% 30% 20% All 10% Bottom 90% 0% 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013 Source: Appendix Table II-B2d.
Since the 1990s, top 1% rises because of capital income 14% Pre-tax capital income of top 1% adult income earners 12% 10% Interest and dividends paid to pension plans % of national income 8% 6% 4% 2% 0% Housing rents 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013 Income from equity Noncorporate profits Interest Source: Appendix Table II-B2b
Labor income concentration has stopped rising since the late 1990s 12% Pre-tax labor income of top 1% adult income earners 10% % of national income 8% 6% 4% Compensation of employees 2% 0% 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013 Labor component of mixed income Source: Appendix Table II-B2b.
The top became younger in the 1980s and 1990s, since the 2000s is growing older 58 Average age by pre-tax income group 56 Top 0.1% 54 52 50 Top 1% Top 10% 48 46 44 42 40 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015 Age Average age in the adult population Source: Appendix Table II-F2.
Decomposing inequality: the role of taxes and transfers
The macro rate of tax rose until the 1960s and has been constant since then 45% 40% Macroeconomic tax rate (Federal + State + local) % of national income 35% 30% 25% 20% 15% 10% 5% 0% 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013 Macroeconomic tax rate Source: Appendix Table II-G1.
Tax progressivity has declined since the 1960s 45% 40% Average tax rates by pre-tax income group % of pre-tax income 35% 30% 25% 20% 15% Top 1% All 10% 5% 0% 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013 Bottom 50% Source: Appendix Table II-G1.
Taxes have increased for the bottom 50% because of payroll taxes 30% Taxes paid by the bottom 50% % of bottom 50% pre-tax income 25% 20% 15% 10% 5% Capital taxes Payroll taxes Individual income taxes 0% 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013 Sales taxes Source: Appendix Table II-G2
Taxes have fallen at the top because of the decline of corporate and estate taxes 45% 40% Taxes paid by the top 1% 35% 30% 25% 20% 15% 10% 5% 0% 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 % of top 1% pre-tax income 2003 2008 2013 Individual income taxes Corporate taxes Estate taxes Sales + residential property + payroll taxes Source: Appendix Table II-G2
Individualized transfers have increased since the 1960s 25% Average transfers: individualized vs. collective consumption expenditure 20% % of national income 15% 10% 5% Collective consumption expenditure 0% 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 Individualized transfers (cash + in-kind) Source: Appendix Table G4
More transfers go to the middle class than to bottom 50% % of average national income 16% 14% 12% 10% 8% 6% 4% 2% 0% 1960 Bot 50% 1965 Average individualized transfer by post-tax income group (excluding Social Security) 1970 1975 1980 1985 1990 All Top 10% 1995 Middle 40% (P50-P90) 2000 2005 2010 2015 Source: Appendix Table II-G4.
More transfers go to the middle class than to bottom 50%, even incl. Social Security 25% Average individualized transfer by post-tax income group (including Social Security) % of average national income 20% 15% 10% 5% Middle 40% Top 10% All 0% 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 Bottom 50% Source: Appendix Table II-G4b.
Comparison with fiscal incomes
Bottom 90% has grown more than in tax data Bottom 90% income growth: Pre-tax income vs. fiscal income Average income in constant 2014 dollars 70,000 60,000 50,000 40,000 30,000 20,000 10,000 National income per adult Bottom 90% fiscal income per tax unit (Piketty-Saez) Bottom 90% pre-tax income per adult +2.0% +1.8% +2.1% +1.4% -0.1% 0 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 +0.8% Source: Appendix Table II-B3 and Piketty and Saez (2003, updated to 2014)
Without fringe benefits, 0 growth for bottom 90% since 1970s 40,000 Average pre-tax income of the bottom 90% Average income in constant 2014$ 35,000 30,000 25,000 20,000 15,000 10,000 5,000 Taxable labor income Capital income 0 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007 2012 Tax-exempt labor income Source: Appendix Table II-B2e
Conclusion
Combining tax, survey, and national accounts data The DINA agenda: Construct new series on the distribution pre- and post-tax income consistent with macro totals Hope will be adopted by govt agencies down the road Results for the United States: Collapse of working-age bottom 50% pre-tax income, 0 growth post transfer Boom of top-end inequality since late 1990s due to capital Spectacular gender gap at the top, not shrinking anymore Gov. has offset only small fraction of the in pre-tax inequality, due to in tax progressivity & limited transfers to bottom 50%
Supplementary Slides
Average income is growing less in tax and survey data than in the economy Average real income growth: national accounts vs. survey vs. fiscal data (1946 = 100) 300 National income per adult 260 CPS income per household (CPI) 220 180 Fiscal income per tax unit (CPI) 140 100 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Source: Appendix Table A0 and Census Bureau.
Three culprits: inflation, fewer marriages, rising non-taxable income 300 260 Average real income growth: national accounts vs. fiscal data (1946 = 100) Fiscal income per adult (national income deflator) National income per adult 220 Fiscal income per tax unit (national income deflator) 180 Fiscal income per tax unit (CPI) 140 100 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Source: Appendix Table A0.
Top 1% income share: pre-tax vs. post-tax Top 1% national income share: pre-tax vs. post-tax 20% % of national income 15% 10% Pre-tax 5% 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 2013 Post-tax Source: Appendix Tables II-B1 and II-C1
For bottom 50% elderly, all the growth comes from health transfers 35,000 Post-tax income of the bottom 50% of elderly Americans (65+) Average income in constant 2014 dollars 30,000 25,000 20,000 15,000 10,000 5,000 Post-tax income Post-tax income excluding health benefits 0 1979 1983 1987 1991 1995 1999 2003 2007 2011 2015 Medicare + Medicaid Source: Appendix Table II-C7c.
Still a lot of inequality between genders, especially for older workers 350% Average labor income of all men aged 20-64 / all women aged 20-64, by age group 300% 250% All 45 to 64 200% 150% 20 to 44 100% 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 Source: Appendix Table F1.
Capital is making a comeback in the US: robust to assuming fixed return on capital 50% The share of capital in pre-tax income, assuming constant 5% return on capital 45% Top 0.1% 40% 35% 30% 25% 20% 15% 10% Top 1% Top 10% All (macro capital share in national income) 5% 0% 1946 1951 1956 1961 1966 1971 1976 1981 1986 1991 1996 2001 2006 2011
Top-end inequality is high and growing at all ages 12% Top 0.1% pre-tax national income share by age group 10% % of pre-tax income 8% 6% 4% 65+ 45-65 All 2% 0% 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015 20-45 Source: Appendix Table II-B11b.
Transfers have softened blow to the middle-class during Great Recession 70,000 Real income of the middle 40%: the role of transfers Average income in constant 2014 dollars 60,000 50,000 40,000 30,000 20,000 10,000 Post-tax income Post-tax income excluding transfers 0 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 Transfers Source: Appendix Table II-C3b.
Pre-tax national income inequality has risen less than fiscal income inequality 50% Top 10% income share: comparison of estimates 45% 40% Pre-tax income per adult 35% 30% Fiscal income per tax unit (Piketty-Saez) 1917 1922 1927 1932 1937 1942 1947 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007 2012 Source: Appendix Table II-B1 and Piketty and Saez (2003, updated to 2014).
Part of the difference owes to income missing from tax returns in 1950s-1970s 50% Top 10% income share: fiscal income vs. pre-tax income 45% Pre-tax income per tax unit 40% 35% Missing income 30% Fiscal income per tax unit (Piketty-Saez) 1917 1922 1927 1932 1937 1942 1947 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007 2012 Source: Appendix Tables II-B9 and Piketty and Saez (2003, updated to 2014)
Rest of the difference owes to adults vs. tax units 55% Top 10% income share: tax units vs. equal-split adults 50% % of total income 45% 40% Pre-tax national income per tax unit 35% 30% 1917 1922 1927 1932 1937 1942 1947 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 2002 2007 2012 Pre-tax national income per adult (equal split) Source: Appendix TablesII- B1 and II-D1.
Bottom 50% income with different treatment of education spending Real post-tax income of bottom 50%: Different allocation of education spending Average income in constant 2014 $ 25,000 20,000 15,000 10,000 5,000 Eduction lump sum per child 0 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 Education proportional to disposable income (excluding health) Source: Appendix Tables II-C3d.