Toward national and regional distributions of personal income.

Fixler, Dennis J. ; Johnson, David S. ; Craig, Andrew 等

THE RELATIONSHIP between macroeconomic growth and income inequality has been the focus of several recent studies (see Organisation for Economic Co-operation and Development (OECD) 2011; Boushey and Hersh 2012; Boushey and Price 2014; Cingano 2014). These studies build on a long research legacy. Almost 80 years ago, Kuznets (1934) in his original report on the national accounts suggested that growth in gross domestic product (GDP) was not sufficient to evaluate the performance of the economy--examining the income distribution was important as well. This view was echoed anew in the recent Economic Report of the President (Council of Economic Advisors 2015) and is the theme of the Report by the Commission on the Measurement of Economic Performance and Social Progress (Stiglitz, Sen, and Fitoussi 2009).

Appropriate statistics to measure income distribution are thus vital. In conjunction with Kuznet's article, in the 1950s, the Office of Business Economics, the predecessor to the Bureau of Economic Analysis (BEA), began producing measures of the distribution of income in the United States. These first estimates were released in 1953 and began with estimates for 1947 (Office of Business Economics 1953). Similar to a method proposed in this article, these estimates used the current population survey (CPS) from the Census Bureau to account for distribution and allocated the measure of personal income to quintiles. These estimates were regularly released in the SURVEY OF CURRENT BUSINESS from 1950 to 1962 (Fitzwilliams 1964), and the last estimates were produced for 1971 (Radner and Hinrichs 1974). The estimates were discontinued because of resource constraints.

In this article, BEA is exploring how to best reissue estimates of the distribution of income. The remainder of this article discusses the following:

* The BEA personal income measure and how it can be reconciled with the Census Bureau money income measure to estimate appropriate personal income distributions. Gini coefficients are computed.

* An analysis and comparison of personal income and money income inequality measures.

* An analysis of regional income distribution for the four Census Bureau regions, using Theil index estimates to measure regional inequality.

* Conclusions drawn from this research effort and a look at future research initiatives.

Constructing Personal Income Data Sources and Methods

Since BEA's published personal income estimate is an aggregate without a distribution, we use the Census Bureau money income measure from the CPS as the basis for the distribution. To do so, it is necessary to bring the Census Bureau money income concept to the BEA national income and product accounts (NIPAs) personal income concept. Briefly, we first reconcile components by adding items that are in personal income but not in money income and removing items in money income that are not in personal income. Because the CPS suffers from underreporting, we adjust the values upward to achieve the national totals.

This paper follows up on the work of two previous papers: Fixler and Johnson (2014) and Fixler, Johnson, Craig, and Furlong (Forthcoming in 2017). These papers construct distributional estimates that are fully consistent with the NIPA personal income concept, closely following the work of McCully (2014), Furlong (2012), and the OECD Expert Group on measuring disparities in national accounts. Previous work by Fixler and Johnson (2014) focused on creating a NIPA-adjusted measure of Census Bureau money income, which kept the definition of money income--and then added other NIPA-specific income components, such as health spending and imputed interest. However, because the Census Bureau definition includes income components that are not included in personal income, such as retirement disbursements, the Fixler Johnson NIPA-adjusted income measure is still conceptually different from personal income, though not as different as the commonly used money income concept. For example, personal income includes, but the NIPA-adjusted money income concept excludes, the following: rental income from owner-occupied housing, employer fringe benefits (for example, retirement contributions and health insurance premiums), and imputed interest on insurance policy reserve funds. Fixler and others (2015) further mapped to NIPA personal income.

To estimate the distribution, we use the annual social and economic supplement of the CPS integrated with the consumer expenditure (CE) survey. The CPS collects data on income while the CE collects data on both income and expenditures. The CPS and the CE surveys are nationwide household surveys designed to represent the U.S. civilian noninstitutional population. There are differences between the surveys in the unit of measure and significant differences in frequency and design (see McCully 2014 for more information on the surveys). The sources used for the NIPA estimates of personal income and outlays are many and diverse; they include sample surveys conducted by the Census Bureau, administrative data from the Social Security Administration, and governmental benefits from other agencies.

To construct distributional estimates, personal income is first decomposed into its underlying detail level, consisting of more than 65 components ranging from wages and salaries and social security disbursements to less obvious components, such as imputed interest on life insurance and pension reserve funds. Each of these components is then matched to corresponding microdata to obtain distributional information. Both CPS and CE surveys are necessary because neither one contains all the information required to fully define personal income. For example, only the CE contains information on the rental value of owner-occupied housing, mortgage interest, and homeowner's insurance, all of which are needed to construct the rental income of owner-occupied housing.

Although both surveys are comprehensive, covering a wide range of income and consumption variables, it is not always possible to find an exact match in the microdata. In these instances, indicator variables are constructed from the microdata and are used to distribute the NIPA aggregates across each household. For example, neither survey contains a variable for employer contributions to pension plans. However, the CPS includes a variable indicating if the person participates in a pension plan or not. This variable is used in combination with a person's wage, which is assumed to be proportional to the employer contribution. Therefore, a person with a higher wage would receive a larger share of the NIPA aggregate than a person with a lower wage, given that they participated in a pension plan. Similarly, the imputed interest received from depository institutions is assumed to be proportional to a household's saving and checking accounts, two variables obtained from the CE.

Because information is used from two surveys, personal income could not directly be estimated for each household in each survey. To overcome this problem, a synthetic data set was constructed using a statistical matching procedure that links housing units in the CPS to units in the CE through the use of 20 common variables in both surveys. A CPS household and a CE household are assumed to be statistically identical if a distance function between the two is minimized for all possible housing combinations.

Another issue with using the survey data for the NIPA personal income distribution is that the underlying populations covered differ. The CPS and the CE survey cover only the civilian noninstitutional population, while NIPA personal income estimates cover the income (and expenditures) of those defined as U.S. residents in the national accounts, which includes nonprofit institutions serving households (NPISHs), the institutionalized population, federal civilian and military personnel stationed abroad, and persons whose usual place of residence is the United States and who are private employees working abroad for a period of less than 1 year. Excluded from the NIPA definition of residents are foreign nationals who work and reside in the United States for part of the year and foreign nationals studying in the United States. In addition, NIPA estimates include the income of those who died during the preceding year and who are not captured in the CPS. Excluding NPISHs' income and accounting for transfers between households and NPISHs gives a measure of household income, which will be used for the integration of the microestimates and macroestimates.

In order to align the NIPA population with that of the household surveys, we adjust the NIPA aggregates to align with the population covered in the household surveys. In most cases, this means removing certain population groups from the estimates, though in a couple of instances, it means adding population groups.

The next step is to construct the totals of each income component defined by the NIPA definition using the synthetic data and calculate scaling factors using the actual NIPA totals. We then apply these factors to the underlying microdata, hence ratio adjusting each income component for each household using the component specific scaling factors.

Specifically, consider household i, with income,

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where the scaling factors, [[alpha].sub.j], depend on the source, j, of income (for example, wages or dividends) and the term [y.sub.ji] is defined as household i's income from source j in the integrated data set. The [[alpha].sub.j] are given by the ratio of aggregate personal income to aggregate income in the surveys (either CPS or CE in the integrated data set); specifically, [alpha] = [Y.sub.j]/[X.sub.j], where [Y.sub.j] is the aggregate for source j in the personal income measure (in the NIPAs) and Xj is the aggregate for source j in the integrated data. This procedure increases each household's income by source and the new scaled household data is then used to obtain distribution measures.

To illustrate, consider only one source of income, such as wages. Then the scaled income for household i is equal to

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Additional sources of income would be similarly calculated and added to the total. This procedure generates a NIPA-based scaled income series for households in the CPS and thereby yields a NIPA-based income distribution.

One limitation of the above approach is that we assume that the levels of underreporting (and the difference between survey reports and NIPA measures) are the same for all households. Hence, every household receives the same scaling factor for each source of income. However, it is likely that different households have different levels of underreporting. Research has shown that there is large underreporting at the top of the distribution (see Sabelhaus and others 2015) and that there is underreporting of government transfer benefits at the bottom of the distribution (Meyer, Mok, and Sullivan 2015).

The main motivation for providing a distribution of personal income is to measure income inequality. A metric that is often used to capture the inequality in an income distribution is the Gini coefficient. A Gini coefficient is based on the Lorenz curve that illustrates how an actual distribution of income differs from one that has an equal distribution (chart 1). The Gini coefficient is the area of A divided by A+B. If the Gini coefficient equals 0, then the Lorenz curve aligns with the line of equality; if it equals 1, then all of the income accumulates to a single person.

The formula used to compute the Gini coefficients that is used in the tables below is given by

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where are observation pairs [x.sub.i], [x.sub.j], and [bar.x] is the mean of observations. (1)

Comparison of Personal Income and Money Income

Since the distributional information mostly comes from the CPS, which is usually used to construct the measure of money income, it is useful to compare them. Chart 2 illustrates the trend in the mean and median for personal income and money income for 2000-2012. The chart also illustrates the level of the differences between the two income concepts: personal income is larger than money income because it includes items such as employer contributions for retirement and health insurance and government transfers in-kind, such as Medicare, Medicaid, and food stamps. Because households differ in size and composition, income levels are adjusted to reflect the attending impact on the use of income. There are several ways of "equivalizing" income, and we use the square root approach that divides each household income by the square root of the number of members in the household; this approach is commonly used in studies of income inequality and poverty measurement. Personal income has a higher mean and median because it is a broader concept of income. In addition, the mean and median for personal income increase over the period; the mean grows 41 percent, and the median grows 39 percent. In contrast, for money income, the mean grows 26 percent, and the median grows 29 percent.

[ILLUSTRATION OMITTED]

It is also of interest to compare the Gini coefficients published by the Census Bureau, using money income from the CPS, with the Gini coefficients constructed from our household-based personal income estimates. Table 1 provides the Gini coefficients, and chart 3 illustrates the differences. The Gini coefficient grows 3 percent, indicating an increase in inequality over the period. Observe that the Gini coefficient for personal income is always below that of money income. It is not surprising that the levels of the personal income Gini coefficients are lower as personal income captures more of the impact of the safety net (essentially government transfer payments) on income. Note that the trends are similar. The table also includes disposable personal income, which will be discussed below, and shows that the Gini coefficients are even lower after personal taxes are deducted.

[ILLUSTRATION OMITTED]

Table 2, which is modeled on NIPA table 2.1, provides a distribution of personal income by component. The first two columns show the difference between the published numbers and what we call the household-based numbers. The difference between these two columns derives from the inclusion of values, such as the income of NPISHs, which are excluded from our household-based estimates. (See the earlier discussion about limiting the analysis to the CPS, which is household based.)

Table 2 is divided into three panels--one for each year. In addition to the published estimates, we provide our household-based estimates, the distribution of the components by quintiles and the quintile shares. As explained above, the household-based estimates remove from personal income the income from NPISHs and the institutionalized population. Observe that the household-based estimates are around 97 percent of the published value of personal income for the 3 years.

[ILLUSTRATION OMITTED]

To simplify the discussion of the table, the focus will be on three primary components of income that are frequently discussed in the context of examining the income distribution: labor income, capital income, and transfer payments. The first will be represented by wages and salaries, the second by personal income receipts on assets, and the third by personal current transfer receipts.

Looking at the estimates for 2000, the share of wages and salaries earned by the top quintile is nearly twice the next quintile and over 15 times the share earned by the first quintile. The pattern is the same for income receipts on assets, but the dominance of the top quintile is stronger, approximately three times the share of the next quintile. For personal current transfer receipts, the first three quintiles have the largest shares, but the distribution of transfer receipts is relatively flat. This result is likely due to the counterbalance between the lower quintiles receiving higher shares of Medicare and Medicaid and the higher quintiles receiving unemployment insurance, veterans benefits and refundable tax credits--the last is embedded in the "other" category.

For 2006, the pattern of wages and salaries is the same as it is for personal income receipts on assets. For personal current transfer receipts, there is a reduction in the share of the first quintile and a reduction in the share of the top quintile--Medicare and Medicaid take a higher share. Note again that the "other" category is significant. (2)

For 2012, the patterns described above remain, though there is some variation, there is no noteworthy departure from trend.

The estimates in table 2 are consistent with the trend for the Gini coefficients for personal income illustrated above. Observe that the shares of the top quintile are stable and that the magnitudes increase slightly. Though we don't provide the details in the table, it is clear from the chart of the Gini coefficients that we would expect the shares of the top quintile to have declined in 2009 and 2010 and that there was a subsequent increase in inequality. Note that the Gini coefficients for 2006 and 2012 are the same.

Regional Analysis

The focus of income distribution analysis is usually at the national level. However, it is also of interest to see how inequality is distributed across the country. Because the decomposability of the Gini index is not straightforward, we use the Theil index to measure regional inequality. The Theil index is based on information theory and was developed in Theil (1967), which also provided an alternative derivation for the Torn-qvist index formula. The Theil inequality measure is given by

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where [y.sub.i] is observation i (income of household), [f.sub.i] is the population share of observation i, and [mu] is mean of observations. (3)

One of the analytically useful features of the Theil index is that it can be decomposed into between-group and within-group effects:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where j refers to the subgroup, [g.sub.j] refers to the population share of subgroup j, and [GE.sub.j] refers to the inequality (Theil) in subgroup j, [[mu].sub.j] refers to the mean in subgroup j, and [mu] refers to the total population mean. Intuitively, the between-group component of inequality measures the level of inequality that would occur if everyone within each group j had income level [[mu].sub.j]; that is, everyone in the group had the mean level of income. We also provide a raw Theil index; this is the Theil index for a member of a subgroup in isolation. The Theil index measures of inequality are similar in magnitude to the Gini coefficients as indicated in table 3 and chart 4.

Chart 4 suggests that there is essentially a level difference between the inequality measures, while the trends are similar. (4) Using the decomposition above, table 4 shows how the national Theil index is decomposed within and between four Census Bureau regions of the United States.

At the national level, the Theil index increases 8.3 percent between 2000 and 2006, while only increasing 4.0 percent between 2006 and 2012; this is likely due to the impact of the recession. Note that the sum of the between-group and within-group components is equal to the national Theil index and that the sum of the regional Theil indexes is equal to the within-group component. The very small value of the between-group component of the Theil index indicates that geography plays almost no role in determining the level of inequality; it is almost entirely coming from the within-group region measure. Note that the South has the highest level of inequality for each year.

[ILLUSTRATION OMITTED]

We next consider each region separately and recompute a Theil index for each state in the region. In tables 5-8, for each state we compute a "raw Theil" index value, which is simply the inequality in the state not considering its membership in the region. To compute the contribution of each state to the region's index, the raw Theil index value is basically weighted by the state's share of population in the region to obtain the contribution to the within-group calculation. Again, the overall Theil index is the sum of the between-group and within-group component.

Table 5 provides the specifics for the Northeast region. The raw Theil index gives a measure of inequality in the state alone, and note that New York is near the top in the level of inequality in 2006 and 2012, but in 2000, the level is higher in Maine. Maine and Vermont start out with high raw Theil measures in 2000, but these steadily decline by 2012. Connecticut and Massachusetts on the other hand have higher values in 2006 and 2012 than they do in 2000. Once population share is considered, New York has the greatest contribution to the within-group measure. Again, the between-group measure is extremely small and indicates that geography plays almost no role; it is all about the inequality in each state.

The Midwest region has a more equal distribution of income, as indicated by the last line of table 6. For some states, the increase in the raw Theil is large. Notably, the raw Theil for North Dakota increases 46 percent between 2000 and 2012--likely because of the growth in mining and fracking. A large growth in North Dakota's personal income was also found in BEA's published state personal income estimates. It should be pointed out that the correlation between BEA's published state personal income estimates and the state personal income estimates computed in this paper is 0.99 for the 3 years examined.

Table 7 gives the state decomposition for the South region. The District of Columbia has high levels of inequality for each year; however, its contribution to the within-group measure is very small because of its low population share. Mississippi also has substantial increases in the raw Theil in 2006 and 2012 relative to 2000.

Finally, table 8 presents the estimates for the states in the West region. The increase in the raw Theil index for New Mexico is eye catching. Also note that the contribution of California to the within-group component is large because of their high population share.

Conclusions

This article has provided distributional estimates of personal income along with measures of inequality both nationally and regionally. It shows that the level of inequality has increased in recent years and that while the level of inequality has increased in the four regions, the rate of increase is not the same across the country.

Our derivation of a distribution of BEA's personal income from the Census Bureau's CPS is not without limitations. In addition to the mapping and scope issues listed above, the main limitation is using the same scaling factor for households to move from the Census Bureau money income concept to personal income. Clearly, the factor for the income receipts on assets is different in the upper quintile households than in the lower quintile households. One way to improve on the factors would be to use federal income tax data to inform the factors. Indeed, such a procedure was experimented with in Fixler and Johnson (2014). The use of tax data may improve other aspects of the mapping. There is a parallel research effort using such data to measure income inequality; see for example, Auten and Splinter (2016). Another limitation is the focus on nominal income. Future research will examine inflation-adjusted (real) personal income and explore the use of BEA's regional price parities to examine regional income inequality.

The distribution of personal income presented above is similar in spirit to the distribution of national income presented in Piketty, Saez, and Zucman (2016). National income is broader than personal income and conceptually includes such categories as corporate profits, taxes on production and imports less subsidies, contributions for government social insurance, business current transfer payments and the current surplus of government enterprises. To create a distribution of national income, they too use microlevel data--data from the CPS and income tax records. And to capture the additional categories listed, they must impute values to the corresponding values for households. Because our focus is on the distribution of household income and its ultimate relationship to personal consumption expenditures, a subject for future research, our use of personal income is appropriate.

Measuring the distribution of household income has received worldwide attention. BEA participates in an Organisation for Economic Co-operation and Development (OECD) Expert Group on Disparities in a National Accounts Framework. The goal of this group is to establish a methodology to construct distributional estimates of income, consumption, and saving consistent with national accounting concepts using microdata. The results of the group's efforts are summarized in Fesseau and Mattionetti (2013) and Zwi-jnenburg, Bournot, and Giovannelli (2016). Generally, the participating countries were able to provide estimates in accordance with the proposed methodology. However, several shortcomings were identified, including the following: (1) the lack of microdata on several national account-specific income components and (2) substantial data gaps between microaggregates and national account totals. Going forward, this expert group will continue to refine the methodology to improve the shortcomings mentioned above, with the aim of eventually establishing a regular publication of distributional results on a per country basis.

References

Auten, Gerald, and David Splinter. 2016. "Using Tax Data to Measure Long-Term Trends in U.S. Income In-equality." Unpublished paper.

Boushey, Heather, and Adam Hersh. 2012. The American Middle Class, Income Inequality, and the Strength of Our Economy: New Evidence in Economics." Center for American Progress Report.

Boushey, Heather. and Carter C. Price. 2014. "How are Economic Inequality and Growth Connected?" Wash i n gton Center for Equitable Growth.

Cingano, Federico. 2014. Trends in Income Inequality and Its Impact on Economic Growth. OECD Social, Employment, and Migration Working Paper No. 163. Paris: OECD, December.

Council of Economic Advisors. 2015. Economic Report of the President, 2015. Washington, DC: U.S. Government Printing Office.

Elbers, Chris, Peter Lanjouw, Johan A. Mistiaen, and Berk Ozler. 2005. "Re-Interpreting Sub-Group Inequality Decompositions." Policy Research Working Paper 3687. Washington, DC: World Bank, August.

Fesseau, Maryse, and Maria Liviana Mattionetti. 2013. "Distributional Measures Across Household Groups in a National Accounts Framework: Results From an Experimental Cross-Country Exercise on Household Income, Consumption, and Saving." OECD Statistics Working Papers No. 2013/04. Paris: Organisation for Economic Co-operation and Development Publishing.

Fitzwilliams, Jeannette M. 1964. "Size Distribution of Income in 1963." SURVEY OF CURRENT BUSINESS 44 (April 1964).

Fixler, Dennis F. and David S. Johnson. 2014. "Accounting for the Distribution of Income in the U.S. National Accounts." In Measuring Economic Stability and Progess, edited by Dale W. Jorgenson, J. Steven Landefeld, and Paul Schreyer. Chicago: University of Chicago Press, for the National Bureau of Economic Research.

Fixler, Dennis F., David S. Johnson, Andrew Craig, and Kevin Furlong. Forthcoming. "A Consistent Data Series to Evaluate Growth and Inequality." Forthcoming in 2017. Review of Income and Wealth.

Furlong, Kevin J. 2012. "Synthetic Data--Unconstrained Statistical Matching: Technical Appendix." Unpublished. Bureau of Economic Analysis.

Kuznets, Simon. 1934. National Income, 1929-1932. Senate Document No. 124, 73rd Congress, 2nd Session. Washington, DC: U.S. Government Printing Office.

McCully, Clinton P. 2014. "Integration of Micro and Macro Data on Consumer Income and Expenditures." In Measuring Economic Stability and Progess, edited by Dale W. Jorgenson, J. Steven Landefeld, and Paul Schreyer. Chicago: University of Chicago Press, for the National Bureau of Economic Research.

Meyer, Bruce D., Wallace K.C. Mok, and James X. Sullivan. 2015. "The Under-Reporting of Transfers in Household Surveys: Its Nature and Consequences." Unpublished Paper. University of Chicago, June.

Office of Business Economics. 1953. Income Distribution in the United States by Size, 1944-1950. Washington, DC: U.S. Government Printing Office.

Organisation for Economic Co-operation and Development (OECD). 2011. How's Life? Measuring Well-Being. Paris: OECD Publishing.

Piketty, Thomas, Emmanuel Saez, and Gabriel Zucman. 2016. "Distributional National Accounts: Methods and Estimates for the United States Since 1913." National Bureau of Economic Research Working Paper 22945.

Radner, Daniel B., and John C. Hinrichs. 1974. "Size Distribution of Income in 1964, 1970 and 1971." SURVEY OF CURRENT BUSINESS 54 (October).

Sabelhaus, John, David Johnson, Stephen Ash, David Swanson, Thesia I. Garner, John Greenlees, and Steve Henderson. 2015. "Is the Consumer Expenditure Survey Representative by Income?" In Improving the Measurment of Consumer Expenditures, edited by Christopher Carroll, Thomas F. Crossley, and John Sabelhaus. Chicago: University of Chicago Press, for the National Bureau of Economic Research.

Stiglitz, Joseph E., Amartya Sen, and Jean-Paul Fitoussi. 2009. Report by the Commission on the Measurement of Economic Performance and Social Progress. New York: United Nations Press.

Theil, Henri. 1967. Economics and Information Theory. Amsterdam: North-Holland Publishing Company.

Zwijnenburg, Jorrit, Sophie Bournot, and Federico Giovannelli. 2016. "Expert Group on Disparities in a National Accounts Framework: Results from the 2015 Exercise." OECD Statistics Working Papers, No. 2016/10. Paris: Organisation for Economic Co-operation and Development Publishing.

By Dennis J. Fixler, David S. Johnson, Andrew Craig, and Kevin J. Furlong

Dennis J. Fixler is Chief Economist of the Bureau of Economic Analysis (BEA). David S. Johnson is Research Professor and PSID Director at the University of Michigan. Andrew Craig is an economist in the Office of the Chief Economist, and Kevin J. Furlong is an economist in the Office of the Director.

(1.) See Elbers and others (2005) for formula

(2.) It should be noted that the ratio of the household-based estimates of Medicare and Medicaid to the published estimates are around 70 percent, much lower than the ratio of the other components because of the "recent death discrepancy"; while the NIPAs capture the medical expenses for the recently deceased, the survey-based CPS does not. It should also be noted that this is also a consequence of controlling to the published personal income estimates. Because government transfers are poorly captured in the CPS, they have large scale factors (low coverage ratios). By scaling the microdata, the monetary value of government transfers assigned to each household will be much higher than what the household actually received. For example, if household "i" in the CPS reported Medicaid of $20,000 and the coverage ratio is 50 percent (scale factor = 2), the adjusted value will be $40,000. As a result, this household will appear to have more personal income than a household that did not receive Medicaid, all else being equal.

(3.) See Elbers and others (2005) for formula.

(4.) Since the Theil index is more sensitive to changes at the top of the distribution, this could contribute to the larger increases between 2010 and 2012. It should be noted that similar increases in the Theil index occur with the Census Bureau data. Table 1. National Gini Coefficients for Money Income, Personal Income, and Disposable Personal Income, 2000-2012 2000 2001 2002 Money income 0.438 0.447 0.444 Personal income 0.424 0.422 0.422 Disposable personal income 0.396 0.398 0.401 2003 2004 2005 Money income 0.445 0.444 0.448 Personal income 0.423 0.430 0.427 Disposable personal income 0.404 0.412 0.411 2006 2007 2008 Money income 0.447 0.441 0.443 Personal income 0.436 0.427 0.433 Disposable personal income 0.421 0.411 0.419 2009 2010 2011 Money income 0.446 0.449 0.457 Personal income 0.419 0.421 0.430 Disposable personal income 0.407 0.402 0.407 2012 Money income 0.456 Personal income 0.436 Disposable personal income 0.416 NOTE. The Gini calculations use equivalized income using the square root of family size. Table 2. Personal Income and Its Disposition Published House-hold based Personal income 8,633 8,395 Compensation of employees 5,857 5,798 Wages and salaries 4,826 4,785 Supplements to wages and salaries 1,031 1,014 Employer contributions for employee pension and insurance funds 686 672 Employer contributions for government social insurance 345 342 Proprietors' income with inventory valuation and capital consumption adjustments 758 751 Farm 32 31 Nonfarm 726 720 Rental income of persons with capital consumption adjustment 188 180 Personal income receipts on assets 1,454 1,425 Personal interest income 1,070 1,050 Personal dividend income 383 375 Personal current transfer receipts 1,083 935 Government social benefits to persons 1,041 893 Social security 401 381 Medicare 219 155 Medicaid 200 140 Unemployment insurance 21 20 Veterans' benefits 25 25 Other 175 173 Other current transfer receipts, from business (net) 42 42 Less: Contributions for government social insurance, domestic 706 699 Less: Personal current taxes 1,232 1,189 Equals: Disposable personal income 7,401 7,205 2000 Billions of dollars Quintiles (levels) 1 2 3 4 5 Personal income 359 761 1,206 1,853 4,215 Compensation of employees 151 481 897 1,466 2,804 Wages and salaries 126 395 734 1,197 2,333 Supplements to wages and salaries 25 86 163 269 470 Employer contributions for employee pension and insurance funds 15 57 109 181 311 Employer contributions for government social insurance 10 30 55 88 159 Proprietors' income with inventory valuation and capital consumption adjustments 1 14 40 95 600 Farm 1 2 3 4 22 Nonfarm Rental income of persons 0 12 37 91 579 with capital consumption adjustment 11 19 25 35 90 Personal income receipts on assets 34 92 162 273 865 Personal interest income 30 81 137 228 573 Personal dividend income 4 11 24 45 291 Personal current transfer receipts 187 220 197 165 165 Government social benefits to persons 186 217 192 154 145 Social security 89 91 80 61 60 Medicare 41 38 31 23 21 Medicaid 24 41 34 27 14 Unemployment insurance 2 4 5 5 5 Veterans' benefits 2 2 5 6 9 Other Other current transfer receipts, 29 41 36 31 36 from business (net) Less: Contributions for 1 3 5 12 20 government social insurance, domestic 26 68 117 183 305 Less: Personal current taxes 13 45 106 228 797 Equals: Disposable personal income 346 716 1,101 1,626 3,418 Percent Quintiles (percent of total) 1 2 3 4 5 Personal income 4 9 14 22 50 Compensation of employees 3 8 15 25 48 Wages and salaries 3 8 15 25 4 Supplements to wages and salaries 2 8 16 27 46 Employer contributions for employee pension and insurance funds 2 8 16 27 46 Employer contributions for government social insurance 3 9 16 26 47 Proprietors' income with inventory valuation and capital consumption adjustments 0 2 5 13 80 Farm 2 6 10 13 70 Nonfarm Rental income of persons 0 2 5 13 80 with capital consumption adjustment 6 10 14 19 50 Personal income receipts on assets 2 6 11 19 61 Personal interest income 3 8 13 22 55 Personal dividend income 1 3 7 12 78 Personal current transfer receipts 20 24 21 18 18 Government social benefits to persons 21 24 21 17 16 Social security 23 24 21 16 16 Medicare 27 25 20 15 14 Medicaid 17 29 24 20 10 Unemployment insurance 10 21 24 22 23 Veterans' benefits 7 10 21 26 37 Other Other current transfer receipts, 17 24 21 18 21 from business (net) Less: Contributions for 3 7 12 28 49 government social insurance, domestic 4 10 17 26 44 Less: Personal current taxes 1 4 9 19 67 Equals: Disposable personal income 5 10 15 23 47 Published House-hold based Personal income 11,389 11,014 Compensation of employees 7,502 7,416 Wages and salaries 6,057 6,006 Supplements to wages and salaries 1,445 1,410 Employer contributions for employee pension and insurance funds 998 967 Employer contributions for government social insurance 447 443 Proprietors' income with inventory valuation and capital consumption adjustments 1,054 1,044 Farm 36 36 Nonfarm Rental income of persons 1,018 1,009 with capital consumption adjustment 208 198 Personal income receipts on assets 1,938 1,901 Personal interest income 1,215 1,193 Personal dividend income 724 708 Personal current transfer receipts 1,610 1,370 Government social benefits to persons 1,588 1,349 Social security 544 516 Medicare 399 282 Medicaid 299 210 Unemployment insurance 30 30 Veterans' benefits 39 38 Other Other current transfer receipts, 277 273 from business (net) Less: Contributions for 22 21 government social insurance, domestic 923 914 Less: Personal current taxes 1,352 1,305 Equals: Disposable personal income 10,037 9,709 2006 Billions of dollars Quintiles (levels) 1 2 3 4 5 Personal income 466 975 1,529 2,380 5,664 Compensation of employees 207 586 1,107 1,868 3,648 Wages and salaries 172 475 887 1,491 2,982 Supplements to wages and salaries 35 111 221 377 666 Employer contributions for employee pension and insurance funds 22 75 152 261 456 Employer contributions for government social insurance 13 36 68 116 210 Proprietors' income with inventory valuation and capital consumption adjustments 4 21 56 148 817 Farm 0 2 3 9 22 Nonfarm Rental income of persons 3 19 53 139 794 with capital consumption adjustment 12 19 27 41 99 Personal income receipts on assets 34 96 176 316 1,278 Personal interest income 31 85 151 254 671 Personal dividend income 3 11 25 62 607 Personal current transfer receipts 243 337 314 249 226 Government social benefits to persons 242 335 311 244 217 Social security 103 130 114 87 82 Medicare 63 74 62 44 39 Medicaid 35 62 55 38 20 Unemployment insurance 3 6 7 8 7 Veterans' benefits 2 5 10 10 11 Other Other current transfer receipts, 36 58 63 58 58 from business (net) Less: Contributions for 1 2 3 5 9 government social insurance, domestic 35 86 152 241 399 Less: Personal current taxes 28 65 131 265 816 Equals: Disposable personal income 438 910 1,398 2,116 4,848 Percent Quintiles (percent of total) 1 2 3 4 5 Personal income 4 9 14 22 51 Compensation of employees 3 8 15 25 49 Wages and salaries 3 8 15 25 50 Supplements to wages and salaries 2 8 6 27 47 Employer contributions for employee pension and insurance funds 2 8 16 27 47 Employer contributions for government social insurance 3 8 15 26 47 Proprietors' income with inventory valuation and capital consumption adjustments 0 2 5 14 78 Farm 1 4 8 24 63 Nonfarm Rental income of persons 0 2 5 14 79 with capital consumption adjustment 6 9 13 21 50 Personal income receipts on assets 2 5 9 17 67 Personal interest income 3 7 13 21 56 Personal dividend income 0 2 4 9 86 Personal current transfer receipts 18 25 23 18 16 Government social benefits to persons 18 25 23 18 16 Social security 20 25 22 17 16 Medicare 22 26 22 15 14 Medicaid 17 29 26 18 10 Unemployment insurance 10 19 24 25 22 Veterans' benefits 6 14 25 26 29 Other Other current transfer receipts, 13 21 23 21 21 from business (net) Less: Contributions for 5 11 16 25 43 government social insurance, domestic 4 9 17 26 44 Less: Personal current taxes 2 5 10 20 63 Equals: Disposable personal income 5 9 14 22 50 Published House-hold based Personal income 13,888 13,394 Compensation of employees 8,607 8,509 Wages and salaries 6,932 6,873 Supplements to wages and salaries 1,674 1,636 Employer contributions for employee pension and insurance funds 1,161 1,128 Employer contributions for government social insurance 514 509 Proprietors' income with inventory valuation and capital consumption adjustments 1,260 1,249 Farm 72 72 Nonfarm Rental income of persons 1,188 1,177 with capital consumption adjustment 533 512 Personal income receipts on assets 2,089 2,049 Personal interest income 1,256 1,235 Personal dividend income 833 815 Personal current transfer receipts 2,351 2,015 Government social benefits to persons 2,308 1,972 Social security 762 723 Medicare 555 392 Medicaid 417 292 Unemployment insurance 84 83 Veterans' benefits 70 69 Other Other current transfer receipts, 419 414 from business (net) Less: Contributions for 43 42 government social insurance, domestic 951 942 Less: Personal current taxes 1,504 1,451 Equals: Disposable personal income 12,384 11,943 2012 Billions of dollars Quintiles (levels) 1 2 3 4 5 Personal income 554 1,177 1,838 2,875 6,951 Compensation of employees 210 627 1,234 2,128 4,311 Wages and salaries 175 508 984 1,690 3,516 Supplements to wages and salaries 34 119 250 438 795 Employer contributions for employee pension and insurance funds 21 81 175 308 542 Employer contributions for government social insurance 13 38 75 129 253 Proprietors' income with inventory valuation and capital consumption adjustments 3 19 49 146 1,032 Farm (0) 1 2 11 58 Nonfarm Rental income of persons 3 18 47 135 975 with capital consumption adjustment 28 50 68 106 261 Personal income receipts on assets 28 88 176 346 1,412 Personal interest income 25 75 146 274 715 Personal dividend income 4 13 30 72 697 Personal current transfer receipts 320 476 464 398 357 Government social benefits to persons 319 474 457 388 334 Social security 128 175 163 135 121 Medicare 79 98 85 69 60 Medicaid 44 85 78 56 29 Unemployment insurance 11 16 20 19 17 Veterans' benefits 3 5 15 21 25 Other Other current transfer receipts, 54 94 95 87 82 from business (net) Less: Contributions for 1 3 6 10 23 government social insurance, domestic 34 86 154 249 418 Less: Personal current taxes 22 58 130 276 966 Equals: Disposable personal income 531 1,119 1,708 2,600 5,985 Percent Quintiles (percent of total) 1 2 3 4 5 Personal income 4 9 14 21 52 Compensation of employees 2 7 15 25 51 Wages and salaries 3 7 14 25 51 Supplements to wages and salaries 2 7 15 27 49 Employer contributions for employee pension and insurance funds 2 7 16 27 48 Employer contributions for government social insurance 3 8 15 25 50 Proprietors' income with inventory valuation and capital consumption adjustments 0 2 4 12 83 Farm -1 1 3 15 81 Nonfarm Rental income of persons 0 2 4 11 83 with capital consumption adjustment 5 10 13 21 51 Personal income receipts on assets 1 4 9 17 69 Personal interest income 2 6 12 22 58 Personal dividend income 0 2 4 9 86 Personal current transfer receipts 16 24 23 20 18 Government social benefits to persons 16 24 23 20 17 Social security 18 23 23 19 17 Medicare 20 25 22 18 15 Medicaid 15 2 27 19 10 Unemployment insurance 13 19 25 23 20 Veterans' benefits 4 8 21 30 36 Other Other current transfer receipts, 13 23 23 21 20 from business (net) Less: Contributions for 2 7 15 23 53 government social insurance, domestic 4 9 16 26 44 Less: Personal current taxes 2 4 9 19 67 Equals: Disposable personal income 4 9 14 22 50 Table 3. National Theil Index and Gini Coefficients, 2000-2012 2000 2001 2002 2003 2004 Theil index 0.372 0.371 0.371 0.372 0.389 Gini coefficients 0.424 0.422 0.422 0.423 0.430 2005 2006 2007 2008 2009 Theil index 0.386 0.403 0.378 0.396 0.369 Gini coefficients 0.427 0.436 0.427 0.433 0.419 2010 2011 2012 Theil index 0.368 0.406 0.419 Gini coefficients 0.421 0.430 0.436 Table 4. National Theil Decomposition 2000 2006 2012 National decomposition 0.372 0.403 0.419 Regional decomposition Between 0.002 0.004 0.003 Within 0.369 0.400 0.416 Northeast 0.076 0.081 0.086 Midwest 0.076 0.083 0.083 South 0.129 0.138 0.145 West 0.088 0.097 0.103 Table 5. Theil Decomposition for the Northeast Region 2000 Contribution Raw to Theil within Maine 0.397 0.009 Hew Hampshire 0.317 0.008 Vermont 0.397 0.004 Massachusetts 0.364 0.045 Rhode Island 0.335 0.006 Connecticut 0.303 0.020 New York 0.391 0.137 New Jersey 0.373 0.061 Pennsylvania 0.346 0.075 Mean 0.358 Standard Deviation 0.033 Within 0.366 Between 0.001 Theil 0.368 2006 Raw Contribution Theil to within Maine 0.275 0.009 Hew Hampshire 0.324 0.008 Vermont 0.376 0.004 Massachusetts 0.436 0.056 Rhode Island 0.402 0.008 Connecticut 0.438 0.034 New York 0.428 0.144 New Jersey 0.376 0.066 Pennsylvania 0.361 0.072 Mean 0.379 Standard Deviation 0.051 Within 0.400 Between 0.007 Theil 0.407 2012 Raw Contribution Theil to within Maine 0.359 0.009 Hew Hampshire 0.374 0.010 Vermont 0.364 0.004 Massachusetts 0.424 0.057 Rhode Island 0.463 0.009 Connecticut 0.437 0.033 New York 0.450 0.150 New Jersey 0.413 0.069 Pennsylvania 0.353 0.074 Mean 0.404 Standard Deviation 0.040 Within 0.414 Between 0.006 Theil 0.420 Table 6. Theil Decomposition for the Midwest Region 2000 Raw Contribution Theil to within Ohio 0.321 0.056 Indiana 0.303 0.026 Illinois 0.345 0.067 Michigan 0.340 0.055 Wisconsin 0.308 0.025 Minnesota 0.343 0.031 Iowa 0.323 0.014 Missouri 0.327 0.030 North Dakota 0.319 0.002 South Dakota 0.335 0.003 Nebraska 0.339 0.008 Kansas 0.310 0.012 Mean 0.326 Standard deviation 0.014 Within 0.329 Between 0.003 Theil 0.332 2006 Raw Contribution Theil to within 0.331 0.054 Ohio 0.346 0.031 Indiana 0.390 0.077 Illinois 0.358 0.053 Michigan 0.391 0.035 Wisconsin 0.343 0.030 Minnesota 0.353 0.015 Iowa 0.414 0.038 Missouri 0.386 0.003 North Dakota 0.362 0.004 South Dakota 0.333 0.009 Nebraska 0.453 0.020 Kansas 0.371 Mean 0.035 Standard deviation 0.369 Within 0.002 Between 0.371 Theil 2012 Raw Contribution Theil to within 0.351 0.054 Ohio 0.311 0.026 Indiana 0.477 0.098 Illinois 0.396 0.058 Michigan 0.424 0.038 Wisconsin 0.352 0.032 Minnesota 0.354 0.016 Iowa 0.397 0.035 Missouri 0.466 0.006 North Dakota 0.430 0.005 South Dakota 0.366 0.011 Nebraska 0.397 0.017 Kansas 0.393 Mean 0.047 Standard deviation 0.396 Within 0.004 Between 0.400 Theil Table 7. Theil Decomposition for the South Region 2000 Raw Contribution Theil to within Delaware 0.317 0.009 Maryland 0.373 0.025 District of Columbia 0.470 0.003 Virginia 0.402 0.036 West Virginia 0.303 0.005 North Carolina 0.337 0.025 South Carolina 0.316 0.012 Georgia 0.366 0.030 Florida 0.333 0.052 Kentucky 0.409 0.017 Tennessee 0.497 0.029 Alabama 0.357 0.014 Mississippi 0.365 0.009 Arkansas 0.334 0.007 Louisiana 0.343 0.012 Oklahoma 0.440 0.014 Texas 0.421 0.088 Mean 0.375 Standard deviation 0.055 0.381 Within 0.008 Between 0.389 Theil 2006 Raw Contribution Theil to within Delaware 0.272 0.009 Maryland 0.349 0.023 District of Columbia 0.495 0.004 Virginia 0.421 0.036 West Virginia 0.376 0.006 North Carolina 0.384 0.030 South Carolina 0.399 0.015 Georgia 0.315 0.026 Florida 0.416 0.075 Kentucky 0.430 0.015 Tennessee 0.375 0.019 Alabama 0.436 0.017 Mississippi 0.490 0.011 Arkansas 0.366 0.008 Louisiana 0.386 0.013 Oklahoma 0.457 0.014 Texas 0.436 0.089 Mean 0.375 Standard deviation 0.055 Within 0.403 Between 0.007 Theil 0.411 2012 Raw Contribution Theil to within Delaware 0.314 0.009 Maryland 0.361 0.024 District of Columbia 0.500 0.005 Virginia 0.454 0.040 West Virginia 0.384 0.006 North Carolina 0.400 0.031 South Carolina 0.393 0.014 Georgia 0.408 0.033 Florida 0.412 0.068 Kentucky 0.310 0.010 Tennessee 0.399 0.022 Alabama 0.464 0.018 Mississippi 0.466 0.009 Arkansas 0.382 0.008 Louisiana 0.383 0.013 Oklahoma 0.371 0.012 Texas 0.426 0.095 Mean 0.402 Standard deviation 0.049 Within 0.409 Between 0.009 Theil 0.418 Table 8. Theil Decomposition for the West Region 2000 2006 Raw Contribution Raw Contribution Theil to Theil to within within Montana 0.344 0.009 0.323 0.009 Idaho 0.346 0.006 0.333 0.006 Wyoming 0.326 0.002 0.353 0.002 Colorado 0.338 0.025 0.399 0.031 New Mexico 0.291 0.006 0.474 0.011 Arizona 0.371 0.028 0.393 0.032 Utah 0.315 0.009 0.293 0.009 Nevada 0.392 0.011 0.380 0.013 Washington 0.446 0.044 0.355 0.035 Oregon 0.394 0.022 0.433 0.023 California 0.379 0.211 0.417 0.224 Alaska 0.265 0.003 0.301 0.003 Hawaii 0.297 0.005 0.355 0.006 Mean 0.346 0.346 Standard deviation 0.048 0.048 Within 0.376 0.400 Between 0.003 0.005 Theil 0.379 0.404 2012 Raw Contribution Theil to within Montana 0.357 0.009 Idaho 0.374 0.007 Wyoming 0.352 0.003 Colorado 0.396 0.030 New Mexico 0.608 0.018 Arizona 0.388 0.031 Utah 0.349 0.011 Nevada 0.368 0.012 Washington 0.346 0.034 Oregon 0.341 0.017 California 0.454 0.242 Alaska 0.329 0.003 Hawaii 0.382 0.007 Mean 0.388 Standard deviation 0.070 Within 0.420 Between 0.005 Theil 0.426