Household Income Inequality in Atlanta by MikeJenny


									                         2008 PROPERTY TAXES ON HOMESTEAD

                                   Kelley Dean & David L. Sjoquist

               Household Income Inequality in Atlanta,
               David L. Sjoquist
               Rayna Stoycheva


               The income distribution in the United States changed significantly over the
               last thirty years. In general, the trend has been one of increasing income
               inequality, whether considering individual wages, family earnings, or
               household income. Family income inequality increased sharply from 1975
               through the 1980s. Inequality continued to grow during the economic
               expansion of the 1990s, but at a slower rate (Gottschalk and Danziger 2005).

               In this study we examine the distribution of income for the Atlanta five
               county area consisting of Clayton, Cobb, DeKalb, Fulton and Gwinnett
               counties, between 1980 and 2007. We consider two measures of inequality,
               the 90/10 ratio of household income and the share of income held by each
               income decile. To construct the 90/10 ratio, household income was sorted
               from lowest to highest. The 90/10 ratio is the ratio of household income at
               the 90th percentile to household income at the 10th percentile. We use the
               income at the 90th and 10th percentile rather than the highest and lowest
               household incomes in order to avoid extreme outliers that would produce a
               measure that is not reflective of the overall distribution of income. The
               income data comes from the public use micro samples available from the
               U.S. Bureau of the Census.

                   Atlanta Issues provides information and research on issues facing the Atlanta region.
                   It is sponsored by the Dan E. Sweat Chair. Atlanta Issues papers are available at
                                                    Dan E. Sweat Chair
                                              David L. Sjoquist, Chair holder
                            Andrew Young School of Policy Studies • Georgia State University
                                      14 Marietta Street NW • Atlanta, Georgia 30303
                                              404.413.0246 •

                    David L. Sjoquist & Rayna Stoycheva

Consider as an example, income for 2007. Using individual household data from the American
Community Survey from the Census Bureau, we determined the household income for the household
at the 10th and 90th percentile in Atlanta. Household income in 2007 at the 10th percentile was $15,420
and at the 90th percentile it was $165,714. (This implies that in 2007, 10 percent of Georgia households
had an income less than $15,420 and 10 percent had an income greater than $165,714.) The 90/10 ratio
is thus 10.75, which is $165,714 divided by $15,420. Thus, household income at the 90th percentile is
10.75 times the income of the household at the 10th percentile. If the incomes of the two households
increased by the same percentage, then the 90/10 ratio would not change. But if the income at that 90th
percentile increased by a greater percentage than the income for the household at the 10th percentile,
then the value of the ratio would increase, and we would say that income inequality increased

The other approach to measuring inequality is to compare the distribution of income. For this we first
array households by income and then divide the population of households into ten equal parts.
(Because the Census Bureau does not reveal that actual income for the very highest income households,
it was necessary to estimate income for those households. The Appendix describes how that was
done.) We then calculate the percentage of total household income that goes to each of the 10 groups. If
the income distribution was absolutely equal, each decile of households would have exactly ten percent
of total income. The more unequal the distribution of income, the greater the percentage of income
held by the higher income households.

Household Income Inequality

Household income at the 90th and 10th percentile, and the corresponding 90/10 ratios for all Atlanta
residents are shown in Figure 1. The values are for Census years 1980, 1990 and 2000, and for 2007.
Income inequality actually decreased between 1980 and 1990, and started increasing after 1990. This
trend is exactly opposite to the national trend. (As reported by Jones and Weinberg (2000), the 90/10
ratio for U.S. household income increased by 9.9 percent between 1980 and 1990, followed by much
slower growth during the 1990s.) This trend is also different from the one found in a similar study for
Georgia, where income inequality did not change significantly after 1990 (Sjoquist and Stroycheva
2009). Also, income inequality between 1980 and 2007 increased only by 5 percent, but the increase
between 2000 and 2007 is 11 percent, which is considerably high when compared to national and state

Figure 2 illustrates how the income distribution in Atlanta changed for each percentile between 1980
and 2007. The share of total income held by the top ten percent increased from 27.9 percent to 34.1
percent, while the share of the other groups decreased (except for the 1st and 9th decile). This amounts
to 22 percent increase, larger than the 14 percent increase in the share held by the top ten percent in

There are two changes underlying the large increase in the share of income held by the richest ten
percent of Atlanta residents. One is that income at the 90th percentile grew faster than income at the 10th
percentile. The second one is that incomes at the 95th and 99th percentiles also grew much faster, so their
share of total income increased more than what can be seen only by examining the 90/10 ratio.

Issue 2, January 2010                                                                                      2
                          HOUSEHOLD INCOME INEQUALITY IN ATLANTA, 1980-2007
                                        David L. Sjoquist & Rayna Stoycheva


                                                                                 $168,082             $165,714
    Household Income

                         120,000                                                                                                10th percentile
                         100,000                                                                                                90th percentile
                                80,000                                                                                          Log ratio
                                                      2.33               2.23                 2.27               2.37
                                               $10,707         $14,670            $17,360              $15,420
                                                     1980                1990                 2000               2007

                                          30                                                                             34.1
                       Percentage Share

                                          25                                                                            27.9
                                          20                                                                     16.6           1980
                                          15                                                           12.5
                                                                                                9.7           16.5              2007
                                                                                6.4     8.1
                                          10                4.0   4.9                                 13.4
                                                      2.6                                     11.2
                                                1.0                                    9.3
                                           5                          7.4
                                               1    2.9   4.3   6.3
                                                1        2     3     4         5   6     7              8        9       10
                                                                           Income Decile

The vast literature examining the determinants of income inequality has identified two sets of factors
responsible for changes in family and household income. Labor economists have examined extensively
the drivers behind wage and earnings inequality. The higher return to education due to technological
change, the decline in the inflation-adjusted minimum wage, de-unionization of industries, outsourcing
of jobs, and the increase in performance pay at the top end of the distribution, are some of the widely
recognized factors that have played a role in the observed change in earnings over the last three

There are also factors that are specific to family and household income inequality. Among these are, the
relative decrease in two-earner families at the bottom of the income distribution, the stronger increase
in labor supply of middle and higher income spouses compared to the bottom tenth percentile, the

3                                                                                                                                           Atlanta Issues
                      David L. Sjoquist & Rayna Stoycheva

higher increase in the wages of second earners at the top of the income distribution, and the relative
importance of other income sources at the top of the income distribution (Lee, 2008).

Inequality by Race and Ethnicity

We also explored how income distribution trends differ by the race and ethnicity of the householder.
The data are available in Table 1 and Table 2. As can be seen from the third column of Table 1, income
inequality among white Atlanta residents increased significantly after 1990. The total increase is almost
40 percent, from a ratio of 7.2 in 1990 to a ratio of 10 in 2007.

Similarly to the changes for all householders above, this increase is much higher that what was
observed for Georgia, where the total increase was about 10 percent. African-Americans and Hispanics,
on the other hand, experienced a drop in income inequality.

These changes in inequality are explained by differential changes in household income at the 10th and
90th percentile for each group. Income at the 90th percentile increased sharply for whites after 1990,
particularly between 1990 and 2000. Income at the 10th percentile increased mostly between 1980 and
1990. For African-Americans and Hispanics, income at the bottom increased rapidly between 1990 and
2000, while income at the top grew at a more moderate rate.

Table 2 illustrates the change in the share of income for each group between 1980 and 2007. The
extreme increase in inequality for white Atlanta residents is confirmed by the fact that the share of
income held by the richest ten percent increased by 24 percent, from 26.9 to 33.3. Everyone else, with
the exception of the 9th decile, experienced a loss in their share of income. For African-Americans, the
bottom four deciles experienced an increase in their share of income, while the higher income deciles
generally experienced a loss in their share of income. The trend is very similar for Hispanics, as well.

 Year                    All         White         American       Hispanic
 1980                   10.2          7.5            14.3            10.5
 1990                    9.3          7.2            12.8            10.7
 2000                    9.7          8.5            10.8             8.1
 2007                   10.7         10.0            10.5             6.7

Issue 2, January 2010                                                                                      4
                    David L. Sjoquist & Rayna Stoycheva

          ---------All---------   -------White-------     American        -----Hispanic-----
Decile      1980       2007         1980      2007      1980     2007       1980     2007
  1           1.0          1.0         1.3       1.2      0.6       0.9        0.9      1.7
  2           2.9          2.6         3.3       2.8      2.0       2.8        2.8      3.3
  3           4.3          4.0         5.1       4.0      3.6       4.4        4.6      4.8
  4           6.3          4.9         6.2       5.1      5.3       6.8        6.2      6.0
  5           7.4          6.4         7.9       6.5      7.0       6.3        7.3      7.5
  6           9.3          8.1         9.4       8.1      9.0       8.7        8.9    10.1
  7          11.2          9.7       11.0      10.0      11.5      11.0      10.8     10.9
  8          13.4        12.5        13.0      12.5      14.1      14.3      13.5     10.6
  9          16.5        16.6        15.9      16.6      18.2      16.4      17.8     17.1
 10          27.9        34.1        26.9      33.3      28.7      28.5      27.3     27.9

Summary and Conclusions

Income inequality for the Atlanta metro area exhibited a much different trend than what has been
observed nationally and for Georgia. During the 1980s, when nationally income inequality was
increasing, income inequality in Atlanta actually decreased, followed by the most substantial increase
between 2000 and 2007. Most likely these changes reflect demographic changes for the Atlanta metro
area as a result of the migration of high-skilled, highly paid professionals in more recent years, as a
result skewing the earnings of the top ten percent towards higher levels. Such individuals tend to be
white, which also explains the significant increase in inequality among whites observed for Atlanta.
The increase in the 90/10 ratio for whites in Atlanta is four times the increase found for whites in
Georgia over the same period, 1980-2007. The increase in the share of income is more similar—24
percent for Atlanta and 18 percent for Georgia. On the other hand, African-Americans and Hispanics
experienced a continuous decline in household income inequality through this period. This probably
reflects the fact that low paying service jobs were increasingly available in the metro area, thus
increasing the share of income held by the lower income deciles. However, individuals from these
groups were less likely to have the education and experience necessary for the high-paying service jobs
associated with the growth in inequality overall. Therefore, the share of income for the top two deciles
did not change much over this period in the case of African-Americans and Hispanics.


Bureau of Labor Statistics (2009). Consumer Price Index. Consumer Price Index Research Series Using
   Current Methods.

Gottschalk, Peter and Sheldon Danziger (2005). “Inequality of Wage Rates, Earnings and Family
   Income in the United States, 1975-2002.” Review of Income and Wealth 51(2): 231-54.

Jones, Arthur F. and Daniel H. Weinberg (2000). “The Changing Shape of the Nation’s Income
   Distribution, 1947–1998.” Current Population Reports, U.S. Bureau of the Census.

5                                                                                              Atlanta Issues
                    David L. Sjoquist & Rayna Stoycheva

Lee, Chulhee (2008). “Rising Family Income Inequality in the United States, 1968 -2000: Impacts of
   Changing Labor Supply, Wages, and Family Structure.” International Economic Journal 22(2): 253-72.

Sjoquist, David L. and Rayna Stoycheva (2009). Household Inequality in Georgia, 1980-2007. FRC Brief No.
    199, Atlanta, GA: Fiscal Research Center, Andrew Young School of Policy Studies, Georgia State

Appendix: Calculation of Income of Very High Income Households.

The Census Bureau protects the confidentiality of individual data in many steps, including top-coding
of very high or very low values. For example, if an individual in a particular county has very high
income it would be easy to determine the identity of the individual based on the Census data. For the
1980 Census, household income is top-coded at $75,000 (in 1979 dollars). In order to be able to calculate
total income and share of income, we need to know the actual income for those individuals with top-
coded values. One approach is to assume that the top-end of the income distribution follows a Pareto
distribution and estimate income above the cut-off point. We use the top 20 percent of household
income in order to estimate the income held by individuals above the cut-off point. We sum the
estimated household income for all individuals above the cut-off point, divide by the number of
people, and obtain the mean household income for this group. The mean is used in place of the cut-off
value in order to determine total income and share of income for each percentile.

About the Authors

David L. Sjoquist is Professor of Economics, holder of the Dan E. Sweat Distinguished Scholar Chair in
Educational and Community Policy, and Director of the Fiscal Research Center of the Andrew Young
School of Policy Studies at Georgia State University. He has published widely on topics related to state
and local public finance and urban economics. He holds a Ph.D. from the University of Minnesota.

Rayna Stoycheva is a PhD student in Public Policy. Her specialization is public budgeting and finance,
and her dissertation research is focused on pension governance and management.

Issue 2, January 2010                                                                                   6

To top