Earlier this week, the Congressional Budget Office (CBO) updated its authoritative data series on household incomes (1979-2005). The new data—highly regarded as a particularly complete source of information on this important topic—reveal a sharp increase in income inequality over the past few years. In fact, the increase in income inequality (both pre- and post-tax) as measured by the change in the shares of income going to different income classes, was greater from 2003 to 2005 than over any other two-year period covered by the CBO data. Over these years, an amazing $400 billion in pre-tax dollars was shifted from the bottom 95% of households to those in the top 5% (all income data in this report are inflation adjusted and in 2005 dollars). In other words, had income shares not shifted as they did, the income of each of the 109 million households in the bottom 95% would have been $3,660 higher in 2005.

Recent inequality developments
In response to the bursting Internet bubble and ensuing loss of capital income, inequality contracted for a few years in the early 2000s, but then reversed course in 2003 and started growing quickly in 2004.

Table 1 shows each household’s share of income for the latest three years of CBO data, including a breakdown of the richest fifth. By 2005, the top fifth held a larger share of both pre- and post-tax income than everyone else in the bottom 80%. On a pre-tax basis in 2005, the top 1%, with 18.1% of total income, held a much larger share of income than the bottom 40% of households, which only received 12.5%.

Figure A plots the changes in each income group’s share in the 2003-05 period. The first five bars reveal a shift in income from the bottom 80% to the top 20%. But by breaking down households within the top fifth, we find it was only the top 5% who saw gains, with small increases at the 95-99th percentiles, but truly large gains going to the top 1%. Together these gains amount to about a 4 percentage-point increase for the top 5%, the largest two-year increase over the history of these data.

Putting this shift in dollar terms reveals a redistribution of $400 billion of pre-tax income over these two years, an average loss of $3,660 per household to the bottom 95% of households. Again, these are the largest income shifts on record over a two-year period (see appendix table below).

Long-term growth and inequality

Figure B shows the long-term growth in real income for each fifth from 1979 to 2005. The “staircase” pattern of growth is evidence of the increase in unequal income growth over this period. Note also that, while the federal tax system raised the income growth of each group of families, it clearly did not turn back the inequality tide. To the contrary, after-tax income growth was much higher for the top 1% than was pre-tax growth (the decline in effective tax rates—the share of income paid in federal taxes—was largest for households in the top 1%).

Real income growth over this period was minimal to moderate for most households. Income growth for the poorest households grew only 1.3% pre-tax but 6.3% post-tax, thanks largely to the increase in the refundable Earned Income Tax Credit over these years. Middle incomes grew 15% pre-tax and 21% post-tax, or less than 1% per year over this 26 year period. Income for the top fifth grew much more quickly in the 1979-2005 period: 75% pre-tax and 80% post-tax. But the most dramatic growth occurred at the very top of the income scale: Households in the top 1% saw their income triple over these years, up by 200% pre-tax and 228% post-tax.

These trends led to stark differences in actual income levels by 2005. In that year, the average after-tax income for households in the bottom fifth was $15,300; for the middle fifth, $50,200; and for the top 1%, just over $1 million. These gaps have led to much greater economic distance between income classes over the years. Back in 1979, the post-tax income of the top 1% was eight times higher than that of middle-income families and 23 times higher than the lowest fifth. In 2005, those ratios grew to 21 (top compared to middle) and 70 (top to bottom), a vast increase in the distance between income classes.

Conclusion

The CBO data reveal the severe depth of our inequality problem. Though overall tax reductions during this period have meant slightly faster post-tax income growth for households in each income group, it has made little difference to the overall picture of inequality and has even exacerbated unequal outcomes over this period.

The problem is particularly stark in recent years. Before the current problems in housing and financial markets developed, the overall economy grew solidly over this recovery, with notably strong productivity growth. As the CBO data reveal, aggregate household income grew $1.1 trillion in the 2003-05 period (see appendix table). But these gains have failed to flow broadly throughout the income scale, and the extent of their concentration at the top of the income scale is historically unique. Just under two-thirds (63%) of the gain in household income from 2003 to 2005 went to just 5% of the nation’s wealthiest households.

Such concentration of income is unsustainable in a democratic society. The distribution mechanisms that have historically worked to ensure much more equitable outcomes appear to be wholly inoperative. Fixing them must be at the heart of any serious economic policy discussion.

Appendix

Table A-1 shows the method for assigning a dollar amount to the change in pre-tax income shares over the 2003-05 period. Aggregate household income is derived by multiplying the number of households by average household income.¹ We then show actual income shares and the income levels that correspond to them. Note that over these years, the aggregate income accruing to the top 5% was much greater than that accruing to the bottom 95%.

Next, we ask what those levels would be if the 2005 shares were unchanged from the earlier year, that is, we apply the 2003 shares to the 2005 aggregate amount. This shows, for example, that instead of going up $396 billion in the 2003-05 period, the aggregate income going to the bottom 95% would have gone up $794 billion. The difference between the simulated and actual changes amounts to $398 billion.

The CBO reports that there were 114.5 million household in 2005. Dividing the $398 billion by 95% of this number allowed us to determine the loss of $3,660 per household in the bottom 95%.

Notes

1. This raises the issue that in the CBO data, negative incomes are including in the average, but not in the shares of income by income class. This will have only a trivial effect on our results.