Who Bears Aggregate Fluctuations and How?

Featured in print Digest

The income [and consumption] of ... rich households [are] now more vulnerable to aggregate fluctuations than [those] of poorer households ...

In Who Bears Aggregate Fluctuations and How? (NBER Working Paper No. 14665), Jonathan Parkerand Annette Vissing-Jorgensen analyze the exposure of high-income households to aggregate booms and busts and find a significant break with the past in regard to who bears aggregate risk. The income -- especially the wage income -- of rich households is now more vulnerable to aggregate fluctuations than that of poorer households and the consumption of high-income households varies more with aggregate fluctuations in part because the income of these households varies more. This has clear implications for the effects of recent recessions on consumption inequality. Specifically, because total inflation-adjusted growth in per capita consumption during the past year was about 3 percentage points below its historical mean, the authors predict that the ratio of consumption of the top 10 percent to the bottom 80 percent of the income distribution has fallen by about 15 percentage points relative to trend.

Parker and Vissing-Jorgensen base this study on two somewhat disparate strands of prior research. One, which documents increases in income and consumption inequality over the past 25 years, focuses on the extent to which income shocks are insured. This literature pays relatively little attention to the extent to which that insurance differs across households. The second line of research involves asset pricing. It has documented that equity risk is borne disproportionately by households with large stock market wealth. This work looks mainly at differences in the relative variations of consumption growth with equity returns, not with aggregate fluctuations more generally. That difference is significant, because the share of aggregate income that comes from labor is roughly double the share coming from capital. As with the literature on consumption inequality, this research is limited by under-representation of households with very high consumption in standard datasets.

Parker and Vissing-Jorgensen present five main results. First, the consumption growth of high-consumption households is significantly more exposed to aggregate fluctuations than that of the typical household in the Consumer Expenditure (CEX) Survey (which was published annually from 1982-2004 by the U.S. Bureau of Labor Statistics.) The exposure to aggregate consumption growth of households in the top 10 percent of the consumption distribution in the CEX is about five times that of the average household.

Second, this pattern predicts that there was a significant decline in consumption inequality over the past year. With real aggregate per capita consumption growth about 3 percentage points less than its historical mean of 2 percent during the past year, the ratio of consumption of the top 20 percent to the bottom 80 percent is expected to fall by about 9 percentage points, relative to its evolution under trend growth.

Third, using income data from tax return studies, the researchers provide evidence on the channels that lead to higher exposure for high-consumption households. In the period covered by the CEX, they suggest, the greater income exposure of rich households to aggregate consumption-and-income fluctuations is a likely contributor to their higher consumption exposure. High-income households (the top 1 percent) earn more than half of their non-capital-gains income from wage income, and their wage income is far more exposed to aggregate fluctuations than that of lower-income households.

Fourth, Parker and Vissing-Jorgensen find even more exposure to aggregate fluctuations for very high-income households (the top 0.01 percent) than for other high-income households. This suggests that the consumption estimates in this study may understate the exposure of very high-consumption households thought to be omitted from the CEX.

Finally, the authors find a striking change over time in the exposure of the incomes of high-income households to aggregate fluctuations. Parker and Vissing-Jorgensen note that prior to the last 25 years, the incomes of high-income households were not more exposed to aggregate fluctuations. High-income households traditionally have had less of their income from wages and more from dividends, relative to the more recent period, suggesting higher exposure of the very high-income households to stock market fluctuations pre-1982. More importantly, in the earlier period the incomes of high-income households have had about the same sensitivity to aggregate consumption as the income of all households, and a lower sensitivity to aggregate income. This is mainly because of lower exposure of the wage income of the rich to changes in aggregate fluctuations in the earlier period and because of lower exposure of non-wage income (disproportionately earned by the rich) to changes in aggregate income in the earlier period.

-- Matt Nesvisky