- The debate about inequality is also a debate about equality of opportunity
“As we acquire more knowledge, things do not become more comprehensible, but more complex and more mysterious.” - Albert Schweitzer
The state of debate on income inequality has reached exactly this juncture. To see why, consider the following. A recent co-authored paper by economists Thomas Piketty and Emmanuel Saez finds that the share of income going to the top 1 percent of the population has more than doubled from 9 percent in 1976 to 20 percent in 2011 . In other words, the top tail of the distribution now enjoys an ever-larger slice of the pie than it has historically done so. In a 2013 paper , however, economist Richard Burkhauser and his co-authors counter that by some measures, the growth in incomes at the top has been significantly lower than in incomes at the middle and bottom, such that the income share of the top quintile has in fact declined since the 1980s. The Congressional Budget Office’s (CBO) analysis shows that incomes of the top 1 percent grew by more than 250 percent between 1979 and 2007. In another recent paper , Greg Mankiw posits that perhaps inequality is not a problem as long as the compensation of the top 1 percent reflects their contribution to society—the “just deserts” principle.
Clearly, the more we know, the more elusive the real answer becomes.
One of the problems that plagues the research on income inequality is the lack of a common definition of income. Piketty and Saez use pre-tax pre-transfer income data from the tax records of filers and include realized capital gains. Thus, they fail to account for transfers and payments like Social Security, Medicare, food stamps etc. In addition, it is also worth noting that by focusing on reported taxable incomes, their data are biased by the fact that taxable incomes respond to changes in tax rates. There is a plethora of literature studying the taxable income elasticity . Essentially, when tax rates are high, reported incomes are lower since people engage in tax avoidance or tax evasion. Burkhauser, on the other hand, argues that the true measure of income should focus on the Haig-Simons definition of income. In other words, we need to include accrued capital gains on housing and other wealth along with earnings and transfer incomes to get at what people actually think of as income. The CBO provides a post-tax and post-transfers definition of income, but does not include accrued capital gains .
So after decades of research on income inequality, do we really know where we stand?
At the heart of this debate, the real conversation is about living standards. Are people enjoying better lives today than they did twenty to thirty years ago? In popular perception, income inequality is bad to the extent that it reflects a general worsening of conditions for the common man. When politicians pitch the top 1 percent against the middle income classes, it engenders a belief that somehow the wealthier families are enjoying their comforts at the expense of the middle and lower class families. But is that the truth? Or is it simply the case that the size of the economic pie has grown over time, and while everyone is enjoying the benefits, a larger share of the benefits are going to the top?
To study this issue further, my colleague Kevin Hassett and I conducted a study to track changes in consumption inequality over time. Why did we use consumption instead of income to track inequality? One thing economists agree upon is that consumption is a better measure of well-being than income . What we buy and consume with our income directly adds to our utility and happiness, and also has a direct impact on our standard of living. Individuals are also better able to smooth consumption rather than income over their lifecycle. While a retired, older individual has low levels of current income, he can still enjoy a high standard of living due to lifetime savings and other forms of wealth. A student with low current incomes can borrow to finance education and household expenses in the hope of earning high incomes in the future from a relatively well-paying job. So one reason why income and consumption are de-linked is the possibility of borrowing and saving. In fact, it is probably rational to assume that at least some part of a poor or low-income family’s consumption is being sustained by indebtedness. However, another reason for the mismatch between income and consumption is likely the tax and transfer system. Many redistributive policies support consumption for low income households and provide transfer payments to them. As we discussed previously, most studies of income inequality are unable to get at these transfer payments.
Before I get to the main findings from our study, I would like to point out that consumption inequality has also been extensively studied in the literature, though perhaps not as much as income inequality. Results from these papers are also mixed. Krueger and Perri find that while income inequality increased during the period 1980-2003, consumption inequality did not. However, Blundell and colleagues show that income and consumption inequality diverged between the 1970s and the 1990s. Other papers find that income and consumption inequality have tracked each other closely since the 1980s. In general, the data used in these studies comes from either the Consumer Expenditure Survey (CEX) or the Panel Study of Income Dynamics. There are two problems with using the CEX to measure consumption inequality, however: measurement errors and a lack of information on the consumption of durable goods. Several authors, including Attanasio et al. and Aguiar and Bils , have attempted to account for the measurement errors in this data, by using techniques that enable them to predict expenditures for the less well-measured items by using information on better-measured items. While we do not criticize or applaud their approach, it does imply that in order to get anything meaningful from the CEX data for measurement of inequality, authors need to rely heavily on modeling assumptions and non-standard approaches, as opposed to simply using the raw data.
Precisely for these reasons, in our study, not only do we work with the CEX data but we also supplement our analysis with the use of the Residential Energy Consumption Survey (RECS) data. The CEX provides a good overview of nondurables’ consumption by American households. In 1984, households in the top income quintile accounted for 37 percent of total expenditures, while households in the bottom quintile accounted for 10 percent. Hence the ratio of top to bottom consumption was 3.7. In 2010, that ratio increased to 4.4. The gap was widest in 2005 when the share of consumption for the top was 39 percent relative to 8 percent at the bottom. In the most recent recession, it appears that households at the bottom increased their share by 1 percentage point while the share at the top either declined or remained steady. In the 2001 recession, the ratio declined as well, suggesting that recessions work towards a more even distribution. On average, over the entire period, the ratio is 4.3 with a standard deviation of 0.22. Therefore, using this measure, we find that consumption inequality has increased only marginally over time.
If we compare these trends in consumption to trends in income using the Current Population Survey data, which is widely used for research on income inequality, we find that the story is strikingly different. As per our analysis, in 1984, pre-tax incomes at the top were more than 11 times incomes in the bottom quintile. In 2010, that ratio rose to 15.4. The average for the entire period is 13.5. Clearly, inequality using annual incomes is significantly higher than when we use consumption, and it has tended to widen over time.
As mentioned earlier, the CEX is not a good source of data on durable goods consumption. Therefore, we worked with the RECS data as well. This survey has questions on household use of appliances such as microwaves, dishwashers, computers, printers and other data. What we find is that the access of low-income Americans – those earning less than $20,000 in real 2009 dollars – to these devices that are part of the “good life” has increased. The percentage of low-income households with a computer rose to 47.7% from 19.8% in 2001. The percentage of low-income homes with six or more rooms (excluding bathrooms) rose to 30% from 21.9% over the same period. Similar increases can be documented for appliances like air-conditioners, dishwashers, microwaves, cell phones and other household items.
In general, we find that people at all income levels now have access to many more material possessions than they did in the 1980s. Moreover, there has been a narrowing of the gap between high and low income classes in terms of ownership of these items. It is hard to argue against the improvement in the standard of living that has accompanied these trends.
Hence, the standard narrative that rising income inequality has somehow hurt the middle and lower income classes is not supported by data. Policies aimed at redistributing incomes from the top to the lower income classes have certainly been responsible for part of this trend. However, we would caution against using this argument for raising marginal tax rates at the top to levels seen in the 1970s. In another co-authored piece , my colleagues and I argue that the Diamond and Saez solution to inequality—a marginal tax rate of 73 percent —is based on unrealistic assumptions relating to how individuals would respond to high tax rates. Their modeling of the optimal rate assumes a “more equality is better” social welfare function and assigns no social value to the marginal dollar of consumption for the rich. Most importantly, it ignores the long-run behavioral responses and consequences of having marginal tax rates that are over 50 percent. In the article, we show that while these assumptions work well in theoretical models that are aimed at catering to an audience of professional economists, these should not be used as the basis of real world public policy formulation.
Whether the explanation for improvement in living standards lies in redistribution policies and the growth of the safety net, or technological improvements that allowed prices of electronics and other durable goods to drop, or real improvements in productivity and wages, the bottom line is: people are better off today than they were twenty or thirty years ago. Households are consuming more and the typical low income household possesses many more appliances and gadgets that have traditionally been considered the preserve of the rich, than at any time in history. Judging by these criteria, inequality is much less of a predicament than most politicians would have you believe.
To conclude, the debate about inequality is also a debate about equality of opportunity. If the problem is disparities in income between the rich and the poor, is there a way that economic and social mobility can at least enable people at the bottom to aspire to be the next Bill Gates? Is equality of opportunity the answer to income inequality? I believe the answer is yes. In a recent column , Arthur Brooks states that 70 percent of Americans believe that everyone should get a chance to succeed or fail, on his or her own merits. But 30 percent prefer a world in which everyone ends up in roughly the same place regardless of abilities and efforts. Like most Americans, I believe in equality of opportunity and not equality of outcome, which is clearly neither desirable nor attainable. By providing the right framework for growth and free enterprise, so that people can aspire to work, earn and make a good living, we can ensure that America remains a land of opportunity for everyone in society. But to penalize the success of the rich for some strange notion of fairness is to aim for a fundamentally different kind of equality than most Americans care about.