Posted on Thursday, January 13, 2011
In response to this post, Brookings's Gary Burtless wrote in with some helpful comments on inequality and median incomes that I want to quote at length. Standard disclaimer: I've made light edits for space and clarity.
I asked what made inequality jump so sharply in 1987. Probably the tax-reform bill, Burtless says, which convinced a lot of wealthy people to stop sheltering their incomes in corporate vehicles. What you're seeing in 1987, in other words, isn't new inequality, but numbers that more accurately reflect existing inequality:
Most public finance economists who have investigated this anomaly in the Piketty-Saez inequality trend attribute it to the Tax Reform Act of 1986 (TRA86). The law sharply cut the top marginal tax rate in the personal income tax, both absolutely and, what is more important, relative to the corporate income tax rate.
Before 1986 individuals with sizeable stakes in a company had a strong incentive to shelter their income inside a corporation because the corporate tax rate was lower than the top personal income tax rate. On the margin, the income earned by a corporation was taxed at, say, 35%, but if the exact same income were directly received by a wealthy private individual it could be taxed (at a minimum) at, say, a 50% rate. The difference between the two rates made it advantageous to earn the profits inside the corporation and then to take the money out of the corporation in an advantageous year by selling stock in the company (paying only the capital gains tax rate on the income, a tax rate that was just one-half the rate on ordinary profits or dividend income). By lowering the top personal income tax rate to below or near the corporate tax rate,TRA86 changed the rich business owner’s best tax strategy. Now it was more advantageous to receive the income directly, report it on one’s personal tax return, and pay a modest rate on the resulting income flow.
As for stagnation among median incomes, Burtless says to watch compensation, which includes benefits, rather than just income:
I tracked down BLS data on the "Median usual weekly earnings - in constant (1982-84) dollars." It shows approximately the same result that you mention: Full-time American workers had $334 of real weekly earnings in 2000 and $342 in the most recent 2010 quarter available. On the other hand, when I go to the BLS productivity Webpage to obtain information about AVERAGE real compensation per hour, I obtain the following variable and results: "Average real compensation per hour" has increased 11.3% since 2000. The difference between the growth of the median real weekly and average real hourly compensation is due to four factors: #1. Real compensation / hour has increased faster than the real hourly wage. This makes sense. Employers’ cost for providing health benefits to their workers (which is in compensation but not hourly wages) has increased faster than wage costs: Compensation has grown faster than money wages. #2. The average number of weekly work hours has slipped, meaning that weekly earnings have increased more slowly than hourly earnings. #3. The worker in the exact middle of the earnings distribution may have done worse than the average worker. This is also easy to believe. If workers in the top 25% of the earnings distribution saw their wages climb 2% a year and everyone else saw their earnings rise just 0.1% a year, the average weekly wage would grow faster than the median weekly wage (which, in my example only rose 0.1% a year). Notice that explanations #1 and #3 are closely connected to one another, although they are distinct. It costs an employer roughly $5,000 a year to pay for health insurance premiums for a single worker and $10,000 or so to insure a worker with dependents. It costs just about the same to insure a worker paid $39,000 a year (approximately the median full-time wage) as it does to insure a worker who is paid $100,000 year. This means that a rapid increase in the cost of paying health insurance premiums will have a bigger percentage effect on the compensation of a worker paid the median wage than on the compensation of a worker paid two or three times the median wage. What I am saying is this: The rapid growth of employer health costs adds more (in percentage terms) to the cost of compensating a worker paid the median wage than it adds to the cost of compensating a worker paid an above-average wage. If employers give high-wage and median-wage exactly the same percentage increase hourly compensation, they would have to give the median-wage worker a smaller increase in hourly wage. Anyway, my guess is that explanations #1, #2, and #3 in some combination all contributed to the fact that median weekly earnings hardly grew at all after 2000, even though productivity increased fairly robustly over that period. #4. Finally, the “labor share” (that is, the part of the total output price that is captured by laborers as opposed to business owners) also fell fairly substantially over the period. According to BLS, the “index” showing labor’s share fell from 105.82 in 2000 to 95.36 in the most recent quarter. Business owners are simply capturing a larger fraction of the revenue being earned from private-sector businesses; workers are capturing less of it. The main reason for this, I think, is that workers are now and have been for the past decade in a very weak bargaining position. (At the moment, 10% unemployment tends to put a check on workers’ wage demands, I think.) Now is a nice time to be a business owner. For your average worker, not so much. The bottom line is this: Median consumption (including consumption of health care) has increased more smartly than median cash and near-cash income. Not surprisingly, the increased overall consumption of health care has absorbed more of the total (cash, near-cash, and non-cash) income growth obtained by the poor and middle class than it has the income growth of the very affluent.
By Ezra Klein | THE WASHINGTON POST