Deriving Your Facts from Your Conclusions

At the Wall Street Journal Laurence Kotlikoff critiques the NYT by Saez and Zucman I commented on last week:

Economists Emmanuel Saez and Gabriel Zucman of the University of California, Berkeley are advising Elizabeth Warren’s presidential campaign and drawing generous media attention for their assertion that the U.S. tax system is flat—that the middle class and poor pay as great a share of their income in taxes as the rich. They’re wrong, and three huge mistakes underlie their analysis.

The biggest mistake is to focus on gross, not net, taxes. They ignore transfer payments, like Social Security, which are disproportionately paid to the poor. In doing so, they mistake language for economics.

To see the problem, consider a simple thought experiment. Joe, a minimum-wage employee, earns $15,000 a year. The government taxes Joe $1,000 and transfers him $600, so that he pays $400 on net. In another scenario, Joe pays $10,000 in taxes and collects $9,600 in transfer payments. Again, he pays $400 on net. But in the first scenario, Messrs. Saez and Zuchman would report Joe’s gross tax rate as a reasonable 6.7% ($1,000 over $15,000). In the second, they’d call it an onerous 66.7%.

The flaw in their method is even more obvious when you consider that American tax laws incorporate benefit subsystems, and benefit laws incorporate tax subsystems. The federal personal income tax, for instance, includes the earned-income tax credit, a major transfer program. The Social Security benefit system includes the earnings test, a major tax. If the focus is on gross taxes, should the EITC be excluded and Social Security’s earnings test included? Economics can’t say. What it does say is look past the language and measure net, not gross, taxes.

Messrs. Saez and Zucman’s second mistake is measuring progressivity on a one-year rather than a remaining-lifetime basis. That ignores the fiscal system’s double taxation: Income earned, taxed and saved this year will be subject to future taxation on interest, dividends and capital gains. This omission disproportionately understates taxes for the rich, who save at a higher rate. The current-year focus also understates benefits paid to the poor, since future benefits are a bigger share of their resources.

Their third mistake is failing to adjust for age. The old have paid most of their lifetime taxes, which makes them now look like tax cheats, particularly those who saved out of previously highly taxed labor income. With changing demographics, this problem will deeply confuse tax progressivity comparisons over time.

Economic theory suggests a better way to measure fiscal progressivity: Include each and every federal and state tax and benefit; measure net, not gross, taxes; focus on each age group separately; add together the present value of each future year’s net taxes; and measure a household’s remaining lifetime net tax rate by dividing the present value of its remaining lifetime net taxes by the household’s remaining lifetime resources—its current net wealth plus its human wealth (the present value of its future labor earnings).

Yes, that’s a mouthful, and making the calculations is even more difficult than describing them. But another UC Berkeley economist, Alan Auerbach, and I have made them. Our findings, based on the Federal Reserve’s 2016 Survey of Consumer Finances, could not differ more from those of Messrs. Saez and Zucman.

I’ll focus on 40-year-olds, but the results are similar for all age groups. Each dollar of pretax remaining lifetime resources of those in the top 1% of the resource distribution is, on average, taxed on net at a 34.5% rate. For those in the top quintile, the average net tax rate is 28.4%. For those in the bottom quintile, every dollar of pre-tax resources is matched by a 46.6% net subsidy. (The tax rises steadily to 4.2% for the second quintile, 12.6% for the third and 18.5% for the fourth.)

Thus, far from being flat, the net tax rate facing middle age Americans rises rapidly with their resources—from negative 46.4% for the bottom 20% to positive 34.5% for the top 1%. The average net rates for the current year only (not including future net taxes) for this cohort understate true progressivity. They range from negative 9.8% for the bottom 20% to positive 38.2% for the top 1%.

Our study also considers inequality in remaining lifetime spending rather than partial measures like inequality in wealth, to which French economist Thomas Piketty has paid so much attention. Spending—including the ability to pass money to your child using bequests and gifts—is, after all, the key concern. Among 40-year-olds, the richest 1% own 34.1% of their cohort’s net wealth, but account for only 14.5% of its remaining lifetime spending. The poorest 20% own only 0.6% of the cohort’s wealth, but are expected to do 7.3% of its spending.

America’s progressive fiscal system, as well as the more even distribution of human wealth, makes spending inequality far smaller than wealth inequality. Nonetheless, average spending by top 1% households is miles higher than average spending by those in the bottom fifth. Like Messrs. Saez and Zucman and Ms. Warren, I am appalled by the degree of U.S. inequality and seek to reduce it. But if economists dramatically understate fiscal progressivity, it may encourage reforms that go far overboard.

Those are rookie errors and they’re exactly what you’d expect from a pair of young elite Frenchmen. Every instinct they have is wrong because it pertains to France rather than the United States. Let’s repeat Dr. Kotlikoff’s most significant observation:

The flaw in their method is even more obvious when you consider that American tax laws incorporate benefit subsystems, and benefit laws incorporate tax subsystems.

We don’t just need tax reform and entitlements reform. We need to disaggregate tax policy, benefits policy, and economic policy, a much taller order. With systems as complex as these have become over time it’s just too difficult to evaluate their effects. You can’t tell whether they’re working or not.