There is an easy story to tell about the Obama Recovery. Devastated by a financial crash, the U.S. launched a historic comeback. The private sector added jobs in 73 consecutive months, the longest stretch ever. Unemployment is lower today than in the month Reagan left office. Real GDP has grown more than 13 percent since its most-recent low in 2009, Obama’s first year in office. That’s more than twice as much growth as in some western European countries, like France. Compared to how countries typically perform after financial crises, the United States has “probably managed this better than any large economy on Earth in modern history,” President Obama told The New York Times Magazine.
But there is an opposite story that is attracting widespread support and millions of votes: The recovery is a failure. Donald Trump is an IMAX projection of white working-class grievances, calling America “a third-world country.” Bernie Sanders’s supporters describe a country where poverty and financial insecurity are not bugs but rather features of a rigged economy. The pessimistic style is not niche: Trump and Sanders have amassed a combined 16 million votes.
Taken together, the two narratives produce a dissonant, but not contradictory, summary of America: On average, everything is getting better, but for many people, nothing is going well.
This fits what statisticians call a “power law” distribution, where exponential equations (hence “power”) deliver extremely unequal rewards. Averages mean little in a power law. Imagine if 100 people enter a lottery, and 90 people win nothing; five people win $10; four people win $100, and one person wins $1,000. The lottery awards $1,450 to 100 people. The average prize is $14.50. But that statistic is nearly meaningless. Ninety percent of the participants won nothing, and one person won 69 percent of the cash. In a normal distribution, “average” is a useful indicator. In a power-law distribution, “average” is misleading.
The U.S. economy’s power-law features, in which averages disguise massive inequalities in outcomes, go a long way in explaining how Obama can tell a story about the economy vastly different from the ones that are propelling some presidential candidates. A prime example is the pattern of income growth. Between 2009 and 2013, most measures of real personal income showed slow but steady improvement. Average hourly earnings for private sector workers grew about 7 percent. But what about the distribution? The top 1 percent saw its disposable income grow by 11 percent. Everybody else got close to nothing. For the bottom 99 percent, income actually declined through the first five years of the recovery. “So far all of the gains of the recovery have gone to the top 1 percent,” the economist Justin Wolfers wrote.
There are several contributors to this power-law dynamic in U.S. wage growth. One is that the top percentiles of earners include a large number of consistently employed executives, lawyers, doctors, financiers, and other highly educated workers whose work and compensation weren’t interrupted by dramatic swings within the labor force. Meanwhile the middle of the labor market has undergone painful technological and demographic shifts. Globalization and technology have "hollowed out" the middle class. Between 1979 and 2011, the income of the second and third quintiles of households grew less than one percent per year. The income of the top one percent grew five-times faster. The quality of many middle class jobs is also eroding. Research by the economists Lawrence Katz and Alan Krueger found that between 2005 and 2015, the number of workers in alternative arrangements without health insurance or paid leave—like home health aides, truck drivers, and call-center workers—grew 66 percent, while the number of standard full-time jobs actually declined a bit.
Another contributor to this power-law economy is that geographical opportunity has become even more spiky. Workers in rich cities earn much more than their counterparts in poorer parts of the country. It’s not just because the price levels of cities are higher. They’re also more productive. In 2001, the 50 richest metros produced 27 percent more per person than the national average. Today, they produce 34 percent more.
Descriptions of average nationwide growth are, like the “average” lottery winner, statistically accurate representations that dissemble more than they reveal. At a time when San Francisco and Austin are bursting at the seams while Cleveland's workforce is actually shrinking, Toledo lost 4 percent of its businesses in the first four years of the recovery, and 61 percent of Detroit's adults aren't working, what sense does it make to say American cities are thriving “on average”?
Even among America’s largest companies, outcomes are diverging. In 2015, Jason Furman, the Chairman of the Council of Economic Advisers, and Peter Orszag, Obama’s former director of the Office of Management and Budget, published a paper on the rise in inequality between large companies. The top 10 percent of publicly traded companies earned 20 percent or more on their invested capital in the 1980s, and 100 percent or more in 2014, five times the average of the typical company. This suggests that “the rise in wage inequality is driven more by a widening gap in the average earnings of workers in different companies than by a widening gap between paychecks inside individual businesses,” wrote Orszag.
These power-law dynamics are fractal, appearing at the family, company, and city level. Assortative mating—rich marrying rich, poor cohabitating with poor—concentrates genetic, economic, and social advantages among children; corporate inequality and the return of monopolies concentrate profits among a handful of companies; geographic inequality concentrates neighborhood effects on both income and behavior. The U.S. smoking rate, for example, fell 50 percent in the last 50 years, but even this “average” change is heavily concentrated among the rich. The poorest families only smoke 9 percent less than they did in the 1960s.
Last week, I published an article called “The Average 29-Year-Old.” The point was to show that media portrayals of Millennials are misleading: “‘College-educated at 29,’ ‘living in a city at 29,’ or ‘married at 29’ all leave out more than 60 percent of the age group.” But as I wrote in the article’s conclusion, even the headline was misleading. “Average” is easy to report. But it’s most useful in normal distributions, like height, where the average looks like the median. This economy is many things. But it sure isn’t normal.