A wage rate is a market price. The hourly wage is what it costs to purchase one hour of labor services of a certain type, in a certain place. Economists generally disbelieve in tampering with market prices unless there are good reasons. Are there good reasons to tamper with wages—specifically, to try to push them up? Maybe so.
First, unlike the prices of gasoline, shoes, or movie tickets, people’s wage rates are the bases of their living standards. Impersonal markets may assign wages to particular “low productivity” people that are so meager that they can’t support themselves, let alone their families. More generally, even when abject poverty is not the issue, market wages may lead to levels of inequality that many in society find intolerable. In such cases, governments may wish to intervene with measures such as minimum wages that are “above market” or so-called tax-and-transfer programs that raise after-tax net wages relative to pretax gross wages, such as the Earned Income Tax Credit.
Second, real wage growth (after accounting for inflation) has been abominable for most American workers for decades. Worse yet, what little wage growth there has been was concentrated at the very top. Wages at the median of the earning distribution and below have fallen or barely risen in thirty-five years. Figure 1 shows the dismaying “staircase” pattern of real wage growth by decile that has characterized the U.S. labor market since 1979—the approximate start of the Age of Inequality. At the 10 percent point of the wage ladder (counting from the bottom), real wages actually declined about 6 percent over thirty-three years. At the 50 percent point (the median), real wages rose a scant 5 percent. Even at the 90 percent point, the real wage increase was less than 1 percent per year. In stark contrast, real wage growth at the top 1 percent point—not shown here because it would be, literally, off the chart—was up 154 percent in thirty-three years.
Third, economists’ standard explanation of how wages are determined—the idea that each worker’s wage equals the value he or she adds to the production process—does not take us far in explaining what has happened to wages. Output per hour in the non-farm business sector (“labor productivity”) rose 93 percent over the thirty-three years covered by Figure 1, for a compound annual growth rate of 2 percent, but real compensation per hour (which includes fringe benefits) rose just 38 percent, a mere 1 percent per year. So even ignoring rising wage inequality, average American workers lost about 1 percent per year in wages relative to their productivity.
Markets are not supposed to work that way. Productivity is supposed to be rewarded in the pay packet. Indeed, markets did just that during the “golden age” of shared prosperity, 1947 to 1973. Doing the same calculation for those twenty-six years shows that labor productivity rose 2.8 percent per year while real hourly compensation rose 2.6 percent.
The entire postwar history of labor’s productivity and compensation is summarized in Figure 2, which shows a widening gap between compensation and productivity since the early 1970s. The divergence is actually pretty minor through 1979, but grows large thereafter. Furthermore, research by U.S. Bureau of Labor Statistics economists Susan Fleck, John Glaser, and Shawn Sprague shows that until the turn of the twenty-first century much of the growing gap between real compensation and productivity was accounted for by differences in “deflation,” that is, in how economists adjust wage and production data for inflation. Thereafter, labor’s share started to erode. But the story is, in some sense, even worse than Figure 2 indicates because part of the rising compensation was merely to cover the increasing costs of health insurance, not to improve living standards.
This analysis of America’s wage problem suggests several questions for policy: How can we raise the productivity of labor, especially that of relatively low-skilled labor? How can we close the gap between labor’s productivity and the compensation workers receive? How can we raise wages after taxes and transfers (net wages) relative to wages before taxes and transfers (gross wages)?
Let’s try to answer each of these questions in turn.
Raising the productivity of labor
There aren’t many mysteries here. A nation raises the productivity of labor by equipping its workers with more and better capital, by improving the technology of production, and by giving them more skills and training—starting, the evidence clearly shows, with pre-kindergarten.
The first method, boosting capital formation, has been the bedrock of U.S. economic growth policy (to the extent we have had any) for decades, leading to the creation of all sorts of tax incentives for saving and investment. The ratio of gross private domestic investment, which includes business investment and home building, to overall gross domestic product is highly cyclical, but it shows some modest upward trend since 1979 (marred by a huge collapse during the Great Recession of 2007-09), thus making a small contribution to productivity growth. (See Figure 3.)
Yet wages do not seem to have benefited much from the “success” in boosting capital formation. As we saw in the two earlier figures, real wage growth was anemic to negative except at the very top.
The second method, improving technology, has traditionally been the biggest source of growth in labor productivity and wages. But some observers, most prominently the coauthors of The Second Machine Age, Massachusetts Institute of Technology economists Erik Brynjolfsson and Andrew McAfee, have claimed that, in contrast to history since the Industrial Revolution, recent technological developments may be hostile to labor’s best interests. Only time will tell if forthcoming innovations in information technology will lead to faster or slower wage growth, but few economists expect technology to be a big game changer in labor’s favor in the short term.
The most direct approach to helping labor is, of course, to increase education, training, and workplace skills—especially for the lower 90 percent of wage earners. We have known for years that the financial returns to formal education, as measured by increased earnings, are high—perhaps even higher than the returns to capital. Economists usually view that fact as suggesting underinvestment in education because enough such investment would have beaten down the financial returns. A few pertinent facts are well known:
- We are miles away from offering all American children high-quality pre-K education, even though research speaks with one voice on how important this is to their future development. The rich make sure their children get what they need; poor children often need what they don’t get.
- The quality of our K-12 education system lags behind those of many other advanced (and some not so advanced) countries. It is also highly unequal, offering much-higher-quality education in higher-income areas.
- As a nation, we invest shockingly little in vocational training and apprenticeships, two aspects of education that would greatly benefit workers near and below the middle. Apprenticeships in the United States cover just 0.2 percent of the labor force, compared to 2.2 percent in Canada, 2.7 percent in the United Kingdom, and 3.7 percent in Australia.
- While some of our colleges and universities are still the best in the world, we no longer lead the world—or even come close—in the fraction of our young people who go to college. The college graduation rate in the United States is now below the average of all developed nation members of the Organisation for Economic Co-operation and Development.
Each of these observations and others offer a lever where policy could help raise the productivity of the lower 90 percent of our workforce. In virtually all cases, the expected returns on such investments are high even if we count only the incremental earnings and ignore human dignity or any other benefits from greater equality. Yet while the potential returns are high, so are the up-front costs—which may be why myopic political decisionmaking shortchanges the future by underinvesting in education.
Closing the gap between productivity and wages
Furthermore, education is supposed to work by raising productivity, which in turn leads to higher wages. But the latter link has been partially broken, and we don’t fully understand why. Part of the reason for the yawning gap between (faster) productivity improvements and (slower) wage gains is that non-wage fringe benefits (especially for health insurance) have grown much faster than wages. But the cumulative gap between productivity and compensation, which includes benefits, is still large and growing. Can (or should) government policy do anything about this?
No one has a full answer, but a few things would almost certainly help. First, we can raise the minimum wage, which has languished at $7.25 per hour for five years. The current proposal, which is dead in Congress, would boost it to $10.10 in three steps over three years. According to recent estimates from the Congressional Budget Office, doing so would directly help about sixteen million low-wage workers—netting out of those who lose their jobs. (Relative to the consensus emerging from many studies of the effects of raising the minimum wage, the CBO’s job-loss estimate seems a bit high to me.) Beyond that, we know that some workers whose wages are near the minimum wage would see their pay rates increase when the minimum wage goes up because employers are either bound to or want to maintain a differential above the minimum wage.
The lagging minimum wage is one important reason why wages around the tenth percentile of the workforce have failed to keep pace even with stagnant median wages. But it certainly doesn’t explain wage stagnation relative to productivity at the median and above. For that, we must look elsewhere for culprits.
The growing gap between compensation and productivity strongly suggests that labor’s bargaining power has fallen relative to that of capital. Why? One important reason is the entry of China, India, and the former Soviet Union into the global economy. This was (and still is) an earth-shattering series of events that effectively doubled the world’s labor force while raising the global capital stock very little. Such a gigantic shock to the global ratio of labor to capital would be expected to reduce wages and raise profits, probably quite a lot—which seems to be what has happened. Unfortunately, there aren’t any good policies to counteract market forces that powerful.
Another force that has probably reduced labor’s bargaining power during the Age of Inequality, the decline of unions, may have been aided and abetted by both government policies and increased business hostility toward unions. Employers, for example, have intensified their anti-union efforts in National Labor Relations Board certification elections. The decline in the rate of private-sector unionization since 1983 is dramatic and almost unbroken. (See Figure 4.) While we lack direct measurements of bargaining power, it is natural to assume that a 60 percent shrinkage in the unionization rate (from 16.8 percent to 6.7 percent), sustained over three decades, indicates a substantial erosion of bargaining power. That same logic, of course, suggests a policy for boosting wages—empowering unions by making it easier to organize—or at least trying to arrest their decline.
I have saved the best option, in terms of likely effectiveness, for closing the gap between productivity and wages for last: speeding up the rate of economic growth and tightening labor markets. It’s not a panacea, to be sure, but running a high-pressure economy is the surest way to boost the demand for labor—often for labor at all skill levels. And steep competition for labor is probably the surest route to higher real wages; firms will pay more for labor when they have to. Lately, with so much slack in labor markets, they haven’t had to.
It is no coincidence that the one recent period in which ordinary workers enjoyed wage increases and inequality stopped rising was the late 1990s, when labor markets grew very tight. Some people mistakenly believe that running the U.S. economy at that high pace will mostly benefit foreign workers as imports flood the nation. But even in today’s globalized economy, imports are under one-sixth of GDP. The rest is “Made in America.”
Figure 5 displays a snippet of evidence in support of the idea that tight labor markets boost real wage growth. It shows changes in inflation-adjusted compensation per hour (measured vertically) and the unemployment rate (measured horizontally) between 1948 and 2013. The scatter of data is clearly downward sloping, meaning that lower unemployment has been associated with faster wage growth. (For the statistically trained, the correlation coefficient is -0.41.)
To assist the eyeball, the chart also displays what is called a “regression line” that fits the data. It has an estimated slope of -0.39 (with standard error 0.11), which implies that a 1 percentage point lower unemployment rate for a year would raise the growth rate of real compensation by about 0.4 percent. If you recall that the observed gap between real compensation growth and productivity growth has been about 1 percentage point per year, it’s clear that labor market slack has been no trivial matter.
Reducing slack in the labor market also tends to result in more equal gains across the wage spectrum. Figure 6 shows a clear positive association between the national unemployment rate and the change in the so-called Gini Coefficient—the most popular measure of inequality—from the previous year. The Gini Coefficient ranges between 0 (perfect equality) and 1 (perfect inequality), and the vertical axis of Figure 6 records “Gini points.” If the Gini ratio rose from 0.410 to 0.415, for example, that would be +5 Gini points. The correlation here is +0.31. Notice that just about every observation with falling income inequality comes when unemployment is below 6 percent.
Thus there is good reason to believe that running an economy with a tight labor market would both narrow the gap between wages and productivity and equalize the income distribution. In recent years, however, the labor market has had lots of slack, and the U.S. government has pursued fiscal policies that move our economy further away from full employment—thereby counteracting the Federal Reserve’s strenuous efforts to do the opposite. Congress, for example, has rejected proposals to invest more in infrastructure or engage in other job-creation programs, while also cutting back sharply on discretionary spending.
Raising net wages relative to gross wages
The final way policymakers can boost wages is to enable workers, especially those on the lower rungs of the wage ladder, to take home more money after paying taxes and receiving transfers in the form of government benefits. It’s simple arithmetic—but difficult politics. An employer pays some gross wage to attract the labor he or she wants to hire. But the government taxes some of this away before it reaches the worker. In some cases, however, the government also augments the worker’s earnings with a tax credit or a transfer payment. So the net wage, after taxes and transfers, could be either higher or lower than the gross wage. For most of us, of course, it must be lower; we are net taxpayers. But if it wants to, the government can make low-wage workers come out ahead—receiving more in transfers than they pay in taxes.
We do this now in the United States in two main ways. First, low-income households were largely removed from the federal income tax rolls via tax cuts enacted under Presidents Bill Clinton and George W. Bush. For a family of four in 2014, personal exemptions and the standard deduction make the first $28,000 of income free of federal income tax. But low-income workers still pay federal payroll taxes—from the first dollar of earnings.
Second, the Earned Income Tax Credit provides transfer payments (via refundable tax credits) to low-income earners. But, oddly, its current structure subsidizes having children more than it subsidizes work. Today’s EITC is actually fairly generous to couples with children. A family of four in 2013, for example, received an effective wage subsidy of 40 percent on its first $13,430 of labor income, leading to a maximum tax credit of $5,370. For couples filing jointly, that credit started to phase out (at a 21 percent marginal rate) once earnings passed $22,870 and was entirely gone once earnings reached $48,373. Thus even families (of four) earning around the nation’s median income received at least some small benefits. If that same couple was childless, however, the wage subsidy rate was a mere 7.6 percent on the first $6,370 of earnings, which implied a puny maximum EITC benefit of $487—less than $10 a week.
It would not be difficult—again, leaving politics aside—to restructure this program to make it a true tax credit for earning income rather than for having children, that is, a wage subsidy. Both President Obama and Congressman Paul Ryan have proposed a first step in that direction by boosting the credit for childless workers. It would be a small step, to be sure; more should be done. But Congress has shown no intention of enacting the Obama-Ryan proposal.
Let me briefly recapitulate my analysis of ways to boost the wages of American workers, especially low- and moderate-income workers:
- Provide quality pre-K education for children whose families could not otherwise afford it.
- Improve the K-12 grade school system, especially in low-income areas.
- Provide much more vocational education and apprenticeship programs.
- Raise the minimum wage.
- Tilt the playing field in favor of, rather than against, unions.
- Run a high-pressure economy via a more stimulative fiscal policy.
- Increase the generosity of the Earned Income Tax Credit, especially for workers with no children.
It’s not exactly an exotic list, bubbling with surprises. But if Congress would enact it, American workers, and especially the most needy American workers, would get a raise.
Return to “American Life: An Investor’s Guide.”