Report | Wages, Incomes, and Wealth

Full employment: Don’t give it up without a fight

Working Paper #122


As the unemployment rate began to tick up in recent months, comments like these began appearing in the press:

“…the economy is moving to a more normal, sustainable unemployment rate after a period of rapid growth.” -Neal Soss, chief economist at Credit Suisse First Boston, quoted in the Washington Post, May 5, 2001

“Unemployment, despite thousands of recent layoffs across a wide range of sectors, is still well below the rate commonly associated with stable inflation and growth.” –New York Times editorial, June 28, 2001

Important policy-making institutions echo these sentiments. According to the Congressional Budget Office (CBO) – the scorekeeper for budget battles for more than two decades – the recent rise in unemployment is simply a return to normal. It views the sustainable unemployment rate as being 5.2%, more than a full percentage point above the 3.9% low hit last year. The influential Organization for Economic Cooperation and Development takes a similar position.

These may sound like the measured, reasonable views of cautious analysts, but, as we see it, they are misguided positions with worrisome consequences. The cost of taking these ideas seriously is high, and it is a cost that falls disproportionately on the working class. In fact, in terms of living standards, working families have no better ally in today’s economy than full employment, and maintaining it should be our foremost goal.

Those who advocate settling for unemployment rates above the low and sustainable rates achieved toward the end of the last recovery are wolves in sheep’s clothing, engaged, perhaps unknowingly, in a subtle breed of class warfare. If, when the economy picks up steam again, we settle into unemployment rates that prevailed over the 1980s and early 1990s, we will be consigning middle- and low-income working families to the raft of economic problems that beset them over these years: stagnant incomes, falling wages, and growing inequality.

What is full employment?
For our purposes, full employment means that virtually everyone who wants a job has one. It doesn’t mean that there are no unsuccessful job seekers, i.e., that the unemployment rate is zero. It allows for frictional unemployment – the notion that a small share of the workforce is seeking jobs and will soon find them. Aside from that, at full employment the number of workers seeking jobs matches up neatly with the needs of employers; the supply of and demand for labor are in equilibrium, and the labor force is fully utilized.

Unfortunately, some groups of disadvantaged persons in the country have very high “structural” unemployment rates, meaning that their unemployment rates are consistently many times that of the overall rate. Yet, as we will show, these are the very persons helped the most by full employment.

What unemployment rate does this notion of full employment correspond to? It is difficult to pin down a precise number, but based on the experience of the past few years, it certainly is no higher than 4.0%. Consider some of the benefits that this low unemployment rate generated.

The post-1996 labor market: the benefits of full employment
The labor market was slow to recover from the last recession. Although the analysts who date recessions said the economy bottomed out in March 1991, unemployment continued to rise until mid-1992. It finally began to fall, and by January 1996 it had leveled off at around five and a half percent, about where it was at the end of the last recovery in 1989. Few expected it to fall much further.

But fall it did. In May 1997 if fell below 5% for the first time since 1973, and in September 2000 it hit 3.9%, the lowest rate in three decades.

The move toward full employment was extremely beneficial to millions of working families whose fortunes had been battered by slack demand for decades. As is usually the case when the labor market tightens up, the least advantaged posted the strongest gains.

For example, while the overall rate fell by 2.1 percentage points from 1994 to 2000 – from 6.1% to 4.0% – the rate for African Americans fell 3.9 percentage points, from 11.5% to 7.6%. The decline for black teenagers was particularly steep, from 35.2% to 24.7%, a drop of over 10 percentage points. For all blacks and for black teens – and for Hispanics as well – the 2000 rates were the lowest since data collection on these groups began in the early 1970s.

While we judge these gains to be evidence of the importance of full employment, we don’t want to lose sight of how high these rates remain. The fact that one-quarter of teenage black job seekers unsuccessfully sought work in the “best economy in 30 years” may seem to some a strange victory. Similarly, the ratio of black to overall unemployment was little changed, suggesting that structural problems persist. Still, though the unemployment levels faced by blacks are more than disconcerting, the trend over this period was impressive.

What about skills? Practically every labor market analyst argued that in the new economy the only thing that would help the least skilled was for them to get more schooling. Yet unemployment rates fell faster for high school dropouts over this period than for any other education group. True, their rates were the highest to start with, so they had more room to move, but this is still a trend that deserves close scrutiny. Did skills suddenly rain from the heavens on them, or had the pundits forgotten about the demand side of the equation? None of this is meant to deny the importance of skills, but clearly labor market tightness also matters.

In some ways, employment rates – the share of the population employed – are more revealing, because they capture the pull of the strong labor market drawing people in who were formerly not even looking for work (and thus not counted among the unemployed). If full employment is as great as we say it is, there should be some reversals evident here.

Employment rates grew fastest for the least skilled, increasing 4 percentage points from 1995 to 2000, while the rates for other education categories was fairly constant (the overall rate grew by 1.6 percentage points). Rates for whites increased by 1.3 points over this period; rates for blacks by 3.7 points. To really see the impact, you have to look at a particularly disadvantaged group, young African American women who didn’t finish high school. Their employment rates rose from 23% to 37%, a whopping 14-point increase.

The fact that this large increase is partly due to the work requirements in the welfare reform law passed in 1996 only underscores the importance of full employment in enabling poor women to find a job when they try to make the move from welfare to work.

The most important result of all this labor market tightening is that these employment gains translated into higher wages and incomes for broad groups of working-class families that had seen their incomes stagnate over the previous few decades.

  • The real hourly earnings of low-wage male workers, after falling at an annual rate of 1% from 1973 to 1995, a 20% cumulative loss, grew 1.5% per year from 1995 to 2000. For low-wage women, wages were flat over the earlier period, but grew 1.8% annually in the latter 1990s. 
  • The real wages of high school dropouts grew 1% per year after 1995, after fallingat about that rate from 1973 to 1995. 
  • After falling 0.6% annually over the 1980s and early 1990s, the real income of the poorest 20% of families grew by 2% per year from 1995 to 1999 (such data are only available through 1999). In real 1999 dollars, low-income families were $400 worse off in 1995 than in 1989. By 1995, their average income had increased by $1,000. 
  • This rise in income for the poorest families led to dramatic declines in poverty rates, especially for African American families, whose poverty rates fell 5.7 percentage points from 1995 to 1999 while the overall rate fell by 2 points (yet by 1999 the black poverty rate was still 23.6%, compared to 11.8% for the overall rate.)

Finally, the tight labor market, and the resulting increase in the bargaining power of low-wage workers, led to a clear slowing in the growth of inequality. The trend hasn’t reversed course, and thus most measures of income or wealth inequality remain at postwar highs. But over the 1980s and early 1990s inequality’s growth seemed inexorable, an unwelcome feature of the “new economy.” Many argued that this trend was the predictable outcome of technological changes, which offered huge rewards for the computer literate and punished everybody else. Now inequality’s upward trek appears to have been at least partly a function of too much slack in the labor market.

We could go on, citing convincing evidence connecting full employment to all kinds of positive developments, from increased home ownership to lower crime rates. But the above evidence still begs the question: what are the people and institutions such as those cited at the opening of this piece thinking? How could they celebrate leaving these beneficent conditions and heading back to the doldrums of a weak labor market?

The answer has a lot to do with a theory that has for decades prevented us from operating a full employment economy. That theory is the non-accelerating inflation rate of unemployment, or NAIRU.

Full employment vs. NAIRU
The reason why some experts take the view that the latter 1990s – the period when, in our view, the U.S. finally moved toward full employment – was unsustainable is because they believe that, when unemployment falls too low, it will lead not simply to faster inflation but to an inflationary spiral. That is, the inflation rate will continue to increase until the unemployment rate goes back up to its sustainable level, i.e., back to the NAIRU.

Furthermore, as noted above, those who subscribe to this doctrine think they know what that sustainable rate is, and they’re willing to use monetary policy (i.e., raise interest rates) to get us back up there if we should be so bold as to fall below.

There are two reasons why we think this theory is wrong. First, it rests on a weak foundation: there is no generally recognized explanation as to why, if we fall below the NAIRU, the inflation rate will continually accelerate. Second, there is no reliable way to figure out the NAIRU with enough certainty to guide policy makers.

The theory fails to make a case for accelerating inflation. It’s easy to show that there will generally be more upward pressure on prices in a period of low unemployment than when the unemployment rate is high. It is also clear that wages will rise more rapidly under such conditions. But there is no reason that this should necessarily lead to ever-accelerating inflation. Some firms will find themselves unable to fully pass on higher costs, and some workers’ wages will fail to keep up with the average rate of wage growth.

Given the problems with the theory, support for the NAIRU rested almost entirely on the ability of NAIRU models to accurately predict the inflation rate. While these models did provide reasonably accurate predictions for most of the postwar period, this changed in 1994, when the unemployment rate dipped below 6.0%, the generally accepted estimate of the NAIRU at the time. Since that point, the inflation rate has remained largely stable, even though the unemployment rate has fallen far below the NAIRU.

Undaunted by the experience of the last seven years, NAIRU proponents fought back by suggesting that the NAIRU changes in unexpected ways through time. Now, instead of having a NAIRU that is 6.0% for all time, we have a NAIRU that, by one recent estimate, was 7.7% in 1980, 6.4% in 1990, and 4.5% last year.
A time-varying NAIRU has appeal, even to non-believers. As discussed above, very low rates of unemployment can pull marginalized people into the labor force, providing them with job experience and skills that will enable them to be more productive workers in the future. This would mean that pushing the unemployment rate to low levels today will allow us to push it to even lower levels tomorrow.

But let us take a closer look at the implications of taking the predictions of the latest incarnation of NAIRU seriously, using a contemporary model by three highly accomplished macroeconomists.1 As just noted, their model calculates a NAIRU of 4.5% in 2000. Suppose for the sake of argument that a Federal Reserve Board chairman, whom we will refer to as Al, wanted to take that prediction seriously.

First, Al would have to note that the accuracy of the model is limited; the NAIRU could easily be 4%, given the lack of precision of the estimate. But let’s say he believed the higher guess (4.5%) and decided to play it fast and loose, letting the unemployment rate drift down without raising interest rates to stop the slide. What might be the cost of such profligacy?

According to the model, not much. The inflation process described by the NAIRU is a very gradual uptick in the rate of inflation. This latest model predicts that the cost of Al’s gambit would be a 0.56 percentage point rise in the rate of inflation over the course of a year. Given the huge gains to society from maintaining low levels of unemployment, the risks of this sort of increase in inflation (e.g., from 2.50 to 3.06 percent) are small.

Suppose we’re wrong, and the inflation rate does actually accelerate as the NAIRU theory predicts. Then, given the gradual and small movements in inflation just described, the unemployment rate can always be pushed back to the NAIRU before inflation reaches uncomfortable levels. Over the last seven years, the country has benefited enormously from the fact that Federal Reserve Board was willing to allow the unemployment rate to remain below the estimates of the NAIRU and to wait and see what happened.

The cost of leaving full employment behind
If we accept the premise of the pundits quoted at the beginning of this piece – that as unemployment rises we are simply heading back to more sustainable territory – we risk sacrificing these gains. And we do so not because the reality of the unemployment/inflation tradeoff demands it, but because of a theory that cannot be reliably used to guide policy.

If we reject the premise that our economy cannot sustain late-90s-style full employment, then the path is clear. When this slowdown plays out, we cannot settle back to the levels of economic activity that prevailed in the 1980s and early 1990s. We need to get back to 4% or lower as soon as possible.

Of course, monetary policy plays the key role. While we applaud the Fed’s recent experiment with full employment, it needs to explicitly adopt the view that unemployment above 4% is as serious a problem as the price pressures it most frequently bemoans. If unemployment is above 4%, the Fed chairman, in his periodic congressional testimony, should say why we’re not at full employment and what the Fed plans to do about it.

Then there’s the fiscal side of the equation. While inveighing against public spending may send valued political signals of fiscal rectitude, it is nonsense on the economics. Keynes imparted lasting lessons about government’s role in stimulating aggregate demand when “negative animal spirits” among consumers and investors drive unemployment up. Similarly, investment in needed public infrastr
ucture, from roads and communication systems to the development of human capital, can help boost economic production and lower unemployment.

Conclusion: learn from success
Rough times are ahead, as the economy recovers from recent excesses in the stock market, business over-investment, and high levels of household indebtedness. But the lessons of the late 1990s must not be lost.

The unemployment rate fell to levels that very few economists thought were obtainable just five years earlier, and the inflation rate did not rise. The 4.0% unemployment rate reached in this period should be the benchmark against which the economy is measured in the future. If anything, we should be looking to push the unemployment rate even lower, not throwing people out of work because of a questionable economic doctrine. This will require a continued commitment from the Federal Reserve Board to maintain low interest rates in order to support economic growth. And, given the slowdown, if lowering the unemployment rate requires expansionary fiscal policy – even deficits – then so be it. There is simply no policy that offers as much benefit to the country, and especially to the disadvantaged, as genuine full employment.

Jared Bernstein is an economist at the Economic Policy Institute. Dean Baker is co-director of the Center for Economic and Policy Research. This paper is drawn from their forthcoming Economic Policy Institute report on the benefits of full employment.

The authors wish to thank the Joyce Foundation, the John D. and Catherine T. MacArthur Foundation, the Charles Stewart Mott Foundation, the Public Welfare Foundation, and the Rockefeller Foundation for their support of this research.

A shorter version of this paper originally appeared in the American Prospect.

1. See James Stock, Mark Watson, and Doug Staiger, “Prices, Wages and the U.S. NAIRU in the 1990s,” January 2001, unpublished manuscript available at <>

See related work on Income and wages | Wages, Incomes, and Wealth

See more work by Jared Bernstein and Dean Baker