Jobs For All

In an age of precarity, a left-wing demand for full employment could be massively popular. But liberalism can’t deliver it.

Line for unemployment benefits in San Francisco, CA in 1938. Dorothea Lange / Library of Congress

The new issue of Jacobin is out now. To mark its release, we’re offering discounted introductory subscriptions.

“Full Employment . . . has become an aim of Conservative policy and the strongest argument against socialist critics.” That’s famed economist Joan Robinson, in 1962 trolling to her left and her right. British unemployment had been below 2 percent for most of the period since the war, without runaway inflation. Keynes had solved the problem of unemployment, converted the Conservatives, and stolen the communists’ best argument. Capitalism apparently didn’t need a reserve army of labor after all.

Just fifteen years later, like a Star Wars opening crawl: “The hopes which accompanied the Keynesian revolution, of reforming capitalism so as to ensure continuous prosperity with full employment, are now all but extinguished. The slide into crisis in the capitalist world has re-established the pre-Keynesian orthodoxy as the conventional wisdom at both national and international levels.” The rentier strikes back.

Two generations have since come of age in a world where getting a job and building a career is a fierce competition against your peers. Even the winners are anxious. Comfortable spots are precarious; the losers have nothing to blame but their CV. In a buyer’s market it seems like the employers are bringing the goods; they create the jobs, we just work in them.

It is in this climate that the demand for full employment is resurfacing with a vengeance. The importance of full employment is not just that, when people’s subsistence depends on selling their labor power, being unemployed sucks. If that were all, cutting a few percentage points off the unemployment rate would be a worthwhile reform but nothing to build a platform around. The bigger point is that the tightness of the labor market affects the whole working class.

A tight labor market is a seller’s market. It reverses the normal order of things. Chris Maisano made the point with the memetic figure of “Old Economy Steven”: “Steven can tell his boss to shove it, walk out and get hired at the factory across the street . . . If he gets fired at the new job, that’s no big deal. He’ll just pick up a new one on the way home. If he wants a raise, he can just walk into the boss’s office and demand one . . . he doesn’t have to bow before anyone to make ends meet.” Old Economy Steven is half myth, but only half.

There is little to add to Maisano’s case for making full employment the centerpiece of socialist strategy: “A full employment economy raises the bargaining power and living standards of the working class in the short run and erodes the relative power of capital, opening up possibilities for radical social transformation.” In case it needs to be said, it doesn’t depend on nostalgia for postwar boom years themselves, whose fruits were unevenly distributed. (“There’s a reason,” Maisano notes, “why a dorky-looking white dude named Steven is used as the avatar of working-class security and agency.”)

Demanding full employment also doesn’t mean glorifying wage slavery: it’s perfectly compatible with converting technological advances into leisure time. Full employment is simply a state in which demand for labor matches supply. If supply falls because working people earn enough to be comfortable from fewer hours of work, that’s great. In fact the aims support one another: full employment gives workers the bargaining power to reap the benefits of productivity growth as they choose, while a managed reduction in working hours spreads the demand for labor around. (A program of work-sharing and voluntary time reductions is a key part of Baker and Bernstein’s 2013 full employment plan.)

Some on the Left have taken the view that full employment is a regressive goal because automation is steadily reducing capitalism’s need for workers. Better to redistribute to the surplus population through a universal basic income. But the idea that machines are about to supplant workers is a trope as old as capitalism itself. It always looks plausible because so many particular tasks are always in the process of being automated, and new wonders are always just around the corner. And yet new jobs have always come along. Maybe this time of driverless cars and machine learning really will be different, but it would be foolish to place our political bets on it, especially before it starts showing up in the productivity growth stats.

Technological change can certainly be very disruptive to particular jobs and the value of particular skills. In recent decades, people have lost stable, well-paying careers in manufacturing and found their replacements precarious and low-paid. Almost all net job growth in the United States since 1990 has been in low-productivity growth sectors: construction, retail, hospitality, health care, education, government, and finance.

Recent times have been no exception to the historical pattern of workers being displaced from some sectors and absorbed by others. This decade, net job growth has been concentrated in education, health care, social assistance, hospitality and retail, sectors with average hourly pay and weekly hours much lower than the economy-wide average. For Matthew Klein in the Financial Times, this suggests “the growth of make-work has been the main thing preventing mass joblessness.”

But there is no technological reason that such jobs be low paid and insecure. The gains to labor productivity growth are not permanently captured in the industries where the advances occur, but spread through the economy in a complex pattern of price and wage changes. Some jobs are simply not as open to productivity improvements as others, especially those like teaching, medicine, hairdressing, waiting, cooking, and so on that (at least so far) depend on a human touch. If wages for such jobs didn’t keep up with wages elsewhere, workers would want to move. Potential labor mobility is supposed to be a force, at least in the long run, for equalizing returns to various kinds of work. But it takes time, and it depends on the overall state of the labor market.

Rather than moving into inherently low paid jobs, the American workforce has been shifting sectors at a time of chronic labor weakness. It has moved from organized industries with longer-standing norms in wages, hours, and conditions to a casualized wasteland. Their lower pay is not explained by lower relative productivity, because their apparent relative productivity depends on the value of the product, which depends on the pay.

Higher pay for retail workers, teachers, nurses, drivers, fast food workers, and so on would mean higher prices for their products, and would certainly change the pattern of demand and of jobs themselves, though in ways that are hard to predict in advance. Our current division of labor is a product of a pattern of demand based on a highly unequal income structure. The problem over the longer run is displaced workers not joining an ever-growing scrap-heap, but spending their lives in insecure service to the winners.

Natural Misery

So what are the prospects for full employment as a project for the Left? First we need to acknowledge that it is not only a project of the Left. No part of the political spectrum proclaims its partiality to unemployment. Conservatives never gave up on full employment, exactly. They just redefined it in terms of price stability instead of jobs. Milton Friedman’s main lasting impact on economics was the notion of the “natural rate of unemployment.” He told a story that Keynesian economists believed about a stable trade-off between unemployment and inflation: governments could manage demand so as to reduce unemployment so long as the public would accept the resulting higher rate of inflation.

But they somehow missed a simple point: once people learned to anticipate the inflation, they would get in ahead of it, raising wages and prices to defend themselves, so that the expected rate of inflation became the baseline, and inflation would have to rise even higher to keep unemployment down. But then that higher rate would feed into expectations: prices would have to accelerate to keep unemployment standing still and eventually runaway inflation would compromise the market system. In the long run, there was no trade-off between inflation and unemployment.

Friedman argued that there was a “natural” rate of unemployment — the one the market would reach in equilibrium, if everyone bargained on the basis of correct expectations. One way or another the economy would settle at that rate, and the only question for policymakers was how painful they wanted to make getting there. Instead of “natural,” economists adopted the awful but more literal “non-accelerating inflation rate of unemployment,” or NAIRU.

Friedman himself never meant to imply it was natural in the sense of an unchanging law of the universe: “many of the market characteristics that determine its level are man-made and policy-made.” The Congressional Budget Office’s estimates for the United States find it varying from 5.3 in the early 1950s, peaking at 6.3 in the late 1970s, and falling below 5 percent in the 2000s, until the crisis.

Why does it vary? Friedman suggested minimum wage rates, various labor laws, and strong unions as factors that “make the natural rate of unemployment higher than it would otherwise be.” The idea that a “freer” labor market is the answer to unemployment is a popular one — often wheeled out to explain Europe’s chronic unemployment in the 1990s and 2000s — but faces a big problem. A number of countries sustained genuine full employment (unemployment at 2-3 percent or even lower) for many years during the postwar long boom, without runaway inflation: Australia, Britain, France, Japan, Sweden, West Germany. They had highly regulated labor markets, sometimes centralized wage bargaining and often strong unions.

But the factors Friedman mentioned are not just any interferences in the labor market, but specifically those that boost workers’ bargaining power. It may be that their labor market institutions helped make full employment compatible with price stability in those countries because they restrained wage growth. That was certainly how they were often seen at the time: in Australia bosses pleaded for the Arbitration Commission to restrain wages and economists lamented inflationary extra pay negotiated out on the market.

Robert Pollin argues that the NAIRU can be seen along the same lines as the Marxian “reserve army of labour.” In Marx’s day, the gold standard prevented chronic inflation so that tight labor markets translated more directly into higher real wages, which could eat into profits only until slowing accumulation lowered demand for labor. Off gold, the drama played out in a more complex way, with rising labor costs feeding inflation, but the story is similar. Pollin points to Robert Gordon’s 1997 analysis of the “time-varying NAIRU.” Gordon notes that the American NAIRU seemed to rise in the 1960s — “a time of labor militancy, relatively strong unions, a relatively high minimum wage and a marked increase in labor’s share in national income” — and fall in the 1990s—“a time of labor peace, relatively weak unions, a relatively low minimum wage and a slight decline in labor’s income share.”

The Friedman story is now often treated as an explanation for runaway inflation in the 1970s: Keynesian policymakers flew too close to the sun in trying to exploit a trade-off. But when economists influenced by Friedman first estimated the Australian “natural rate” of unemployment in the early 1970s, the numbers they came up with were between 1.7 and 2.3. This is remarkably low only in retrospect: it was not out of line with the unemployment rates actually enjoyed, without inflation, for the previous two decades. By the end of the decade, unemployment was above 6 percent and inflation around 10 percent.

Later backward-looking estimates of the Australian NAIRU suggest that it leapt upward suddenly around 1974, in the recession that marked the end of the long boom. The evolution of the American NAIRU is less dramatic, because postwar unemployment was never so low. But in any case, the Friedmanite “expectations-augmented Phillips curve” gives no explanation for the 1970s crisis, because it doesn’t explain shifts in the NAIRU itself.

As James Forder has shown in a recent book, it is a myth that a Phillips curve trade-off guided macroeconomic policy before Friedman showed up to point out the problem. Neither economists nor policymakers were willing to accept inflation as a price of low unemployment in the postwar period. The idea that inflation could take off in a wage-price spiral was nothing new; it was a commonplace that feeding inflation was playing with fire.

From the outset, economists worried about the inflationary potential of full employment. For conservatives, this was a reason not to pursue it in the first place; for social democrats a reason to supplement demand management with other measures. Some called for extensive industrial planning, widening bottlenecks rather than pulling back demand. More common was a call for centralized wage determination. Jan Tinbergen, the first economics Nobelist and liberal arch-technocrat, argued that combining full employment with monetary stability demanded policy control over wages.

The basic economic rule was that nominal wage growth should not outpace productivity growth. That way real wages would rise in line with labor productivity without fueling inflation. Economists lectured the labor movement that workers had nothing to gain from aiming higher: labor’s income share was determined by technical factors and faster money-wage growth would only be lost to inflation. The question was whether labor would restrain itself, whether restraint could be guaranteed through centralized wage bargaining, or whether it would take a taste of unemployment.

In the event, the Keynesians were as surprised as anyone by the success of full employment. William Beveridge had suggested that full employment would mean 2-4 percent in practice. According to Joan Robinson, “Keynes thought this much too optimistic” and she herself had suggested 5 percent. The actual British experience of sub-2 percent surprised even the most ardent. Most astonishing to those who expected a return to slump after the war, it was underpinned by surging private investment. A growing public sector and “cheap money” surely helped, but deliberate Keynesian macroeconomic policy was often turned to restraint, rather than stimulus.

Michal Kalecki’s 1943 paper on the “Political Aspects of Full Employment” is well known on the Left as a prophecy of the end of full employment. He argued that the macroeconomic revolution had solved unemployment as an economic problem, but that capitalists would thwart it for political reasons. Full employment levels of demand were good for profits, and most firms had the pricing power to recoup what they paid out in higher nominal wages, so their objection was not one of cold calculation.

Instead, full employment threatened the power capital had over the state: governments no longer depended on petulant “business confidence” but could make up any shortfall in private investment on its own. Governments would be tempted out of the narrow bounds of roads, schools, and hospitals and expand onto capital’s territory.

Most frightening, “‘the sack’ would cease to play its role as a disciplinary measure. The social position of the boss would be undermined and the self-assurance and class consciousness of the working class would grow . . . ‘discipline in the factories’ and ‘political stability’ are more appreciated by the business leaders than profits.” Captains of industry would join forces with rentiers threatened by inflation, “and they would probably find more than one economist to declare that the situation was manifestly unsound.”

Prescient indeed, but the distinction between economic and political barriers to full employment is actually difficult to draw. Political barriers limit the “technical” means available to policymakers, and economic problems can easily turn political. Look at the ingredients in Kalecki’s recipe for full employment: interest on the National Debt paid via a wealth tax; subsidies to firms facing higher wages, in lieu of price increases; “a sufficiently high income tax”; price controls; expansion of the public sector and subsidies to private consumption. These are not technical adjustments, but a complete politicization of the economy.

In the end, it was full employment’s economic problems that killed it, precisely because the stagflationary crisis became a political crisis. For social democracy, the golden weather of the 1950s and 1960s was a kind of double-edged good luck, breeding complacency by making things all too easy. Leftist economists like Kalecki and Robinson had expected the public demand for full employment to fundamentally transform capitalism. Conservatives came to embrace it because it didn’t.

For two decades, the distributional effects of full employment were manageable. Rapid productivity growth allowed wages to grow well beyond the expectations of workers who still remembered the years of depression and war, without feeding inflation or eating into profits. Gradually, into the 1960s, expectations were raised. Kalecki’s predictions began to come true: fear of “the sack” faded and a generation blessed with economic security began to set its sights higher. And then, the slump. The oil shocks — part of a broader commodities boom — arrived with labor at the height of its confidence. But slower real growth meant there was no longer room for rapid wage rises.

If full employment really had been underpinned by the kinds of institutions thought necessary by Kalecki — or even the liberal Tinbergen — who knows? The distributional struggle might have been managed in other ways. It certainly would have been politicized, fought out for what it was, a class conflict, and that is ultimately why it couldn’t be left to the technocrats. In the United Kingdom at least, the socialist left urged the labor movement that it needed to supplement its economic power with political power, to launch a project for a more fundamental transformation, before a conservative project was launched upon them. Alas.

What Liberalism Can’t Deliver

Thatcherism and Reaganism came and went. Friedmanite monetarism failed on its own terms but fulfilled its historical mission of smashing the labor movement. It left behind the NAIRU, the reframing of unemployment as a technical problem shared by conservatives and liberals alike. Its pervasive message is not to set our sights too high, to accept a certain level of slack in the labor market as, if not “natural,” at least necessary for price stability.

The decades since have been pretty dismal for labor. In the United States the median real wage has barely risen in twenty-five years. Here, where employment was only briefly really “full” in the 1960s, the post-1970s have been marked by only intermittent visits to below 5 percent unemployment. Job growth has typically been slow in recoveries, and especially so since 2010. Only for a few years in the late 1990s could the labor market plausibly be called tight — and this was unsurprisingly also a brief period of real wage gains, declining poverty, and a slowdown (not a reversal) of the rise in income inequality.

In Jacobin, J. W. Mason has shown how central bankers in this period have openly and persistently used wage growth as a signal for tighter policy. They have not taken NAIRU estimates for granted. Alan Greenspan was prepared to let unemployment drop lower in the 1990s because wages seemed less responsive to rising demand for labor: “a major surprise to most analysts,” as he put it in 1997. “If labor demand continues to outpace sustainable increases in supply, the question is surely when, not whether, labor costs will escalate more rapidly.” But it was something to experiment with. A “heightened sense of job insecurity” had brought “subdued wage gains” and so made lower rates of unemployment sustainable.

It is against this background that we should think about the idea that the Federal Reserve has become a champion of full employment. Even before the crisis, current chair Janet Yellen was associated with the view that globalization and labor weakness had flattened the short-run Phillips curve, making high demand less likely to spark inflation. She has even argued that rising wages are critical to recovery, allowing households to pay down debt.

In September Fed board member Lael Brainard gave a speech on the “new normal” economic climate, noting that inflation has not responded as unemployment has fallen as low as 4.9 percent, and speculating that “estimates of the natural unemployment rate may move even lower . . . The case to tighten policy preemptively is less compelling.”

For Mike Konczal, this “new thinking at the Fed” is an example of the “new liberal economics.” Indeed, liberal voices in macroeconomics are at the height of their confidence; conservative warnings of high rates of structural unemployment and the inflationary consequences of loose monetary policy have been discredited. With monetary policy in “pushing on a string” territory, expansionary fiscal policy has returned to the economic mainstream.

But for the Left, it is important to remember that notions of a lower “natural rate of unemployment” are predicated on labor’s weakness. In spite of the low unemployment rate (for the United States), genuine full employment is some way off. Though the unemployment rate is back where it was in 2005, the labor force participation rate is more than three percentage points lower: some of that reflects demographic trends, but most is due to discouraged workers giving up the job search. The Bureau of Labor Statistics’s broader measure of unemployment, which includes workers who want work but have not been actively job-seeking as well as part-time workers who want full-time employment, has been stagnant at almost 10 percent for most of the past year, still two percentage points off its pre-crisis level.

As soon as genuine labor market tightness begins to raise expectations and bargaining power, we can be sure that the monetary hawks will be back in the ascendance. The Left needs to reach higher. Restoring labor’s economic power is the point of full employment — and at that point our interests diverge from the liberal technocrats.

This divergence is a good thing. Genuine, lasting full employment could be a massively popular aim for a revitalized left. It is hard to explain to the populace that a strong economy needs just the right amount of anxiety, but that has long been the position of the technocratic center. We should be calling liberals’ bluff here. Basic economic security for all — freedom from fear — hardly seems too much to ask in a wealthy, technologically advanced society. But liberalism cannot deliver it.

There are many ways to get there, and they all have the advantage of being desirable policies in themselves: public investment in education, health, infrastructure; green transformation; job guarantees; income transfers — the precise mix will depend on the movement’s preferences and the circumstances. (Two books — one by Robert Pollin, and one by Dean Baker and Jared Bernstein — have sketched plans for full employment in America.)

The postwar boom showed us that full employment is possible in a capitalist society, but only under fortunate conditions that are not themselves under policy control. Once the luck runs out, or if the luck is not there in the first place, full employment demands nothing less than a politicization of distribution, and of production itself. It is, in Gorz’s terms, a “non-reformist reform,” dysfunctional for the system. Such a reform sooner or later faces a crisis, and either the system spits it out or is reshaped around it. Once there, a whole new game begins.

Celebrate the new issue of Jacobin, The Party We Need,” with a discounted subscription.

If you like this article, please subscribe or donate.