The Empirical Evidence for the Brenner Hypothesis

In the inaugural editorial for Catalyst, Robert Brenner notes that

Since 1973, the economies of the advanced capitalist countries have performed ever more poorly. The growth of GDP, investment, productivity, employment, real wages, and real consumption have all experienced a historic deceleration, which has proceeded without interruption, decade by decade, business cycle by business cycle, to the present day.

The source of this loss of dynamism has been the deep fall, and failure to recover, of the economy-wide rate of profit, a process that took place mainly from the late 1960s to the early 1980s and derived largely from the relentless buildup of overcapacity across the global manufacturing sector. 

Brenner identified the ultimate cause of the loss of Western economic dynamism as the relentless buildup of overcapacity across global manufacturing. It is hard to find numbers on capacity utilization rates at the global level. But we can instrument that with US manufacturing capacity utilization rate since utilization is unlikely to be high in the world’s preeminent advanced manufacturing power if there is significant global slack. In other words, we can proxy global overcapacity by US capacity utilization.


Figure 1: Capacity utilization rate of US manufacturing. (Source: Haver Analytics.)

Figure 1 displays the capacity utilization rate of US manufacturing. We see that there has been a secular rise in overcapacity since the 1960s. However, Brenner is wrong that this buildup of overcapacity “took place mainly from the late 1960s to the early 1980s.” To the contrary, after a brief respite in the 1990s, the buildup gained renewed strength in the 2000s and the 2010s. In fact, peak capacity utilization rates in the 2000s failed to match even the lows reached in the 1990s, and the post-Great Recession recovery is even more tepid.

Predictably, ever-intensifying overcapacity in the manufacturing has generated a secular decline in the profit rate of this sector. Figure 2 displays the profit rate in US manufacturing. For this sector, Brenner’s right. Profit rates fell relentlessly through to 1985 and failed to recover thereafter. The largest declines occurred in the mid-1950s, late-1960s, and early-1980s. We know that the last was the consequence of the super-strong dollar that attended Volcker’s sky-high interest rates. The first was perhaps the understandable consequence of the recovery of European manufacturing firms that brought down highly elevated rates of profit for US firms who briefly had the global market to themselves in the late-1940s and early-1950s. The second, the sharp decline of the late-1960s, can be traced to the Japanese onslaught as suggested by the weight of cross-sectional evidence: the decline in profit rates was concentrated in the six sectors most exposed to Japanese competition.


Figure 2. Profit rate of US manufacturing defined as the ratio of profits after capital adjustment to net stock of fixed assets. The break reflects the fact that BEA has three different series for 1948-1987, 1988-2000, and 1998-2017. I display all three but don’t connect them because they are strictly speaking not comparable.

But manufacturing now accounts for only 12 percent of value-added in the United States. Is it true, as Brenner claims, that there was a “deep fall, and failure to recover, of the economy-wide rate of profit”? No, Brenner’s wrong. Figure 3 displays the rate of profit of all US corporations in 1948-2017.


Figure 3. US corporate profit rate. (Source: BEA)

Contrary to Brenner’s claim, profit rates fell straight through 1966-1982, but then recovered much of the lost ground in the last three business cycles. How were US corporations able to restore profitability despite persistent industrial overcapacity? Simple. They confiscated the bulk of the growth in labor productivity. Figure 4 displays output per hour in US nonfinancial corporations and real labor compensation per hour.


Figure 4. Labor compensation and labor productivity in the US corporate sector.

A structural break in the productivity and labor compensation series can be detected in the 1970s. In 1947-1975, output per hour grew 125 percent, while labor compensation per hour grew 90 percent. Since 1975, output per hour has doubled, while labor compensation per hour has grown by only 30 percent. Equivalently, the annual rate of productivity growth fell from 2.7 percent in 1947-1975 to 1.8 percent in 1975-2016. Meanwhile, the average growth rate of real wages fell from a robust 2.2 percent in the first period to a measly 0.7 percent in the second.

All in all, in the neoliberal era, labor secured only 30 percent of the gains in productivity, whereas in the postwar New Deal-era, labor secured fully 72 percent of the pie. This is the secret sauce of restored corporate profitability despite persistent overcapacity. This is a major achievement of the neoliberal counterrevolution which must be understood as a political movement aimed at the restoration of profit rates and upper class wealth and power.

This, of course, doesn’t invalidate Brenner’s key insight that the loss of Western dynamism and the attendant disappearance of broad-based growth must be traced to persistent systemwide overcapacity. For while the neoliberal assault may have restored the fortunes of US corporations and the upper classes, the absence of dynamism is all too real. As I noted earlier on Facebook today in the context of the decline of Europe’s social democrats,

What the social democrats have failed to do is to articulate is a credible solution to the principal problem of contemporary Western political economy: the end of broad-based growth. It is this failure hidden in plain sight that has wrecked the center and empowered political forces opposed to any further pursuit of the open society on both the left and the right.

It’s also amply clear why the center-left has everywhere paid the price. For the center-left abandoned the pursuit of protection against market discipline when it folded in the face of the neoliberal ascendancy. Having surrendered that pursuit, the center-left failed to articulate how its vision differed from that of the center-right, which never championed protection against market discipline anyway and was therefore always the natural claimant of the neoliberal center.

But the fundamental problem goes deeper. For it is not at all clear that a solution to the core problem in fact exists. None certainly exists within the parameters of contemporary governmentality; that is to say, the neoliberal toolkit. And no credible solution outside the toolkit has so far been proposed. Even in France “the gist of Macron’s labor market reform” is to “push back the state and facilitate growth, and aim to reduce unemployment by making it easier to hire and fire people.” Good luck with that.

The contemporary impasse is a historical echo of the 1970s when Keynesianism—the preferred technique of macroeconomic management in the era of social market democracy—failed in the face of the stagflation crisis (which was itself blamed on excessive entitlement claims generated by social market democracy). That impasse led to the neoliberal counterrevolution. How and when we will come out of this one is the most important question of contemporary Western political economy.

Bonus charts. 


Figure 5. Capital’s portion of the productivity pie.

The bottom chart in Figure 5 shows capital’s portion of the “pie” of labor productivity growth (roughly growth in output per hour). It is a function of both labor productivity growth per se (dotted line in top chart) and the bargaining solution between labor and capital (who gets how much of that pie). The solid line in the top chart is labor’s portion of the pie (roughly growth in real wages per hour). [See Technical Notes at the end.]

I think the following three observations are called for. (1) There are brief respites in the early-1980s and the mid-1990s where productivity growth is decent if still modest and wage growth is suppressed. But there is a veritable bonanza in the 2000s. Not only is the respite from Brenner’s crisis strongest in the neoliberal era (ie, productivity growth briefly approaches postwar levels), wage growth is nonexistent. (2) Bites taken by capital are somewhat countercyclical. They fall to their lows in booms (late-1960s, late-1980s, late-1990s, mid-2000s), and are high early in the recoveries (early-1960s, early-1980s, mid-1990s, early-2000s). (3) Capital seems to be suffering under postcrisis stagnation even more than labor. It has been downhill since 2010. And 2015 was the worst year for capital since 1957.


Figure 6. Productivity drives both wages and unemployment.

Figure 6 marshals the empirical evidence for the causal diagram “wages ← productivity → unemployment”. That is, productivity drives both wages and unemployment. The interpretation being that times are good when the economy is experiencing positive productivity shocks. Growth in the pie allows wages to rise. And since output grows with productivity, unemployment ought to fall. The monocausal model has clear implications for the conditional probabilities. Specifically, if negative unemployment shocks cause wage growth independent of productivity, then the bottom-left would have a distinct pattern and a high R^2. On the other hand, if positive productivity shocks cause wage growth independent of the unemployment rate, then the bottom-right would have the observed pattern. In short, the empirical correlations are what you would expect given the monocausal model that productivity drives both unemployment and wages.

Here’s another version of Figure 5.


Figure 7. American pie.

Technical Notes: 

Instead of looking at annual growth rates which are very noisy for both output per hour and labor compensation per hour, we stochastically detrend the two series by deducting from them their 4-year trailing averages. These detrended series are what are referred to as “shocks” in the charts above. They can be interpreted roughly as smoothed growth rates. When we write “X orthogonal to Y”, it is shorthand for contemporaneous variation in X orthogonal to variation in Y in the usual sense, ie we project X onto Y and obtain the OLS residuals.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s