There’s a controversy aflame in the left–liberal blogosphere around a revelation in Ron Suskind’s new (and apparently error-riddled) book, Confidence Men. (Brad DeLong has the page.) Suskind reports on tense high-level meetings within the Obama administration as it became clear that the StimPak wasn’t really working. Unemployment was drifting higher, and the Keynesian faction—Christina Romer, then chair of the Council of Economic Advisors, and later Lawrence Summers, then resident wise man—was calling for more stimulus. Obama said no. It was politically impossible, but Obama also argued that the productivity revolution has made workers obsolete. Against that, a few hundred bill in further stimulus—which would be DOA in Congress in any case—would do little.

Romer and Summers (who, it hurts to say, has been looking pretty good recently) argued that productivity need not create unemployment. After Obama supported an orthodox briefing by former OMB head (and inexplicable sexpot to some) Peter Orzag (who also recently said that we need less democracy to have a sensible fiscal policy), Romer objected, urging a more expansive fiscal policy. Obama cut her down in what Suskind calls “an uncharacteristic tirade.” A few months later, with unemployment now over 10%, Summers essentially made Romer’s earlier argument, and Obama listened “respectfully.” After the meeting, Romer said to Summers, “Larry, I don’t think I’ve ever liked you so much.” He told her that the feeling would pass—but then noted that Obama had been a lot more “generous” with him than her. When you’re on the wrong side of Larry Summers on feminist issues, you’ve got a serious problem.

Matt Yglesias assures us that his conversations with administration officials report that Summers and Romer won the productivity=unemployment argument eventually, and convinced him that the problem was low demand, not high productivity. But Yglesias himself also reports on an interview that Obama gave in June in which he blamed ATMs for high unemployment. And airport self-check-ins too. Stuff like that, that any globetrotter sees. So apparently Obama still believes that there’s only so much gov can do because all these business geniuses are using machines so cleverly.

Respectable Keynesians argue that the problem is demand, which is a cyclical argument more or less, while the productivity story is more structural. But maybe both sides are right in a way. Productivity is way up but wages are flat. Workers have seen little of the productivity revolution.

The canonical tech-driven productivity acceleration took hold around 1995. From then until the cyclical peak year of 2007, productivity growth averaged 2.6% a year—4.3% in manufacturing. Compensation, which includes fringe benefits (see graphic caption), rose 1.7% a year—and 1.5% in manufacturing. Direct pay overall rose 0.9% a year—0.3% for factory toilers. During the recession, productivity growth slowed, employment collapsed, and wages rose modestly. In the “recovery years” of 2010 and 2011, productivity growth—this time not from a tech revolution but from sweating a shrunken workforce harder—has resumed growing at a 2.6% pace (2.1% is the very long-term average, by the way). But both compensation and direct pay have fallen by an average of 0.3% a year.

If workers were paid better, there wouldn’t be so much of a demand problem, would there? But then we stumble upon a contradiction: the entire recovery in corporate profitability that began in 1982 came from squeezing wages and workers. The trajectory of profit expansion is very uplifting—returns are roughly double what they were in the early 1980s:

After the profitability collapse of the 1970s, things began to turn around as the Volcker recession and the PATCO firing transformed capital–labor relations. Corporations enjoyed dramatically improving returns from the early 1980s through the late 1990s. Profitability took a major hit in the early 2000s recession, with the bursting of the tech bubble, but it recovered, only to stumble again in the 2007–2009 recession. But it’s staged a remarkable recovery. In the second quarter, nonfinancial firms were more profitable than they were at any time since 1997, in the frenzy of the dot.com moment.

For most of the last 30 years, much of the working class was able to borrow what it wasn’t earning to make ends meet. Household debt rose strongly:

Mortgage debt rose from 41% of after-tax income in 1983 to 100% at the 2007 peak. It’s since come down hard, and the economy and political mood show it. Consumer credit—auto loans and credit cards and such—went from 16% to 25%. It’s come down too.

So here’s the demand problem: there’s no longer an endless supply of easy credit to make up what’s not in the paycheck. The greatest product of the productivity revolution is the production of profits, which has enabled a vast upward distribution of income and wealth. No one really wants to touch that one. So what’s a capitalist order under no serious political challenge to do except dither and squabble?