I try not to write too much about either the slowdown in productivity growth or Donald Trump. In both cases, the market is already pretty saturated, the topic isn’t very pleasant, and no one really knows what to do about either problem.
But since I at least kind of understand the productivity story, and haven’t revisited it on this page for a while, let’s get an update.
Productivity is output per hour and its growth is the primary way economies generate higher living standards. That’s intuitive: if we can produce more of the things people want and need in fewer hours, in theory, we can have more time to enjoy the benefits of growth. Right off, you see one problem: who’s “we?” In the age of inequality, the gap between productivity growth and the typical workers’ paycheck is wider than ever.
But today, I’d like to explore a different problem: the significant downshift in productivity growth. Yes, in the presence of higher inequality, we can’t count on faster productivity growth to raise living standards. But it is a necessary condition.
The basic story of the last 40 years or so goes like this. Productivity settled into a groove of about 1.6 percent per year from the mid-1970s through the mid-90s. At that point, while some of it looked wasteful and bubbly (fiber optic to nowhere), many hypothesized that the investments in computers and the Internet ought to show up as faster productivity growth.
Productivity growth did significantly accelerate between the mid-90s and the mid-2000s, to almost 3 percent annually. Then it unexpectedly slowed, seriously downshifting to about less than 1 percent per year. The average annual growth rate since 2012 is a lamentable 0.6 percent.
When it comes to explaining these changes, economists tend to be uncharacteristically humble, as we should be (about this and much else, IMHO). But we are not totally at a loss, and we’ve usefully broken down productivity growth into three parts: capital per worker, labor quality, and what’s left over. We call the last bit “total factor productivity” (TFP) because “what’s left over” sounds much less smart.
Even so, that’s not as abstract as it sounds. The diffusion of productive machinery and the skill and experience of the labor force working with that machinery should reasonably influence how efficiently we turn our inputs into the goods and services we want and need. Add the fudge factor — which could be more innovative ways in which we’re using those inputs — and you’ve got productivity.
Of course, legitimate abstractions abound, especially in measuring service-sector productivity, or the output from technology, like free apps, wifi and so on. In fact, as the economy has shifted from easier- to harder-to-measure sectors and processes, some have argued that we’re just failing to measure elusive contributions. Suffice it to say, extensive research says no. If anything, I suspect we’re probably doing a slightly better job measuring inputs and outputs. If so, that implies the recent slowdown in productivity growth might even be a bit sharper than the official stats are telling us (see here and here for a deeper dive into these measurement issues).
I’m afraid the slowdown is real (it’s wrestling that’s fake). And as I showed in my last doleful post on the topic, it’s happening in other advanced economies as well.
I think there are 5 reasons: slower growth of capital per worker, slower TFP, capital misallocation, the absence of full employment, and dysfunctional government (labor quality has been pretty constant, so it isn’t much implicated in the slowdown).
A recent analysis by researchers at Goldman Sachs made an important point regarding slower “capital deepening,” a major reason for the slowdown. While business investment has been positive, albeit tepid, in this recovery, job growth has been very strong. As GS puts it, the fact that employment growth has been so much more “procyclical than the growth rate of the capital stock” has slowed capital deepening and thus productivity.
Regarding the fudge factor, a.k.a. TFP, government statisticians tell us it contributed an average percentage point less per year to productivity growth in 2007-2014 compared to 2000-07. That’s a huge loss. Productivity optimists believe that history shows this series will “mean revert,” i.e., eventually tick back up to a less depressed pace.
You ask me, “mean reversion” is a pretty thin reed. I’ve got three better ideas.
First, as I detail here, there’s The Big Short problem: we have been misallocating capital to non-productive finance. The fact that investment capital is and has been cheap for a long while won’t help us boost productivity growth if it flows to “innovative” finance versus actual productivity enhancing innovations.
Second, we are failing to tap a full employment productivity multiplier (FEPM). Among the pantheon of wrong-headed economic theories is the one that says: firms failing to operate at the edge of their productivity potential will be competed out of business by more productive firms. Unfortunately, in slack labor and credit markets, inefficient firms can handily maintain profit margins by squeezing workers and rolling over cheap loans. At full employment, workers have more bargaining clout, labor costs go up, and inefficiencies become more costly (Josh Bivens agrees and offers some evidence.)
Third, analysis by Barry Eichengreen et al suggests that dysfunctional government eventually grinds down productivity growth. This strikes me as intuitive: an $18 trillion economy requires a government that can efficiently diagnose problems and prescribe solutions in areas of climate, infrastructure, education, innovation, social insurance, poverty and more. Our government, on the other hand, tends to engage in aimless votes to defund Obamacare and shutter the EPA and IRS.
So, how can we better allocate capital, move toward full employment and restore functional government?
Especially given the unreliability of current and recent private investment, public investment is a natural. A deep infrastructure dive — in transportation, medical research and repairing our public schools — would be a big twofer, both on the productivity and full employment fronts. And by tightening the job market, there’d be positive feedback impact from the FEPM.
Given the combination of low interest rates and the big bang for the buck from this sort of investment right now, I wouldn’t break a sweat if we added such investment to the deficit. But here’s another idea with multiple benefits in this space: pay for these productivity enhancing investments with a small tax on financial transactions. That would both raise revenues needed for public investment and raise the cost of non-productive, “noise” trading.
As best I can, I cover the path back to functional democracy in the final chapter of my latest book, yours for a mouse click. The key is recognizing dysfunction as a lethal, ultimately self-defeating but highly effective conservative strategy.
OK, in good Buddhist fashion, I think I’ve now tied together pretty much every single thing I think about, so I’d better stop. And we didn’t even get into the Trump problem at all. But like I said, that’s something I can’t bear to think about.