Explaining (away) the productivity miracle

There’s been a lengthy debate, in the blogworld and elsewhere about the causes of recent growth in US labor productivity, and the associated fact that reasonably good output growth has produced little or no employment growth. Brad de Long, in particular has discussed this at length. My preferred explanation, which I put forward here is that

Beginning with productivity, it’s only labour productivity that’s grown rapidly and seemingly anomalously. Capital productivity has declined markedly, as has multifactor productivity (a weighted average of capital and labour productivity) In part this reflects the economics of embodied technical change – as computing power has become cheaper it has been applied more intensively. But there’s also a big hangover effect from the bubble and bust, when crazy signals from capital markets led lots of firms to undertake unprofitable investments. Once some semblance of reality returns, the natural response is to cut back and it’s much easier to sack the least productive workers than to reduce capital stock. So labour productivity rises fast, but output growth is weak.

This NYT piece gives lots evidence, admittedly msotly anecdotal, to back up my analysis. The key statement is the

the slow process of working through a glut of boom-era investment that continues to litter the economy with underused factories [ is weighing on job creation]

Along with lots of case studies of soap factories is the point that

Not since the severe recession of the early 1980’s has capacity use in manufacturing stayed so low for so long, government data show. Production as a percentage of total capacity fell precipitously in the aftermath of the last recession, which ended in 2001, and 23 months into the recovery, the upturn has still not come. On average, manufacturers are using less than 73 percent of their capacity.

If multifactor productivity were really rising we’d expect to see new investment and hiring as US producers displaced competing imports, especially with the dollar depreciating. But there’s no sign of this so far.

Updated Brad de Long has responded to the same article with a restatement of his main argument

Louis Uchitelle frames the issue the wrong way around: there is no such thing as “overinvestment,” there is only too little aggregate demand. If you think that there is “overinvestment,” dropping the interest rate will cure you of that belief. At least, it will until you hit a liquidity trap… But it looks like that is no longer a dangerous possibility.

There’s a disconnect here between the macro arguments Brad is making and the micro arguments I’m making. My intuition is that the size of the US CAD indicates that deficient aggregate demand is not the problem. But I need to think more about this.

7 thoughts on “Explaining (away) the productivity miracle

  1. Let’s get the pedantry out of the way first. I think you might want to make these italicised changes in your opening: “There’s been a lengthy debate, in the blogworld and elsewhere about the causes of recent growth in US labor productivity, and the associated fact that reasonably good output growth has produced little or no employment growth.”

    Fixed, thanks (JQ)

    While I think you are describing the productivity changes accurately, I don’t think those are causes so much as symptoms. I’ve already expressed my own view, that there is an externality causing a bias against labour use (and which flows through to more capital use, “exporting jobs”, etc.). the externality is from Vagrancy Costs (for which Social Security costs are a proxy in many countries), and I believe that Kim Swales’ approach is effectively a Pigovian solution. I prefer a Pigovian solution here partly because we already have what it needs in place and partly because the Coasian solutions are slow and impractical in the short to medium term (distributism) or undesirable on other grounds (slavery, euthanasia).

    Ah well, maybe one day the message will fall on fertile ground. It did for Cato the Censor.

  2. I don’t think we know why labor productivity has followed the pattern that it has in the late American recession and stagnant recovery–it is a great micro mystery. But it does seem very clear that the economy could use another couple of shots of aggregate demand…

  3. I certainly agree that the fiscal stimulus that’s been given could be retargeted to make it more effective. And if that were done, there might be room for more stimulus in aggregate.

    But I don’t think this offsets the point that, in microeconomic terms, there can be “overinvestment” in the sense of lots of unprofitable investments that will take years to unwind.

  4. It’s not all that hard to reconcile these positions.

    An aggregate demand problem means there is too little spending, at present. Overcapacity, at present, in one sector (tech & telecomms) means there was too much investment in that sector, in the past.

    The problem is fixed by having more consumption, more investment (but not in the sector with overcapacity) and more exports.

    The CAD might get bigger or smaller depending on which of these compnents of spending you get.

  5. I think de Long’s dispute with Uchitelle can be clarified by making two distinctions.

    First, between ex-ante and ex-post ‘excess capacity’: There is such a thing as an optimal capital stock, beyond which steady-state consumption actually falls as that stock increases. It is irrational to accumulate beyond that point. But if the capital is already in place it’s a sunk cost and it can’t be rational from a social point of view not to use it, provided there is some substitutability between capital and labour. If the technology is close to being Leontief then capacity utilisation is limited by available labour, but this is not the case here: indeed, growing unemployment is the point of departure for the whole discussion. Therefore a failure to use it implies a deficiency of aggregate demand, and I think this is all de Long is saying.

    Second, between the social viewpoint and the firm’s: Proctor and Gamble do not control aggregate demand, so it makes sense for them and the other individual firms to treat the level of demand as given. From their point of view there is surplus capacity.

    John’s own argument about labour productivity growth is quite convincing, and like Milton and Derider I can’t see any incompatibility with deficient demand as an explanation for unemployment.

    On the other hand, John’s objection in relation to the current account puzzles me. Isn’t the simultaneous presence of unemployment and external deficit, the very essence of the notion of an over-valued currency?

    To tie the threads of this sermon (my longest comment ever, I fear) together: The excessive capital stock is obviously a reason for the decline in aggregate spending: gross fixed capital formation has declined through most of 2001-02 after nine years of continuous growth. In principle measures could be taken to increase consumption as a countermeasure, but this would be unwise with the current account in its present state. So the only solution is an increase in net exports, and this brings us back to the mystery of the over-valued dollar.

Comments are closed.