What’s Different?

Go on over to this truly interesting post at a Canadian econblog:

Okun’s Law tells us that labour productivity, crudely measured as GDP/employment, and ignoring subtleties like hours worked and quality of labour, normally falls in a recession. Because the percentage fall in GDP will be two or three times as big as the percentage fall in employment. And it did fall in all the other countries. But in the US it didn’t fall at all. Labour productivity actually increased. GDP fell a little over 4%, peak to trough, and employment fell nearly 6%, so the GDP/employment ratio increased by over 1%.

The US is an even bigger puzzle if you think that business cycles are caused by productivity shocks. Sure, you could always argue that US firms and workers were expecting even bigger productivity growth, so when it actually came in at only 1%, that was a negative shock to productivity. But you would have to work hard to convince me that that’s plausible. And what were all the other countries expectations for productivity growth — chopped liver?

Why did US productivity increase during the recession? Why doesn’t your explanation also apply to the other 6 countries?

Why is the US an exception?

Over the last few months I’ve presented a host of explanations for why this might be including the consequences of high levels of corporate bureaucracy and demographics. I think that demographics remains the most likely answer.

Let me suggest a few other possible others. Is it possible that in an environment of nearly free and virtually instantaneous communication and highly portable means of production that Okun’s rule of thumb no longer holds? Is it possible that there’s something wrong with the way we’re calculating GDP and that the U. S. for reasons of scale, differences in the structure of our economy, or what have you reflects that while no other country does?

Are there others?

10 comments… add one
  • Perhaps we’re measuring productivity wrong?

    I’m a scientist, not an economist, so my understanding of how productivity is measured is far from complete, but is it possible that the methodology has become inaccurate because of changes in technology over the years?

  • Generally speaking productivity is output per labor hour. To calculate productivity wrong in this particular way you’d need to be seeing output as higher than it was or labor hours as lower than they were.

    It’s possible that both may be true. It’s essentially what I said in the post, just phrased another way.

  • Its more of a statistical relationship that anything derived from economic theory. Much like the unemployment/inflation trade off embodied in the Philips curve. The latter persisted long enough for models to be built around this statistical artifact and when it fell apart so did the models. Or to put it differently, Dave could very well be right, at least in a broad sense, that while a handy rule of thumb there is no reason it has to hold in all circumstances.

  • Even if it’s a rule of thumb, I think that the question of why it’s working everywhere in the OECD except for here is a good one.

  • steve Link

    Mish had claimed, about a year ago (?), that we do calculate our GDP differently. I would note that my employees showed up earlier and worked later w/o complaint. They also failed to ask for a raise for the first time during this recession. The consequences of job loss are harsher in the US. I wonder what corporate profits would look like if charted on those graphs?


  • I’d be focusing my detective skills on international transfer pricing, supply chain, and nature of layoff issues.

    -Who got laid off in disproportionate numbers? Line or staff?
    -How easy is it to layoff people in the US versus other OECD countries?
    -How much of final product value is determined by operations in the US which is actually the result of labor productivity in other countries due to US import accounting procedures?
    -If corporations want to capture profit overseas then it behooves them to inflate the value of American contributions to the product or service in order to minimize profit earned in America. By engaging in this tactic the value of US labor is over-counted and and the productivity per labor unit of contribution is boosted. Laying off employees doesn’t much affect this productivity measure because productivity, to the extent that it is boosted by transfer pricing decisions, is determined by labor.

  • PD Shaw Link

    Since I think the traditional 40-hour work week (or the 37.5 hour variant) is not necessarily a productive assumption in certain semi-skilled service jobs, I believe companies were able to use furloughs, mandatory vacation days and flex-time to erode the meaning of full-time employment. I don’t think the BLS data is fine-tuned enough to test this assumption, and it’s going to be job and sector specific anyway.

    Some OECD countries have lower work hour assumptions (Netherlands was 75% of the U.S. in 2004) and may not have any labor slack to give. Others might have stricter work rules that make it difficult to reduce hours to save money.

  • Sam Link

    -How easy is it to layoff people in the US versus other OECD countries?

    I think this is key, as well as which people get laid off – the least senior or the lowest performing.

  • john personna Link

    When GM makes 1/3 of its money overseas, a US GDP/employment ratio makes more sense than when GM makes 2/3 of its money overseas.

    They tell us that globalization isn’t new, or has ebbed and flowed over the centuries, but any of these national measures assume a stasis.

  • john personna Link

    (It was near an “industrial policy” for the US to pursue financialization as a response to globalization. Washington oriented toward it starting at least as early as the Clinton years. Perhaps earlier.)

Leave a Comment