Thursday, February 4, 2010

Productivity is up 5.1%, unit labor costs down 2.8% in 2009

Today, the Bureau of Labor Statistics reported that productivity in the U.S. was up 5.1% for 2009, the largest increase since 2002. Meanwhile unit labor costs are falling rapidly, since wage gains have been muted.

I just met recently with a business leader who used the downturn to layoff 10% of his workforce. He said he did not lose 10% of his productive capacity when he did this, since he kept the best workers. Efficiencies, driven by process reorganization and new technologies, make it possible to run leaner than before.

Throughout the economy, many firms are finding that they can increase output without hiring more workers. This portends more strong productivity growth in 2010, but also a relatively jobless recovery.


  1. Makes you think how the costs of worker retraining, relocation, job search etc should be allocated when studying the productivity impacts of technology.

    If only the short-run and internal costs of technology implementation are considered when studying the impact of technology on firms, isn't that ignoring the fact that these other costs are borne by society (through unemployment benefits, broken communities/networks, etc) and ultimately by firms, as they pay higher payroll taxes, worker retraining etc? So, ultimately higher productivity may be offset by other difficult-to-calculate costs

  2. @harmindersingh: I agree. There are a lot of unmeasured costs (and benefits) from technology implementation. When firms squeeze costs by shifting the expenses to other firms, to individuals or to the government, then that's not a true productivity improvement.

    In the longer run, I suspect that many of the externalities are positive, however, as others learn from successful productivity improvements, imitating and building on these successes.

  3. It would be worth doing a study to test this hypothesis, now that we've had computing in businesses for ~40 years. Teasing out the effects would be quite difficult but the outcomes would help resolve the IT capital vs. labor replacement problem. That info would be especially useful for developing countries