Tyler Cowen has some comments on the GDP-Less recovery

I don’t myself see an additional short-run fall in productivity (I don’t much trust the short-run statistics in any case), though of course I have been a productivity pessimist more generally.

First and foremost, I see the very latest data as evidence for the Garett Jones hypothesis.  Employers are going back to the idea of investing in workers who build up the future of the company, but who may not produce much output now.

Second, higher exports and moderating health care costs (the latter over the last two years) mean that “real gdp” is higher than traditionally measured gdp; see for instance Matt’s remarks.  This supports Michael Mandel’s view about the importance of offshoring and, presumably, reshoring, to the extent that is going on.  In general we undermeasure the gdp gains of successful export nations, because their outputs tend to have lower percentages of rent-seeking expenditures and more real stuff of value.

There are a couple of issues here but key is terminology.

Now ultimately it doesn’t matter which way this confusion is settled so long as it is settled but I am going to give a strong initial push for my way. If its just un-grokable then I will demure.

We don’t under-measure GDP unless we are unfaithful to our methodological procedures. That’s because GDP is that thing which we measure using our methodological procedures.

There is no “real” GDP. There is no truth of the matter beyond the BEA report.

Moreover, the things that we care about are grounded in people’s emotions. A vase is valuable because it is valuable to someone one. There is nothing intrinsic about the vase in which there is value.

Thus you are entering down the road of massive confusion and miscommunication if you start talking about how statistics are showing us “the real GDP” or how “productivity gains are lying to us”

Just take the statistics for what they are. As long as the methodology is consistent with the published methodological standards that’s it.

Then say, I think this important feature is not captured in the stats.

Or say, be sure to remember that productivity is a residual .It’s not a “measurement” in the sense that it is the independent product of some counting procedure.

This has certain practical implications.

One is that the stories I often here about employers panicking and shedding a bunch of employees and now hiring them back to doesn’t seem to correspond to the dynamics of the recession. In part because those stories are so “office work centric” but office employment doesn’t really look strange in anyway.

Here is professional and business employment versus retail sales

FRED Graph

and indeed if I take out temps the response to the recession flattens

FRED Graph

So, maybe I am misunderstanding something but there doesn’t seem to be much story to tell. Businesses seem to respond to business conditions and try to do so by hiring less temps or more temps when they can.

If you look at manufacturing – one of the hardest hit and slowest growing sectors employment-wise then you see this

FRED Graph

Which might look odd until you look at growth rates

FRED Graph

Construction employment is the same.

The only potential oddity is that the residual is not behaving as folks expect

FRED Graph

But, then if you note, look the residual doesn’t actually measure the productivity of anyone it’s just part of an adding up constraint and it turns out to be fairly sensitive to import mix changes.

Then you say, oh okay, and we cruise right along.


I can’t help but mention that there is a truth of the matter with regards to inflation or at least more truth. That’s because inflation does allege to measure something particular. However, that truth is grounded in the truth of Nominal GDP measures.

So to put it in Scott Sumner-esque terms. Some Nominal GDP would have actualized if there had been no change in currency stock over say 50 years. That nominal GDP did not actualize and so there is a thing to be estimated. Whether or not we can estimate it is another story.