Arnold Kling has a new post discussing the issue of wages and productivity. He argues that productivity is increasingly difficult to measure:

We are not in a GDP factory. As the share of GDP devoted to health care and education goes up and the share devoted to manufacturing goes down, we are giving more weight to a sector where real output and the quality of labor input are extremely difficult to measure.

This is also my view if GDP means RGDP (which, in context, it clearly does.) Indeed Kling and I are probably out on the extreme, in terms of being especially skeptical (relative to other economists) of the usefulness of measures of RGDP, real productivity, the price level, etc.

Over at Econlog, I argue that while the problem of measuring real productivity is very real, it’s largely unrelated to the so-called “wage decoupling issue—which is mostly about the gap between nominal wages and nominal productivity. Fortunately, errors in measuring real productivity have no impact on measures of nominal productivity.

So while I entirely agree with the Kling quotation above, I also strongly believe the following to be true (please read carefully):

We are in an NGDP factory. As the share of NGDP devoted to health care and education goes up and the share devoted to manufacturing goes down, we are giving more weight to a sector where nominal output and the quantity of labor input are relatively easy to measure.

So why is this second claim so different from the Kling quotation? Because RGDP and NGDP are radically different concepts, almost unrelated. Thus we can say with some confidence that the nominal health care industry has expanded from (say) $36 billion to $3.6 trillion since the mid-1960s, but I have absolutely no idea how much real health care output has grown, as I don’t even know what the output of the health care industry is. What are they trying to produce? (Recall that Robin Hanson says that health care is not about health.)

And even if health care is about health, how much of the improvement in health is due to health care, and how much is due to less smoking, better nutrition, better sanitation, etc.

In many ways, NGDP is not a very interesting variable (think Zimbabwe), except when it shows short-run volatility. In that case, it destabilizes labor and financial markets, because there are lots of nominal wage and debt contracts. In those situations, the “NGDP factory” is a useful concept, even though NGDP is also measured imperfectly. But NGDP is at least an order of magnitude more clearly defined and more easily measured than RGDP.

I do occasionally refer to RGDP, as despite its flaws it tells us something about business cycles and international comparisons. A sudden drop in RGDP usually indicates a recession, as all those imponderables associated with measuring “real health care” don’t change much from one year to the next. Think “law of large numbers in BEA errors”. The BEA bureaucrats are making similar errors, one year after another. Thus it’s a problem if measured RGDP suddenly falls by 5%, despite the many flaws in RGDP data.

As far as international comparisons, a country with a $50,000 per capita GDP is not necessarily richer than one with a $45,000 per capita GDP, but it is almost certainly richer than a country with a $5000 per capita GDP. So RGDP has some value if used with care.

Most importantly, don’t ask any statistic to do more than it can.