Struck by Mankiw’s Struck-ness

I am struck by the fact that the prominent Harvard economist N. Gregory Mankiw, former Chief Economic Adviser under George W. Bush,  is struck by the following graph (due to the equally prominent John Taylor).


So, when unemployment is high (in the US) investment, as a share of output, is depressed. And when unemployment is low, investment is buoyant. Most people will surely find that blindingly obvious. As Mankiw himself notes, the link runs both ways.

The question is: why does this come as a surprise to the likes of Mankiw and Taylor?

The answer, I think, is that they tend to see things through the lens of price theory at the micro level.  That is individuals making decisions under the prime influence of relative prices. The proximate explanation for high unemployment is then that labour is ‘too expensive’ (relative to what it produces). And a key mechanism, at least in large, more or less closed economies, is capital-labour substitution. If labour is expensive firms will use more capital (which in relative terms is cheap) and less labour to produce the goods demanded, displacing labour from the production process and driving up unemployment. (This is one reason why the demand curve for labour – the quantity of labour demanded at a given price (i.e. wage) – is drawn as downward-sloping. In a small open economy the negative impact on international competitiveness comes on top of this effect.)

Yet if this mechanism were really an important part of how our economies work, then you would expect to see investment high when unemployment is high (because of all those purchases of labour-saving capital).

So if you see the world in terms of allocation decisions driven by relative prices then you are struck by the graph. If you see things through a Keynesian lens, in which aggregate quantities are what really drive economic outcomes, you can only be struck by their struck-ness.