This is just an idea that's been floating uselessly in my mind for some time. I have never been able to apply it. But maybe someone out there could.
This isn't an earth-shattering post. Maybe someone else has already thought of the same idea. Or even applied it, and I just haven't heard about it. But I might as well throw it out there, just in case they haven't, and someone reading this might figure out how to pick up the ball and run with it.
There are two main theories of business cycle fluctuations:
1. Productivity shocks cause fluctuations in output and employment. (Real Business Cycle theory.)
2. Aggregate Demand shocks cause fluctuations in output and employment. (Keynesians and Monetarists etc.)
We've know for a long time that in a recession, output generally falls by a greater percentage than does employment (both relative to trend). We call this empirical regularity "Okun's Law".
Real Business Cycle theorists say that this is because there has been a negative productivity shock, because productivity is output per unit of input, and that same negative productivity shock is what caused the recession.
Aggregate Demand theorists say that a fall in Aggregate Demand is what caused the recession, but firms don't like to lay off workers, so they hoard some workers even if they don't really need them to produce goods right now, which is why measured productivity falls.
How can we test between the two theories? Was there really a productivity shock? Or did measured productivity appear to fall because of labour hoarding?
Suppose we found an industry where we knew for sure that productivity really never changed. Because that industry was just so very traditional that we knew that workers with the exact same skills and using the exact same tools and raw materials had been producing the exact same goods in the exact same way for the last century or two.
And suppose we had very good data on outputs and inputs (like hours worked) in that traditional industry.
And suppose we found exactly the same sort of measured productivity shocks in that traditional industry that we found in the aggregate data, and that the two series were highly correlated and about the same size.
We would know for sure (by assumption) that the measured productivity shocks in our traditional industry had to be labour hoarding rather than real productivity shocks.
And that would strongly suggest that the measured shocks in aggregate productivity also had to be labour hoarding rather than real productivity shocks.
Is there any such industry?
Just throwing this out there.