This summarises a Sunday afternoon conversation with Roger Farmer on Twitter. Apologies for multiple drafts and no links: this is one- fingered on an iPhone from hospital.
Roger has a new microfounded macroeconomic theory that has the prediction that the unemployment rate is a random walk, and cointegrated with the stock market. He looks at the data and finds that the unemployment rate is a random walk and is cointegrated with the stock market.
Is the natural rate hypothesis, at the heart of mainstream modern macro dead, or, as Roger puts it, ‘well past its sell by date’? That’s the inference, at face value.
Another explanation, is that there is slow moving institutional change that drives both the stock market and the unemployment rate. Institutions related to benefits, employment protection legislation, taxes that affect labour demand and supply, unions and collective bargaining have all been in slow motion change, which over small samples may look non-stationary (or even be thus), and may have affected the labour share and the value of a claim on the remainder. These changes one would have expected too to have pushed around the equilibrium unemployment rate.
On this view while naive manifestations of the natural rate that exclude institutional change may be past its sell by date, those that embrace it may still be edible.
Of course at this point I have done no more than raise a counter conjecture to compare with Roger’s econometrics but it is at least worth investigating.
Roger has been urging large scale intervention in the stock market to raise equilibrium employment. In fact dovish natural rate believers can find common cause with this and related policies right now, although the cost -benefit analysis will differ somewhat. So even though until entirely disproven I would urge we keep some eggs in thematically rate basket, at present the policy prescriptions from the two perspectives are not in direct opposition.