@Alon @SteveRoth @jamisonnbishop @ryanlcooper In a world where employers perceive high costs to finding employees and so hold onto employees through the business cycle, measured productivity declines overall. during periods of lower demand, labor hours are expended for the nontransactional benefit of maintaining relationships, which does not appear in GDP, reducing apparent productivity. 1/

@Alon @SteveRoth @jamisonnbishop @ryanlcooper In a world where employers fire very freely, we see the morally counterintuitive result that productivity rises in recessions. Businesses fire employees whose hours contribute least to revenue generating transactions, raising the average of revenue per hour. Plus they can work those they retain harder, as "the sack" becomes a very credible threat. 2/

@Alon @SteveRoth @jamisonnbishop @ryanlcooper Treating productivity as we measure it as an unmitigated virtue is perilous. Yes, we want to maximize welfare generated per hour worked. But productivity doesn't measure welfare, it measures GDP, the macro analog of revenue. There are obvious wedges between welfare and revenue when firing someone doesn't much hit revenue but condemns a person to penury. /fin

@Alon @SteveRoth @jamisonnbishop @ryanlcooper (The better matching you describe in boom periods could counter to some degree my claim of lower measured productivity over the cycle, but to the degree there are matching gains to be had from job-switching, there might be tensions between the worker scarcity and so security “labor hoarding” yields, encouraging switching, and the relationship-building attached to that, discouraging it even in booms.)