In our model, people advance along at least locally optimal career paths in expansions, and then have to find a new one in recessions. So you’d expect job tenure, marked in green, to reflect the business cycle – people accumulate it during expansions and lose it in recessions – and that’s precisely what we see. In 1996-2000, when unemployment dropped sharply, it was a strongly negative contributor to wage growth. After that, it began to be a positive contribution as the new hires progressively accumulated tenure and advanced along their career paths. We also see a bit of this after the .com crash. However, it didn’t become a big negative item after the great financial crisis, perhaps because unemployment didn’t rise as much as expected.
The effect of change in qualifications has been quite surprising; it was negative for most of the boom, and then very positive immediately post-crisis.
From 1999 to 2007, workers changing between occupations seems to have been a significant contributor to wage growth (about +0.2% a year). Between 1996 and 2002, workers changing between industries was a positive contribution, but it then swung negative between 2002 and 2006, before becoming positive again in 2007.
During the great financial crisis, it was significantly negative, and it then became positive in the recovery. Since then, it’s disappeared as a factor. Change between occupations, however, was strongly positive in the crisis, erratic and noisy in the recovery, and since Q1 2013, has become very strongly negative. So has the effect of job tenure. At the moment, the combination of tenure and occupational change accounts for -0.75 percentage points of wage growth. The strong negative tenure effect is comparable to that in the late 90s expansion, implying significant net hiring. The occupational change effect is, however, unprecedentedly awful, and it is increasing.
This is consistent with the perverse selection I proposed in the original post. The big difference between now and the 90s experience, though, is that the occupational shift effect is much bigger.
Which is also consistent with making Jobseekers’ Allowance claimants stand around Finsbury Park station wearing a hi-viz vest to no particular purpose.