REPLY
| Published date | 01 May 1989 |
| Author | Blanchflower, Oswald |
| Date | 01 May 1989 |
| DOI | http://doi.org/10.1111/j.1468-0084.1989.mp51002003.x |
OXFORD BULLETIN OF ECONOMICS AND STATISTICS, 51,2(1989)
0305-9049 $3.00
REPLY
Blanchflower and Oswald
We welcome this microeconomic study but did not expect conclusive results
to emerge. Our model rests in part on non-linear aggregation effects which
one might not expect to show up on individual data. Nevertheless, the results
look encouraging rather than the reverse. It is not in the least surprising that
for four years' data, a combination of dummies for time and regions washes
out the house price effect.' It also seems unsurprising that the coefficient on
the three year change in log unemployment is not well determined, though
the point estimate is surprisingly close to ours. Among the control variables
are 'ever unemployed over preceeding five years', 'redundancy expected' and
'number of workers in the plant expected to rise etc. over the next 12 months'.
Regional aggregates of these are likely to be fairly correlated with the regional
unemployment changes.
Bariy McCormick
Barry McCormick has given us a set of thought-provoking comments that
raise interesting issues for future research. As he says, our theoretical model
does not explicitly distinguish manual and non-manual workers and we agree
with much of his interesting discussion on pp. 10-11 what differing behaviour
one might expect. He does not quite appreciate the crowding-out effect from
the lower occupancy rates reflected in high South East house prices of those
for whom portfolio demand considerations are very important (see our reply
to Patrick Minford).
He questions our use of non-manual earnings in constructing RD (which
incidentally is incorporated as a three year moving average not a single year
effect as he claims). We also experimented with an alternative in which we
weighted manual (MW) and non-manual earnings (NMW) by cross-section
estimates of their migration rates. This giyes roughly 0.4 ln MW+ 0.6 in
NMW. Using this instead of in NMW gives very similar results but a slightly
worse fit. We suspect skilled manuals, who are more mobile that unskilled
manuals, have earnings that move more with NMW than with the earnings of
unskilled manuals.2
Incidentally, we regret the inclusion of time dummies in such rich specifications.
2Like the non-manual/manual differential, skill differentials narrowed in the early 70's and
widened to new levels in the 1980's.
159
To continue reading
Request your trial