effects of closures on the data. When we remove the plants which have
exited, we find that there is overall net job creation rate of 9.6 per cent.
There is a 20 per cent net job creation rate for large R&D spending plants,
compared with under 12 per cent for non-R&D spending plants. Small
R&D-spending plants have a net job destruction rate of 20 per cent (see
Table 10).
(Table 10)
(B) The Quality of Employment Created
We can attempt to measure the quality of employment created using
measures of job persistence (Davis and Haltiwanger 1996: 21-26). We
define quality in terms of the duration that a new job will exist, since our
data do not allow us to account for quality in terms of the skilled or
unskilled nature of the employment. Essentially whenever a quantity of
jobs is created, we estimate the percentage of those jobs that still exist in
1,2,3 or 4 years time. A limitation of this analysis is that we cannot be sure
that they are precisely the same jobs that survive over a given time period.
We are limited to examining to what extent the company maintains its new
employment levels which contain the newly created jobs. If all of the jobs
created survive, we note that there is a 100 per cent job persistence rate. If
the plants employment level falls to a level below that which existed
before the jobs were created, we say that there is zero per cent persistence
rate.
Table 11 is based upon a cohort analysis of all plants with ten or
more employees in 1986. Overall the average persistence of jobs created is
high. After four years almost 95 per cent of jobs persist or survive when
the category considered is all plants. However, when we look at plants
disaggregated by the sector in which they operate, we note that the
24