318
International Labour Review
Correlation of employment outflows with business cycle
As explained above, demand for labour in the industrialized countries
increases in periods of economic upswing. At such times, employers also tend
to offer higher wages in order to attract new and more qualified workers. The
latter, in turn, are more inclined to avail themselves of better job opportunities
so that, besides new labour market entries and re-employment of previously
unemployed persons, job-to-job moves also accelerate. In contrast, in periods
of economic decline workers become more hesitant to change their jobs vol-
untarily for fear of eventually remaining jobless. Enterprises endeavour to cut
production costs in order to maintain or restore their competitiveness, includ-
ing, if necessary, by means of redundancies, early retirement schemes and
other such measures to cut their workforce.
The evolution of GDP and employment outflows, including job-to-job
moves, for four transition countries is shown in figure 2. In order to bring out
the strength of the correlation between GDP growth rates and flows from
employment to each destination, the corresponding correlation coefficients
have been calculated as well (see table 10). The underlying assumption is that
job-to-job moves are usually voluntary (unlike in the CIS countries, as men-
tioned above), while outflows from employment to unemployment are
usually involuntary, which also largely applies to flows from employment to
inactivity (forced withdrawals from employment, de-registration from unem-
ployment, early retirement, etc.). One might therefore expect the incidence of
job-to-job moves to be positively correlated to GDP growth and that of moves
to unemployment and inactivity to be negatively correlated to it. Moreover,
while job-to-job moves typically take place without much delay, employment
protection rules delay involuntary moves from employment to unemploy-
ment and to inactivity. The latter flows are therefore also correlated to the
GDP growth rates that preceded the employment outflow data by one year.
Again, the results should be taken with caution, because the data series are
rather short and incomplete, with differences between countries as far as
coverage of the transition period is concerned.
Table 10 shows strong negative correlations between GDP growth and
exits to inactivity for Estonia, Poland and Slovenia, even when exits to inac-
tivity are time-lagged. In the Czech Republic, by contrast, moves from
employment to inactivity seem to be pro-cyclical, though the time-lagged
correlation tends to show a counter-cyclical pattern (similar to that observed
in the other countries selected).
Also in line with expectations, outflows from employment to unem-
ployment have a strong negative correlation with the economic cycle in the
Czech Republic, Poland and Slovenia, regardless of whether the time lag is
considered or not. For Estonia, however, the relationship seems to be rather
weak, possibly because the results are to some extent affected by the first
three or four years of the decade, when the behaviour of enterprises and
workers was still strongly influenced by past practices. Indeed, administrative