25
have the disadvantage of high firing costs but entail less need to engage in costly search for job
applicants. In the Appendix we show by means of a simple model that firms are more prone to
use temporary contracts in periods of generally low profitability of hiring labour. A tighter
labour market makes it relatively more advantageous to hire workers on long-term contracts
since those “insure” the firm against worker separations that are more costly when workers are
more difficult to find. 15
The third candidate explanation is a conjecture that revolves around the firms’ incentives to
use fixed-term contracts as a screening device during periods of high unemployment. Suppose
that recruiting firms in general can extract more information about job applicants that search
while employed than about unemployed applicants. Offering a temporary job may be a means to
gain information about the worker’s productivity. The fraction of unemployed searchers in the
pool of job applicants is much higher in a recession than in a booming economy, a phenomenon
that may trigger the use of fixed-term contracts. This conjecture is not rejected by a simple
regression exercise. We have taken the ratio between the number of temporary and permanent
hires as dependent variable and regressed it on the number unemployed and the number of on-
the-job searchers over the period 1989-1998. A typical result looks like the following16:
ln(HT ∣ Hp )
= 0.23ln U - 0.66 ln OJS + 0.31 lnHT/HP )_ 1 ...
(3.1) (3.9) (2.0) -1
where HT is the number of new hires on temporary contracts, Hp the number hires on open-
ended contracts, U the number of unemployed, and OJS the number of on-the-job searchers.
Absolute t-values are shown underneath the estimated coefficients. A constant and seasonals are
15 Our discussion takes wages as given. In the model of Wasmer (1999), wages are endogenous and it is shown that
the equilibrium may involve both temporary and permanent jobs. A decline in the growth of labour productivity
brings about a rise in temporary employment. The mechanism is essentially a variant of the “capitalization effect” of
growth discussed in Pissarides (1990/2000) and Aghion and Howitt (1994). With slower growth, firms are less
inclined to hire and also less eager to retain workers by offering them long-term contracts. Other contributions to the
theoretical analysis of temporary and permanent employment include Bentolila and Saint-Paul (1992), Booth
(1997), Cabrales and Hopenhayn (1997), Guell (2000) and Blanchard and Landier (2000).
16 We used quarterly data for hires in the private sector and quarterly data for total open unemployment from the
labor force surveys. Data on on-the-job search are only available on an annual basis (from the retrospective labor