state regulatory activity and union presence. Hiring and firing costs could be lower for MNEs because
they can adjust employment by simply shifting employees from one subsidiary to another. Moreover,
being MNEs relatively footloose and having a higher bargaining power (associated with the threat of
shifting production to alternative locations), they may face fewer constraints than NEs with respect to
government, political parties or unions. This would translate in general into lower procedural costs. The
possibility of shifting employment across different locations in response to wage changes and the
greater ability to bypass country-level labour market regulations may create an asymmetry in the
behaviour of MNEs versus NEs not only in terms of the speed of employment adjustment, but also
with respect to the extent of the desired adjustment. Thus, we expect that the same reasons that can
explain lower adjustment costs of employment in MNEs would lead to higher wage elasticities in these
firms, keeping equal the skill composition of employment.
To find some empirical support to these arguments we have correlated the ratio of the long-run
wage elasticities in MNEs and NEs to a set of alternative country-level proxies of labour market
conditions. Hiring costs are proxied by the educational attainment in the population (source Barro and
Lee (1993)), on the expectation that the higher the proportion of educated labour force, the lesser are
the costs of recruiting a skilled labour force and the lower are hiring costs. The most used measures for
labour market rigidities are the employment protection indicators obtained from OECD. The different
aspects of labour market regulation are summarised in an “overall indicator of the strictness of
employment protection legislation”, available for the period “late 90’s”.17 It is not immediate to think of
proxies for procedural inconvenience. We have considered the “overall indicator of product market
regulation” (from Nicoletti, Scarpetta and Boylaud (2000)) and measures of union presence (from
Checchi and Lucifora (2002)).18
Cross-country correlations between the ratio of long-run wage elasticities for NEs and MNEs
and labour market indicators are reported in Table 4. Figure 1 reports plots of this ratio on selected
labour market indicators. From the correlation of the ratio of wage elasticities with the share of
population with college degree it emerges that the relative long-run wage elasticity in MNEs is
positively affected by the availability of skilled labour. This evidence gives indirect support to the view
that the structure of employment is more skill-intensive in MNEs. As hiring and firing skilled personnel
is less costly when skilled labour is relatively abundant, skill intensive firms like MNEs tend to have a
this as reinforcing the result that the main difference is not in the short run adjustment (either wage or output elasticities),
but in their speeds of adjustment.
17 With respect to regular employment, OECD introduce the variable “overall strictness of protection against dismissal for
regular employment”, obtained as weighed average of “regular procedural inconvenience”, “notice and severance pay for
no-fault individual dismissals” and “difficulty of dismissal”. We also tested “overall strictness of regulation for temporary
work” (weighed average of “fixed-term contracts” and “temporary work agency”) and “overall strictness of collective
dismissal”. The overall indicator of the strictness of employment protection legislation is a weighed average of these three
indicators. All these scores range from 0 to 6, with higher values representing stricter regulation.
18 The indicator of product market regulation summarises information on “State control”, “Barriers to entrepreneurship”,
“Barriers to trade and investment”, “Economic regulation” and “Administrative regulation”; data are referred to 1998.
13