One reason for the higher level of investment subsidies may simply be that employment and investment subsidies have different underlying policy objectives. While employment subsidies clearly have the function to raise employment, investment may be subsidised for reasons not directly related to labor markets. Most importantly, investment subsidies are also an instrument to raise economic growth. The endogenous growth literature has identified various externalities arising from private investment in physical capital. These externalities may justify investment subsidies.6 Investment grants may also be a result of tax competition, where national governments offer fiscal incentives to attract internationally mobile capital.

It is a shortcoming of these explanations, however, that they call for broadly-based investment incentives. Most existing programmes, in contrast, are highly selctive and are typically designed to favour investment in certain sectors or regions. Moreover, the preferred regions are often those suffering from high rates of unemployment. One prominent example is, of course, Eastern Germany; another one is the Atlantic Provinces in Canada. The officially declared objective of subsidy programmes also typically includes job creation. This raises the question of why governments prefer investment to employment subsidies, given that unemployment is almost always a key problem of the supported regions. In fact, in a standard labour market model with rigid wages, it is a straightforward exercise in welfare analysis to show that employment subsidies strictly dominate investment subsidies (Begg and Portes (1993), Sinn (1995)). The policy implications of this theoretical benchmark case stand in marked contrast to the policy actually pursued by most governments. In the following, we will try to resolve this investment subsidy puzzle. click here

Before turning to our own model, we should briefly discuss two other possible explanations. First, many practitioners argue that, while there may be a theoretical case for wage subsidies, they are much more difficult to administer than investment subsidies. A second possible explanation has been given by Torsvik (1993). In the model of his paper, employment subsidies represent a first-best policy. The first-best, however, cannot be implemented because there is a time inconsistency problem. In his model, the government cannot commit to future payments of employment subsidies promised to firms in the present.