In this paper we study how the presence of unions and decentralized labour contracts (two-tier bargaining) may affect physical capital investment decisions of firms. More generally, we study how the above factors can influence allocative efficiency, i.e., generate a gap between the (marginal) productivity and the cost of factors of production. On the one hand, unions increase the bargaining power of workers and thus push their wages up. If firms expect this happens after the capital investment has been made (and it is therefore considered as irreversible), they will reduce investment ex-ante. On the other hand, decentralized labour contracts introduce some degree of flexibility in the wage schedule establishing a more direct link between wages and productivity, thus favoring investment of firms.
In the empirical application, we use RIL (Rilevazione Longitudinale su Imprese e Lavoro) firm level data provided by ISFOL (INAPP) for a representative sample of the population of both the limited liability companies and partnerships in the private (non-agricultural) sectors in Italy for the year 2010 to study the effects of interest. In order to overcome some econometric problems related to the presence of many zeros in the dependent variable (about 30% of firms report zero investment), we provide evidence on the effects of unions and two-tier bargaining on investment per worker estimating the models by Poisson quasi-maximum likelihood techniques.
We use two different measures of union power, which is proxied alternatively by union density (share of union members over total employees) or by a binary variable equal to one for firms with an established worker council (RSU-RSA). We find that the existence of a RSU-RSA is associated to a reduction in investment per worker of about 24%, while an increase in union density by one standard deviation tends to reduce investment per worker by about 15%. In firms with a two-tier bargaining agreement we note that investment per worker is higher by a similar amount. In other words, because firms with a decentralized agreement are generally also unionized, our results suggest that the presence of a decentralized agreement tends to exactly counteract the negative effect that unions seem to exert on investment per worker. We also investigate possible interaction effects: we find that the interaction term is positive, thus suggesting that unions have a negative effect on investment per worker only when there is no two-tier agreement within the firm; in turn, the positive effect of a two-tier agreement seems to exist only in unionized firms.
In each regression, we also include variables that take into account the role of different sectors of activity, regional areas and firm size. Moreover, we check the robustness of our results controlling for several other determinants of investment per worker and considering the role of collective national contracts. First, we include variables related to the firm workforce composition, such as the share of workers by age group, education level, gender, training status and presence of fixed-term contracts. Then, we add binary variables for firms applying a national collective contract and for firms that belong to an employee confederation. Finally, we consider additional variables that are meant to capture other firm characteristics that could be important determinants of the levels of investment per worker, such as whether the firm has already off-shored some of its activities, a binary variable for exporting firms, for firms that are run by a manager and, finally one for firms where a Cassa integrazione' schemes applies, which is a proxy for firms that have been facing tough economic and financial conditions. Our results show that these regressors are generally not statistically significant, with the exception of the Cassa Integrazione dummy which, unsurprisingly, displays a negative coefficient, and the share of workers under a fixed term contract, which in turn seems to be positevely correlated to investment per worker; moreover, when we proxy unionization with union density, there is some evidence that firms with an higher share of trained workers tend also to have higher levels of investment per worker.