آموزش و تراکم اقتصادی: برخی شواهد از ایالت های ایتالیا
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|16126||2008||23 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Labour Economics, Volume 15, Issue 1, February 2008, Pages 118–140
In this paper we use a search and matching model to investigate the economic relationship between training and local economic conditions. We identify two aspects of this relationship going in opposite directions: on the one hand, the complementarity between local knowledge spillovers and training generates a positive correlation between training and local density; on the other hand, higher wages and labor turnover in denser areas reduce training. Overall the relationship can be either positive or negative, depending on the relative strength of these two effects. Our empirical analysis, based on a sample of Italian firms, shows that training is lower in provinces with higher labor market density, measured as the number of employees per squared kilometer.
The productivity gains associated to local economic density are documented by an increasing body of empirical research (Henderson, 1986, Ciccone and Hall, 1996 and Ciccone, 2001). Understanding the sources of these gains is important for policy, especially in the light of the Lisbon strategy, which aims at making of Europe a highly competitive and productive region of the world. One source identified by the literature is the positive spatial externalities associated to the physical proximity of workers and firms, which more than offset the negative congestion effects originated by the intense use of capital and labor (Ciccone and Hall, 1996). An additional channel linking density to productivity operates via training and its positive influence on productivity:1 on the one hand, denser areas encourage firms to invest in human capital, because knowledge spillovers are better exploited by skilled workers, and trained employees are more productive. On the other hand, density increases wages and turnover, which discourage training investments and reduce productivity. If the former effect prevails on the latter, higher density is associated to more training, and thus to higher productivity. The purpose of this paper is to explore this second channel, both theoretically and empirically. We believe that the study of the relationship between density and training is important because it helps us understand how local agglomeration patterns influence productivity. The complementarity between knowledge spillovers, innovation and skills has already been emphasized in the relevant literature. Acemoglu (2002), for instance, argues that the ability of each firm to adapt new technologies and ideas developed by other firms is strictly related to the skills of its own labor force. On the empirical side, Moretti (2004), finds that productivity gains from human capital spillovers are relevant for manufacturing establishments in the US. Skills after labor market entry are generated by training and experience. Training is important both because it increases the ability to perform well the relevant task and because it improves the ability to understand and process the flow of information from the productive environment, and to translate this information into higher productivity on the job (Jaffe et al., 1993 and Anselin et al., 1997). The complementarity between skills and local knowledge spillovers suggests that firms located in denser areas have stronger incentives to invest in training: by increasing the skills of the labor force, firms benefit more from the positive externalities associated to density and are more productive. This complementarity, however, is not sufficient to generate a positive relationship between local economic density and training incidence, because denser areas have also higher wages (see Glaeser and Mare, 2001 and Ciccone and Cingano, 2003, for a review) and higher labor turnover, which discourage investment in training.2 It follows that the relationship between local agglomeration effects, measured by local employment density, and employer-provided training, can take either sign. In their empirical analysis of UK data, Brunello and Gambarotto (in press), show that the balance of positive and negative effects is tilted in favor of the latter, and that training incidence is lower, ceteris paribus, in denser economic areas. In this paper, we present a model of employer-provided training in search equilibrium which illustrates the steady - state relationship between training, density and productivity in local labor markets. The model also provides guidance to the empirical investigation which follows. We use data on more than 1000 Italian manufacturing firms, drawn from the Survey of Italian Manufacturing (Indagine sulle Imprese Manifatturiere) conducted by Mediocredito Centrale, and estimate the relationship between training incidence, measured as the percentage of trained employees in each sampled firm during the year of reference, and local labor market density, where the local market is identified with the province. We find that the estimated relationship is negative and statistically significant. While the nature of the data at hand warn us against easy generalizations, we confirm the qualitative findings obtained by Brunello and Gambarotto (in press), for a different country and with a different dataset. The combined evidence of a negative relationship between training and local density and a positive relationship between productivity and density suggests that the productivity gains to geographic proximity are not attained because firms in denser areas train more their employees. Since higher density influences productivity both directly — by stimulating innovation — and indirectly — by altering skills, our results suggest that the uncovered relationship between density and productivity would be even stronger in the absence of the negative impact of density on training. One of the advantages – identified by authors such as Marshall, Arrow and Romer – which accrue to firms locating near other producers in the same industry is that geographic proximity helps spreading information and the exchange of ideas, the discussion of solutions to problems, and the awareness of other important information (Feldman, 1993). In Marshall's words “..The mysteries of the trade become no mysteries, but are as they were in the air”. While this view focuses on the intra-industry transmission of knowledge, according to Jacobs (1969), the most important source of knowledge spillovers is external to the industry in which the firm operates, and knowledge externalities are especially promoted by the variety of the local economic system.3 Italian small and medium firms have often been scrutinized because of the important role played by industrial districts, or clusters of firms involved in the production of homogeneous goods. We find that firms operating in such districts, and belonging to the same industrial sector characterizing the district, invest more in training. On the other hand, specialization, measured by the ratio of employment in the own industry and area and employment in the area, does not seem to have any significant additional effect on training. We interpret these results as evidence that, conditional on local density, the production of skills is favored by the marked cooperative behavior typical of industrial districts (see Brusco, 1982 and Becattini et al., 1990), which reduces the risk of poaching and increases the returns to training. Turning to the policy implications, the natural question to ask is whether institutional design can affect the negative impact of local density on training. Our results suggest that the development of institutions that foster the combination of competition and cooperation, a typical feature of industrial districts, can help reducing the risk of poaching and the negative congestion effects of local agglomeration, and by so doing promote training and productivity. The paper is organized as follows. Section 1 illustrates the theoretical model. Section 2 describes the data used in the empirical analysis and provides some descriptive statistics. In Section 3 the empirical model is specified. The main results are discussed in Section 4, and a few extensions are presented in Section 5. Conclusions follow.
نتیجه گیری انگلیسی
Economic density, according to a growing body of research, encourages shared learning among individuals and firms and generates positive knowledge spillovers. Since the ability to translate information in real economic advantages is strictly related to the skills of the labor force, firms located in dense labor markets are encouraged to invest in training. However, density not only provides benefits, it also generates economic costs — identifiable with higher wages and a higher turnover risk. In the model proposed in this paper, these effects are analyzed in a matching and search framework. While from the theoretical point of view the effect of economic density on training investment can be either negative or positive, our empirical analysis suggests the prevalence of a negative effect. Using data on a sample of Italian manufacturing firms, we show that training is higher in provinces with lower employment density. This result confirms the evidence presented by Brunello and Gambarotto (in press), for the UK, and more recently by Muehleman and Walter (2006), for Switzerland. Our findings refer to off-the-job training provided by external organizations specialized in training provision. An open question is whether they can be extended to all training. While our data do not allow a direct answer, we speculate that the exclusion of on-the-job training should not substantially modify our results for two reasons: first, our sample includes mainly small firms with less than 50 employees, which find it too costly to provide their own training facilities; second, there is no clear correlation between the type of knowledge transmission (on-the-job versus off-the-job, formal versus informal) and the transferability of knowledge. While firms might substitute general with specific training, the empirical evidence shows that to a large extent on-the-job training is general, so that the threat of turnover applies as well. Our results suggest that the productivity gains associated to local economic density are not attained because firms in denser areas train more their employees. If anything, they train less, we believe because congestion effects – such as higher wages, poaching and turnover– are at work which dampen the positive effects of clustering on training. For the viewpoint of policy, the key question is whether firms in dense economic areas are training too little. It is hard to say. The economic literature on the under-provision of training is rather inconclusive, mainly because of relevant measurement issues: to establish a case for under-provision requires information on the private and social costs and returns to training, a formidable empirical task with the data at hand.