سیاست های استخدام، شیوه های استخدام و عملکرد شرکت

In this paper we investigate how active labour market policy programmes affect firms’ hiring strategies and, eventually, firms’ performance. We focus on counseling and monitoring which may reduce search costs for employers, but which may also have ambiguous effect on the employer-employee matching quality and thus on firms’ performance. Using a large scale experiment which was conducted in Denmark in 2005-2006 and induced a greater provision of activation, we find that small firms hiring in the districts where the social experiment was conducted changed their hiring practices in favor of unemployed workers and experienced greater turnover than other firms. Treated firms also experienced no change or a marginal reduction in value added and total factor productivity during the first years after the experiment. These results are consistent with the idea that monitoring creates compulsion effects which counteract the possible improvement in the matching process expected from job search assistance.

In most European countries active labor market policy programs (ALMPs) were largely developed over the past decade. The rich literature evaluating these programs almost exclusively focuses on outcomes related to individuals participating in the interventions (see the meta-analysis by Kluve, 2010 or Card et al., 2010). Recently, also general equilibrium effects of ALMPs have been evaluated (Crépon et al., 2012, Ferracci et al., 2010 and Gautier et al., 2012) with a focus on nonparticipants indirectly affected by these programs. Although equally important, the evidence on how ALMPs may affect firms is very limited (Gautier et al., 2012 and Lechner et al., 2012). While there is a growing empirical literature showing that diversity in the workforce demographic composition has an effect on firms' outcomes,1 there is a lack of empirical evidence on the effect of workforce diversity in terms of ALMP participation history on firms' performance, except for Lechner et al. (2012). In this paper we offer to fill the gap in the literature by providing an empirical analysis of how two important components of ALMPs, namely job search assistance dedicated to the unemployed workers and their monitoring, affect firms' hiring decisions and performance. By investigating firm-level outcomes, we provide the missing input for complete evaluation of costs and benefits of ALMPs. Labor market institutions have strong potential to affect firms' economic performance by influencing their hiring and firing strategies. For example, employment protection legislation has been both theoretically and empirically documented to increase frictions in the labor market and thus impede hiring (see for instance Autor et al., 2007 and Pries and Rogerson, 2005 or Kugler and Saint-Paul, 2004). On the other hand, the effect of active labor market policies, such as counseling and monitoring, is not clear. Acting as an intermediary on the labor market, the public employment service has the potential to affect the employer–employee matching process through several channels. First, if effective, job search assistance helps the unemployed at finding job offers, and monitoring helps at maintaining certain level of search activity. As a result, unemployed workers send more applications. Labor market tightness is reduced, so firms open more vacancies (Pissarides, 2000) and are more likely to meet and hire unemployed workers. Second, job search assistance may help the beneficiaries at directing their applications to offers which suit their qualifications best. If so, ALMPs generate better employer–employee matches and have a positive effect on firms' performance. On the other hand, compulsory participation in activation programs creates a leisure tax on the participants (Rosholm and Svarer, 2008) so that monitoring and potential sanctions generate a threat effect. This compulsion effect lowers the value of unemployment (Van Ours, 2007) and pushes people back to work without concerns about the match quality. This may counteract the positive effects of counseling on the employer–employee match quality and damage firms' productivity. Thus, the effect of ALMPs on firms' performance becomes an empirical question. We offer to empirically evaluate the effect of ALMPs on firms' performance in the context of Denmark. Denmark is known for its flexicurity system introduced in 1994, which combines flexible hiring and firing regulations, generous unemployment insurance and social security net, and extensive active labor market policies (Andersen and Svarer, 2007). In this setup, intensification of activation policies has been consistently shown to have a strong positive effect on exit rates from unemployment for participants (see for example Rosholm, 2008). But the existing empirical evidence of both the compulsion effect on participants (Graversen and van Ours, 2008) and the displacement effect on nonparticipants (Gautier et al., 2012), suggests that counseling helps the beneficiaries at the expense of other participants of the labor market, be it the employing firms or the nonparticipating unemployed. This sheds doubt on the hypotheses that firms hire more unemployed workers and have better market performance when the provision of ALMPs increases. We use a Danish linked employer–employee dataset, which provides a unique panel of firms and workers, combined with accounting data to test the above hypotheses. Because we expect firms' workforce composition to be endogenous to their performance, we use a large-scale social experiment conducted in Denmark in Winter 2005/2006 that provided exogenous increase in the provision of ALMPs for firms hiring in the treated regions. We implement a panel Difference-in-Difference strategy to evaluate whether this increased provision of ALMPs changes hiring strategies of firms in favor of the unemployed, which, in turn, would lead to changes in firms' workforce compositions and affect various measures of firms' performance. We run the analysis separately for small, medium-size and large firms, as they may be affected differently by the experiment. We find that within a year following the experiment small firms which labor markets were significantly affected by the experiment hire slightly more workers from the pool of unemployed and experience greater workforce turnover. Regarding firms' performance, we find that the experiment had no or only marginal detrimental effects on firms' productivity in the short run. We find no effect for larger firms, which hiring strategies seem not to be affected by the increased provision of ALMPs. These results run along with those found by Lechner et al. (2012) for the German case, and are consistent with the hypothesis that the compulsion effect acts against the counseling effect, so that intensive counseling and monitoring improve the matching process, but not firms' performance. The rest of the paper is organized as follows. Section 2 reviews the literature and proposes a stylized matching model to underline the theoretical mechanisms. Section 3 presents the empirical question and strategy and Section 4 discusses the results. Lastly, Section 5 concludes.

In this article, we extend the evaluation of active labor market programs to firms' outcomes. The empirical evaluation literature tends to show that ALMPs affect both unemployment durations and reemployment stability of individual participants, indicating that ALMPs help improve the matching process, the match quality and/or the employability of unemployed workers. Although understanding the behavior of the demand side of the market is also crucial for a comprehensive evaluation of the ALMPs, there are only very few studies that investigate how ALMPs affect firms' hiring behavior and, especially, their economic performance. We present a simple theoretical framework to explain how ALMPs may change the terms of recruitment decision in a context where there is ex ante unobserved heterogeneity of unemployed job seekers. Our theoretical predictions indicate that if the intensity of counseling and monitoring increases, firms hire more workers from the pool of unemployed. The effect on their economic performance is, however, not clear. This is why we turn to the empirical analysis. Using an exogenous variation in the intensity of activation provided by a large-scale social experiment that was implemented in Denmark in 2005/2006, we evaluate the effects of increased job search assistance and monitoring for small, medium, and large firms. First, we check whether the theoretical prediction of increased hiring from the pool of unemployed under an increased provision of ALMPs can be observed in the data. We show that the large experimental increase in counseling and monitoring leads to an increase in the share of unemployed workers among the newly hired for the affected small and medium-size firms. It also generates higher workforce turnover in small firms. These results are in line with our theoretical predictions as well as with the previous literature, suggesting that ALMPs improve employability of the treated unemployed individuals, but they are not conclusive about the effect of ALMPs on the average match quality. Assuming that job tenure is a measure of match quality, the documented high workforce turnover in small firms could be driven by the compulsion effect lowering the quality of newly formed matches. While we are not able to directly test whether this is the case, our further analysis of the effects of ALMPs on firm economic performance gives further evidence in this direction. When evaluating the effect of the experimental intensification in ALMPs on firms' economic performance, we find that medium firms experience a marginal increase in the value added per worker due to treatment. Moreover, small firms are marginally negatively affected by increase in counseling and monitoring in their hiring regions, while medium and large firms appear not to be affected by the experiment. These mainly non-positive effects are only partially driven by the fact that the affected firms hire more unemployed job seekers, suggesting that we observe two channels. First, the affected firms hire more unemployed and thus might be forced to spend more resources on training their new workers. Second, the workers hired by the affected firms are, on average, less productive due to the compulsion effect. If so, a policy that emphasizes less on monitoring and sanctions and more on counseling could give more positive effect on firm performance. However we cannot test this hypothesis, as the experiment consists of a package of ALMPs and we cannot disentangle between the two components of the program. Moreover, one cannot conclude since a less strict unemployment insurance system (when the level of sanctions and monitoring is reduced) can have a negative effect on exit rates from unemployment, leading to longer unemployment durations and possible loss in employability. Further evidence is required here. Altogether, our results imply that the experimental increase in counseling and monitoring of the unemployed induces firms to rely more on the unemployed as the potential hiring pool. As small firms benefited most from the improvement of the matching process, our results tend to advocate in favor of hiring support targeted at small firms. We find that intensification of counseling and monitoring not only increases hiring of unemployed workers, but also increases workforce turnover. Thus, we confirm effectiveness of these policies in pushing more individuals into employment, at least in the short run. Additionally, we show that, on average, firms are not harmed by the analyzed ALMPs, which suggests that the previously estimated benefits of such policies are valid. The analysis presented in this paper has several limitations. First, the identified effects are responses to one shot experimental increase in the provision of ALMPs in the setup where firms did not know about the intervention. While this allows us to obtain robust estimates of these effects, it is not clear whether our findings can be generalized. Second, due to data availability we estimate only short-term effects. A longer observation period would allow us to test if the non-positive effects observed in the short-term are the signs of the adaptation period of the newly employed and would improve over time once this transition period is over. Finally, the analyzed experiment took place during the economic boom. As the matching literature states, when unemployment is low, firms encounter more difficulties in hiring as it is harder to meet an unemployed worker. As a result, an employment policy, such as counseling and monitoring which improves the matching function will be particularly beneficial for prospective firms. This favorable economic context may explain why the experiment had a noticeable effect on hirings, in particular for small firms which may be less visible to job seekers than larger firms and would therefore benefit from the prospective support provided by caseworkers. Furthermore, the economic context may contribute to the non-positive effect on firms' performance, as the average productivity of unemployed workers during booming years is likely to be lower than during a recession. We would then expect smaller effects on firm-level hiring outcomes and stronger positive effects on firm-level economic performance if the experiment was implemented during a recession instead of a boom. Testing whether our results are specific to the economic context (that is to a pool of unemployed consisting of, on average, very unproductive workers), or are driven by lower employer–employee match quality induced by the compulsion effect, requires to decompose the effect on firm's productivity into a specific firm effect, a specific worker effect and the match effect. The estimation of a model with both firm's and worker's fixed effect is then the natural extension to this research.