An Evaluation of the “Jeune Entreprise Innovante” scheme An application of the sensitivity analysis model of Rosembaum

Set up in 2004, the Jeune Entreprise Innovante” (JEI) scheme allows newly created firms, whose research and development expenditures are sufficiently high, to benefit from social and tax reliefs, in particular labor tax cuts for jobs dedicated to research. This study proposes an ex post evaluation of the effects of this scheme on salaried employment (total or dedicated to research) and on wages.
For this, we rely on a matching of the JEIs with non-beneficiary companies presenting “similar” socio-economic characteristics, in order to control for selection bias in the use of this device. However the identification of a causal effect with such an approach relies on the respect of the strong hypothesis of conditional independence, which is no longer valid when there is an unobserved characteristic affecting both the selection and the chosen performance variable. The sensitivity analysis model proposed by Rosenbaum (2010, 2007, 2002c) that we implement here consists precisely in evaluatingthe impact of a relaxation of this conditional independence hypothesis by considering, for example, that after matching, one of the two companies (not necessarily the beneficiary firm) is still twice aslikely to use the JEI system. More precisely, the approachimplemented here tests the hypothesis of the existence of an effect on employment (or the salary) of the social and tax reductions from which the JEIs benefit and quantifies the extent of the unobserved selection bias that would lead to the disqualification of any causality in the correlation revealed under the conditional independence hypothesis.
Assuming that, after matching, one of the two firms is always (at most) twice as likely to use the scheme as the other, our results show that, for a potentially small but significant proportion of the beneficiary companies evaluated, there is a causual effect of the use of the JEI scheme on total and research salaried employment, which is not accompanied by an effect on the wages paid to employees.