Labour market responses to taxes: a structural approach – Structtax

The project evaluates the effect of tax policy on the labour market, for example the reductions in social security contributions that were introduced in the 1990s. The evaluation is based on an equilibrium search model of the labour market and uses very rich administrative data on wages and social security contributions from France. The model takes into account the decision of unemployed individuals whether to search for a job, as well as firms’ recruitment choices. Furthermore, the model takes into account the time that workers need to find a job and that workers and firms may fail to agree on terms of recruitment when a worker searching a job finds a vacancy. In particular, the wage negotiation process is made explicit, as well as the productive complementarities that may make productive individuals over-proportionally productive in productive firms.
Individuals may be inactive, searching for a job or employed. They are able to search for a job and receive competing job offers when they are already employed. The theoretical contribution of the project is to include both minimum wage and labour taxation in this kind of model. The data are the used to estimate the structural parameters of the model. The results provide the basis for a more general study of tax policies, helping to identify the most efficient and fairest policies.

Measures such as the crédit d’impôt pour la compétitivité et l’emploi (CICE) are likely to have more limited effecs on employment than reductions in social security contributions that are more closely targeted at low-wage jobs.

Reductions in social security contributions may generally encourage more employment, but may also have unintended consequences. In particular, firms are incited to create more low-paying jobs of low quality and with little value added. This negative effect can be mitigated if the social security reductions do not only apply to the lowest wages. According to our estimates, reductions to social security contributions should be reduced progressively for wages above the minimum wage, falling to zero for wages two times the level of the national minimum wage.

The objectives of the project have been achieved. The funding made it possible to recruit a high profile young researcher in order to carry out a very technical and ambitious project. The project proposed a new model to understand the effects of taxation on the labor market as a whole, as well as the practical ways to finely estimate the structural parameters of this model by exploiting the very rich panel data that are available nowadays. The project has broadened the conceptual framework and models for understanding the global effects of taxation. Doing so it has contributed to developing a field of research that until then was very limited.

The main output of the project is a scientific article in English that will soon be submitted to an important international journal:
Breda, T., Haywood, L. and H. Wang (2017) «Labor Market Responses to Payroll Tax Reductions«, Working paper accessible at
www.parisschoolofeconomics.com/breda-thomas/working_papers/fl_bargain_v53_TB_HW.pdf

A French-language article has already been published:
Breda, T., Haywood, L. and H. Wang (2017) «Un modèle calibré pour évaluer l'effet du CICE sur l'emploi«, Revue Française d'Economie, Vol. 32, n° 1, pp. 16-58.

The project has also given rise to numerous presentations in conferences and seminars and should spin off and lead to some follow-ups.

We propose to investigate the long-run impact of taxes on employment, working hours and wages. The objective is to relate these key labour market outcomes in France and Germany with differences in taxes, both at firm and household levels. Few studies have considered these margins jointly. The structural model we propose allows not only a joint estimation of most tax impacts on the labour market, but also simulation of any potential tax reform and its welfare effects for different individuals.

We suggest developing a realistic, general equilibrium model of the labour market to study the effects of labour market taxes. The model draws on the most recent contributions in the job search literature. It includes frictions (workers and firms do not match instantly), a wage-setting mechanism through bargaining, transitions from unemployment to employment, and job-to-job transition. So far taxes have not been included in these state-of-the-art models, despite their quantitative importance and the related policy debates on their incidence. The key value-added of the model is to allow us to study jointly the responses of employment, working hours and wages to tax changes, while the standard partial equilibrium approaches require shutting down (at least) one of those responses to study the other ones. Indeed, the typical challenge for the study of labour market taxes is that labour supply reacts to net wages, while labour demand reacts to labour cost. One thus needs to know first how both net wages and labour cost are impacted by tax changes in order to assess the subsequent effect of these changes on supply and demand. But the impact of tax changes on net and labour cost itself depends on how elastic supply and demand are in response to a change in wages. Our proposed model makes explicit the matching between firms and workers and the subsequent process of wage-determination. This implies that supply, demand and wages are jointly determined, naturally avoiding the typical issues accruing with partial-equilibrium studies.
The model also offers a natural tool to study together all types of labour market taxes and contributions: income tax, corporate tax, payroll taxes and social security contributions. In our framework, wages (and profits) depend on the surplus that can be created from the specific match between a given worker and a given firm. We model simultaneously how each type of tax affects this surplus. Doing so, we can assess the relative effect of each type of tax, and how it depends on the level of the other taxes.
Our last purpose is to study redistributive and welfare effects of potential tax reforms. To this end, we allow for individuals to differ in ability and firms to differ according to productivity. We then allow firms and workers to sort according to their productive characteristics. Since it does not assume that workers in different firms are similar, the model is able to capture the idea that if taxes affect only certain firms, they may also affect only certain types of workers.

We propose to use within-country variation in tax rates (across time and regions) to estimate the model. The structural estimates are then used to simulate various labour market policies and test how workers and firms are affected. Using the specific estimated parameters for each country, we finally discuss potential labour tax reforms and characterize welfare-enhancing tax policies.