In this paper we study the employment effects of a budget neutral restructuring of taxes levied on employers and employees. We derive conditions for taxes levied on workers to have the same employment effects as taxes levied on firms under standard processes of wage determination.
I INTRODUCTIONMost countries finance their expenditures and the social security by a mix of taxes paid by workers and by employers. Generally the taxes are related to the wages received by employees. This raises the question of the equivalence of these taxes: do taxes levied on wages have the same impact on employment if they are paid by workers or by firms? Put in another way, does a switch between taxes paid by employers and taxes paid by employees affect employment? Clearly the answer is no in a perfectly competitive market: replacing an employer's tax by an employee's tax of equal magnitude has no effect on the net wage, the gross wage or the employment level (Symons and Robertson, 1990). Still empirical evidence suggests that the answer is yes (see e.g. Lockwood and Manning, 1993;Holm et al., 1994; TyrvaÈ inen, 1995;Muysken et al., 1999), and union leaders would surely not be indifferent to an increase in the employees' taxation rate that would be compensated by a decrease in the employers' taxation rate. Therefore there is a need to depart from perfectly competitive markets to fit the facts. 1 This is what is done in a union framework by Rasmussen (1993), Holm et al. (1995 or Koskela and Schob (1999). However these models heavily rely on a non linear taxation to explain the non-equivalence result. In this paper we show that even in the case of linear taxation, both taxes may not be equivalent and we extend the results to nonunion frameworks. 2 Our model is partly related to that of Bovenberg and van der 1 This departure from perfectly competitive markets is also needed to explain why a tax per head has a different employment impact than a tax on the wage bill (see e.g. Pisauro, 1991;Rasmussen, 1997Rasmussen, , 1998. 2 Our paper encompasses results of Goerke (2000) who focuses on efficiency wage models.