Minimum Wages: Experimental Evidence Of The Impact On Productivity

New experimental research by Jordi Brandts and Gary Charness finds strong evidence that imposing a minimum wage lowers productivity at all wages, and also decreases the likelihood that a high wage will be paid. This result, published in the July 2004 Economic Journal, has important implications for wage-setting policy, as it suggests that government intervention may be literally counter-productive.

Brandts and Charness explore behaviour in experimental ”gift-exchange” markets in which firms make wage offers that are displayed to workers, who have the opportunity of accepting them. Workers then choose the levels of effort they will put into their work. Holding effort constant, higher wages yield lower payoffs for firms and higher ones for workers; higher effort levels have the opposite effects, holding wages constant.

The researchers consider the impact of competitive imbalance in the market, by varying whether there is an excess supply of firms or an excess supply of workers. They also impose a minimum wage in the market with an excess
supply of workers, and study the overall effect on wages and productivity. The results exhibit a clear pattern of reciprocal actions, with effort provision (productivity) increasing with higher wages but falling with a minimum wage. The state of competitive imbalance does not have a major impact on behaviour.

In the experiment, each firm and worker receives 10 units. Firms make wage offers between 0 and 10; if an offer is accepted, the wage is subtracted from the firm”s endowment and five times the wage is added to the worker”s endowment. The worker then chooses effort between 0 and 10, with this effort subtracted from the worker”s earnings and five times this effort added to the firm”s earnings. With an excess supply of firms, there are 12 firms and 8 workers; with an excess supply of workers, there are 8 firms and 12 workers. In the minimum-wage condition, there is an excess supply of workers and wage offers are 5 or greater.

The standard prediction is that workers will choose the lowest possible effort level; anticipating this, firms will offer the lowest possible wage. But these researchers observe average wage offers between 7 and 8 in all circumstances.

When there is no minimum wage, the ratio of aggregate effort to aggregate wage is .527 with excess firms and .516 with excess workers. But effort appears to be substantially lower when there is a minimum wage, with the effort/wage ratio reduced by 35%, to .345. The urge to offer high wages is eroded by the imposition of a minimum wage – a wage of 10 was offered 60% of the time with excess firms and no minimum wage, but only 25% of the time when a minimum wage was imposed.

The focus of this study is related to the more general theme that preferences depend not only on the outcomes that follow from certain choices, but also on information concerning the process leading to these outcomes. Information pertaining to the process may matter because it offers inferences about the intentions or disposition behind the actions of others. In this experiment, behaviour is not greatly affected by market forces (the state of competitive balance), but there are negative effects on labour outcomes from imposing a minimum wage. Perhaps policy-makers should take such results into account.

”Do Labour Market Conditions Affect Gift Exchange? Some Experimental Evidence” by Jordi Brandts and Gary Charness is published in the July 2004 issue of the Economic Journal. Brandts is at the Institut d”Anàlisi Econòmica (CSIC), Barcelona; Charness is at the University of California Santa Barbera.

Gary Charness

University of California | +1-805-893-2412 | charness@econ.ucsb.edu

Jordi Brandts

+34-93-580-6612 | Jordi.Brandts@uab.cat