Identifying and measuring the economic effects of unfair dismissal laws

Theory cannot provide an unambiguous prediction regarding the economic effects of employment protection laws. Such laws confer benefits on employees and shift the labour supply curve to the right. But they also impose costs on business and therefore shift the labour demand curve to the left. The net effect on employment is ambiguous and depends on the magnitudes of the costs and benefits as well as the elasticities of labour supply and labour demand. The net effect on welfare is also ambiguous. However, since many businesses did not provide protection against unfair dismissal in the 1990s one can argue that that indicates that the costs exceed the benefits and thus the unfair dismissal laws introduced in 1993 reduced employment and welfare. Reduced form models that use employment or unemployment as the dependent variable are useless for identifying and measuring the economic effects of unfair dismissal laws. Structural models or survey based evidence is required to answer these questions. Evidence from a survey of 1800 businesses establishes that unfair dismissal laws impose significant costs on businesses and cause them to make major changes to the way in which they hire and fire workers. An estimated lower bound on these costs is $296 per full time employee. This is a lower bound in part because some business said unfair dismissal laws raised their costs but were unable to quantify by how much. Also, the opportunity cost in terms of lost productivity of continuing to employ those workers whose performance is unsatisfactory is excluded from the calculation above. Evidence from the 1990 and 1995 AWIRS survey shows that dismissal rates for cause declined from 4.4 per cent to 2.1 per cent and may have declined even further in the past decade. This suggests that the lost productivity from retaining unsatisfactory employees is likely to be high. The AWIRS data shows that in 1990 small and medium sized businesses were more likely than larger business to dismiss employees for cause. The 1995 AWIRS survey shows that small and medium sized businesses made much larger adjustments to their firing practices and thus shouldered more of the burden of these laws. Discussion of unfair dismissal laws has ignored the fact that these laws increase the risk born by businesses. Small business is unable to pool this risk and so it poses a much greater cost for such businesses. This feature may also explain why small businesses reported that they spent more on complying with and reducing their exposure to unfair dismissal laws. These considerations suggest that the Government's policy of exempting businesses with fewer than 100 employees from the unfair dismissal laws will most likely not cause major resource allocation costs. But a better policy would be abolish the unfair dismissal laws. Using a labour demand elasticity of 0.7 percent I estimate that the existing unfair dismissal laws reduce employment by at least 0.46 per cent (about 46,000 employees). Freyens and Oslington question my findings. There are two mistakes in their paper that account for their position. First they underestimate by a factor of 10 the probability that a worker is dismissed for cause. Second, they exclude the costs incurred by business in avoiding exposure to the law and focus only on the cost of complying with the law. Both their paper and my paper can be criticised for underestimating costs because we exclude the foregone productivity that arises where businesses retain some employees whose performance is unsatisfactory and who would have been dismissed under an employ-at-will regime.