L'auteur passe en revue certaines caractéristiques générales des études économétriques des effets du salaire minimum sur l'emploi et présente un bilan des résultats des études canadiennes et québécoises sur le sujet tout en référant à l'occasion à la littérature empirique américaine.
The employment effects of changes in the (statutory) legal minimum wage still remain a controversial issue despite the bulk of empirical studies that have attempted to measure such effects in the last fifteen years or so. Although econometric, especially time-series, studies became gradually more popular than survey studies in the 70's, it is not yet obvious that much progress has been accomplished in supplying researchers and policy makers with confident and useful measures of the employment effects of changes in the minimum wage. In Canada, most econometric studies have been conducted at either the national level or for the Province of Quebec (national measures of the minimum wage are usually weighted provincial averages since in Canada the federal jurisdiction remains very limited in terms of industries and employment covered). These are reviewed here.
The purpose of the review is to identify those results that emerge with some sufficient amount of confidence to be considered «solid» or «robust». The paper has two parts. The first part discusses the empirical implications of various difficulties that characterize econometric studies with a particular emphasis on problems or issues raised in Canadian studies such as the measures of employment and of minimum wage used, the nature of the statistical series, the specification of the models, etc.
The second part presents a summary of major conclusions that emerge from the review of individual studies. Those are organized according to the category of employment studied: youth employment, employment of young persons 20-24, adult employment, part-time employment and employment in low-wage industries. This organization is similar to that used by Brown, Gilroy and Kohen in their major and extensive review of American studies of the last fifteen years. Not surprisingly, our conclusions have much in common with theirs.
Overall, very few estimates are found to be unproblematic. As already noted by Brown, Gilroy and Kohen, the major difficulty in assessing econometric estimates of minimum wage effects on employment remains the quasi-absence of sensitivity analysis. Indeed, studies that have attempted some form of sensitivity analysis report unstable and often contradictory estimates. At a more specific level, the following observations can be made:
1) it now seems clear that minimum wage increases, whether measured in real or in relative terms, do have a negative impact on youth employment with an estimated elasticity more likely between -0.1 and -0.3;
2) negative effects would also be the case for those 20-24 years old but the number and the quality of the estimates are more limited here and the possibility that the effects are zero may not yet be rejected;
3) the few estimetes produced for those 25 years and over are so problematic that no conclusion can be reached so far;
4) with respect to the length of the workweek, the suggestion that the minimum wage has a negative impact on part-time employment relative to full-time employment in Canada is not empirically supported while it seems more likely that weekly hours of work in low-wage industries show negative effects from increases in the minimum wage;
5) with respect to employment in low-wage industries, results are less clear than in the case of hours of employment.
Download the article in PDF to read it.