The Changing Labour Force Participation of Canadians, 1969-96: Evidence from a Panel of Six Demographic Groups

Cyclical, policy changes, and structural factors have been put forward to explain the decline in labour force participation in Canada in the 1990s. In the first article in the symposium, Pierre Fortin and Mario Fortin attempt to determine the relative importance of these three types of factors by estimating a participation rate equation where the independent variables are an index of job availability (the Help-Wanted Index), the real wage, the real minimum wage, an index of unemployment insurance generosity, real social assistance benefits, and a time trend to capture other structural influences. The equation is estimated for six demographic groups (men and women 15-24, 25-54 and 55 and over) over the 1969-96 period. They find that the drop in the aggregate participation rate in the 1990s was equally due to the three factors: poor macroeconomic conditions, policy changes in unemployment insurance and the minimum wage, and structural transformations. The explanatory power of the equations is found to be good for the younger and middle-age groups, but poor for the older groups. The equations show no significant shift over the 1990-96 sub-period, implying that the decline in the participation rate in the 1990s should not be attributed to a new structural relationship. In terms of the variables, the authors find that young people, and to a less extent, middle-aged persons, respond positively to cyclical variations in job opportunities, but older persons do not; minimum wages affect negatively the labour force participation of the younger age groups and middle-aged women; UI generosity affects positively only youth participation; and social assistance affects negatively the participation of middle-aged women. The authors use their equations to simulate the impact of changes in cyclical, policy, and structural variables on labour force participation of the six groups in the 1990-97 period. The model can account for 2.6 percentage points of the total 2.7 percentage point decline in the aggregate participation rate between 1989 and 1997. Reduced job availability was responsible for 37 per cent of the decline, increases in the real minimum wage 15 per cent, decreased UI generosity 19 per cent, falls in social assistance benefits made a 4 per cent negative contribution to the decline in participation, and structural factors accounting for the remaining 30 per cent of the decline. These structural factors included rising school attendance, the changing roles of men and women in society, the changing demand and supply of skills, the expansion of public pension plans, and the rising average age of the 55 and over group. The simulations explained quite well the evolution of labour force participation for all groups except older males. Perhaps surprisingly, none of the large decline in participation of this group was found to be due to reduced job availability. The finding that only about 40 per cent of the decline in labour force participation in the 1990s is cyclical has important implication for the path of potential output. The authors find that whatever the value of the non-accelerating inflation rate of unemployment (nairu), in a non-inflationary cyclical recovery, the participation rate is unlikely to rise more than 0.8 percentage points from the 1997 level of 64.8 per cent.