1. Objectifs

Intentions :
The first chapter «[...] seeks to exploit the statistical variation generated by these 10 natural experiments brought about by changes in provincial welfare policies to measure the effects of new reform strategies on welfare participation» (p. 1) «The first goal of [the second chapter] is to investigate the extent to which the provincial-level analyses (Kneebone and White, 2009; Berg and Gabel, 2010) can be replicated using individual level microdata. The second goal is to exploit the demographic information in the microdata to describe which subpopulations were most and least sensitive to new reform strategies. [...] A third goal [...] is to estimate the effects of new reform strategies on welfare entry and exit, using transition models estimated separately for different six-year longitudinal waves of the Survey of Labor and Income Dynamics (SLID)» (p. 51-52)

2. Méthode

Échantillon/Matériau :
«Data [relative to the first chapter] on welfare participation, welfare benefit levels, as well as earnings and asset exemptions, were obtained from the National Council of Welfare (1987) and its Welfare Incomes series published nearly every year from 1990 through 2008. Provincial population data, demographics, unemployment rates, real GDP, and unemployment insurance transfers were provided by numerous data files compiled by Statistics Canada. Finally, minimum wage rates were obtained from the Minimum Wage Database compiled by Human Resources and Skills Development Canada (2009).» (p. 22-23) The second chapter «[...] uses the Survey of Labour Income Dynamics (SLID) with observations spanning the years from 1993 to 2007.» (p. 52) The chapter also uses «province-level data on unemployment rates, real GDP growth, lags of these macroeconomic variables, and unemployment insurance transfers, obtained through Statistics Canada [as well as] provincial minimum wage rates provided by the Minimum Wage Database at Human Resources and Skills Development Canada [HRSDC] (2009).» (p. 55)

Type de traitement des données :
Analyse statistique

3. Résumé

«Heterogeneous welfare reform policies and timing of those policies among Canadian provinces reveal new information about the roles of different policy tools in accounting for declines in welfare participation. Using province-level data we exploit the information in these 10 natural experiments and find that work requirements, diversion, earnings exemptions, and time limits— referred to as new reform strategies—appear to explain at least 16 percent of observed welfare participation declines in Canada from 1994 to 2005, more than eligibility requirements and benefit levels explain. Conservative estimates imply that welfare participation falls by 1.9 percentage points (21 percent relative to mean welfare participation) in provinces and years with stringent combinations of new reform strategies in place. [...] Welfare exit was initially the main channel through which participation declined, whereas reducing entry became important later. Province-specific welfare reform accounts for at least a 1.1 percentage point decline in the welfare participation rate. This large effect size survives alternate model and distributional specifications. Immigrants, native Canadians, single parents and disabled individuals experienced the largest adjustments in their rates of welfare participation as a result of policy reform.» (p. v-vi)