In progress

Drinks and/or Drugs: Evaluating the substitutability of alcohol and drugs (with S. Lluis)

Using the MLDA (Minimum Legal Drinking Age) as an instrument, we examine the relationship between alcohol consumption and use of cannabis products using a large sample of Canadian youth. While other studies of this nature have found alcohol and drugs to be substitutes, we find a complementarity between the use of alcohol and cannabis. We extend our analysis to include all drugs, and also investigate whether the actual age at which the MLDA is set impacts the relationship between drug use and alcohol. The findings suggest that interventions seeking to reduce drug use should consider the intersection of these two substances, and provide further evidence that the sign of the elasticity of substitution between alcohol and drugs is highly context-dependent.

Family ties: Exploring fertility choices and parental marital disruption

This work proposes to examine the exposure to divorce of the parents and subsequent reduction in fertility as an adult based on a model where perceived stability of a partnership influences number of children and is influenced by parents’ divorce. An empirical analysis finds that parental divorce is correlated with negative selection in the choice to have any children, but, conditional on having some children, there is a small positive correlation between parental divorce and fertility but only when the woman’s parents are divorced. This result
can be applied to models of bargaining to conclude that some more unitary household bargaining models can be rejected in the Canadian context, however, the results cannot differentiate between frameworks making the
same predictions regarding the assignment of decisions to individuals within a partnership.

David vs. Goliath, or David and Goliath: Evaluating the Impact of a Regional Poverty Initiative

Can municipalities make a difference for individuals receiving benefits from programs at higher levels of government, or is their spending cancelled out by mechanical adjustments in policies that are often designed formulaically to reduce administrative costs? This article seeks to examine the impact of a regional poverty-reduction initiative enacted in Niagara Region in Ontario which saw a large increase in municipal funding for charities with the goal of creating new programming for low income families. The paper evaluates the potential conflict between regional government strategies and programs at higher levels of government, where the regional strategy could incentivize greater dependence on the higher level government’s programs, or could completely crowd out government spending. The use of administrative data permits testing of a new strategy in estimating treatment effects in duration models where there exists regional variations in the baseline hazards by using an extension to the synthetic control method within a duration model difference in difference framework. This method has broad applicability to duration models with treatment effects where stratification exists in a way that is not randomly assigned. I find that the poverty initiative resulted in a small increase in the expected duration of an Ontario Works spell, but had no effect on the growth of the caseload and an unknown effect on spending. These findings suggest that the initiative did not increase overall dependence on social assistance, and, under reasonable assumptions, that the policy was welfare enhancing for the social assistance
recipients. The proposed method of analysis was also able to reduce bias in the estimated treatment effect in the duration model over estimates conducted without matching.


This isn't flavour country anymore - Canada's menthol ban and youth smoking uptake

Early evidence has suggested that Canada's menthol ban was unsuccessful in reducing the rate of youth smoking. I explore the effect of the ban on underage smokers and the possible reasons that it was unsuccessful.

Healthcare Access Among Canada's Elderly

This research program will measure the vulnerabilities to negative health shocks in Canada's elderly population, identify geographic patterns in gaps in elderly care, as well as quantify and map inequities in the labour market disruption caused by reliance on (or choice of) informal care in lieu of professional care. 

This study will examine the implications for Canada of using different methods to calculate inequality. Measurements using both revealed and stated access methods will be provided across provinces and different socio-economic groups.