Essays in Wealth Effect, Family Structure, and Female Labour Supply

Abstract

This thesis consists of three self-contained essays evaluating topics in family structure, household wealth, and married women's labour decisions using Canadian data.

The twentieth century has seen significant changes in family formation and dissolution in Canada. Chapter 1, co-authored with Ana Ferrer, investigates the role of family structure (family disruption or reconstitution) on cognitive outcomes of primary school Canadian children. We focus on reading and math scores of these children and look into differential effects by gender as well as child’s cultural background, which is an important dimension to consider in diverse societies. Using the rich panel data information from the National Longitudinal Survey of Children and Youth (NLSCY), collected biennially since 1994, we find substantial disadvantages in reading, but not math, scores among children in single parent families, relative to children in intact families. However, we find that single parenthood seems to affect boys more than girls in terms of their reading performance, but girls’ math performance suffers more than that of boys when in step families. In addition, when looking into differential effects across cultural/religious affiliations of family structure on cognitive performance, we typically observe differential effects in math, but no reading scores. These results suggest that exploring the heterogeneity of children’s performance responses to family disruption might be an important factor in assessing the benefits of programs aimed at helping children to cope with family disruptions.

It is worth noting that changes in marital status of parents not only affect their children's performances but also influence their own welfare. The spouse (typically the wife), who usually has less labour market attachment compared to the other spouse (typically the husband) due to the traditional gender roles, is less likely to accumulate much assets during the marriage. Therefore, this spouse with less assets might have less intra-household bargaining power and could potentially face worse financial conditions in the event of a divorce compared to the other one. Chapter 2, co-authored with Stéphanie Lluis, studies a reform of the marital property law following the amendment of the Civil Code of Quebec to improve economic equality between spouses by imposing an equal division of the family assets when a marriage ends. This change created an unexpected shift in the bargaining power of the spouse with relatively lower investment in the family assets, usually the wife. We explore whether and if so how the changes in this redistributive divorce law impacted female spouses' labour market decisions and individuals’ marital decisions. We use a difference-in-difference approach and exploit detailed information on female labour supply and marital status from the Canadian Labour Force Survey (LFS) data to analyze outcomes before and after the reforms in Quebec, relative to other provinces which did not experience marital property law changes over that time period. We find that the reform of marital property law that improve economic equality between spouses in Quebec reduced married women’s hours of work and the adverse employment effect is relatively stronger for less educated women (the most disadvantaged spouse) and among couples with larger wealth as measured by the ownership of the couples’ property. At the extensive margin, we find that the redistributive law change significantly decreased the labour force participation of the relatively more educated married women but increased the labour force participation of the relatively less educated women (among married women who stayed married). This differential result by education among married women suggests that the labour supply impact of the redistributive law change likely depends on the decision to stay married as marital decisions are also part of the household bargaining outcome. We investigate this question by studying the Quebec amendment impact on divorce rates and the decisions of whom to marry. We find that the redistributive law change had no impact on overall divorce but significantly increased the likelihood of divorce/separations among less educated spouses. In addition, over the sample of young individuals deciding whether or not to marry, the Civil Code amendment contributed to increasing the proportion of marriages in which the wife is more educated than the husband.

The intra-household bargaining position is not the only factor that could affect female labour supply as well as people's marital decisions. The wealth of a household is also another important factor that might influence spouses' decisions in the labour and marriage markets. Chapter 3 examines the impact that changes in household wealth due to the house price variations during the 1990s and 2000s had on the labour market behaviour of Canadian married women. House prices in Canada have tripled over the past decades. This dramatic rise has essential effects on households' wealth and the wealth effects might be different on house owners versus renters (potential house buyers). I use time-series average house prices data from the Canadian Real Estate Association's Multiple Listing Service data set (CREA MLS) which covers the entire Canada, 102 real estate boards (REBs), and provides detailed geographical variations in house prices in both urban and rural areas. Then, I link these house prices to each respondent in the confidential longitudinal household files - the Survey of Labour and Income Dynamics (SLID). Estimating the causal effects of housing wealth changes on female labour supply is challenging. For instance, The life-cycle theory of the labour supply emphasizes that unexpected gains in wealth should decrease household labour supply. However, wealth changes due to rising house prices could be anticipated by a household. Thus, there might be no effect if the household was forward looking and incorporated these expected wealth changes into their decisions. In addition, the reverse causality between house prices and female labour supply has been highlighted in literature. Rising housing prices induce more female spouses to participate in the labour market to offset the future housing purchase costs if their families intend to enter homeownership or balance rising rental prices. Nonetheless, it is also plausible that more working women in one area, which contributes to a higher proportion of two-earner households with stronger payment capacities, may bid up the house prices there. Therefore, I apply two strategies to overcome these challenges. My first strategy is to calculate a measure of house-price shocks which is aimed at capturing unexpected variations in local house prices, rather than variations that could be anticipated by people. My second strategy is constructing comprehensive and exogenous topography instruments to address the reverse causality between the house prices and female labour supply. After capturing unexpected changes in local house prices, among house owners, I find that an increase in (positive) house-price shocks causes a reduction in the likelihood of participation of married women. At the intensive margin, I find that an increase in the house price shocks induces a decrease in annual work hours of a woman at the low percentile. Additionally, I find heterogeneous effects of house-price shocks on women's labour supply depending on their education level and residence locations. These results are consistent with the prediction of family labour supply and life-cycle models, which indicates that unexpected gains in wealth should decrease household labour supply. There is no evidence showing that house-price shocks have labour effect on renters in this study, which might suggest that they choose to delay to enter homeownership or find a cheaper residence instead of adjusting their labour supply when an appreciation of house prices occurs. The IV approach which uses the fraction of buildable land and the difference in elevation as the instruments also provides consistent results as the house-price shock approach does.