Xiaodi Xie

Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Economics / Economic Policy


Professor M. Browning


This thesis consists of three essays on empirical household behaviour; in particular, on demand and female labour supply decisions. The first essay examines empirically whether the condition for aggregation across goods and the condition for aggregation across individuals are accepted using Canadian family expenditure surveys. A new demand system is developed to test both conditions. Both conditions are rejected by the data. This places some doubt on the use of single good, representative agent models in macroeconomics.

The second essay is concerned with the impact of children on labour supply decisions of married women using the 1975 labor supply data from Panel Study of Income Dynamics. Some previous research treated the number of children as a continuous variable and has found that the number of children is exogenous in the hours of work equation of married women, suggesting that the number of children is determined independently of hours worked. This essay finds that treating the number of children as a discrete integer may be important when testing for their individual exogeneity in the hours equation. It also finds that children and labour force participation are a joint decision for married women. The findings in this essay emphasize the importance of considering both the potential endogeneity and the discrete count data nature of the number of children when estimating the policy-related parameters in labour supply equations, especially in the participation equation.

The third essay investigates the cost of children through equivalence scales using Canadian family expenditure surveys. It generalises Blackorby and Donaldson's (BD) condition concerning the structure of preferences. Both the BO condition and its generalization determine the equivalence scale uniquely from the demand data. It is found that the restriction implied by the BO condition regarding the budget share for the 'children only' good is rejected, whereas that of its generalization is not. A new rank three demand system is developed to examine the testable implications of the generalized BO condition, which are then rejected. Nevertheless, equivalence scales are estimated under the generalized SO condition. It is found that the cost of a child increases with both the age of a child and the labour force involvement of the female adult, and decreases with the income level of the household.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Included in

Economics Commons