Overview

Abstract (English)

In my dissertation I concentrate on two themes. first, I use consumption data and look at the evolution of Canadian standards of living over the last few decades. Second, I investigate the accuracy of commonly used information and measuring practices. ^ The first chapter evaluates the impact of the 1995-1998 Ontario welfare reforms on the standard of living among single mothers. I look at how single mothers’ consumption varies in response to income shocks. Within a difference in difference natural experiment design I compare the changes in the standard of living among Ontario single mothers to changes in the standards of living of three control groups. This study considers the effects of the welfare reforms in Ontario, in comparison with other Canadian jurisdictions at the same time. The results are most consistent across measures of consumption and across methods of estimation when the comparisons are made against the control group that is demographically identical and geographically distinct. The results indicate an initial decrease in the relative standard of living among Ontario single mothers. The immediate policy impact has been reversed with the complete implementation of the welfare reforms and the introduction of National Child Benefit program. The results obtained through the propensity score matching suggest that the traditional regression adjusted methods may overstate the initial negative policy effects and understate the subsequent recovery. ^ The second chapter examines the effectiveness of the CPI as a measure of inflation faced by Canadians. I show that when incomes and food expenditures are deflated by the official price index, the mapping of food shares to income shift progressively to the left as time goes by (and GDP grows). This result indicates that the Consumer Price Index overstates the true inflation by about 50%. Consequently, any benefits indexed by CPI grew in real terms. The rate of growth of the cumulative bias overstating inflation was higher in the 1970s and 1980s then in the 1990s. Since the most important benefits indexed by the CPI are pensions, special attention is paid to senior households. The patterns of accumulation of the CPI bias vary marginally between senior and non-senior households. The final results for the 1978-2001 bias are nearly identical for both groups. ^ The third chapter investigates the properties of different ways of collecting expenditure information. The Canadian Food Expenditure Survey allows us to compare recall and diary based expenditure data collected from the same individuals. Assuming the diary measures are ‘true’ food consumption, we observe substantial errors in measures of recall food consumption. These errors do not have many of the traditionally assumed properties of the classical measurement error—their expected value is positive (non-zero), they are neither mean independent of true consumption nor conditionally homoscedastic. They are also not well approximated by either a normal or a log normal distribution. In contrast to related research, we find that household size elasticities of per capita food consumption based on the two kinds of expenditure data are very similar. Consequently, the phenomena known in literature as “Deaton – Paxson puzzle” (an empirical contradiction of the predictions dictated by the household economies of scale model) cannot be explained by the survey design. ^

Abstract (French)

Please note that abstracts only appear in the language of the publication and might not have a translation.