The Determinants of Life Satisfaction in Newfoundland and Labrador Douglas May and Philip Hoskins, September 2014 | Get paper (PDF)
Abstract: Happiness, subjective well-being and utility have been at the centre of the neoclassical microeconomic analytical framework for a considerable period of time. In its strictest form the utility that an individual obtains is dependent upon the combination of goods and services that he or she consumes. The actual or perceived utility of these goods or services is ...the very characteristic that makes the individual want to consume them. At the same it has been recognized that there are aspects of our lives beyond consumption that help to determine our subjective well-being. These concerns of micro-economists have spilled over to those interested in macroeconomics and the System of National Accounts, which focuses on the production of goods and services within an economy or by its nationals. For example, in 1974 Richard Easterlin noted that happiness does not seem to increase within a country such as the United States as GDP per capita increases. This seems to pose a paradox (called the "Easterlin paradox") for economists. Over the past decade there have been a number of studies that have attempted to measure the determinants of subjective well-being or life satisfaction and ours is one of them. These studies have increasingly relied on access to appropriate cross-sectional data on individuals and to more sophisticated econometric techniques. Our study follows that tradition by using 2012-12 Canadian Community Health Survey (CCHS) data from Statistics Canada for Newfoundland and Labrador to estimate the determinants of subjective life satisfaction using a generalized ordered logistic regression model using a gologit3 command written by Richard Williams for use in the STATA statistical package. Our results demonstrate the added value of using such a procedure including a non-linear response of subjective life satisfaction categories to changes in income as traditional economic theory would predict. Surprisingly, we also find that gender does not make a great deal of difference in explaining subjective well-being for our sample.