Oil prices and life satisfaction: asymmetries between oil exporting and oil importing countries

Abstract

In spite of general consensus on the importance of oil prices for objective measures of economic well-being across countries, almost no research has been carried out to analyse the effects of oil prices on subjective well-being internationally. Using the World Values Survey (2009), we help fill this gap by studying the effects of oil prices on life satisfaction for two groups of countries, oil importers and oil exporters. Although some previous studies have shown negative effects of oil prices on subjective well-being of one oil importing country the United States, since it is an outlier in terms of dependence on automobiles and in gasoline consumption per capita, these findings may not be representative of other oil importing countries. Our results show that, in fact, oil prices have quite strong negative effects on life satisfaction in a sample of over 40 oil importing countries. By contrast, for oil exporting countries for which there have been virtually no previous quantitative studies, but theoretical analyses suggest the results could be ambiguous, we find strong positive effects on life satisfaction. Hence, our results reveal quite strong asymmetries in the effects of oil prices on life satisfaction between oil importers and oil exporters.

Keywords:

• life satisfaction
• oil prices
• Dutch disease
• governance

JEL Classification:

• D60
• I31
• O13
• O57

View correction statement:

Erratum

Notes

1 A number of explanations have been offered for the declining strength of these effects but thus far there would seem to be no consensus on which is the most important.

2 Unfortunately, however, the latter conclusion is limited by the fact that their sample of oil exporting countries was only two.

3 Due to the political and geographic changes taking place in Europe during the years 1980–2005, comparable values of GDP per capita were not easily available for some countries such as Serbia. Additionally, due to inconsistent sampling between waves, observations from the second wave had to be eliminated for Nigeria and from all waves for Colombia.

4 For more information on the microeconomic explanatory variables, please see Table 1.

5 Information on the approximate months of each survey, as well as on the identities of countries for which the survey month was not available is provided in Appendix Table A1.

6 The GDP series used corresponds to the Real GDP per capita (Constant Prices: Laspeyres), derived from growth rates of domestic absorption, reported in Version 6.3 of the PWT (Heston et al., 2011). When GDP per capita was not available from the PWT, the World Development Indicators Database from the World Bank (2010) was used to fill in missing values. The unemployment rate was obtained mainly from the World Development Indicators Database, but estimates from the International Labor Organization, the CIA Factbook and the national departments of statistics were also employed. For a few countries and years for which data on unemployment rates were not available, trends in adjacent years were used to extrapolate the missing observations.

7 Following Easterlin and Sawangfa (2010), the dates of these macroeconomic variables were matched, not with the actual survey dates, but rather with those for which they are most likely to affect life satisfaction. That is, surveys conducted in January–April of a given year were matched with the GDP and unemployment rate values of the previous year, the surveys from May–August to an average of the values from the present and previous years, and the surveys from September–December to the values of the present year. For surveys for which information on the exact month in which it was conducted was not available, the values of the macroeconomic variables for the year of the survey were used.

8 Since the 60% cutoff was arbitrary, in results available on request, we also repeated the analysis with different cutoff levels but found the results unaffected.

9 In the interest of space we report only on the variables of primary interest since the other results are largely unaffected.

10 The relatively long-lasting effect for oil exporting countries is consistent with negative effect of the long trough in oil prices from 1986–2000 on the average well-being index of Algeria even in the 2003 survey reported by Tiliouine et al. (2006). This supports our explanation for the hypothesized positive effect of oil prices in oil exporting countries based on the relative importance of indirect effects on the quality and quality of social services.

11 These changes in the results are mostly due to the exclusion of wave 1 in this sample, which implies that the variability of oil prices in the period analysed is greatly reduced (as oil prices were exceptionally high during wave 1).