Do methodical traps lead to wrong development strategies for welfare : a multilevel approach considering heterogeneity across industrialized and developing countries

Subjective well-being (SWB) is becoming increasingly important as welfare concept in both scientific research and politics, as it comprises additional welfare aspects compared to the GDP per capita. Consequently, it becomes important to explicitly identify its driving forces and clarify still ambivalent findings of the literature. For this purpose, with a multilevel model we investigate the extent to which individual-level and national variables together influence subjective well-being. Moreover, we expect that life satisfaction of people in developing countries is determined differently than life satisfaction of people in industrialized countries. The database used includes both individual and national variables and is split into two subsamples of 40 industrialized countries and 41 developing countries. The results show that the national environment is highly important for a person's SWB. Thus, neglecting this national level would generate biased estimates. Moreover, the split into industrialized and developing countries shows that statistically significant and substantial differences in the effects on life satisfaction exist. Important differences are found for example regarding the income variables. We identify a saturation effect of income on the individual level, whose level is however different depending on the development status of the countries. Moreover, on the aggregated level a significant impact of GDP per capita is found for the developing but not for the industrialized countries. Thus, this study indicates that multilevel modelling approaches are necessary to obtain robust results and that the impact of macroeconomic variables diverges in dependence of the country's development status.