The relative role of economic growth vis-a-vis public action in raising living standards in developing countries has been a point of contention for quite some time now. The arguments on both sides are usually based on some estimated relationship between indicators of living standard and other variables. A critical review of the existing studies throws up some methodological issues among which misspecification of the model is most crucial. An alternative approach, viz. the nonparametric regression method, has been shown to be superior in taking care of this problem. Analysing the data for 88 developing countries we note that per capita income has positive significant effect on the life-expectancy at birth. However, we have not observed any relationship between the improvement in life-expectancy and change in income as well as the level of income, unlike some earlier studies. This study has the implication that well-targeted public policies may be successful in improving the standard of living in poor economies in the short-term. But for sustainability in the long-term, growth-based strategies are necessary.