Life-Cycle Models and Cross-Country Analysis of Saving

This paper develops a rational expectations life-cycle model designed as a framework for the cross-country analysis of (private) saving decisions. It is shown that a broad range of life-cycle models that have been used in the literature to study aggregate time series on consumption and saving fail to deliver plausible predictions for the purpose of analyzing saving decisions across countries as they imply that the level of saving has a constant mean and that the long-run saving rate may tend to zero. Introducing a utility specification that ties the long-run evolution of consumers' aspired consumption paths to that of aggregate labor income, an analytically tractable life-cycle model is proposed that has plausible long-run properties, including the implication that the net asset-labor income ratio, the saving rate, and the consumption-labor income ratio have meaningful long-run distributions. The moments of the long-run saving rate are shown to depend in a precise way on various characteristics of consumers' preferences, the real rate of interest, the growth rate and volatility of labor income, the government consumption-labor income ratio, and the government debt-labor income ratio. Employing a data set on saving rates and asset holdings across OECD economies and using techniques for the estimation of dynamic heterogeneous panels, the paper will also adduce empirical evidence assessing the model's ability to explain differences in the saving patterns across these economies.