The aim of this paper is to model consumption behavior of Swedish, Danish and Japanese households. Since stationarity is a big issue in time-series studies, series are tested using unit root tests. Those tests usually yield contradictory results. Appropriate differencing is applied, since consumption and income exhibit a common positive trend. Consumption and disposable income are checked for cointegration to see whether there is a short-run (SR) adjustment towards the long-run (LR) equilibrium. The Campbell-Mankiw model is run to investigate a possible division among consumers - those following the Permanent Income Hypothesis (PIH), and others who behave in a myopic way. Instumental Variables (IV) procedure is applied to eliminate the endogeneity problem. The studies are sensitive to possible structural breaks in some of the economies. At the end, the models for the three countries are compared in terms of the underlying theory behind them, economic meaning and magnitude of the parameter estimates.
Consumption stationarity Augmented Dickey-Fuller test Cointegration Error Correction Mechanism Distributed lag Models