The permanent income hypothesis combined with rational expectations has led to an important insight into saving as a way of offsetting the expected decline in labour income.This saving-for-a-rainy-day motive is tested and confirmed in Finland.The paper concentrates on evaluating the effects of financial liberalization on the saving rate.It is found that the saving rate decline in the late 1980s was largely a reflection of increased indebtedness.The collapse in the saving rate was due to expenditure on housing and durables as a consequence of relaxed liquidity constraints (downpayment ratio and reference rates) and increased income expectations (eg. lowered taxes).The current recovery in the saving rate has also been affected by the exceptionally rapid increase in unemployment that reflects increased uncertainty about earnings. Changes in the aggregate saving rate are motivated by both backward and forward looking behaviour.According to estimates, the share of credit-constrained consumers could be as high as 30 50 percent.Habits and rules of thumb tend to make saving correlate with current and past income, whereas forward looking behaviour is underlined in the dependence of saving on income expectations and the real interest rate.Based on these three factors affecting the saving rate, we estimate an empirically well-fitting saving rate model.This separation, emphasizing both backward looking and forward looking determinants, can be understood intuitively by separating households into liquidity constrained and forward looking consumers.Therefore, saving rate changes can be explained as a reaction to ability to save (past and current income), willingness to save (income expectations and income uncertainty) and intertemporal price of consumption (real interest rate). This paper also evaluates the criticism aimed at the rational expectations version of permanent income hypothesis for various anomalies, including excess smoothness and excess sensitivity of consumption.Empirical tests reject almost unambiguously the REPIH restrictions including orthogonality of consumption revisions to income changes and the smoothness condition for consumption variance.Excessive sensitivity of consumption to income changes is found, which casts doubts on stringent REPIH formulations of consumption functions that attempt to model saving without explicitly modelling liquidity constraints.The rest of the paper evaluates the precautionary buffer-stock theory of saving as an alternative to the PIH.
