The Life-Cycle Permanent Income Hypothesis can be formulated so that consumption in a given time period is expressed as a function of a constant, consumption in the previous time period, and a disturbance which is uncorrelated with the information known to the consumer in the previous period. In particular, current consumption is not affected by the expected value of disposable income (given the information available in the previous time period).
Empirical investigations have found, however, that disposable income does play a role in explaining consumption in equations in which consumption is regressed on its own lagged value. This result has been attributed to the absence of perfect credit markets, an assumption which underlies the theory. If a consumer cannot borrow and lend at the same interest rate whatever amount he needs to carry out his optimal consumption plan, his desired consumption in some periods will be constrained by current resources. If the `excess sensitivity' of consumption to income stems from liquidity constraints, the implications for fiscal policy can be important: the response of consumption to transitory taxes and transfers will exceed that predicted by the theory, allowing more scope for anti-cyclical fiscal policy. In Discussion Paper No. 244, Tullio Jappelli and Research Fellow Marco Pagano present new evidence on the causes of this excess sensitivity. If liquidity constraints are indeed the source, one would expect the departures from the predictions of the theory to be comparatively larger in countries with more imperfect capital markets.
Jappelli and Pagano therefore estimate a model in which consumption depends on its own lagged value and the current and lagged values of disposable income. They use annual data for seven countries in which the markets for consumer credit and mortgage loans are at varying stages of development: Sweden, the United States, the United Kingdom, Japan, Italy, Spain and Greece. The coefficient on the disposable income terms in the equation represents the excess sensitivity of consumption to income. It is significantly different from zero for all countries except Sweden, and its magnitude varies widely across countries, attaining the highest values for Italy, Spain and Greece, the lowest for Sweden and the US, and intermediate ones for Japan and the UK.
The authors then compare the estimates of excess sensitivity to measures of the outstanding stock of personal consumer loans and of housing mortgages (relative to total consumption expenditure). The ranking of countries by the sum of consumer debt and housing mortgages has an almost exact relationship to that of excess sensitivity: the countries where aggregate consumption exhibits more pronounced excess sensitivity to current income are also those where consumers borrow less from capital markets.

Differences in the amount of consumer debt could, however, be due to variations in the propensity to borrow by households rather than to rationing of the supply of credit. Jappelli and Pagano therefore provide a more detailed analysis of the factors that affect the supply and demand for consumer debt in the seven countries. They simulate the amount of consumer debt that should be observed in each country if capital markets were perfect, given the observed age structure and earnings profile of the population. Jappelli and Pagano compare the theoretical debt-consumption ratios from these simulations across countries but find that age and earnings profiles do not explain the observed borrowing patterns across countries.
Jappelli and Pagano conclude that there is no evidence of any demand factor capable of explaining the large international differences in consumer debt, and that pure rationing of the supply of credit can account for such differences. The low levels of consumer debt observed in countries where the excess sensitivity of consumption is high therefore suggest that liquidity constraints are at the root of the empirical failures of the life-cycle theory.

Consumption and Capital Market Imperfections: An International Comparison Tullio Jappelli and Marco Pagano

Discussion Paper No. 244, June 1988 (IM/ATE)