In Discussion Paper No. 257, Tullio Jappelli and Research Fellow Marco Pagano use cross-section data to evaluate the empirical relevance of borrowing constraints on consumer behaviour in the Italian economy. The methodology consists of splitting the sample into low- and high-saving households and using only data relative to the latter group to estimate a reduced-form equation. Here, with no liquidity constraints, desired consumption is a function of observable variables such as age, income, wealth and family size. For liquidity-constrained consumers, however, actual is less than desired consumption. The `gap' between desired and actual consumption is then used as a summary measure of the effects of liquidity constraints.
In Discussion Paper No. 244, Jappelli and Pagano reported evidence based on time-series data on the characteristics of credit markets, in order to measure the relative importance of liquidity constraints in a number of countries (see Bulletin No. 26). The previous paper concluded that the excess sensitivity of consumption to current income fluctuations is higher in countries where credit market imperfections appear to be more pervasive; in particular, borrowing constraints appear more severe in Italy than in the United States.
In the present paper, Jappelli and Pagano investigate whether this macroeconomic conclusion is supported by microeconomic evidence on household behaviour. The authors use data on a sample of 2,515 households drawn from a 1984 survey, which contains information on consumption, income, wealth, and demographic characteristics. Again they find support for the idea that in Italy liquidity constraints significantly affect individual consumption and are more severe than in the United States. The `gap' between the desired consumption/income ratio calculated by the authors (0.884) and the actual mean of this ratio in the sample (0.786) is about 10%. In a comparable analysis, Hayashi obtained a value of 3.4% for the United States on 1963-4 data. The discrepancy between the two estimates is consistent with other evidence: during the years 1980-5 in the United States, total consumer credit has averaged 22% of personal consumption expenditure, while it has only been 4.1% of consumption in Italy; even in 1963-4, consumer credit was 20.7% of personal consumption expenditure in the United States.
The data also allow Jappelli and Pagano to investigate the characteristics of households that face borrowing constraints. The estimated gap between actual and desired consumption is highest for households headed by people less than 30 years old and lowest for those headed by people over 50. Jappelli and Pagano also find the consumption gap to be particularly large for the unemployed and for non-home-owners. This accords with the notion that banks view unemployment and lack of home-ownership as negative signals about the worthiness of their credit applicants. The gap between desired and actual con- sumption is also comparatively large for households in the South, where banks have fewer branches relative to the population, charge a considerably higher lending rate and face a higher risk of default.
These findings have clear implications for policy, Jappelli and Pagano observe; if liquidity constraints are more stringent for well defined groups of the population, the fiscal multipliers and the welfare merits of transfers aimed at these groups are likely to be substantial

Liquidity-Constrained Households in an Italian Cross-Section
Tullio Jappelli and Marco Pagano.

Discussion Paper No. 257, August 1988 (ATE)