Consumption and uncertainty: a panel analysis in Italian Regions

Abstract

This work tests precautionary saving theory in Italian regions, examining the relationship between consumption growth and income uncertainty in a panel of five-year averages in the period 1981 to 2000. The analysis used two different measures for income uncertainty. The first is a naive measure given by the variance of GDP growth rates while the second is a more sophisticated one obtained by computing the conditional variance by means of the expectation of GDP growth. This expectation is obtained by using the best ARMA process for each region. The results obtained confirm the importance of the precautionary saving motive in consumption decisions.

Notes

¹ To my knowledge Italian regional data have been used in consumption analysis only by Scorcu (1998), to examine the different problem of regional risk sharing. Note also that these data come from a different and smaller dataset previously provided by ISTAT.

² Guiso et al . (1992) analyse precautionary saving in Italy using a measure of earnings uncertainty based on household answers in two questions regarding; the probability distribution of the rate of growth of their earnings, and inflation in the subsequent year. However, as suggested by Lusardi (1998), this analysis, although strongly coherent with the theory, contains a potential problem since this kind of questions might not be easy for respondent to understand.

³ See also Drèze and Modigliani (1972).

⁴ See, for instance, Deaton (1991).

⁵ Campbell and Mankiw (1989) showed that a large fraction of US income accrues to individuals who follow the ‘rule-of-thumb' of consuming their current income.

⁶ Note that, coherently with Equations 3 and 4, all regressions use data of logarithm of consumption and income.

⁷ The choice of the best model is obtained by comparing different ARMA(p, q) with reference to Akaike and Schwartz criteria.