by Antonio Bassanetti and Francesco Zollino (Bank of Italy)
Abstract
In this paper we focus on Italian aggregate quarterly time series covering the sample period 1980-2006 and test the presence and size of wealth effects on consumption, taking separately account of financial assets and residential property on the basis of new stock measures.
We find sound evidence in favour of a cointegrating relation, in which both wealth components display the expected, positive effect on households’ consumption. In particular, our results point to a lower marginal propensity to consume out of housing than out of non housing net worth, with a respective size laying in the range of 1.5-2 and 4-6 cents. Following the estimate of a vector error correction model, we discuss the role played by
transitory and permanent shocks in driving changes in the variables we consider. We have found that both consumption and wealth respond almost exclusively to permanent shocks, which are the sole determinants of the common stochastic trend in our theoretical set-up. As a result, we are confident that our estimates of wealth coefficients in the cointegrating vector match very closely the true long run marginal propensity to consume for Italian households.