The behaviour of the permanent and transitory economic shocks for different levels of households’ welfare is studied using both consumption and income measures. After testing for heteroskedasticity of the economic shocks, we use local polynomial regression models to estimate the variance of the shocks conditional on welfare level. Italian data covering the period 1980-2004 show evidence of heteroskedasticity of both the transitory and the permanent economic shocks, with the poor experiencing higher variances. The permanent shocks seem to have a more uniform effect at all welfare levels.
Centre for microdata methods and practice