We use household-level data from the Household Budgetary Survey to study the saving patterns of Maltese households. We first establish a set of key observations from the data, showing that the likelihood of saving rises with income, age and education. We then estimate a Logit model of the probability of saving on a set of household characteristics. Our results support the observations from the raw data on income and age, but show that when we control for a host of factors, the probability of saving is actually highest among households with low education. We interpret these findings through the precautionary and bequest motives as predicted by theory. Other household characteristics, such as house tenure, marital status and the number of individuals in a household also affect the probability of saving. An extended Generalised Order Logit model sheds further light on the saving intensity of households based on their characteristics. These findings can inform both macroeconomic model development as well as guide policy formation.
saving rate Household Budgetary Survey Logit model precautionary saving