This article analyses the determinants of housing savings (the contractual saving scheme) in Croatia by applying Vector Autoregression (VAR) methodology on monthly data for the period 2000-2013. The results show that the selected macroeconomic indicators (wages, interest rate, stock exchange index-CROBEX, availability of loans and unemployment rate) do not influence the housing savings market dynamics. Moreover, the housing savings system proves to be resilient to economic downturn. On the other hand, the findings are that the system primarily depends on government incentives (GI) (premiums), as confirmed by the intervention analysis. Each of the three interventions in the structure of incentives (reductions in 2005 and 2013 and the decision to freeze the payment of premiums for 2014) resulted in a significant decline in the number of new housing savings contracts. Since the findings differ from empirical research on determinants of private savings, this article emphasises that housing savings are a specific financial product, very sensitive to institutional changes. Therefore, frequent regulatory modifications, especially regarding government incentives, can easily destabilise the whole system.