We investigate whether households adjust retirement savings decisions in response to changes in the means-tested public pension plans. The policy in question lowered the taper rate of the assets test on the age pension in Australia in 2007. We use HILDA, a detailed micro panel data-set for Australian households and focus on the age group between 50 and 64 in 2006, prior to the reform. We compare savings behaviours of those who were constrained to increase nancial wealth because of the assets test prior to the reform with those who were not constrained, and nd that assets tests do have a perverse impact on saving.