We adopt the habit utility specification of Campbell and Cochrane (1995) to estimate the Australian equity premium: the return on a market portfolio of equities in excess of the risk-free rate.We use Australian quarterly data for private household consumption, population, equity returns, risk-free asset returns, dividend yields and price dividend ratios taken from Datastream InternationalTM over a 28-year period from January 1973 to June 2002 providing 118 observations. A habit utility specification is able to reproduce an equity premium that is comparable to the actual equity premium. A variance ratio test rejected the hypothesis that the variance of estimates of the habit-based Arrow–Debreu equity asset prices were the same as those based on estimates using CRRA to assess volatility or on the Hansen–Jagannathan lower bound. The habit model is able to account for more of the variability in equity prices than the CRRA model. The smooth consumption puzzle is not as severe in the Australian context when the habit model is applied to this data set. However, the habit model still does not completely resolve the equity premium puzzle in an Australian context — stock volatility is still too high compared to consumption volatility and the coefficient of risk aversion is unreasonable.