Studies of the relative income hypothesis find positive and negative effects of relative income. To facilitate comparisons we use a large household panel and highlight the sensitivity of the relative income effect to the definition of the reference group and to the estimation strategy employed.