The standard model of permanent and transitory income is known to be misspecified. Estimates of income volatility in the model differ depending on the type of data moments used---levels or differences---and how these moments are weighted in the estimation. We propose two changes to the standard model. First, we account for the time-aggregated nature of observed income data. Second, we allow transitory shocks to persist for varying lengths of time. With only one additional parameter, our proposed model consistently recovers the parameters of the income process irrespective of the estimation method. To the extent that researchers employ the standard model, we advise special caution with the use of first-difference moments.