This paper investigates consumption behavior within an intertemporal optimization model of the representative household. Our dataset consists of deposits and withdrawals from individual household checking accounts that received paychecks by direct deposit. We construct samples of panel data for households with weekly, biweekly, and semi-monthly pay-periods and form two different measures of consumption for each sample. GMM estimates of structural parameters provide mixed evidence for habit formation or durability and limited support for the permanent income hypothesis. The results instead point to “rule of thumb” consumption under liquidity constraints, where the household consumes its current disposable income each pay-period with possible debt servicing. These findings are uniform with regard to estimation of sub-samples split according to age or household income.