When Transfers Arrive: Payment Timing, Liquidity Constraints, and Labor Suppl

The 2021 expansion of the Child Tax Credit raised annual benefits to $3,000–$3,600 per child and, for the first time, delivered half the credit as monthly advance payments rather than a lump sum at tax filing. Average employment effects were near zero, but low-income married mothers with three or more children showed a substantial increase in employment. We develop a twoperiod structural model in which working requires paying childcare and fixed costs before wages are received, creating a financing constraint that depends on the timing of transfer delivery. The timing parameter ϕR is determined by statute and is not estimated, yielding a testable crosscell prediction. We show analytically that the timing effect operates only through liquidity constraints, that transfers can increase or decrease employment depending on constraint status, and that the aggregate timing response admits a sufficient-statistic representation in financinggap distributions. Estimating by SMM on 2018–2021 CPS data, the model captures 86% of the targeted parity-3+ low-income response and locates the binding-share collapse at (L,3+) from 12.7% pre-ARPA to 6.5% post. Within the estimated model, the timing channel accounts for 80% of the model-implied parity-3+ cross-quartile difference-in-differences (DD3+); since DDn is not targeted in estimation, this is a structural decomposition rather than a reducedform decomposition of the empirical DDdata 3+ = 0.070. Bootstrap inference yields a tight 95% interval for the timing increment DD3+ − DDϕ=0 3+ of [0.014,0.044] (point estimate 0.030); the implied timing share is more imprecise as a ratio. Disbursement frequency is a low-cost lever for reducing financing frictions among constrained households.