We study how daily labor supply responds to unanticipated earnings shocks among Singapore's taxi drivers using a novel identification strategy that uses idiosyncratic variation in booking cancellations and passenger no-shows (CNS) that drivers repeatedly receive. Our results provide new and more compelling evidence in support of the income-targeting model of labor supply. Not only are the average responses on the extensive margin consistent with the income-targeting model, but the responses on the intensive margin and the heterogeneous responses at different income levels and across driver characteristics are as well. We find that drivers work longer and earn more per hour following CNS, and the effects are robust after controlling for rich fixed effects, market supply and demand conditions, and drivers' sunk cost of time. The CNS effects on ending a shift exhibit a U-shaped pattern, are strongest when cumulative income is close to the average shift income, and become insignificant when the income level is too low or too high. The effects are most pronounced in the first hour of CNS and fade away quickly afterward. Drivers achieve higher productivity by reducing break time, taking more jobs, driving faster, driving to places with more earning opportunities, and having more time with passengers on board. Drivers choose the response strategies that are complementary to their abilities and circumstances such as schedule flexibility and potential for productivity improvement: those with flexible working schedules tend to prolong their shifts, whereas those with flexible earnings rates tend to increase their subsequent productivity. Our novel identification strategy strengthens the empirical literature on daily labor supply, and our findings of heterogeneity effects offer new insights on income-targeting behaviors. This paper was accepted by Matthew Shum, marketing. Funding: This work was supported by a Singapore Ministry of Education Social Science Research Thematic Grant [GrantMOE2016-SSRTG-059]. Supplemental Material: Data and the online appendices are available at https://doi.org/10.1287/mnsc.2022.4349. [ABSTRACT FROM AUTHOR]