A large share of the growing demand for care is met informally by relatives. Many family caregivers also work, and the empirical evidence suggests that burden of caregiving interferes with employment. Although it is understood that caregiving ultimately results in lower labor force attachment, the dynamic employment trajectories around the start of a caregiving spell are not well understood. We use retrospective caregiving reports from the Survey of Income and Program Participation (SIPP) linked to administrative records to create a high-frequency panel dataset and estimate the dynamic relationship between work and caregiving for new caregivers. We find that the probability of employment declines by approximately 2 percentage points immediately following the start of caregiving and this decline is sustained for at least two years. The decrease in employment is accompanied by a drop in (unconditional) earnings, and is driven largely by exit from the labor force, rather than by transitions into unemployment. We find evidence that female caregivers return to the labor force approximately two years after the start of a caregiving spell, and this return is characterized by lower hours and higher likelihood of self-employment. Labor force exit among male caregivers appears to be permanent. Finally, we find increases in Social Security Disability Insurance (DI) claiming among all caregivers following the start of a caregiving spell.