We investigate whether removing or reducing the severity of a previously obtained criminal record improves employment outcomes. Using quasi-experimental strategies and a field experiment, we estimate the causal impact of criminal record remediation laws that have been widely enacted with the goal of improving employment opportunities for millions of individuals with records. We find consistent evidence that removing or reducing the severity of an existing record does not improve labor market outcomes, on average. There are some notable exceptions for gig work and for individuals with relatively recent felony convictions. The evidence is consistent with records initially scarring labor market trajectories in a way that is difficult to undo later.