This paper is the first to present empirical evidence consistent with models of signaling through unemployment and to uncover a new stylized fact using the 1988-2006 DWS, namely that, among white-collar workers, post-displacement earnings fall less rapidly with unemployment spells for layoffs than for plant closings. Because high-productivity workers are more likely to be recalled than low-productivity ones, they may choose to signal their productivity though unemployment, in which case the duration of unemployment may be positively related to post-displacement wages. Identification is done using workers whose plant closed as they cannot be recalled, and no incentives to signal arise.
