In the US, almost half of unemployment spells end through recall. In this paper, we show that the probability of being recalled is much higher among unemployment benefit recipients than non-recipients. We argue that a large part of the observed difference in recall shares is accounted for by the design of the unemployment insurance financing scheme characterized by an experience rating system. We develop a search and matching model with different unemployment insurance status, endogenous separations, recalls and new hires. We quantify what would have been the labor market under alternative financing scheme. In the absence of the experience rating, the hiring and separations would have been higher in the long run and more volatile. Experience rating system contributes significantly to the difference in recalls between the recipients and the non-recipients.