Better to have led and lost than never to have led at all? Lost leadership and effort provision in dynamic tournaments

In this paper, we develop and test a behavioral theory of lost leadership. Using insights from the literature on goals as reference points and prospect theory, we predict that former leaders exert more effort compared to otherwise identical competitors. We test this prediction using two contexts. The first, quasi-laboratory, data comes from an educational business game. The second setting draws on field data from a two-month banking sales contest. We find that provision of effort increases following the loss of leadership. We also explore whether past leaders exert more effort in general or simply shift effort from other, potentially less-salient goals. We find evidence of both mechanisms. Finally, we investigate the temporal effects and find that having been a leader has an attenuating effect on subsequent behavior.