Using the 2017 National Hockey League Expansion Draft as a natural experiment we analyze the decisions of General Managers to protect or expose their players to selection by the expansion team and the selection preferences of the expansion team. We find 1) The presence of endowment effects or divestment aversion by existing General Managers in that players with pre-existing relationships with their General Manager are 1.6 times more likely to be protected than players without such prior relationships . 2) Players who play fewer minutes relative to their position are less likely to be protected. 3) Players past their prime are less likely to be protected. 4) Player selection is predominantly determined by time-on-ice. Implications and future research are discussed.