Rawlsian fairness is motivated by two devices: the veil of ignorance that s
eparates individuals from knowledge of their ultimate positions in society
and the necessity of irrevocable consent. After making constitutional choic
es, individuals must live with the consequences of the distributive princip
les they have chosen; they cannot withdraw from society. Yet, in many probl
ems of institutional choice, people can withdraw from cooperative arrangeme
nts after the allocation of benefits has been made. The relationship betwee
n fairness and stability in general problems of institutional choice is exa
mined. In symmetric problems, fairness maximizes the stability of instituti
ons. When one individual has a higher valued outside option, maximizing ins
titutional stability requires assigning to that individual a correspondingl
y large share of the benefits of cooperation. An individual for whom the be
nefits of cooperation are more prone to risk must generally receive a large
r share of the benefits of cooperation if institutional stability is to be
maximized.