It is becoming increasingly difficult to satisfy the growing demands on spectrum with the conventional policy of fixed spectrum allocation. To overcome this problem, flexible and dynamic spectrum sharing methods that can significantly improve utilization of the spectrum have gained increasing interest recently. In this paper, resource allocation in decentralized spectrum sharing systems applying the existing policy rule referred to as "inequality-averse policy" is focused on. The problem related to the impact of inaccurate estimation of future traffic demands on the performance of such systems is addressed. To alleviate the problem, a novel policy termed misallocation-averse policy, which introduces adjusting factors and is resistant to such estimation errors, is proposed. The effectiveness of the proposed policy rule is verified through computer simulations.