Risk management under incomplete information: exact upper and lower bounds for the probability to reach extreme values
Faculty of Applied Economics
Antwerp :UA, 2006
Research paper / UA, Faculty of Applied Economics ; 2006:19
University of Antwerp
A key problem in nancial and actuarial research, and particularly in the eld of risk management, is the choice of models so as to avoid systematic biases in the measurement of risk. An alternative consists of working with incomplete information, by xing only a number of parameters instead of a complete distribu- tion, which results in bounds instead of unique results. In the present contribution, we present upper and lower bounds for tail probabilities or probabilities to reach extreme values, in case the information about the underlying distribution is restricted to successive moments, and possibly the mode. Part of these results were already published earlier, but we present them here in a uniform and clear way, and we add results for the case of three moments.