Abstract
This paper characterizes the principle of first order stochastic dominance in a multivariate discrete setting. We show that a distribution f first order stochastic dominates distribution g if and only if f can be obtained from g by iteratively shifting density from one outcome to another that is better. For the bivariate case, we develop the theoretical basis for an algorithmic dominance test that is easy to implement
Originalsprog | Engelsk |
---|---|
Udgiver | Department of Economics, University of Copenhagen |
Antal sider | 25 |
Status | Udgivet - 2007 |
Emneord
- Det Samfundsvidenskabelige Fakultet