| An examination and analysis of modern utility theory—along with the methods used to test it, begins with a look at the history of utility theory. The first modern formulation of utility theory was Expected Utility (EU). However, it has been shown that there are difficulties with the theory. To a large degree EU has been replaced by newer utility theories, e.g., Subjective Expected Utility (SEU) Savage (1954), Rank and Sign Dependent Utility (RSDU), Luce and Fishburn (1991, 1995), cumulative prospect theory, Tversky and Kahneman (1992), and a recent theory based on associative joint receipts, Luce (1997). In particular, Luce and Fishburn (1991, 1995) are compared to Luce (1997). Furthermore, Parameter Estimation by Sequential Testing (PEST, a procedure adapted from psychophysics) has been used to test RSDU. There has been recent doubt about its accuracy; so, this problem has been carefully examined. Furthermore, there is some evidence that certainty equivalents; the monetary equivalent of a gamble, are not order preserving. However, a new test of certainty equivalents affirms the order preserving property. Finally, several candidates for the weighting function are empirically tested. An investigation of these problems follows a historical look at utility theory. |
