Postprint version. Published in Review of Accounting and Finance, Volume 4, Issue 4, January 1, 2005, pages 34-49.
NOTE: At the time of publication, the author Sanjiv Jaggia was not yet affiliated with Cal Poly.
The definitive version is available at https://doi.org/10.1108/eb043436.
One of the key elements of survival models is that they enable the researcher to determine whether the length of time an individual (or economic entity) spends in a particular state affects the probability of exiting that state. Natural applications in economics and finance include the analysis of unemployment spells, corporate bankruptcies and mortgage pre-payments. The distinguishing feature of most applications is the definitive event that marks the transition from the origin to the transition state. We believe that limiting the use of survival analysis to applications in which the event duration appears to be 'naturally' available is an unnecessary constraint. For example, the date of emergence from Chapter 11 bankruptcy protection is a subjective management decision and the true event duration, though treated as definitive, is in reality quite ambiguous. We propose that survival models can and should be extended to analyze researcher-defined events such as the length of time a stock takes to reach a pre-set price target. We illustrate our point with an examination of IPO aftermarket behavior.