My current research is focused on stochastic dynamic optimization in the context of fishery management. In other words, I have spent the last few years deriving optimal fishing policies under uncertain and variable environments. Of late, frequent incidents of fishery collapse have raised concern among both ecologists and economists. Among the many causes of this phenomenon, one of the leading ones has been linked to increased variability in fishery environments. Due to the very nature of the resource, fisheries are wrought with uncertainty and are subject to stochastic shocks from numerous sources.
The broader scientific issue that I have addressed in my research is whether fishery policy is qualitatively different across deterministic and stochastic environments; my research shows that there is indeed a significant difference. One part of my work has focused on obtaining analytical results. To facilitate the derivation of analytical expressions, I have had to keep the number of stochastic shocks to a minimum. Even as the usefulness of these results for real life policy-making is limited since real life fisheries are subject to multiple shocks and my results may be sensitive to model specification, some of my conclusions are surprising and counter-intuitive. This should serve as a warning sign to policy-makers about jumping to ‘obvious’ conclusions and using rules-of-thumb to design fishery policy. Currently, I am extending these results by relaxing the assumptions.
The computational part of my research handles multiple shocks and incorporates some new techniques to solve the multiple uncertainty problem in a highly efficient manner through the twin use of Markov chains and value function iteration. Over the next few years, I would like to extend my work in the following ways:
The stochastic variables in my current work follow a uniform distribution. In the real world, agents almost never have beliefs that are consistent with the uniform density. Therefore in the short run I would like to generalize my results to distributions that are more plausible, such as the truncated normal or the log normal.
There is significant concern about northwest Atlantic fisheries, including the Georges Bank fishery as well as those off Newfoundland. It should be possible to obtain sufficient amounts of data from agencies likes the Canadian Department of Fisheries and the American Fisheries society that will permit an empirical test of the predictive ability of both analytical as well as computational models.
The growth equation used in my current work is based on the logistic function. It is has been shown in the biology literature that the growth parameter of this equation can lead to deterministic chaos. I would like to study the implication on policy of the interaction between deterministic chaos and stochastic shocks.
In the longer run, I would like to focus on environmental problems that deal with natural resources other than fisheries. One of the resources that interests me deeply is groundwater and I would like to extend the stochastic dynamic optimization techniques I have developed so far to analyze groundwater management problems.