Current/Ongoing Research
Reports
"Assessing the Economic Impact of Legal Aid in British Columbia - Promising Areas for Future Research" with Yvon Dandurand (UFV). The Law Foundation of British Columbia (2012).
OVERVIEW - In light of the serious challenges faced by legal aid service providers in trying to respond to the needs of British Columbians and the decreasing financial support for the provision of these services, more attention needs to be given to the economic impact of legal aid. It is imperative to better understand not only the social and economic impact and other benefits of legal aid, but also the social and economic costs and consequences of failing to provide adequate legal aid services. The direct and indirect economic benefits of providing different types and levels of legal aid services through different service delivery approaches are poorly understood and mostly undocumented. The present review considers previous attempts, in Canada and abroad, to identify and measure the social and economic benefits of legal aid and the social and economic costs of failing to make that service accessible to people who face various legal problems (justiciable issues). It offers a discussion of the feasibility of different options for conducting a methodologically sound study (or studies) to fill these gaps and produce a credible and hopefully uncontroversial assessment of the economic impact of legal aid in British Columbia. The goal is to identify what kinds of economic benefits analyses are methodologically possible and realistically feasible given the state of existing data on the legal services provided, the legal services needs of the British Columbians, the functioning of the justice system, and the impact of unaddressed legal needs of various segments of the population. The review therefore concludes with concrete suggestions about distinct studies which could be undertaken in the short to medium term to demonstrate the cost-effectiveness and benefits of further public investments in the provision of legal aid.
Submitted Papers
ABSTRACT - This work is concerned with the possible impact binary encoding of strategies may have on the performance of genetic algorithms popular in agent-based computational economic research. In their recent work, Waltman, et al. (2011) consider binary encoding and its possible contribution to a phenomenon referred to as premature convergence; the observation that different individual runs of the genetic algorithm can lead to very different results. While Alkemade et al. (2006, 2007, 2009) argue that premature convergence is caused by insufficient population size, Waltman et al. argue that this phenomenon depends crucially on strategies being encoded in binary form. This conclusion is based on their illustration that premature convergence can be avoided even in simulations with small populations so long as real, rather than binary, encoding of strategies is utilized. Utilizing their methodology, we return to the consideration of the cause of premature convergence. After robustness checks with respect to the length of the binary string used for encoding, the fitness function, and the form of mutation, it is concluded that an alternative specification of mutation may also alleviate the occurrence of premature convergence. It is argued that this alternative form of mutation may be more appropriate in a wider range of problems where real encoding of strategies may not prove sufficient.
ABSTRACT - The focus of this work is to advance the standard cobweb market model testbed for learning algorithms in two significant ways. The two changes to the baseline environment are interdependent and each is introduced in order to create a model more in line with reality. First, similar to the information constraint that burdens firms in previous work, in this environment buyers must also formulate their consumption decisions prior to acquiring full knowledge of all relevant data. Second, the environment considered is one where there are multiple goods over which agents must make production and consumption decisions and whose demand is interdependent. It is assumed that agents may engage in transactions over only one good at a time. As such, formation of expectations over goods for which transactions have not yet occurred is particularly important for individual consumer demand. In this sense, expectations are important for both sides of the market and, similar to supply, demand evolves according to the formation of expectations governed by the learning algorithm being employed. Price changes in one market affect the demand firms face in all other markets which is a result of evolving consumer expectations and is not exogenous or static. Taken together, these two changes to the standard model produce one in which multiple interdependent markets may evolve simultaneously. Two forms of evolutionary learning are examined in this extended framework: social evolutionary learning (SEL) and individual evolutionary learning (IEL). The likelihood of convergence towards general equilibrium and the characteristics of this convergence are examined and compared for these two algorithms under various learning parameters. (Presentation Slides)
"Particle swarm optimization in agent-based economic simulations of the Cournot market model"
ABSTRACT - The numerous variations over the particle swarm optimization algorithm (PSO) originally proposed by Kennedy and Eberhart (1995) have proven to be powerful optimization methods that rely on exploiting simple analogues of social interaction. In this work, the particle swarm optimization algorithm is adopted in lieu of the social and individual evolutionary learning algorithm as a model of individual adaptation in an agent-based computational economic model of the simple Cournot market framework. In this examination each agent's individual strategy evolves according to the PSO algorithm. This agent-based model is one in which agents' strategies must adapt within a much more dynamic economic environment than in previous implementation of the PSO algorithm; the fitness associated with individual positions within a population (or swarm) evolving according to the algorithm is interdependent. The dynamics and convergence properties associated with this model are compared to those associated with those where the social and individual evolutionary learning algorithm governs the evolution of individuals' expectations.
Work in Progress
"Learning Organization and Competition", with Jasmina Arifovic (Simon Fraser University) and Gerard Ballot (University Paris II and ERMES)
ABSTRACT - While much of the previous microeconomic and industrial organization research has focused on price competition, product differentiation and process innovation, a decisive competitive advantage for firms may lie in finding the efficient form of organization. This paper examines the manner in which heterogeneous firms choose their organization of work in a competitive setting in which the optimal choice cannot be computed. In this work, the selection of the dominant organization type(s) may be impacted through various potential channels including the adoption of successful types by other firms (imitation), the experimentation with new types, the exit of firms which learn too slowly and the constraints determined by the endogenous labor market. Our analysis of the determinants of organizational structure occurs in a dynamic aggregate setting in order to formalize a model of how the organizational choices evolve and to analyze in a such a framework if there is convergence to one or several types of organization.