Agent-based Labor Market Models With Applications

Agent-based Computational Economics is a method including genetic algorithm, classifier system and artificial intelligence which used to research on economic problems, and it embodies the complex and adaptive system. Different researchers on economics have applied ACE to different fields and made many economic problems to be understood extensively. In 1990s and the beginning of this century, some researchers have began to research on problems in labor market with ACE. Aimed to explain various styled facts in labor market, a few ACE models are built. Because labor market is a complicated and adaptive system, it is difficult for classical regression methods to explain the impact of public policy on aggregate variables and labor market performance. Freeman(1998) suggested that agent-based computational modeling might offer a promising additional way to explain the aggregate variables and labor market regularities. Along this line, many foreign researchers begin to research on the effect of labor institutions on aggregate variables and labor market performance (Tesfatsion, 2003, Gabriele, 2002, Fagiolo, 2004).At the end of last century, many researcher in China began to investigate and build ACE models to analyze the economic dynamics in China and assess public policies, but these models do not include labors'decisions about working behaviors and alternative labor institutions in these model so that they can not be used to analyze labor market in microscopically.In the second part, I use genetic algorithm to simulate the impact of two different inflation policies on economic outcome and unemployment dynamics. The model includes a policymaker and several public agents. The policymaker sets a target inflation using two different methods including discretion and rules. The public anticipate the target inflation according to last period real inflation using different methods. The results suggest that when the policymaker selects the discretion inflation policy, as only as the public can make some degree anticipation, then the economy will reach to Nash equilibrium inflation and other focal points. However if the policymaker keeps a lower inflation consistently, then the public can anticipate the lower inflation rate and the economy will reach to the lower inflation equilibrium. The payoff of the lower inflation equilibrium is higher than that of Nash equilibrium by analyzing the results of experiments. So if the policymaker concerns long time payoff, then he will keep economy running on the lower inflation rate.In the third section, I discuss the impact of different non-unemployment payoff terms on employers and employees. The model based on microeconomic individuals can be used to analyze bottom-up the effect of different non-employment institutions on macro-economic unemployment, efficiency and utility of the interacting two sides which are developed by employers and employee. I discussed three types of labor markets (1) Equal market in which the number of employers is equal that of employees. (2) Buyer market in which the number of employers is larger than that of employees. (3) Seller market in which the number of employers is smaller than that of employees. Employees and employers will search and match preferred partners according to utility anticipation. Once employers and employees are matched, they engage in work-site interaction which is modeled as a prisoner's dilemma game. Each worker and employer uses genetic algorithm including elitism, mutation and recombination to evolve. Elitism ensures that the most successful work-site rules retain unchanged from one generation to the next. Mutation ensures employees and employers continually to experiment with new work-site rules. Recombination ensures employers and employees continually to engage in mimicry.The results suggest that different non-employment payoff terms have no significant impact on employers and employees in equal market and buyer market. However, in the seller labor market, the term differences indeed have significant impact on both sides. As the increase of terms, employees'utility, efficiency and unemployment rate increase, employers'utility, efficiency decrease, and employers'vacancy rate increase. Longer non-employment payoff terms maker the velocity that employers and employees match lower. In seller labor market, long non-unemployment payoff term increase the market power of employees, and the risk that employer can not employ workers increase, so employer take more cooperation. So I can conclude that the same social security policy in different labor market will have quite different effects. Social security policies shouldn't concentrated on seller labor market rather than Equal and buyer labor market. Our economy is in the key stage of transition, and Social security resources are limited. Social security policies should aim at the right object which is divided according to labor market structure in order to use the limited resources.