Policy and Complex Systems
5 - Discussion and perspectives
This agent-based model ( ABM ) is
intended as a first prototype of an agent-based regional labor market model with sector-specific requirements . This paper , therefore , proposes a general methodology , agent-based computational economics ( ACE ), aimed to identify and quantify the effects of different situations and hypotheses together .
We have highlighted a mechanism of the job search in a regional labor market characterized by a complex environment where firms , worker agents , and policies interact in a systemic way . At this point , the model is still experimental and the definition of the initial conditions is preliminary ( the initial conditions on nSect , nFirm , nationality , skills S i
, and productivity P i are based on empirical data , coming from a Swiss regional labor market analysis ), but it already allows to approximate quite a number of stylized features of a regional labor market . The benefits of this type of model are potentially many . The most obvious advantage is that it permits to simulate at an individual level the political impact of highly differentiated policies . This advantage has its counterpart in terms of results . Such a model that simulates a sample of individuals or firms can give messages on the evolution of a very large number of output variables , as employment rates , unemployment , labor turnover , length of stay in different states .
Human capital accumulation within the model is effective on its own to account for many formalized facts . Indeed , the model implies that skill improvement is essential in facilitating the matching process ; furthermore , the flexibility level is a key part of an efficient labor market .
Moreover , the actual implementation of the model is effective in testing different scenarios useful in the evaluation of different policy setting . In fact , the tool permits to set model parameters on worker ’ s nationality , agent ’ s wealth , agent ’ s strategy , worker ’ s “ learning constraint ,” flexibility of the labor market , and differences between sectors . As an example , an improvement of the sectorial flexibility of the labor market can significantly modify the employment rate .
However , the approach of evaluating labor market policies with agent-based models ( ABMs ) entails strong assumptions and some issues . The model specification is a key factor that can involve difficulties and the definition of the initial conditions is still imperfect . In such models , some issues such as empirical initialization , the limitations of data collection , empirical validation , or the role of data in the design must be addressed . The model should be considered as a starting point for further improvements .
As an example , in the study of labor markets , gender , age , education , and nationality ( local-foreigners ) inequalities are the four most relevant issues to have in mind ; we are aware that the inclusion of a greater heterogeneity of agents ( gender , age , formal education ) is a crucial issue as it affects the labor demand function / decision rule in several ways . By introducing an empirically grounded agent-based modeling technique , it should be possible to address some challenging issues in modeling of complex regional labor markets phenomena .
We finally consider enlarging the analysis to a wider geographical area in order to cover a greater labor market ; the integration of a GIS module would make the policymaker self-confident with the geographical context .
27