Agent Simulation of Structural Change: The Venezuelan Case

C. Domingo (IEAC, Faculty of Economics and Social Sciences)
M.E. Fargier, J. Mora, S. Quiroz (Faculty of Economics and Social Sciences)
V. Ramírez (CESIMO, Faculty of Engineering)
G. Tonella (CESIMO, Faculty of Engineering and Univ. della Svizzera Italiana)
A. Rojas (Faculty of Forestal and Environmental Science)
University of Los Andes, Mérida, Venezuela

(Simulating Societies 1997 Workshop, Cortona, Italy)


Since a year ago a new economic policy has been tried in Venezuela. It is essentially an opening to foreign investment -specially in the oil sector- and privatizations in the huge public economic sector. As the process is developing in a pluralistic political environment a set of influential sectors are involved: government, the oil state monopolic enterprise, central bank, exporters, importers, national and foreign enterprises, new oil foreign and national enterprises, state enterprises, population groups, international bank and foreign pressure groups. These actors have different goals and sometimes opposite interest. This fact, the lack of a definite, consensual developing plan, and the adoption of the important economic decisions mentioned above make the process highly indeterminated. The use of simulation to throw some light on the situation is justified by the multiplicity of interacting factors. But usual simulation, based in relationships between variables descriptive of a stable structure is not enough. The dynamic of the system is now highly dependent on decisions of the actors and these are determined by their goals, available tactics, perceptions of the situation, that include perceptions of the tactics of the others, and past experiences. The process that might be simulated is this mutual interaction among behavior changing actors and a changing system that presents new restrictions and opportunities to them. In this work a model is described that tries to generate scenarios of structural changes in the socio-economic system produced by the interactions of the mentioned actors that use changing tactics, according to their goals, the state of the system and past experience. The scenarios, it is hoped, may help to uncover dangers and point to unexpected solutions. The character of the simulation experiments is exploratory rather than predictive. In the model an initial structure is defined and the normal actions of the actors are assumed: work, investment, spending, savings, monetary controls, international transactions in goods, services and currency. When certain economic measures are taken or certain variables assumed new values (inflation, foreing current reserves, exchange rates, rates of interest, profit margins, wages, income of population groups, country risk, investment opportunities, tactics of other actors) different actors select tactics of their repertories. A knowledge base is attached to each actor. This have a set of rules that relates a set of the critical values of those variables with the available tactics. So new tactics may be assumed and the corresponding actions are taking.The rules can be changed by meta-rules during the simulation run to express the actor learning. The GLIDER language, developed to deal with structural simulation is used. In the work a detailed description of the model is made, some instances of rules and meta-rules are given and some of the generated scenarios are discussed.