Rule-based modeling of labor market dynamics: an introduction

Abstract

Labor market dynamics are not fully understood today. Their formal description often assumes homogeneity among individual agents in the labor market. Agent based frameworks allow for the faithful description of heterogeneous systems and are thus well suited for a detailed description of labor market dynamics. Here, we introduce rule-based modeling as a novel approach of agent-based modeling with regard to the description of labor market dynamics. We discuss its advantages and limitations, and we demonstrate that BioNetGen, an implementation of rule based modeling, can efficiently describe socio-economic interactions and that it compares favorably to other agent based softwares.

Appendices

Appendix 1: Summary statistics, quarterly data

See Tables 3 and 4.

Table 3 Standard deviations (SD), autocorrelation (AC) and correlation coefficients of the number of unemployed ($u$) and vacancies ($v$), the labor market tightness ($v/u$) and hiring rate ($h$) and separation rate ($s$) for the period after first shock and before second shock to the separation rate ($\hbox {pPlJ}=0.08$, baseline: $\hbox {pPlJ}=0.04$)

Table 4 Standard deviations (SD), autocorrelation (AC) and correlation coefficients of the number of unemployed ($u$) and vacancies ($v$), the labor market tightness ($v/u$) and hiring rate ($h$) and separation rate ($s$) for the period after the second shock to the separation rate ($\hbox {pPlJ}=0.02$, baseline: $\hbox {pPlJ}=0.04$)

Appendix 2: The model in BNG syntax

The syntax of BioNetGen differs from $\kappa$ in a few aspects. Besides details in the syntax of rules, the structure of a BNG-model differs slightly because Parameters are assigned names which are then used in the rules. Perturbations to a model can be implemented by stopping the simulation of a model, changing parameter values and restarting the simulation. Here, we omit the description of perturbations in BNG. For details of the BNG syntax, please refer to Faeder et al. (2009).