Self-employment status: the role of state dependence and initial circumstances

This paper uses British longitudinal data to model self-employment status. In contrast to previous studies, the modelling approach accounts for state-dependence and unexplained heterogeneity effects. The paper concludes that state dependence is an important influence on self-employment choice. Someone self-employed last year is, controlling for observable and unobservable influences, 30 percentage points more likely to be self-employed this year than someone who was in paid employment a year ago. We also find significant individual heterogeneity in the probability of self-employment, with significant explained influences operating through gender, educational attainment, occupation, spouse's self-employment, and parental and educational background. Significant, though quantitatively smaller influences come though initial financial circumstance and current house price movements. Local labour market shocks do not appear significantly to influence self-employment choice. This we conclude that the autoregressive nature of self-employment time-series would appear to be a structural rather than a cyclical phenomenon