Abstract:
We present a dynamic occupational choice model with a learning algorithm simultaneously capable to explain entrepreneurial entry, exit, and survival. According to our model, those individuals decide to become entrepreneurs who expect their productivity to be highest when managed by themselves. As we further assume that individuals have incomplete information about their own non-cognitive skills, which are relevant for entrepreneurial processes, entrepreneurial entry in our model is driven by overconfidence in the own skills---in line with earlier empirical findings. After entry, entrepreneurs receive noisy feedback from the market. Depending on a set of traits different from those driving the entry process into the market, entrepreneurs decide to either stay or leave the market. Our learning-based model generates survival rates decreasing at decreasing rates and captures findings on the earnings puzzle according to which median entrepreneurs do not earn more than median wage workers.
