Exploring Heterogeneity in Employer Learning

In this thesis I attempt to uncover the degree of employer learning in the Norwegian private sector. Employer learning refers to the process where employers learn about the initially unobserved characteristics of their workers by observing their output. In employer learning literature this process have traditionally been identified through two mechanisms. First, returns to natural ability is increasing over labor market experience. This is because as employers observe the output produced by each worker he adjusts wages according to productive capabilities. Second, returns to schooling is decreasing over labor market experience. This is because as employers rely more on output in wage-setting the importance of education level decrease, reducing the initial signaling effect of schooling. I estimate a canonical model of employer learning for a sample of full-time Norwegian private sector workers for the period 2004-2014. In addition to this, I study whether existing theories regarding sources of heterogeneity in employer learning can be applied to the Norwegian labor market. Specifically, I study whether there are differences in employer learning across education level, occupation and gender. Last, I estimate the speed of employer learning following the framework provided by Lange (2007). I find similar trends as the existing literature in terms of returns to ability, but not schooling. Although my results suggest there is evidence of employer learning, the effects are weak. I attribute this to specific characteristics of the Norwegian labor market and education system. I also argue that technological advancements could reduce employer learning by providing additional tools making employers better able to observe ability early in the employment period. All numerical results in this thesis were obtained using STATA