LEARNING BY DOING AND THE CHOICE OF TECHNOLOGY

This is a one-agent Bayesian model of learning by doing and technology choice. The more the agent uses a technology, the better he learns it s parameters, and the more productive he gets. This expertise is a for m of human capital. Any given technology has bounded productivity, whi ch therefore can grow in the long run only if the agent keeps switchin g to better technologies. But a switch of technologies temporarily red uces expertise: The bigger is the technological leap, the bigger the l oss in expertise. The prospect of a productivity drop may prevent the agent from climbing the technological ladder as quickly as he might. I ndeed, an agent may be so skilled at some technology that he will neve r switch again, so that he will experience no long-run growth. In cont rast, someone who is less skilled (and therefore less productive) at t hat technology may find it optimal to switch technologies over and ove r again, and therefore enjoy long-run growth in output. Thus the model can give rise to overtaking.