Technology adoption under embodiment: a two-stage optimal control approach

We use two stage optimal control techniques to solve some adoption problems under embodied technical change. We first solve a benchmark problem without learning behavior. At the date of switching, the consumption level is shown to drop, as the relative price of capital goes down (obsolescence). In such a case, the economy sticks to the initial technology, or immediately switches to a new technology with a higher level of embodiment, depending on how the obsolescence costs compare to the induced growth advantage. In a second step, we introduce learning. The learning curve involves fixed costs and incentives to wait as well. Adoption is shown to depend on the growth advantage of switching net of obsolescence and learning fixed costs. The economy will switch if and only if this indicator is positive. If it is big enough to "compensate" the option of waiting, then the economy switches immediately. Otherwise, the economy waits.