Beschreibung:
We investigate intertemporal strategic interactions if monopolies, cartels or oligopolies benefit from firm internal as well as external learning by doing. The analysis is carried out for learning cost curves based on the standard power and a linear function. The first allows for multiple long run outcomes separated by thresholds. The latter implies a linear quadratic differential game with surprising properties of the linear Markov perfect equilibrium (LMPE): First and highly policy relevant is non-existence except for very few firms and sufficiently large spillovers, relative to learning, although the open loop as well as the collusive (cartel) equilibria and the monopoly solution exist. Second and of theoretical interest is that the linear (and symmetric) Markov perfect equilibrium need not be unique