BSB vs SKY number 4
The central issue in Ghemawat's final case-study is this: Do price-wars of the sort in which BSB and Sky engaged imply irrationality on the part of one or both of the firms?
What does Ghemawat mean here by `irrationality'? This is more complicated than it seems - or than Ghemawat indicates.
It is sometimes supposed that application of an economic model requires the assumption that all agents act in such a way as to maximize money profits. This is not the case. Economists presume rationality in only a very thin sense: to treat something as an economic agent is only to assume that its choices are consistent with an acyclical ordering of preferences over states of the world. Where firms are the agents, this typically involves the assumption that states of the world involving higher money profits are preferred to states that involve lower ones, but this is not essential. (E.g., BSB could have been maximizing its reputation for toughness, for the sake of prestige. This would not render it `irrational', in that its behaviour could still have been modeled by the theorist, or by its competitors, in game-theoretic terms.)
However, suppose that BSB was internally divided, in that its manager (i.e., Simmonds-Gooding) was trying to maximize one thing (in this case, status in corporate history as a `winner') while his shareholders were trying to maximize another (in this case, money profits). In that instance, we could not treat BSB as a single agent; our game would involve at least three players - Simmonds-Gooding, the BSB shareholders, and Sky.
This is why Ghemawat wonders whether the case casts doubt on the extent to which game-theoretic modeling can safely treat firms as "unitary agents". This question is exactly equivalent to asking whether we can treat firms as rational. (There is nothing peculiar to game theory here; we could say exactly the same thing with respect to the applicability of any economic model.)