There are three divisions in a firm, A, B and C. Each makes a different kind of product. Resources are allocated from Center and the three divisions compete for them. Over time the members of each division generate ideas for new products that need funding. The ideas may be good or bad and the division members get an accurate signal of the quality of the idea. The members of other divisions get a noisy signal or no signal at all. The three divisions have to send a vote to the Center which will determine whether to fund the idea or not. The greater the number of divisions supporting an idea, the more likely it is to be funded by Center.

Division A is “honest”. They only push their ideas if they are good. They support the ideas of other divisions if and only they become convinced by objective arguments that they are good.

Division B is an “empire builder”. They push all their ideas as if they are all good. They thrash other divisions’ ideas if they feel threatened.

Division C is “honest yet strategic”. They  push their ideas if and only if they are good. How should they vote on other divisions’ ideas? Division A is likely to be on their side when they push an idea. After all, Division C only support their own ideas if they have a good signal. They can then convince Division A of the strength of their case. To convince Center, it would be even better to have a unanimous decision with Division B on board. So, Division C supports Division B’s ideas. If Division A proposes an idea for funding and Division B opposes it, Division C sides with Division B. A quid-pro-quo equilibrium develops between Divisions B and C.

In the long run, Division A will die out. This is bad for Center as Division A’s good products are good for profits and Division B’s bad products are not. So perhaps Center will intervene. Or Division A may also become strategic. They should deliberately destroy some of Division C’s good potential products and persuade them to switch their support to them over Division B.