Kollusion, also Preisabsprachen zwischen Anbietern, sind ein beliebtes Thema in der Ökonomie. Preisabsprachen haben für Anbieter den Vorteil, dass der Marktpreis und der Wettbewerb, der sich damit verbindet, umgangen, und Abnehmer mit einem höheren Preis gemolken werden können. In diesem Sinne wirken Subventionen wie Preisabsprachen, aber natürlich sind Subventionen erlaubte Preisabsprachen die z.B. dem Schutz der Landwirtschaft dienen. Im Rahmen der Industrial Organization, also dem Versuch, Ökonomie weitgehend auf Grundlage der Spieltheorie zu betreiben, sind die interessantesten Ansätze zu Preisabsprachen zu finden. Wir haben bei Luis Cabràl die geheimen Preisreduktionen nachgelesen, die zeigt, dass nicht nur Anbieter an Kollusion beteiligt sind und es zeigt vor allem, dass es gar nicht so einfach ist, Kollusion aufrecht zu erhalten.
“Ready-mixed concrete and ocean shipping are examples of customer markets. These are industries where each customer is sufficiently large that prices are negotiated on a case-by-case basis. For this reason, collusive agreements are difficult to monitor: Although firms may agree on what prices to set, the temptation to secretly cut prices for a particular customer is large. In fact, what deters forms from cheating on a collusive agreement is the threat of reversion to a ‘bad’ equilibrium. But if deviations from the prescribed equilibrium cannot be directly observed, then the deterrence effect is greatly decreased.
Suppose that demand fluctuates and that these fluctuations cannot be perfectly observed. All that each form can observe is the price it sets and the demand it receives. If a firm receives permanently low demand, it is faced with a guessing problem: Its low demand may result from low overall demand, or it may result from some rival having undercut prices with respect to the agreement. Should the firm punish its rival when it receives low demand? Could it not be punishing an innocent firm?
Suppose form i decides not to punish form j on the assumption that its low demand resulted from a market downturn, not firm j’s cheating. Can this be an equilibrium? Clearly not. If that were firm i’s strategy, firm j would be better off by secretly cutting prices and blaming firm i’s low demand on market conditions. Suppose instead that each time firm i receives low demand it reverts to an infinite price war, on the assumption that low demand resulted from firm j’s cheating. Such harsh punishment would most likely suffice to keep firm j from offering price cuts. But that is little consolation,for, sooner of later, a market downturn would imply low demand for firm i regardless of firm j’s behavior. The industry would revert into an indefinite price war even if no cheating on the agreement had occurred.
Finally, consider an intermediate solution. Each time firm i or firm j receives low demand, both firms move into a price war for T periods, upon which they revert back to pricing at the collusive level. Let T be sufficiently large such that no firm has an incentive to undercut the rival. In fact, if the future is sufficiently important for each firm (i.e., if the interest rate is sufficiently low), then such T will indeed exist. We thus have an equilibrium with collusion phases alternating with price war phases, just as empirical observations suggests. Notice that, although price wars occur in equilibrium, no firm cheats in equilibrium. That is, price wars are a necessary evil of equilibrium collusion: If firms never engaged in price wars, the incentives for cheating would be too great for the collusive agreement to be stable”
(Cabràl, 2000: Industrial Organization. pp.133-135.)