We try to better understand the comparative advantages of structural and behavioral remedies of deregulation in electricity markets, an eminent policy issue for which the experimental evidence is scant and problematic.Specifically, we investigate theoretically and experimentally the effects on competition of introducing a forward market which the European Commission classifies as a behavioral remedy. We compare this scenario with its best alternative, the structural remedy of adding one more competitor by divestiture. Our study contributes to the literature by introducing more realistic cost configurations, by teasing apart competition effect and asset effect, and by investigating competitor numbers that reflect the market concentration in the European electricity industries. Our experimental data suggest that introducing a forward market has a positive effect on the aggregate supply in markets with two or three major competitors, configurations typical for the newly accessed and the old European Union member states, respectively. Introducing a forward market also increases efficiency. 1 Corresponding author
2/101In contrast to previous findings, our data furthermore suggest that the effect of introducing a forward market is stronger than adding one more competitor. We thus provide evidence that behavioral remedies may be more effective than structural ones. Our data hence suggest that competition authorities are well advised, in line with EU law (European Commission, 2006a, p.11), to focus on introducing, and facilitating the proper functioning of, forward markets rather than on lowering market concentration by divestiture., -, -. Rearranging this equation yields , -( , -) ( ) (( ) ) ( , -, -) , -.
5/101when none of them had sold. This prisoner's-dilemma type result is standard textbook fare (e.g., Binmore 2007, chapter 10).Allaz and Vila (1993) model the competition as a one-shot game. The Nash equilibrium of a one-shot game and that of a repeated game with predictable ending are theoretically (albeit not necessarily behaviorally; see Selten (1978Selten ( , 1991 identical. This makes the Allaz and Vila model a fitting theoretical benchmark for an experiment with a fixed, or predictable, number of periods. Of course, in the real world the number of periods may not be fixed, or predictable, and the Nash equilibrium of the stage game is one of many equilibria, some of which may be collusive: see Liski and Montero (2006). Since all treatments have the same number of rounds, between-treatment results ought to be nonetheless informative.We compare the results of introducing a forward market with those of the best alternative remedy: reducing market concentration by divestiture. We do so for competitor numbers that arguably reflect better the market concentration in the old European states than previous literature has done. We also use realistic cost configurations and tease apart competition and asset effect.We show that, theoretically and behaviorally, the effects of introducing a forward market might be larger than...