Do Pay-as-bid Auctions Favor Collusion? Evidence from Germany's Market for Reserve PowerZEW Discussion Paper No. 13-035 // 2013
Using bidding data on a firm-level basis which we received from the German Federal Network Agency (Bundesnetzagentur) we analyze a drastic price increase in the market for negative secondary reserve power in Germany between 2009 and 2010. As there were no conspicuous movements on the spot market of the European Energy Exchange (EEX) which serves as a substitute market for the owners of generation capacity eligible for producing reserve power, the price increase did not appear to be driven by increased costs. Against the background of several market-design modifications aiming to increase competition and liquidity in Germany's market for reserve power during recent years and an increasing number of market participants the price increase was even more surprising.
We study market structures and market power in the first stage and apply common measures like HHI, Concentration Ratios as well as measures developed for electricity markets like Pivotal Supplier Index (RSI) and Residual Supply Index (RSI). We identified a concentrated market with highly pivotal suppliers in an environment with completely price-inelastic demand and high entry barriers. Subsequently we analyzed individual bid strategies with regard to the present pay-as-bid auction design and its "guess the clearing price" principle. Our results suggest that the price increase can be traced back to a reduction in supply of the most dominant supplier and the interaction of the two most dominant suppliers' bidding strategies afterwards. Thereby a spiraling price increase was initiated by repeated pretended "bad guessing" and prices maintained on a higher level even after supply was brought back to its initial volume again.
We demonstrate the absurdity of the "guess the clearing price" idea behind discriminatory auctions in an environment with highly pivotal suppliers, inelasticity of demand, repeated auctions and the absence of demand uncertainty. While the first three characteristics can be found in virtually all energy markets, validity of our findings for energy markets in general depends on the degree of demand uncertainty and the frequency of similar market situations. Thus, given regulatory threats, the deemed main advantage of pay-as-bid auctions over uniform price auctions and the popular belief that they reduce dominant suppliers' withholding incentives and diminish their ability to tacitly collude can be rebutted as general rule. In fact, while strategic capacity withholding immediately leads to higher average prices for all suppliers in uniform price auctions, in pay-as-bid auctions bidders do not directly benefit from it but can generate extra-profits lagged by one period. We conclude that proving abusive behavior becomes harder in discriminatory auctions even when cost estimation is feasible since suppliers can hide behind the "guess the clearing price" principle and guessing prices wrong is not sufficient for legal prosecution.
Heim, Sven and Georg Götz (2013), Do Pay-as-bid Auctions Favor Collusion? Evidence from Germany's Market for Reserve Power, ZEW Discussion Paper No. 13-035, Mannheim.