Polish experience from first-ever spectrum auction

The Polish 4G spectrum auction was a watershed event for the Polish telecommunications market. For the first time in history, the Office of Electronic Communications decided to award spectrum by using an auction process. Polish regulators implemented a simultaneous multiple-round ascending-bid auction, which was widely used worldwide for selling spectrum. The process leading to the auction and the auction itself was lengthy and eventful. Due to the wrong auction design, the auction dragged on indefinitely, causing a delay in spectrum distribution. The auction ended only as a result of the government's intervention, after 8 months (513 rounds) of bidding that raised revenue of over PLN 9 billion (over EUR 2 billion). This paper describes in detail the rules of the Polish 4G auction, analyses bidding behaviour and auction dynamics of the crucial phases of auctions, and presents the final outcomes. It also draws lessons that could help policy makers in Poland and other countries in auction design.     

Bidding asymmetries in multi-unit auctions: Implications of bid function equilibria in the British spot market for electricity

This paper introduces and tests Bid Function Equilibria (BFE) in the British spot market for electricity. BFE extend von der Fehr and Harbord's (1993) multi-unit auction model of wholesale electricity markets by allowing firms to have heterogeneous costs for different generating units. Pure-strategy equilibria in BFE predict asymmetric bidding by producers: a single firm (the “price-setter") bids strategically while other firms (“non-price-setters") bid their costs. We test for asymmetries in firms' bid functions in the British spot market between 1993 and 1995 and find strong empirical support for the theory. We conclude that BFE have important implications for the design and governance of electricity markets.     

Sorry Winners

Bidders who receive both “common-value” and “private-value” signals about the value of an auction prize cannot fully infer their opponents’ information from the bidding. So bidders may overestimate the value of the prize and, subsequently, regret winning. When multiple objects are on sale, bidding in an auction provides information relevant to the other auctions, and sequential auctions are more vulnerable to overpayment and winners’ regret than are simultaneous auctions. With information inequality among bidders, the seller’s revenue is influenced by two contrasting effects. On the one hand, simultaneous auctions reduce the winner’s curse of less informed bidders and allow them to bid more aggressively. On the other hand, sequential auctions induce less informed bidders to bid more aggressively in early auctions to acquire information.      

Spectrum auction design: Simple auctions for complex sales

Following the successful PCS Auction conducted by the US Federal Communications Commission in 1994, auctions have replaced traditional ways of allocating valuable radio spectrum. Spectrum auctions have raised hundreds of billion dollars worldwide and have become a role model for market-based approaches in the public and private sectors. The PCS spectrum was sold via a simultaneous multi-round auction, which forces bidders to compete for licenses individually even though they typically value certain combinations. This exposes bidders to risk when they bid aggressively for a desired combination but end up winning an inferior subset. Foreseeing this possibility, bidders may act cautiously with adverse effects for revenue and efficiency. Combinatorial auctions allow for bids on combinations of licenses and thus hold the promise of improved performance. Recently, a number of countries worldwide have switched to the combinatorial clock auction to sell spectrum. This two-stage auction uses a core-selecting payment rule. The number of possible packages a bidder can submit grows exponentially with the number of licenses, which adds complexity to the auction. For larger auctions with dozens of licenses bidders cannot be expected to reveal all their valuations during such an auction. We analyze the impact of two main design choices on efficiency and revenue: simple “compact” bid languages versus complex “fully expressive” bid languages and simple “pay-as-bid” payment rules versus complex “core-selecting” payment rules. We consider these design choices both for ascending and sealed-bid formats. We find that simplicity of the bid language has a substantial positive impact on the auction?s efficiency and simplicity of the payment rule has as a substantial positive impact on the auction?s revenue. The currently popular combinatorial clock auction, which uses a complex bid language and payment rule, achieves the lowest efficiency and revenue among all treatment combinations.     

Synergistic valuations and efficiency in spectrum auctions

In spectrum auctions, bidders typically have synergistic values for combinations of licenses. This has been the key argument for the use of combinatorial auctions in the recent years. Considering synergistic valuations turns the allocation problem into a computationally hard optimization problem that generally cannot be approximated to a constant factor in polynomial time. Ascending auction designs such as the Simultaneous Multiple Round Auction (SMRA) and the single-stage or two-stage Combinatorial Clock Auction (CCA) can be seen as simple heuristic algorithms to solve this problem. Such heuristics do not necessarily compute the optimal solution, even if bidders are truthful. We study the average efficiency loss that can be attributed to the simplicity of the auction algorithm with different levels of synergies. Our simulations are based on realistic instances of bidder valuations we inferred from bid data from the 2014 Canadian 700 MHz auction. The goal of the paper is not to reproduce the results of the Canadian auction but rather to perform “out-of-sample” counterfactuals comparing SMRA and CCA under different synergy conditions when bidders maximize payoff in each round. With “linear” synergies, a bidder's marginal value for a license grows linearly with the total number of licenses won, while with the “extreme national” synergies, this marginal value is independent of the number of licenses won unless the bidder wins all licenses in a national package. We find that with the extreme national synergy model, the CCA is indeed more efficient than SMRA. However, for the more realistic case of linear synergies, SMRA outperforms various versions of CCA that have been implemented in the field including the one used in the Canadian 700 MHz auction. Overall, the efficiency loss of all ascending auction algorithms is small even with high synergies, which is remarkable given the simplicity of the algorithms.     

Optimal reserve prices in sealed-bid auctions with reference effects

The reference effect and loss aversion are incorporated into the buyer’s utility in the symmetric independent private value models of sealed-bid auctions. The buyer’s equilibrium bidding strategy and the seller’s optimal reserve price are derived for the first-price and second-price sealed-bid auctions. In both auction mechanisms, the seller’s optimal reserve price and expected revenue are increasing in the reference point. We compare the seller’s expected revenues as well as the optimal reserve prices in the two auctions. The results show that the seller will set a higher optimal reserve price but obtain lower optimal expected revenue in the second-price auction compared to the first-price auction. Further, we extend the model to the gain-seeking case, and endogenize the reference point as the ex-ante expected price of the item in equilibrium. In contrast to the loss-averse case, the seller will set a lower optimal reserve price but obtain higher optimal expected revenue in the second-price auction compared to the first-price auction if the buyers are gain-seeking. With an endogenous reference point, similar results are obtained in terms of revenue comparison between the two auctions.     

Estimating demand from eBay prices

This paper presents results for identification and estimation of the value distribution from eBay auction prices. The paper presents results for eBay type auctions with independent private values and unobserved participation. It is first shown that the distribution of values is identified from observing the distribution of prices and knowing the distribution of potential bidders. The main identification result presents conditions for which the distribution of values and the distribution of potential bidders are simultaneously identified. Not surprisingly, the intuition is similar to the standard results for identifying demand from observed equilibrium prices. The estimation method suggested by the identification results is used to estimate the value distribution for the “C5” Chevrolet Corvette sold on eBay. The results suggest that a simple OLS model on prices will over estimate the mean value of the item. The estimation results are then used to calculate the optimal reserve price for these cars. The estimated optimal reserves are compared to the actual reserves. Actual hidden reserves are set much higher than actual non-hidden reserves. The evidence suggests sellers set Buy-It-Nows and hidden reserves optimally to account for re-listing opportunities.     

Spectrum Auction Design

Spectrum auctions are used by governments to assign and price licenses for wireless communications. The standard approach is the simultaneous ascending auction, in which many related lots are auctioned simultaneously in a sequence of rounds. I analyze the strengths and weaknesses of the approach with examples from US spectrum auctions. I then present a variation-the combinatorial clock auction-which has been adopted by the UK and many other countries, which addresses many of the problems of the simultaneous ascending auction while building on its strengths. The combinatorial clock auction is a simple dynamic auction in which bidders bid on packages of lots. Most importantly, the auction allows alternative technologies that require the spectrum to be organized in different ways to compete in a technology-neutral auction. In addition, the pricing rule and information policy are carefully tailored to mitigate gaming behavior. An activity rule based on revealed preference promotes price and assignment discovery throughout the clock stage of the auction. Truthful bidding is encouraged, which simplifies bidding and improves efficiency. Experimental tests and early auctions confirm the advantages of the approach.     

Opportunity cost of spectrum for mobile communications: Evaluation of spectrum prices in Bangladesh

The purpose of this study is to focus attention on the choice of the reserve price made by the telecom regulator to explain the under-performance of spectrum auctions in Bangladesh. The overestimation of the reserve prices for several auctions in Bangladesh including the auction of the 3G – 2100 MHz band in 2013, the 2G - 1800 MHz band in 2015, and the 4G – 1800 and 2100 MHz bands in 2018 is established by comparing the reserve prices to the economic value of spectrum, determined by using the engineering valuation and production function approaches. Possible reasons for the inflation of the reserve price are discussed.     

Recent progress in the empirical analysis of multi-unit auctions

I discuss some recent progress in the empirical analysis of multi-unit auctions, through which a number of important commodities (Treasury bills and bonds, electricity, emission permits, monetary infusions by some central banks) are allocated.     

The 5G spectrum auction in Chile

In 2021, the Chilean government implemented a first-price package auction to allocate electromagnetic spectrum for 5G mobile services. The auction was run sequentially for different spectrum bands, allowing firms to exploit band complementarities. It was a combinatorial auction, so firms could bid for any combination of blocks within a band. It contemplated spectrum caps – upper limits on the spectrum for each firm – to ensure competitiveness. The beauty contests used in previous processes became obsolete, as there was a need to promote competitiveness and transparency in the telecommunication sector. Four incumbents and one potential entrant participated in the auction. The auction raised more than USD $450 million, which was six times more than the sum of the revenues of all previous contests in the country. We discuss this experience and show how different aspects of the context justified our design choices.     

Last-Minute Bidding in Sequential Auctions with Unobserved, Stochastic Entry

Using a dataset of calculator auctions on eBay, we first show that last-minute bidding is not merely the result of bidders’ going to the next-to-close auction. Instead, bidding is concentrated at the end of the period in which the auction is the next to close, suggesting the existence of strategic last-minute bidding. Then, we model repeated, ascending price auctions for homogeneous goods with unobserved, stochastic entry. We show that the dynamic game has a pure-strategy symmetric equilibrium in which entrants always reveal themselves by bidding in the auction in which they arrive, and bidding occurs at the last minute.     

The value of 1800 MHz and 2100 MHz spectrums in India and implications for auction design

The sequence of events leading up to the upcoming auction of 1800 MHz spectrum in India has led to the auctions acquiring an extraordinary significance for the future of the Indian mobile industry. A key feature of the auction design proposed by the regulator TRAI is the benchmarking of the reserve price of 1800 MHz to the price of 2100 – 3G spectrum revealed in the 2010 auction. In the context of the low number of LTE devices available and the fragmentation in the 1800 MHz band, this paper proposes reducing the duration of spectrum holding to ten years (from the current level of twenty years), and calibrating the reserve price of 1800 MHz with its value with GSM deployment. An economic model is used to compute the value of startup and incremental 1800 MHz spectrum. The estimated values are shown to differ from the value of 2100 MHz spectrum at a pan-India level and also in their distribution across circles. A new set of reserve prices are computed based on the estimation. The estimated values are also shown to be close to the AGR-adjusted price revealed in the 2001 auction. A reserve price based on the 2001 auction is also provided. Concomitant features of the auction are suggested to give coherence to the auction design.     

Bids for the UMTS system: An empirical evaluation of the Italian case

The aim of this paper is to analyze the licensing of the telecommunication spectrum as a public good and the search for equilibrium prices through bilateral bargaining and multilateral bidding. It develops a general model of price setting under dynamic uncertainty and applies it to the Italian auction for Universal Mobile Telecommunications System (UMTS). The empirical application shows that the model can be used both to determine the base price as well as other desirable characteristics to organize an auction and to better understand, after the auction is closed, what really happened in terms of the critical factors involved. After recalling some basic concepts on spectrum rights and reviewing the general experience with UMTS auctions in Europe, the formal model and its application are presented. The results confirm certain views on the Italian auction, which are widely shared but were never tested before, namely that: (i) given the initial price, the number of licenses offered for the bidding should have been fewer, or alternatively, (ii) given the number of licenses, the base price should have been higher and (iii) the main bidder underpaid for the license. The model also allows us to quantify the bidders’ reservation price and the State and the bidders’ implicit bargaining powers.     

Raising rivals’ cost in multi-unit auctions

The objective many telecom regulators want to achieve when they decide to auction spectrum is that acquiring firms pay a market price (based on the opportunity cost principle). The simultaneous ascending auction may fail in this respect, as it provides bidders with an opportunity to engage in strategic demand reduction. This paper asks whether the combinatorial clock auction (CCA) fares better in this respect. We show that the answer to this question depends on the objectives bidders have. If bidders have only the slightest preference to raise rivals’ cost, they will use the opportunities the CCA provides them to engage in strategic demand expansion. This is even the case when the clock phase ends with excess demand.     

Market Mechanism Choice and Real Estate Disposition: Search Versus Auction

We propose a model of mechanism choice in the disposition of real estate assets where we consider two alternatives: a search market and an auction. Within the search framework, we derive an equilibrium whereby buyers incur search costs and sellers incur holding costs for the period during which the property is not sold. In the auction alternative, the seller joins an existing pool of sellers in undertaking a multiple–object auction and pays a commission upon sale. Buyers and sellers freely choose their mechanisms, which in equilibrium are optimal given each group's conjectures about the mechanism choice of their counterpart. In equilibrium, an agent cannot benefit from deviating from his choice and each agent's beliefs are consistent with the equilibrium outcome. It is shown that (a) buyers with high search costs will choose auctions because the auction payoff imposes an upper bound on buyers' gains from search, and (b) prices at auctions will be higher. Using vacant lot sales data and a method–of–moment estimator which accounts for the presence of an endogenous discrete mechanism choice variable, we estimated a hedonic regression to detect the price effect. It was determined that, on average, lots sold for $1.44 per square foot more in auctions than in the search market, as predicted by our model.     

Auctions with quantity externalities and endogenous supply

This paper studies the design of license auctions when the number of licenses allocated in the auction determines structure of the post-auction market. I first show that a sequence of conditional reserve prices that specify minimum acceptable bid at each supply level can be used to determine supply endogenously. Then I construct a static auction called multi-dimensional uniform-price auction and a dynamic auction called Walrasian clock auction that allow the auctioneer to condition reserve price on supply and allow bidders to condition bids on supply. I show that both proposed auctions can implement the efficient market structure that maximizes total surplus in the post-auction market in a dominant strategy equilibrium. I next characterize the optimal auction and show that the two proposed auctions can yield the optimal revenue under a sequence of optimal reserve prices.     

How Do Land Auction Formats Influence the Market Structure and Aggregate Surplus of Real Estate Development?

This article proposes a duopoly dynamic game theoretic model to investigate the market structure and aggregate surplus of real estate development when land is sold in a sealed‐bid first price auction vis‐à‐vis an open English auction. It relaxes the assumption of symmetric bidders. The land values have common value and private value components. We find that the sealed‐bid first price auction introduces competition in the real estate development market. The open English auction leads a monopoly market. State agencies are recommended to increase the aggregate surplus of real estate development by publishing past bidding information under the sealed‐bid first price auction and reducing information asymmetry between bidders.     

Auction Platform Design and the Linkage Principle

This paper examines an auction platform in which the monopoly platform maximizes profits by adjusting participation fees and choosing an auction format. The seller has private information on the quality of the good, and each participating buyer receives a private signal about his valuation of the good. The choice of auction format determines the allocation of trading surplus between the seller and buyers. This paper shows that when the seller's type is affiliated with the buyers' signals, the platform can charge higher participation fees on both sides by choosing a first‐price or descending auction than a second‐price or ascending auction.     

Effects of Minimum Bid Increments in Internet Auctions: Evidence from a Field Experiment

I study the role of minimum bid increments (MBI) in internet auctions using natural field experiment data from a Finnish online auction site Huuto.net. Internet auctions are typically viewed as second‐price, implying truthful bidding. However, due to the presence of the MBI, equilibrium bidding behavior involves bid shading. I sell otherwise identical gift cards while varying the MBI and test between truthful bidding and bid shading. Truthful bidding is rejected. Consistently with equilibrium behavior, bid shading increases as the MBI increases. This implies that real‐world bidders are capable of reacting rationally even to seemingly small variations in their strategic environment.