Winner's curse: can we avoid the ‘curse of the winner’?

Texte

The ‘Winner's curse’ arises mainly in four situations in sports economics: when candidates for the award of a mega-event overestimate its value to win at the expense of their competitors, which results in cost overruns; when North American cities try to attract a major league franchise from another territory by means of subsidies or tax exemptions, this relocation being generally irrelevant from an economic standpoint; when television channels try to acquire broadcasting rights for competitions by paying sums higher than their real value; when teams outbid each other to buy players at too high a price for their financial capabilities [Andreff, 2014]. We will analyse the first case, namely the winner's curse of the awarding of the Olympics. Three other similar ‘cursed’ situations are developed throughout the book (see in particular 'Sports arenas', 'Televised sport', 'Bosman ruling', 'Club deficits', 'Transfers').

The centralised award procedure for the Olympic Games: auctions and information asymmetry

The winner's curse reflects a recurring phenomenon in international sport with the overrun of the costs foreseen for the organisation of the Olympic Games (or any other mega-sporting event), mainly in terms of capital expenditure. The difference between ex-ante anticipated costs (before the IOC vote) and ex-post observed costs (at the end of the Games or later) is due to the bidding and auction process itself when there are several cities in the running to host the event. Each candidate city is obliged to bid for the event with a more ambitious project than its competitors (the auction principle) and accept to pay a higher cost than expected. The economic analysis of the IOC's method of awarding the Games shows that cost overruns are inevitable and that a deficit is likely.

The IOC is the owner of the Olympic Games, and as such, it defines the specifications of the competition imposed on the candidates: technical characteristics of the sports infrastructure (swimming pool, athletics stadium, dojo, skating rink, etc.) and non-sports infrastructure (airports, telecommunications systems, motorways, tramways, etc.). The IOC alone has the power to entrust its organisation. It aims to offer the best possible Olympic Games to its direct and indirect clients (media, sponsors, spectators, and television viewers), without worrying about the eventual oversizing of Olympic facilities concerning the needs of the local population. The latter will have to bear the long-term costs (investment, maintenance) without having the use of the said facilities, which are calibrated for high-level practice.

In a process in which the object of the auction has an uncertain value (no one knows the actual market value of a designation as a host city), the winner is the one who has overestimated the value of the object and thus won the tender (as the highest bidder) by outbidding all the other competitors. The winner of the auction loses financially, as his final bid exceeds the real value of the object won [Thaler, 1988].

Firstly, this situation corresponds to an anti-selection in information asymmetry, a cause of 'market failures' and inefficiency, since this phenomenon prevents the development of relationships between the IOC and the host city that are mutually beneficial. This is because the IOC almost systematically chooses the best project for itself, i.e., the most grandiose and therefore the most expensive. The IOC seldom chooses the cheapest project, as reducing the investment cost of the Olympic Games is not an objective for the IOC since it does not pay for them. For example, the file presented by Madrid received the fewest votes from IOC members for the 2020 Games with a "low cost, reasonable and responsible" project.

The IOC naturally has an interest in the bidding of the candidate cities against each other. And it is also in the interest of these same cities to outbid each other to increase their chances of winning. The chosen city is the one that offers the best project. The 'best deal’ in terms of social benefits - generally overestimated - and financial costs - generally underestimated - [Andreff, 2012]. Anti-selection is in full effect: most expensive Olympic games, uncontrolled investment budgets and very high deficit risk. Other indications reflect the existence of the winner's curse: delays in the works programme thus generating price increases from contractors, financial extensions, lobbying and corruption.

Secondly, the appearance of the winner's curse in an auction can be explained by the opportunistic behaviour of candidate cities. On the one side, they know their project perfectly well and deliberately underestimate the cost of the infrastructure relative to the actual costs so as not to 'weaken' the project with prohibitive costs and gain acceptance from the population, the contractors and the voters. For the same reasons, they systematically overestimate the positive economic impact expected from the Olympic Games: in reality, the economic impact for the country of the host city is very low, and at worst, zero or even less. Candidate cities also hide negative externalities: the possible absence of taxpayers' willingness to pay contributions to the Games; the opportunity cost of the Games; the crowding-out effect with less frequentation by regular tourists fearing the saturation of the host city; gentrification; land speculation and an increase in property prices as well as increased pollution and traffic jams, etc.

On the opposite side, the IOC is not in a position to verify the reliability of the information provided. Exchanges between IOC officials and candidate cities or visits to the proposed Olympic venues are insufficient to reduce this information asymmetry. Moreover, the IOC members' vote is based more on their personal and political judgement of the bids than on their technical and financial aspects.

The winner's curse operates when the successful auction bid generates an abnormal or even negative return on investment, contrary to the theory of rational investment choice. The most significant examples of the negative consequences of the winner's curse of the auction are probably, and chronologically: Montreal (1976), Athens (2004), Beijing (2008), Sochi (2014) and Rio (2016)

The extent of the winner's curse should be assessed with an - ex-ante and then ex-post - economic cost-benefit analysis, using the same methodology. Further analysis should be carried out systematically and in addition to the cost-benefit analysis during the pre-bid study phase. The opportunity cost should be approached by calculating what other investments the host city could have made with the same amount of money spent on the Games, to verify whether alternative projects would not be more socially useful (i.e., with a higher net social benefit) for the well-being of the population (schools, hospitals, social centres, etc.).

Cost overruns as a sign of the winner's curse

However, in the absence of cost-benefit analysis and opportunity cost calculations for the Olympic Games, one indicator of the 'curse of the auction winner' is the underestimation of costs and its inevitable consequence of cost overruns. Cost overruns are systematic as soon as there are several bids (from 2 to 6 bids depending on the Olympiad, except in 1984, when there was only one). On average, for the summer and winter Olympics from 1960 to 2016, the ex-ante/ex-post cost overrun rate is 172 % [Flyvbjerg, Budzier and Lunn, 2020]. A study of all the Summer Olympics from 1972 to 2012 and all winter Olympics from 1980 to 2014 showed that the winner's curse is the rule with one exception, the 1984 Summer Olympics in Los Angeles [Andreff, 2015].

For the Summer Olympics, the overrun rates range from 32 % to 1130 %: 32 % for Atlanta in 1996, 93 % for Sydney in 2000, 108 % for Seoul in 1988, 109 % for Athens in 2004, 127 % for London in 2012, 156 % for Barcelona in 1992, 171 % for Munich in 1972, 247 % for Rio in 2016, 385 % for Montreal in 1976 and 1130 % for Beijing in 2008 [Andreff, 2015; Andreff, 2012].

For the winter Olympics, the magnitude of the overruns is admittedly smaller, but remains very high, ranging from +17 % to +495 %:

+17 % for Vancouver in 2010, +29 % for Salt Lake City in 2002, +56 % for Nagano in 1998, +59 % for Calgary in 1988, +82 % for Turin in 2006, +135 % for Albertville in 1992, +173 % for Sarajevo in 1984, +201 % for Grenoble in 1968, +277 % for Lillehammer in 1988, +321 % for Lake Placid in 1980 and +495 % for Sochi in 2014 [Andreff, 2015; Flyvbjerg, Budzier and Lunn, 2020].

The only exception was the 1984 Olympic Games, which did not result in any investment cost overruns and therefore made a profit for the organising committee. Los Angeles, being the only candidate, did not have to overbid and underestimate its expenses to be chosen by the IOC due to the lack of interest from other cities after the financial disaster of the Montreal Olympics (1 billion dollars of deficit financed by the taxpayer from 1976 to 2006). In addition to the absence of competition, and therefore of bids and overbidding, it should be noted that Los Angeles had the necessary sporting and non-sporting infrastructure to host the Games even before its candidacy.

For the 2024 and 2028 Summer Olympics, several cities have withdrawn their bids - Hamburg, following a negative referendum, Boston, Budapest, Rome and Toronto following a decision by the public authorities – as a result, the IOC awarded the 2024 Olympics to Paris and the 2028 Olympics to Los Angeles. In fact, and a priori, the conditions for the emergence of a winner's curse have not been met. A first indication lends credence to this hypothesis: the ex-ante cost of the Paris 2024 Olympics is 6.8 billion euros (the cheapest since the Sydney Olympics in 2000), with a risk of over-costing identified by the General Inspectorate of Finance of 500 million euros, i.e., a virtual overrun rate of 7 %. These figures seem rather reasonable, if they are confirmed ex-post, compared to the previous five editions of the Games: Tokyo (€ 13 billion, December 2020 figure not definitive taking into account the additional costs generated by the postponement of the Games to 2021 due to the COVID-19 pandemic.), Rio de Janeiro (€ 16 billion, 2016), London (€ 12 billion, 2012), Beijing (€ 32 billion, 2008) and Athens (€ 11 billion, 2004).

How to remove the winner's curse?

What solutions could eliminate the winner's curse or, at least, limit the additional costs [Andreff, 2015]? The first option would be to abandon the method of awarding the Olympic Games by auction with a competition between several candidate cities and drawing on the lessons of the Los Angeles Olympic Games in 1984. The event could always be organised in the same city (one for the Summer Olympics, another for the Winter Olympics). This would put an end to cost overruns and underestimates and would considerably reduce the necessary investments since there would no longer be any need to build the infrastructures imposed by the IOC in different cities for each Olympiad. Only the costs of maintenance, upkeep and modernisation of the sporting and non-sporting infrastructure would have to be borne. This is both a radical solution and a utopian one, as it would run counter to the commercial interests of all the sporting and economic actors who have much to gain from a change of location that would generate more turnover: the IOC and international federations, television channels, sponsors, marketing companies, equipment manufacturers, construction companies, tourism agencies, etc.

A very hypothetical alternative would be to let the IOC bear the entire cost of the Games. There would be no more winner's curse and no more socialised long-term costs. The IOC would continue to choose the venue and the budget of the Games would be calibrated to the level of its revenues (5-6 billion euros net on average per Olympiad). However, it is difficult to imagine that the IOC would agree of its own accord to give up the economic advantages of its monopoly on the supply chain of the Games.

Another solution would be for the IOC to announce that it would choose the project that, while respecting the specifications of the Games, would be the least expensive. This hypothesis seems unlikely, as it too would run counter to the IOC's desire to have the most beautiful Games possible every four years. This initiative would also be counterproductive, as it would encourage the candidate cities to underestimate the costs even more. This again would amplify the effects of the winner's curse of the auction. Instead, the introduction of a ceiling on expenditure, that cannot be exceeded, to curb the growing gigantism of the event would deprive the IOC of an economic criterion for selecting the host city.

More realistically, a compromise could change the rules for competitive bidding and the eligibility of candidates. Thus, the consequences of the winner's curse would be limited by a double restriction. A rotation by sufficiently narrow geographical zones (a dozen in the world) would considerably reduce the number of cities able to host the Games. A ban on bidding more than once in a century would also limit the effects of bidding. The IOC could thus frequently find itself with only one candidate city, a necessary configuration for the disappearance of the winner's curse [Andreff, 2013].

In 2019, the IOC, concerned about the scarcity of bids, adopted two reforms to the selection procedure, as part of its Agenda 2020, intending to try to reduce the cost of the bidding process and the organisation of the Games. A combined bid will be able to be carried by several cities, regions or countries (until now only one city was entitled to apply). A referendum will have to be held systematically in the candidate cities if the national legislation allows it. These measures will likely be insufficient to reduce the winner's curse.

Despite a tendency to systematically overrun costs, with chronic overspending of sporting infrastructures, recurrent social deficit, overestimation of economic by-products and more than questionable opportunity costs, the obvious question begs to be asked: Why do cities continue to bid to host the Olympic Games (or other mega-events)? According to the results of a study of five editions of the Olympic Games plus three football World Cups, the answer is not economic, but rather political (patriotism, belonging to the same community, electoralism), geopolitical (soft power), and psychological (sense of pride, happiness of the population) [Mitchell and Fergusson Stewart, 2015].