Why the Oxford study’s nominal 324% cost overrun is under scrutiny

There are two figures in circulation about the Olympic Games and both deserve examination. The first is the cost overrun: a much-cited University of Oxford study put the average deviation from Olympic budgets between 1960 and 2012 at 179% in real terms and 324% in nominal terms. The second is the benefit: the economic boost that justifies the bid.

What makes this case interesting is that both have been challenged with data. A re-analysis published in 2022 could not replicate the Oxford study’s results for the period 2000-2016 and obtained cost overruns of 13% to 178% on direct investments, noting as well that the cost of the Games represents less than 0.28% of the host country’s GDP. And the benefit, when measured against the appropriate counterfactual — the cities that bid for the Games and lost — turns out to be 3 to 4 percentage points of regional GDP in the event year and the year before, with long-run effects that are not statistically robust.

And one specific case no discussion should omit: Sochi 2014 cost $55 billion, 4.5 times the bid budget, with 96.5% public funding and a subsequent annual burden of around $1.2 billion.

Summary

It is advisable to verify for every single row whether the long term or the event year is under evaluation, alongside checking if the focus lies on the Summer or Winter Games. These represent four entirely separate scenarios, and their outcomes fail to align.

What you will find here

• The world’s most-cited cost overrun figure and why a re-analysis could not reproduce it.

• What happens when the host city is compared against the city that lost the vote.

• Why the Summer and Winter Games produce opposite results.

• What is left of the tourism legacy when it is measured with air passenger data.

• The Sochi case, with all its figures.

• Despite what the literature suggests, municipalities continue to submit bids.

Disputing the cost overrun amount

Start with the one everybody cites. Sascha Schmidt summarised it in 2017 — in a book chapter, not a peer-reviewed article — with what has become the standard formulation: between 1960 and 2012, the Olympic Games consistently overran budgets, on average by 179% in real and by 324% in nominal terms, as researchers from the University of Oxford found out.

His complementary observation describes the state of the debate well: although politicians and interest groups continually justify the overarching costs by promising future economic benefits, economists have a hard time finding any economic impact from hosting the Games when engaging in ex post analysis of indicators like GDP, employment rate and trade volume.

Back in 2022, Holger Preuss from Johannes Gutenberg University Mainz re-evaluated that figure within the pages of the International Journal of Sport Policy and Politics. It is important to be clear regarding the exact nature of this piece: a commentary and methodological dispute, rather than an independent measurement, while his extensive background working inside the Olympic realm forms part of the background to his stance.

That said, his objections are concrete and checkable. He starts from the fact that the so-called Oxford study shows cost overruns of 170% to 750% with a substantial risk of further overruns in future, and sets out to re-analyse them using a method in accordance with the international convention of comparing across time and geographies, with a dataset containing abundantly more governmental data.

His results: he could not replicate the Oxford study’s results for the period 2000-2016. His own calculation yields, for the Games from 2000 to 2014 and for direct non-organising-committee investments, cost overruns in the range of 13% to 178%. And he adds the figure that sets the discussion in macroeconomic scale: the costs of the Olympic Games represent less than 0.28% of GDP; hence it is clear that the Games cannot cause a severe national economic crisis.

Preuss also identifies six typical «misperceptions» in this debate and disputes the assumption that host cities give a blank cheque to the International Olympic Committee, arguing that the demands are fixed in the Host City Contract. That the two estimates differ so widely — 170-750% against 13-178% — does not mean one is false and the other true: it means the answer depends on which line items are counted as a cost of the event. And that accounting decision is, in practice, the heart of the controversy.

Comparing against whoever lost the vote

The central methodological problem of this literature is obvious once it is named: the cities that host Olympic Games are not just any cities. They bid because they have ambition, growth, fiscal capacity and political support, and all of that affects their economy in its own right. Comparing a host against the national average credits the Games with what was in fact the city’s profile.

The elegant solution exists and two studies have used it: comparing against the cities that also bid and lost. They are cities with the same ambition and the same profile, minus the event.

Matthias Firgo applied it in Regional Science and Urban Economics in 2021 to examine both immediate and long-term impacts on the regional economy, utilizing runners-up in the Olympic bidding process to construct the counterfactual for Olympic host regions.

His results, by type of Games:

The final row corresponds to Stephen Billings and Scott Holladay, who published in Economic Inquiry in 2012 the toughest study in this group. Their design: matching Olympic host cities with cities that were finalists for the Games but were not selected by the International Olympic Committee, and applying a difference-in-differences estimator to examine post-Olympic impacts between 1950 and 2005.

Their outcome, literally: regression results provide no long-term impacts of hosting an Olympics on two measures of population, real gross domestic product per capita and trade openness.

It is important to read exactly what that means and what it does not. It does not say the Games have no effect: it says that twenty or fifty years later, the host city is indistinguishable from the city that lost the vote on those four variables. The structural economic legacy, measured that way, does not appear.

What the most recent measurement does find

David Boto-García and María Santana-Gallego published in Tourism Management in 2025 the analysis with the most modern econometric apparatus: a panel from 1991 to 2023 with dynamic difference-in-differences estimators that distinguish anticipation from legacy effects, account for staggered treatment timing and accommodate heterogeneous effects.

Their outcomes outshine previous ones and ought to be documented: moderate anticipatory effects on tourism and notable impacts on employment and GDP per capita, with the Summer Games exhibiting a stronger impact.

The authors frame their contribution by flagging precisely the field’s problem: empirical evidence on the overall impact remains inconclusive, largely due to methodological challenges, and they highlight the limitations of standard approaches. That a better method yields a more positive result is relevant and has to await replication.

The tourism legacy

The tourism argument is the second pillar of any bid: the world sees the city and then comes to visit it. Steven Moss, Kathleen Gruben and Janet Moss tested it in Journal of Policy Research in Tourism, Leisure and Events in 2019 with the most direct data possible: monthly international air passenger arrivals for nine Olympic host cities, with a time series methodology controlling for problems noted in prior studies — pre-existing trends, external events, seasonality and crowding out.

Their result: the results show little evidence to support the legacy effect in the short run.

Two studies focusing on Brazil alongside major events in general moderate that conclusion more favourably. In 2018, Roberto Meurer and Hoyêdo Nunes Lins examined the impacts of the 2014 World Cup and the 2016 Games on Brazilian international tourism receipts within Tourism Economics, whereas Christian Vierhaus estimated the international tourism outcome of staging Olympic Games and FIFA World Cups in the exact same journal that year. Measurable tourism impacts are highlighted by their findings, which remain entirely compatible with a lack of legacy reflected in airport statistics, given that an outcome can manifest in revenue rather than passenger numbers, or simply concentrate during the event itself.

Eleni Kasimati and Peter Dawson provided the temporal framework for Athens 2004 that encapsulates this entire body of work, featuring an aggregate macroeconometric model released in Economic Modelling in 2008. Their research uncovers evidence indicating that the Olympics can successfully stimulate the host city’s economy by producing gains that exceed the expenses of preparation. Furthermore, they append the concluding note: whilst the impact effects are quite strong during the preparation phase and the year the Games took place, the long-term economic legacy effects appear to be quite modest.

Strong during preparation and the event, modest afterwards: that is precisely the profile Firgo discovers when employing finalists as the counterfactual, aligning seamlessly with the long-run zero established by Billings and Holladay. Three separate studies, utilizing three distinct approaches, yield the exact same curve.

Sochi, with all its figures

Martin Müller published the financial records of the Sochi Winter Games in Eurasian Geography and Economics back in 2014, establishing what remains the most comprehensively documented instance across the entire academic literature. It is worthwhile to present these numbers plainly and without embellishment:

And the balance of outcomes: the benefit from this high cost is limited. Extensive construction led to hotel overcapacities, investors defaulted on state-backed loans, and there is no coherent plan for the after-use of venues and some of the largest infrastructure projects.

The annual burden Müller quantifies includes operation, maintenance, and foregone interest and tax revenue. And the two non-economic effects usually invoked do not appear either: the event also did not manage to improve the image of Russia in the world, and among the domestic population support dropped over the seven years of its implementation, most notably among the local population.

That final remark warrants emphasis. Backing dropped steepest among residents of the host city. The alleged civic enthusiasm, typically deployed as a fallback justification when financial rationales collapse, actually produced the opposite effect.

What the literature says as a whole

Two reviews allow everything above to be placed. The most cited is by Robert Baade and Victor Matheson in Journal of Economic Perspectives in 2016, and it sorts costs and benefits into three categories each. The costs: general infrastructure such as transportation and housing, specific sports infrastructure required for competition venues, and operational costs including general administration as well as the opening and closing ceremony and security. The benefits: the short-run benefits of tourist spending during the Games, the long-run «Olympic legacy» — infrastructure improvements and increased trade, foreign investment or subsequent tourism — and intangible benefits such as the «feel-good effect» or civic pride.

Their conclusion is the most-quoted sentence in the field: the overwhelming conclusion is that in most cases the Olympics are a money-losing proposition for host cities; they result in positive net benefits only under very specific and unusual circumstances. And they add a relevant asymmetry: the cost-benefit proposition is worse for cities in developing countries than for those in the industrialised world.

The 2018 study by Pasquale Scandizzo and Maria Rita Pierleoni, published in Journal of Economic Surveys, broadens the perspective on the Games by viewing them as cultural assets and catalysts for endogenous growth, while also reviewing various methodologies, concepts, and approaches. According to the authors, their broad conclusions are controversial, with some hints of positive overall effects, but also with a well-documented tendency to exaggerate the benefits and underestimate the costs in the ex ante versus the ex post studies.

And one observation that opens a door: ex post cross-country econometric studies tend to catch sizeable differential and persistent benefits ignored by individual studies. That is to say, the method of measurement partly determines the result, just as it did with the cost overrun.

Sorted by type of question, the literature turns out to be considerably less contradictory than it looks at first glance.

What that table shows is that the real disagreement concentrates in a single cell — the magnitude of the cost overrun — and in the exception of the 2025 measurement. On everything else, twenty years of work using different methods point to the same curve: a peak in the event year and little or nothing afterwards.

Why cities keep bidding

If the overwhelming conclusion is that it is a money-losing proposition, the obvious question is why cities compete for it. Baade and Matheson pose it explicitly in closing their review: they discuss why what looks like an increasingly poor investment decision on the part of cities still receives significant bidding interest, and whether changes in the bidding process of the International Olympic Committee will improve outcomes for potential hosts.

Schmidt points the same way in noting that politicians and interest groups justify the overrunning costs with promises of future benefits. And Müller supplies the fact that changes the politics of this question: in Sochi, public support fell over the seven years of implementation, and more sharply among local residents. Preuss, in his 2018 article in the same journal, had proposed a legacy measurement framework precisely because, in his terms, while the sport organisations gain large revenues, cities lose, and the referendums held in the preceding years returned negative results consistently.

Put differently: the economic literature and the electorate of candidate cities have been in agreement for a decade, and bids are decided somewhere else.

One point of fairness is owed to the bid documents, because the literature does not say the Games do nothing. It says the Games produce a large, short, concentrated burst of activity and then return the host to the trajectory it was already on. For a city with a specific piece of infrastructure it wanted anyway — a metro line, a cleaned-up waterfront, a transport interchange — an Olympic deadline is a way of getting it built in seven years instead of twenty, and that is a real benefit that none of these studies measures, because it is not an Olympic effect but an acceleration effect.

The honest version of the bid, then, is not «the Games will pay for themselves». It is «we are buying a deadline, and here is what the deadline will deliver», with the overrun priced in at whichever end of the 13-to-178 per cent range the city can absorb. That is a defensible case. It is also a much harder one to put to a referendum, which may be part of why the referendums kept failing.

How to handle this, practically speaking

• Always ask which line items are in the cost. The difference between a 170-750% overrun and a 13-178% one is largely an accounting decision about what gets attributed to the event.

• Demand comparison against the finalist cities. It is the only honest counterfactual, and with it the long-run effect disappears in population, GDP per capita and trade openness.

• Do not treat summer and winter as the same case. The Summer Games give +3 to +4 points in the event year; the Winter Games, no positive impact and possibly a temporary decline.

• Discount the tourism legacy. Air passenger data from nine host cities show little evidence supporting the short-run effect.

• Budget for subsequent maintenance, not just construction. Sochi carries around $1.2 billion a year in operation, maintenance and foregone revenue.

• Distrust pre-event studies. The 2018 review documents a tendency to exaggerate benefits and underestimate costs in ex ante studies relative to ex post ones.

Frequently asked questions

Are the Olympic Games profitable?

According to the most-cited review, no: in most cases they are a money-losing proposition for host cities, and they produce positive net benefits only under very specific and unusual circumstances. The proposition is also worse in developing countries.

To what extent do budgets get exceeded?

It depends what is counted. The Oxford study puts the 1960-2012 average at 179% real and 324% nominal; a re-analysis with more governmental data could not replicate it for 2000-2016 and obtained 13% to 178% on direct investments.

Does it provide a genuine economic stimulus?

Indeed, and with distinct limits. Relative to runner-up cities, the Summer Games elevate regional per capita GDP by 3 to 4 percentage points higher than the nationwide figure during both the hosting period and the preceding year. The enduring impact remains positive, although lacking robust statistical significance.

And twenty years later?

The study matching hosts with finalists between 1950 and 2005 finds no long-term impact on two measures of population, real GDP per capita or trade openness.

Are visitor numbers higher later on?

Monthly international air passenger arrival data for nine host cities show little evidence to support the short-run legacy effect. Other studies detect effects on travel receipts, which is not the same variable.

What happened in Sochi?

It cost $55 billion against the bid’s $12 billion, with 96.5% public money, left hotel overcapacity, defaults on state-backed loans and no coherent after-use plan, and carries around $1.2 billion a year in burden.

What about civic pride?

It is the argument of last resort and in the best-documented case it behaved the other way round: support dropped over the seven years of implementation, most notably among the local population.

What we do not know

• What the correct cost overrun is. The two estimates differ by an order of magnitude and the discrepancy is an accounting one: which general infrastructure items are attributed to the event.

• Why the more recent methods give more positive results. The 1991-2023 panel finds notable impacts on employment and GDP per capita where the finalist-based designs see none in the long run.

• Whether there is a city-size threshold below which it never pays. The asymmetry between developed and developing countries is documented; the threshold is not.

• What the intangible benefit is worth. Baade and Matheson include it as a category and nobody has quantified it comparably against the costs.

• Whether shifts in the International Olympic Committee’s bidding procedure enhance the ultimate result. That remains the inquiry concluding the 2016 evaluation, and it is still empirically unresolved.

References

1. Baade, R. A. and Matheson, V. A. «Going for the Gold: The Economics of the Olympics». Journal of Economic Perspectives, 2016. 10.1257/jep.30.2.201

2. Scandizzo, P. L. and Pierleoni, M. R. «Assessing the Olympic Games: The Economic Impact and Beyond». Journal of Economic Surveys, 2018. 10.1111/joes.12213

3. Firgo, M. «The causal economic effects of Olympic Games on host regions». Regional Science and Urban Economics, 2021. 10.2139/ssrn.3500856

4. Billings, S. B. and Holladay, J. S. «Should Cities Go for the Gold? The Long-Term Impacts of Hosting the Olympics». Economic Inquiry, 2012. 10.1111/j.1465-7295.2011.00373.x

5. Preuss, H. «Re-analysis, measurement and misperceptions of cost overruns at Olympic Games». International Journal of Sport Policy and Politics, 2022. Commentary article. 10.1080/19406940.2022.2037685

6. Schmidt, S. L. «Hosting the Olympic Games». 2017. Book chapter. 10.1007/978-3-319-47458-8_51

7. Müller, M. «After Sochi 2014: costs and impacts of Russia’s Olympic Games». Eurasian Geography and Economics, 2014. 10.1080/15387216.2015.1040432

8. Boto-García, D. and Santana-Gallego, M. «Old questions, new methods: Revisiting the economic effects of hosting mega-sport events». Tourism Management, 2025. 10.1016/j.tourman.2025.105324

9. Moss, S. E., Gruben, K. H. and Moss, J. «An empirical test of the Olympic tourism legacy». Journal of Policy Research in Tourism, Leisure and Events, 2019. 10.1080/19407963.2017.1418750

10. Kasimati, E. and Dawson, P. «Assessing the impact of the 2004 Olympic Games on the Greek economy: A small macroeconometric model». Economic Modelling, 2008. 10.1016/j.econmod.2008.06.006

11. Vierhaus, C. «The international tourism effect of hosting the Olympic Games and the FIFA World Cup». Tourism Economics, 2018. 10.1177/1354816618814329

12. Meurer, R. and Lins, H. N. «The effects of the 2014 World Cup and the 2016 Olympic Games on Brazilian international travel receipts». Tourism Economics, 2018. 10.1177/1354816617746261

This piece synthesizes peer-reviewed academic research regarding the financial consequences of large athletic tournaments, specifically highlighting the referenced book chapter and commentary piece. The data reflect the specific editions and timeframes investigated and cannot be generalized to other happenings. Furthermore, this content does not constitute public policy guidance, nor does it serve as an assessment of any particular candidacy.