Re-analysis, measurement and misperceptions of cost overruns at Olympic Games

Abstract

The so-called ‘Oxford Study’ by Flyvbjerg et al shows Olympic cost overrun of 170-750% with a substantial risk of a further cost overrun above this range in the future. The aim of this paper is to re-analyse the results by using a method in accordance with the international convention of comparing across time and geographies. The dataset used has abundantly more governmental data than in the Oxford Study. The results of the Oxford Study for the period of 2000–2016 could not be replicated. It is argued why the statistical analysis and their data reliability are misleading, thus leading to a discovery of six typical misperceptions. In contrast to those authors, this paper shows cost overruns (Olympics 2000–2014) for direct non-OCOG investments in the range of 13–178%. Further, it shows that the costs of the Olympic Games represent less than 0.28% of the GDP; hence, it is clear that the Games cannot cause a severe national economic crisis. This paper provides a discussion on the underlying assumptions of the Oxford study, thus providing valuable insights for politicians who intend to submit an Olympic bid, and contextualises cost overruns caused by fixed deadlines and long planning horizons. I disagree with the claim by Flyvbjerg et al. that hosts give a blank cheque to the IOC and I contend that demands and a tight coupling are fixed in the Host City Contract, thus negating the blank cheque assertion.