A Multinomial Logit-based Statistical Test of Association Football Betting Market Efficiency

Henri Nyberg

Research output: Contribution to journalArticleScientific


This paper presents a novel statistical test for the informational efficiency of bookmakers' odds in the association football betting market. The new test has several advantages in terms of interpretation and statistical properties over the previously considered match outcome-specific tests. According to simulation experiments, the empirical size of the classical likelihood-based test statistics is satisfactory, whereas the power is dependent on the degree of market inefficiency and the sample size. A large English football dataset reveals that the efficient market hypothesis cannot be statistically rejected except for the Premier League games.
Original languageEnglish
JournalHECER discussion papers
Pages (from-to)1-35
Number of pages35
Publication statusPublished - Jun 2014
MoE publication typeB1 Journal article

Fields of Science

  • 511 Economics

Cite this