Can sporting success in Norwegian football be predicted from budgeted revenues?

ABSTRACT

Research question: Although previous research has shown that there is an association between financial strength and sporting outcome, whereby teams with larger budgeted revenues typically perform better in sports than other teams, this was not supported by a newspaper article describing the 2012 season in Norwegian football. Drawing on the Norwegian football league over the period 2011–2013, this paper sets out to explore the association between financial strength and sporting outcome in detail.

Research methods: To examine fully the association between financial strength and sporting outcome, a wide array of different statistical methods is adopted, ranging from simple t-tests to regression analysis and fixed effects regression analysis.

Results and findings: A duality is present in the relationship between budgeted revenues and sporting outcome, as evidence is found suggesting that budgeted revenues are a significant driver of sporting outcome among the bottom-half teams but not among the top-half teams. Moreover, the static and dynamic regression models, as well as the fixed effects panel data models, support the notion of budgeted revenues being a driver of sporting outcome.

Implications: The duality in the results is also supported by the fixed effects models, indicating that competitive advantages other than financial advantages are relevant. An interpretation of these findings is that money is a significant driver of sporting success, but only to a certain extent (i.e. avoiding relegation). In other words, a focus on sports is still important (i.e. Moneyball).

KEYWORDS:

• team sports
• budgeted revenues
• correlations finances and sports
• regression analysis
• fixed effects

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Gerrard (2007, p. 207):

Michael Lewis' bestseller, Moneyball: The Art of Winning an Unfair Game (2003), tells the story of how the Oakland Athletics in Major League Baseball (MLB) have achieved a sustained competitive advantage over an eight-year period despite being one of the lowest wage spenders.

2. The marginal revenue product is a concept used to describe a player's contribution to the team's revenues (see Scully, 1974, for a comprehensive description of this concept).

3. To address causality, the Granger causality test is often used. However, this requires time-series data. Thus, this test is not applicable to this data set.

4. Although histograms are a very useful starting point for addressing deviation from normality, ‘they do not tell us whether this deviation is large enough to be important’ (Field, 2000, p. 48).

5. The Kolmogorov–Smirnov test was originally a non-parametric test. However, a modified version can be used to test the normality of the distribution. Other alternatives are the Shapiro–Wilk test and the Anderson–Darling test. According to Stephens (1974), these tests are actually more powerful than the Kolmogorov–Smirnov test when it comes to addressing normality.

6. Normality could not be assumed for the total sample (for single seasons), the top 8, and the first quartile. Because all the tests draw on one or more of these sub-samples, non-parametric tests were used.

7. The sports economics literature (e.g. Daly & Moore, 1981) has applied the SRCC when measuring performance persistence in the context of competitive balance. Note that Groot (2008) argued for Kendall's tau in the context of competitive balance.

8. More specifically, this applies to the analysis drawing on the first quartile against the rest of the teams (in both 2012 and 2013). Note that in these cases, the two-sample Kolmogorov–Smirnov test is used.

9. For a definition of Levene's test statistic, see, for example, Brown and Forsythe (1974).

10. Non-parametric two-sample tests are conducted using the Wilcoxon–Mann–Whitney rank sum test of two populations by Kanji (1993), in which we want ‘to test if two random samples could have come from two populations with the same mean’ (p. 86).

11. See Gerrard (2006) for a description of the possible persistence effects in this context.

12. See, for example, Wooldridge (2009). Note that the functional form that we call semi-log is called ‘level-log’ by Wooldridge (2009).

13. Hervik, Ohr, and Solum (2000) find the R² to be close to .74 when using the logarithmic sporting success on average over the period 1997–1999 as the independent variable and the logarithmic revenues as deviation from the average revenues. Applying an additional season, Gammelsæter and Ohr (2002) measured the R² as .770.