407 ECONOMIC ANALYSIS & POLICY, VOL. 39 NO. 3, DECEMBER 2009 Towards a New Dynamic Measure of Competitive Balance: A Study Applied to Australia’s Two Major Professional ‘Football’ Leagues * Liam J. A. Lenten School of Economics and Finance, La Trobe University, Victoria, 3086, Australia (Email: l.lenten@latrobe.edu.au) Abstract: A new measure for competitive balance between seasons is proposed, which takes the form of a mobility gain function, based on each team’s win ratios from the current and previous seasons. This ‘dynamic’ function measures competitive balance within a one- period change framework. While it is not suggested that this measure replace useful existing within-season measures, such as the widely used actual-to-idealised standard deviation (ASD/ISD) ratio, this measure does overcome one of the shortcomings of within-season measures – that is, the ability to pick up uncertainty of outcome from season to season, rather than merely from round-to-round. Hence, it is suggested that this measure could be used in conjunction with within-season measures in time-series analysis. An application to Australia’s Australian Football League (AFL) and National Rugby League (NRL) over a century of data reveals numerous interesting comparisons. I. INTRODUCTION Competitive balance (CB) refers simply to the degree of evenness in sports leagues. However, the idea of measuring and quantifying CB is far from clear-cut, given the diversity of defining it precisely. Most notably, complications arise when one considers the distinction between the three often-cited dimensions of CB. Firstly, there is the notion of uncertainty of outcome of any single match/contest, to which the ‘uncertainty of outcome’ hypothesis referred originally. Secondly, we have the concept of parity or otherwise in terms of the distribution of wins between teams in any given season, or ‘within-season’ CB. Finally, the idea of an * Earlier versions of this paper were presented at: (i) the Staff Developmental Workshop, Department of Economics and Finance, La Trobe University, 28 September 2006; (ii) the Seminar Series, Centre for Operations Research and Applied Statistics, Salford University, UK, 11 October 2006; (iii) the Seminar Series, Department of Economics, BI Norwegian School of Management, Norway, 30 May 2007; and (iv) the Australasian Meeting of the Econometrics Society, University of Queensland, Brisbane, 3-6 July 2007. The author would like to thank the various participants of the workshop, seminar and conference for their comments and suggestions, especially David Prentice and David Forrest, as well as Suzanne Sommer, Ishaq Bhatti and Andrew Raponi for some preliminary input.