The Relative Demands for Sports and Leisure in England

Abstract

This paper contributes to an emergent but growing literature, which is focusing on large-scale data analysis of mass participation sport. Distinct from the existing literature, a model is proposed in which sport demand is examined relative to the demand for other leisure, a structure suggested by micro-economic theory. Based on new data for England, which has been collected by the government department responsible for policy promoting Sport, Media and Culture, it is shown that the demand for sports is dependent upon the demand for leisure for the activities defined by the data collection exercise. This qualifies the existing literature and has implications for current policy.

Acknowledgements

This paper has benefited from previous discussion with Alistair Dawson, Mick Green, Joseph Riordan and participants at a session of the economics of sports participation at the16th EASM conference, at the University of Heidelberg, 10–13 September 2008.

Notes

1. At this point it should be noted that there are complex issues surrounding claims concerning the validity and efficacy of public policy promoting sports participation. In the limit, economic theory is founded upon voluntary action, which makes policy attempts to change behaviour either unethical or unlikely to work in the presence of well-developed property rights unless one argues that market failures are present. This paper consequently comments upon policy as an entity whose existence presupposes the relevance of policy action, without necessarily sharing this view.

2. The use of a Cobb–Douglas utility function is purely for exposition to produce parametric demands to infer the direction of relationships. One quantitative limitation of the function is that it restricts the elasticity of substitution between sport and leisure to a unitary value. A given percentage change in sport relative to leisure demand is driven by the same percentage change in the marginal rate of substitution, that is preparedness of an individual to switch sport and leisure whilst leaving utility constant. A more general form of utility function is given by the Constant Elasticity of Substitution utility function. A simple version would be:

This would produce demand functions of the form

and a marginal rate of substitution of

The exponent of this last expression is the elasticity of substitution. If this term is “1” then the Cobb–Douglas solution is obtained. Estimates of this elasticity are provided for Flanders for leisure and non-leisure activities and between leisure activities by Kesenne (1981, 1983), respectively.

3. In this respect the “extensive” margin of substitution only is explored.

4. Walking is included in this survey but only in so far as it is for pleasure/rambling. General walking is not listed as a sport or other free-time activity constituting leisure. It is omitted from the analysis therefore because of its scale and ambiguous classification.

5. “0” reported percentage participation rates do not necessarily correspond to “0” levels of participation, but very small numbers.

6. One could attempt to impose theoretical restrictions, or use uncorrected structural estimates to suggest restrictions to identify the equations. The economic theory discussion above, however, indicates that this would be an arbitrary exercise as there is no reason to believe that certain factors contribute to leisure rather than sports demand and vice versa. Likewise, any prior data analysis would involve using uncorrected and biased estimates to suggest the required corrections, which would be illogical.

7. As there are no data on individual wage rates, or data on work hours from which a wage rate could be interpolated from income data, a direct measure of the shadow price of time could not be included in the analysis.

8. In this research as sport represents a smaller set of cases, sports is placed in the numerator to maximize the sample size actually estimated upon. As discussed further below, the estimator chosen then controls for issues of sample selection.

9. It is very important to note that this decomposition of the coefficients applies for a given value of the independent variables and should not be generalized. In the paper the mean value of the metric variables is used and the switch of a dummy variable from the value 0 to 1 (see Kang, 2007).

10. This is, of course, to recognize that the activities described as sport may vary and include elements of leisure as defined here for a few activities.

11. This is perhaps indicated by the statistical significance of the variable, and viewing the t and z statistics as standardized effect sizes.

12. This is notwithstanding any potential debates about the likely policy efficacy of intervention. By definition, policy efficacy will be presupposed by the DCMS.