Hostname: page-component-cd9895bd7-hc48f Total loading time: 0 Render date: 2024-12-25T08:38:52.340Z Has data issue: false hasContentIssue false
  • English
  • Français

The Relationship between Public Expenditures and Services: A Longitudinal Analysis

Published online by Cambridge University Press:  27 January 2009

B. Guy Peters
Get access Rights & Permissions [Opens in a new window]

Extract

In summary, we do find that levels of services and expenditures are more closely related across time than they are in cross-sectional studies and that, further, these relationships are functionally specific. At the same time, however, it would appear from these data that the extension of services precedes the expansion of the budget. It may, therefore, be argued that the extension of services can be used as a rational means of expanding the budget available to an agency rather than as a response to the increased availability of funds. This expansion may also be seen as a function of demand creation on the part of the agencies in increasing these services.

A second point to be made is the importance of differentiating different categories of service. For those indicators of service that concern output or a service rendered, expenditures and policy are positively related. For those indicators of service that are more related to impacts – or needs – there is a negative relationship – i.e. lower quality associated with higher expenditure. Therefore, we find that the policy-making system of Sweden may have been making policy in relation to important objective changes in its service environment. It further leads us to expect quite different relationships between services and expenditures dependent upon the type of service measure being used.

The research site and the measures of services used here are obviously different from those of Sharkansky. However, these findings do speak to the general theoretical point of the connection between public expenditures and services. They demonstrate the need to examine this important research question within a more dynamic framework and in a variety of social and cultural settings. Sharkansky's findings for the American states should be tested, for example, in an examination of the same question for sub-national units in Sweden and other European systems. Through the gradual accretion of these findings, we can elaborate theoretically the linkages between expenditures and policy outputs.

Type
Notes and Comments
Information
British Journal of Political Science , Volume 6 , Issue 4 , October 1976 , pp. 510 - 517
Copyright
Copyright © Cambridge University Press 1976

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 This assumption has been the basis of equating expenditures with ‘output’ in the majority of the empirical analyses of public policy. For a conceptual distinction between various aspects of public policy, see Sharkansky, Ira, ‘Environment, Policy, Output and Impact: Problems of Theory and Method in the Analysis of Public Policy’, in Sharkansky, Ira, ed., Policy Analysis in Political Science (Chicago: Markham Press, 1970), 61–79.Google Scholar

2 Sharkansky, Ira, ‘Government Expenditures and Public Services in the American States’, American Political Science Review, LXI (1967), 1066–77.CrossRefGoogle Scholar

3 Sharkansky, , ‘Government Expenditure and Public Services in the American States’, pp. 1074–5.Google Scholar

4 For example, the effects of previous capitalization tend to be taken into account directly, and there is relatively less variance in the effects of private spending.

5 Sharkansky, , ‘Government Expenditure and Public Services in the American States’, p. 1068.Google Scholar

6 Sharkansky, , ‘Environment, Policy, Output and Impact: Problems of Theory and Method in the Analysis of Public Policy’.Google Scholar

7 The major data sources are the following: Sveriges officiella statistisk, Sammandrang (Stockholm: Kungliges Boktryckeriet, 1867–1914)Google Scholar; Statistiska Centralbryan, Statistisk arsbok for Sverige (Stockholm: Kungliges Boktryckeriet, 1914–1970)Google Scholar; Statistiska Centralbryan, His-torisk statistisk for Sverige, Vol. I, 1 (Stockholm: Kungliges Boktryckeriet, 1968)Google Scholar; Statistiska Centralbryan, Bidrag til Sveriges officiella statistisk: Valstatistisk (Stockholm: Kungliges Boktryckeriet, 1872 to present)Google Scholar; Finans department, Riks-stat och intilldess nasta statsreglering vidtager (Stockholm: Kungliges Boktryckeriet, 1866 to present)Google Scholar; Lindahl, Erik, Dahlgren, Einar and Koch, Karin, National Income in Sweden 1861–1930 (Stockholm: P. A. Norstedtoch Soner, 1937).Google Scholar

8 There are several standard means of detecting the presence of autocorrelations, but we will rely on the most commonly used of these – the Durbin-Watson statistic. This statistic is used to test a null hypothesis of no serial autocorrelation, with values below the significance points allowing the rejection of the null hypothesis of no significant positive autocorrelation. Assuming that a significant value of the Durbin-Watson statistic is found, we must then proceed with a process for elimination.

The residuals of the regression are first used to compute a new regression in the form

The resultant regression coefficient q is then used to compute a regression in the form

If the autoregressive structure is of the first order, the above regression should yield a randomly distributed set of residuals, and provide a ‘true’ estimate of the relationships between the variables. We have used this procedure here and the values reported are for the first iteration without significant autocorrelation. See Johnston, J., Econometric Methods, 2nd edn. (New York: McGraw-Hill, 1972), pp. 243–66Google Scholar; Durbin, J., ‘Estimation of Parameters in Time-Series Regression Models’, Journal of the Royal Statistical Society, Series B, XXII (1960), 139–53Google Scholar; Grilickes, Z. and Rao, P., ‘Small Sample Properties of Several Two-Stage Regression Methods in the Context of Autocorrelated Errors’, Journal of the American Statistical Association, XXXVI (1969), 253–72.Google Scholar

9 See Miller, W. L. and Mackie, M., ‘The Electoral Cycle and the Assyfnetry of Government and Opposition Popularity: An Alternative Model of the Relationship Between Economic Conditions and Political Popularity’, Political Studies, XXI (1973), 263–79.CrossRefGoogle Scholar

10 There are other techniques, most notably spectral analysis, which can be used to isolate patterns of temporal relationships, but the method of cross-lagging would appear more appropriate for the problems encountered in this data. For the use of spectral analysis see Fishman, G., Spectral Methods in Econometrics (Cambridge, Mass.: Harvard University Press, 1969).CrossRefGoogle Scholar

11 See Deutsch, Karl, The Nerves of Government (New York: The Free Press, 1963).Google Scholar

12 The necessary zero-order correlations were calculated by the same iterative method detailed above. The GNP measure is in the same constant monetary units as the expenditure variable. Urbanization is defined as the proportion of the population living in cities of 20,000 or more.

13 Wildavsky, Aaron, The Politics of the Budgetary Process (Boston, Mass.: Little, Brown, 1964), pp. 111–13.Google Scholar

14 Niskanen, William A., Bureaucracy and Representative Government (Chicago: Aldine, 1971).Google Scholar

15 See, for example, Marmor, Theodore R., Wittman, Donald A., Heagy, Thomas, and Kudrle, Robert, ‘Politics, Public Policy and Medical Inflation’Google Scholar, in Zubhoff, M., ed., Health: A Victim or Cause of Inflation? (forthcoming).Google Scholar