Organizational Task Environments and Performance: An Empirical Analysis

ABSTRACT

Organizational theory and research suggests that organizational performance may be affected by environmental circumstances. However, the available evidence on this important issue for public organizations relies primarily on “objective” archival measures of the environment drawn from secondary data sources. Perceptual measures gauging managerial “subjective” views on the nature of the task environment are also likely to be important determinants of organizational outcomes. In this paper, a comprehensive model of the impact of “objective” and “subjective” task environments on the performance of local government service departments is formalized and tested. The results suggest that both objective and subjective environmental munificence are positively related to performance, and that objective and subjective environmental dynamism are negatively related to service achievements. In addition, objective task complexity is negatively related to performance. The theoretical and practical implications of the findings are discussed.

Notes

⁺ p ≤ 0.10.

*p ≤ 0.05.

**p ≤ 0.01 (two-tailed tests).

+ = significant positive association; − = significant negative association; NS = no significant association.

1. This performance indicator was not collected in 2002/03. Thus the organizational effectiveness measure for that year is made up of only two housing PIs.

2. Spending per capita for the local government as a whole is used as expenditure data for this service area are not available.

The use of the terms “objective” and “subjective” might be deemed provocative, but they are widely understood in both the social science measurement and the management literatures. See Andrews, Boyne, and Walker (2006b) for a discussion of this issue.

Some performance indicators are inverted (e.g., the percentage of rent written off as not collectable) so scores above the mean always indicated higher performance.

By using z-scores it was possible to pool the data for different services, because service effects were removed from the scores on the indicators (Andrews, Boyne, and Walker 2006b). To ensure that the analysis was not unduly influenced by particular indicators, the aggregation method weighted each indicator equally. Factor analysis was not used to create proxies for each performance dimension because the number of cases per service area is too small to create reliable factors. Similar statistical results were obtained when the analysis was repeated using a performance measure which gave one “key” indicator for each service area a weight equal to the total number of indicators in that area.

This is the standard population-weighted deprivation measure used by the U.K. central government.

To gauge levels of demographic diversity it is necessary to establish the relative size of different demographic sub-groups in relation to each other and the population as a whole. The percentages of the population in each local area falling within relevant age, ethnic, and social class sub-groups identified in the national census of 2001 (such as Chinese, Clerical and Secretarial, children aged 0–4) were therefore squared, before being summed and subtracted from 10,000 – the maximum possible sum of squares when using percentages as whole numbers. This measure is equivalent to the Herfindahl index economists use to gauge market concentration, and in this instance the subtraction of the sum of squares from its maximum possible score gives a proxy for the relative “heterogeneity” of the organizational environment, with a high score on the index representing a high level of complexity (see Trawick and Howsen 2006).

This measure is used as the Indices of Multiple Deprivation were unavailable in 1991. It represents a useful proxy for deprivation as there is a strong positive correlation (.67) between lone parent households in Wales recorded in the 2001 national census and levels of multiple deprivation in 2000.

Ten was added to the aggregated task instability and turbulence measures to ensure that the interacted measure combined two positive scores.

Questionnaire available on request from author.

Nonetheless, it is still conceivable that the findings are limited by the possibility that the 70% of survey non-respondents may have provided different responses to those that were received.

Following 2000/01, coverage of service expenditure data is less comprehensive in the NAWPIs. Nevertheless, research has shown that relative spending levels in U.K. local government departments vary little year on year (Sharpe and Newton 1984).

This is borne out in a positive correlation (.26) between objective and subjective munificence.

Additional information

Notes on contributors

Rhys Andrews

Rhys Andrews ([email protected]) is a Research Fellow at Cardiff Business School. He received his PhD in Politics from University of Wales. His research interests focus on organizational environments, civic culture, and public service performance.