Human resources outsourcing in Canadian organizations: An empirical analysis of the role of organizational characteristics, transaction costs and risks

Abstract

The main objective of this study is to evaluate the influence of three classes of determinants using a sample of 271 Canadian Organizations, namely: (1) strategic and organizational characteristics; (2) transactional variables; and (3) risk factors on human resource outsourcing. Using an ordered probit statistical analysis, the results show that a strategic implication on the part of the human resources department, support from top managers, prior experience with outsourcing and extensive use of benchmarking in human resources management are positively related to the intensity of Human Resource Outsourcing (HRO). We found some support for the transaction-specific variables, most notably for the complexity and the specificity of transactions. Finally, the study reveals that human and business risks appear to be important drivers of a firm's decision to outsource.

Keywords:

• benchmarking
• human resources
• outsourcing
• outsourcing risks
• top management support
• transaction costs

Acknowledgements

The authors wish to thank CIRANO, Social Sciences and Humanities Research Council of Canada (SSHRC) and the Fonds pour la Formation de Chercheurs et l'Aide à la Recherche du Québec (FCAR) for their financial support and Mrs. Michelle Lacombe, Mr. Jérémy Luchetti and Mr. Jean-François Houde for their outstanding research assistance.

Notes

1. For instance, Lever (1997) asked if the extent of outsourcing was ‘low’, ‘moderate’ or ‘high’, and did not distinguish between design and day-to-day administration of programmes.

2. Initially, we also considered a fourth mode, i.e. outsourcing of 100% of the activity; but we had less than 2% of our observations at that level. Our results are consistent with those of Klaas et al. (2001) that a minority of firms have opted to outsource an entire HR activity.

3. In opposition to a multinomial probit where the different alternatives reflect choice variables that have no ordinal meaning (i.e. a multinomial probit could be used to explain the choice of a transport mode). See: Wooldridge 2002, ch. 15.

4. Ideally, we would like to use a multivariate ordered probit with six equations, but this would not be tractable as it would imply the evaluation of high-order multivariate normal integrals. Some progress has been made on trivariate integration, but existing results are not sufficient to allow for higher-order integration (Greene 2003, p. 911).

5. We included in the variance function the UNION (dichotomous) variable and the industry dummies.

6. The four items are: ‘The human resource direction is strongly involved in the strategic decisions of the firm’; ‘The human resource function is considered by top managers as a key function in the firm’; ‘Human resource officers are actively involved in strategic planning of organization activities’; and ‘Top management is convinced that the human resource function has a great part in the organization's success’.

7. The variable takes the value 1 if the percentage of unionized workers is 0; 2 if it is between 1 and 50%; 3 if it is between 51 and 80%; 4 if it is between 80 and 99% and 5 if it is 100%.