Date of Award


Document Type



Edith Cowan University

Degree Name

Doctor of Philosophy


Faculty of Business and Public Management

First Supervisor

Colin Dolley

Second Supervisor

Keith Houghton

Third Supervisor

Gary Monroe


In recent years, the push for reform in the Australian public sector audit has placed the Office of the Auditor-General (hereafter OAG) in a more contestable or market- like environment, where the OAG is accountable for an efficient and effective provision of public sector audit. The purpose of this study is to compare the cost efficiency of in-house and contract-out arrangements to deliver financial audits in the public sector. It empirically tests whether there are audit cost and audit fee differences between in-house providers (i.e., the OAG) and contractors (i.e., public accounting firms). The secondary aims of this study are to develop audit cost and fee models for the public sector. The unit of analysis is audit cost/fee at the audit engagement level. The data for this study is collected for a sample of financial statement audit engagements for year-end 1998, at the state level in Western Australia. The data is extracted from publicly available and private sources. The audit cost and fee models are used to test for the cost differences between in- house providers and contractors. Prior audit production and audit fee studies in the private and public sectors provide the basis for the development of the two models. The results indicate that agency size, complexity and risk are positively associated with audit costs and audit fees. In addition, the total advice provided to the agencies by the OAG and agency type are also significantly associated with audit costs and audit fees. Overall, by incorporating these factors into the models, the audit cost model explains 82 percent of the variance in audit costs, while the audit fee model explains 86 percent of the variance in audit fees. More importantly, the main findings suggest that contract-out audits are more costly than in-house audits. However, this finding is conditional on agency type. Further analysis reveals that the type of audit arrangement is significantly associated with audit costs for the statutory authority audits only. There is no significant difference in audit costs between contract-out and in-house arrangement for hospital audits. This analysis shows that the statutory authority audits are driving the significance of the interaction between type of audit arrangement and agency type. Specifically, the costs of contract-out audits are, on average, significantly higher than in-house audits. This result is attributed to the contractor's lack of expertise in auditing statutory authority as there is no equivalent of this agency type in the private sector. As such, the OAG bas the greater advantage of delivering a lower audit cost for statutory authority audits compared to the contractors. However, the non-significant interaction term in the audit fee model suggests that cost differences between in-house and contract-out audits for the statutory authority audits are not reflected in audit fees billed to agencies. Further analyses, using audit hours as the dependent variable, generally corroborate the findings from the audit cost and audit fee models. Sensitivity analyses on the OAG's supervision costs reveal that these costs have a significant effect on the interpretation of the cost efficiency results. By excluding supervision costs from contract-out audits, there are significant changes in the results for the total sample and the two sub-samples (partitioned by agency type). Generally, these changes favour the contract-out audits for all groupings, where contract-out audits are now more cost efficient than in-house audits for hospitals, and not significantly different in costs for statutory authority audits. Additional tests to investigate the determinants of the GAG's supervision costs in contract-out audits reveal that agency size, risk, reliance on internal control, total advice provided by the OAG and packaged audits (a single contract for two or more audits) are significantly associated with the supervision costs of contract-out audits. The main contribution of this study is to add to the growing literature on audit market efficiency (see Dopuch, Gupta, Simunic & Stein, 2000; Knechel & Payne, forthcoming). It provides evidence on the production function of different type of suppliers in the public sector and their relative efficiency in providing audit services. This study contributes to the recent discussions on the changing nature of public sector audit market towards a market-based provision of public sector audits. The evidence from this study allows researchers and policy-makers to compare the two types of audit arrangement to undertake public sector audits. In pan, this study also contributes to the line of inquiry that examines the difference between government auditors and public accounting firms in US municipalities (Copley, 1989; Dwyer & Wilson, 1989; Rubin, 1992). The secondary contribution of this study is to develop and test the audit cost and fee models in the public sector and provide validity on the transferability of audit models from the private and public sectors. This study adds to the literature that examines the public sector audit market. More importantly, it is one of the few non-US studies that examine the public sector audit market and the findings from this study suggest that the public sector audit studies from the US are generalisable to Australia. These findings add to our understanding of the range of market conditions under which it is so far known to hold.