A corporation is a nexus of incomplete contracts that gives a demand for control right allocations. Whether firms should centralize or decentralize control rights depends on local information and coordination. This study uses the audit industry in China as a laboratory to shed light on to what extent control right allocation within an organization affects individual economic actors’ incentives when balancing coordination cost and local information acquisition cost. The audit industry in China provides a rare setting for researchers to understand the costs and benefits of a centralized/decentralized organization. Coordination and local information acquisition challenges are acute in the audit industry. It is widely recognized that effective coordination in profit/cost-sharing and standardized training is crucial for audit firms to ensure high-quality audit services. This force gives rise to a need for centralization based on Williamson (1996).
In addition, centralized systems potentially ease the costs of separating the principal, who receives the firm’s profits, and the agents who generate the profits. Audit firms face shirking because each partner bears the total cost of their effort but reaps only a fraction of the benefit. Accommodating clients’ aggressive financial reporting requests can be another manifestation of the shirking problem. Yet, audit services rely on localized and sensitive information. Partners at branches have superior knowledge regarding local demand than headquarters-based managers, tilting the comparative advantage towards a decentralized control system. The distinct institutional structure in China provides additional benefits for analysis. Hayek (1945) posits that strategic control right allocation can help solve the problem of “rapid adaptation to changes.” Unlike the accounting markets in developed economies with large market shares by Big 4 auditors, the audit market in China is highly dynamic and competitive. Disturbances in the China audit market make it critical that audit firms incorporate local information. They must adapt to changes in their environment and increase the interdependence of decisions, making coordinated adaptation more acute than in stable professional service markets.
We investigate the audit firm’s centralization and professional service quality using private data from The Chinese Institute of Certified Public Accountants (CICPA). Adopting Greenwood et al. (1990)’s framework of centralization, we focus on the following:
- centralized market-financial control, proxied by centralization of the revenue system and cost system,
- centralized strategic control, proxied by royalty payments and centralized recruitment and training,
- centralized operating control, proxied by centralized corporate governance and quality control (inspection).
We further construct centralization scores based on this information set to examine the association between centralization and audit incentives (audit effort and adjustment) and audit outcome (incidence of restatement) and uncover how such association varies with the comparative importance of local information acquisition. Relying on this novel, control-right-allocation information on Chinese audit firms from 2010 to 2012, we first show that our centralization scores of audit firms are negatively associated with publicly observable audit firm operation networks. In particular, audit firms with more operational branches are associated with a lower centralization score. These findings help validate our measure as it is fully consistent with Hayek’s argument on the importance of decentralization in utilizing dispersed local information. Specifically, when an audit firm’s operational network is more dispersed, it should allocate more control rights to the local audit office.
Then, we examine the effect of audit firm centralization on auditor incentives. We find that local engagement auditors of more centralized audit firms are less likely to compromise in the audit. In particular, they exercise more effort when auditing local clients. In addition, they are more likely to adjust the reported earnings of their clients downward. Third, the incidence of restatement is less likely if a local office of a more centralized audit firm audited the statement. Fourth, the above results do not present if we restrict samples to those of audit firms’ headquarters, suggesting that local auditors’ distorted incentives drive the variations in audit quality. Collectively, our findings support the notion that an audit firm’s centralized control helps mitigate the coordination problems among local offices, supporting Williamson (1996).
We note that this evidence is based on statistical correlations and could stem from economic factors omitted from the analysis. Nevertheless, the patterns we document are robust across different specifications. We first use the number of Confucian temples nearby the audit firm headquarters as an Instrumental Variables for centralization. The underlying rationale is that Confucian temples are vehicles for educating offspring and achieving ordered progress toward a society controlled by the central ruler. We indeed show that a substantive number of Confucian temples cultivate a strong preference for centralization, leading audit firms to centralize control rights. The second step of the regression yields a consistent result as our main finding. Second, we use whether they headquartered the audit firm in a rice farming region as an alternative Instrumental Variables for its centralization. Rice farming is labor intensive and involves water sharing, requiring the building of dikes and canals. As a result, rice farmers must work together to develop and maintain infrastructure, increasing the demand for centralized control. In contrast, wheat farmers depend more on themselves, leading to a more individualistic mindset. Therefore, headquarters in a rice farming region, relative to a wheat region, gives firms a tendency to adopt centralized control. Results based on this alternative IV are also consistent with our main finding. Third, we rely on audit firm M&As as a unique setting to examine how changes in the degree of organizational centralization reshape auditors’ incentives at the local office level. Our findings reveal that an increase in the degree of control right centralization from audit firm M&A activities results in more audit effort, a higher likelihood of downward adjustments in clients’ reported earnings, and a lower likelihood of restatement at the local office level.
We next examine how the allocation of control rights and audit incentives/audit outcomes vary with the relative importance of local information acquisition (Hayek, 1945). A lower information acquisition cost would tilt the comparative advantage towards centralized control systems, whereas a more complex local environment would weaken the beneficial effect of centralization. First, we use a high-speed railroad station opening in a local city to measure information acquisition costs. Our analysis reveals a lower information acquisition cost because introducing high-speed railroad stations strengthens the association between centralization and auditor incentives/audit quality. Second, information technologies (IT) development facilitates internal information transfers and allows the central authority to control their local operations. More investment in internal IT can increase the central authority’s capacity to manage the branches, leading to a stronger association between centralization and auditor incentives/audit quality. We find a consistent result. Third, we examine how the importance of local information can tilt the comparative advantages toward decentralization. We show that if the client firms are more complex and require local auditors to exercise extensive individual efforts to collect client-specific information, the beneficial outcomes associated with centralization are attenuated.
In summary, relying on proprietary data of internal control right allocation information of Chinese audit firms, we examine the beneficial effects of centralized/decentralized control systems on local engagement auditors’ incentives and audit quality. Our tests reveal that local engagement auditors of more centralized audit firms provide more effort than their decentralized peers. Auditors in centralized firms are also more likely to adjust reported earnings downward and produce better-quality audited financial statements than in decentralized firms. These findings collectively support the notion that an audit firm’s centralized control helps mitigate the coordination problems among local offices.
Qihui Gong is an Associate Professor of Accounting at the Zhejiang University School of Management.
Yupeng Lin is an Associate Professor of Accounting at the National University of Singapore.
David M. Reeb is a Professor at the National University of Singapore.
This post was adapted from their paper, “To centralize or not: Control right allocation and auditor incentives,” available on SSRN.
