SHOULD WE STAY OR SHOULD WE GO?
STATUS ANXIETY IN CLIENT DEFECTIONS FROM ARTHUR ANDERSEN 2002
Michael Jensen
Corporate Strategy and International Business
University of Michigan Business School
701 Tappan Street
Ann Arbor, MI 48109-1234
Phone: (734) 764 2313
Fax: (734) 764 2555
Email:
ROUGH DRAFT: FOR SEMINAR USE ONLY
PLEASE DO NOT CITE OR CIRCULATE
April 2004
SHOULD WE STAY OR SHOULD WE GO?
STATUS ANXIETY IN CLIENT DEFECTIONS FROM ARTHUR ANDERSEN 2002
This paper focuses on the role of status anxiety in the dissolution of interfirm relationships, focusing specifically on what determined the timing of client defections from Arthur Andersen during its dramatic collapse in 2002. Status anxiety refers to concerns about being devalued because other actors question the quality of your partners, and is triggered in two different ways. First, since firms are exposed to different levels of monitoring by market institutions, they differ in the extent to which audit quality is important to them, thus suggesting they would react differently to deteriorating audit quality (demand heterogeneity). Second, when firms make decisions in highly ambiguous contexts, they often model their decision on the decisions already made by other firms, thus suggesting exposure to prior adopters affected defections (social contagion). Event-history analyses provide strong support for both arguments: firms with higher demands for audit quality and firms exposed to other defectors were indeed more likely to defect.
While firms typically form exchange relationships with each other to gain access to specific resources, exchange relationships are simultaneously important market signals that affect how third parties evaluate firms. When it is difficult to determine the future value of a firm prior to investing in it, investors are likely to shift attention to the status and reputation of the exchange partners used by the firm (Podolny, 2002). For example, external endorsements by prestigious exchange partners have been shown to benefit private firms going public (Carter and Manaster, 1990), day care centers (Baum and Oliver, 1992), investment banks (Podolny, 1994), and biotechnology firms (Stuart, Hoang, and Hybels, 1999). Not only does the value of relationships with prestigious firms increase the demand for these firms as exchange partners, but it also gives prestigious firms strong incentives to protect their status and reputation by avoiding affiliations with firms that could threaten it (Podolny, 1994). For example, investment banks prefer to rely on other investment banks of similar status as themselves to manage underwriting syndicates (Chung, Singh, and Lee, 2000), and medium-sized firms prefer high-status audit firms even if other audit firms have the capacity to audit them (Han, 1994). Securing and protecting one’s own status and reputation through selective formation of relationships at the micro-level result in self-reproducing stratified markets at the macro-level where status and reputation for quality form important mobility barriers (Podolny, 1993).
This study shifts attention from the benefits firms obtain from using prestigious exchange partners with strong reputations for quality to the dissolution of these relationships when firms’ reputations for quality are questioned. Since status is latently transferred through relationships with other firms (Podolny and Phillips, 1996), relationships with firms whose reputation is under pressure is likely to cause status anxiety, or concerns about being devalued because other actors question the quality of your partners, which, in turn, could motivate disassociation. Using the dramatic collapse of Arthur Andersen following the Enron audit failure in 2002 as the empirical setting, I focus on the dissolution of prestigious relationships and ask specifically what determined the timing of defection by publicly traded clients from Arthur Andersen. I develop two sets of arguments to account for differences in defection rates. First, since firms are exposed to different levels of monitoring by market institutions, they differ in the extent to which audit quality is important to them, thus suggesting that they may react differently to deteriorating audit quality (demand heterogeneity). Second, when firms make decisions in highly ambiguous contexts, they tend to model their decisions on the decisions already made by other firms, thus suggesting that exposure to prior adopters affected when firms defected (social contagion). Together, these arguments suggest that demand heterogeneity accounts for early adoption among firms with higher demands for audit quality, which then triggers a social contagion process through which defections diffuse to firms with lower demands.
Although some research has examined the dissolution of audit relationships, it has focused on isolated dissolutions, which are an integral part of normal business activities and typically have few long-term consequences for the involved firms (Levinthal and Fichman, 1988). However, the epidemic defections from Arthur Andersen represent a unique opportunity to study how processes of delegitimation driven by economic and social forces can have catastrophic consequences for individual firms and restructure relationships throughout an entire industry. While the selective formation of relationships to gain status is well understood (Podolny, 2001), the dissolution of relationships to protect one’s own status has received less attention. This study suggests that firms that normally benefit from their reputation for quality are also more likely to collapse if their reputation is jeopardized because discerning firms with higher demands for quality are more likely to defect early. Client defections from Arthur Andersen might, ironically, have been less dramatic had Arthur Andersen not been a high-status audit firm.
Finally, this study also provides insights into the dynamics of business collapses by using event-history analysis to model a seemingly chaotic process in which hundreds of relationships dissolved in a very short period of time: what seemed like chaos was, however, at least partly determined by demand heterogeneity and social contagion.
DEMAND HETEROGENEITY AND CLIENT DEFECTIONS
Audit Industry Structure
The audit industry has traditionally been divided into three different tiers of audit firms: international, national, and local (see Doogar and Easley, 1998, on industry structure). The first tier included in 2001 the Big Five audit firms, i.e., PricewaterhouseCoopers, Deloitte & Touche, Ernst & Young, KPMG, and Arthur Andersen.[1] All the Big Fives had more than 1,000 partners and a strong international presence and dominated the industry with a combined market share of more than 85% SEC clients (Public Accounting Report, 2001). The second tier included approximately 10 audit firms, of which the most important were Grant Thornton and BDO Seidman, with a national presence and between 100 and 500 partners. The third tier included all the other audit firms, most of which only had a local presence (some had a regional presence) and less than 50 partners. The most important differences between the Big Fives and the second and third tier audit firms are their audit capacity and audit quality. The capacity of an audit firm refers to the size, complexity, and breadth of firms it is capable of auditing, whereas quality refers to the joint probability of detecting and reporting material errors in financial statements (DeAngelo, 1981).
While some Big Fives have slightly more experience in some industries and geographical areas than other Big Fives, they all have the capacity to audit firms of any size and complexity operating in any industry, institutional setting, or geographic area. The capacity of the other audit firms is, in contrast, more limited, and they typically restrict themselves to auditing relatively smaller firms with simpler operations operating in a particular geographical area. The average quality of the audits performed by the Big Fives is also likely to be higher than the average quality of the audits performed by the other firms because of their size and brand name recognition. Larger audit firms have an incentive to provide higher quality when incumbent auditors earn client-specific quasi rents, i.e., have a cost advantage due to auditor startup and client switching costs, because they risk losing more if it becomes known that they have failed to discover and report material accounting errors (DeAngelo, 1981). Similarly, firms are explicitly motivated to develop and protect their reputation for quality to secure and protect the quasi-rents arising from their brand name and the quality-assuring premium it allows them to charge, thus suggesting that firms with more reputable brand names are less likely to disappoint on quality (Klein and Leffler, 1981). The Big Eights have indeed been shown to charge higher audit fees for comparable audits, have lower litigation (due to audit failure) occurrence rates, and are more likely to constrain aggressive and opportunistic reporting that could threaten their reputation than other audit firms (Simon and Francis, 1988; Palmrose, 1988; Francis, Maydew, and Sparks, 1999).
Audit Demand Heterogeneity
The demand for audit services reflects the audit industry structure and is heterogeneous on the capacity and quality dimensions. While all the Big Fives are involved in different forms of consulting, the main service they provide to their clients is auditing and certifying their financial statements as required by the Securities Act of 1934. As firms grow and their operations become more complex due to, for example, product or geographical diversification, smaller local audit firms become increasingly less likely to have the capacity to audit their financial statements. While audit capacity is a common reason to switch to the Big Fives, it is less likely a reason to dismiss Big Fives or switch between them as they all have the audit capacity to service even the largest and most complex firms. More importantly, firms also differ in their demand for quality. Some firms demand higher quality audits and are therefore willing to pay the premium the Big Fives require, whereas other firms are less concerned about quality as long as it meets the minimum standards required by the Securities Exchange Commission (SEC).
Since audits are a means of reducing the agency costs of separating ownership from control in modern corporations (Jensen and Meckling, 1976; Watts and Zimmerman, 1983), firms with the potential for more agency problems are also more likely to demand higher audit quality. Firms may voluntarily demand high-quality audits as a signaling mechanism when they raise capital in the equity market by issuing stocks. The existing owners of firms that are going public have an incentive to minimize underpricing (i.e., the difference between the offering price and the market price at issuance) because it transfers wealth to the new investors. The degree of underpricing is a positive function of the ex ante uncertainty of the future value of a firm (Beatty and Ritter, 1986), which gives firms with low ex ante uncertainty an incentive to hire credible auditors to reduce uncertainty through certification of their financial statements. Consistent with these arguments, firms audited by the Big Eights tend to experience less underpricing when they go public (Beatty, 1989) and firms are more likely to switch to the Big Eights before going public (Titman and Trueman, 1986; Menon and Williams, 1991).
Firms may also have high-quality audits imposed on them as a monitoring mechanism by shareholders, debtholders, and institutions seeking to reduce agency problems. For example, debtholders (bondholders and commercial banks) have an interest in limiting agency problems that may reduce the value of their bonds. They are therefore likely to make sure firms agree to produce detailed financial statements and to have their accuracy certified by an external auditor (Jensen and Meckling, 1976). DeFond (1992) reported that firms were indeed more likely to switch to Big Eights after increasing their leverage (debt ratio) as well as before they increased it, thus suggesting that debtholders impose high-quality audits and that firms anticipate this requirement and shift before increasing their leverage. Krishnan (2003) found that reported earnings, which may be subject to aggressive and opportunistic accounting by managers, were more informative if firms had been audited by the Big Six as witnessed by the stronger association between reported earnings and stock returns for these firms, thus suggesting that audit quality is important to capital markets.
Demand Driven Defections
The heterogeneity in demands for audit quality implies that firms with higher demands should react faster to declines in audit quality and therefore dissolve their relationships with troubled auditors before firms with lower demands. I suggest specifically that firms exposed to monitoring by institutional investors and security analysts are more likely to defect when audit quality is questioned. The dispersion of ownership among smaller shareholders has traditionally been viewed as an important source of managerial power (Berle and Means, 1932). Smaller shareholders do not have as strong incentives as larger shareholders to collect information and monitor the management as larger shareholders, such as institutional investors (Shleifer and Vishny, 1986). While institutional investors rarely are directly represented on corporate boards, they have the necessary expertise and economies of scale to effectively monitor management (Useem, 1996). As the shareholdings of institutional investors have increased in size, it has become more difficult for them to rely on the “Wall Street Rule” and sell the shares of poorly performing firms without depressing their price (Heard, 1987; Pound, 1992). Institutional investors have therefore adopted a more activist approach to corporate governance issues and begun seeking to influence management more directly. While the direct effect of institutional ownership on firm performance is negligible (Kang and Sørensen, 1999), institutional ownership has been shown to affect specific corporate decisions, such as adopting antitakeover amendments (Brickley, Lease, and Smith, 1988) and the level and mix of CEO compensation (David, Kochhar, and Levitas, 1998).
Regardless of the direct effects of institutional ownership, the threat itself of shareholder activism has forced management to become more attentive to scrutiny by large institutional investors (Useem, 1996). Davis and Robbins (2003) argued that firms appoint prestigious directors because they provide a signal of quality and found that scrutiny by institutional investors, as measured by proportion institutional ownership, increases the likelihood firms appoint prestigious directors. I suggest similarly that firms are more likely to react to perceived decreases in audit quality by replacing their auditor, the more potential scrutiny they face from institutional investors. Some anecdotal evidence suggests that institutional investors indeed do pay attention to audit quality and put firms audited by troubled auditors under heightened scrutiny. For example, Patrick MCGurn, vice president of Institutional Shareholder Services, predicted that institutional investors were “going to look more closely at Andersen companies” (Solomon, 2002: C7), thus suggesting firms may seek to avoid institutional investors questioning their financial statements by defecting from Arthur Andersen. The following hypothesis is suggested on this basis: