1.0 INTRODUCTION

The recent mergers of public accounting (CPA) firms that reduced the Big 8 to the Big 6 refocus concern about the behavior of suppliers in the market for audits [Rankin, 1989]. In an early attempt to present a profile of larger CPA firms from publicly available data, Zeff and Fossum [1967, p. 306] noted the opportunities for future research by stating, "The data presented...barely scratch the surface of a potentially fascinating area of study, namely, how public accounting firms grow, develop, and acquire personalities."

One aspect of this area of study that has been considered since Zeff and Fossum is the nature of competition in the market for audits and CPA firms' behavior in this competitive environment. Concerns range from recurring charges that large, dominant CPA firms prevent effective competition to fears that destructive competition will erode auditor independence. The implication of these concerns is that regulation is needed to control supplier behavior.

This paper synthesizes empirical results about auditor concentration, audit fee, and auditor choice to produce a picture of what is known and what remains to be learned about the audit market structure, determinants of that market structure, and CPA firm behavior and performance in the United States. The economics of industrial organization provides a conceptual framework for organizing this topic.

Any discussion of an economic market must define the market's buyers and sellers. Although it is convenient to define the U.S. audit market as all audit clients, most results indicate that suppliers' performance and behavior differ in the large-and small-client segments of that market. Similarly, suppliers are grouped into two types: the Big 8 firms, which participate in both market segments, and all other CPA firms, which generally participate only in the small-client segment. The distinction between the two market segments and the two supplier types is maintained in our discussion.

The next section provides a brief overview of the economics of industrial organization and the approach we take to organize the accounting literature. In sections 3-6, we review the empirical evidence. Section 7 contains a summary, conclusions, and directions for future research.

2.0 THE ECONOMICS OF INDUSTRIAL ORGANIZATION

Industrial organization stems from microeconomics and is the applied economics of supply [Shepherd, 1990]. The primary concern of industrial organization is whether firms in the market are behaving in a manner that promotes social welfare. A central assumption of structuralists is that effective competition depends on a market's structure and results in performance that provides the greatest benefit to society. Effective competition involves a striving among comparable rivals who exert a mutual pressure so strong that all competitors must apply maximum efforts. Effective competition requires two internal market conditions: a reasonable degree of parity among competitors and enough competitors to prevent effective collusion among them to rig the market.

A market's structure is measured by the size distribution of competing firms and reflects the internal conditions of the market. A firm's market power is measured by its market share, which determines whether there is a monopoly in the market. Concentration is the combined market share of leading firms and represents the degree of oligopoly in the market. Combined market share is a dilution of the effect that a single firm with the same market share would have. High market concentration usually favors collusion and increases leading firms' ability to earn excess profits by raising price above cost.

External entry conditions can substitute for internal competitive conditions. Even with few rivals, a market may behave competitively if entry from the outside is a realistic threat. However, barriers to free entry (such as economies of scale product differentiation, or diversification) may determine structure.

Mainstream researchers hold the structuralist view that (1) conditions of supply and demand determine market structure and (2) firm behavior and performance are the result of market structure. Firm behavior can be monopolistic, collusive, or competitive, and firm performance is reflected in the firms' economic patterns operational efficiency, and rates of innovation. Good performance is promoted by a market structure with low concentration and competitive firm behavior. However a firm's superior performance may justify monopoly in some cases. The fundamental pattern is that a firm's market power varies mainly with its market share, and price differences may reflect a firm's monopolistic behavior.

Although structuralists recognize that cause and effect are mixed to some degree, causation is presumed to flow primarily from structure to behavior and performance. The school of thought associated with the University of Chicago and UCLA assumes that the direction of causation is reversed (see figure 1). (Figure 1 omitted) Each firm's relative efficiency is said to be the real determinant of its behavior and market structure. The view can be summarized as follows:

1. Monopoly reflects superior performance.

2. The costs of attaining monopoly commonly use up any possible monopoly profits.

3. Market dominance has only minimal harmful effects.

4. Collusion is the only pure form of market power, and it quickly collapses from cheating.

Analysts within this school assert that monopoly is rare, weak, transient, and justified by efficiency [Demsetz, 1974].

Proponents of both views agree on the following core issues:

1. Firms seek higher market shares to gain higher profits.

2. When these firms' mutual strivings hold each other in check, effective competition exists.

3. If one or several firms gain high market shares, they can obtain high profits by setting prices above costs and restricting output. Both monopoly and oligopoly impose social costs.

4. The costs of monopoly may be offset if there are large enough scale economies, product differentiation, diversification, or superior performance by the dominant firms.

To assess the nature of competition in the market for audits, empirical evidence is organized with the following approach: In section 3, Big 8 market shares and audit market concentration are used to divide the audit market into two categories. Market performance expected in each market category is compared to the evidence of actual performance. In section 4, evidence of economies of scale is considered in a similar manner. Sections 5 and 6 contain discussions of the empirical evidence regarding product differentiation and firm diversification, respectively. A summary of the sources for empirical results included in this review is provided in table 1.

3.0 STRUCTURE OF THE MARKET FOR AUDITS

Markets are categorized according to evidence of a monopoly, dominant-firm, tight oligopoly, and loose oligopoly. Monopolistic and dominant-firm markets are characterized by a single firm with no close rival. Since at least six to eight firms compete as relatively equal rivals in the audit market, the critical issue is whether collusion is likely or not, the distinction between tight and loose oligopoly.

The difference is one of degree. A tight oligopoly has fewer rivals, higher concentration, stable market shares, and medium to high barriers to entry. A loose oligopoly has lower concentration, unstable market shares, and low entry barriers. Generally, loose oligopoly is the best all-around structure. It provides firms with significant degrees of market presence, continuity, and stability while permitting the realization of scale economies in production, innovation, and other activities. Because concentration is low, the likelihood of collusion is also low.

Burton and Roberts [1967] reported that, in 1952, 92 different CPA firms audited the 620 companies listed at least once as a Fortune 500 company from 1955 to 1963. In 1965, only 37 different CPA firms were involved in these audits, and 599 of the companies were audited by just 16 different CPA firms. This rapid reduction in the suppliers of audits caused concern that the structure of the audit market had changed and that the competitive behavior of CPA firms in the market was affected.

3.1 EVIDENCE OF HIGH MARKET CONCENTRATIONS

Early work employed descriptive measures of individual CPA firms' market shares at points in time. Zeff and Fossum [1967] described the market share of individual CPA firms in the market composed of companies listed in the 1965 Fortune Directory, extrapolating market share from client revenues, assets, and income by industry and industry groupings. They found that the Big 8 firms dominated the market generally, but the extent of domination varied across industries. Rhode et al. [1974] and Schiff and Fried [1976] compared Zeff and Fossum's results with similar computations, employing the 1972 and 1973 Fortune Directory companies, respectively. No notable changes were reported by either study.

These studies share the following limitations.

1. Each considers only the 500 largest corporations in the United States as a surrogate for the total audit market.

2. None of the studies is able to allocate portions of audits among CPA firms when the auditor of the parent company relies on the opinion of other auditors.

3. Each study presents measures of concentration in terms of number of clients and three client characteristics: revenues, total assets, and net income. Revenues are emphasized in discussions because the researchers considered client revenues to be the least ambiguous accounting measure. Theoretical justification for this is absent.

4. The classification of companies into industries is problematic. Most companies of this size do not function within only one industry and thus are difficult to categorize. The problem becomes more difficult when multiple time periods are considered, since a company's industrial classification may change more frequently than its auditor.

5. An unusually large portion of an industry's total revenue could be contained in one or a few large companies. This concentration of market power in the underlying industry would be reflected in the CPA firm's market share and could misrepresent the market power of the firm.

6. None of these studies combines market shares of the largest CPA firms or provides any other summary concentration statistics. The only analysis is performed by "eyeballing" the data for "obvious" interpretations.

The principle contribution of these pioneering studies was to identify the problem and focus attention on the issue. Also, the approach used in these studies became an important guide for future studies. However, the emphasis was on individual firms and monopoly of the market rather than on the degree of oligopoly in the market. The conclusions of these early studies should be considered in light of their limitations and the absence of theoretical guidance.

Gilling and Stanton [1978] studied the structure of the audit market in Australia. Two improvements on the previous studies were introduced: the sample was not limited to the largest Australian companies, and summary concentration measures were computed. The two concentration measures were the n-firm concentration ratio (CRn) and the Herfindahl index (H). The CRn ratio expresses concentration as the percentage of total activity in a market that is accounted for by the n most active firms. The H index is a marketwide concentration measure that is sensitive to the number of firms active in the industry as well as to the variances in activity level across firms.

Table 2 contains concentration measures for the U. S. audit market since 1955. When possible, CR4, CR8, and H indices are provided. (Table 2 omitted) The source of client data employed by the various studies and the surrogate used to measure audit fees are identified in table 3. (Table 3 omitted) Early studies reported only total client revenue data by auditor for Fortune 500 companies. We computed concentration indices from the data. Later studies used much larger data bases and, except for Tomczyk and Read [1989], computed the square root of client revenue as a proxy for audit fees.(1) Tonge and Wootton [1991] reported concentration figures for 1988 by assuming the Big 8 had already merged into the Big 6 with no loss of clients.

Shepherd [1990, p. 61, 661 suggests that a CR4 (H) above 60 percent (O.18) is indicative of a tight oligopoly and below 40 percent (O.10) of a loose oligopoly. As seen in table 2, the CR4 and H indicators consistently give ambiguous signals about the type of oligopoly in the U. S. audit market over the entire time period and across the various measurement methods. High overall concentration levels increase the possibility of collusive behavior and regulation [Demsetz, 1973]. Concentration in the audit market is high enough that collusion is feasible but small enough to make collusion difficult.

3.2 EVIDENCE OF FIRM PERFORMANCE

Market structure is a concern because of its effect on firm behavior and performance. The primary distinction between a tight and a loose oligopoly is the ability to collude. Collusion's classic effect is to raise price above cost. In a tight oligopoly, collusion usually is tacit; firms build up a degree of shared knowledge and trust that allows price fixing without a system of explicit controls and penalties. Market power, for example, enables dominant firms to engage in patterns of price discrimination, in which differing price-cost ratios are achieved across clients. Loose oligopolies, in contrast, rarely achieve tacit collusion. Homogeneity of products, stable industry conditions, and long industry experience aid tacit collusion [Tirole, 1988].

Dopuch and Simunic [1980] tested the stability of market share among CPA firms and found that, even when only Fortune 500 companies were considered, the dominant Big 8 firm changed in 24 of 32 industries studied from 1965 to 1975. When the data source was expanded to all companies listed in the 1976 Who Audits America(2), the smaller client market exhibited lower Big 8 market share overall, and individual Big 8 firms' market shares differed by auditee size categories. Generally, instability of relative market share within industries and auditee size categories among the Big 8 does not facilitate maintenance of tacit collusion.

The primary result of tacit collusion is leading firms' ability to charge fees in excess of costs for some clients. Studies of audit fees provide a test of price discrimination by attempting to find a consistent relationship between fees and cost factors. Survey evidence has demonstrated that fees are an important consideration to corporate executives in their choice of auditor [Bedingfield and Loeb, 1974; Eichenseher and Shields, 1983; Hobgood and Sciarrino, 1972]. Audit fees should be competitive and, in the absence of collusion, should reflect marginal cost factors. Researchers of the audit market obtain fee data directly from clients or from public sources. However, cost data have not been available. Instead of direct measures, researchers have developed surrogates for audit cost factors.

Simunic [1980] assumed that the benefit of an audit is liability avoidance. He developed an economic model that relates audit fees to potential third-party losses and to relative costs of production. The model's three components are:

1. client loss exposure, which represents the business risk assumed when the CPA firm is associated with a given client,

2. client ability to share losses, which represents the client's ability to absorb some of the auditor's liability, and

3. auditor's production function, which represents the process cost of an audit relative to other audits.

Several studies that employ various empirical measures of these cost factors as control variables showed that much of the variation in U. S. audit fees is explained by the model. Table 4 contains a summary of these results. In a competitive market with homogenous products and production processes, a strong association of cost factors with fees is expected.

Generally, authors of fee studies audit fees by a measure of company size because size because size is closely related to fee.(3) Loss exposure has been measured by empirical surrogates for audit complexity, audit problems, and industry. Empirical surrogates for corporate performance and audit opinion (receipt of a qualified opinion during the last two years) measure loss-sharing ability. The auditor's production function is measured by the auditor's tenure and empirical surrogates for the auditor's work load.

Loss exposure variables that surrogate for complexity are always significant and, when considered, audit problems and industry also are significant. Except for Colson et al. [1988], surrogates for loss sharing, corporate performance measures, and audit opinion are significant when considered. Production function variables give mixed results. Although the empirical surrogates appear to be incomplete and researchers often do not accurately specify what the variables represent, the high explanatory power found across these results is a strong indicator of price competition.

Further evidence of competitive behavior is provided by evidence on audit fee behavior. DeAngelo [1981b] argues that low-balling, which occurs when auditors price initial audit engagements at a lower amount than they could sustain in the long run, is a competitive response to the existence of quasi-rents to incumbent auditors. Simon and Francis [1988] and Turpin [1990] found evidence of price cutting on initial audit engagements that persists through the third year of the engagement. Simon and Francis [1988] estimated the discount to be 24 percent in the initial year and 15 percent in the next two years. Evidence of low-balling persisted when only auditor switches between firms of the same type were considered, clients receiving qualified opinions were removed, other fees received by the CPA firm were controlled, and client size was controlled.

3.3 SUMMARY

High concentration levels in the audit market create concern that competition is compromised by collusive behavior. However, instability of relative market share is not expected from a well-behaved oligopoly. The dominant Big 8 firm changed over time in large-client submarkets defined by industry, and, across submarkets defined by client size, individual Big 8 firms' market share varied. Empirical evidence of persistent low-balling and an association between audit cost factors and audit fee indicates that price competition may be the dominant market behavior. High concentration levels may be the result of underlying market conditions rather than collusive behavior. Empirical studies have produced evidence related to three determinants of market structure: economies of scale, product differentiation, and diversification.

4.0 ECONOMIES OF SCALE

Free entry is a market condition that reinforces effective competition. Entry barriers give established firms a chance to raise price above long-run average cost without attracting new competition from outside. Common causes of barriers to entry are presented in table 5. (Table 5 omitted)

The existence of economies of scale could explain a market structure that is dominated by the largest firms. Economies of scale imply that the lowest average cost is attained only in a range of large volume output. The advantage provided by scale economies is superior operating efficiency that should be reflected in higher profits or lower fees. Social welfare is increased if scale economies are realized.

A direct test for scale economies would require data on CPA firms' costs. In the absence of cost data, researchers have investigated possible scale economies in the U. S. audit market by employing the survivorship approach and by considering audit fee behavior.

4.1 EVIDENCE OF BIG 8 ECONOMIES OF SCALE

The survivorship approach assumes that only cost-effective firms will gain market share over time if the market is characterized by homogeneous products, price competition, and price-inelastic aggregate demand [Stigler, 1958]. The cost effectiveness may be attributable to economies of scale. Researchers have applied the survivorship approach to the audit market by investigating the association between market share, gains in market share, and opportunities for scale economies in various audit submarkets.

Eichenseher and Danos [1981] compared auditor concentrations across 54 industries while controlling for underlying industry concentration. They found a significant association between auditor concentration and two industrywide indicators of opportunities for economies of scale: regulation and the extent of initial public offerings (IPOs) in the industry. These results imply that economies of scale can be obtained by auditors through specialization in regulated industries and IPOs. Danos and Eichenseher [1982] looked for evidence of economies of scale in the audit market by considering the relationship between market share at a point in time and subsequent changes in market share. Empirically, they considered whether a CPA firm's change in market share for a particular industry from 1972 to 1979 is a function of the firm's 1972 overall market share and industry market share (IMS). Only IMS and its interaction with overall market share were significant. These results imply that, if economies of scale are available in the audit market, they are industry-specific rather than across industries.

Danos and Eichenseher also analyzed the 249 companies in their sample that switched auditors during the period to help interpret their results. However, since change in concentration levels between two points in time can be the result of (1) a different mix of clients at the two times, (2) differential growth rates of continuing clients during the period, and/or (3) auditor switches, data on auditor switches alone cannot reliably explain the changes in concentration. Haskins and Williams [1988], for example, considered mergers among client companies and found significant evidence of acquisitions of other CPA firms' clients by Big 8 clients. Also, CPA firms with clients in industries with high merger activity tended to be growth leaders.

The empirical evidence of economies of scale consists of positive associations of auditor concentration with regulated industries, IMS, and IPOs. Relatively lower auditor fees reported in regulated industries also indicates possible scale economies [Colson et al., 1988; Palmrose, 1986a; Simunic, 1980; Turpin, 1990]. Studies of the effect of industry expertise on choice of auditor have not reported significant results if Big 8 and other CPA firms are included in the sample [Eichenseher and Shields, 1983; Palmrose, 1984]. However, since industry expertise was significant in studies that considered only Big 8 firms [Francis and Wilson, 1988; Shockley and Holt, 1983; Williams, 1988;1, economies of scale may be available in the large-client market.

Results from studies of IPOs are not as supportive of economies of scale. Simunic and Stein [1987], Balvers et al. [1988], and Beatty [1989] reported that more reputable underwriters prefer clients with Big 8 auditors and that IPO companies using Big 8 auditors are less leveraged and their new issues perform better in the market. Beatty also reported that Big 8 firms receive higher fees for IPO work. The combination of greater demand and higher fees generally indicates product differentiation rather than economies of scale. However, Beatty and Balvers et al. report some differences in performance among Big 8 firms in the IPO market, which implies the possibility of scale economies in the large-client market.

4.2 EVIDENCE OF BIG 8 FEE PREMIUMS

Simunic [1980] assumed that the small-client audit market should exhibit characteristics of a competitive market because of the large number of CPA firms active in that market. He used his audit-pricing model to make inferences about overall levels of price competition in the audit industry by comparing results in the small-client market to results in the large-client market. Table 6 contains the interpretations of comparative results provided by Simunic. (Table 6 omitted) The assumptions underlying this model are that (1) the greater number of suppliers in the small-client market implies more competition, (2) audit cost factors are adequately described, and (3) empirical measures adequately surrogate for these variables. Two additional studies have employed Simunic's model to compare large and small clients in the U. S. audit market. The results are presented and interpreted in a manner consistent with Simunic in Table 7. (Table 7 omitted)

All three studies found support for a competitive market in general. Palmrose [1986a] and Francis and Simon [19871 also found different fee behavior in the small-client markets, even though their empirical measures of large and small differ. Simunic [1980] argued that the combination of a competitive market and Big 8 fee premiums indicates Big 8 product differentiation.

4.3 SUMMARY

Although direct evidence of economies of scale is unavailable, possible scale economies are indicated in regulated industries, IPOs, and large-client segments of other industries. However, higher concentration and a general trend toward Big 8 firms apparently are occurring in conjunction with higher Big 8 fees in the small-client market. Cost advantages available to the Big 8 firms as a result of economies of scale do not explain higher fees. A premium would be paid to the Big 8 firms only if clients were receiving some additional benefit. Big 8 product differentiation could explain the fee premium if the differences indicate product superiority [Klein and Leffler, 1981]. Researchers have investigated this possibility.

5.0 PRODUCT DIFFERENTIATION

The audit product is a report that contains the CPA firm's opinion on the fairness of financial statement presentation. Prior research has considered two related features of the audit product that may differ with CPA type: (1) the degree of assurance provided to external parties by the audit opinion and (2) the use of CPA firm type as a signal by management and investors. If audit firms convince clients that their products differ, their prices can vary. Studies of auditor choice investigate factors that affect auditor quality, switches, and selection. Most investigate the demand side of the audit market by associating client characteristics with auditor choice. However, the empirical results also can be evaluated in terms of product differentiation.

The amount of assurance provided is a direct consequence of the quality of an audit opinion. Although audit opinions are governed by standards and do not vary in form between CPA firms, opinion quality can vary as a function of (1) the effort and expertise expended to gather supporting evidence and (2) the auditor's willingness to report truthfully [DeAngelo, 1982]. Empirical research has focused on consequences of the audit report because direct indications of opinion quality, such as audit working papers and auditor thought processes, are not observable.

The ability to serve as a signal may be another distinguishing feature of the audit product among CPA firms. Because asymmetry of information between managers and investors may cause investors to be unsure of the relative quality of various investments, managers have incentives to project superior future performance to distinguish their company from others. Bar-Yosef and Livnat [1984] proposed that the type of auditor may convey a signal about management's assessment of future cash flows to market participants.

5.1 QUALITY OF THE AUDIT OPINION

Audit opinion quality is a term frequently used in the literature to refer to differences between audits by Big 8 and other CPA firms. Many studies, for example, attempt to relate auditor selection and auditor switching to agency costs and assume that Big 8 and other CPA firm types appropriately surrogate for high and low audit opinion quality, respectively [Albrecht, 1991; DeFond, 1992; Eichenseher and Shields, 1989; Francis and Wilson, 1988; Healy and Lys, 1986; Palmrose, 1984; Simunic and Stein, 1987]. Empirical results indicate that loan officers share this perception [Knapp, 1988], but financial analysts do not [Imhoff, 1988]. DeAngelo [1981a] provides an analytical basis for a relationship between CPA firm size and audit opinion quality. For purposes of her analysis, DeAngelo assumes that lower audit quality causes loss of clients. This assumption is not valid in the audit market unless differences in audit quality are observable.

Three observable indicators of quality differences in audit opinions have been researched: industry expertise, litigation experience, and opinion shopping. Industry expertise could represent better evidence-gathering ability, litigation experience is a surrogate for failure to perform a quality audit, and opinion shopping is an indicator of the auditor's inability to resist client pressure and report truthfully.

5.1.1 INDUSTRY EXPERTISE

The Big 8 firms may differ from other CPA firms in terms of their industry expertise. Evidence from empirical studies indicates that industry market share is positively associated with client acquisition [Danos and Eichenseher, 1982; Eichenseher and Danos, 1981; Williams, 1988], although Palmrose [1984, 1986a] found no significant association between industry market share and audit fee. Industry market share was discussed as a possible indicator of scale economies in section 4.1. If greater efficiencies are not realized by clients in the form of reduced fees, client gains associated with industry market share may represent perceptions of greater assurance being provided for the same fee.

The use of industry market share to surrogate for industry expertise in all of the empirical research allows an alternative interpretation of the results. If industry market share represents a degree of knowledge transfer between clients in the industry, its value could vary with underlying industry market conditions and the specific client's relative position in the industry. Under certain conditions, such as regulation, clients may be indifferent to knowledge sharing within the industry. Thus, regulated industries support higher auditor concentrations [Danos and Eichenseher, 1986]. However, the potential for knowledge sharing may be a competitive disadvantage for the auditor in some industries, like soft-drink manufacturers [Berton and Niebuhr, 1990]. Industry expertise is discussed again in section 7.1.1.

5.1.2 LITIGATION

Palmrose [1988] compared the litigation activity of CPA firms as a means of assessing audit opinion quality. She argued that an inverse relationship between litigation activity and audit opinion quality should exist. Empirically, comparisons of total legal actions and resolutions of those actions among Big 8 firms and between Big 8 firms as a group and other CPA firms as a group were made for all actions identified between 1960 and 1985. Although differences among the Big 8 were dependent on the measures employed, all analyses produced statistically significant evidence that Big 8 firms as a group had a better litigation record than other CPA firms as a group. These results suggest that the Big 8 provide a higher quality audit opinion.

However, differences in client characteristics also could explain the difference in litigation records among CPA firms. Palmrose [1987] further analyzed her sample and found that litigation against the auditor is associated with business failures involving management fraud. St. Pierre and Anderson [1984] reported other situational characteristics of the client and the client's industry (such as size, regulation, and complexity) that are associated with client litigation.

In her analysis, Palmrose [1988] scaled actions involving public companies by an estimate of the number of public companies audited by CPA firms and total actions by estimated CPA firm revenue. With the assumptions that (1) the Big 8 client base includes companies that are larger than those of other CPA firms and (2) business failures and frauds occur more frequently among smaller business, scaling litigation records by number of public companies audited and CPA firm revenue could bias results in favor of the Big 8. Palmrose [1991] reported evidence indicating that, for cases involving public clients, CPA firms may be more willing to litigate, and this litigation posture may lower the incidence of lawsuits against the auditor. A comparison of litigation activity may reflect underlying litigation postures of CPA firms as well as audit quality.

5.1.3 OPINION SHOPPING

Levinthal and Fichman [19881 examined the effect of the duration of an auditor-client relationship on the client's propensity to switch auditors. They found that the rate of switching was relatively low but increasing during the first three years of a relationship, was highest during the fourth year, and declined thereafter. They also reported that clients receiving a qualified opinion are more likely to switch auditors, and the likelihood of receiving a qualified opinion follows the same pattern as the rate of switches. Levinthal and Fichman's findings support the general contention that a qualified opinion is an indicator of conflict in the auditor-client relationship.

Other studies of auditor switches have directly considered the possibility that switching companies shop for audit opinions. Two research questions have been addressed: does opinion shopping occur, and are other CPA firms more susceptible to opinion shopping than the Big 8? The first question is important because opinion shopping must occur before the second question can be addressed.

Research has reported little indication of opinion shopping. Specifically, researchers have reported no significant association between auditor switch and(l) change in accounting principles [Burton and Roberts, 1967], (2) change in management [Chow and Rice, 1982; Schwartz and Menon, 1985; Williams, 1988], (3) receipt of a qualified opinion by a company prior to bankruptcy [Schwartz and Menon, 1985], or (4) subsequent financial performance reported by the client [Williams, 1988]. Bolten and Crockett [1979] reported no association between number of qualified opinions issued and CPA firm. Smith [1986] identified only five companies that switched auditors between 1977 and 1982 for which the successor auditor issued an unqualified opinion and the predecessor auditor may have continued to qualify.

One study reported evidence of opinion shopping. Mangold [1988] found that companies switching auditors had received significantly more qualified opinions in the two years prior to, and in the second year following, a switch than companies that did not switch. Switchers and nonswitchers did not differ significantly in the year of the switch or in the year immediately following the switch. She also found that companies with poorer earnings have a greater tendency to switch auditors and that earnings were higher than expected after the switch. Although Mangold interprets these results as opinion shopping, the possibility exists that, among her sample companies, performance actually improved around the time of the auditor switch and, as a result, audit opinions legitimately differed.

Although evidence of its occurrence is weak, perceptions of opinion shopping persist. Smith and Nichols [1982], Mangold [1988], Teoh [1989], and Albrecht [1991] all report significant negative market reactions to auditor switches following disclosure of disagreements with auditors and/or receipt of a qualified opinion. Although alternative explanations of these results, such as additional legal costs [Smith and Nichols, 1982] or economic signalling [Fried and Schiff, 1981], are possible, a perception of opinion shopping also is a viable explanation. Opinion shopping could explain a Big 8 fee premium if other CPA firms are more susceptible to opinion shopping. Several empirical studies have considered CPA firm type when examining opinion shopping.

Chow and Rice [1982] examined the possibility of an association between qualified opinions and auditor switches. They found a significant association between receipt of a qualified opinion and a tendency to switch auditors within a year. Next, they investigated whether Big 8 firms differed among themselves, and from other CPA firms, in their tendency to issue qualified opinions. Three Big 8 firms had a significantly greater tendency to issue qualified opinions. Other CPA firms, as a group, were not significantly different from the remaining Big 8 firms.

Having established an association between switching and receipt of a qualified opinion and having found a difference in the propensity of firms to issue qualified opinions, Chow and Rice then investigated whether clients that switch are likely to choose an auditor who gives fewer qualified opinions and whether clients who switch after receiving a qualified opinion tend not to receive a qualified opinion from the subsequent auditor. They found little support for either proposition.

McConnell [1983] investigated whether the Big 8 firms differed from other CPA firms in their propensity to report disagreements with clients in the 8-K filing when auditor switching occurs. Although he found that the Big 8 disagree more often than other CPA firms, he also reported that the Big 8 accept more clients with disagreements. In fact, McConnell reports a significant positive correlation within the Big 8 between firms that accept clients with disagreements from their prior auditor and those that gave few disagreements to their prior clients.

The difference between Chow and Rice's and McConnell's results may be attributable to the event of interest. Chow and Rice considered the audit opinion directly. McConnell considered the auditor's propensity to report a disagreement to the Securities and Exchange Commission (SEC) and assumed a link between that action and the audit. His assumption may not be valid. Two of the three firms reported by McConnell as having a lower tendency to disagree were reported by Chow and Rice as having a higher tendency to issue qualified opinions. Both studies provide support for more difference in the quality of audit opinions among the Big 8 firms than between the Big 8 and other CPA firms.

Wyer et al. [1988] compared audit opinions issued to public clients by smaller CPA firms (those with ten or fewer public clients) to Big 8 audit opinions. No statistically significant difference in the percentage of qualified audit opinions was reported.

5.1.4 SUMMARY AND CONCLUSION

Overall, there is little evidence that the quality of the audit opinion differs between Big 8 and other CPA firms. In Palmrose's comparison of litigation activity between CPA firms, Big 8 firms as a group had a better record than other CPA firms as a group and may produce higher quality audit opinions. However, litigation activity can be influenced by client characteristics and the CPA firm's litigation posture. Palmrose did not control for these factors. If audit opinion quality drives the Big 8 fee premium as DeAngelo suggests, quality differences should be observable. The empirical studies of industry expertise, litigation, and opinion shopping have not provided convincing evidence of an observable quality difference.

5.2 SIGNALING

A pure signal sends a message about an attribute but does not assist in the achievement of the attribute. Analytical models demonstrate that the type of auditor can serve as a signal that affects the price of IPOs [Titman and Trueman, 1986] and the percentage of retained ownership [Datar et al., 1991]. To serve as a useful signal to investors, auditor type should represent levels of audit quality, with higher quality auditors receiving fee premiums and possessing a comparative advantage in providing information of value to owners and investors [Titman and Trueman, 1986]. In other words, the marginal benefit provided by the signal must equal or exceed its marginal cost. Since the Big 8 receive fee premiums in the small-client market, evidence that they serve as a useful signal to investors would be consistent with their providing greater value. Two sources of evidence that the Big 8 serve as a signal have been investigated: the effect of auditor choice on the price of IPOs and the association between client company characteristics and auditor choice.

Beatty [1989] directly tested an analytical prediction of Titman and Trueman [1986] and found that IPO clients of larger CPA firms earned lower initial returns for investors than other IPOs. The choice of a larger audit firm seemed to increase the price obtained by the IPO client.

Palmrose [1984] searched for an association between type of CPA firm employed as auditor and company characteristics. She found only size significantly associated with the use of Big 8 auditors. However, her sample of 276 companies included only 39 that were audited by other CPA firms. The lack of variability in type of auditor may have reduced her ability to detect associations. In addition, she was comparing current measures of company characteristics with a choice that may have been made years before.

Three studies of the relationship between CPA firm type and company characteristics have examined companies that switch auditors. These studies considered various measures of client size, leverage, and growth during the period immediately preceding and following the switch and changes in these measures during these periods. All found that a switch to a Big 8 firm is associated with increasing size, greater leverage and higher growth [DeFond, 1992; Eichenseher and Shields, 1989; Francis and Wilson, 1988].

Johnson and Lys [1990] found a significant association between auditor size and (1) pre-switch growth in total assets, (2) post-switch ratio of new debt plus new equity to total assets, and (3) the pre-and post-switch ratio of cash flow to total assets. Healy and Lys [1986] investigated clients' reaction to the merger of their auditor's CPA firm with a Big 8 firm. They report that clients staying with the Big 8 firm are larger and exhibit higher asset growth over the three years prior to the merger. Simunic and Stein [1987] found that larger and less risky clients were audited by Big 8 firms.

Although the variables differ, all results indicate that companies selecting Big 8 firms as auditors tend to be growing and less risky. These results are consistent with the hypothesis that a switch to a Big 8 firm is a signal. Investigations of investors' reactions to auditor switches also provide evidence consistent with signaling.

Eichenseher et al. [1989] found significant negative reactions to changes from Big 8 to other CPA firms and to changes between the same CPA firm type. They also report that, in the two years after a switch, companies that switch to the Big 8 have higher total assets, companies that switch from the Big 8 have lower total assets, and those that switch without changing CPA firm type remain about the same. Danos and Eichenseher [1986] found that the long-run equilibrium market share for the Big 8 is highest among companies that are large, growing, or operate in regulated industries.

The empirical evidence consistently indicates that type of auditor is associated with client performance. Such an association is consistent with successful signaling. However, rather than serving as a pure signal, a Big 8 auditor could assist in the achievement of client performance by providing advisory services to management or providing "deep pocket" insurance to investors.

5.3 SUMMARY

The empirical evidence on product differences between the Big 8 and other CPA firms is inconclusive. Audit opinion quality differences are not apparent. Palmrose presented some evidence of a better litigation record among Big 8 firms, but client base-rate differences were not controlled, nor were potential differences in CPA firms' litigation postures. Evidence of opinion shopping in the audit market is scarce, as is any evidence that its rate of occurrence differs between CPA firm types. The empirical associations between lower initial returns and auditor type in the IPO market, and between client characteristics and choice of auditor type, indicate that auditor type may serve as a signal. However, rather than serving as a pure signal, auditor types may differ in their ability to assist clients' achievement of signaled characteristics.

The primary beneficiaries of the audit product are users of financial statement information. Management benefits secondarily from the assurance that the audit opinion provides users. An auditor can diversify product offerings by providing services that benefit management primarily, and investors secondarily, by improving management's performance. These products are called advisory services and will be considered next.

6.0 DIVERSIFICATION

Diversification can be product extension, geographic extension, or pure. Since diversification is neither horizontal nor vertical, some economists doubt that it affects competition and performance in the market at all. The simplest hypothesis is the view that diversification does not reduce competition but, instead, can provide true economies by facilitating technology transfer, capital allocation, and risk pooling. However, competition can be harmed if diversification results in cross-subsidizing.

The quality, quantity, and nature of advisory services provided directly to client management by a CPA firm (e.g., the extent of product diversification) may vary between firms. To the extent that audit and advisory services subsidize each other, advisory services may effect audit fees, costs, and the structure of the audit market. In particular, if knowledge gained on the audit also can be used for advising management, the resulting knowledge spillovers increase the value of the audit and the audit fee. Greater knowledge spillovers could explain the Big 8 audit fee premium.

Analytically, Antle and Demski [1991] examined conditions affecting the bundling of advisory and auditing services. The knowledge spillover was information regarding the cost of performing audit and advisory services for a particular client. Such information may give rise to economies of scope. As a result, the client may benefit by using a single source for both activities rather than multiple sources. Antle and Demski found the determining factor to be the ex ante relation among costs of the services. A positive relationship between the costs prevents economies of scope while a negative relationship allows economies of scope.

Empirically, Simunic [1984] used the audit fee model from Simunic [1980] to examine a sample of companies that employed a Big 8 firm as both consultant and auditor, with performance of advisory service included as another factor. He found that audit fees are higher when advisory services also are performed. Palmrose [1986b], however, found that audit fees were higher for clients that purchase advisory services regardless of who provides the service. Knowledge spillovers cannot occur when the regular auditor does not provide the advisory services. Beck et al. [1988] suggested that recurring and nonrecurring advisory services may have different spillover effects, but they were unable to report strong evidence of any knowledge spillovers empirically.

The purchase of advisory services may be associated with company-specific problems or changes rather than with audit knowledge spillovers. The same problems or changes within the company may create complexities for the audit and, consequently, increase the audit fee. However, the potential for economies of scope may vary with the type of advisory service. DeBerg et al. [1991], for example, found that the amount of systems work purchased from the auditor significantly decreased after an auditor switch. Generally, the evidence does not support an explanation that the provision of advisory services is an additional product for which an audit fee premium is charged by the Big 8 in the overall audit market.

7.0 SUMMARY, DIRECTIONS FOR FUTURE RESEARCH AND CONCLUSIONS

This article employs empirical results about auditor concentration, audit fee, and auditor choice to investigate CPA firms' behavior in the U.S. market. Mainstream economic researchers believe that market concentration is closely related to profitability and that a high degree of oligopoly is likely to cause excess profits. High and increasing levels of auditor concentration raise concerns about oligopoly and competition in the audit market. However, classification of markets as tight or loose oligopolies only on the basis of supplier concentration levels is insufficient for determining supplier behavior [Shepherd, 1990]. Audit fee research has presented evidence of price competition in which the Big 8 are able to earn fee premiums. Research efforts attempting to explain the combination of high concentration and fee premiums have met with little success.

Product differences investigated by prior research have exhibited little evidence of quality differences in audit opinions between Big 8 and other CPA firms. Industry expertise is confounded by the underlying industry characteristics. Comparisons of litigation activity between CPA firms are difficult to interpret because data are incomplete and relevant differences in client characteristics are not controlled. Little evidence of opinion shopping has been detected by researchers despite the volume of anecdotal evidence supporting opinion shopping.

Although the use of auditor type as a signal of client performance would explain associations between CPA firm type and client characteristics, a pure signal would not assist client performance. Big 8 audit fee premiums may represent the value of signaling, but additional fees also may represent the value of advisory services that help client performance. However, Big 8 audit f premiums persist even when advisory services are purchased from another supplier. The purchase of advisory services may indicate audit complexity rather than knowledge spillovers, although no direct tests of an association between the purchase of these services and audit complexity have been reported.

The remainder of this section provides some direction for future investigations of the organization of the audit industry and concluding remarks.

7.1 MARKET DEFINITION

A market is a group of buyers and sellers exchanging services that are highly substitutable for one another. Within a larger market, submarkets can exist that define the relevant range of choices for many buyers. Market and submarket boundaries usually are established on the basis of substitutability of services at competitive prices. We suggest that dividing the audit market into submarkets by seller and/or buyer type could further our understanding of the entire audit market.

7.1.1 BUYERS

Studies that divided the U.S. audit market into small-and large-client submarkets have not defined the small-client market consistently (see table 7). Most data have been collected in the small-client market. Securities regulations create inelastic demand for external auditing in the large-client market, a condition that raises concern about the possibility of collusive behavior. However, large clients are characterized by complex corporate governance systems that include audit committees, internal auditors, management compensation packages, and strong control structures, any of which may substitute for external audit services.

In the large-client market, the Big 8 primarily compete with each other and the various company substitutes for external audit services. Wallace [1984], for example, reported a significant negative association between internal audit cost and external audit fees. However, cross-elasticities of demand for audit and audit substitutes are difficult to measure. Empirical surrogates for internal auditing have not been well specified, and examinations of other external audit substitutes have not been controlled.

Shockley and Holt [1983] found that bankers use industry expertise to distinguish between Big 8 firms, but mixed results have been reported for the effect of industry expertise on decisions to employ Big 8 or other CPA firms [Eichenseher and Shields, 1983; Lynn, 1987]. Industry expertise may be an important competitive factor in the large-client market, where Big 8 firms compete with each other, but less of a factor in the small-client market when the competition is between Big 8 and other CPA firms.

More research could be directed toward the effect of CPA firm industry expertise and specific market conditions within the industry on the audit process. For example, when the banking industry was highly regulated, the audit market in that industry was highly concentrated, and the dominant CPA firms developed audit processes aimed specifically at that industry [Kelly et al., 1986]. As banking is deregulated and the competitive environment within the industry changes, banks may be less willing to share the same auditor and, as a result, CPA firms may be less willing to develop specialized audit processes for the industry.

Future research also should consider measures other than industry market share to surrogate for industry expertise. If the minimum level of industry expertise required by the auditing standards is sufficient for competitive purposes, a dichotomous measure that indicates at least one client in the industry may be an appropriate surrogate. The percentage of a CPA firm's total revenue generated by clients in the industry also may surrogate for industry expertise.

7.1.2 SELLERS

Generally, research has grouped CPA firms into two types: Big 8 and other. Although the difference between CPA types is important to understanding the nature of the audit market, comparisons between individual Big 8 firms also could aid understanding. Some studies [e.g., Beatty, 1989; Francis and Wilson, 1988; Johnson and Lys, 1990] have utilized relative size to discriminate among CPA firms, rather than rely on the dichotomous typing (Big 8 and other). The results indicate that Big 8 firms differ among themselves, but how they differ has not been fully explored.

An examination of auditor switches among the Big 8, for example, presents a different picture of winning and losing firms than does a comparative picture of Big 8 firms' overall market share gains [Kauffman and Yardley, 1991]. Market share gains among the Big 8 appear to be driven by growth of continuing clients. Haskins and Williams [1988] found that client mergers may be a primary source of continuing client growth. Change in overall market share in the small-client market probably is more sensitive to auditor switches than to client growth because any single client represents a small share of the market.

Future research should consider the relationship between CPA firm size and client size. The need for a large CPA firm in the large-client market is obvious. A similar fit may exist in the small-client market. Diseconomies of scale, reduced or nonexistent quasi-rents, and disadvantages of a structured audit approach may give smaller CPA firms a competitive advantage in the small-client market.

7.2 MARKET STRUCTURE

Although market shares of individual Big 8 firms and market concentration have been measured for several decades (see table 2), measurement problems underlie all results. Typically, market share is measured as a portion of an industry's total sales revenue. Audit market researchers have not had access to audit revenue figures and, instead, surrogate audit revenue with functions of client revenues. Mozier and Turley [1987] and Tomczyk and Read [1989] empirically examine the error inherent in this surrogate (see n. 1).

Economists typically analyze the effect of market structure on firm performance by comparing market share to rate of return. Profitability data are unavailable for firms in the audit market. Assuming CPA firms are effectively competing, a stronger case could be made if more data were available.

The set of mergers that reduced the Big 8 to the Big 6 may have important implications for the structure of the audit market. In other industries, mergers are often blamed for increasing concentration. Ravenscraft and Scherer [1987] report that the main reasons for mergers are market power, profits, technical economies, and pecuniary economies. They found that, on average, operating efficiency fell following a merger, and they conclude that many mergers reflect empire building rather than efficiencies. Few empirical studies (see Tonge and Wootton [1991] and Minyard and Tabor [1991]) have considered the effect of CPA firm mergers on market concentration and consideration should be given to the effect on efficiencies as well.

Innovation may be a major effect--perhaps the major effect--of market structure. In the audit market, technological progress is primarily in the form of such process innovations as statistical techniques for analytical review [Stringer and Stewart, 1986], which raise the ratio of total outputs compared with total inputs. Generally, effective competition optimizes innovation.

Few audit innovations have been identified and researched. An exception is the structured audit approach. Cushing and Loebbecke [1986] were the first to describe a difference in audit structure among large CPA firms. Subsequently, Kinney [1986] confirmed the difference among the Big 8. Research on the degree of structure in the audit process has focused on differences among the Big 8, but it is probable that other CPA firms are far less structured than the least structured Big 8 firm. Although the effect of structure has been found to have some behavioral implications [Bamber and Snowball, 1988; Bamber et al., 1989; Mutchler and Williams, 1990], there are only weak indications of differences in audit efficiency [Newton and Ashton, 1988; Williams and Dirsmith, 1988].

If audit processes differ in a manner that decreases other audit-related costs to the client [Turpin, 1990], differences in audit processes could explain fee premiums earned by the Big 8 in the small-client market. For example, if a structured audit requires less client assistance or causes less disruption of the client' s business, the client may be willing to pay additional fees for the avoidance of these costs. Future research could consider total costs of the audit to the client and the effect of innovations in the audit process.

7.3 DETERMINANTS OF MARKET STRUCTURE

Barriers to entry are the main determinants of high market concentration. Although research has considered economies of scale and product differentiation, neither these nor other ideas related to barriers have been fully explored.

7.3.1 ECONOMIES OF SCALE

Scale economies occur at the range of output where the average cost curve is lowest. Two types of cost conditions, technical and pecuniary, can lead to scale economies. Technical economies of scale arise from genuine increases in efficiencies as a result of specialization, physical laws, or management efficiencies. Pecuniary economies of scale arise from the lower input prices paid by a larger firm, such as with volume discounts or lower finance charges. Only technical economies produce true social gains.

Methods for measuring economies of scale include (1) expert estimates of optimal scale and costs, (2) actual average cost data, and (3) survivor technique. Only the third method has been applied to the audit market. However, the survivor technique fails to distinguish between technical and pecuniary economies. In other industries, net pecuniary gains to size appear to be significant. Studies done in the 1960s of various U.S. industrial markets provided little justification for market shares above 20 percent on the basis of technical efficiency [Shepherd, 1982]. Although cost data are unavailable in the audit market, expert estimates have been employed in a number of economic studies (e.g., Bain, 1956; Weiss, 1976) and may be a feasible technique to apply in audit studies.

7.3.2 PRODUCT DIFFERENTIATION

Wallace [1980] suggests that the insurance provided by the auditor's "deep pockets" is valued by external parties. Although prior research has employed CPA firm size as a surrogate for audit opinion quality, size also could surrogate for insurance. Wallace [1989] failed to find indirect support for the insurance hypothesis. She reported that audit fees, unlike insurance premiums, are not adjusted for risk. A more direct test of the insurance hypothesis may be possible by studying auditor behavior, auditor insurance premiums, cost of capital, and audit fees in jurisdictions where "deep pocket" verdicts have been issued.

7.3.3 OTHER IDEAS

Control of entry is a type of collusion that can lead to limit pricing. The limit price is the highest price that existing firms can charge without causing new firms to enter a market. Contestability theory, or ultra-free entry, proposes that entry threat dominates industry pricing; structure is secondary in importance to possible entry from outside [Baumol et al., 1982]. Theoretically, a contestable market meets three criteria: (l)entry is free and without limit, (2) entry is absolute, and (3) entry is perfectly reversible. Under ultra-free entry, the threat of entry will drive prices down and guarantee efficiency even if there is just one firm in the market. The idea of several CPA firms' practicing a unified limit-pricing policy seems implausible. However, if each client is considered a market, the criteria for a contestable market are almost met, and limit pricing may provide the best explanation of CPA firms' pricing behavior.

7.4 DIVERSIFICATION

Product diversification affects audit competition only to the extent that profits from advisory services are used to subsidize audit fees. Future consideration of advisory services should consider the type of service being purchased. Advisory services are so different that the availability of economies of scope probably varies with the advisory service performed. Investigation of the relationship between audit costs and costs of advisory services for different types of service should be undertaken. Beck et al., [1988] and DeBerg et al., [1991] report differences in the effect of different advisory services on auditor-client relationships. Consideration also should be given to the possible interaction between industry specialization and the kind of advisory service provided.

7.5 FIRM PERFORMANCE

According to the Chicago-UCLA school of thought, superior performance is the primary cause of high market concentration. Superior performance is usually stated in terms of lower costs, and one indicator is higher profitability. In the audit market, only fees and surrogate measures of fees are available. Thus, direct comparisons of costs across firms are not possible, and even relative profitability cannot be measured.

Another measure of firm performance is X-efficiency, or the ability of the firm to keep costs at a minimum level. Measures of X-efficiency are primitive, but most economists recognize that X-inefficiency exists to a varying extent between competitors [Scherer, 1970]. Attempts to measure X-efficiency in CPA firms, perhaps by employing realization rates, could be interesting.

CPA firms often complain of severe price competition (see, e.g., Berton [1991l). Maher et al. [1992] report a significant decrease in real audit fees from 1977 to 1981 that could not be explained by changes in client size, risk, or complexity. Destructive competition may occur when demand falls below full-capacity levels for the firm and the firm cuts prices to marginal cost. If overhead costs are high and marginal costs are low, sustained price cuts cause failures that may reduce industry capacity below the efficient long-run level for the industry. Shepherd [1990, pp. 340-41] discusses two main cases of possible destructive competition: cyclical shifts in demand and long-term "sick" industries. In either case, price cutting is usually healthy and complaints usually have little economic substance. Accounting researchers should consider whether the seasonal nature of auditing creates a cyclical demand that is destructive to competition, or whether the audit industry is "overbuilt" and shrinking.

7.6 FIRM BEHAVIOR

An aspect of CPA firm behavior that has been given little attention in the literature is advertising. Advertising, which is an alternative to price cutting as a competitive tactic, is classified as informative or persuasive. Informative advertising can be socially beneficial, but persuasive advertising merely tries to change consumer preferences or to divert attention from facts to images. The larger effect of advertising is to raise entry barriers. A tight oligopoly tends to breed intensive advertising because it is a safer method of competition than price-cutting. The advertising-to-sales ratio is commonly used to indicate advertising intensity.

Advertising can be a powerful device by which new or small firms succeed, or it may just be a process by which leading firms invest in customer loyalty. Accounting researchers could evaluate advertising in the audit industry to determine whether most is informative or persuasive and whether Big 6 or smaller CPA firms have the higher advertising-to-sales ratio.

7.7 CONCLUSIONS

The analysis of industrial organization is a complex undertaking that can benefit regulators, educators, and professionals by providing a better understanding of the audit market. No single measurement contains sufficient information to determine the nature of competition in the industry, and market structure, determinants of structure, firm behavior, and firm performance are each difficult to study in the audit market. We are not suggesting that all aspects of industrial organization must be considered in a single study; we suggest only that any interpretation of results should consider other aspects of industrial organization and make clear the theoretical perspective of the author.

Researchers should continue to evaluate the audit market from the traditional points of view of structuralists and the Chicago-UCLA school of thought. However, unique characteristics of the audit market and researchers' intimate knowledge of audit processes may provide an opportunity for audit research to contribute to the study of industrial organization, rather than simply borrow from it. Real examples of contestable markets are rare, and the audit market may serve as an example. Economists developed their theories by studying industrial markets, but their theories may not apply to markets for services. Services cannot be wholesaled, and each product is somewhat unique. The extent of product differentiation, price discrimination, and buyer knowledge of the product may be different in service markets. The audit market may serve as an example of first-degree product differentiation and price discrimination, with little opportunity for buyer retaliation (see Pigou [1920]).

Economic theory provides a useful framework for evaluating the audit market, but other disciplines provide alternative perspectives. Levinthal and Fichman [1988] view the affiliation between auditor and client as a dyadic interorganizational relationship. Social exchange theory, as opposed to strictly economic exchange theory, considers investments in relation-specific assets, such as knowledge of the client's accounting system and interpersonal relationships, that develop commitment between exchange partners.

Simunic and Stein [1990] use portfolio theory to model the auditor's client-selection process. Their model may help explain the auditor-client commitment as a two-way fit rather than a simple buyer-seller relationship. If large CPA firms have private information about client performance, they may be more selective in their choice of client companies than smaller CPA firms. This is a shift in emphasis from auditor selection to client selection.

The audit market is not yet well-understood. Continued research provides hope for increased understanding of the behavior of CPA firms in the market as well as a possible contribution to the underlying discipline.

ANNOTATED BIBLIOGRAPHY

1. Beatty, R. P. 1989. Auditor reputation and the pricing of initial public offerings. The Accounting Review (October): 693-709.

Investigated the impact of auditor reputation on the pricing of IPOs, Beatty hypothesized that a higher audit fee is a premium for reputation capital and that companies will pay the premium to increase the price received for their IPOs. A regression analysis on 1,807 IPOs from 1975 to 1984 revealed that (1) Big 8 firms receive a fee premium, (2) increased investor confidence is associated with the Big 8, and (3) lower ex post return deviation is associated with higher auditor fees. Beatty also partitioned the CPA firms into "largest five, middle six, and smallest nine" and found similar results, which indicates that firms may differ even within the Big 8.

2. Chow, C. W., and S. J. Rice. 1982. Qualified audit opinions and auditor switching. The Accounting Review (April): 326-35.

Chow and Rice examined the possibility of an association between qualified opinions and auditor switches. Their analysis included all public companies that received a qualified opinion in 1973 or 1974 and/or switched auditors in 1973. A logit analysis showed that qualified opinions were significantly associated with switches and that different Big 8 firms have different propensities to issue qualified opinions. However, clients that switched auditors were not likely to chose an audit firm with a lower propensity to issue qualified opinions, and clients who switch after receiving a qualified opinion tended to receive a qualified opinion from the subsequent auditor.

3. Danos, P., and J. W. Eichenseher. 1986. Long-term trends toward seller concentration in the U.S. audit market. The Accounting Review (October): 633-50.

Danos and Eichenseher examined factors that may affect the trend toward seller concentration in the market for audits. Big 8 market share in 16 submarkets defined by factorial combinations of client size, growth, capital market activity, and industry regulatory status was examined from 1965 to 1972 and from 1973 to 1980. Results indicate that the long-run equilibrium Big 8 market share is high and perhaps increasing in all subsections of the market. However, increases were greater among large and growing companies from 1964 to 1972 and among regulated companies from 1972 to 1980.

4. Francis, J. R., and D. T. Simon. 1987. A test of audit pricing in the small-client segment of the U.S. audit market. The Accounting Review (January): 145-57.

Francis and Simon examined the relationship between audit fee and auditor firm size for audits of small, publicly traded companies. Information about audits conducted in 1984 and 1985 was gathered from 210 companies with sales of less than $125 million. A regression model that utilized audit cost factors and categorical variables for Big 8, second nine, and nonnational CPA firm types indicated that Big 8 audit fees are significantly higher than other CPA firm types.

5. Johnson, W. B., and T. Lys. 1990. The market for audit services: Evidence from voluntary auditor changes. Journal of Accounting and Economics 12(1-3): 281-308.

Johnson and Lys tested whether changes in clients' financing, investing, and operating activities result in a switch to specialized and lower-cost auditors. Their study included 603 companies that switched auditors from 1973 to 1982. Univariate and multivariate analysis revealed that factors related to client expansion, financing, and profitability explain client switches to larger CPA firms. Johnson and Lys also reported no significant stock market reaction to an 8-K filing of an auditor switch. Generally, companies switching auditors were less profitable than a similar group of companies not switching auditors and had negative cumulative average market returns.

6. Palmrose, Z. 1986a. Audit fees and auditor size: Further evidence. Journal of Accounting Research (Spring): 97-110.

Palmrose investigated whether there is a relation between audit fee and auditor size. She gathered 1980 information from 361 companies. Audit cost factors, auditor industry market share, and audit firm type (Big 8 or not) were regressed on the log of audit fees for each client industry. All variables were significantly associated with audit fee except industry market share. Results also showed that hours of audit time, as estimated by clients, were higher for the Big 8.

7. Palmrose, Z. 1996b. The effect of nonaudit services on the pricing of audit services: Further evidence. Journal of Accounting Research (Autumn): 405-11.

Palmrose investigated the effect of provision of management advisory services (MAS) on audit fees using data from Palmrose [1986a]. She found that larger clients were more likely to purchase MAS and that audit fees were higher for these clients, whether the MAS were provided by the auditor or another party.

8. Palmrose, Z. 1988. An analysis of auditor litigation and audit service quality. The Accounting Review (January): 55-73.

Palmrose analyzed 472 legal actions and 183 resolutions of actions in the United States from 1960 to 1985 against the largest 15 CPA firms. The Big 8 as a group had a significantly smaller ratio of legal actions involving public companies to number of public companies audited during 1960-85 and significantly fewer total legal actions per dollar of firm revenue than the other CPA firms as a group. Rankings among Big 8 firms were sensitive to the scaling factor employed.

9. Simon, D. T., and J. R. Francis. 1988. The effects of auditor change on audit fees: Tests of price cutting and price recovery. The Accounting Review (April): 255-69.

Simon and Francis examined whether price cutting occurs on initial audit engagements and, if so, when the price returns to normal. Analysis was performed on 214 (226) public companies that switched (did not switch) auditors between 1979 and 1984. Expected audit fees from a cost-based regression model were compared to actual fees. Switches within CPA firm type, qualified opinions, MAS, and client size were controlled. Discounts of 24 percent in the initial year and 15 percent in the next two years were measured before fees returned to expected levels. In all years, a Big 8 fee premium persisted.

10. Simunic, D. A. 1980 The pricing of audit services: Theory and evidence. Journal of Accounting Research (Spring): 161-90.

Simunic searched for empirical evidence of noncompetitive pricing in the market for audits. He gathered audit fee data from 397 companies. Having shown lower Big 8 concentration in the small-client audit market (sales < $125 million), results from a regression analysis employing small-client data served as a benchmark for measuring competition in the large-client audit market. The regression model included three audit cost categories (loss exposure, loss sharing, and auditor production function) and a dichotomous variable for CPA firm type (Big 8 or not). Simunic concluded that competition prevails in both large and small client segments of the audit market.

1 Use of the square root of revenue as a surrogate for audit fees was supported in work by Elliot and Korpi [1978] and Simunic [19791. but two subsequent studies investigated the biases inherent in surrogates for audit fees. Moizer and Turley [1987] reported that the square root of client sales provide a lower bound of market concentration based on audit fees. Tomczyk and Read [1989] reported opposite results. They computed lower average concentration measures by using direct estimates of the 28 largest CPA Firms' audit revenues than with measures of the square root of clients sales as surrogates for audit fees (see table 3).

2 Who Audits America includes information on more that 8,000 publicly held companies.

3 Studies consistently report that company size is closely associated with audit fee. Simunic [1980] reported that a regression of assets on fees provides an R sup 2 of 0.57; Firth [1985] reported 0.64. Simunic [1980] reported a coefficient of 0.45 for the log transformation of assets; Francis [1984] and Palmrose [1986a] reported coefficients of 0.46 and 0.47.

REFERENCES

Albrecht, W. D. 1991. The determinants of the market reaction to an announcement of a change in auditor. Ph.D. dissertation, Virginia Polytechnic Institute and State University.

Blacksburg. Antle R., and J. Demski. 1991. Contracting frictions, regulation, and the structure of CPA firms. Journal of Accounting Research (Supplement): 1-24.

Bain J. S. 1956. Barriers to new competition. Cambridge, MA: Harvard University Press.

Balvers R. J., B. McDonald, and R. E. Miller. 1988. Underpricing of new issues and the choice of auditor as a signal of investment banker reputation. The Accounting Review (October): 605-22.

Bamber E. M., and D. Snowball. 1988. An experimental study of the effects of audit structure in uncertain task environments. The Accounting Review (July): 490-504.

--, D. Snowball, and R. M. Tubbs. 1989. Audit structure and its relation to role conflict and role ambiguity: An empirical investigation. The Accounting Review (April): 285-99.

Bar-Yosef, S., and J. Livnat. 1984. Auditor selection: An incentive-signalling approach. Accounting and Business Research 14: 301-9.

Baumol, W. J., C. Panzar, and D. Willig. 1982. Contestable markets and the theory industry structure. San Diego, CA: Harcourt Brace Jovanovich.

Beatty, R. P. 1989. Auditor reputation and the pricing of initial public offerings. The Accounting Review (October): 693-709.

Beck, P. J., T. J. Frecka, and I. Solomon. 1988. A model of the market for MAS and audit services: Knowledge spillovers and auditor-auditee bonding. Journal of Accounting Literature (7): 50-64.

Bedingfield, J. P., and S. E. Loeb. 1974. Auditor changes--An examination. Journal of Accountancy (March): 69.

Berton, L. 1991. Accounting profession, once a staid field, is torn by incivility. Wall Street Journal (July 24): A1.

--, and R. G. Niebuhr. 1990. Ernst & Young is forced to drop PepsiCo as client. Wall Street Journal (February 26): B8.

Bolten, S., and J. Crockett. 1979. How independent are the independent auditors? Financial Analysts Journal (November-December): 7-8.

Burton, J. C., and W. Roberts. 1967. A study of auditor changes. Journal of Accountancy (April): 31-6.

Chow, C. W. and S. J. Rice. 1982. Qualified audit opinions and auditor switching. The Accounting Review (April): 326-35.

Colson, R. H., M. W. Maher, A. J. Broman, and P. Tiessen. 1988. Audit pricing models for regulation research: Reliability and stability. Research in Accounting Regulation 2: 61-79.

Cushing, B. E., and J. K. Loebbecke. 1986. Comparison of audit methodologies of large accounting firms. Studies in accounting research No. 26. Sarasota, FL: American Accounting Association.

Danos, P., and J. W. Eichenseher. 1982. Audit industry dynamics: Factors affecting changes in client-industry market shares. Journal of Accounting Research (Autumn): 604-16.

--, and J. W. Eichenseher. 1986. Long-term trends toward seller concentration in the U.S. audit market. The Accounting Review (October): 633-50.

Datar, S., G. A. Feltham, and J. S. Hughes. 1991. The role of audits and audit quality in valuing new issues. Journal of Accounting and Economics 14 (1): 349.

DeAngelo, L. E. 1981a. Auditor size and audit quality. Journal of Accounting and Economics 3(3): 183-99.

--. 1981b. Auditor independence, "low balling," and disclosure regulations. Journal of Accounting and Economics 3(2): 113-27.

--. 1982. Mandated successful efforts and auditor choice. Journal of Accounting and Economics 4(3): 171-203.

DeBerg, C. L., S. E. Kaplan, and K. Pany. 1991. An examination of some relationships between non-audit services and auditor change. Accounting Horizons (March): 17-28.

DeFond, M. 1992. The association between changes in client firm agency costs and auditor switching. Auditing: A Journal of Practice and Theory (Spring): 16-31.

Demsetz, H. 1973. The market concentration doctrine. Stanford, CA: American Enterprise Institute-Hoover Institute on War, Revolution and Peace.

--, 1974. Two systems of belief about monopoly. In Industrial concentration: The new learning, edited by H. J. Goldschmid et al. Boston, MA: Little, Brown & Company, 1684.

Dopuch, N., and D. Simunic. 1980. The nature of competition in the auditing profession: A descriptive and normative view. In Regulation and the accounting profession, edited by J. W. Buckley and F. Weston. Belmont, CA: Lifelong Learning Publications, 77-94.

Eichenseher, J. W., and P. Danos. 1981. The analysis of industry-specific auditor concentration: Towards an explanatory model. The Accounting Review (July): 479-92.

--, M. Hagigi, and D. Shields. 1989. Market reaction to auditor switches by OTC companies. Auditing: A Journal of Practice and Theory (Fall): 29-40.

--, and D. Shields. 1983. The correlates of CPA-firm change for publicly-held corporations. Auditing: A Journal of Practice and Theory (Spring): 23-37.

--, and D. Shields. 1989. Corporate capital structure and auditor "fit." Advances in Accounting (Supplement 1): 39-56.

Elliott, R. K., and A. Korpi. 1978. Factors affecting audit fees. In Cost-benefit analysis of auditing, Commission On Auditors' Responsibilities, Research Study No. 3, edited by M. Shakun. New York: AICPA, 17-25.

Firth, M. 1985. An analysis of audit fees and their determinants in New Zealand. Auditing: A Journal of Practice and Theory (Spring): 23-37.

Francis, J. R. 1984. The effect of audit firm size on audit prices: A study of the Australian market. Journal of Accounting and Economics (2): 133-51.

--, and D. T. Simon. 1987. A test of audit pricing in the small-client segment of the U.S. audit market. The Accounting Review (January): 145-57.

--, and E. R. Wilson. 1988. Auditor changes: A joint test of theories relating to agency costs and auditor differentiation. The Accounting Review (October): 663-82.

Fried, D., and A. Schiff. 1981. CPA switches and associated market reactions. The Accounting Review (April): 326-41.

Gilling, D. M., and P. J. Stanton. 1978. Changes in the structure of the accounting profession in Australia. ABACUS (June): 6-80.

Haskins, M. E., and D. D. Williams. 1988. The association between client factors and audit fees: A comparison by country and by firm. Accounting and Business Research 18(70): 183-92.

Healy, P., and T. Lys. 1986. Auditor changes following big eight mergers with non-big eight audit firms. Journal of Accounting and Public Policy 5(4): 251-65.

Hobgood, G., and J. A. Sciarrino. 1972. Management looks at audit services. Financial Executive (April): 26-32.

Imhoff, E. A. 1988. A comparison of analysts' accounting quality judgments among CPA firms' clients. Auditing: A Journal of Practice and Theory (Spring): 182-91.

Johnson, W. B., and T. Lys. 1990. The market for audit services: Evidence from voluntary auditor changes. Journal of Accounting and Economics 12(1-3): 281-308.

Kauffman, N. L., and J. A. Yardley. 1991. The big-8 audit market. Unpublished manuscript, Virginia Polytechnic Institute and State University, Blacksburg.

Kelly, K. P., G. S. Ribar and J. J. Willingham. L986. Interim report on the development of an expert system for the auditor's loan loss evaluation. Auditing Symposium VIII, edited by R. P. Srivastava and N. Ford. Lawrence, KS: University of Kansas, 167-81.

Kinney, W. 1986. Audit technology and preferences for auditing standards. Journal of Accounting and Economics 8(1): 73-89.

Klein, B. and K. B. Leffler. 1991. The role of market forces in assuring contractual performance. Journal of Political Economy (August): 615-41.

Knapp, M. C. 1988. Toward a more coherent regulatory policy for auditor changes: An empirical study of their impact on financial statement credibility. Research In Accounting Regulation 2: 41-59.

Levinthal, D. A., and M. Fichman. 1988. Dynamics of interorganizational attachments: Auditor-client relationships. Administrative Science Quarterly (September): 345-69.

Lynn, S. A. 1987. Segmenting the CPA services market. Advances in Accounting 4: 239-66.

Maher, M., P. Tiesen, R. Colson, and A. J. Broman. 1992. Competition and audit fees. The Accounting Review (January): 199-211.

Mangold, N. R. 1988. Changing auditors and the effect on earnings, auditors' opinions, and stock prices. Ann Arbor, MI: UMI Research Press.

McConnell, D. K. 1983. Auditor changes and related disagreements. Auditing: A Journal of Practice and Theory (Fall): 44-56.

Minyard, D. H., and R. H. Tabor. 1991. The effect of big eight mergers on auditor concentration. Accounting Horizons (December): 79-90.

Moizer, P., and S. Turley. 1987. Surrogates for audit fees in concentration studies. Auditing: A Journal of Practice and Theory (Fall): 118-23.

Mutchler, J. F., and D. D. Williams. 1990. The relationship between audit technology, client risk profiles, and the going-concern opinion decision. Auditing: A Journal of Practice and Theory (Fall): 39-54.

Newton, J. D., and R. H. Ashton. 1988. The association between audit technology and audit delay. Paper presented at eighth symposium on auditing research, University of Illinois at Champaign-Urbana.

Palmrose, Z. 1984. The demand for quality-differentiated audit services in an agency-cost setting: An empirical investigation. In 1984 auditing research symposium, edited by A. R. Abdel-Khalik and I. Solomon. Urbana, IL: University of Illinois Press, 229-52.

--. 1996a. Audit fees and auditor size: Further evidence. Journal of Accounting Research (Spring): 97-110.

--. 1986b. The effect of nonaudit services on the pricing of audit services: Further evidence. Journal of Accounting Research (Autumn): 405-11.

--. 1987. Litigation and independent auditors: The role of business failures and management fraud. Auditing: A Journal of Practice and Theory (Spring): 90-103.

--. 1988. An analysis of auditor litigation and audit service quality. The Acounting Review (January): 55-73.

--. 1991. Trials of legal disputes involving auditors: Some empirical evidence. Journal of Accounting Research (Supplement): 149-85.

Pigou, A. C. 1920. The economics of welfare. London: Macmillian.

Rankin, K. 1989. Minority coalition attacks big mergers. Accounting Today (August 28): 1,17.

Ravenscraft, D. J., and F. M. Scherer. 1987. Mergers, sell-offs, and economic efficiencies. Washington, D.C.: Brookings Institute.

Rhode, J. G., G. M. Whitsell, and R. L. Kelly. 1974. An analysis of client-industry concentrations for large public accounting firms. The Accounting Review (October): 772-87.

St. Pierre, K., and J. A. Anderson. 1984. An analysis of the factors associated with lawsuits against public accountants. The Accounting Review (April): 242-63.

Scherer, F. M. 1970. Industrial market structure and economic performance. Skokie, IL: Rand McNally.

Schiff, A., and H. D. Fried. 1976. Large companies and the big eight: An overview. ABACUS (December): 116-24.

Schwartz, K. B., and K. Menon. 1985. Auditor switches by failing firms. The Accounting Review (April): 248-61.

Shepherd, W. G. 1982. Monopoly profits and economies of scale. In Industrial organization, antitrust and public policy, edited by J. Craven. Boston, MA: Kluwer Nijhof.

--. 1990. The economics of industrial organization. Englewood Cliffs, NJ: Prentice-Hall.

Shockley, R. A., and R. N. Holt. 1983. A behavioral investigation of supplier differentiation in the market for audit services. Journal of Accounting Research (Autumn): 545-64.

Simon, D. T., and J. R. Francis. 1988. The effects of auditor change on audit fees: Tests of price cutting and price recovery. The Accounting Review (April): 255-69.

Simunic, D. A. 1979. The pricing of audit services: Theory and evidence. Ph.D. dissertation, University of Chicago.

--. 1980. The pricing of audit services: Theory and evidence. Journal of Accounting Research (Spring): 161-90.

--. 1984. Auditing, consulting, and auditor independence. Journal of Accounting Research (Autumn): 679-702.

--, and M. T. Stein. 1987. Product differentiation in auditing: Auditor choice in the market for unseasoned new issues. Research Monograph No. 13. Canadian Certified General Accountants Research Foundation.

--. and M. T. Stein. 1990. Audit risk in a client portfolio context. Contemporary Accounting Research (Spring); 329-43.

Smith. D. B. 1986. Auditor "subject to" opinions, disclaimers, and auditor changes. Auditing: A Journal of Practice and Theory (Fall): 95-108.

--, and D. R. Nichols. 1982. A market test of investor reaction to disagreements. Journal of Accounting and Economics 4(2): 109-20.

Stigler. G. J. 1958. The economies of scale. Journal of Law and Economics (October); 54-71.

Stringer, K. W., and T. R. Stewart. 1986. Statistical techniques for analytical review in auditing. New York: John Wiley & Sons.

Teoh. S. H. 1989. The impact of auditor switches on accounting and market measures of value. Unpublished manuscript. University of California at Los Angeles.

Tirole, J. 1988. The theory of industrial organization. Cambridge. MA: MIT Press.

Titman, S., and B. Trueman. l986. Information quality and the value of new issues. Journal of Accounting and Economics 8(2): 159-72.

Tomczyk, S., and W. J. Read. 1989. Direct measurement of supplier concentration in the market for audit services. Auditing: A Journal of Practice and Theory. (Fall): 98-106.

Tonge, S. D., and C. W. Wootton. 1991. Auditor concentration and competition among the large public accounting firms: Post-merger status and future implications. Journal of Accounting and Public Policy 10: 157-72.

Turpin, R. A. 1990. Differential pricing on auditors' initial engagements: Further evidence. Auditing: A Journal of Practice and Theory (Spring): 60-76.

Wallace. W. A. 1980. The economic role of audits in free and regulated markets. Wilton, CT: Touche Ross Foundation.

--. 1984. Internal auditors can cut outside CPA costs Harvard Business Review (March-April): l20.

--. 1989. Are audit fees sufficiently risk adjusted? Advances in Accounting (Supplement 1): 3-38.

Weiss. L. W. 1976. Optimal p]ant size and the extent of suboptimal capacity. In Essays on industrial organization in honor of Joe S. Bain. edited by R. T. Masson and P. D. Qualls. Cambridge. MA: Ballinger. chap. 7.

Williams. D. D. 1988. The potential determinants of auditor change. Journal of Business, Finance and Accounting 15(2): 243-62.

--, and M. W. Dirsmith. 1988. The effects of audit technology on auditor efficiency: Auditing and the timeliness of client earnings announcements. Accounting, Organizations, and Society 13(5): 487-508.

Wyer, J. C., G. T. White. and E. C. Janson 1988 Audits of public companies by smaller CPA firms: Clients, reports, and quality. Auditing: A Journal of Practice and Theory (Spring): 1673.

Zeff, S. A., and R. L. Fossum. 1967. An analysis of large audit clients. The Accounting Review (April): 298-320.

TABLE 1

SUMMARY OF SOURCES OF EMPIRICAL EVIDENCE (REFERENCED TO PARTS OF THE PAPER)

SECTION 3.0 STRUCTURE OF THE MARKET FOR AUDITS

3.1 EVIDENCE OF HIGH MARKET CONCENTRATIONS

Burton and Roberts [1967]

Zeff and Fossum [1967]

Rhode et al. [1974]

Schiff and Fried [1976]

Dopuch and Simunic [1980]

Eichenseher and Danos [1981]

Tomczyk and Read [1989]

Tonge and Wootton [1991]

3.2 EVIDENCE OF FIRM PERFORMANCE

Dopuch and Simunic [1980]

Hobgood and Sciarrino [1972]

Bedingfield and Loeb [1974]

Eichenseher and Shields [1983]

Simunic [1980]

Wallace [1984]

Palmrose [1986a]

Francis and Simon [1987]

Colson et al. [1988]

Simon and Francis [1988]

Turpin [1990]

SECTION 4.0 ECONOMIES OF SCALE

4.1 EVIDENCE OF BIG 8 ECONOMIES OF SCALE

Eichenseher and Danos [1981]

Danos and Eichenseher [1982]

Haskins and Williams [1988]

Simunic [1980]

Palmrose [1984, 1986a]

Colson et al. [1988]

Turpin [1990]

Eichenseher and Shields [1983]

Shockley and Holt [1983]

Williams [1988]

Francis and Wilson [1988]

Simunic and Stein [1987]

Balvers et al. [1988]

Beatty [1989]

4.2 EVIDENCE OF BIG 8 FEE PREMIUMS

Simunic [1980]

Palmrose [1986a]

Francis and Simon [1987]

SECTION 5.0 PRODUCT DIFFERENTIATION

5.1 EVIDENCE OF AUDIT OPINION QUALITY DIFFERENCES

5.1.1 INDUSTRY EXPERTISE

Eichenseher and Danos [1981]

Danos and Eichenseher [1982, 1986]

Williams [1988]

Palmrose [1984, 1986a]

5.1.2 LITIGATION

Palmrose [1987, 1988, 1991]

St Pierre and Anderson [1984]

5.1.3 OPINION SHOPPING

Levinthal and Fichman [1988]

Burton and Roberts [1967]

Chow and Rice [1982]

Schwartz and Menon [1985]

Williams [1988]

Bolten and Crockett [1979]

Smith [1986]

Mangold [1988]

Smith and Nichols [1982]

Teoh [1989]

Albrecht [1991]

McConnell [1983]

Wyer et al. [1989]

5.2 EVIDENCE OF SIGNALLING DIFFERENCES

Beatty [1989]

Palmrose [1984]

DeFond [1992]

Francis and Wilson [1988]

Eichenseher and Shields [1989]

Johnson and Lys [1990]

Healy and Lys [1986]

Simunic and Stein [1987]

Eichenseher et al. [1989]

Danos and Eichenseher [1986]

SECTION 6.0 DIVERSIFICATION

Simunic [1984]

Palmrose [1986b]

Beck et al. [1988]

DeBerg et al. [1991