Printer Friendly

Earnings response to auditor switches using a multi-tiered auditor classification.

ABSTRACT

Previous studies have provided evidence indicating that the securities market values audits performed by large audit firms more then audits performed by smaller audit firms. This may be due to a perception that large audit firms provide higher quality audits or that large audit firms provide greater insurance to investors in the event of a loss. Findings in this study are based on studies between big-five (big-six, big-eight) audit firms and smaller audit firms. While the market may value big-five audit firms, it would be unreasonable to expect the market to treat all other audit firms equally. This study provides evidence that the market reacts differently to earnings when an audit firm of a different classification is associated with the financial information provided by the firm and is not limited to changes from or to the big-five audit firms.

INTRODUCTION

The fact that large audit firms enjoy a reputation as the premier quality auditors has been accepted in audit research as a valid construct. The recognition of this size relationship has been operationalized in research as a two-class system consisting of the "quality" big-five (The term big-five will be used in the remainder of the paper except for descriptions of previous studies. This term will also include big-eight and big-six auditors as appropriate for the time period auditors and all other audit firms.) While firms like Arthur Andersen and KPMG enjoy international reputations which are shared by only a select group of firms, other firms like McGladrey & Pullen, BDO Seidmans and Grant Thornton, enjoy national reputations that are not shared by smaller regional firms. Therefore, it may be logical to expect investors to find value in a change from a small regional firm to a non-big-five firm with a national reputation. This study evaluates the market reaction to auditor switches using a multi-tiered classification based on the number of audit clients for each audit firm. This market reaction may be the result of an increase in information quality or an increase in the insurance provided by the larger audit firm.

The recognition of a different market reaction to distinct classifications of auditors is important for two reasons. First, audit researchers have dichotomized audit quality between big-five audit firms and all other audit firms. This classification has been utilized in many studies that investigate auditor-client relationships. If the market views auditors as a multi-classification and reacts differently to each classification, this information may facilitate future research in auditor-client relationships. Second, if the market views auditors as a multi-classification, clients may need to give more consideration to the markets perceptions when selecting an auditor.

The remainder of the paper is organized in five sections. The first section outlines significant prior research in the area of auditor switching. The next section provides the theoretical development of the hypotheses being tested. This is followed by the section that describes the sample selection. The next section provides the discussion of the results. This is followed by the Conclusions of the study.

PRIOR RESEARCH

Previous studies regarding changes in auditors have focused on the reasons companies change auditors and the market reaction to changes in auditor type (i.e. big-five and non-big-five). Most of this research utilizes a dichotomous variable, where 1 is assigned to big-five firms and 0 to non-big-five firms, as either the dependent or independent variable. As an independent variable, this classification is usually used as a proxy for audit quality.

Francis and Wilson (1988) tested whether a positive relationship exists between a firm's agency costs and its demand for a quality-differentiated audit. The authors utilized two models;

1) a brand name model where the dependent variable of big-eight/non-big-eight, and

2) a continuous size model where the dependent variable was defined as the natural logarithm of the ratio of combined sales of the public companies audited by the new auditor to that of the old auditor in the year of the auditor change.

Results of the brand name model supported the agency cost relationship, however, the results of the continuous size model did not support the hypothesis.

Johnson and Lys (1990) evaluate whether changes in clients' financing, investing, and operating characteristics are related to auditor changes. They evaluate auditor changes between and among big-eight audit firms and non-big-eight firms. The authors used a cross-sectional regression and logit analysis to study the relation between relative audit firm size and the change in client characteristics for the period 1973-1982. They maintain that auditor changes are a response to shifts in the client's financing and operating characteristics that result in an auditor-client mismatch. The authors also presented an event study that evaluated common stock returns at the time of the auditor change. The results of the event study provide no statistical evidence of a market reaction to auditor changes.

In a 1993 article on perceived audit quality, Teoh and Wong provided an analysis of the market reaction to firms changing from big-eight to non-big-eight or non-big-eight to big-eight. This study evaluated market reaction to earnings during a period prior to the change in auditors with the period subsequent to the change in auditors. The results of this part of Teoh and Wong study were inconclusive with regards to a market reaction to changes in auditor.

Krishnan (1994) examined auditor switching as a function of auditor conservatism. The author concluded that switching is triggered by conservative treatment rather than by the issuance of qualified opinions. Krishnan used an ordered probit regression that includes a dichotomous independent variable BIG6. This variable is not defined as to its representation in the equation, but appears to proxy for the quality of the auditor. Krishnan et. al. (1996) indicate that auditor switching is more likely to occur when the auditor issues a qualified opinion, however, the authors find no support that a change in auditor influences the opinion provided. In the research, the authors use an independent variable BIG6 to proxy for auditor quality and reputation.

This paper extends the market reaction to auditor switching studies of Johnson and Lys (1990) and Teoh and Wong (1993) by looking at a multi-tiered classification scheme based on the number of clients for an auditor.

THEORETICAL DEVELOPMENTS

Audit Quality

The users of financial information desire an independent audit as means of monitoring financial information to ensure that information is reliable. Information reliability incorporates the characteristics of precision and bias. Precision implies that stated measurement methods were properly applied, while bias indicates that the measurement results were correctly displayed (Kinney, 2000). The users of financial information require a quality audit to ensure that numbers are precise (within the confines of materiality) and free from bias.

The quality of an audit may be defined as the market-assessed joint probability that an auditor will discover a financial reporting impression or bias, and report the situation to the information users (DeAngelo, 1981). Although audit quality is not directly observable, users develop proxies that they believe are associated with audit quality (Wilson and Grimlund 1990; Palmrose 1991). One such proxy is audit firm size.

It has been stated in many studies that large audit firms provide a higher quality audit then smaller firms (DeAngelo 1981; Chow and Rice 1982; Schwartz and Menon 1985). According to DeAngelo (1981)
 .... the larger the auditor as measured by the number of clients,
 the less incentive the auditor has to behave opportunistically and
 the higher the perceived quality of the audit.


In addition, large audit firm investments in specialized resources such as training and technology yield economies of scale and scope for audit services (Johnson and Lys 1990). If investors believe that large audit firms provide a better quality audit then smaller audit firms, and that this quality results in improved reliability of financial information, changes to larger audit firms could lead to a positive share price reaction around the announcement of the auditor change.

Insurance Hypothesis

Prior research indicates that investors perceive auditors as providing a type of implicit insurance to users and investors (Hill et. al. 1993). The auditors are deemed to be a "deep pocket" because CPA firms often carry malpractice insurance or, in many cases, are the only solvent defendant in a lawsuit. Therefore, the auditor is considered a potential indemnifier to investors and creditors if a loss is experienced. Menon and Williams (1994) assert that the legal right to seek indemnification from an auditor for losses sustained is assigned a value by investors and is a component of the stock price of publicly traded companies. The insurance hypothesis indicates that investors will react to changes that may affect their ability to collect damages from the auditor. Changes to larger, more solvent audit firms could lead to a positive share price reaction for audit clients around the announcement of the auditor change.

The Timing of the Market Reaction

Many auditor switching studies have used the announcement of a change in auditor as the event date. Studies such as Smith (1988), Mangold (1988), Teoh (1989), and Johnson and Lys (1990) do not find a market reaction to the announcement of an auditor switch. Fried and Schiff (1981) find a negative cumulative abnormal return for a 21-week period following the filing of an 8-K report for an auditor change. Wells and Louder (1997) find evidence that the market views an auditor resignation as bad news and a resultant negative price reaction occurs.

It should be noted that at the announcement date of the change in auditor, the successor auditor has not yet performed any work for the client. Legally the successor auditor cannot be held responsible for the work performed or information provided by the predecessor auditor. Therefore, should the investors find fault in published financial information, it is still the predecessor auditor that is liable. This may lead to a perceived difference in the risk of relying on financial information associated with the successor auditor. In addition, when a new auditor is announced, investors may anticipate an audit of higher quality from the successor auditor. However, until information prepared or audited by the new firm is made available to the market, there is no "product" that the market can assess. This may lead to a perceived difference in the reliability of the financial information associated with the successor auditor. For these reasons, the market may react differently to earnings announcements when a new auditor is associated with the financial statements. This reaction may be in addition to any reaction that the market has to the announcement of the auditor change.

Hypothesis Development

The theory that large audit firms provide higher quality audits and that large audit firms provide greater insurance protection for investors and creditors does not indicate that size advantages are limited to the big-five audit firms. Indeed, it is only the perpetuation of prior methodology that has resulted in an audit quality proxy as a dichotomy between big-five firms and all other firms providing audit services. The following hypotheses will be tested to determine if there are cumulative abnormal returns at the earnings announcement date when clients change auditor class in a multi-tiered classification system.
H1: Clients that change from a non-big-five audit firm to big-five
audit firm experience positive cumulative abnormal returns in the
market place.

H2: Clients that change from a big-five audit firm to a smaller audit
firm experience negative cumulative abnormal returns in the market
place.


If the market place perceives a quality difference in a big-five auditor, the results of H1 should be statistically significant with a positive sign, while the results of H2 should be statistically significant with a negative sign.
H3: Clients that change from one big-five audit firms to another
non-big-five audit firm experience cumulative abnormal returns in the
market place.

H4: Clients that change from a non-big-five to another non-big-five
audit firm experience cumulative abnormal returns in the market place.


Hypotheses 3 and 4 are designed to indicate that the market does not react to auditor changes when auditor changes do not involve changes in auditor class. The results of the testing of these hypotheses should not be statistically different from zero.
H5: Clients that change to a larger audit firm experience positive
cumulative abnormal returns in the market place.

H6: Clients that change to a smaller audit firm experience negative
cumulative abnormal returns in the market place.


Hypotheses 5 and 6 are designed to test whether the market reacts to changes in audit firm size when the auditor is not a big-five auditor.

SAMPLE SELECTION

Six regressions are estimated with samples of firms obtained from the COMPUSTAT industrial tapes, which include firms listed on the New York Stock Exchange (NYSE), the American Stock Exchange (AMEX), and the National Association of Security Dealers Automated Quotations (NASDAQ). The sample is selected from files of the 1999 annual industrial tapes, and is limited to firms with earnings information in each year of the period 1989-1998. As a measure of unexpected earnings, we use consensus analysts' forecast, therefore, we require the sample firms to be followed by the Institutional Brokers Estimate System (IBES) similar to Baginski, Hassell and Waymire (1994) and Stunda (1996). We also require that firms have daily stock returns data available on tape from the Center for Research in Security Prices (CRSP) during the period under study.

The first sample is a control sample that contains industry-matched pairs of firms to control for differences in the information environment. Industry matching is accomplished by matching four-digit, three-digit, and two-digit SIC codes for each firm audited by one of the big-five firms versus a firm in the same industry audited by non-big-five firm. For the period under study, 1,485 firms were observed that met the sample criteria.

The second sample contains firms that have switched from a non-big-five firm to a big-five audit firms during the study period. The sample is selected from COMPUSTAT and must meet the data availability criteria. For the study period, 147 firms were observed which met the sample parameters.

The third sample contains firms that have switched from one of the big-five firms to a non-big-five audit firms during the study period. For the study period, 42 firms were observed which met the sample criteria.

The fourth sample contains firms that have switched from a big-five audit firm to another big-five audit firms during the study period. For the study period, 26 firms were observed which met the sample criteria.

The fifth sample contains firms that have switched from a non-big-five audit firm to another non-big-five audit firm during the study period. For the study period, 31 firms were observed which met the sample criteria.

The sixth sample stratifies firms that have switched auditors during the study period. Four groups of audit firms are created using the number of clients identified for each firm. This means of classification is appropriate, since this study utilizes many smaller audit firms. Information, such as total audit firm revenue, is not readily available for such clients. Information concerning the number of COMPUSTAT clients for an audit firm is objective and measurable for all firms.

Group 1 consists of the five largest firms (big-five). These firms average more than 2,000 clients as reported on COMPUSTAT for the years 1989-1998. Group 2 consists of audit firms with an average number of clients between 500 and 2,000 as reported on COMPUSTAT for the years 1989-1998. These firms proxy for the large national firms. Group 3 consists of audit firms with an average number of clients between 200 and 400 as reported on COMPUSTAT for the years 1989-1998. These firms proxy for the large regional audit firms. Group 4 consists of audit firms with less than 200 clients as reported on COMPUSTAT for the years 1989-1998 of small regional firms. These cut-offs are arbitrary in nature, but based on the data appear reasonable.
audit group # of audit firms # of sample firms

1 (Big 5) 5 535
2 (non-Big 5 national) 6 326
3 (large regional) 16 418
4 (small regional) 10 206
Total 37 1485


Dummy variables are utilized and an analysis is made of sample firms that switch as follows:

Change from group 1 auditors to group 2 auditors

Change from group 1 auditors to group 3 auditors

Change from group 1 auditors to group 4 auditors

Change from group 2 auditors to group 3 auditors

Change from group 2 auditors to group 4 auditors

Change from group 3 auditors to group 4 auditors

Change from group 2 auditors to group 1 auditors

Change from group 3 auditors to group 2 auditors

Change from group 3 auditors to group 1 auditors

Change from group 4 auditors to group 3 auditors

Change from group 4 auditors to group 2 auditors

Change from group 4 auditors to group 1 auditors

Methodology

The first regression assesses the relative information content of unexpected earnings in matched pair samples of firms in comparable industries audited by big-five and non-big-five audit firms. This regression is run as a control from which subsequent regressions will be compared. The following model is used to evaluate information content:

[CAR.sub.it] = a + [b.sub.1][UE.sub.it] + [b.sub.2][D.sub.it][UE.sub.it] + [b.sub.3][MB.sub.it] + [b.sub.4][D.sub.it][MB.sub.it] + [b.sub.5][LMV.sub.it] + [b.sub.6][D.sub.it][LMV.sub.it] + [b.sub.7][N.sub.it] + [b.sub.8][D.sub.it][N.sub.it] + [b.sub.9][B.sub.it] + [b.sub.10][D.sub.it][B.sub.it] + [e.sub.it] [1]

Where: [CAR.sub.it] = Cumulative abnormal return for firm i, time t

a = Intercept term

[UE.sub.it] = Unexpected earnings forecast for firm i, time t

[D.sub.it] = Dummy variable, 1 for NB5 client, 0 for B5 client

[MB.sub.it] = Market value to book value as a proxy for growth and persistence

[LMV.sub.it] = Natural log of market value as a proxy for firm size

[N.sub.it] = Number of analysts' forecasts included in IBES as a proxy for noise in the predisclosure environment

[B.sub.it] = Market value slope coefficient as a proxy for systematic risk

[e.sub.it] = Error term for firm i, time t

The coefficient "a" measures the intercept. The coefficient "[b.sub.1]" is the earnings response coefficient (ERC) for all firms in the sample (both big-five and non-big-five clients). The coefficient [b.sub.2] represents the incremental ERC. Therefore, b2 captures the difference in the information content for firms that are big-five clients versus those who are not. The remaining coefficients are contributions to the ERC for all firms in the sample. To investigate the effects of the information content of unexpected earnings, there must be some control for variables shown by prior studies to be determinants of the ERC. For this reason, variables represented by these coefficients are included in the study.

Unexpected earnings ([UE.sub.it]) is measured as the difference between the actual earnings and the security market participants' expectations for earnings proxied by consensus analysts' forecast as per IBES. The unexpected earnings are scaled by the firm's stock price 180 days prior to the forecast:

[UE.sub.it] = Actual Earnings - Expected Earnings/Price

For each disclosure sample, an abnormal return ([AR.sub.it]) is generated for event days -1, 0, +1, where day 0 is defined as the date of the earnings disclosure identified by the Dow Jones News Retrieval Service (DJNRS). The market model is utilized along with the CRSP equally-weighted market index and regression parameters are estimated between days -290 and -91. Abnormal returns are then summed to calculate a cross-sectional cumulative abnormal return ([CAR.sub.it]).

Regressions two through five address the switching of client firms among and between audit firm groupings. These switches are observed as follows:

1. NB5 to B5

2. B5 to NB5

3. B5 to B5

4. NB5 to NB5

The following model is used to evaluate information content among the switched groups:

[CAR.sub.it] = [a + [b.sub.1][UE.sub.it] + [b.sub.2][MB.sub.it] + [b.sub.3][LMV.sub.it] + [b.sub.4][N.sub.it] + [b.sub.5][B.sub.it] + [e.sub.it] [2]

Variables and model parameters used are the same as those utilized in equation 1, except for the elimination of the dummy variable. The above equation is run four times, substituting switched groups in each run.

For regression six, the following equation is used:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] [3]

Where: b1 = Variable for change from group 1 auditors to group 2 auditors

b2 = Variable for change from group 1 auditors to group 3 auditors

b3 = Variable for change from group 1 auditors to group 4 auditors

b4 = Variable for change from group 2 auditors to group 3 auditors

b5 = Variable for change from group 2 auditors to group 4 auditors

b6 = Variable for change from group 3 auditors to group 4 auditors

b7 = Variable for change from group 2 auditors to group 1 auditors

b8 = Variable for change from group 3 auditors to group 2 auditors

b9 = Variable for change from group 3 auditors to group 1 auditors

b10= Variable for change from group 4 auditors to group 3 auditors

b11= Variable for change from group 4 auditors to group 2 auditors

b12= Variable for change from group 4 auditors to group 1 auditors

b13= Market value to book value as a proxy for growth and persistence

b14= Natural log of market value as a proxy for firm size

b15= Number of analysts' forecasts included in IBES as a proxy for noise in the predisclosure environment

b16 = Market value slope coefficient as a proxy for systematic risk

All parameters are the same as used in the first two regression equations.

DISCUSSION OF RESULTS

Table 1 provides the results from the first regression of matched pair firms. As can be seen, none of the variables contained in the regression are significant in explaining the CAR. This is similar to results found by Teoh and Wong (1993).

Table 2 provides results from the first switch sample where client firms switched from non-big-five audit firms to big-five audit firms. As can be seen from the table, the unexpected earnings variable is positively significant in providing information content relative to CAR. The implication is that unexpected earnings contain information content for firms switching to big-five auditors, and this information is positively correlated. This result confirms hypothesis 1 and indicates that the marketplace adds value to a publicly traded stock when the company changes to a big-five auditor.

Table 3 provides results from the second switch sample where client firms switched from big-five audit firms to non-big-five audit firms. As can be seen from the table, the unexpected earnings variable is negatively significant in providing information content relative to CAR. The implication is that unexpected earnings contain information content for firms switching to non-big-five auditors, and this information is negatively correlated. These results support hypothesis 2 and indicate that the marketplace reduces value to the stock when a change from a big-five auditor is made.

Tables 4 and 5 provide results from the remaining two switch samples between big-five firms and non-big-five firms respectively. As expected, no significant results were noted in these switch samples. While the lack of significance cannot imply the acceptance of the alternative hypothesis that the marketplace adds no value to changes among the big-five or among the non-big-five auditors, it is comforting that the results were as expected.

Table 6 provides results of firms that have switched between auditor classes for the period under study. As can be seen, the unexpected earnings variable is positive and significant in providing information content relevant to CAR in switches from smaller firms to larger firms (i.e., variables b7 through b12). These results are consistent with hypothesis 5 and indicate that the market adds value to firms that switch to a larger firm, even if the larger firm is not a big-five audit firm. This result may be associated with an audit quality or an insurance factor associated with a larger audit firm.

The variables b1 through b6 (excluding b5 which did not include any samples) represent switches from larger audit firms to smaller audit firms. While the signs for these switches are negative, as expected, none of the coefficients are statistically significant at any of the conventional levels. It should be noted that many of these groups contained few, if any, audit switches. Where only a few firms exist, a contrary reaction to even one switch can greatly skew the data. Table 3 presented the results for switches from big-five auditors (group1) to non-big-five auditors (groups 2, 3, and 4 combined). Using that grouping, the changing from a large auditor to a smaller auditor was statistically significant. The limited number of firms in these groups is a limitation of this study and may account for the lack of statistical significance for variables b1 through b6.

CONCLUSION

It has been noted in many studies that the financial statement users value large audit firms because they perceive these firms as either providing a higher quality audit or greater insurance in the event of a financial loss. These studies have resulted in audit quality as a dichotomous variable where big-five auditors represent the quality firm. This study provides some evidence that the market does value a larger audit firm, even if that firm is not a big-five firm.

REFERENCES

Baginski, S., J. Hassell, and G. Waymire, 1994, Some Evidence on the News Content of Preliminary Earnings Estimates, The Accounting Review, January, pp. 265-271

Chow, C. W. and S. J. Rice, 1982, Qualified Audit Opinions and Auditor Switching, Accounting Review, April, pp. 326335

DeAngelo, Linda, 1981, Auditor Size and Audit Quality, Journal of Accounting and Economics, pp. 183-199

Francis, J. R. and E. R. Wilson, 1988, Auditor Changes: A joint Test of Theories Relating to Agency Costs and Auditor Differentiation, Accounting Review, October, pp. 663-683

Fried, D. and A. Schriff, 1981, CPA Switches and Associated Market Reactions, The Accounting Review, April, pp. 326341

Hill, J. W., M. Metzger, and J. Schatzberg, 1993, Auditing's Emerging Legal Peril Under the National Surety Doctrine: A Program for Research, Accounting Horizons, March, pp. 12-28

Johnson, W. B. and T. Lys, 1990, The Market for Audit Services: Evidence from Voluntary Auditor Changes, Journal of Accounting and Economics, January, pp. 281-308

Kinney, Willian R., 2000, Information Quality Assurance and Internal Control for Management Decision Making, Irwin McGraw-Hill:Boston

Krishnan, J., 1994, Auditor Switching and Conservatism, Accounting Review, January, pp. 200-215

Krishnan, J., J. Krishnan, and R.G. Stephens, 1996, The Simultaneous Relation Between Auditor Switching and Audit Opinion: An Empirical Analysis, Accounting and Business research, 3, pp. 224-236

Mangold, N. R., 1988, Changing Auditors and the Effect on Earnings, Auditors' Opinions, and Stock Prices, UMI Research Press

Menon, K. and D. Williams, 1994, The Insurance Hypothesis and the Market Prices, Accounting Review, April, pp. 327342

Palmrose, Z. , 1991, An Analysis of Auditor Litigation Disclosures, Auditing: A Journal of Theory and Practice, Supplement, pp. 54-76

Shwartz, K. B. and K. Menon, 1985, Auditor Switches by Failing Firms, Accounting Review, April, pp. 248-261

Smith, D.B., 1988, An Investigation of Securities and Exchange Commission Regulation of Auditor Change Disclosures: The Case of Accounting Series Release No. 165, Journal of Accounting Research, Spring, pp/ 134-145

Stunda, R., 1996, The Credibility of Management Forecasts During Corporate Mergers and Acquisitions, The American Academy of Accounting and Financial Studies Journal, December, 352-358

Teoh, S.H., 1989, Auditor Independence, Dismissal Threats, and the Market Reaction to Auditor Switches, Journal of Accounting Research, 30, pp. 1-25

Teoh, S. H. and T. J. Wong, 1993, Perceived Auditor Quality and the Earnings Response Coefficient, Accounting Review, April, pp. 346-366

Wells, D.W. and M.L. Loudder, 1997, The Market Effects of Auditor Resignations, Auditing: A Journal of Theory and Practice, Spring, pp: 138-144

Wilson, T. and R. Grimlund, 1990, An Examination of the Importance of an Auditor's Reputation, Auditing: A Journal of Theory and Practice, Spring, pp. 43-59

Ronald A. Stunda, Birmingham-Southern College

David H. Sinason, Northern Illinois University
Table 1: Summary of Pair-Matched Samples 1989-1998
n = 1,485 client firms

[CAR.sub.it] = a+[b.sub.1][UE.sub.it]+[b.sub.2][D.sub.it][UE.sub.it]+
[b.sub.3][MB.sub.it]+[b.sub.4][D.sub.it][MB.sub.it]+[b.sub.5]
[LMV.sub.it]+[b.sub.6][D.sub.it][LMV.sub.it]+[b.sub.7][N.sub.it]+
[b.sub.8][D.sub.it][N.sub.it] +[b.sub.9][B.sub.it]+[b.sub.10]
[D.sub.it][B.sub.it]+[e.sub.it]

 Mean Median

Variable B5 NB5 B5 NB5

UE -0.0286 -0.0180 -0.0009 -0.0011
MB 2.7190 2.4881 1.6021 1.7872
LMV 4.1911 4.2832 4.2901 4.0098
N 5.2218 5.0190 4.0000 3.0000
B 1.3856 1.3249 1.2019 1.2001

 T-
Variable Coeff statistic p-value

UE -0.1103 -0.6475 0.4362
MB 0.0829 0.2183 0.5768
LMV -0.0541 -0.3389 0.4976
N 0.0423 2.9090 0.1586
B -0.0218 -1.1391 0.4027

Table 2: Summary of Client Firms Switching From NB5 to B5 Audit Firms
n = 147

[CAR.sub.it] = a+[b.sub.1][UE.sub.it]+[b.sub.2][MB.sub.it]+[b.sub.3]
[LMV.sub.it]+[b.sub.4][N.sub.it]+[b.sub.5][B.sub.it]+[e.sub.it]

Variable Mean Median Coeff. T-statistic p-value

UE 0.0329 0.0190 0.0921 2.3284 0.0219
MB 2.2802 1.2081 0.0538 0.4492 0.6304
LMV 4.5890 4.2891 -0.0322 -0.1938 0.7984
N 4.5890 4.0000 0.0725 1.5947 0.2609
B 1.2819 1.1947 0.0198 1.1149 0.2531

Table 3: Summary of Client Firms Switching From B5 to NB5 Audit Firms
n = 42

[CAR.sub.it] = a+[b.sub.1][UE.sub.it]+[b.sub.2][MB.sub.it]+[b.sub.3]
[LMV.sub.it]+[b.sub.4][N.sub.it]+[b.sub.5][B.sub.it]+[e.sub.it]

Variable Mean Median Coeff. T-statistic p-value

UE -0.0217 -0.0209 -0.1053 2.7562 0.0118
MB 2.1546 1.8253 0.0486 0.5417 0.6728
LMV 4.3862 4.1170 -0.0251 -0.1764 0.8194
N 3.7652 3.0000 0.0665 1.7621 0.4153
B 1.1597 1.2089 0.0249 1.0018 0.2764

Table 4: Summary of Client Firms Switching From
B5 to B5 Audit Firms n = 26

[CAR.sub.it] = a+[b.sub.1][UE.sub.it]+[b.sub.2][MB.sub.it]+[b.sub.3]
[LMV.sub.it]+[b.sub.4][N.sub.it]+[b.sub.5][B.sub.it]+[e.sub.it]

Variable Mean Median Coeff. T-statistic p-value

UE 0.0568 0.0419 0.0291 0.8915 0.4956
MB 1.9976 1.9541 0.0447 0.4876 0.6219
LMV 4.1876 4.0091 -0.0210 -0.2018 0.7847
N 4.8431 4.0000 0.0655 1.6291 0.2987
B 1.3196 1.2922 0.0200 1.1551 0.3281

Table 5: Summary of Client Firms Switching From NB5 to NB5 Audit Firms
n = 31

[CAR.sub.it] = a+b1[UE.sub.it]+b2[MB.sub.it]+b3
[LMV.sub.it]+b4[N.sub.it]+b5[B.sub.it]+[e.sub.it]

Variable Mean Median Coeff. T-statistic p-value

UE 0.1521 0.1019 0.0313 0.4987 0.6636
MB 2.0198 1.9827 0.0521 0.5121 0.7147
LMV 4.2819 4.0989 -0.0198 -0.4942 0.7767
N 3.8219 3.0000 0.0715 1.8431 0.2262
B 1.4003 1.3821 0.0249 1.0089 0.3724

Table 6: Summary of Client Firms Switching Auditor Types
n = 220

[CAR.sub.it]=a+[b.sub.1][UE.sub.it]+[b.sub.2][UE.sub.it]+
[b.sub.3][UE.sub.it]+[b.sub.4][UE.sub.it]+[b.sub.5]
[UE.sub.it]+[b.sub.6][UE.sub.it]+[b.sub.7][UE.sub.it]+
[b.sub.8][UE.sub.it]+[b.sub.9][UE.sub.it] +[b.sub.10]
[UE.sub.it]+[b.sub.11][UE.sub.it]+[b.sub.12][UE.sub.it]+
[b.sub.13][MB.sub.it]+[b.sub.14][LMV.sub.it]+[b.sub.15]
[N.sub.it]+[b.sub.16][B.sub.it]+[e.sub.it]

 Number
Variable Switching Mean Median

b1 28 -0.1021 -0.1035
b2 11 -0.1085 -0.1062
b3 3 -0.1407 -0.1318
b4 8 -0.1182 -0.1168
b5 0 N/A N/A
b6 3 -0.1201 -0.1159
b7 87 0.0739 0.0801
b8 10 0.0901 0.0889
b9 51 0.0995 0.1010
b10 6 0.0840 0.0809
b11 9 0.1014 0.1041
b12 4 0.0785 0.0798
b13 2.1608 2.1554
b14 4.1005 4.1001
b15 3.5821 3.5618
b16 1.5109 1.5007

 T-
Variable Coeff. Statistic p-value

b1 -0.0943 1.2846 0.3015
b2 -0.0827 1.3102 0.2795
b3 -0.0915 1.4519 0.2102
b4 -0.0622 1.3422 0.2820
b5 N/A N/A N/A
b6 -0.0449 1.1950 0.3102
b7 0.0591 2.3515 0.0211
b8 0.0774 2.2412 0.0372
b9 0.0820 2.3802 0.0146
b10 0.0338 2.1921 0.0486
b11 0.0516 2.2056 0.0437
b12 0.0249 2.3601 0.0203
b13 0.0418 0.6072 0.7519
b14 -0.0302 0.2019 0.8209
b15 0.0467 1.8721 0.2398
b16 0.0128 1.1526 0.4001
COPYRIGHT 2002 The DreamCatchers Group, LLC
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:MANUSCRIPTS
Author:Stunda, Ronald A.; Sinason, David H.
Publication:Academy of Accounting and Financial Studies Journal
Geographic Code:1USA
Date:Sep 1, 2002
Words:5628
Previous Article:Accounting for acquisitions and firm value.
Next Article:The dual role of audit evidence.
Topics:

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters