Printer Friendly

How Effective Is Joint Public and Private Monitoring? The Case of the AICPA Auditor Change Notification Letter.

The AICPA SEC Practice Section (SECPS) requires a member firm to notify its former client and the Chief Accountant of the SEC in writing within five business days of the date the auditor resigns or is dismissed. During 1993-1995, the SECPS conducted an educational effort to make members aware of this responsibility. We evaluate the effectiveness of the SECPS notification letter and educational effort. Specifically, we answer these questions:

1. Does faster auditor filing of the notification letter influence clients to file their Form 8-Ks more quickly?

2. Have either auditors or clients filed more quickly since the SECPS conducted its educational effort?

To address these questions, we investigated 307 auditor changes that occurred before (1991-1992) and 399 changes after (1995-1996) the SECPS education initiative. We find that both client and auditor filing performance improved during the period, and that the improved client performance is likely due in part to improved auditor performance.

The auditor's notification letter represents an example of a recent trend requiring auditors to help the SEC monitor client compliance with required filings. The results suggest that, given adequate education, auditors can make an effective contribution toward client compliance.

The Securities Exchange Act of 1934 provides the SEC with the authority to develop and enforce accounting standards, and requires periodic filings by companies under its jurisdiction. The SEC traditionally has looked to the private sector for development of accounting and auditing standards, input regarding the nature and timing of required filings, and self-regulation of professional services. The SEC recently has begun also to depend on the private sector for help in monitoring registrant compliance with required filings. Another example in which accounting professionals are relied upon for monitoring assistance is the corporate fraud disclosure requirement of the Securities Reform Act of 1995.(1) Because of the recency of this Act, its effectiveness cannot yet be evaluated.

A joint monitoring arrangement for which there are sufficient observations for empirical inquiry is the AICPA SEC Practice Section (SECPS)(2) membership rule that requires a member firm to notify its former client, and the Chief Accountant of the SEC, in writing when the firm determines that the client-auditor relationship has ended. The auditor's notification letter provides the SEC with timely information about the auditor change that is independent of the SEC's receipt of the registrant's (client's) Form 8-K filing announcing the change.(3) Enacted in response to SEC concerns over registrants' late filings, the notification requirement reflects the accounting profession's belief in the importance and effectiveness of monitoring in the private sector.

Discussions with representatives of the SEC indicate that attempts are made to match the SECPS auditor's letter with the Form 8-K filing, and to follow-up situations where an SECPS letter is filed and no Form 8-K can be found. Client knowledge of the auditor's notification responsibility is intended to provide an incentive for prompt Form 8-K filing and thereby help the SEC ensure timely public disclosure of auditor changes. However, the Commission has not formally evaluated the letter's effect. An incentive for auditors to follow the SECPS rule exists because compliance is evaluated as part of the SECPS's mandatory peer-review program, which is the cornerstone of the profession's self-regulatory effort.

This study evaluates SECPS member compliance with the auditor change notification rule, and investigates how this compliance is related to registrant compliance with disclosure-timeliness criteria found in SEC auditor change regulations. In so doing, the paper provides evidence on the effectiveness of a self-regulatory rule requiring accounting professionals to assist the SEC in monitoring registrant compliance. The effectiveness of self-regulation by professional associations is currently being evaluated by the Panel on Audit Effectiveness (1999) appointed by the Public Oversight Board. We investigate the following research questions:

RQ1: Are auditor compliance (timeliness) and registrant compliance (timeliness) positively associated after controlling for other factors?(4)

RQ2: Have compliance rates (or timeliness) improved for either auditors or registrants during the time the notification rule has been in effect?

The paper is organized as follows. In section two, we review the history of SEC auditor change regulations, and describe how the SECPS letter supports the disclosure process. We also briefly review prior literature that studied auditor changes. Section three states the research questions. Section four describes the empirical models, samples, variables, and tests. Section five summarizes the results. Section six presents conclusions, recommendations, and suggestions for future research.

BACKGROUND

SEC and SECPS Requirements

In 1971 the SEC issued the initial rules requiring Form 8-K disclosure of auditor changes (SEC 1971). The stated purpose of these rules was to discourage "opinion shopping" and thereby to strengthen auditor independence. Regulatory attention refocused on the timeliness of auditor change disclosures when a registrant accused of filing fraudulent financial information (ZZZZ Best) took 15 calendar days (the full period then allowed) to file a Form 8-K disclosing that the relationship with its auditor had ended. The auditor then took the full 30 calendar days allowed to comment on the former client's explanation of the auditor change.

Concerned that such delays could harm investor confidence and well being, the SEC (1989) issued Financial Reporting Release (FRR) No. 34 which shortened the time for Form 8-K reporting of an auditor change from 15 calendar days to 5 business days. FRR No. 34 also shortened the number of days the former auditor had to comment from 30 calendar to 10 business days after the registrant filed the Form 8-K announcing the auditor change.

Reasoning that it would be easier to identify registrants who file late if auditors independently informed the SEC of the end of client-auditor relationships, the SEC encouraged the SECPS, whose members audit more than 90 percent of publicly traded companies, to enact a notification rule. The SECPS responded with a requirement that a member of the SECPS notify the Chief Accountant of the SEC within five business days after it determines that its audit relationship with an SEC registrant has ended.

Members of the SECPS committee who drafted the requirement described it as a quality control practice needed to help preserve the public's perception of professional integrity and objectivity (Milan 1993). This SECPS regulation is a significant departure from past practice forbidding direct communication between the registrant's auditor and the SEC. Previously the auditor's only communication with the SEC was a letter commenting on its former client's Form 8--K explanation of the circumstances of the auditor change. The current SECPS regulation requires direct timely communication between the auditor and the SEC when the auditor-client relationship ceases. The auditor sends the letter that this paper studies directly to the SEC to give the SEC independent notification that an auditor change has occurred. The auditor is required to send the registrant a copy of the letter at the same time the auditor sends it to the SEC. What makes the SECPS regulation unique is that it was made and is enforced by a private organization (the AICPA) to assist a public organization (the SEC) to perform its charge of ensuring full, timely disclosure of relevant registrant information.

The goal of this paper is to investigate the effectiveness of joint public and private monitoring using the case of the AICPA SECPS auditor change notification letter. To reach this goal we first investigate whether auditor compliance (timeliness) and registrant compliance (timeliness) are positively associated after controlling for other factors.

The uniqueness of the SECPS regulation also provides the opportunity to investigate empirically whether educational and enforcement efforts by a private organization can assist a public organization to be more effective. This investigation provides evidence related to the debate concerning when self (AICPA) regulation works and when there is a need for greater outside (SEC) regulation. SECPS member-auditors were initially either unaware that filing the SECPS letter was mandatory, or they were slow in filing. This period, prior to 1993, is the period from which the samples used in previous research investigating auditor change filing lags were drawn (e.g., Schwartz and Soo 1996). Beginning in 1993 and continuing into 1995, the SECPS mounted a significant educational effort to make member firms aware of the need to comply with the SECPS auditor change regulation. In this paper we use a sample drawn from both the pre-SECPS educational effort period (1991-1992) and from the post-SECPS educational effort period (1995-1996). This sample allows us to investigate how effective a private professional organization (the AICPA) can be in prompting its members to monitor and disclose information that assists a public organization (the SEC) to fulfill its charge. We do this by investigating whether compliance rates (or timeliness) improved for auditors and, more importantly, registrants during the time the notification rule has been in effect.

Prior Auditor Change Research

Prior research has focused primarily on identifying variables associated with auditor changes. Variables examined include those representing auditor-client disagreements (Dhaliwal et al. 1993; Smith and Nichols 1982; Fried and Schiff 1981); motives for auditor resignations (DeFond et al. 1997); and client financial distress (Schwartz and Soo 1995; Schwartz and Menon 1985; Whittred and Zimmer 1984; Lawrence 1983).

Schwartz and Soo (1996) use data from 1988-1993 to investigate whether the 1989 enactment of FRR No. 34, which, as noted previously, reduced the Form 8-K filing deadline to five business days, affected registrant compliance rates and timeliness. Although they mention the 1989 enactment of the SECPS notification rule, they do not test for association between auditor and client compliance rates or filing delays. This is an important distinction between their work and ours.

Schwartz and Soo's (1996) results indicate that compliance rates before and after the enactment of FRR No. 34 are not statistically different, but tend to increase after FRR No. 34. They find that post-FRR No. 34 filing lags are approximately four days shorter, and that compliance rates are higher and filing delays are shorter for large companies and those with Big 6 auditors. Noncompliance is higher among financially distressed companies. Schwartz and Soo (1996, 570) conclude that their "study suggests a need for additional research on the incentives of registrants to manage the disclosure of information and the effectiveness of existing regulatory mechanisms." Carter and Soo (1999) find that only firms that file Forms 8-K on time experience a significant market reaction at the filing date. Carter and Soo (1999, 120) interpret their finding as indicating that the timely filing is the "single most consistent determinant of the 8-K's informativeness, and as providing support for the SEC's efforts to shorten Form 8-K filing periods." Our paper provides evidence on the success of the SECPS notification letter in improving the timeliness of auditor change disclosures.

RESEARCH QUESTIONS

RQI: Are auditor compliance (timeliness) and registrant compliance (timeliness) positively associated after controlling for other factors?

Schwartz and Soo (1996) provide information on the compliance (timeliness) of registrants' Form 8-K filings. From the standpoint of the SECPS and others within the AICPA, however, it is the auditors' behavior that is of initial interest and that they are best able to influence. This behavior is not routinely tracked by the SECPS, but must be understood if the SECPS is to evaluate the efficacy of the current rule. Previous research has shown that registrant compliance (timeliness) is associated with the size of the registrant and of the auditor, and with events surrounding the end of the auditor-client relationship. If the auditor's notification letter is performing the role envisioned by the SEC, the addition of an auditor compliance (timeliness) variable should increase the explanatory power of models of registrant compliance (timeliness). Registrant compliance (timeliness) should be positively related to auditor compliance (timeliness).

RQ2: Have compliance rates (or timeliness) improved for either auditors or registrants during the time the notification rule has been in effect?

Observing that auditor and registrant compliance are positively associated would suggest that the SECPS notification letter is having some effect on registrant compliance and timeliness of filing. The inference of causality should be strengthened if changes in auditor compliance over time are accompanied by corresponding changes in client compliance.

MODELS AND RESEARCH DESIGN

Empirical Models

We use three models to investigate our research questions. The first model, Model (1), tests whether the likelihood of registrant compliance is positively associated with auditor compliance. The dependent variable for Model (1) is the Form 8-K compliance variable (C8-K). This variable equals 1 if the client filed the Form 8-K within five business days, and equals 0 when the filing lag exceeded five business days. One independent variable of interest in Model (1) is the SECPS letter compliance variable (CSECPS). The variable is coded 1 if the auditor filed the SECPS letter with the SEC within five business days after the change and is coded 0 when the filing lag exceeded five business days or the SECPS letter was not filed (nonfilers). A second variable of interest is YEAR. This is coded 1 if the sample observation is drawn from the pre-SECPS education period and 0 if the sample observation is from the post-SECPS education period. YEAR captures any shift in the model's intercept over time.

Model (1):

C8-K = f(CSECPS, YEAR, control variables).

If CSCEPS is omitted from Model (1), then YEAR should have a significant, negative coefficient if client compliance increased from 1993 to 1996 (RQ2). When CSECPS is added to Model (1), the coefficient of YEAR should decrease in absolute magnitude if the change in client compliance is explained by increased auditor compliance (RQ2). The coefficient on CSECPS, when entered in Model (1), should be positive if auditor compliance encourages client compliance (research question No. 1). Research Question 2 also is addressed using descriptive statistics to compare auditor and client filing compliance, C8-K and CSECPS, for the pre- and post-education effort periods.

Our second model, Model (2), examines whether the Form 8-K filing lag is shorter if the SECPS letter is filed on time. The dependent variable for Model (2) is the number of business days required for registrants to file the Form 8-K, DAYS. DAYS is the number of business days from the auditor resignation or termination date (found in the text of Item 4 of Form 8-K) to the date the SEC receives the Form 8-K. The date-stamp on the face of the Form 8-K provides this date. Similar to Model (1), the independent variables of interest in Model (2) are the SECPS letter compliance variable (CSECPS) and the dichotomous temporal variable, YEAR.

Model (2):

DAYS

= f(CSECPS, YEAR, control variables).

If CSECPS is omitted from Model (2), then YEAR should have a significant positive coefficient if clients' filing lags were longer in the earlier period (RQ2). When CSECPS is added to Model (2), the coefficient of YEAR should decrease if the change in client compliance is explained by increased auditor compliance (RQ2). The coefficient on CSECPS, when entered in Model (2), should be negative if auditor compliance encourages faster client filing (RQ1).

Model (3) examines the relation between the Form 8-K filing lag in days and the SECPS letter-filing lag in days, SDAYS. SDAYS provides a continuous independent variable for examining the relation between the SECPS letter-filing lag and the Form 8-K filing lag, after controlling for other factors. Results from this model indicate whether the Form 8-K filing lag in days is shorter when the SECPS filing lag in days is shorter. We exclude auditor changes where the auditor did not file the SECPS letter from the estimation of Model (3) because it is impossible to calculate the SECPS letter-filing lag for nonfiling auditors.

Model (3):

DAYS = f(SDAYS, YEAR, control variables).

As with Model (2), the coefficient on YEAR should be positive and significant when the auditor behavior variable, SDAYS, is not in the model. The coefficient on YEAR should decrease when SDAYS is entered into the model, if changes in auditor filing behavior over time explain changes in client filing behavior. A positive coefficient on SDAYS would suggest that clients' filing lags are due in part to auditors' filing lags (RQ1). Again, RQ2 is addressed using descriptive statistics to compare auditor and client filing delay, DAYS and SDAYS, for the pre- and post-education effort periods.

Probit is used to estimate Model (1), which has a dichotomous dependent variable. Ordinary least squares is used to estimate the models with continuous dependent variables, Models (2) and (3).

Control Variables

Our models include other (control) variables that potentially affect registrant compliance with SEC Form 8-K filing rules, irrespective of whether the auditor files the SECPS letter on time. Schwartz and Soo (1996) found that large registrants tend to comply with the Form 8-K rules more often than do small registrants. Consistent with Schwartz and Soo (1996), we define size as the log of total assets (LASSET). Size is expected to be positively associated with compliance (negatively associated with filing lag) because larger companies probably provide better continuing education for their employees, and have employees with greater financial and regulatory expertise. Furthermore, to the extent that late filing draws the attention of the SEC, larger companies may have more to lose (e.g., eligibility to use short-form registration statement forms). Consistent with previous research (Skinner 1994, 1997) related to disclosure of "bad news," we would expect a positive relationship between auditor-client disagreements (DISAGREE) and Form 8-K filing compliance. Skinner (1997) finds modest evidence that disclosing adverse earnings news early via, for example, an earnings forecast, lowers expected legal costs. A final control variable expected to be positively associated with registrant compliance is having had a Big 6 auditor (BSIX). Big 6 auditors are expected to be more knowledgeable about SEC regulations and to be more aggressive in educating their clients about these regulations, which include filing a Form 8-K within five business days of the auditor change.

Several control variables are expected to be negatively associated with compliance (positively associated with clients' filing lags). First, if the registrant received a going-concern qualification or a modifying paragraph in its auditor's report within two years of the auditor change, we expect this to be negatively related to SEC Form 8-K filing compliance (MOD). Schwartz and Soo (1996) previously used this variable as a proxy for firm financial difficulty. They interpreted their results as evidence that financially stressed companies are slower at filing documents with the SEC. Second, we expect a negative relationship between SEC Form 8-K filing compliance and auditor resignation (RESIGN). Previous research (DeFond et al. 1997) indicates auditor resignations are related to declining company cash flows and financial difficulty. Consistent with Schwartz and Soo (1996), we expect that companies whose auditors resign are slower with their SEC filings.

Table 1 summarizes the dichotomous and continuous variables described above. The first variable in each group is the dependent variable. The information for several of the control variables is collected from disclosures in the Form 8-K. These included DISAGREE, MOD, and RESIGN. We obtained total asset amounts (LASSET) from either Compact Disclosure or LEXIS/NEXIS. We took natural logarithms of total assets so the distribution would be less skewed.
TABLE 1

Description of Variables

Dichotomous Variables

    C8-K 1 = registrant notified SEC of auditor change within five
             business days,
         0 = registrant did not notify SEC of auditor change within five
             business days;

  CSECPS 1 = predecessor auditor notified SEC within five business days,
         0 = predecessor auditor did not notify SEC within five business
             days;
    YEAR 1 = auditor change occurred during 1991-92,
         0 = auditor change occurred during 1995-96;

    BSIX 1 = predecessor auditor was a Big 6 firm,
         0 = predecessor auditor was a non-Big 6 firm;

DISAGREE 1 = auditor-client disagreement(s),
         0 = no auditor-client disagreement(s);

  RESIGN 1 = predecessor auditor resigned or declined to stand for
             reelection,
         0 = predecessor auditor was dismissed; and

     MOD 1 = predecessor auditor gave the registrant a going-concern
             opinion or included a modifying paragraph in its audit
             report within two years of the change,
         0 = predecessor auditor gave a "clean" audit opinion.

Continuous Variables

    DAYS = number of business days between the date that the
           auditor-client relationship ceased and the Form 8-K filing
           date;

   SDAYS = number of business days between the date that the
           auditor-client relationship ceased and the SECPS letter is
           filed;

  LASSET = log of total assets; and

   ASSET = total assets.


Sample Selection

Our initial sample was selected through a citation search of the LEXIS/NEXIS Company Form 8-K library for two time periods: (1) July 1, 1991 to June 30, 1992 and (2) July 1, 1995 to June 30, 1996. We focused on these two periods for several reasons. First, by concentrating effort on two one-year periods, we were able to gather a comprehensive sample of all auditor changes occurring during each year, rather than just the changes for which data are most readily available. We were able to enlist the help of the SECPS committee (all Big 6 auditors at the national office level), the SEC, the AICPA, and a number of smaller auditing firms in the collection of our sample, and in other aspects of the project. Second, prior to 1994 the SECPS Committee believed firms failed to comply with the SECPS auditor change regulation because they were unaware of it. The SECPS educational program carried out from 1993 to early 1995 potentially changed auditor behavior. One objective of our study is to investigate the effectiveness of joint public and private monitoring. Using samples that span the SECPS's educational effort enables us to examine whether that effort appears to have improved the effectiveness of joint public and private monitoring. Third, the two years were chosen so that the initial year (1991-92) was long enough after the initial enactment of the SECPS rule to ensure that auditors should have had a reasonable chance to be familiar with the rule. The second year (1995-96) was long enough after the SECPS Committee began its educational efforts to expect to find greater compliance in timely reporting if these efforts were effective. Finally, we used the July to June year because this corresponds to the year for Compustat data.

Our search yielded 1,603 Form 8-K Item 4 citations (auditor changes). We gathered as many of the Forms 8-K as we could from EDGAR and then ordered the remainder from the SEC. The SECPS letters were obtained from the SEC under the Freedom of Information Act. Each letter is numbered and date stamped when the SEC receives it. We matched each SECPS letter to the appropriate Form 8-K as it is listed in the cite search.

Table 2 summarizes sample attrition. Our final sample contains 706 auditor changes for which we have evidence (Form 8-K or both the Form 8-K and SECPS letter) that (1) an auditor change took place within our testing period and was documented by a Form 8-K filing, (2) the Form 8-K contains a readable SEC stamp date, and (3) the auditor's letter (if received) contained a readable SEC received stamp date.
TABLE 2
Sample Selection and Attrition for Form 8-K and SECPS
Letter Compliance Test

                                             1995-96  1991-92  Total
                                             Sample   Sample   Sample

Original lists                                 720      883    1,603

Eliminations
 Auditor changes occurred before July 1         40       41
 Registrant was corporate subsidiary,
  or not a corporation                          14
 SECPS letter was dated earlier than
  the auditor change date given in the
  Form 8-K                                      26
 No evidence of auditor change                  39       75
 Change was due to a merger of auditors          6
 Form 8-K announcing auditor dismissal or
  resignation was not available                 42      271
 Predecessor auditor was not an AICPA
  member                                        45       12
 Subtotal of eliminated auditor changes        212      399
 Forms 8-K and/or SECPS letter did not have
  readable date stamps                         109      177
 Sample for compliance tests                   399      307     706


RESULTS

Table 3 provides means of study variables for each of the two subsamples, and tests for differences in means between subsamples. Means for the dichotomous variables are multiplied by 100 and can be interpreted as the percentage of observations coded 1. Client compliance with the Form 8-K filing deadline increased from 61 percent in 1991-92 to 76 percent in 1995-96. Auditor compliance increased more dramatically from 26 percent to 62 percent. The clients' average Form 8-K filing lag decreased from about nine days to six days, while the auditors' mean letter-filing lag decreased from about nine business days to five business days. The data indicate that compliance and timeliness improved both for clients and auditors over the period examined. The OLS and Probit results reported below investigate whether the improvement in client filing behavior is likely related to the improvement in auditor filing behavior.
TABLE 3
Comparison of Variables for Registrants Filing Form 8-Ks in the 1991-92
Period vs. Registrants Filing Form 8-Ks in the 1995-96 Period

                              1995-96 Sample  1991-92 Sample
                                (n = 399)       (n = 307)      t-value

Variables of Interest

Dichotomous Variables (in %)
 C8-K                               76              61        4.2(***)
 CSECPS                             62              26       10.3(***)

Continuous Variable
 DAYS                                6.3             9.3     -3.6(***)
 SDAYS                               5.4             8.9     -3.7(***)

Control Variables

Dichotomous Variables (in %)
 BSIX                               61              63        0.7
 RESIGN                             14              16        0.7
 DISAGREE                            9              10        0.6
 MOD                                21              28        2.0(**)

Continuous Variable
 ASSET                             657.2           195.6      2.2(**)
 LASSET                              3.1             2.4      3.5(***)

(**), (***) Significant at the 0.05 and 0.01 level, respectively,
two-tailed.


The circumstances of the auditor changes remained fairly constant over the time periods examined. The percentage of clients audited by the Big 6, and the percentage of auditor changes involving resignations and disagreements, did not change significantly over time. The percentage of changes involving prior qualified opinions decreased over time, while the size of clients increased. Descriptive statistics are not provided for YEAR because it is coded 1 for all auditor changes in the earlier period, and is coded 0 for all changes in the later period.

Table 4 presents results for Probit Model (1). The first column provides results with the auditor compliance variable excluded from the model. As expected, the coefficient on YEAR is negative, indicating lower client filing compliance in the earlier period. This is consistent with the univariate statistics in Table 3, but the Probit results indicate that the change in client compliance is not likely due to changes in those circumstances captured by the control variables. The coefficients of variables BSIX and LASSET are positive as expected, indicating that clients of the Big 6, and larger companies, are more likely to comply. Other control variables are not significant.
TABLE 4
Summary of Probit Results
Compliance with 8-K Filing Rules (C8-K) is Dependent Variable

Independent             Model (1)(a)       Model (1)(b)
Variable                  (n = 706)          (n = 706)

Intercept                 0.25(**)           0.04
YEAR                     -0.37(***)         -0.20(**)
BSIX                      0.31(***)          0.22(**)
DISAGREE                  0.08              -0.01
RESIGN                   -0.10              -0.18
MOD                      -0.07              -0.07
LASSET                    0.11(***)          0.11(***)
CSECPS                                       0.50(***)
Log Likelihood         -401.5             -392.2
 (prob.)                 (0.000)            (0.000)

(*), (**), (***) Significant at the 0.10, 0.05, and 0.01 level,
respectively, one-tailed.
(a) Model does not contain CSECPS.
(b) Model contains CSECPS.

Variable Definitions:

    C8-K 1 = the client filed the Form 8-K with the SEC within the
             required five business-day period, 0 otherwise;
    YEAR 1 = auditor change occurred during 1991-92, 0 if auditor change
             occurred during 1995-96;
    BSIX 1 = predecessor auditor was a Big 6 firm, 0 otherwise;
DISAGREE 1 = auditor-client disagreement(s), 0 otherwise;
  RESIGN 1 = the predecessor auditor resigned or declined to stand for
             reelection, 0 otherwise;
     MOD 1 = the predecessor auditor gave the registrant a going-concern
             opinion or included a modifying paragraph related to
             uncertainty, 0 otherwise;
  LASSET 1 = log of total assets; and
  CSECPS 1 = the predecessor auditor filed the SECPS letter with the SEC
             within the required five business-day period, 0 otherwise.


When the auditor compliance variable CSECPS is entered into Model (1), as reported in the second results column, it has a positive, significant coefficient. Auditor compliance with SECPS filing requirements is positively associated with client Form 8-K filing compliance. Furthermore, when the SECPS variable is added to Model (1), both the magnitude and significance of YEAR approach 0 more closely. These results suggest that the improvement in clients' Form 8-K filing compliance over time is partially, but not fully, explained by improvement in auditors' SECPS letter filing compliance.

Table 5 presents OLS regression results for Models (2) and (3). The results can be interpreted similarly to those in Table 4, but with expected coefficient signs reversed since client-filing lag, DAYS, is inversely related to client compliance. The first column provides results with the auditor compliance variable excluded from Model (2). As expected, the coefficient on YEAR is positive, indicating longer client filing lags in the earlier period. The coefficients of variables BSIX and LASSET are negative as expected, indicating that clients of the Big 6, and larger companies, file more quickly on average. Clients that received prior qualified opinions take longer to file. Other control variables are not significant.
TABLE 5
Ordinary Least Squares Regression Results
Form 8-K Filing Lag (DAYS) is Dependent Variable

Independent    Model (2)(a)  Model (2)(b)  Model (3)(a)  Model (3)(c)
Variable        (n = 706)     (n = 706)     (n = 509)     (n = 509)

Intercept       9.39(***)     10.5(***)     7.90(***)     6.28(***)
YEAR            2.42(***)      1.46(**)     1.57(***)     0.80
BSIX           -2.94(***)     -2.46(***)   -2.09(***)    -1.67(**)
DISAGREE       -0.97          -0.48        -2.21(**)     -1.82(**)
RESIGN          1.34           1.73(*)      1.82(**)      2.68(***)
MOD             1.74(**)       1.75(**)     1.72(**)      1.66(**)
LASSET         -0.57(***)     -0.55(***)   -0.28(**)     -0.26(**)
CSECPS                        -2.62(***)
SDAYS                                                     0.21(***)
Adj. R-Square  0.080          0.092        0.066         0.130
F              11.21          11.25         6.97         11.89

(*), (**), (***) Significant at the 0.10, 0.05, and 0.01 level,
respectively, one-tailed. Variable Definitions:
(a) Does not contain CSECPS or SDAYS.
(b) Contains CSECPS.
(c) Contains SDAYS.

    C8-K 1 = the client filed the Form 8-K with the SEC within the
             required five business-day period, 0 otherwise;
    YEAR 1 = auditor change occurred during 1991-92, 0 if auditor change
             occurred during 1995-96;
    BSIX 1 = predecessor auditor was a Big 6 firm, 0 otherwise;
DISAGREE 1 = auditor-client disagreement(s), 0 otherwise;
  RESIGN 1 = the predecessor auditor resigned or declined to stand for
             reelection, 0 otherwise;
     MOD 1 = the predecessor auditor gave the registrant a going concern
             opinion or included a modifying paragraph related to
             uncertainty, 0 otherwise;
    LASSET = log of total assets;
  CSECPS 1 = the predecessor auditor filed the SECPS letter with the SEC
             within the required five business-day period, 0 otherwise;
             and
     SDAYS = auditor's SECPS letter-filing lag in days.


When the auditor compliance variable CSECPS is entered into Model (2), as reported in the second results column, it has a negative, significant coefficient. Auditor compliance with SECPS filing requirements is positively associated with more timely client Form 8-K filing. When the SECPS variable is added to Model (2), the YEAR variable becomes less significant and the RESIGN variable becomes marginally significant. These results suggest that the decrease in clients' Form 8-K filing lags over time is partially, but not fully, explained by improvement in auditors' SECPS letter filing compliance.

The final two columns in Table 5 provide results for Model (3). The first of the two columns presents results with the auditor filing lag variable excluded from Model (3). As expected, the coefficient on YEAR is positive, indicating longer client filing lags in the earlier period. The coefficients of variables BSIX and LASSET are negative as expected, indicating that clients of the Big 6, and larger companies, file more quickly on average. Clients that received prior qualified opinions take longer to file, as do clients whose auditors resign. Consistent with expectations, clients that express a disagreement with the auditor in the Form 8-K file more quickly. A possible explanation for this result is that managers seek to avert liability through timely disclosure of this bad-news event.

When the auditor lag variable SDAYS is entered into Model (3), as reported in the final results column, it has a positive, significant coefficient. Less timely filing of the auditor's SECPS letter is associated with less timely client Form 8-K filing. When the SDAYS variable is added to Model (3), the coefficient of YEAR does not differ statistically from 0. Results are otherwise similar to those in the third results column. The increase in [R.sup.2] and the significance of the coefficient related to SDAYS indicate that the decrease in clients' Form 8-K filing lags over time is almost fully explained by reductions in auditors' SECPS letter-filing lags.

One procedure that could potentially affect the results is the dichotomous coding of the CSECPS variable as in Models (1) and (2). We re-coded the CSECPS variable to be "on-time filers" (CSECPS1), "late filers" (CSECPS2), and "no filers" (CSECPS3). Since all three variables sum to 1, only two can be used in any regression. We estimated probit Models (la) using different combinations of the three variables, and the results provide a rationale for combining the "late filers" and "no filers" into a single group as in Tables 4 and 5. Table 6 presents probit results, using the trichotomous variables, in the last three results columns. The first two results columns replicate Table 4 for comparison purposes.
TABLE 6
Probit Results Using Different Forms of the CSECPS Variable
Compliance with 8-K Filing Rules (C8-K) is Dependent Variable

Independent    Model (1)    Model (1)(a)   Model (1a)(b)
Variable       (n = 706)      (n = 706)      (n = 706)

Intercept      0.25(**)         0.04           0.15
YEAR          -0.37(***)       -0.20(**)      -0.24(**)
BSIX           0.31(***)        0.22(**)       0.27(***)
DISAGREE       0.08            -0.01           0.00
RESIGN        -0.10            -0.18          -0.21(*)
MOD           -0.07            -0.07          -0.06
LASSET         0.11(***)        0.11(***)      0.11(***)
CSECPS1                         0.50(***)      0.36(***)
CSECPS2                                       -0.25(**)
CSECPS3

Independent   Model (1a)(c)    Model (1a)(d)
Variable       (n = 706)         (n = 706)

Intercept        -0.09           0.51(***)
YEAR             -0.24(**)      -0.24(**)
BSIX              0.27(***)      0.27(***)
DISAGREE          0.00           0.00
RESIGN           -0.21(*)       -0.21(*)
MOD              -0.06          -0.06
LASSET            0.11(***)      0.11(***)
CSECPS1           0.61(***)
CSECPS2                        -0.61(***)
CSECPS3           0.25(**)     -0.36(***)

(*), (**), (***)  Significant at the 0.10, 0.05, and 0.01 level,
respectively, one-tailed.
(a) Contains CSECPS 1.
(b) Contains CSECPS1 and CSECPS2.
(c) Contains CSECPS1 and CSECPS3.
(d) Contains CSECPS2 and CSECPS3.

When only SECPSI is in the regression, it is equivalent to having the
dichotomous variable CSECPS in model (1).

    YEAR 1 = auditor change occurred during 1991-92, 0 if auditor change
             occurred during 1995-96;
    BSIX 1 = predecessor auditor was a Big 6 firm, 0 otherwise;
DISAGREE 1 = auditor-client disagreement(s), 0 otherwise;
  RESIGN 1 = the predecessor auditor resigned or declined to stand for
             reelection, 0 otherwise;
     MOD 1 = the predecessor auditor gave the registrant a going-concern
             opinion or included a modifying paragraph related to
             uncertainty, 0 otherwise;
    LASSET = log of total assets;
 CSECPS1 1 = the predecessor auditor filed the SECPS letter with the SEC
             within the required five business-day period, 0 otherwise;
 CSECPS2 1 = the predecessor auditor filed the SECPS letter with the SEC
             but not within the required five business-day period, 0
             otherwise; and
 CSECPS3 1 = the predecessor auditor did not file the SECPS letter with
             the SEC within the required five business-day period, 0
             otherwise.


The third result column uses separate dummy variables for auditors that file on time (SECPS1) and that file late (SECPS2). The contrast subsample impounded in the intercept is auditors that never file. The coefficients on the two dummies are intercept shifts relative to the subsample in the intercept. Results indicate that clients of timely filers are more likely to comply (coeff. = 0.36) compared to clients of nonfilers. Clients of late filers are less likely to file on time than are clients of nonfilers (coeff. = -0.25).

The fourth and fifth results columns tell the same story. In the fourth column, the contrast subsample is late-filing auditors. Results indicate that clients of timely filing auditors are more likely to file on time than are clients of late-filing auditors (coeff. = 0.61). Clients of auditors that never file also are more likely to file on time than are clients of auditors that file late (coeff. = 0.25). The fifth column results are consistent with the third and fourth column results. In fact, the fifth column coefficient for CSECPS2 is identical to that for CSECPS 1 in column four, but opposite in sign. The same is true for CSECPS3 in column five and CSECPS1 in column three. The changes in coefficient signs are due to the changes in the contrast sub-samples across models.

In summary, the third, fourth and fifth results columns indicate the following. (1) Clients of auditors that file on time are most likely to file on time themselves. (2) Clients of auditors that file late are least likely to file on time. (3) Clients of auditors that never file lie between these two end points. The third observation deserves further comment. Clients that have not received a copy of the auditor's SECPS letter five days after a dismissal cannot know whether their auditor will file later or never. The motivational effect of potentially receiving such a letter after the five-day period (and knowing the SEC has received it) is ex ante identical for both late-filing and nonfiling auditors. This led us to believe clients whose auditors filed on time would be more likely to file on time than either those whose auditors filed late or those whose auditors file never. That result continues to be apparent in Table 6 and justifies the dichotomous classification of auditor performance underlying Tables 4 and 5. Second, our priors were that the filing performance of clients whose auditors filed late would be the same as for clients whose auditors never filed. However, in samples where performance is not the same, as is the case above, we do not think it possible to specify ex ante which subsample will perform better. When performance differs between clients of late filing and nonfiling auditors, the ordering of performance between the two subsamples should be random. This is due to a feature of the client's decision setting that we wish to stress again: The client cannot know whether an auditor that has not yet filed will file later or not at all. Therefore, the ex ante effect of these ex post distinctions on the client's actions is the same. This reasoning also supports our dichotomous classification of auditor performance underlying Tables 4 and 5.

CONCLUSIONS, IMPLICATIONS, AND SUGGESTIONS FOR FUTURE RESEARCH

Summary and Conclusions

This paper investigates the effectiveness of joint public and private monitoring using the case of the AICPA SECPS auditor change notification letter. The notification letter is the monitoring component of the SEC Practice Section (SECPS) requirement that member firms notify the SEC within five business days of the date the auditor determines that its auditor-client relationship with an SEC registrant has ended. The objective of the SECPS regulation is to help the Commission monitor SEC registrants to ensure that they file Form 8-K within five business days. The SECPS's notification rule was a response to concerns that registrants were depriving investors of important information by delaying Form 8-K filings.

We first investigate whether auditor compliance (timeliness) and registrant compliance (timeliness) are positively associated after controlling for other factors. Our results indicate a significant relationship between auditor timely compliance with the SECPS rule and registrant timely compliance with the SEC auditor change Form 8-K filing requirements. They also indicate that auditor filing timeliness is significantly positively related to registrant filing timeliness. We find these relationships after controlling for factors that others studies show are related to registrant compliance and timeliness.

The uniqueness of the SECPS regulation provides the opportunity to empirically investigate whether educational and enforcement efforts by a private organization can assist a public organization to be more effective. This investigation provides evidence related to the debate concerning when self (AICPA) regulation works and when there is a need for greater outside (SEC) regulation. SECPS member-auditors initially were unaware that filing the SECPS letter was mandatory, or they were slow in filing. Beginning in 1993 and continuing into 1995, the SECPS mounted a significant educational effort. In this paper we sample from both the pre-SECPS educational effort period (1991-1992) and from the post-SECPS educational effort period (1995-1996). This sample allows us to investigate how effective a private professional organization (the AICPA) can be in prompting its members to monitor and disclose information that assists a public organization (the SEC). We do this by investigating whether compliance rates (or timeliness) improved for auditors and if these improved compliance rates seemed to affect registrant-filing behavior between the pre and post periods.

We found auditor compliance with the SECPS rule improved from 26 percent to 62 percent during the period. This finding suggests that the SECPS educational program was effective. Timely Form 8-K compliance rates for registrants increased from 61 percent to 76 percent. These results indicate that auditors' independent evaluation of a regulatory question of fact appears to have an effect on client compliance (e.g., speed of filing in the specific case examined). More generally, the results suggest that monitoring requirements imposed on auditors by self-regulation (e.g., the AICPA's SEC Practice Section in the specific case examined) can assist a regulatory body (the SEC) in its duties.

Limitations

This study examines changes over time in client compliance with Form 8-K filing rules. We attribute improvements in compliance in part to the SECPS Committee's educational efforts. However, other factors that changed over time might also account for greater compliance. These include possible greater SEC diligence and greater familiarity by clients with the law, regardless of actions by the SECPS Committee. Another limitation is that the two subsamples were collected at different points in time, and it was not possible to use identical procedures at every step of the data gathering process.

Implications

In recent years, the Chairman of the SEC has expressed growing concern over the effectiveness of auditors in enhancing the quality of financial reporting. The AICPA responded by working with the SEC to form an Independence Standards Board (ISB) in 1997 to establish independence standards for public entities. Similarly, the New York Stock Exchange and the National Association of Securities Dealers formed the Blue Ribbon Committee (BRC 1999) on Improving the Effectiveness of Corporate Audit Committees in 1998.

The SEC recently proposed implementation of several BRC recommendations. One would require that proxy statements include a report from the audit committee indicating, among other things, whether the committee discussed independence issues with the auditor and received the communications required by ISB Standard No. 1, Independence Discussion with Audit Committees. The standard directs the auditor to disclose in an annual letter to the audit committee all relationships between the auditing firm and the company that could bear on independence. The letter should include a confirmation of the auditor's independence. This requirement is in essence a call for additional self-regulation that would put responsibility on the auditor and client to regulate their relationship so that independence is preserved. The results of our study suggest (by analogy and extension) that this type of monitoring requirement self-imposed on auditors may be effective in helping the SEC and the public ensure auditor independence.

Future Research

Because the SEC is not the only regulatory body that depends on the accounting profession for monitoring assistance, other opportunities for research on the effectiveness of other joint monitoring arrangements exist. One is the National Association of Insurance Commissioner's' (NIAC 1999) requirement that auditors report material misstatements of financial condition or failure to meet minimum capital and surplus requirements to state insurance commissioners if the insurer's board of directors fails to do so within five business days of notification by the auditor. Another is the U.S. General Accounting Office (U.S. GAO 1999) Government Auditing Standards (the Yellow Book) requirement that under certain circumstances the auditor must report illegal acts to the appropriate outside entities (e.g., funding agency or law enforcement agency). To the extent that sufficient data for empirical inquiry exist, these requirements provide opportunities for researchers to study the effectiveness of joint monitoring arrangements in situations for which the event to be reported is less objectively identifiable than the auditor notification requirement examined in this paper.

The authors appreciate the financial support of Ernst & Young LLP and the Union Pacifc Corporation. They also appreciate the valuable suggestions of Paula Panik and the research assistance of Tony Lendez and J. Roddy of the AICPA, Mike Geary, Carolyn Hartwell, Jerry Sullivan, Robert Bums, and Scott Bayless, especially Jeff Blankenship, Thomas L. Milan, and the two anonymous reviewers.

(1) When an auditor becomes aware of information that an illegal act may have occurred, the Reform Act requires the auditor to determine whether an illegal act has occurred, determine its possible effect on the financial statements, and inform the appropriate level of management. If management does not take appropriate remedial action, then the auditor must report its conclusions to the board of directors. The company is required to notify the SEC of the auditor's conclusions within one business day. If the company does not do so, then the auditor is required to notify the SEC the next business day.

(2) The objective of the SEC Practice Section is to improve the quality of practice by CPA firms before the Securities and Exchange Commission through the establishment of practice requirements for member firms. All auditors who are AICPA members must be members of the SECPS if they audit SEC registrants.

(3) The letter (the first auditor letter), which is the topic of this paper, is filed by the auditor directly with the SEC, to give the SEC independent notification that an auditor change has occurred. When the auditor sends the first letter to the SEC, the auditor also must send the client a copy. This letter is an exception to normal practice in which the auditor does not have direct contact with the SEC related to registrant affairs. The auditor also provides a second letter to its former client. That letter expresses agreement or disagreement with the former client's comments in the auditor change Form 8-K. The auditor may not file the second auditor letter directly with the SEC, but the former client is required to provide the SEC with a copy. The first letter, which is the subject of this study, should not be confused with the second letter.

(4) For the auditor, compliance refers to fulfilling the SECPS rule that the auditor files a notifying letter with the SEC within five business days of an auditor change. For the registrant, compliance means that it files an auditor change Form 8-K within five business days of the dismissal or resignation of the auditor. Timeliness refers to quickness in filing. We measure compliance using dichotomous variables (you either comply with the regulation or you do not) while timeliness is measured with a continuous variable, which is the number of days to file.

Submitted November 1997 Accepted January 2000

REFERENCES

Blue Ribbon Committee (BRC) on Improving the Effectiveness of Corporate Audit Committees. 1999. Report and Recommendations of the Blue Ribbon Committee on Improving the Effectiveness of Corporate Audit Committees. New York, NY: The New York Stock Exchange and the National Association of Securities Dealers.

Carter, M. E., and B. S. Soo. 1999. The relevance of Form 8-K reports. Journal of Accounting Research 37 (Spring): 119-132.

DeFond, M., M. Ettredge, and D. Smith. 1997. An investigation of auditor resignations. Research in Accounting Regulation 11: 38-62.

Dhaliwal, D. S., J. W. Schatzberg, and M. A. Trombley. 1993. An analysis of the economic factors related to auditor-client disagreements preceding auditor changes. AUDITING: A Journal of Practice & Theory 12 (Fall): 22-38.

Fried, D., and A. Schiff. 1981. CPA switches and associated market reactions. The Accounting Review 56 (April): 326-341.

Lawrence, E. C. 1983. Reporting delays for failed firms. Journal of Accounting Research 21 (Autumn): 606-610.

Milan, T. L. 1993. Reporting a change in auditors. Journal of Accountancy 175 (May): 105.

National Association of Insurance Commissioners (NAIC). 1999. Annual Statement Instructions--Property-Casualty. Washington, D.C.: NAIC.

Schwartz, K. B., and K. Menon. 1985. Auditor switches by failing firms. The Accounting Review 60 (April): 248-261.

--, and B. S. Soo. 1995. An analysis of Form 8-K disclosure of auditor changes by firms approaching bankruptcy. AUDITING: A Journal of Practice & Theory 14 (Spring): 125-136.

--, and --. 1996. Evidence of regulatory noncompliance with SEC disclosure rules on auditor changes. The Accounting Review (October): 555-572.

Securities and Exchange Commission (SEC). 1971. Notification of Adoption of Amendments to Form 8-K, Form 7-Q, Form 10-Q, Form 10-K, and Form N-1 Q. Washington, D.C.: SEC.

--. 1989. Acceleration of the Timing for Filing Forms 8-K Relating to Changes in Accountants and Resignation of Directors, and Amendments to Regulation S-K regarding Changes in Accountants. Financial Reporting Release No. 34. Washington, D.C.: SEC.

Skinner, Douglas J. 1994. Why firms voluntarily disclose bad news. Journal of Accounting Research (Spring): 38-60.

--. 1997. Earnings disclosures and stockholder lawsuits. Journal of Accounting and Economics (November): 249-282

Smith, D. B., and D. R. Nichols. 1982. A market test of investor reaction to disagreements. Journal of Accounting and Economics 4 (December): 109-120.

U.S. General Accounting Office. 1999. Government auditing standards (the Yellow Book). Washington, D.C.: General Accounting Office.

Whittred, G., and I. Zimmer. 1984. Timeliness of financial reporting and financial distress. The Accounting Review 59 (April): 287-295.

Michael L. Ettredge is an Associate Professor at the University of Kansas, David B. Smith is a Professor at Iowa State University, and Mary S. Stone is a Professor at The University of Alabama.
COPYRIGHT 2001 American Accounting Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2001 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Ettredge, Michael L.; Smith, David B.; Stone, Mary S.
Publication:Auditing: A Journal of Practice & Theory
Article Type:Statistical Data Included
Geographic Code:1USA
Date:Mar 1, 2001
Words:8228
Previous Article:Further Evidence on the Auditor's Going-Concern Report: The Influence of Management Plans.
Next Article:An Examination of Auditors' Reporting Intentions when Another Auditor is Offered Client Employment.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters