The Introduction of State Regulation and Auditor Retendering in School Districts: Local Audit Market Structure, Audit Pricing, and Internal Control Reporting Free

In response to the public outcry following an $11 million embezzlement at the Roslyn Union Free School District in the State of New York, state lawmakers responded with regulations aimed at improving the disposition of public tax funds. Among the regulations enacted was an audit retendering mandate under which school districts became required to document auditor characteristics including an auditor's specialized knowledge, independence, experience, and anticipated staffing levels on the engagement in making the school board's auditor selection. The retendering was to be conducted every five years. Unlike an auditor rotation mandate, school districts could reappoint the existing audit firm, encouraging auditors to actively build a quality-based reputation, while the documentation process required in the mandate precluded districts from lowest-bid auditor selection common within governmental audit markets (GAO 1987). We report the effects of implementation of the regulations on the school district auditing market within New York State.

After the 2004 discovery of the Roslyn scandal, the Office of the New York State Comptroller (OSC) conducted a state-wide series of inspections of school district finances and discovered widespread weaknesses in the design of internal control systems as well as numerous other frauds throughout the state. Interestingly, control deficiencies were not detected and reported, despite a long-established, state-mandated annual independent audit requirement, or even through expanded internal control testing mandated for recipients under the Single Audit Act (SAA).¹ The OSC discovered the vast majority of school boards did not periodically review their auditor selection, which along with inherently low audit pricing were interpreted as factors contributing toward complacency and entrenchment within the audit market.

The extent of non-Big 4 auditor specialization within governmental audit markets varies across the country, and factors influencing auditor demand within the governmental and private company audit markets are not fully understood. In the governmental sector, regulators such as the U.S. Government Accountability Office (GAO) have encouraged governments to include auditor retendering (or rotation) as a component of a strong control environment and as a tool ensuring quality auditor selection through competition for an audit award (AICPA 2017; Elder, Lowensohn, and Reck 2015; Copley and Doucet 1993; GAO 1986). Research finds differing use of specialist auditors in governmental audit markets with some research finding evidence of dominant specialists (Chase 1999; Ward, Elder, and Kattelus 1994), while other researchers find evidence of extensive use of small, nonspecialized auditors within the markets (Lowensohn, Johnson, Elder, and Davies 2007). The results in our study are important to accounting practitioners interested in understanding the factors driving an audit market toward specialization, and our findings are especially important in understanding local markets not dominated by the Big 4 (Yebba and Elder 2019; López and Peters 2010; Hackenbrack, Jensen, and Payne 2000).

Our research summary highlights questions exploring the effects of auditor retendering mandates on the types of auditors serving a local audit market and presents evidence of greater use of specialized audit firms with retendering policies. We also examine audit effort and find that with implementation of a set of quality-focused policies, auditors' effort increases as entrenchment is reduced. Finally, we explore the effects of regulations on audit outcomes, including internal controls reporting, and summarize evidence of enhanced disclosure of control deficiencies within a regulated market containing greater concentration of specialist auditors. Overall, our research findings contribute to the literature by demonstrating how public policy focused on strengthening municipal internal control systems can ensure greater disposition of public assets, and consequently restructure an audit market along quality lines. While our findings are based on a sample from one governmental audit market, our research summary presents important findings useful to accounting professionals in other settings, as limited empirical research settings exist to explore the effects of retendering and other regulatory policies in other commercial and governmental audit markets.

Public funding for school districts within New York comes from a separate assessment of property tax, and rates are approved annually by local voters.² Immediately after discovery of the Roslyn scandal, school district budgets across the state were increasingly rejected by voters as public confidence in school district officials, state regulators, and the independent audit process was lost. Like most scandals, the exact start date of the Roslyn fraud is unknown, but most believe the $11 million was stolen by the district's superintendent and other school officials over a period of at least eight years, and throughout this time, school rankings for Roslyn were among the highest in the nation—taxpayers were satisfied exchanging greater taxes for highly ranked schools and were distracted from financing details by the educational success.³

Roslyn's independent auditor was widely recognized as the most regarded specialist within the nuances of New York school district reporting (Akhtar 2005). The firm reported on at least 50 other districts and provided additional services through its own specially developed accounting software program. The administration at Roslyn relied on a flaw in the development of this accounting software to run a vendor coding scheme where false expenditures could be posted to the general ledger with seemingly legitimate vendor names (e.g., Houghton Mifflin Harcourt), but the subsequent cash disbursement could be altered with a check paid to another vendor for a fraudulent expense (e.g., American Express Company).

The fraud was first brought to the auditor's attention in 2002 when the firm was engaged to conduct an investigation of misappropriations fraudulently recorded by the assistant superintendent for business. At this time, the district's board decided not to disclose this investigation to the public but did accept restitution and terminated the assistant superintendent (OSC 2005). Subsequent-year audit documentation, including audit planning documents, neglected to mention knowledge of the fraud or how the accounting software had been manipulated in the fraud's concealment. Two years later, Roslyn's audit firm became complicit in the concealment of the recurring fraudulent activities at their client as allegations surfaced that an administrative login, reserved for software personnel, was being used in recording a series of transactions. These transactions were later determined to be fraudulent, almost all of which were authorized by the district's superintendent and recorded under the accounting firm's administrative login (OSC 2004).

In response to public outcry at the scandals, the OSC became increasingly involved in the school district audit market, including the auditing of internal controls at every district. Upon review, the OSC was outraged by the lack of due care school district officials exercised toward custody of taxpayer resources and was especially enraged by the independent audits, many of which did not follow professional standards (Havesi in Lambert [2005]). Following passage of several legislative acts, the OSC began inspecting all school districts, and these inspections went beyond review of auditees and included inspection of audit workpapers of at least ten independent CPA firms. These inspections resulted in the dissolution of the large Long Island-based specialist servicing Roslyn—raising the question as to whether all specialist auditors are equal and whether traditional market-based measures of specialization common in research adequately measure audit quality.

As OSC inspections progressed through the state, inspectors found repeated instances of auditors failing to follow professional standards including the lack of audit planning and fraud risk assessment, testing of controls, and fully satisfying substantive evidence requirements in forming an opinion. In addition to Roslyn's auditors, several firms were reported to the New York board of public accountancy for substandard attestation work. Overall, at the time of the scandal's discovery, the audit markets contained a mix of audit firms, including sole practitioners and several specialist auditing practices. These firms competed in a market where federal reporting guidelines under the SAA complicate the reporting of school districts receiving federal awards, and lowball pricing strategies and a lack of local-level oversight pressured reporting quality. A summary of regulations adopted during 2005 in response to the Roslyn scandal is contained in Figure 1.

We first explore the effects of the regulations on the types of auditors serving the audit market. Governments with strong controls will competitively bid expenditures, including auditing services, and the pressures that bidding places on auditing firms will likely result in some of the most qualified firms winning market share. We predict that the effects of enhanced control environments from the regulations will result in a restructured audit market, served primarily by large, specialist auditors. We state the first hypothesis as follows:

Second, we explore the effects of the regulations on overall auditor effort. A complacent audit market will be characterized by auditors not fully complying with professional standards—in the absence of a scandal, the unobservability of auditing would mask this activity. After the markets become regulated, we anticipate the quality focus defined under the regulations will result in increased effort observable through greater overall audit pricing and greater amounts of time required to complete the audit; quality will be demonstrated through greater detection of control deficiencies. We state our second set of hypotheses as follows:

Finally, we explore the effects of auditor specialization on these audit outcomes. Audit quality is dependent on both independence and detection expertise, but empirical measurement is difficult. While specialization may have become more valued under the Act, the increased regulation and board oversight may have resulted in an environment of uniform conformance with audit standards, and no observed effort or pricing differences may be found between audit firms. However, our study predicts greater demand for specialized audit firms and those audit firms that choose to specialize in this niche market will have the ability to charge a premium for their increased effort and reputation. We also predict a greater association of the regulations with observed audit outcomes, including the reporting of control deficiencies by specialists. We state our third set of hypotheses as follows:

Issues of audit procurement and audit quality have long been of concern in the governmental sector. One argument is that active procurement leads to increased competition and greater likelihood of selection of a quality auditor. However, active bidding is also often associated with lower audit fees that may be detrimental to audit quality. The increased oversight of auditors in New York school districts as a result of the Acts provides a unique opportunity to evaluate how regulation combined with active procurement affects audit markets, audit fees, and audit quality. In the years before the scandal, all school districts were required to obtain an independent audit, and many school districts were subject to compliance auditing under the SAA with many audits conducted by local specialist auditors. Despite this, the lack of direct oversight and interest by taxpayers and a state comptroller's office that placed reliance on independent audit reports may have contributed toward complacency, and audit evidence may not have been satisfactory.

While heightened regulations should be associated with greater audit effort, the nature of school district financial reporting was largely unchanged both before and after the regulation's passage, and this includes mandates under the SAA. As the frauds made school boards more aware of the importance of independent auditing, and because Roslyn's auditor was a leading specialist in the market, there may be greater or lesser demand for established specialist auditors as school boards become more involved in auditor selection. Like most governmental audit markets, lowball audit pricing was commonplace within this market; auditor competition through mandated retendering may produce either continued lowballing, unchanged audit pricing, or greater pricing matched with greater auditor effort and increased use of specialist auditors.

The primary data reported on in this summary are drawn from New York school districts in the Federal Single Audit Clearinghouse database of the U.S. Census Bureau for fiscal years 1998–2012.⁴ Observations from these data were combined with a dataset of fiscal information available under the OSC's transparency initiatives, and these data include independent audit pricing and a series of financial statement variables used as controls in our analysis. Of the 700 school districts in New York, our analysis represents an annual sample of approximately 425 (61 percent) districts over the 15-year sample period. We restrict our results to districts with observations appearing at least 13 times during this 15-year period, and our sample includes 2,910 and 3,443 observations from the pre-regulation and post-regulated markets, respectively. We summarize our sample in Table 1.

Table 2 presents descriptive statistics on variables used in presenting the summarized results of our hypothesis testing. Variables are partitioned between the pre- and post-regulated audit markets, respectively, and include identification of variable use and source. Hypotheses testing includes regression models that test our predictions using an indicator variable for the post-regulated audit market and for auditor type (i.e., specialization). Our models also include the control variables summarized in Table 2 to capture effects of other school district characteristics; all models also control for changes in time trends. Our main test is for significant differences between the pre- and post-regulated audit markets, respectively, and between specialized and nonspecialized auditors. We present and discuss tabulated results for the research variables and graphical illustrations of the main findings. Results for control variables are not included in the tables for brevity; these control variables measure factors such as school district revenue, budget surplus, involvement within the SAA, and an efficiency ratio measuring direct educational spending scaled by non-educational spending. Other control variables are focused on compliance within the SAA and include audit report type and the history of and current reporting of material weaknesses.

In Figure 2, we present descriptive results illustrating changes in the types of audit firms servicing this market and summarizing the effects of testing H1. For brevity, we do not present regression results detailing these findings. We identify four market-based categories of audit firms in our analysis, and these categories are determined annually. Low market share auditors conduct five or fewer audits, medium market share auditors conduct 6–20 audits, and high market share auditors conduct greater than 20 audits annually; we consider these firms to be specialists. In addition, we analyze a fourth category representing auditors cited for low audit quality by the OSC; one of the three auditors in this category is the large specialist complicit in the Roslyn scandal.

Consistent with the first hypothesis, Figure 2 indicates increased audit market concentration following the passage of the Acts. In the beginning of the sample period, the audit market contained 100 mostly smaller audit firms, many of which served a single district in a year. The market did contain several specialist auditors, but these auditors collectively served an average of just 20 percent of the districts, compared to the 45 percent served by smaller firms. By the end of 2012, most districts had retendered their audit twice, and we find that by this point, 55 audit firms remained in the market. The majority of the school districts (50 percent) were serviced by specialist auditors, several of which grew their market share throughout the sample period. We also find that cited auditors that served 8 percent of the unregulated market largely disappeared by the end of the sample period.

In Figure 3, we present trends in auditor pricing. Audit pricing significantly increased in the post-regulation audit market, consistent with H2a. Audit fees increased despite the active retendering as the audit market underwent a quality-focused reform, reflecting greater auditor effort and the effects of school districts documenting their retendering decisions.

Table 3, Panel A includes summarized regression results of the effects of regulation and auditor specialization on audit fees. Specialized (i.e., high market share) auditors did not earn significant premiums in the unregulated market, and the smallest audit firms tended to discount their services, while the auditors cited for low audit quality by the OSC tended to price at low levels despite containing a highly regarded local specialist. After the regulations were implemented, use of specialists increased and those auditors with the highest market shares began earning a significant premium, consistent with H3a. In untabulated testing, we explore whether specialist auditors that grew market share over the sample period price similarly with established market leaders. Interestingly, we find evidence that growth-oriented specialist auditors priced below market levels in the unregulated market but priced above the levels of established market leaders once becoming specialists, an important finding illustrating that fee premium growth strategies can emerge in a setting containing a quality-focused retendering strategy.

Consistent with evidence of increased audit effort in a regulated market, Figure 4 shows increases in the amount of time (in days) auditors require to submit audit reports after the regulations, consistent with H2b. Combined with the results shown in Table 3, Panel B, we find that specialist auditors required more time compared to nonspecialized auditors and auditors cited for low audit quality, consistent with H3b. After the regulations, there is evidence specialist auditors had significantly longer reporting lags while smaller auditors and the cited firms continued to issue their reports with less reporting lag, indicative of less effort.

Figure 5 shows a large spike in reported control deficiencies following enactment of the regulations, consistent with H2c. If the independent auditing process was effective, we would have observed the detection and reporting of control deficiencies, regardless of heightened regulation. In the unregulated period, state regulators relied upon the independent auditing process to sufficiently oversee controls at school districts, and from the perspective of the OSC, they would have believed the lack of reported control deficiencies reflected strong controls, not a weak auditing process. The decline in the trend lines toward the end of the sample period reflects the overall effects of remediation of deficiencies initially reported following the regulations.

In Table 3, Panel C, we summarize regression results of the reporting of control deficiencies by auditor type, including by firms that were specialist auditors at the beginning of the sample period and those that grew their practices throughout the sample period. Both reportable conditions and material weaknesses were more likely to be reported post regulation. Contrary to H3c, we find evidence specialists auditors are associated with reduced reporting of material weaknesses, and this effect is greater for long-standing, original specialist auditors. In untabulated results, we find an inverse association between original specialist auditors and material weakness reporting; this finding is comparable across the unregulated and regulated periods. We interpret this result as consistent with an association between specialist auditors and the delivery of quality reporting; a particularly interesting finding in that material weakness reporting is unchanged for established specialist auditors, despite increased internal control inspections by local regulators and enhanced awareness of these matters by school district boards. This finding suggests greater reporting quality with specialization, regardless of OSC regulations.

It is important to understand determinants of auditor demand in non-Big 4 markets. Despite their smaller size, these firms provide a large amount of assurance services within the United States. However, data constraints usually preclude research exploring determinants of demand and pricing. The governmental audit markets are an especially valuable research setting, as these markets are served primarily by local auditors and most states have unique reporting requirements resulting in geographically constrained markets useful in conducting an analysis of non-Big 4 auditors at the local level.

Our paper explores the effects of a bundle of regulatory enhancements to a local market passed in response to the discovery of a large fraud, including an audit retendering mandate. Although regulators have explored auditor rotation as an enhancement to quality, our study suggests increased competition for winning or retaining an audit award may be an alternative means of reducing entrenchment and enhancing quality while not forcing inefficiencies associated with mandatory rotation.

First, we find evidence the regulations resulted in a restructuring of the audit market. Auditors with a negligible market share were replaced with specialized auditors, including new specialists that grew through the retendering. Contrary to the expectation that active retendering is associated with downward audit pricing, we find fees increased with the regulations. In addition, we find that specialist auditors began earning fee premiums and, importantly, auditors that grew market share in response to the regulations were able to grow their client base while earning premiums above those of established specialists. These premiums likely reflect the effects of specialization and billing for increased auditor effort illustrated through longer reporting lags.

We find that in the absence of regulatory oversight, smaller audit firms participate in the market, despite its complex reporting requirements, while discounting fees and quickly issuing reports. Without oversight, specialized auditors may feel pressures to perform substandard auditing services in exchange for the opportunity to upsell client services such as software design and maintenance. An important finding of our research is the idea that not all specialist auditors are equal. Traditional, market-based measures used in research treat all specialists the same and have produced mixed research findings.

The involvement of regulators in an audit market can most certainly affect quality, and our paper finds extensive evidence of this. An open question is whether dual inspection of auditees, by both an independent auditor and by a state regulator, can improve quality. Interestingly, our paper finds little evidence of an association between internal control inspections performed by state regulators and auditor selection and measures of audit pricing and quality. This finding suggests that while the school district audit market in New York was not effective without regulation, the solution may not be secondary levels of auditing. Rather, an environment encouraging competition within the market and documentation of audit quality determinants during retendering may be sufficient in improving market quality while making audit pricing sustainable for audit firms serving this market. While we recognize the inherent empirical challenges in measuring outcomes over a time series, we performed sensitivity testing using control groups unaffected by the Acts, and we found consistent evidence supporting the findings presented in this paper.