Skip to main Content
Site Search

Advanced Search

  • Mondo Visione
  • Mondo Visione - Worldwide Exchange Intelligence
Member Login

Member Login

Forgotten your password?

Corporate Governance Technology Applications to Accounting and Audit Quality Surveillance

Date 07/06/2004

Carsten Friedrich, Donald Stokes and Stephen Taylor [1]

Capital Markets CRC Ltd in collaboration with the University of NSW and University of Technology Sydney.

Introduction

The critical role of accounting and audit surveillance in corporate governance has recently been highlighted through a number of well known corporate collapses. Following cases like Enron and Worldcom in the United States and HIH Insurance in Australia, accounting and audit surveillance is increasingly in demand as an important toolset in decision making by market participants that include company directors and auditors, insurers, lenders, analysts, and investors.

This article outlines the value of and underlying demand for surveillance applications of research-based corporate governance technology solutions that facilitate benchmarking of company accounting and audit quality. The Capital Markets CRC (www.cmcrc.com)is a cooperative research centre across universities and industry established with a mandate from the Australian government (along with its funding support) to undertake research into corporate governance surveillance and to work with business to deploy surveillance technology solutions worldwide. We overview some of the key research upon which the technology is being constructed and explain how the technology can be used to maintain surveillance over aspects of accounting and audit quality of interest to various market participants.

Heightened demand for accounting and audit quality surveillance

As with other jurisdictions such as the United States and Europe, Australia is experiencing increased regulatory surveillance of accounting and audit quality with reform of the institutional arrangements governing accounting and audits of company accounts. The increased interest in surveillance has largely been based on a tightening of the laws and regulations governing company disclosures and audits. Hamilton and Stokes (2004) [2] and Culvenor, Stokes and Taylor (2002) [3] summarise many of these developments. These papers note that in Australia, the key developments have been the release of the Ramsay Report, [4] the report of the HIH Royal Commission into the failure of HIH Insurance [5] and various reports and submissions at government level, including the CLERP 9 discussion paper and the subsequent draft CLERP (Audit Reform & Corporate Disclosure) Bill. [6] [7] Internationally, in the wake of corporate failures such as Enron, the most significant reforms include the US Sarbanes-Oxley Act (2002). [8]  

Accounting risk surveillance systems and, to a lesser extent, audit quality surveillance systems, are used by many organisations worldwide. They contribute to the identification of questionable aspects of periodic accounting information, as well as possible instances of poor quality auditing. Such systems are relatively unsophisticated with almost a complete absence of any systematic, let alone scientifically based, approaches for identifying questionable accounting and audit quality and technology for implementing such approaches. The lack of effective surveillance processes results in losses, errors, manipulation, fraud, and inappropriate accounting and auditing practice. In Australia, the government subsidises investors in this process via the operation of the Australian Securities and Investments Commission (ASIC) and the Australian Prudential Regulatory Authority (APRA) and their associated international affiliates as well as via legislation and inquiries as outlined above.

This practice is not restricted to Australia. We see similar levels of subsidisation in other countries, e.g., in the United States in the wake of Enron’s collapse. The cost of accounting and audit quality surveillance delivery to government, and ultimately to investors and consumers, has risen sufficiently to be a driver for ASIC, APRA and other organisations (including the Australian Stock Exchange (ASX), companies being regulated, their auditors, investors and analysts) involved in relying upon or establishing corporate governance, to better control accounting and audit quality and better predict and manage their exposure. There are significant expectations on government and regulatory organisations to oversee improvements in the quality of accounting and auditing, and this in turn will continue to increase the pressure on regulators to more efficiently identify and report instances of poor quality financial reporting and/or auditing. Likewise, the organizations that are being monitored (i.e., corporates, their directors, auditors and advisors) in turn face pressure to manage and control their accounting and audit quality risk. These problems exist internationally and effective solutions are being sought by international equivalents of ASIC and APRA and the other organisations to ensure their long-term viability.

In an environment of increased regulatory interest, scientific based technology for surveillance of company accounting and auditing is useful for market participants, companies, auditors and regulators in particular, for at least four reasons.

Improved compliance and control

Detection of potential problems is needed to alert regulators and other organisations to accounts and audits which should be rejected or at least challenged. Automated assessment of accounts and audits is needed as soon as possible after the release of accounts so that regulators and other organisations can identify high risk accounts and audits and, if necessary, intervene. Investors and taxpayers, companies and their advisers need to know that effective checking processes are in place. The deterrent effect can be expected to be significant.

Improved assessment of accounts and audit provider behaviour

At this stage, assessment of company accounting and audit behaviour is limited because of a lack of transparency on the process, to a certain extent the lack of atomistic data quality and piecemeal approaches to addressing aberrant provider behaviour. A service that automatically assesses the accounting and audit quality of all providers, including listed and unlisted companies, is needed to ensure regulators and those relying upon the accounting and audit quality are able to properly target their greatest risks and manage them.

Reduction of “failure” rates

Yearly increases in the budgets of regulators have at best only paralleled the progressive increased expectations about their ability to effectively identify declines in accounting and audit quality. This increases the cost to government and taxpayers as well as consumers directly through appropriation bills and indirectly through higher compliance costs and insurance costs for business each year. Government, consumers and investors, organisations and their advisers have incentives to control these costs, and one way in which this can be achieved is through better compliance and control of the potential for corporate failures due to poor accounting and audit quality.

Improved investment risk management

Analysts, investors, insurers, and company directors need improved means of predicting and controlling their investment and insurance risks by better understanding accounting and audit risks and provider behaviour through profiling of their quality. This will enable better design and modelling of investments to better meet the needs of consumers while eliminating those that perpetuate errors, manipulation, fraud, abuse and inappropriate services.

Accounting and audit quality defined

Modelling of accounting quality and audit quality can be employed to find and characterize outlier accounting and auditing quality to serve these four objectives. While there is no standard definition, accounting quality is, ultimately, the extent to which published, periodic accounting data provides a “reliable” indication of the underlying performance of the reporting entity. Some examples of accounting quality concepts include equivalence to underlying cash flow, the extent of accrual manipulation, or even the deliberate incorporation of a degree of “conservatism” in the numbers. Audit quality is defined by the competence and independence of the auditor, given the demand and supply conditions governing an audit. Competence is the probability an auditor will discover an error or irregularity in the accounts (including low accounting quality) and independence is the probability that, having made such a discovery, an auditor reports the findings to shareholders and other parties required under contracts or by law. Some examples of audit quality concepts include the extent to which audit fees charged meet expectations about normal behaviour, and the extent to which fees for non-audit services relative to audit fees meet expectations about normal behaviour.

Operationalising these concepts to generate corporate governance technology solutions requires research that models accounting numbers and audit phenomena (like audit fees charged) and sets up expectations about normal behaviour in those numbers and phenomena and then establishing scores that benchmark companies on the extent of their unusual behaviour.

Some accounting quality modelling evidence

Modelling thus far has shown that for a very large sample of ASX listed firms between 1993 and 2002, there is some evidence consistent with “benchmark beating” behaviour by these firms. Benchmarks examined were zero earnings (avoiding a loss) and last year’s earnings (avoiding an earnings decline). Benchmark beating behaviour, although statistically testable, is also a very “visual” concept lending itself to visualization technology.

Research sponsored by the Capital Markets CRC (Coulton, Taylor and Taylor, 2004) [9] has used the ASPECT financial database as the primary data source for a comprehensive examination of the properties of the distribution of earnings for a sample of Australian publicly listed companies. For all available firm years, the authors obtained operating income, [10] cash flows, sales revenue and asset data, for the period 1993-2002.

Figure 1 reports operating profit after tax for year t (OPATt) scaled by total assets from year t-1 (TAt-1). The distribution of scaled OPAT is graphed to identify possible discontinuities in the distribution consistent with earnings management. The initial sample consists of 7,298 observations, following basic data checks but without removing or controlling extreme observations. In addition to graphing the distribution of current earnings, the distribution of the change in earnings (OPATt – OPATt-1) is shown in Figure 2. This ‘earnings change’ sample contains 5,950 observations. Both figures show a distribution from –0.25 to +0.25 with bin width of 0.01. Observations lower than –0.25 or greater than 0.25 are in the extreme bands. The Figure 1 graph of OPATt/TAt-1 shows a lower than expected frequency of observations falling immediately below 0 and is consistent with the theory that some earnings management occurs to avoid the zero earnings benchmark. Although not so strong, this trend is also present in the distribution of changes in earnings shown in Figure 2. Hence, there is at least preliminary evidence of some degree of benchmark beating behaviour by Australian firms.

However, there are many other ways of capturing potentially low accounting quality. For example, one widely accepted approach in the academic literature is to estimate the “unexpected” component of periodic accrual adjustments. Given that earnings are comprised of cash flows which are subsequently adjusted by accounting accruals, this approach is premised on the assumption that accounting manipulation is more likely to occur via the accrual process than by manipulating cash flows. A simple example of this approach is to model the expected value of the accrual component of earnings, viz:

E(Acc) = a + B1(∆REV-∆REC) + B2(PPE) + e (1)

In equation (1) the accrual component of earnings [E(Acc)] is modelled as a function of the change in revenue after adjusting for the change in receivables (∆REV-∆REC), and the level of property, plant and equipment (PPE). Unexpected accruals are therefore the difference between the expected value from a model such as equation (1) and the actual value of accruals (calculated by subtracting operating cash flow from income). Dechow Richardson and Tuna (2003) [11] discuss a number of extensions to this model (e.g., controlling for current performance), and Coulton et al (2004) show that these can improve the basic model in equation (1). More importantly, they show that for Australian firms, benchmark beating of the type highlighted in Figures 1 and 2 is associated with significantly more positive unexpected accruals (when compared to all other firm years), and this is robust to a number of different models of unexpected accruals and to methods of estimating these models. However, there is also some evidence that unexpected accruals are also significantly higher for firms that “just miss” the benchmark when compared to all other firms, and this brings into some doubt the value of “simple” visualizations of earnings (or earnings changes) distributions as a method of identifying specific instances of low accounting quality. The research also highlights a number of issues in “calibrating” measures of unexpected accruals so as to get meaningful measures of possible accounting manipulation. Some of these issues have received scant attention in the research literature, but are of critical importance if such models are to be applied to firm-specific measures of accounting quality.

Researchers have also examined whether the stock market appears to understand the different properties of the cash flow and accrual components of earnings. Given that a key application of the accounting and audit surveillance is to investment evaluation, a logical place to start is the extent to which the market’s evaluation of earnings might display the same “naivety” as has been documented in the United States (e.g., Sloan, 1996). [12] Preliminary results of research by the Capital Markets CRC (Bird, Taylor and Thosar, 2004) [13] suggests that similar, although somewhat weaker results hold for Australian firms. Using a similar data set as depicted in Figures 1 and 2, the authors demonstrate that firm years with relatively extreme positive (negative) accrual components of earnings are overpriced (under-priced), leading to negative (positive) unexpected returns. As an example, Figure 3 summarizes evidence that the Australian market appears slow to appreciate the different persistence of the cash flow and accrual components of earnings. In Figure 3, firm years are sorted on the accrual component of earnings (deflated by total assets) and placed into quintiles, with quintile 1 having the smallest accruals and quintile 5 the largest. Returns (using market weighting) are calculated for each portfolio, and compared over a 48 month period (-24 to +24) relative to the earnings announcement date (month 0). It seems clear that the market reacts with some enthusiasm to accrual-induced earnings, but subsequently (and predictably) becomes disappointed when earnings cannot be sustained.

As part of its suite of corporate governance technology solutions, the Capital Markets CRC has constructed accounting quality surveillance applications that can assist in identifying poor quality earnings (www.cmcrc.com/Products/accountingx.html). Poor quality earnings can be of concern to investors for precisely the reason highlighted above, namely the failure of markets to adequately distinguish even the most basic properties of accounting, such as the different behaviour of the cash flow and accrual component of earnings. However, while metrics such as those being developed by the Capital Markets CRC can be of use to investors as a means of identifying opportunities to take advantage of markets “slowness” to react, they can also help protect investors from falling victim to questionable accounting. The technology currently available enables the following two dimensional comparisons using essentially any group of comparable firms as defined by the user:

  • Earnings quality and cash flow
  • Earning quality and price-earnings ratios
  • Earnings quality and market-book ratios
  • Earnings quality and credit risk
  • Earnings quality and past stock returns
  • Earnings quality and measures of audit quality
  • Earnings quality and corporate governance mechanisms

Some audit quality modelling evidence

One audit quality concept is the extent to which audit fees charged meet expectations about normal behaviour. The initial modelling of audit fees dates back to the pioneering work of Simunic (1980) in the USA and by Francis (1984), Francis and Stokes (1986) in Australia through to more recently Craswell, Francis and Taylor (1995), Ferguson and Stokes (2002) and Ferguson, Francis and Stokes (2003). [14]

Audit fee models estimate the dependence of audit fees (expressed as a function of both quantity of audit work and price) upon an audit client’s characteristics and assumptions about the nature of competition taking place in the audit services market that allows for the competence and independence delivered in the audit. The models assume that avoidance of third party liability motivates the design of audit client control systems in which external auditing is part of the design solution. The variables control for the quantity of auditing supplied in three categories namely, audit client size, complexity and risk and the effects of competition in the market for audit services captured by using a variable to indicate the type of auditor (e.g., Big 4). The principal differentiating characteristic of the audit service is likely to be the identity of the auditor. Big 4 auditors (previously Big 5, Big 6 or Big 8) enjoy visibility and brand-name recognition among buyers and they invest heavily in establishing infrastructure to deliver a quality differentiated service (greater audit competence and independence) for which they expect a normal rate of return from clients demanding their services. This results in higher prices for the hours delivered on the audit work compared to non-Big 4. A further differentiating feature of the market is that those auditors (typically the Big 4) who are recognised as industry leaders are assumed to have invested heavily in establishing the infrastructure to delivered greater audit competence as industry specialists. They seek a normal rate of return on that investment from clients valuing this greater competence leading to higher prices for the hours delivered on the audit work compared to non-specialists.

The basic empirical relationships between audit fees and audit client size, complexity and risk and auditor identity have been documented in several countries since 1980, including the US, Canada and Australia. The audit fee model generally accepted as offering the best performance in the Australian audit market in recent years has achieved good explanatory power with adjusted R2 s of 67 percent and higher. This implies that the client characteristics and audit quality factors explain in combination at least 67 percent of the variation in audit fees. The research has shown that the large Big 8/6/5/4 audit firms command premiums of around 30% on top of what non-Big 8/6/5/4 audit firms charge and that industry specialists pick up a further premium of approximately 20% on top of what the non-specialists charge.

As part of its corporate governance technology solutions, the Capital Markets CRC has constructed audit quality surveillance applications, one of which uses the two concepts identified above - the extent to which audit fees charged meet expectations about normal behaviour (using the fee estimation models), and the extent to which fees for non-audit services are derived relative to audit fees (www.cmcrc.com/Products/auditx.html). Departures of audit fees from that predicted by the modelling combined with larger ratios of fees from non-audit services to fees from audit services, implies something unusual about the audit in that year (after allowing for the audit quality demanded in the audit). This, in combination with increased economic significance of other services income that could lower the independence (perceived or in fact) and therefore the quality of the auditing being delivered, provides a filter to identify cases that could help protect users from falling victim to questionable audits. Figure 4 shows a visualization from the Capital Markets CRC audit surveillance prototype depicting a sample of companies in the Sydney audit market in this two-dimensional space.

The technology solution being offered in this application allows a company’s position in this two dimensional space to be further (visually) assessed relative to a number of benchmarks such as:

  • Other companies in the same industry
  • A company’s immediate competitors
  • All companies in the market
  • Other companies audited by the same audit firm
  • Other companies audited by the same city office
  • Other companies audited by the same audit partner

Benchmarking behaviour, although statistically testable, is a very “visual” concept which lends itself to automating the generation of visualizations in conjunction with reports for each company as an output from the Capital Markets CRC technology.

Conclusions

This article has outlined the heightened interest being generated in accounting and audit quality and the demand for technology based surveillance solutions that facilitate benchmarking of company accounting and audit quality. We describe one approach being taken to develop technology solutions in this space that is based upon scientific research into accounting and audit quality. We have overviewed some of the key research being used to model accounting and audit quality and explained how these models can be used in benchmarking as part of the technology solutions.

Further research is focused on understanding the circumstances where companies and their auditors have incentives to lower accounting and audit quality below what would be expected as normal behaviour. In addition, work is focused on developing better models of normal behaviour for accounting quality and audit phenomena. Scientific based enhancements to the modeling increase the veracity of the benchmarking applied in the technology solutions and ultimately leads to improved compliance and control, improved assessments of accounts and audit provider behaviour, reduction of company ‘failure’ rates and improved investment risk management.

The authors can be contacted at www.cmcrc.com or by post to Capital Markets CRC Ltd, PO Box 970, NSW 2001 Australia. Ph + 61 2 9233 7999. Email: Carsten@cmcrc.com, Donald@cmcrc.com,  or S.Taylor@unsw.edu.au.

Figure 1
Figure 1: Distribution of OPAT scaled by TASS.
no. Observations = 7298

Figure 2
Figure 1: Distribution of change in OPAT scaled by TASS.
no. Observations = 5950

Figure 3: Returns for a 48 month period around the earnings announcement date (0) for 5 quintile portfolios of Australian companies based on accruals deflated by total assets (Q1 – smallest, Q5-largest)

Figure 4
Figure 4: Distribution of listed companies in the Sydney audit market on two dimensions- non-audit services fees to audit fees and actual to expected audit fees

[1]   Carsten Friedrich and Donald Stokes are Post-doctoral Fellow and Professor of Accounting respectively at the University of Technology, Sydney and Stephen Taylor is Professor of Accounting at the University of New South Wales. Donald Stokes and Stephen Taylor are the Chief Scientists undertaking research into corporate governance surveillance for the Capital Markets CRC and Carsten Friedrich is the Data Visualisation Expert leading this surveillance technology development for the Capital Markets CRC.

[2] Hamilton, J., and D. Stokes, 2003, “Contracting for audits – the role of markets, client firms, audit firms, audit partners, the profession and regulation”, Academy of Social Sciences in Australia Ethics and Auditing Workshop, Australian National University, December 2003.

[3] Culvenor, J., D. Stokes and S. Taylor. 2002. A Review of the Proposals for Reform of Independence of Australian Company Auditors. Australian Accounting Review, Vol.12, No. 2, Issue 27, pp. 12-23.

[4] I. Ramsay (2001) Independence of Australian Company Auditors – Review of Current Australian Requirements and Proposals for Reform, Commonwealth of Australia.

[5] HIH Royal Commission (2003) The Failure of HIH Insurance, Commonwealth of Australia.

[6] CLERP 9 (2002) Corporate Disclosure – Strengthening the Financial Reporting Framework, Commonwealth of Australia.

[7] The Australian reform process is summarized in CLERP (Audit Reform & Corporate Disclosure) Bill, (2003) Commentary on the Draft Provisions, Commonwealth of Australia, p.1.

[8] Overseas developments in auditor independence regulations prior to the release of Ramsay (2001) include the International Federation of Accountants’ (IFAC) proposals for updating its ethical requirements on audit independence; the European Commission’s release of a consultative paper containing proposals designed to achieve greater uniformity among member states of the European Commission; and the United States Securities and Exchange Commission’s (SEC) rules on auditor independence which were released in November 2000.  

[9] Coulton, J., S. J. Taylor and S. L. Taylor, 2004, Do benchmark beaters manage their earnings: Australian evidence, Working paper, University of New South Wales and Capital Markets CRC. 

[10] Data has been collected for OPAT, Net Extraordinary Items, OPBT, OPBAT (op. profit before abnormals and taxes) so a number of alternative accruals based performance measures may be considered.

[11] Dechow, P. M., S. Richardson and I. Tuna. 2003. Earnings Management and Costs to Investors from Firms Meeting or Slightly Exceeding Benchmarks. Working Paper. University of Michigan Business School.

[12] Sloan, R., 1996, Do stock prices fully reflect information in accruals and cash flows about future earnings, The Accounting Review 71, 289-315.

[13] Bird, R., S. Taylor and S. Thosar, 2004, Earnings quality and stock prices: Australian evidence, Working paper, University of Technology, Sydney and Capital Markets CRC.

[14] Simunic, D. 1980. The pricing of audit services: Theory and evidence. Journal of Accounting Research 18 (1): 161-190. Francis, J. 1984. The effect of audit firm size on audit prices: A study of the Australian market. Journal of Accounting and Economics 6 (2): 133-151.Francis, J., and D. Stokes. 1986. Audit prices, product differentiation, and scale economies: Further evidence from the Australian audit market. Journal of Accounting Research 24 (2): 383-393. Craswell, A., J. Francis, and S. Taylor. 1995. Auditor brand name reputations and industry specializations. Journal of Accounting and Economics 20 (3): 297-322. Ferguson, A., and D. Stokes (2002). Brand name audit pricing, industry specialisation and leadership premiums post big 8 and Big 6 mergers. Contemporary Accounting Research  19: 77-100. Ferguson, A., J. Francis and D. Stokes (2003), ‘The Effects of Firm-Wide and Office-Level Industry Expertise on Audit Pricing, The Accounting Review,  April, 78, No. 2, 429-448.