This is a rather long post, for which I apologize in advance. It is an adaptation of a much more diplomatic comment letter that I wrote to the PCAOB, criticizing its proposed, and ultimately adopted, definitions of 'material weakness' and 'significant deficiency' in AS 5. As you will soon see, my critique addresses a larger problem having to do with how poorly uncertainty (for example, contingent liabilities) is dealt with in key financial reporting standards.
Let's start with a common example outside of the realm of financial reporting. We all pretty much know that the threshold burden of proof to find someone guilty of committing a crime is 'beyond a reasonable doubt.' What does that mean? Methinks that individual jurors have their own personal definitions for 'reasonable doubt.' However, to express uncertainty as a quantitative probability might create more problems for the jury-serving public than it would solve.
It has long been a puzzlement to me that the SEC, FASB and PCAOB take the same vaguely-worded approach as the criminal law--as if accountants and managers were innumerate. Yet, all business schools and accounting programs have required courses in statistics and the use of quantitative probabilities in decision making.
Here are some of the major examples of weasel words standing in for probabilities in financial reporting standards:
- MD&A requirements for forward looking information concerning uncertainties that could affect future profitability, liquidity or capital resources is based on a probability threshold vaguely specified as 'reasonably likely'.
- FAS 5 on the accrual of contingent liabilities can require an accrual when the likelihood of occurrence is 'probable.' Disclosure only is required if the contingency is less than probable, but more than 'reasonably possible'.
- AS 2 (now replaced by AS 5) required an auditor to conclude that a material weakness existed if the probability of a material misstatement was 'more than remote.'
- AS 5 replaces 'more than remote' with 'reasonably likely' in its definitions of 'material weakness' and 'significant deficiency'
Incidentally, the PCAOB stated in its proposing release for AS 5 (page 9) that by changing terminology, it would not be changing meaning, but merely providing clarification. This is supposedly because (1) auditors were familiar with how to apply the weasel words of FAS 5, as they have been around since 1975; and (2) 'more than remote' is not a term used by FAS 5. The first reason is disingenuous, because everybody is aware that auditors have been ineffective in curbing abuses of FAS 5 by their clients. The second reason is simply false (see FAS 5, paragraph 3b). In FAS 5, 'more than remote' is indeed a lower threshold than 'reasonably possible.' Evidently, weasel words can beget weasel logic.
I now want to explain why weasel words to express uncertainty inevitably result in low quality standards of financial reporting. I'll be discussing FAS 5 and AS 5, but as with my comment letter, my focus will be on AS 5.
Analysis
The points A, B and C in the diagram below denote unspecified probabilities that must, of necessity, demarcate the ranges of uncertainty used to apply SFAS 5, AS 2 and AS 5:
Before proceeding further, it is important to note that Points A, B and C do not change. In other words, the points are unaffected by the facts and circumstances of a particular transaction or internal control; by way of confirmation, no publication of the PCAOB or FASB that I am aware of provides any indication that either of these august bodies believe that the points should vary across audit engagements or particular events.
Since the FASB is the fount of the weasel words that are the subject of this quixotic diatribe, it is also important to note that the FASB did not disclose any information concerning the process by which “probable” and other qualitative terms for describing uncertainty were selected in the Basis for Conclusions section of SFAS 5; or whether quantitative probabilities were even considered. Arguably, many of the well-known problems in application of SFAS 5 have resulted from the absence of explicit points of demarcation—particularly Point C in the above diagram. The ambiguity and inevitable disagreement between auditors, preparers and users as to the appropriate demarcation Points B and C has had two effects: (1) substantial lack of comparability of financial statements, and (2) windfalls to auditors and preparers by allowing them to avoid being held to account for misleading financial statements.
But, regardless of the FASB’s motives for promulgating SFAS 5 as it did, it was neither in the public interest, nor is it consistent with the PCAOB’s mission, to continue to follow their conveniently vague approach to dealing with uncertainty set forth in SFAS 5. In particular, investor protection is less than it could be due to the ambiguation through weasel words of the point between “more than remote” and “reasonably possible” in AS 5 (i.e, Point B in the above diagram). Blurring terminology adds judgment and cost to financial reporting while providing no discernible purpose that is consistent with the mission of the PCAOB.
Some may argue that quantitative probability thresholds constitute bright-line rules that are contrary to the notion of principles-based standard setting. This is wrong, because weaselly wording runs contrary to normative economic principles on the application of judgment in decision making. These overarching principles require that subjective probabilities be quantified. These principles have been widely applied for generations, taught in all accredited schools of business and accounting (Managerial Economics 101), and incorporated into more recent accounting standards.
Note on recent accounting standards: SFAS 144 on impairment of long-lived assets recognizes that probability-weighted cash flows may be used to test the recoverability of long-lived assets (¶17). SFAS 109 on income taxes specifies a probability threshold of 0.5 when measuring the deferred tax asset valuation allowance (¶17). Perhaps most germane is the auditing literature, wherein it is stated in AU Section 350 on sampling, “…the auditor should determine an acceptable audit risk and subjectively quantify [emphasis supplied] his or her judgment of the risk of material misstatement.” (¶20).
Now, let's set FAS 5 aside and speak only of auditing. The auditor's assessment of risk is inherently quantitative and structured, even though an assessment of materiality may be more judgmental and dependent on facts and circumstances. Along these lines, the Board’s contention that “evaluation of whether a control deficiency presents a reasonable possibility of misstatement can be made without [emphasis supplied] quantifying the probability of occurrence as a specific percentage or range” [¶73 of proposed AS 5] runs counter to norms of rational decision making. Here is an illustrative example of what I mean:
Assume that Point B in the earlier diagram represents the probability 0.4. In the terms of proposed AS 5, this is the lower bound of “reasonably possible.” Further assume that the auditor determines the materiality threshold for a misstatement of revenues to be $1,000,000. Therefore, $400,000 (0.4 x $1,000,000) represents the maximum allowable expected misstatement (given that a misstatement is at least reasonably possible) such that an internal control weakness would not be disclosed as material.
The PCAOB is simply wrong to expect an auditor to obtain reasonable assurance for its opinion within the framework of AS 5 without undertaking a process substantially similar to the one described by the above example. Stated another way, as AS 2 was written, and as proposed AS 5 is currently written, it should be unacceptable for auditors to adopt different threshold probabilities for different clients, or even for different financial statement amounts (although materiality thresholds may reflect case-specific factors). The unavoidable conclusion from the PCAOB’s language in these auditing standards is that it should not be necessary, or required, for each auditor and client to come to separate conclusions on each engagement, and negotiate the threshold probability for “reasonably possible.” Yet, the weaselly specification of Point B is an invitation for such costly and counterproductive negotiations to occur.
Note: To illustrate another problem of logic in AS 5 (some might call this a loophole), consider the following extension of my numerical example: if a particular control over revenues had a probability of misstatement of 0.39, the control would never be reportable as a material weakness even if the resulting misstatement would be significantly greater than $1,000,000. Thus, thresholds per se in proposed AS 5 lack foundation in principle.
Summary and Looking Ahead
For both FAS 5 and AS 5, probability thresholds can be, and therefore should be, explicitly quantified. For example, a change from qualitative terminology (i.e., “more than remote” in AS 2, or “reasonably possible” in proposed AS 5) would simplify auditing standards, increase reliability of ICFR audits, and reduce audit and compliance costs. Such a change would better protect the interests of investors and further the public interest through greater clarity and transparency of auditing and financial reporting. I know of no reason consistent with investor protection for intentionally blurring the lines with ambiguous language when precise thresholds are feasible.
Looking to the near future, the FASB may soon require recognition at fair value of contingent liabilities assumed in a business combination. But, how will the FASB specificy which contingent liabilities would be recognized--'probable' contingent liabilities, or some other weaselly-worded standard for recognition? Will a bad accounting rule beget another bad accounting rule? Indeed, the FASB has long known that FAS 5 is not working (nearly universal understatement of environmental liabilities being just one example), and their stated intention is to consider contingent liabilities as part of the project to revise the conceptual framework. But, don't hold your breath; if improved recognition criteria for contingent liabilities should ever come to pass, it may not be until the year in which both the Cubs win the World Series and the Rangers win the Stanley Cup.
Recent Comments