A recent interview by the Insights by Stanford Business, titled “In Financial Disclosures, Not All Information Is Equal” (all references are supplied below and all emphasis in quotations is mine) touched upon a pivotal issue of quality of information available from public disclosures by listed companies - the very heart of the market fundamentals.

The interview is with Stanford professor of accounting Iván Marinovic, who states, in the words of the interviewer, that “financial statements are becoming less and less relevant compared to other sources of information, such as analysts and news outlets. ...there is a creeping trend in financial disclosures away from the reliance on verifiable assets and toward more intangible elements of a business’s operations.”

In simple terms, financial information is being gamed. It is being gamed by increasing concentration in disclosures on ‘soft’ information (information that cannot be verified) at the expense of hard information disclosures (information that can be verified). More parodoxically, increasing gaming of information is a result, in part, of increasing requirements to disclose hard information! Boom!


Let's elaborate.

In a recently published (see references below) paper, Marinovic and his co-author, Jeremy Bertomeu define ‘hard’ and ‘soft’ information slightly differently. “The coexistence of hard and soft information is a fundamental characteristic of the measurement process. A disclosure can be soft, in the form of a measure that “can easily be pushed in one direction or another”, or hard, having been subjected to a verification after which “it is difficult to disagree”."

For example, firms asset classes can range "from tangible assets to traded securities which are subject to a formal verification procedure. Forward-looking assets are more difficult to objectively verify and are typically regarded as being soft. For example, the value of many intangibles (e.g., goodwill, patents, and brands) may require unverifiable estimates of future risks.”

The problem is that ‘soft’ information is becoming the focus of corporate reporting because of coincident increase in hard information reporting. And worse, unmentioned in the article, that ‘soft’ information is now also a matter of corporate taxation systems (e.g. Ireland’s ‘Knowledge Development Box’ tax scheme). In other words, gamable metrics are now throughly polluting markets information flows, taxation mechanisms and policy making environment.

Per interview, there is a “tradeoff between reliability and the relevance of the information” that represents “a big dilemma among standard setters, who I think are feeling pressure to change the accounting system in a way that provides more information.”

Which, everyone thinks, is a good thing. But it may be exactly the opposite.

“One of the main results — and it’s a very intuitive one — shows that when markets don’t trust firms, we will tend to see a shift toward financial statements becoming harder and harder. [and] …a firm that proportionally provides more hard information is more likely to manipulate whatever soft information it does provide. In other words, you should be more wary about the soft information of a firm that is providing a lot of hard information.”

Again, best to look at the actual paper to gain better insight into what Marinovic is saying here.

Quoting from the paper: “...a manager who is more likely to misreport is more willing to verify and release hard information, even though issuing hard information reduces her ability to manipulate. To explain this key property of our model, we reiterate that not all information can be made hard. Hence, what managers lose in terms of discretion to over-report the verifiable information, they can gain in credibility for the remaining soft disclosure. Untruthful managers will tend to issue higher soft reports, naturally facing stronger market skepticism. We demonstrate that untruthful managers are always more willing to issue hard information, relative to truthful managers."

Key insight: "...situations in which managers release more hard information are also more likely to feature aggressive soft reports and have a greater likelihood of issuing overstatements.”

As the result, as noted in the interview, “…we should expect huge frauds, huge overstatements precisely in settings or markets where there is a lot of credibility. The markets believe the information because they perceive the environment as credible, which encourages more aggressive manipulations from dishonest managers who know they are trusted. In other words, there is a relationship between the frequency and magnitude of frauds, where a lower frequency should lead to a larger magnitude.”

In other words, when markets are complacent about information disclosed and/or markets have greater trust in the disclosures mandates (high regulation barrier), information can be of lower quality and/or risk of large fraud cases rises. While this is intuitive, the end game here is not as clear cut: heavily regulating information flows might be not necessarily a productive response because markets trust has a significant positive value.

Let’s dip into the original paper once again, for more exposition on this paradox: “We consider the consequence of reducing the amount of discretion in the reporting of any verifiable information. The mandatory disclosure of hard information has the unintended consequence of reducing information about the soft, unverifiable components of firm value. In other words, there is a trade-off between the quality of hard versus soft information. Regulation cannot increase the social provision of one without reducing the other.”

Now, take European banks (U.S. banks face much of the same). Under the unified supervision by the ECB within the European Banking Union framework, banks are required to report increasingly more and more hard information. In Bertomeu-Marinovic model this can result in reduced incidence of smaller fraud cases and increased frequency and magnitude of large fraud cases. Which will compound the systemic risks within the financial sector (small frauds are non-systemic; large ones are). The very disclosure requirement mechanism designed to reduce large fraud cases can mis-fire by producing more systemic cases.

In its core, Jeremy Bertomeu and Ivan Marinovic paper shows that “certain soft disclosures may contain as much information as hard disclosures, and we establish that: (a) exclusive reliance on soft disclosures tends to convey bad news, (b) credibility is greater when unfavorable information is reported and (c) misreporting is more likely when soft information is issued jointly with hard information. We also show that a soft report that is seemingly unbiased in expectation need not indicate truthful reporting.”

So here is a kicker: “We demonstrate that …the aggregation of hard with soft information will turn all information soft.” In other words, soft information tends to fully cancel out hard information, when both types of information are present in the same report.

Now, give this a thought: many sectors today (think ICT et al) are full of soft information reporting and soft metrics targeting. Which, in Bertomeu-Marinovic model renders all information, including hard corporate finance metrics, reported by these sectors effectively soft (non-verifiable). This, in turn, puts into question all pricing frameworks based on corporate finance information whenever they apply to these sectors and companies.



References for the above are:

The Interview with Marinovic can be read here: https://www.gsb.stanford.edu/insights/financial-disclosures-not-all-information-equal.

Peer reviewed publication (gated version) of the paper is here: http://www.gsb.stanford.edu/faculty-research/publications/theory-hard-soft-information

Open source publication is here: Bertomeu, Jeremy and Marinovic, Ivan, A Theory of Hard and Soft Information (March 16, 2015). Accounting Review, Forthcoming; Rock Center for Corporate Governance at Stanford University Working Paper No. 194: http://ssrn.com/abstract=2497613.