Negative screening and supervising environmental liabilities under IAS 37

This is an excerpt of an essay I wrote for a course on Sustainable Finance. Its goal was to advocate for more pressure on the disclosure of contingent liabilities as a way of nudging institutions into better negative screening (to reduce the risk of losses due to environmental litigation).

I know accountancy and auditing are not the most exciting subjects (and, honestly, I'm no expert either), but I believe this problem instanciates an area where decision-making (for companies and investors) could be improved (which might interest people in IIDM). And I hope it could contribute to ESG metrics, too - in fact, I decided to share this after pondering on Sanjay's posts on effective ESG (which I totally recommend). I suppose that: a) firms that are more socially responsible tend to have lower legal risks, and so would likely have lower contingent legal liabilities - which means this could work as a useful proxy (to be confronted with other corporate responsibility data), so that careful investors can better screen the most responsible ones (it's still just what Sanjay calls "inwards impact", though); b) by disclosing their best estimates of legal liabilities, firms anticipate information about future losses, so mitigating some incentives problems (like a firm profitting from harm caused in the present - which would be compensated only in the future).

Thanks to Gavin Taylor and Jekaterina Ijina, who kindly discussed with me about this subject and read excerpts of this read.

Introduction

The Amazon Rainforest is one of nine tipping points in the global climate system; yet, the whole world was shocked in the last years as the sky in São Paulo went dark because of clouds formed by huge forest fires. Despite the government promising swift action, deforestation broke new records in 2020, so increasing the country’s net carbon emissions – in contrast with the global decrease caused by the Covid-19 recession. To make things worse, the forest is usually burned to be turned into pasture for cattle – a main source of methane emissions. Notice, however, this is more a problem of supervision and law-enforcement than a lack of norms: environmental regulation places several constraints on economic activities in protected areas[1], and torts law holds polluters strictly liable for causing environmental harm.

            Since responsibility should be assigned to those well-placed to deal with a problem (as implied by principles of extended porducer responsibility, for instance), this issue has implications for financial institutions. In capitalism, financial markets are deemed as efficient institutions to aggregate and process information and to “pump” capital to where it is most needed; so these markets can better internalize uncertain future costs, share the corresponding risks, and also leverage impact by influencing other economic sectors. In addition, financial systems have a tradition of international governance to coordinate collective action, with expert assistance; this can be seen in the worldwide alleged support from financial institutions and regulators for the fight against climate change, gathered in international organizations such as the Network for Greening Financial Systems (NGFS).

In the last years, the literature has put due emphasis on the challenge of defining and managing green portfolios (to include investments that are provably environment-friendly, and exclude polluting activities - i.e., whitelisting and positive screening); however, in this text, I will focus on negative screening as, at least in Brazil, where capital markets are still underdeveloped and investment is still more reliant on debt (specially banking loans) than equity, it is still a relevant (and underdeveloped) strategy for financial institutions to fulfill their obligations related to sustainability, so deterring their customers from causing harm to an ecosystem important for the global climate. Legislation imposes this duty on them: Resolution 3545, issued by the National Monetary Council, forbids banks from lending rural credit funds to customers implicated in environmental infractions in the Amazon biome. In addition, Resolution 4327 defines corporate social responsibilities for financial institutions, including an obligation to map, in their social responsibility policies, risks (even if just reputational) of losses due to environmental harm. Finally, banks might incur in legal sanctions for failing to comply with those standards and for funding blacklisted activities and enterprises[2].

In that case, the corresponding obligations due to fines or compensations should be (when uncertain) regarded as environmental and legal contingent liabilities, according to IAS 37 – as recommended by the Task-Force on Climate Disclosures (TCFD) in 2017; and, as we shall argue, their disclosure is part of the incentive schemes to deter investments on harmful activities. With this, we are joining the literature emphasizing the need to further integrate environmental reporting with IFRS/IAS, and pointing out routes that managers, auditor and supervisors could further explore in the near future.

Example: Banco Santander and its environmental liabilities

            In 2016, IBAMA (a Brazilian environmental agency) fined the Banco Santander in R$ 47.5 million for funding the production of seeds in areas under environmental embargo (due to deforestation). The story was replicated by news outlets, and it is duly disclosed by the firm in its Reference Form (trans. Formulário de Referência) – a document with extensive information for investors, defined by the Normative Instruction 480, issued by CVM (Brazilian Security and Exchanges Commission). According to the bank, the administrative litigation is still going on - so the fine is not yet due, although its payment is already considered probable.

Santander comprises itself as the “most responsible bank” of the world, according to the 2019 Dow Jones Sustainability Index. The sustainability report of the holding Santander Group (Banco Santander, S.A.) details ESG actions and projects (descriptive qualitative information), as does the financial statements of the Brazilian operation; nevertheless, unlike its Reference Form, none of these documents mention any legal suit related to environmental harm. This confirms the results of previous research that had failed to find disclosures of environmental liabilities in statements from financial firms listed in Brazil’s stock exchanges – even among companies listed in sustainability indexes (ISE). This raises suspicion over the CSR field: unless firms provide adequate disclosure of their environmental risks and liabilities, we might be justifiably concerned that social responsibility (and business ethics, in general) might end up reduced to ethics washing – that companies use CSR initiatives exclusively as marketing campaigns[3]. Adopting full disclosure of environment-related activities and liabilities would be a first step to counter this problem.

IAS 37 – where are environmental contingent liabilities?

IAS 37 defines contingent liability as an obligation that, though deriving from a past event, will only be confirmed after future events, which are not in the control of the entity; the paradigmatic case is that of a legal suit. Because of this contingency, it might be hard to estimate the odds of a loss; depending on such estimate, this future event might be considered:

• Probable: when it is more likely than not that the event will occur, so that the firm must recognize a provision.
• Possible: when, though not probable, this probability is not remote either, so that the firm must disclose the contingency in its balance sheet notes;
• Remote: IAS 37 never defines ‘remote’, but it is implied that this refers a substantially small probability; in this case, the entity must not make a provision, and does not have to disclose the contingency.

The first problem is that contingencies often fall under the category of Knightian Uncertainty - where we lack enough information to measure frequentist probabilities, or to build a model and estimate (reliable) subjective probabilities. The second one is the existence of disagreement among professionals about those very uncertainty concepts; not only the application of IFRS standards (and particularly of environmental disclosures) varies according to sector and firm size, but also across countries. For instance, despite the apparently mathematical clarity of the definition of probable (“more likely than not” seems to imply any odds higher than 1:1), and the implied exceptionality of the notion of remote, here are some examples of surprising disagreement our brief literature review has found:

This results in two risks for the disclosure of environmental legal contingencies[4]:

1. Concealment: given the vagueness of these concepts (with no definite criteria to allow for tests and convergence), contingent liabilities may end up hidden if they are considered of “remote possibility”; or
2. Dilution: if environmental liabilities are lower than the value corresponding to other legal suits or fines, they are not deemed as “relevant” or “material,” and so are aggregated to other liabilities and end up not being explicitly mentioned in financial statement notes – even if their payment is recognized as possible or probable.

Unlike other sectors more involved in environmental litigation (such as oil, infrastructure, or public utilities companies), in the financial sector, other torts and administrative sanctions (particularly tax-related litigation) are (when aggregated) more frequent and costly than environmental cases (such as suits filed by customers, tax-collectors and workers). And, of course, if, in a competitive context, a lack of transparency is regarded as normal, a firm has no incentives to unilaterally disclose their environmental contingencies, as this would harm its reputation.

Supervising environmental contingencies: next steps

Currently, testing compliance with IAS 37, specially regarding environmental liabilities, requires a lot of effort: it implies gathering data on all the (i) legal suits of each firm, confronting it with their corresponding (ii) reference forms and (iii) financial statements, and then analyzing how those contingencies are classified under IAS 37. However, (i) is not an open database (the justice system does not openly share its data, only allowing single queries in their websites), (ii) is in the CVM’s system (so not directly available for BCB supervisors) and only applies to companies listed in stock exchanges, and (iii) is in non-structured text format. Because of that, unlike credit supervision, this subject must be tested on a case by case basis, without automated monitoring tools.

A change in international accounting standards would be welcome. First, environmental legal liabilities should be regarded as a special class, and so disclosed separately from others - so blocking the problem of dilution. Second, uncertainty concepts should be defined more precisely, by, e.g., referring explicit probabilities - even if it leads to made-up statistics; this wouldn't solve, but could help mitigate the problem of concealment. Finally, testing could become way more efficient in the near future: with more reliable natural language processing, we could automatize the reading of financial statements, and confront them with data on legal suits. Besides improving estimates of legal risk for companies, it could provide inputs for ESG / CSR metrics (assuming that increased litigation correlates with low compliance with ESG policies - a conjecture I'd like to see tested).

Conclusion

            In this brief essay, we argued that: (a) given the importance of the Amazon forest, preserving it is one of the main ways the country can contribute to fighting climate change; b) given the role of the financial system in the economy, it is well-placed to contribute through negative screening; c) though legislation could be improved, it already establishes responsibilities to preserve and to abstain from funding harmful activities; d) the current challenge is in effective law-enforcement, and in aligning incentives so we can reliably assume the systems is performing its role; e) other reports aside, a bank’s environment legal liabilities would currently fall under IAS 37, and should be disclosed in financial statements, or even justify a provision, if its payment is deemed possible or probable; f) financial supervisors must develop tools to test compliance with this.

            Of course, this is not a solution to the climate change problem, and it is likely less impactful than discussions about green securities, or on using monetary policy to fund a green transition. Nevertheless, we need all the help we can get today; and this is a straightforward way to nudge the financial sector to improve its negative screening practices.

 Notes

[1] R. Schneider and Imazon researchers propose a model to understand regional economic cycles “from boom to bust” in the Amazon: first, wood cutters develop basic (often illegally) infrastructure to extract valuable trees, being followed by low productivity cattle and agriculture ranchers (attracted by cheap land), and finally by service providers. In a few years, all economic value is gone, and the economy will remain dependent on transfers from the federal government.

[2] One must notice that, in most cases where a borrower is blacklisted, the bank will claim this occurred after the loan, so there would have been no violation of environmental norms; even in those cases, though, the borrower’s credit risk should be updated to reflect their inability to seek further funding (akin to being blacklisted by a credit bureau). However, this seldom occurs, as banks usually keep apart their credit and environmental risk assessments. Both the operational risk for the lender (which we discuss in the rest of the text) and the increase in the borrower’s credit risk due to litigation would fall under what NGFS has called liability risks.

[3] It is even worse if, besides “cheaply” signaling virtue,  firms end up capturing the field, as when tobacco companies funded academic research aligned with their interests – so delaying or influencing regulation, to preserve their profits. Thus, one of the challenges in business ethics is to align incentives in this “reputational game” and encourage firms to take effective actions, instead of only displaying an appearance of ethically aligned.

[4] IAS 37 has been under revision for years, and the proposed changes deal with procedures to classify the chances of a loss in a legal suit and to provide more transparency to contingencies relevant for CSR and ESG. One must notice, though, that disclosing one’s forecast about the outcome of a legal suit may reduce the information asymmetry between the litigants and so undermine the firm’s position in the suit (especially if it is aiming for a settlement).