The EU’s insurance regulator suggests climate capital charges that even it’s not fully sold on

This is a free-to-view edition of Unpacking Climate Risk. Become a paying subscriber to get new posts sent to your inbox each week 👇

Not quite ready to make the leap but liked the article? Show your support by buying us a coffee👇

The debate on whether financial institutions should hold additional capital against climate risks has been raging since I first started on this beat some five years ago.

A new report from a top EU regulator suggests they’re a step closer to becoming a reality – at least on first glance.

Last month, the European Insurance and Occupational Pensions Authority (EIOPA) published its final recommendations on the prudential treatment of sustainability risks in Solvency II, the EU-wide capital and risk management framework for insurers. These included a proposal to introduce “additional capital requirements for fossil fuel assets” held by insurers to “accurately reflect the high risks of these assets.”

On the surface, it’s a big move – and one a long time in the making. EIOPA kicked off the discussion on sustainability risks way back in 2022. Two separate consultation windows for industry and other stakeholders were held in 2022 and 2023. Eighty-four bodies submitted feedback, including insurers, advocacy groups, and private individuals. You can’t say they didn’t follow a measured and deliberative process.

And yet and yet and yet….EIOPA’s own documentation suggests introducing the charges would do little to alter insurers’ asset allocations. Moreover, the regulator’s justification for introducing the charges rests on shaky foundations, and highlights an ongoing tension in climate-related financial risk regulation between trusting in the past, and betting on the future.

Let’s get into the substance.  EIOPA calls for insurers to hold extra capital against equities and bonds issued by fossil fuel companies to compensate for the alleged heightened market risk of these positions. 

For equities, it proposes three options.  The first is no additional charge.  The second, a reclassification of fossil fuel equities as “Type 2” equities, putting them on a par with developing country stocks and global equity funds. This would increase their capital requirements from 39% (the charge for “Type 1” equities) to 49%. The third option would see a dedicated 17% add-on applied to these holdings. So, an Exxon stock would attract a 56% charge (39%+17%=56%). Note that this translates to a real term 44 percentage point increase in the current capital charge – quite the hike!

EIOPA also presents three policy options for bond holdings. Again, the first is no additional charge. The second, a “downgrading” of fossil bonds for the purposes of calculating the relevant capital charge – meaning AAA bonds would be treated as AA bonds, A- bonds as BBB bonds, and so on. The third would see a multiplication factor applied to each “risk bucket” used to calculate a fossil bond’s charge. In plain terms, this would lift the charge for in-scope bonds up 40% from the baseline – an increase commensurate with that proposed for equities.

Let’s be clear – the third option for bonds and stocks, respectively, would be material from a regulatory capital standpoint.  After all, insurers would find it hard to justify tying up so much more capital for investments that are unlikely to generate returns that beat the market as a whole. Over the last five years, iShares chief oil and gas ETF has appreciated 76%, whereas the iShares Core S&P 500 ETF has increased 90%. Presumably, fossil stocks and bonds would have to guarantee returns at least 40% greater than their non-fossil peers to make the cut under a pure capital allocation lens.

This being unlikely, insurers would be incentivized to divest their fossil holding and reallocate to “cheaper” investments from a capital standpoint. This is the outcome sought by climate activists and policymakers who want to redirect capital away from climate-harming activities to climate-friendly ones.  Personally, I’m all for it.

Still, I can’t pretend that the application of even the most punitive options on the table would be anything more than symbolic.  EIOPA itself found European insurers’ direct exposure to fossil stocks to be “very low” on average, meaning even the +17%  charge would have a minimal effect on undertakings’ solvency ratios.  Estimated impacts from the change in fossil bonds’ treatment was also found to be slight.  

On the plus side, this suggests that introducing the charges would not upset financial stability, because any divestments would be small enough to be absorbed by the market.  In the minus column, if the change fails to signal to insurers – and in turn the wider market – that fossil investments are persona non grata, then what’s the point?

It doesn’t help that bank regulators have shied away from climate capital add-ons of their own.  While EU lenders do have to consider ESG risks in their internal risk models, there are no special risk-weights for oil, gas, and coal assets under the revised Capital Requirements Regulation.  The impact of EIOPA’s proposal would, therefore, be isolated to insurers if implemented.  And what with policymakers seemingly so against the idea of “unlevel playing fields” in financial regulation, it’s easy to see European officials tabling the proposal on grounds that the insurance industry shouldn’t be unfairly singled-out.

Now onto the really interesting stuff – EIOPA’s internal wrangling on whether their own capital proposals make sense. In a revealing cover letter accompanying the report, EIOPA cites “the range of views” within its Board of Supervisors on approving the recommendation (my emphasis):

While some Members expressed full support for the proposed measures, others expressed concerns about the methodology and data set used in the analysis. Other Members supported the analysis but raised concerns in relation to wider considerations which are deemed relevant for the decision.

To me, it seems kind of remarkable that a supervisory body would essentially undercut itself in so public a way – especially as it’s recommending significant changes to a capital framework that shapes one of the biggest and most important industries in Europe.

Going back to the report, it’s clear where the methodology and data concerns arise.

EIOPA says it conducted a backward- and forward-looking risk assessment “to conclude whether there is sufficient evidence for a dedicated prudential treatment of transition risks”.  Regular readers should know where Tony and I stand on forward-looking assessments by now (if you don’t, check out the articles here, here, and here).  For its part, EIOPA found there were some respondents to their feedback sessions who were more enamored with forward-looking approaches than others.  From the report:

Some respondents mentioned that the use of a model-based assessment can be subject to technical bias due to the model assumptions taken, and corresponding findings should be treated with caution regarding the conclusion on potential prudential implications.

EIOPA also recognized itself that the forward-looking approach – based on scenarios and “what-if” conjecture – is at odds with the rest of the insurance regulatory framework, which relies on backward-looking value-at-risk metrics to set capital charges. The following paragraph shows how the report’s authors had to tie themselves in knots to justify using a forward-looking approach in their analysis at all:

Although the calibration of Solvency II’s capital requirements for market risks typically relies on historical data, a purely historical perspective might not be sufficiently informative about the dynamic materialization pattern of climate-related risks. This is because the realisation of pure transition risks is difficult to observe in historical data, as transition risks can be considered as events not broadly represented in historic data, and be accompanied by further market risk drivers (e.g., inflation). In this context, the challenging question arises as to whether to rely on historic asset price data to conduct an empirical risk analysis (backward-looking) or to use model-based risk assessments, typically in terms of stress scenarios (forward-looking), or a combination of both.

EIOPA opted to use both approaches to gauge potential transition risks. As you may have guessed, the backward- and forward-looking analyses produced very different outputs.

For example, the backward-looking equity risk analysis provided no “clear finding” on whether transition risk-related risk differentials exist across issuers with different transition risk profiles.  And while it found fossil fuel-related stocks were riskier over the period 2010 to 2021, certain economic stress events (COVID-19 being one) played some role here.

In contrast, the forward-looking analysis showed “material” differentials for equities linked to mining, oil refining, electricity generation, and the composite fossil fuel sector – although these varied wildly depending on the annual probability of a disorderly transition scenario occurring:

Estimates range from -2% at low annual probabilities for a disorderly scenario to as much as -50% at higher probabilities for the mining and electricity generation activities. The total 0.5% VaRs for these activities reaches 90% at higher probabilities, reflecting the fact that in some transition scenarios the mining and electricity assets will be largely stranded.

Given these results, you can understand why some on the EIOPA Board of Supervisors were skittish about the methodology – and presumably about the recommendations flowing from it, too.  After all, we’re talking here about changing capital rules that a) penalize certain assets over others, and b) put insurers on an unlevel playing field vis-a-vis other financial institutions based on their risk characteristics.

While I’m all in favor of climate-related capital charges as a policy measure – meaning as part of a broader regulatory effort to accelerate the low-carbon transition – authorities that claim such charges are necessary from a financial safety and soundness perspective have to clear a high bar when it comes to the credibility of their data and methodologies.

In this case, members of EIOPA’s own board do not seem to believe this bar was cleared.  Why, then, should the rest of us?