One direction: which way to effective climate stress tests?

Existing estimates of climate risk impacts rely on 'top-down' or 'bottom-up' assessments. Could there be another way?

Want Climate Risk Review in your inbox every Monday and Friday? Then please consider subscribing:

Traditional stress tests ask financial institutions how they’d fare if subjected to a prolonged economic downturn. The macroeconomic shock is the question; the firm’s post-stress balance sheet the answer. 

Usually, the better the question, the more useful the answer. But what if the question has never been asked before? And what if the question itself cannot be defined?

Climate change represents just such an unanswerable question. 

How do you ask what a bank’s balance sheet would look like if 19% of the world became an “unlivable hot zone”, or if oil was displaced as its chief fuel source? What would happen to a corporate loan book on the sudden imposition of a $100 per ton carbon tax?

There are no precedents that banks can reach back to in order to model the answers to such questions, and no way for regulators to judge the macroeconomic impacts of such events. 

This doesn’t mean they’re not trying. 

Top-down, bottom-up, or something else?

Global standard-setter the Financial Stability Board (FSB) took stock of regulators’ efforts to quantify climate risks in a new report. It identified two distinct approaches: ‘top-down’ and ‘bottom-up’. 

Each has its pros and cons. The top-down approach sees a supervisor estimate the magnitude of the climate shock itself, and then ask its supervisees to map the effects to their assets, liabilities and capital. This has the benefit of putting the analysis’ parameters in the hands of the watchdogs themselves — meaning they get to define the terms of the assessment. It’s also less resource-intensive for the financial institutions subjected to them.

However, the FSB notes that the top-down approach is inherently broad-brush, as a watchdog has to use a generic dataset to estimate a climate shock that can cover a broad assortment of supervisees.  

There are other downsides, too. If the regulator prescribes the shock, then the outcome for each tested institution relies exclusively on its own judgement. The concept of ‘garbage in, garbage out’ applies here: if a regulator applies a shock that ill-fits the business model of the firms under its watch, the results of the test will be of limited use.

In addition, firms may be less convinced by the findings of a top-down assessment than one they define themselves — and therefore less likely to fully take on board the insights gleaned from the analysis.

The bottom-up approach is the mirror-image of the top-down. Here, the impacts of climate risks on firm exposures are estimated by the institutions themselves, rather than prescribed by the regulator. In some cases, a watchdog will hand down the rough outlines of one or more climate scenarios, and ask institutions to fill out the details. In others, the firms are asked to come up with “institution-specific” simulations themselves.

The merits of a bottom-up approach, in the FSB’s eyes, is its ability to better capture firm-specific dynamics. Because climate-related effects and, where applicable, the scenarios themselves, are tailored to each institution’s business model, the results provide a more comprehensive picture of each entity’s vulnerabilities.

However, the democratised nature of the bottom-up approach may also be a weakness. By putting firms in the driving seat, they have freedom to “game” the analysis, for example by underestimating the asset-level impacts of climate shocks, or defining the shocks in such a way as to avoid undesirable outcomes.

A bottom-up approach also advantages big, sophisticated firms over smaller, simpler ones. Those with access to cutting edge modelling technology, and with cash to spend on granular climate data, should be able to produce more detailed stress test outputs than their less flush counterparts. This means regulators get an unbalanced assessment of the system-wide impacts of climate change. 

On the other hand, by placing the onus of producing climate stress assessments on firms themselves, a bottom-up approach should drive innovation in modelling and data gathering at the institutional level.

Throw it in reverse

Neither a top-down not bottom-up approach will produce “perfect” climate assessments. By their very nature, scenario analyses can only ever provide a blurry picture of a firm’s strengths and weaknesses.

Right now, regulatory authorities are muddling through, groping for best practices. It’s right that they, and the firms they supervise, experiment with different ways of quantifying climate-related risks.

There’s help available, too. The Network for Greening the Financial System (NGFS), a club of central banks committed to fighting climate change, issued a series of climate change scenarios, alongside a guide for supervisors, just last month. Representative and alternative scenarios included in this package could help facilitate both top-down and bottom-up assessments.

But beyond these, there’s another approach that could yield illuminating insights, one that wasn’t in the FSB’s stocktake: the “reverse stress test”.

This sort of assessment asks an institution to imagine scenarios that would break its business model and cause the market to lose confidence in its ability to function. These are, by their nature, extreme, but can be useful for identifying vulnerabilities that may otherwise be overlooked. 

There’s no reason why firms shouldn’t be asked to cook up catastrophic climate change scenarios as part of a reverse stress testing process.

For example, a coastal mortgage bank could ask itself: what climate-related event would cause my portfolio to implode? The answer could be a rapid uptick in severe weather events, subjecting seaside towns to floods and storm damage that cause homes to become unsellable. Mortgage borrowers could simply walk away once their properties become uninhabitable, defaulting on their loans and saddling banks with impaired collateral. 

Transition-related climate risks could also be captured in reverse stress tests. Right now, troubled shale oil producers are forcing banks to set billions of dollars aside to cover expected losses. For example, of Wells Fargo’s $12.6 billion of oil and gas loans, 11% were delinquent as of end-June.

A reverse stress test could ask the bank: what climate-related event would cause 50% of this portfolio to fall into arrears? One answer may be the imposition of a high carbon tax, or a permanent collapse in oil prices caused by its replacement as an auto fuel by hydrogen or electricity.

Compelling financial institutions to imagine far-out climate scenarios that would put them in their graves may sound unreasonable. After all, firms are supposed to use the outcomes of climate assessments to take action in the here and now to mitigate their exposures. If they don’t believe reverse stress test scenarios are credible, why would they use them to inform their climate-related decisions?

Then again, policymakers openly admit that existing stress-testing models and methodologies are poorly suited to capturing climate-related economic impacts effectively. The Bank for International Settlements wrote that an “epistemological break” is needed in risk management to properly tackle the looming climate crisis — one which embraces uncertainty and fat-tailed probability distributions.

Certainly, reverse stress tests suffer from similar shortcomings as the bottom-up approach. Firms need advanced modelling prowess to construct and implement climate doomsday scenarios, and regulators may find it hard — if not impossible — to compare assessments across entities.

However, it would be tough for banks to “game” such analysis, since if the shock fails to break the firm, then it can’t be considered an effective stress.

Perhaps, then, it’s time regulators stopped thinking just in terms of top-down and bottom-up, and threw risk management into reverse for a change.

Thanks for reading! Why not share this post with your colleagues?


Please send questions, feedback and more to

You can catch climate risk management updates daily on LinkedIn

The views and opinions expressed in this article are those of the author alone

All images under free media license through Canva