Weathering the Storm: Climate Risk Stress Tests

those existing from geopolitics and volatile capital markets. To counter this climate threat, some central banks have been creating and conducting climate risk stress tests in order to make sure financial institutions are capable of riding out potential impacts on businesses and investments. But what do these climate risk stress tests entail? What are the approaches that different central banks take,

and which markets are taking this climate threat seriously. Today, I'm joined by Tiff and Brandley Bloomberg, NIF's head of Transition Risk and Alignment, who draws from findings from his note Climate Risk Stress Test Review, which B and EF clients can find at B and EF go on the Bloomberg Terminal or on BNF dot com. On our website, B and EF clients can also find our Transition Risk Model TRACKED, which assesses around eighty thousand companies and is

now integrated into Bloomberg's MARS climate stress testing solution. Also on today's show, we are joined by a special guest, at Osh, the head of Climate, Nature and Regulatory Financial Solutions at Bloomberg, where he and his team provide data and analytics that help investors incorporate climate and nature related risks in investment decisions and comply with sustainable disclosure requirements.

Prior to Bloomberg, Edo was a manager of the Climate Scenarios team at the Bank of England, where he led the development of the NNGFS scenarios, and while working at the Dutch National Bank, he was part of the first ever climate stress test conducted by a central bank. Examples of his team's work be found on the terminal at WSL Physical Risk or WSL Water. All right, let's get talking about climate risk stress tests with Tiffin and Edo. Tiffin, welcome to the podcast today, please to be here. Tom

and Ado welcome on board as well. Thanks Tom, so. I know that you both come into this question of climate stress testing from quite different angles and have had different experiences in the space. So my first question was, you know, to talk about maybe the history of climate stress testing. But before we even do that, for those of us and I include myself in this that are not really familiar with this concept, can you explain what a climate risk stress test is or even just what is a stress test?

So this can be the transition risk, so the risk arising from climate policies, shifting consumer preferences, or technology disruption, and also physical risks, so this is sort of your classic natural disasters or the chronic risk rising temperatures, overall sea level rise and so on.

analyze scenarios that are severe but plausible. So it's explicitly not about you know, business as usual risk management, but really thinking to more extreme cases that could occur, such as the two thousand and eight financial crisis, and make sure we're better prepared for such events in the future.

of what we were trying to do. There was already in academic literature kind of separating out climate transition risks from climate physical risks as different risk channels to look at and some insight onto what those risk transmission channels could look like. So what we really needed to then do is think about how could we model these risk transmission channels in a scenario setup that's going to be

relevant for financial institutions. But yeah, as you can imagine with anything new, we did run into a number of challenges along the way trying to do that.

of financial institutions. Well, then you need to start thinking about what are the GHD emissions of those financial institutions, How could thoseg emissions be the gg emissions of the companies that financial institutions are invested in, And how could a carbon price impact the cost structure of these companies, and how could that ultimately affect things like the credit worthiness of the companies. That's going to have a bearing

on the financial institutions stability and safety and soundness. So you have to think backwards, and then you realize that at that time especially, there wasn't very good data on things like GHD emissions, and there wasn't really a president of how do you model the impact of a common price? How does that play too on the cost structure of that company? Could they pass it to to upstream suppliers

downstream consumers. So there were a lot of new questions we uncovered in the process, and a lot of the stress tests that have happened over the year, have you know, as part of their objectives, have had capacity building to just try and understand these questions better and kind of learn ways of answering them.

that then sort of tails off. Is that just because a phase of stress tests amongst certain central banks has been complete or is you know, I'm presuming this isn't because they were in fashion now they've gone out of fashion. It's just they've been done for now. Am I reading it right?

he really set the conceptual framework for stress testing. The Dutch Central Bank DNB was the first one to publish twenty seventeen twenty eighteen a transition risk and then a physical risk stress test. And it's no surprise that DNB actually published the first physical risk stress test because the Netherlands has a lot of exposure to physical risk from

sea level rise. Twenty nineteen is also a very important date for stress testing because this is when the NNGFS so the Network for Greening the Financial System, which is

an alliance of central banks and financial regulators. This is when the MNGFS launched its open source scenarios that really created the blueprint for a lot of central banks around the world to build their own stress test upon and so you know, in twenty twenty one, twenty twenty two, our stress test review shows that, you know, there's almost thirty exercises that have been published. Some of them are sort of first of a kind or one off, and some others have a frequency of you know, a year

two year. What we see is a lot of central banks integrating climate stress testing into their own standard risk review. It's not necessarily a part yet of what we call the capital requirements. So usually, you know, a stress test, one of the main goals would be to derive a number that would quantify the amount of money that an institution has to set aside to face a certain shock.

So climatory stress testing is not currently leading to capital adequacy requirements, but central banks are mainly trying to push to raise awareness around this new source of risk and also build cap city within their regulated entities in terms of the modeling, making sure that banks hire different teams that are able to look at these questions.

The main topic that stood out and the main work that stood out out of the NGFs is the NGNGFS scenarios. So these are climate scenarios that look across physical risk and transition risk. The long term scenarios are looking out to twenty one hundred on a five year basis. And when you think of you know, the nngfsnios. This is quite different actually from the scenarios that BNF can publish

or the International Energy Agency can publish. The NNGFS scenarios are not normative, so they don't tell you what you should do in order to achieve a certain climate outcome,

like a net zero emissions pathway. The nngfsnios are really here to explore different plausible futures and they're interested in sort of exploring those extremes as well, you sort of extreme a physical risk and extreme transition risk, and what happens if the transition is delayed for a long time and then all of a sudden, you know, in the twenty thirties, you have a policy shock that imposes a very strangent pathway to the economy to decarbonize very fast.

There are six to seven scenarios depending on the phase of the nngfsnios you're looking at, and these are really here to explore this world of risk from a physical and transition standpoint.

decalbonize fast. We don't accumulate as much greenhouse gases into the atmosphere, and so natural disaster are a little less frequent than they would have otherwise been. Disorderly transition is looking at what I already describe. So this is assuming that the transition does not happen in the short and mid term, but then by a certain date you start seeing more climate policies coming and disrupting part of the economy.

So you can think of for example, bands internal combustion engine cars being implemented pretty heavily across the world and so disrupting one part of the economy. And then the hot house world is a very high intensity, physical risk world where we do not decarbonize fast enough and so we continue accumulating a lot of emissions in the atmosphere, and this leads to very frequent and high amplitude and natural disasters and the physical risk impacts.

local jurisdictional features that they would want to incorporate. So they're very much as starting point, and in fact the Bank of England's climate be annual exploratory scenario or the stresses that they ran in the UK. The scenarios that were used actually deviated from the ng festerinaris in a few places, even though the anti festerinaris were used as the baseline, if you will, But then certain tweaks were made,

for example, around physical risks. There was an understanding that the NNGFS scenarios, due to modeling constraints that are that are well restrewed across climate modelers, there was a risk that the scenarios didn't capture the fullest extent of physical risks that may actually occur. So the Bank of england took steps to mpop some of the physical risks in their exercise, essentially to get to that severe but still plausible a level of stress that you want to analyze.

So Europe really played a role in that first wave of stress tests, so between twenty twenty one and twenty twenty two, and then we saw this spread across the world. So APAC really took the lead towards the end of twenty twenty two and twenty twenty three with a lot of new stress tests that were published on the back of the NNGFS scenarios. Overall, we found that between sixty and seventy percent of transition risk and physical risk stress test were conducted on the back of the NNGFS scenarios.

Another source of climate scenario data is the IPCC, and you have also different ways of capturing the risk. You know, different models, but NNGFS has really been the blueprint for a lot of countries to develop their own stress test.

sector level climate policies. And I think this is also this sets a very interesting precedent for other stress testers to look at, not just running the analysis in a very generic way, but running those stress tests in a specific policy orientated way.

for example. So I thought was this was a very promising episode. But actually right before the Trump administration came into office, So in Jazz Neurary twenty twenty five, the Federal Reserve actually exited the NGFs And so this is sort of a reaction potentially to the current politics and the current environment in US politics. But nevertheless, physical is are very important today in the US. So you're seeing heightened natural disasters. There's a wild fire season now in California.

You have zones across the US that have become unensurable from a residential and commercial real estate standpoint. And so while the US is sort of backing away from stress testing potentially for a while, there are actual effects that are being felt across the economy.

out with a set of recommendations in twenty seventeen. The same year the first under My stress test was published, and they set out recommendations for corporations to publish more information on various particularly around their climate related risks and how they were managing climate related risks. And these recommendations actually included a recommendation to apply scenario analysis to identify what those risks might be over the short, medium and

long term. And the TCFD recommendations have evolved the bit

since then. They have been formalized by the International Sustainability Standards Board, the ISSB, who has actually merged with the TCFD or vice versa, and has published a more formalized set of reporting requirements which have been formally adopted by jurisdictions across the world, which require companies, financials and non financial companies to disclose on their climate related risks and implicitly or explicitly, there's an askut in there, depending on

the jurisdiction that is, but there's an askut in there to apply scenario analysis to identify those risks. So we've seen, if you will, a kind of trend of private sector scenario analysis in peril level to the more broader central bank exercises. So that's been helping further establish climate stress

testing as a discipline across the world. Now, what we did notice with the Trump administration coming in is that the SEC which was set to mandate such a climate disclosure rule in the US, has currently stopped their efforts to get that rule across It was already being contested in a number of US courts and in Canada the regulator had a similar climate disclosure requirement that they were about to bring into force. And they've actually pressed pause

on that in response to developments in the US. It seems so we do see a bit of a setback

there as well. But a final thing I just quickly want to mention is that during all of these years of capacity building, my sense of the industry is that there's been a growing intrinsic appreciation of the risks brought forth by climate change, and as Tiffan pointed out, especially the physical risks which are which are very visible in some jurisdictions, and you know, with very real con sequences that we can see today financial consequences as well as others.

And so even though the regulatory push in Americas may subside somewhat, I think that might be compensated in part by an innate push by institutions to manage these risks regardless.

to support the insurance market. And so this is something we see across a number of states where public money is going to come to support locally some future homeowners that are looking to insure their house and so, you know, the collapse of the private insurance market is actually a very good proxy for you know, how critical physical risks

are to homeowners today. So you know, physical risks are becoming a real risk for households and companies, and so this is critical to still have, you know, an insurance market locally that can support those businesses. A collapse of the insurance market is also a collapse of the financial market locally. Climate risk is a manageable risk from a financial standpoint. You know, if you think about the full economy,

banks have a diversified portfolio. It is true that locally the market for insurance and the market for mortgages has already collapsed and public money has to step in already.

specifically you know, gelt towards credit exposure. We've seen also, you know, floods in Mexico being one of the main targets of one analysis across Brazil this is mainly a drought have pushed the financial supervisors to publish their own stress test. In China, we've seen also the regulator publishing transitionary stress test that was also in twenty twenty two.

So there's a lot of activities around the world and this really speaks to the sort of broadhouse strategy the NGFs has implemented, and going forward, we definitely expect a lot of regulators across Africa also to publish their first stress test using an MGFS blueprint mainly.

What can they do and more importantly, what can't they do and what's the gap.

And so one thing that I'll mentioned on the transition mix side is really the use of shadow carbon prices. So shadow CAB is a modeling artifact that the NNGFS models are using essentially to capture the intensity of climate policies overall. This is an approximation and basically says in the modeling, essentially you have these carbon prices that are applied to every single economic activity across the world and are driving up the cost of carbon intensive activities and

sort of helping also climate solutions to come into the system. Now, in reality, what we see in practice is only a quarter of global emissions are subject to an actual carbon price, and outside of the most carbon pricing schemes or taxes are well below thirty dollars pert on. So this is a level that's not necessarily extremely material for corporates, and they're often applied to corporate activities or industrial activities. So the central point in the transitional scynalysis of NGFs or

those carbon prices. But actually when you think about the transition and the weight it manifests, you know, it might be for example, the cost competitiveness of electric vehicles that is driving market adoption, and in turn this means you know, the sales of internal combustion engine costs are going down, So this is a financially material risk for a number of automakers, and it's not currently captured in the nngfscenios.

So what we've seen is up until phase five of the ngfsscenios, the scenarios really looked at transition risk in the power sector and also transition risk coming from you know, higher carbon prices across the economy, but it didn't necessarily assess the risk across you know, all the different sectors you can think of transport, but also transition metals that are going to play a key part in the transition just on the on the physical risk side as well.

The NNGFS being a bit of a global baseline for scenarios, it looks at physical risk very much through a macro economic lens. It tells us something about the macroeconomic impacts in a location in a country that is, you know, the kind of jurisdiction where we'd measure GDP, so that that's countries, what the impacts from physical risk may be at the aggregate level. In practice, physical risk is something

that tends to be highly localized. An intense hurricane impacting New York City could be very different from a financial standpoint from that same hurricane impacting forward. In both cases, we may, you know, there may be significant impacts, but the types of institutions that are impact which companies, which households differs quite a bit, and hence, you know, also the level of insurance of those impacts may differ, and ultimately what the impact of the financial institution is could

impact quite a bit. Similarly, physical risks, you know, so there's this direct damage depending on the location where the impact occurs. Damages can also kind of shocked to a company's supply chain. So a company itself may not be directly impacted by a physical risk, but if a critical supplier, say Transition metals, is impacted by physical risk, and that chokes the you know, the supply and the distribution of

those metals, that could impact downstream companies. So there's a lot of complexity to physical risk, which requires an understanding of the locations where companies operate and the supply chains of those companies as well. Now, naturally, NDFS, you know, being a global open source model, won't be able to

capture all of these specificities. So it's an important area where financial firms, if they really want to have a good grasp of the risk, likely will need additional sources of data to supplement that analysis.

is an issue. Whereas if the model was say, like let's in said, say you are an electric vehicle manufacturer, then I think that you reasonably could if you were running a scenario exercise to identify risks, you reasonably could be expected to capture, you know, the risks to your

supply chain. So does this system need a sort of a bottom up complement to provide metrics that can feed into the big picture model rather than putting the onus on central banks to try and figure everything out themselves.

have all the information. So if companies themselves, who understand their own business model better than anyone in their supply chains and their operating locations, if they can analyze the risk and report on that, that can then help other stakeholders, including their investors and their lenders to assess the risks

on their part. So certainly that bottom up component is there, and this is why we see a lot of push for climate related reporting in Europe for example, and in across asas specific with the ISSB standards that are being implemented.

came from Europe. Is there an issue that maybe the NNGFS is a little bit too eurocentric in how it thinks about the world, and that might be part of the limitation. And you know that the shadow carbon price issue is just an example of that.

are the basis for the scenarios themselves. And the tradition in integrated assessment modeling in this particular strand of academic literature is to use carbon prices as a metric for policy intensity. Now, where you know where you might have a debate is whether this metric is suitable for the use in stress testing, as has been applied by many central banks. And I think that's tiffense commons. You know,

strike a good court. They may not be as reliable and in fact, in the stress test themselves, in some of them you'll actually see acknowledgments by the central banks about the limitations of using a carbon price because it

won't accurately reflect the true policy environment. It was very much acknowledged as a shortcut that could help determine and pockets of risk, and you know, I think some of the subsequent analysis by Bloomberg and e F and others has shown that this shortcut may actually miss significant drivers of risk, like the automotive example that Tiffin mentioned.