Before coming to Jupiter Intelligence, I spent most of my career at a bank. One of my favorite former executives introduced me to the concept of the “Cs of Credit.” I have since found lists of many lengths referencing the Cs of credit in subsequent research. I quickly found articles claiming between three and eight “Cs of credit.” The quantity of Cs isn’t relevant, however. The critical point is none of these lists of Cs include the most important C — The C for Climate.

To start us off, let’s review the original five Cs (I learned it with five, so let’s start there):

1. Capacity: The ability of a borrower to repay a given loan.
2. Capital: The down payment. How much “skin in the game” does your borrower have?
3. Collateral: A secondary source of repayment in the event a borrower defaults. Your home is the collateral for your mortgage, which is why the bank gets your home in foreclosure if you stop paying.
4. Conditions: What is the present condition of the market? For a mortgage: are home prices increasing? For a small business loan: is your product going to sell? These are essential things to consider when approving a loan.
5. Character: What are the characteristics of the borrower? This one is tricky because federal regulations restrict what a lender can consider, but some items are allowed. Considerations like educational background, personal credit history, and work experience can all be used to evaluate borrowers.

Climate is Changing the Credit Landscape, Too

These Cs all make great sense to a legacy (pre-climate change) credit landscape. They address a common existing risk related to consumer and business lending. But today’s loans are issued in a physical climate that is changing — and changing fast. For example, consider the size of damages to the mortgage market predicted by the recent ECB stress test recommendations (a whopping single year 45% loss), the 2018 UCS report, and the recent HKMA climate stress testing results. Or, consider the billions in damages created by the Texas power outages last winter. These are just studies I happen to have read recently. Finding additional examples of massively expensive weather-related disasters is easy.

NOAA recently put out this excellent graphic:

Created by the National Center for Environmental Information (NCEI) at NOAA, this image is a masterwork of an infographic. I hope Edward Tufte would approve; I certainly do. The above image is only a small part of a massive online repository of data from NOAA. This figure focuses on the location of the twenty $1B+ disasters in 2021. The full dataset provides information about the relevant metrics for the intensity and frequency of weather-related disasters since 1980. The online portal they have created is excellent: it summarizes and presents these data in a very user-friendly format. I highly suggest spending some time browsing the site. Spend any amount of time with these data, and one point is clear: the cost of weather and climate-related disasters is increasing significantly.

A deeper inspection of the NOAA data reveals that the total cost of disaster events in 2021 was $145 billion. The total cost for the last five years was $742.1 billion. This figure is over a third of all climate damages experienced since 1980. Climate change is here, and we are experiencing increasingly damaging disasters that often have roots connected to that change. Given that we are still far behind the emissions reductions needed to meet the Paris climate agreement’s 2-degree warming target, it is worth emphasizing (yet again) the clear and present nature of these risks.

Climate Risks: Clear, Present, and Increasing

The risks are not only clear and present but increasing. For example, consider another impactful graphic from NOAA that is often cited in media emphasizing climate-related risks:

This chart has been around for a while and it is regularly updated. An older version of this chart is one of the first things that encouraged me to switch my career from building real estate mortgage models to one focused entirely on climate issues. The relationship between time and disaster frequency in the chart is pretty flat over the first half of that chart. Then, after 2005, the trend increased dramatically. To paraphrase Dr. Katherine Hayhoe: we had weather disasters every three months in the ’80s and ’90s that cost us over a billion dollars; today, we have weather disasters on that scale every three weeks. Below is a figure I created containing a much-simplified version of the same data:

In this figure, I have included trendlines for two periods — the period before 2005 and the period after 2005. I pick 2005 here as it represents a nice visual break in the trends, and I dub the pre-2005 period “Baseline.” Before 2005, the trendline slopes slightly upward, but the slope is not statistically different from zero (i.e., we can’t claim with confidence this line isn’t flat). The line slopes dramatically upward in the second period and is highly significant. This simple figure isn’t a rigorous statistical analysis, but it is highly motivating.

Another critical thing to note is that this is not a chart showing the path of climate change. This chart is essentially about accounting. It’s dollars and cents. The word climate isn’t even on the chart. I want to be careful with this point, as attaching the blame for weather disasters to climate change is a whole branch of climate science, called Attribution Science, and I am making no serious attempt to join that field with this blog. Yet, the results certainly are motivating. There has been a clear regime shift in the scale of damage from the weather over the years. As the retired head of the National Weather Service, Louis Uccellini, recently pointed out:

“There’s still some debate about exactly how it’s related to climate change, but there’s some fundamental physics going on here, of a warmer atmosphere and warmer ocean providing more fuel for these hurricanes.”

That quotation is about hurricanes in particular but the logic applies to many types of weather events that can be related to climate change.

With that in mind, let’s consider another exciting chart. This figure displays the global mean surface temperature change according to historical observations and various climate change scenario projections. It is from the IPCC working group 1 report released last August. This report was specific to physical science and climate change. It is only the first of three reports to be released during this round of IPCC reporting, known as AR6 since it is the sixth round of reporting:

The scale of this chart may mask one significant fact. There is a large increase in the slope between the baseline period and the present. That may be hard to see as the increase continues colorfully and dramatically past the current year. But the increase between 2005 and 2020 is visible in the grey line. Hovering near 0.5 degrees C in the baseline period, it is well past one degree at this moment. Some back-of-the-envelope math shows us this is a 100% increase in this fundamental climate metric over that period. If the change in the slope of observed data in the above chart is not convincing enough, consider this zoomed-in view from another figure in the summary for policymakers of the same AR6 WG1 IPCC report:

In the second panel of the above figure, the dramatic increase in surface temperature changes before 2020 can easily be seen.

I am very aware that correlation is not causation, and one should be careful to note that society will adapt over time to some of these risks. Even so, it’s hard not to see a similarity between these two charts. They both show a trend in data that is increasing and increasing fast. Observe the jump between the value in 1995 and 2020 in both charts.

Notably, this is a realized actual change in outcomes in both charts, not a modeled climate projection. With the relatively small existing and recent jump in global temperature, we have seen billion-dollar disaster frequency increase by over 300 percent compared to the period around 1985–2005. Over the next 30 years, this temperature metric is likely to rise up to another 100% relative to today. Imagine the devastation if the current pace of billion-dollar disasters increases at this rate for 30 more years. Again, simple math shows us that, absent significant adaptation or mitigation, 60+ billion-dollar events may occur in the United States alone by 2050. It would be both morally and financially unsound not to be concerned about the financial system’s stability with such a large quantity of annual shocks potentially on their way. The climate services industry, and Jupiter, exist to provide insights on where those potential 60+ billion-dollar events are likely to happen.

Climate Metrics Exist Now to Inform Credit Evaluations

Luckily (as I have pointed out), it is not all bad news. Forward-thinking finance professionals can adapt, thrive and help mitigate the risks of climate change. We can do this today because (as I have also pointed out) it is possible to obtain incredibly precise and granular information about climate risks and use these to inform credit decisions. They just need the proper data to do the evaluations. And that data include one more C, that is the C for Climate. Jupiter exists to provide the best version of this climate risk data to the financial system.

With that, let us revisit the list of the Cs of Credit:

1. Capacity: The ability of a borrower to repay a given loan.
2. Capital: The down payment.
3. Collateral: A secondary source of repayment in the event a borrower defaults.
4. Conditions: What is the present state of the market?
5. Character: What are the characteristics of the borrower?
6. Climate: Given the best available data and the assumptions made by the lending institution about climate scenarios, what are the relevant climate metrics for this loan?

Conclusion

I’m making a simple point: Climate change is an emerging threat to financial stability. Plus, it is not necessary to rely on models of the future with all their uncertainties—the climate change signal is already in the observation records as defined in the figure above. Climate change results from corporate entities emitting carbon without cost and treating our oceans and sky like an open sewer. As it results from unpriced climate pollution, climate change is an economic problem. This blog describes why I strongly believe physical climate risk is also a financial problem.

Ryan Vaughn, PhD is a Technical Product Manager at Jupiter Intelligence. Learn more about Jupiter at jupiterintel.com.

References

NOAA National Centers for Environmental Information (NCEI). U.S. Billion-Dollar Weather and Climate Disasters (2022). https://www.ncdc.noaa.gov/billions/.

Smith, Adam B. (2020). U.S. Billion-dollar Weather and Climate Disasters, 1980 — present (NCEI Accession 0209268). NOAA National Centers for Environmental Information. Dataset. https://doi.org/10.25921/stkw-7w73.

IPCC Sixth Assessment Report, The Physical Science Basis: https://www.ipcc.ch/report/ar6/wg1/.

The Next C of Credit was originally published in Jupiter Intelligence on Medium, where people are continuing the conversation by highlighting and responding to this story.

How do we know what we know?

High water risin’, the shacks are slidin’ down.

Folks lose their possessions and folks are leaving town.

- Bob Dylan, High Water (2001)

It is clear that we should be concerned about how the housing market is responding to climate change. Bob Dylan sang about it as far back as 2001, and Sean Becketti from Freddie Mac noted the following in 2016:

“It is only a matter of time before sea-level rise and storm surge become so unbearable along the coast that people will leave, ditching their mortgages and potentially triggering another housing meltdown — except this time, it would be unlikely that these housing prices would ever recover.”

-The New York Times, 11/24/2016.

But it is much less clear if the average home buyer is concerned (yet). Recent research from Redfin demonstrates that, counterintuitively, migration is strongest into some of the highest climate risk regions. Additionally, the Wall Street Journal profiled strong demand for homes in fire-prone Napa Valley. So the titans of rock and roll and the lead thinkers of the mortgage industry are worried, but the average buyer’s behavior seems to indicate little to no understanding of these risks. This article is the first in a two-part series that attempts to explain the current state of research into how climate is impacting housing. In this first part, I focus on the question “How do we know what we know?” In the follow-up, I’ll summarize what we actually know.

What are the Major Challenges?

Housing is diverse, and housing demand is rapidly changing.
It is conceptually complex to figure out what houses are worth. Things like cars are just the sum of their parts. You can take them with you wherever you go, so the face value of a car is its total value. By contrast, a house is fixed in place. Valuing a home involves understanding how consumers value a bundle of goods that includes not only the home itself but also the amenities (or disamenities) offered by the area around the house. Things like school quality, crime rates, walkability to bars and restaurants, and flood risk are all part of a home’s value. Estimating how things like the weather impact the home’s value is therefore complex because the home itself and its surroundings influence each other. Consider that homes on the Gulf Coast tend to be built on stilts to avoid floodwaters. Adaptation measures like these exist most often in areas with a history of flood risk. Therefore, a region with more historical flood risk is likely to be more resilient to increases in future flood risk. We can’t compare flooded homes with large levee improvements in a post-Katrina New Orleans to flooded homes in nearby areas without major levee improvements like Biloxi.

Remote work has completely changed the game.
As remote work becomes more common, we are likely to see a combination of risk- and amenity-loving consumers moving towards coasts and woodland areas. The areas that are exposed to the highest levels of climate change-related risks. The migration towards risky areas is happening simultaneously while climate-risk-averse buyers are moving away from those same amenities. In fact, the local beliefs in climate change are a significant factor driving its impact on home sale prices. In this increasingly complex landscape of real estate, it is no wonder a hard-working professional real estate appraiser can earn well over six figures. The market value of a home is the complex interaction of the home’s size, construction quality, local schools, crime rates, climate change, and animal spirits all at once.

Good data on housing is hard to get.
For most of the world, housing data are limited at best. Datasets exist for the US, Australia, the UK, and a few other parts of Europe, but for most of the rest of the world, they are an unknown quantity. Quality and geographically explicit housing data are most readily available in the United States, and thus most studies focus on the US housing market. In the US, housing is big business, and there are multiple substantial and very successful companies dedicated to collecting, cleaning, and reselling information about the US housing stock. In addition, there is a lot of money in the housing market and, thus, strong demand for good analytics. For example, Freddie Mac forecasts USD $1.9 trillion in new purchase originations in 2021 alone. Unfortunately, all this demand means that there is money to be made, and thus these data are not free. Some companies like Zillow and Corelogic offer their data to university-affiliated researchers for free, but these data are not perfect and only available to academics. Further, even after obtaining relevant data, researchers could be prohibited by copyright law from publishing their data along with their results, so new researchers cannot challenge the findings of previous studies.

How Do We Measure Impacts? Events versus Boundaries

In this short essay I focus my attention on flood risk, as it is the most studied of the existing climate perils for real estate. I classify the existing studies on flood risk into two broad types: event-based studies (hurricanes and floods damaging things) and studies exploiting existing floodplain-related boundaries that impact the sale of a home (these are “information effects”, i.e. “My flood insurance is how much?”). The first can be thought of as measuring the response to actual weather events, and the latter as measuring the impact of information about the probability of weather events in the future. Event-based is the larger of the two categories, with Hurricanes Sandy and Harvey representing the two most studied events.

Event-based studies examine the impact of significant weather events on the price of the local housing stock.
These studies compare homes at the edge of the affected region that survive the event unscathed with homes inside the affected area that were damaged. Think of a randomized control trial (RCT) experimental design. By happenstance or accident, homes really close to the border of a disaster ended up there nearly at random. For example, when a hurricane hits the coast, it typically doesn’t destroy the entire region. Some homes will survive untouched, some will be damaged, some destroyed. Using information about price changes for similar damaged and undamaged homes, or homes in impacted and nearby “near miss” areas, the author can estimate the impact of the weather event on home prices confidently. A recent paper by Francesc Ortega and Suleyman Taspinar from CUNY Queens on the impact of Hurricane Sandy on New York real estate is an excellent example of this type of study. Each study of this type is then uniquely responsible for justifying the external validity of its estimates to as-of-yet-unrealized disasters. This validation is often a daunting task and is a significant weakness of this type of study.

Floodplain boundary-based studies exploit the fact that artificial legal boundaries often define homeowners’ amount of, and types of information about, risk(s) to which a particular property is exposed.
For example, most often, the only information about flood risk that’s available to consumers comes in the form of a written warning from their lender that the buyer must purchase flood insurance as collateral for securing their loan. In the US, this requirement is determined almost entirely by whether the home is located inside a FEMA-designated special flood hazard area (SFHA). Hino and Burke (2020) exploit changes in these SFHAs over time to study the impact on home prices for homes that were moved from outside to inside these boundaries after those boundaries were revised. Other studies, like Keys and Moulder (2020), abstract from floodplains and use census tracts with associated measures of high or low Sea Level Rise risk to each census tract and examine the distribution of home prices within each area.

Concluding Thoughts

I have laid the foundation here for an understanding of how researchers approach the questions of flood risk, climate change, and the housing market. There are certainly studies that do not fall into my broadly defined dichotomy. But these two methods describe the most common techniques at a level sufficient to support a dinner-party conversation about how exactly economics approaches this question. In Part 2 I will dive deeper into the results of these studies — attempting to parse what we actually know into a useful rule of thumb that can be used to approximate the likely impact of climate-change-related risks on the housing market in general.

Ryan Vaughn, PhD is a Technical Product Manager at Jupiter Intelligence. Learn more about Jupiter at jupiterintel.com.

Climate Change, Weather Risk, and Residential Home Prices — Part 1 was originally published in Jupiter Intelligence on Medium, where people are continuing the conversation by highlighting and responding to this story.

As a product manager, I try to stay abreast of the metrics and techniques used by my fellow public and private “climate translators:” credible experts who interpret climate projections; understand their assumptions, strengths, and weaknesses; and may derive metrics from them. Sometimes I’m impressed with a new way to think about, say, wildfire prevalence…and sometimes I’m worried that what’s being put out there hasn’t been generated with sufficient scientific rigor.

In previous posts I’ve generally discussed a few places where climate analysis can go astray, such as a reliance on old models, inappropriate metrics, or vague scores. Today, I’m diving into the very first step: how should users evaluate a climate translator’s selection of the source climate models that underlie their metrics? What are the pitfalls that could harm metric quality? Unfortunately, a wide range of quality is available, but, assuming a data provider is open about their methods — and they should be — we can learn to identify both red flags and the signs of sound analysis.

General Principles for Model Selection

No matter the metric, climate data users should hunt for metrics generated using a few key principles.

• Using the models from several modeling groups increases the likelihood that the results represent the broader scientific consensus, as opposed to relying on the view from a single group. And including data derived from independent sources helps maximize the information content that is input into any quantitative analysis.
• Prior to use, each individual model should be assessed for its skill in simulating the peril of interest in the region of interest. All models have strengths and weaknesses; these can be sussed out by surveying academic papers on the topic and by comparing a model’s simulations of historical climate to the actual prior conditions to identify bias.
• Uncertainty should be clarified where possible. This may be a qualitative description about different climate projections, based on how they are produced. It may also be a quantitative analysis based on underlying data samples.

Following these principles is a good step towards producing a quality metric about the future, particularly if that metric is a chronic peril. Chronic perils are those that occur frequently, such as drought, thunderstorm, and average temperatures. Because they are common, they are well represented in a future projection. It’s still necessary to examine multiple years of data, such as the targeted projection year ±5 years — this helps tease out the underlying climate change signal from the effects of shorter-term climate oscillations, such as the El Niño–Southern Oscillation (ENSO), that also affect climate projections — but the statistics are relatively manageable.

Chronic perils do have their pitfalls, but, as we’ll see, they present a more straightforward challenge to climate translators than acute perils.

Model Selection for Acute Perils

Acute perils occur at the “tails” of a probability distribution: the extreme winds, floods, and rain events that can cause severe damage. By definition, acute events are rarely seen in climate models. For example, a 100-year flood only has a 1-in-100 (1%) chance of appearing in a climate projection of a particular year at a particular place. The World Meteorological Organization (WMO, 1989, 2007) suggests that at least 30 years of data be used to fit a distribution that can characterize acute metrics; for climate data, that means 30 simulated years.

So: how can climate translators find 30+ data samples from simulations that are relevant to, say, Miami’s wind speeds in 2040? I’ve seen a few techniques:

• (Valid) Use data from a range of years surrounding the target point. As mentioned earlier, this helps isolate the climate signal from sources of interannual variability like ENSO. However, getting to 30 samples by using the central year ±15 years is too wide of a range: because the climate is changing, metrics for the target year can be skewed by results from distant (warmer) future years, particularly when the focus is on those extreme events. This technique can only get you so far.
• (Valid) Use models with an ensemble of projections. Ensembles are generated by running the same climate model with slight differences in its initial conditions, such as local temperature and pressure. Each ensemble member follows the same pattern for carbon dioxide concentration and radiative forcing — this is determined by the climate scenario — but the initial perturbations allow for different weather to evolve. Ensembles were rarely found in the CMIP5 generation of models, but they are more prevalent in CMIP6 and in standalone experiments such as CESM-LENS (Community Earth Systems Model — Large Ensemble). It’s statistically sound to draw samples from a climate model’s ensemble members to fit a distribution and thus “fill in” the tail of the curve.
• (Invalid) Pull data from multiple climate models and treat them as ensembles of one system. Doing so implies that each sample is being drawn from the same distribution, but that is not a statistically sound assumption. Climate models can be vastly different: they can have different spatial resolutions, predict or ignore different aspects of the earth system, have different model physics, and other factors. In short, each model has its own climate (and no model’s climate is the same as the real climate). Statistical methods that require the assumption of identically distributed data — a common assumption in many statistical theories — are invalidated, and we cannot expect trustworthy results from using this technique.

At Jupiter, we use a combination of the first two techniques: 11 simulated years and only those global climate models (GCMs) with at least three ensemble members. For each GCM, the 33+ years of simulated data that this process generates is enough to define a distribution and calculate that GCM’s view of an extreme peril, after which it can be combined with the perspectives of other GCMs. We believe that this strikes the right balance: require too many ensembles and the pool of acceptable GCMs is too small, require too few ensembles and you have to sample from a larger set of years and potentially skew your results. While these techniques limit the GCMs that we can include for our extreme perils, it still leaves us with plenty of GCMs to work with and results in a more responsible understanding of tail risk.

Implications for Climate Data Users

Several months back, I wrote about the importance of commencing any physical climate risk analysis with the “right” data, stressing the value of data appropriateness: selecting the right metric for the job. For example, if you want to understand the impact of future hail risk to your solar farm, you should find a way to measure future hail risk directly, instead of using a proxy like atmospheric instability.

Occasionally a climate translator will sacrifice metric appropriateness because their methods don’t allow them to produce high quality metrics describing the tail of the curve. For example, they’ll model “extreme wind events” by calculating the maximum wind gust at the 1-year return period: the hardest the wind will blow during a year. While such a metric may not require a complex model ensemble approach, the metric itself isn’t appropriate for understanding a peril like a tropical cyclone or European winter storms, which don’t strike the same location every year. Instead, it is critical to look further out on the tail — to the 50-year, 100-year, and more extreme frequencies. And that can only be done responsibly with techniques like those described here.

Worse still, an unqualified climate translator won’t employ these techniques, but they will produce an extreme tail metric anyway. And they might get away with it — for now. These are tricky concepts to grasp, and it’s easier for users to assume that the data is credible than to dig in and possibly discover it to be highly suspect. But as climate data moves into the mainstream and users become more comfortable with these concepts, they will be demanding more from their data providers. And we should be ready to step up.

Meghan Purdy is a Senior Product Manager at Jupiter Intelligence. Learn more about Jupiter at jupiterintel.com.

Consider the Source: Climate Metrics and Their Underlying Models was originally published in Jupiter Intelligence on Medium, where people are continuing the conversation by highlighting and responding to this story.

One of my favorite podcasts, Freakonomics, often interviews luminaries in economics and other fields. They often repeat their questions across interviewees. My favorite question is: “What’s something that you believed to be true for a long time, until you found out that you were wrong?”

I would answer that question by returning to my college years. I spent four years rolling my eyes at the various Women in Science and Women in Business clubs. Why did women need a “safe space” to talk about science or business or our careers? Why did we need a leg up in these industries? Then I moved to Silicon Valley.

I’d sit in meetings with a 2:17 ratio between women and men. I’d watch women be repeatedly talked over — including myself. I received a performance review with heavily gendered language about being too ambitious. A top executive was challenged on our paltry record of women in leadership and snapped back that he didn’t see a problem. (He later apologized.) And that’s just what I could see from my perspective as a white woman; Silicon Valley and the sciences don’t have a great track record when it comes to many other underrepresented groups. Clearly something was wrong. What could be done?

When I came to Jupiter nearly two years ago, I found a much more welcoming and congenial environment, but it was obvious that we still had a long way to go. So I leapt at the chance when my former colleague Dr. Betsy Weatherhead, Nobel prize-winner and pioneering scientist, asked me to join the newly-formed Diversity, Equity, and Inclusion committee. It was the middle of a pandemic, but we tasked ourselves with envisioning what we wanted Jupiter to be, and how we were going to get there. As Jupiter exits its Series C and jumps into a period of rapid growth, how do we make sure that the company we are building is one that we are proud of — where people of all backgrounds are supported and can collectively succeed?

We divided our initiatives between:

1. attracting and hiring job candidates from underrepresented groups, and
2. retaining employees with a supportive and inclusive culture at Jupiter.

Finally, we decided that we would hold ourselves accountable by publicly committing to these efforts in a detailed, action-oriented DEI statement. With the full support and enthusiasm from our leadership team, the first iteration has been completed, and you can read it on our website. Here I’ll highlight a few of our more thought-provoking discussions and initiatives. While I am far from an expert on these issues, I hope this illuminates practical concerns when it comes to implementing new initiatives at a company.

Broadening our talent pool

Hiring can be fraught with challenges, including ugly perceptions that someone is a “diversity hire” for whom standards were lowered. Our goal at Jupiter remains to hire the best candidate for a role, but we are learning two things:

• Not all candidates have the same access to our job postings. For example, if we always return to our networks when a position opens, we will likely hire people who look and think a lot like ourselves. We need to broaden our reach by posting to groups like Black Girls Code and Women in Machine Learning, and we need to ask our recruiters to dig deeper when bringing qualified candidates to us.
• We are all biased in our interviewing techniques. For example, if I saw on someone’s resume that they also liked to trail run, I would have an easy time breaking the ice and chatting all about it. But trail runners are far from a diverse bunch — we’re likely to be white, young, and live in affluent areas with easy access to trails. I would have a harder time warming up to someone who looks and has hobbies that are different from me. And so we need to acknowledge that our biases affect the conversation and work harder to move past them. We need to make our interview process more consistent by asking similar questions and providing more structured candidate assessments. And, of course, all interview committees should include employees from a range of backgrounds.

With these initiatives, and a few others described in the statement, we hope to attract and hire the best candidates, full stop.

Supporting our colleagues

Our second area of focus is the retention of people from all backgrounds by providing an inclusive working environment.

• Everyone has unconscious biases that can cause us to act badly at times. We can give employees the language to identify these situations by defining Jupiter’s values and defining what diversity, equity, and inclusion mean to us. Only then can we identify and learn from our errors and move forward.
• People from underrepresented groups may be less likely to raise their hand and ask for a promotion, a raise, or a leadership opportunity. We can remove the burden of asking from them by having a uniform cadence for promotions/raises, clear responsibilities at each level of the career ladder, and a more open mind when assigning leadership roles. Buddy programs for new hires, mentoring, and training programs are also critical to ensuring that everyone’s career can grow.

Keeping us accountable

The last part of our statement defines how we will track the success of our initiatives and define new places where growth is needed. No statement is complete without it, because how else can we be held accountable?

I’ll close with a sincere thanks to my colleagues — Rahul Dubey, John Exby, Elizabeth Perry, Justin Rogers, Craig Simmons, and Eric Wun — who all showed great dedication and care in defining these initiatives and putting them into words.

The first draft of our DEI statement may be complete, but it is not a time to be self-congratulatory. It is time for the real work to begin.

Meghan Purdy is a Senior Product Manager at Jupiter Intelligence. Learn more about Jupiter at jupiterintel.com.

Building an Inclusive and Equitable Jupiter for Everyone was originally published in Jupiter Intelligence on Medium, where people are continuing the conversation by highlighting and responding to this story.

Something Could be Finer: Making GCM Data Actionable

The climate crisis is unfolding as I write these words. Recently a climate change-induced “heat dome” smashed temperature records in historically temperate areas all over North America. Climate-crisis-related disasters are crossing some very interesting and very scary thresholds. For example, during the recent heatwave, portions of Canada exceeded the all-time record high temperature recorded in well-known hotspot Las Vegas, Nevada. Yet even with these record-breaking events, some sites will make it through relatively unscathed. In another example, in New York City, heavy rains are flooding streets in places that don’t typically flood. But not all of NYC was negatively affected. Some parts of the city are quite capable of coping with the recent floodwaters. These differences are absolutely a social justice problem. Yet, before we address societal root causes, we must understand the geography of the risks. And therein lies the crux of this article. If our goal is to assess the risk of the climate crisis on specific regions, or specific assets, then granularity matters.

The United Nations International Panel on Climate Change (IPCC) collects and freely distributes multiple “Global Climate Models” or GCMs. These data are freely available. In recent years, many companies have been founded to collect and rescale these GCM data to be more granular and user-friendly. The publicly available GCM data are the combined output from hundreds of scientists from all over the globe. Given the amount of effort already invested in producing these datasets, why would private companies need to spend years and tens of millions of dollars refining the output?

The answer is that weather is what policymakers need to understand, and climate is what the GCM data are designed to study. The climate is not the weather. Climate is a description of a long-run average over a large area, and the weather is the realization of climate in a small geographic and time scale. Questions such as What is going to happen to my town? and What is going to happen to my house? are questions about the weather. Understanding how climate and weather interact is complicated. To understand how difficult this is, consider that companies (for example, the one at which I work) spend millions on massively complex cloud computing resources to take what the GCMs say about the climate and transform it into something we can say about the weather. This practice is called “downscaling,” and it is no small task.

A Relatable Example: Assets and Decisions Look Different at 1.5-Million-X Higher Resolution

Consider the risk faced by Starbucks locations in Florida. Starbucks makes delicious caffeinated beverages that increase my productivity (and therefore, I do not mind giving them some free analysis of their climate risk). The choice of Florida is obvious; “The Sunshine State” is subject to significant levels of flood risk, and flood risk among all climate perils is the most spatially variable. I picked Starbucks locations for this example because I like coffee, they are everywhere, usually similarly sized, and as a playful nod to the below image that spawned a viral photoshop battle a few years ago.

As noted above, GCM models need some help to be decision-useful. They typically simulate the Earth’s climate at one degree (or greater) spacing in latitude/longitude (approximately 110 km). That level of aggregation has important implications for the usefulness (or lack thereof) of these data. To quote a recent article in the journal Nature from Alice C. Hill:

Without information about where and how damaging events are likely to unfold, choosing the right adaptations to invest in is little more than guesswork. The smallest resolution of climate models is generally at a scale of 100–150 square kilometres. That sort of area can span several towns and extreme differences.

The exact extent of 110km grid cells may be hard to grasp. To get an idea of just how large an area the raw GCM outputs cover, I plot a grid of red one-degree latitude/longitude lines over Florida in the figure below. Note that only two of these large grid cells (and thus two risk levels) exist to describe all 6,297 km² (2,431 square miles) of Miami-Dade County. Anyone seriously concerned about local conditions in Miami will need better data than what’s offered by 1-degree grid-cell resolution. The figure callouts demonstrate how climate service providers fill this gap. Combining the GCM output with expert science and local ground conditions, they refine (downscale) the GCM outputs to the 90m grid cell level. For context, within a single 1-degree latitude/longitude cell, there are approximately 1.5 million 90m grid cells.

In my analysis, l use Florida counties to represent aggregate level data instead of 1-degree grid cells. A county is usually slightly smaller than a 1-degree grid cell, but it makes for a more naturally understood aggregation. Plus, even with this somewhat smaller aggregation, it is still very apparent that the granularity of the data matters.

Both high and low-risk regions are well represented in the sample. Of the 302 Starbucks sites sampled, roughly 30% are located in the very-high-risk Miami-Dade and Broward counties. The next most Starbucks-dense county is the relatively much safer Orange County. The figure below summarizes the risk level at each location and county. The green color indicates a lower risk for both counties and points, and the red color indicates higher risk.

Aggregate Information Masks Realities “On the Ground”

The main takeaway from this analysis is easy to see from a simple visual inspection of the figure. Not all the locations in very risky counties are at high levels of risk, and not all places in ostensibly safe regions are safe. Therefore, focusing only on the aggregate information would often lead to inefficient or — worse — flat-out wrong conclusions. Furthermore, errors from aggregation are often correlated with things we are interested in measuring — for example, the level of damage to structures. This can lead to biased estimates when the climate projection data are used in downstream models.¹

If we zoom in further to Miami-Dade County, the pattern is even more obvious. Multiple Starbucks locations are colored green, indicating relatively lower risk, and many others are dark red, indicating higher risk relative to the county median. Even in Miami, the region widely regarded as being one of the riskiest, there is broad variance in outcomes at a hyper-local level. For any asset or property manager, this is crucial information to understand.

Taking another point of view, consider the scatter plot below showing the risk score of all our Starbucks locations plotted against the average risk in each county. The color in this chart indicates whether each site is above or below the 45-degree line, indicating higher or lower risk on average than the surrounding area. The plotted circles nearly form a square, indicating that the total range of the risk within each county is only loosely correlated with the average risk in that county. In other words, the maximum and minimum risk scores don’t change much from the safest to the riskiest counties, and just knowing the risk of a county is a potentially poor predictor of a given location in that county.

A Resolution to the Granularity Challenge

In this playful example, I show that the scale at which we view climate change data, a.k.a its granularity, really does matter. This blog highlights a source of measurement error in climate analytics that is very likely to introduce unwanted bias in any models using these data. When we know the location of assets at a granular level, we should evaluate climate perils at that same level. Fortunately, the granularity problem is easy for analysts to solve; simply contact your friendly neighborhood climate service provider. There are tools in the marketplace available today that provide this level of resolution. So why not start there?

Ryan Vaughn, PhD is a Technical Product Manager at Jupiter Intelligence. Learn more about Jupiter at jupiterintel.com.

Footnotes

¹ As noted by Solomon Hsiang in Climate Econometics, “Coarse climate measures introduce large measurement errors that will cause attenuation bias, leading to under-rejection of the null hypothesis”

Granularity Matters was originally published in Jupiter Intelligence on Medium, where people are continuing the conversation by highlighting and responding to this story.

Climate change is alarming. It is indeed a climate crisis. Regardless of the actions we take to decarbonize, changes are coming. The seas will rise, storms will strengthen, the planet will warm, crops will die. The future we face is uncertain and challenging. But humans are intelligent. We have tools to adapt and reduce the damage. Forward-thinking finance professionals can adapt, thrive, and help mitigate the harm of these coming changes. It is not all bad news.

Early thinking about the financial risks of climate change dubbed them “unhedgeable.” In finance, hedging is a strategy to limit risk exposure. Projections of future physical climate risk from Jupiter Intelligence can provide insight into which areas are riskier than others.

A risk that can be defined is a risk that can (usually) be hedged.
Consider Florida; after years of “sunny day flooding,” Florida is the iconic example of sea-level rise risk. However, we have no reason to expect that the entire state of Florida will ever be entirely underwater. As John Englander notes in his book Moving to Higher Ground, “Florida is neither flat nor about to ‘slide beneath the waves’…the tourism mecca of Orlando is a solid 80 feet above sea level and not in any direct jeopardy” even in a world with ten feet of sea-level rise. Some areas are always safer than others.

If the entire distribution of risks is understood, then the market can price them.
When prices reflect risks, markets will naturally drive resources to safer areas. If the risks are known, newly informed migrants to Florida will “vote with their feet” and target safer regions at higher elevations.¹ This migration will drive up home prices in safer areas, and builders will shift their attention to these new higher profit zones. Companies interested in locating in Florida will go where the highly skilled population is and look to the safer, higher elevation cities. New capital will naturally shift to safer areas, builders will profit, homeowners will enjoy safer homes. These natural shifts in the market landscape will reduce the overall social damage from the climate crisis. This mechanism driving us to better outcomes is the “invisible hand” economics talks about so often.

The Importance of Adaptation

In a departure from previous international agreements, the Paris Agreement codified adaptation and flexibility into international cooperation on climate change. Two key features of the Paris Agreement, differentiating it from its predecessors, are local autonomy and the importance of supporting adaptation efforts. Each country must set its own goals for emissions reductions and how it will reach those goals in the agreement.

This flexibility to local conditions is great for financial markets and local communities already planning for resilience. It means capital can flow where it is likely to generate the highest returns. The era of significant infrastructure investments will return. The focus of the Paris Agreement on adaptation will direct resources to massive engineering projects — like the Maeslantkering in Holland or the MOSE barrier in Venice — that will inspire future generations and provide jobs and investment opportunities for years to come.

The climate crisis is an existential threat, but there is an irrational level of pessimism about our ability to adapt to future challenges. Not only have we as a species explored the moon and Mars, but recall that, well before the age of modern computing, human engineering changed the direction of the Chicago River. While there are some issues with that particular bit of geoengineering today, it successfully adapted Chicago to a physical peril for more than a century. Extending the life of our greatest cities by more than a century is certainly an excellent investment.

With modern technology and a minimum of $100B committed annually in the Paris Agreement to help developing countries reduce their emissions and adapt to climate change,² some very innovative projects are likely on the horizon.³ To accomplish them, substantial amounts of money will need to be directed towards its best uses. Financial markets are the vehicle in which these funds will travel, and reasonable climate risk projections are the signposts they will follow.

Quantifying the Financial Impacts of Physical Risks

Climate risks are typically split into two categories: transition risk and physical risk. The former are those associated with the transition away from a fossil fuel-based economy. A primary example of transition risk is that many companies count as assets oil reserves that are still underground, even though, in a low carbon economy, these reserves may be worthless. Physical risks are those associated with emerging climate perils. The rising sea levels that threaten our coastal cities are a primary example of physical risks. Both transition and physical risks are global in scope and potentially devastating in scale, but the financial system must adapt to each in different ways.

Transition risks depend on political action and their occurrence is uncertain. They are relatively more straightforward to model as financial impacts.⁴ In contrast, some degree of physical risk is nearly certain to occur. No amount of mitigation today will prevent our climate from changing for decades to come. But physical risks are tough to model as financial impacts.

Jupiter Intelligence provides the data to help close this gap, yet Jupiter’s models do not remove uncertainty; they quantify it. Physical risks can be thought of as increasing the variance, or volatility, of the value of any asset. Luckily, finance is good at dealing with uncertainty and volatility; one might say financial markets exist because of them. The game played in markets is to capture new information that reduces uncertainty and quantifies volatility. Quantified volatility can be hedged, and a hedge reduces risk. Reducing — not removing — uncertainty and understanding volatility are all that is needed to hedge risk and increase returns.

Consider the investment mechanisms such as credit derivatives and default swaps that proliferated during the 2003–2007 period that preceded the housing market crisis of 2008–2012. Instruments like these are essential to keeping liquidity in financial markets when tail risk is large. Given good data on where climate risks are and are not, these same instruments can be adapted to hedge climate risks. Ironically the very financial innovations that once brought the worldwide economy to its knees may save us from a climate-related economic meltdown.

A Case Study: Mortgage Loan Fulfillment

There is no way to relocate real estate assets to safer ground quickly. Investments in real estate assets are also long-term investments. A typical home, once built, can last for a century. The average home buyer in the United States will remain in their home for a decade or more⁵ and loans on residential real estate amortize over 30-year horizons. Any investor lending money towards the purchase of real estate is assuming that the risks faced by that region will be somewhat stable into the future. Knowing where to invest requires validated, transparent, and granular information about which places are safe, and which are not. Jupiter Intelligence is well-positioned to help investors make these investment decisions about real estate location.

Research has shown that climate change matters most for home prices when the risk of weather-related damage is “new news” to the homeowner. Today the only information a buyer receives about this risk comes in the required insurance policies for flood or fire. Yet neither of these are perfect signals of risk. Flood insurance, in particular, is based on an outdated and imperfect system.⁶ With better information about risks, rational home buyers would ask for a discount on riskier properties. Without this information, climate events will shock both homeowners and lenders, and thus negatively impact the housing market. If climate disclosures become a part of mortgage loan transactions, the risk will be priced in at the beginning. The resulting changes will be expected, gradual, and properly hedged. Jupiter can provide exactly this information to the loan fulfillment process.

Mirror Images: Opportunity and Risk

We at Jupiter believe it is important to highlight these opportunities alongside the risks.

As President Biden noted in the Executive Order on Climate Related Financial Risk, “this global shift presents generational opportunities to enhance U.S. competitiveness and economic growth, while also creating well-paying job opportunities for workers.”⁷ The promise of future possibilities is more motivating than the risk of doom and gloom. As a motivating example, note that in his presentations on climate change, Al Gore always devotes at least a third of his slide deck to a hopeful message. He calls this allocation the “hope budget.”

To move forward, we need a hope budget. The financial system should prepare for the worst. Yet our mindset should remain hopeful, focused on the coming opportunities. The financial system exists to move resources as quickly and efficiently as possible to their location of best use. The promise of opportunity and good information about risks will move resources towards a smooth and just transition.

Ryan Vaughn, PhD is a Technical Product Manager at Jupiter Intelligence. Learn more about Jupiter at jupiterintel.com.

Footnotes

¹ Barth, Molly, and Rollins, John. Climate Gentrification and the Role of Flood Insurance. Milliman, Inc. (September 2019)

² International Affairs 92, №5. (September 2016)

³ See, for example, blue carbon and nature based coastal resilience initiatives in the Ocean Based Climate Solutions Act

⁴ As Robert Pindyck of MIT has often pointed out, there are still complications in modeling macro-economic transition scenarios.

⁵ Katz, Lily. One in Four U.S. Homeowners Have Lived in Their Homes for Over 40 Years — The Highest Share on Record. Redfin (21 January 2021)

⁶ Jupiter Intelligence Special Report. The Danger in Relying on FEMA Flood Maps For Risk Management.(29 April 2020)

⁷ The White House. Executive Order on Climate-Related Financial Risk. (20 May 2021)

It’s Not All Bad News was originally published in Jupiter Intelligence on Medium, where people are continuing the conversation by highlighting and responding to this story.

Hampered by a historical focus and other shortcomings, cat models can’t adequately quantify future risk in a rapidly changing climate. Analytics based on best-in-science climate models give insurers a far-sighted assessment of climate-driven impacts.

I was captivated by catastrophe modeling from the moment I heard of its existence at a presentation by RMS at my college career fair. By Christmas I had accepted a job offer, and, less than a month after graduation, I had embarked on what was ultimately a 10.5-year career at the company before coming to Jupiter. As an earth and atmospheric sciences major, I was fascinated that the principles instilled in my Harvard ivory tower could be put to such practical use.

For over thirty years, cat models (primer available here) have quantified natural disaster risk, helping re/insurers confidently underwrite and trade risk. Insurers could craft underwriting rules based on their risk tolerance, and reinsurers could diversify portfolios of Prague Flood, Gulf Coast Hurricane, Japan Earthquake, and Australia Bushfire.

But as the climate changes, our exposure to these high-impact events is growing. The cracks are starting to show. The Geneva Association recently stated: “While highly useful for today’s and next year’s portfolio management, catastrophe models are limited in providing decision-useful quantitative information over a longer term, impeding progress toward a forward-looking climate change and exposure/vulnerability landscape.” Seven-time Leading London Underwriter award winner Richard Trubshaw of MAP was more blunt in his December 2020 Managing Agent’s Report:

The problem is, what [reinsurers] are being fed (via the counterparty’s intermediary) is risk-output from a severely under-cooked model. I find it particularly ironic that industry veterans will regularly opine on global warming and climate change…yet their organisations are still pricing to the same mis-calibrated metrics that were in place five years ago, despite the heightened catastrophic frequency since 2016, in both the Atlantic and Pacific, despite deterioration in loss estimates, despite the proliferation in wildfires in California and Australia — and that is observed reality, let alone whether the future will actually be even more extreme.

What can be done about this “under-cooked” view of present-day risk and the expected future changes? With cat models thoroughly entrenched in the insurance underwriting and risk management frameworks, it would be extremely convenient to tweak a few settings and create a “climate adjusted view of risk” that is valid for present-day and the 5- to 30-year time horizons for which boards of directors and regulators are increasingly asking. But do these tweaks truly address the short-sightedness of present-day views of risk? What is gained and lost with these methods?

Adjusting Cat Models for Climate Risk: Current Approaches

The foundation of cat models is a set of stochastic events: simulated hurricanes, earthquakes, floods, and other events. Models assign explicit frequencies to each event, and these frequencies become the basis for simulated timelines of how these events may occur over the next 1–5 years. For example, in a 10,000-year simulation of Florida hurricanes, Miami might be unscathed in 9,400 of those years, but, during the other 600 years, be struck by a mix of direct hits and glancing blows from tropical storms through major hurricanes. For each event, wind speeds, rainfall, and storm surge are simulated, damage is estimated, and the financial consequences to insurance policies and reinsurance treaties are quantified. By examining the full, 10,000-year simulation, an exceedance probability curve can be calculated.

To account for climate change, cat model vendors and users often either adjust event rates to simulate new Year Loss Tables (YLTs) or resample complete years into new YLTs. For example, they might raise the frequency of Category 5 hurricanes that strike Miami or reshuffle the YLT so that the resulting precipitation ranges match the expected trends from climate models, as RMS did in a recent European flood assessment. But this has several drawbacks:

It only represents the events and years that we experience today.
Most importantly, it is impossible to incorporate events and simulated years unlike those we’ve seen before. Cat model events are loosely constrained by what we’ve seen in the past, but just four years ago, Hurricane Harvey announced that the storms of the future, with unprecedented rainfall, might look very different from those of today. In addition, events may cluster and compound differently in the future than they do today, which a reweighted YLT can’t capture.

It doesn’t incorporate broader shifts in the climate.
Event rate adjustments also over-simplify how several perils may change over time. Does solely changing event frequency account for changing event intensity? Wildfire severity, for example, is significantly affected by the available fuel sources. As prolonged drought turns forests into grasslands, we can expect our fire seasons to evolve differently.

Limited perils and regions.
Cat modelers focus their efforts on so-called Tier 1 perils and regions, which tend to drive insurance premiums. But climate change will have much broader effects. According to Swiss Re, 70% of insured losses in 2020 were from secondary peril events such as severe convective storms and wildfires. Climate change will also raise the prominence of previously overlooked perils, such as business interruption claims when a mine can’t operate in extreme temperatures or agricultural losses because laborers can’t work enough hours due to a high heat index. As insurers consider climate risk, they need to think more broadly than the most high-profile events.

Slow to update.
Cat models are impressively intricate, but they also evolve slowly. RMS’ US Flood model was in development for most of my 10+ years at the company, for example; after release, most models were updated only every five years at best. During these long stretches, not only is the climate continually changing, but so is the state of the science. As a result, cat models can be several years behind even when they are first released. This gap is widening as the pace of innovation in climate science accelerates, led by thousands of climate scientists who have learned to take advantage of big data and cloud computing to improve models and tease out the impacts of climate change. For example, the land cover projections that affect wildfire fuel have only appeared in Global Climate Models (GCMs) released in the last year, and there is emerging research on changing hail risk. A new paradigm is needed that is more connected to the scientific community so that insurers can benefit from all of this crucial work.

Where can climate models help?

Compared to cat modeling firms, climate modelers take a fundamentally different approach to understanding the impact of climate change on perils. The core principle is one of nonstationarity: that weather conditions, land cover, sea level, and many other factors are always changing, and they will collectively affect how events form, propagate, and intensify. Many of these changes are captured in GCMs, which are released by institutions around the world and represent the consolidated work of thousands of scientists. Starting with GCMs, as opposed to Regional Climate Models (RCMs), ensures that data and methods are equally applicable in Tier 1 countries and the developing world. (RCMs like CORDEX have their own uses and misuses, as my colleague Dr. Josh Hacker recently discussed.) Occasionally these models have exactly what we need, such as the daily max temperature for our heat metrics, but Jupiter often takes GCM output many steps further. For example, hail probabilities are not explicitly modeled, but we can build our own projections based on downscaled values of the freezing level, Convective Available Potential Energy (CAPE), and other factors. The critical thresholds that cause hail in the future are identical to those of today; it’s the weather conditions themselves that are changing.

Jupiter has moved far beyond what GCMs can offer regarding tropical cyclone risk. Tropical cyclones are poorly represented in climate models; even with dozens of ensembles from dozens of GCMs, there aren’t enough simulations to thoroughly blanket our world’s coastlines with an adequate mixture of possible tropical cyclones with realistic physical characteristics. Instead, Jupiter has layered thousands of our own tropical cyclones — with accompanying wind, rain, and surge — based on the supportive environmental conditions from GCMs. With scientific consensus about changing tropical cyclone wind and rain intensities only recently being established (Knutson et al., 2020), our next release of ClimateScore Global will change the mixture of tropical cyclones in our event set that is applied in each future year and scenario that we model. This gives us 52 distinct event sets: a historical baseline plus three scenarios for each of the 17 future year periods that we model. This fall, we’ll also incorporate emerging research on event clustering and future track migration. These innovations ensure that our customers benefit from the latest that science has to offer.

Even with this event simulation approach, our solution is not, nor is it intended to be, a fully-fledged cat model. Our partners in insurance use us to augment cat models for longer-term portfolio planning, asset management, risk management, and regulatory reporting and to identify where cat models may be “under-cooked” today. We are working to bring climate modeling and cat modeling together in a way that adheres to our core principles: climate-conditioned from the outset, nimble and swift incorporation of scientific innovations, globally applicable methods, strong verification and validation, and transparent documentation. While it may not be as convenient as tweaking event rates and recalculating an EP curve, we believe this to be a more accurate, appropriate, and defensible method to quantify climate risk.

Meghan Purdy is a Senior Product Manager at Jupiter Intelligence. Learn more about Jupiter at jupiterintel.com.

Cat Models: Convenient, Familiar — But Flawed for Projecting Physical Climate Risk was originally published in Jupiter Intelligence on Medium, where people are continuing the conversation by highlighting and responding to this story.

When we think about physical climate risk, what makes a location risky? It’s a question that my colleagues at Jupiter and I often ponder. We’ll execute a climate analysis on millions of locations on behalf of a customer. Our ground-up approach means that, for each location, we produce thousands of metrics: the flooding, hail, extreme heat, and other perils found historically, today, and in the future, and how they might change under various climate scenarios. But before we get swept up in economic loss modeling or aggregating to a portfolio level, we like to take a breath and understand what our climate output is telling us. What’s the story here?

Last spring, as we were preparing to release ClimateScore Global, we spent hours building dashboards and trying to identify the “right” slices of the data to uncover the story. One sticking point emerged quickly: certainly the change element of climate change is what our users care about most, right? But is it relative change, or absolute? Who is at higher heat risk: Dubai, which is expected to pile on more extreme heat days onto its already lofty total; or Paris, where air conditioning is rare, and very hot days (> 35°C/95°F) could be 10x higher by the end of the century?

Obviously, Dubai needs to be flagged for high heat risk. But I’m also worried about Paris, which will have to make a transition from an occasional extremely hot day to weeks of extremes.

Future Shock: Sometimes, It’s All Relative

We can differentiate between Dubai-like risk and Paris-like risk only if we separate the concept of present-day risk and the expected change in that risk. Helpfully, these can be plotted on a chart:

• Bottom-left quadrant. If present-day risk and the level of future change is low, this is labeled as “Good.” For example, heat risk in a place like Edinburgh might be low today, and be expected to remain low in the future.
• Top-right quadrant. If present-day risk and the level of future change is high, this is labeled as “Bad.” A place like Bangkok already experiences high heat, and its risk is climbing quite rapidly in the future.
• Bottom-right quadrant. Present-day risk is high, but it isn’t changing too quickly in the future. These locations are likely to be accustomed to high levels of risk and may have already deployed adaptation measures. For that reason, this quadrant is labeled “Scary but manageable.” For example, Dubai already has many days per year of extreme heat, and while that is rising in the future, its deployment of air conditioning throughout the city will make it more resilient to change.
• Top-left quadrant. Present-day risk is low but is expected to rise very quickly, relatively speaking, in the future. The end state for these locations won’t approach that of the “Bad” and “Scary” quadrants, but it will be high compared to their current risk level. For example, Paris currently experiences about 1.5 days per year of extreme heat (defined as a high temperature exceeding 35°C / 95°F); that’s low enough to make air conditioning uncommon. That number could rise to 20 days per year by the end of the century. While that’s not as high as Bangkok or Dubai, it will be quite disruptive to the population.

Scores Tell You Where to Dig

As climate analytics proliferate, scores are becoming a useful tool to guide non-scientists in the “right” direction when weighing the riskiness of their assets. But users should be aware that scores are inherently a judgment call. An asset’s characteristics, its value, its expected holding time or remaining useful life, and other factors may dramatically affect how consequential the risk actually is and the timeline over which we want to measure how that risk will change. Those factors can only be assessed by looking at the underlying metrics. For that reason, a useful score is one that highlights the riskiest locations so that users know where to dig further. Thus, there are certain characteristics of “good” and “bad” scoring methodologies that users should assess before jumping on board with a tool:

At Jupiter, despite my own product being named ClimateScore Global, we initially held back from publishing scores. We wanted our users to gain confidence in our underlying peril metrics, and we wanted to be careful about asserting our own judgment on the inherent riskiness that our metrics reveal. But as I write today, our own understanding of riskiness has matured just as the skill and the need for scores have emerged. The sixth and newest release of ClimateScore Global, v2.2, now includes “risk,” “change,” and “overall” scores per peril. While the above example only discusses extreme heat days, each peril’s score is based on several underlying metrics per peril, including uncertainty. Over the coming months, I’ll be talking about how each peril brings unique challenges to the problem of quantifying climate riskiness, and I’ll be revealing the insights that emerge from it.

Meghan Purdy is a Senior Product Manager at Jupiter Intelligence. Learn more about Jupiter at jupiterintel.com.

Get the Story Behind the Score was originally published in Jupiter Intelligence on Medium, where people are continuing the conversation by highlighting and responding to this story.

As Jupiter Intelligence described in a new special report in November, regulators worldwide are designing and implementing new standards for quantifying climate risks for investors, insurers, and corporations, with the Task Force on Climate-related Financial Disclosure (TCFD) recommendations becoming the standard.

The Bank of England, widely considered the foremost regulator on systemic impacts from climate change, is requiring stress tests and the use of climate scenarios to apply to a firm’s short- and long-term business outlook. With the United Kingdom recently announcing that most sectors of its economy will be subject to new climate disclosure regulations, it’s more important than ever that companies ready themselves for this new reality.

Where to start? The climate data, of course. It’s a given that this data should be grounded in the latest consensus climate science from the scientific community. But it’s equally important that the right metrics are extracted from climate data to answer the right questions, such that firms’ economic models are fed by metrics that make sense. Only then can companies and organizations adhere to regulations, adjust business strategies as needed, and improve their resilience to the material risks to come.

Transition Risk vs. Physical Risk

To date, climate analysis has largely focused on the quantification of transition risk.

In some ways, transition risk is an easier leap for firms to make. For firms used to forecasting consumer preferences in order to understand demand shifts, or modeling oil shocks to understand supply risk, transition risk is a natural extension. Instead of a shift away from gas-powered vehicles being driven by consumer preferences, the shift may occur due to government vehicle efficiency standards. Or instead of a tariff on steel driving up supply costs, the steel is more expensive because it must be produced with a smaller carbon footprint. Firms typically have teams of analysts creating forecasts and pumping figures into their enterprise risk management framework. Transition risk is no different.

Conversely, physical risk assessment is outside most firms’ wheelhouse and is only recently moving from black to grey swan territory, as I discussed in October.

Certainly, a firm’s engineers and economists are well-versed in how their assets are affected by flooding, high winds, lightning, extreme heat, and other perils. But that is only Step Two in physical climate risk; Step One requires an accurate measurement of that future peril. Climate risk is not immune to the “garbage in = garbage out” concern. No amount of impressive economic modeling in Step Two can make up for foundational peril data that are full of cracks; if a utility prepares for one meter of water instead of two, the wrong resiliency measures will be taken. That foundation can only be solidified if climate scientists are involved to ensure that these peril metrics are both accurate and appropriate.

Defining Accurate, Appropriate, and Transparent Climate Data

Let’s examine what it means for peril data to be accurate. It might be tempting for firms to grab climate data directly from a public source, but analysts need to ensure:

1. The data is assessed for bias
   It’s occasionally seen in climate modeling that bias in the model is larger than the climate signal itself. For example, a model could be underestimating temperatures in the Indian monsoon season by 2°C and thus obscure a temperature rise of 1.5°C. Or a model might be fairly predictive in the mid-latitudes but weak near the equator. While climate scientists have established methods to identify and correct bias, data consumers need to confirm that these methods have been responsibly applied to their chosen data sources.
2. The data is downscaled
   Global Climate Models (GCMs) provide data at a resolution of about 100km (60 miles), but numerous environmental factors determine how that weather tends to manifest at the local level. For example, surface roughness from trees and tall buildings alters wind risk, and local ground cover affects the fuel available to wildfires. Low-resolution data can mask many extremes that may make specific neighborhoods too hot to inhabit or severely flood-prone. (Look out for a post on downscaling authored by Jupiter Intelligence’s head of earth and ocean systems, Josh Hacker.)
3. The data can model extremes
   Firms are often asked to perform climate stress tests, i.e. potential losses under extreme situations. By definition, these events appear very infrequently in climate data. For example, a 100-year flood event should only occur once in 100 years of simulated weather. Climate scientists solve this data availability problem by using statistical methods, blending multiple models, and using ensemble approaches, in which the same GCM provides multiple simulations of the future. Data consumers who are performing stress tests should be particularly attuned to these pitfalls.
4. The data is up to date
   As I discussed last November, given the rapid rate of improvement in climate modeling, it’s crucial to stay attuned to the latest developments in the scientific community. Climate metrics should be based on models from the sixth and most recent phase of the ongoing Coupled Model Intercomparison Project (CMIP).

In addition to accurate data, firms must also consider whether their chosen metrics are appropriate for the task. Imagine an electric utility that needs to understand future hail risk at their photovoltaic plants. They know the vulnerability of their assets to hail, but first they need to know how often they will be exposed to it. Unfortunately, hail isn’t directly output by GCMs, so they might be tempted to use a proxy like Convective Available Potential Energy (CAPE) instead. That method will introduce a lot of uncertainty into the process: few GCMs directly measure CAPE, so the assessment will be prone to bias; and while CAPE approximates the potential for convective storms, those storms don’t always cause hail, so the potential for error with this assessment will be high. It would be more appropriate to directly project hail risk by building a model from several metrics such as CAPE, wind shear, and mixing ratio. That model’s performance against historical hail events should then be assessed to determine its bias and predictive abilities before the utility uses it for decision-making. It’s a complex task, but it’s absolutely necessary for a high quality analysis.

The challenge of data appropriateness also extends to the choice of scenario. As my 2020 post also notes, the inclusion of new Shared Socioeconomic Pathways (SSPs) into CMIP6 is an important development. This helps us understand how certain socioeconomic conditions can lead to radiative forcing, and how that forcing in turn causes the climate to change. Said another way, it gives users a consistent lexicon to tie the actions of our society to consequences for the planet. This is critical for companies conducting scenario analysis and, in particular, for gaining a complete view of the interconnectedness of physical and transition risk.

Finally, climate data must be transparent. Methods should be open so that data consumers can assess the quality of results based on the issues outlined above and, where relevant, convey that to regulators. Transparency also extends to openness about uncertainty: despite the many leaps that scientists have made in projecting climate risk, significant uncertainty remains. Data providers need to be open about the sources of that uncertainty and quantify it wherever possible. That quantification is critical to help firms convey the range of possibles that may occur.

Working with Data Partners

Physical climate risk can be challenging to quantify. Raw climate data is widely and freely available, but it requires significant resources and expertise to extract, de-bias, downscale, and process into appropriate metrics to feed firms’ economic models. And unlike transition risk, firms are unlikely to have this expertise in-house. As firms decide which climate data will underpin their analyses, they must demand transparency from data providers and assess the metrics’ accuracy and appropriateness for their needs. Only with this foundation in place can firms feel confident that their climate journey is starting off on the right foot.

Meghan Purdy is a Senior Product Manager at Jupiter Intelligence. Learn more about Jupiter at jupiterintel.com.

It All Comes Down to the Data was originally published in Jupiter Intelligence on Medium, where people are continuing the conversation by highlighting and responding to this story.

These models, which will power next year’s sixth report from the Intergovernmental Panel on Climate Change, represent seven years of innovation in the climate community. They also shift climate scenarios from Representative Concentration Pathways (RCPs), which considered Earth’s response to varying levels of radiative forcing, to Shared Socioeconomic Pathways (SSPs), which tie the world’s actions to radiative forcing and then to climate change. These SSPs illustrate the interconnectedness of transition and physical risks, and they are a powerful framework for corporations seeking to understand the breadth of climate risks and opportunities that they face.

Introduction

It’s an exciting time in the climate community. With next year’s sixth report from the Intergovernmental Panel on Climate Change inching closer, the global climate models that will support its conclusions, from the sixth Coupled Model Intercomparison Project (CMIP6), are just now releasing results. The CMIP6 data archives are much richer than those of CMIP5, with more models and ensembles contributing their perspective on Earth’s changing climate. And not only do these models represent more than seven years of innovation — a lifetime in such a dynamic field as climate science — but the models also incorporate Shared Socioeconomic Pathways, which provide additional context for how greenhouse gas emission scenarios are tied to socioeconomic conditions globally. These new climate scenarios will help users think more broadly and consistently about the climate crisis.

This October, Jupiter released version 2.0 of ClimateScore Global, which models physical climate risk everywhere in the world. With this release, Jupiter became the first climate data provider to put these new CMIP6 models and scenarios in the hands of users as the bedrock for our models of future flooding, extreme wind, heat, hail, drought, wildfire, and precipitation. So how will CMIP6 make a difference for users? What is to be gained? Here, we discuss the implications of the update and in particular the new Shared Socioeconomic Pathway. Next month, we’ll discuss how this information can be turned into the foundation for disclosures in the TCFD, PRA, and other frameworks.

Updated Science

Given the rapid advances in the field, it is important for climate data users to seek out the latest consensus science from the community. For example, the newest global climate models (GCMs) expect between 0.4°C and 0.9°C of additional warming by the end of the century, even when holding constant Earth’s radiative forcing (a measure of Earth’s ability to trap heat based on greenhouse gas concentrations, land use, and other factors). Even the RCP 2.6 scenario, which was designed under CMIP5 to keep warming well under 2°C by 2100, now results in a mean warming of 2.1°C across CMIP6 models. Users trying to estimate how their business could be affected by climate change — and in particular, those running stress tests — should pay special attention to these effects.

Updated Data

The newest GCMs also benefit from seven further years of data. CMIP6 models commence simulations in 2014, while CMIP5 start from 2007. These seven additional years of emissions, land use, sea level, and more have led to different models for radiative forcing trajectories. In particular, achieving an RCP 2.6 scenario will require stronger mitigation efforts in later years, given the weaker-than-expected efforts to limit emissions over the past 13 years. On the other hand, scenarios resulting in higher warming are more likely. The updated models provide much more realistic estimates for how climate change might play out, which helps users make more practical business decisions today and provide more realistic estimates to their shareholders and regulators.

Updated Scenarios

The CMIP effort standardizes GCMs in ways that make their results easier to compare. It includes a few practical features (e.g., data is available from a central repository and variable names are consistent), and, critically, it specifies scenarios for how the remainder of this century and beyond could play out.

Under CMIP5, these scenarios were specified as Representative Concentration Pathways: how the “radiative forcing” would change over time. Scenarios such as RCP 8.5 are so named because they result in an additional 8.5 Watts per square meter of radiative forcing by the end of the century. Each GCM took the “pathway” of rising radiative forcing as an input to their simulations; the outputs were the resulting changes in temperature, precipitation, and other factors. This helped the GCMs speak the same language: researchers could view several opinions for how the same amount of forcing could lead to different air temperatures, ice melting, precipitation patterns, and more around the world.

A key feature of the CMIP6 models is the inclusion of new Shared Socioeconomic Pathways (SSPs). In CMIP6, we can see how certain socioeconomic conditions can be linked with radiative forcing, and how that forcing in turn causes the climate to change. In CMIP5, the radiative forcing trajectory was taken as a given, and the only variable was the degree to which the climate changed. This has given CMIP6 models much more flexibility to explore how humans might affect the climate: how much energy is generated from fossil fuels vs. renewable sources? How much do we preserve our rainforests? How well do we succeed at mitigating climate change or are our efforts focused on adapting to it? How do our institutions and populations respond, and how does this differ between developing and developed countries? These factors were distilled into five socioeconomic pathways.

With the SSPs determined, modelers then had to determine which SSPs could lead to future radiative forcing scenarios: the old RCPs. By combining economic, social, and simplified climate models, researchers evaluated what conditions could lead to emissions that are consistent with particular physical climate change scenarios. For example, SSP5 — in which the world continues to burn fossil fuels at a rapid rate — was found to be the only scenario that could result in RCP8.5 levels of radiative forcing and a warming of more than 4°C by 2100; all other scenarios lead to less forcing and less warming. Conversely, SSP1 — in which the world prioritizes sustainability — could result in as little as 1.9 or 2.6 Watts per square meter of forcing and less than 2°C of warming. To express these relationships, pairs of socioeconomic pathways and representative concentration pathways are written together: SSP5–8.5, for example.

Towards a Combined View of Physical and Transition Risk

The new SSP-RCP framework is incredibly powerful because it gives users a consistent language to tie the actions of the world to the consequences for the planet. This is critical for companies conducting scenario analysis. Under the old RCP framework, users conducting stress tests were guided to the most extreme scenario (RCP 8.5) simply because it stressed their balance sheet the most. But with SSP-RCP pairings, users can create more realistic situations: for example, they can use the SSP to determine how their firm is affected by conforming business practices to a < 2°C world and then couple that information with the physical risk of a world that will eventually be 2°C warmer. In other words, it helps users build scenarios that harmonize varying degrees of transition risk with the resulting physical risk. In the past, users discussed transition risk scenarios like “hothouse world” and “disorderly transition,” and the physical risk was linked purely to the amounts of radiative forcing. But now, users can see the collective impact on their balance sheet of both sources of risk: how sharper reductions in carbon usage (higher transition risk) lead to lower radiative forcing (lower physical risk) and vice versa.

The world faces complex choices as it seeks to blunt the impact of climate change while also allowing our economies to thrive. With CMIP6 and the new SSP-RCP framework, these trade-offs are laid out alongside each other for the first time. Jupiter is excited to incorporate these new models as a key component of our climate projections and put years of scientific research in the hands of our users so that they can understand these risks.

Meghan Purdy is a Senior Product Manager at Jupiter Intelligence. Learn more about Jupiter at jupiterintel.com.

Jupiter and CMIP6: Unlocking the Next Generation of Global Climate Models was originally published in Jupiter Intelligence on Medium, where people are continuing the conversation by highlighting and responding to this story.