Dr. Michael Ferrari has spent the last two decades working with — and in one notable instance, against — some of the world’s largest organizations to better understand the intersection of climate science and the economy. Reflecting his unique ability to see both sides of this particular equation, Dr. Ferrari joined Climate Alpha in January as both its Chief Scientific Officer and Chief Commercial Officer, in addition to his prior seat on the company’s board of directors.
In these roles, he works closely with Climate Alpha’s technology and growth teams to further refine how state-of-the-art climate models inform the risk-adjusted valuations of its Climate Price™, the scenarios of its patent-pending Scenario Forecaster™, and Alpha Finder™ location analysis tool.
Prior to joining Climate Alpha, Dr. Ferrari was the Chief Science Officer and Chief Data Scientist at Engine No. 1, the activist investment firm best known for successfully campaigning to elect three new members to Exxon-Mobil’s board of directors in 2021.
Before that, he had built and led data science teams at multiple companies, including Syngenta, Point72 Asset Management, IBM, Mars, and the Coca-Cola Company, as well as with several startups in the technology, fintech, and commodity sectors. He has also been an affiliate scientist at the National Center for Atmospheric Research and the MIT Media Lab, and is currently senior fellow at The Wharton School.
We sat down with Michael to ask why this is the year adaptation takes center stage in climate efforts, how both climate models and Climate Alpha’s models are constantly improving, and what a mature version of the firm looks like under his watch.
Why join Climate Alpha now? And why is 2023 the year adaptation comes to the fore in the climate conversation?
Honestly, every year is the year. Climate Alpha is built around issues I’ve been working on for 20 years — and for 19 of them, I’ve pushed this idea just based on conversations with prospects and partners and clients; only now is the market ready for adaptation, both from a physical readiness perspective and potential financial opportunities. We’re finally at a point where we have enough data and a few cases to demonstrate there’s a way to do this and still generate returns for investors whether they have a short-, medium-, or long-term horizon.
Look at California, where they recently received a year’s worth of rainfall in only a few days. Every time we have an episode like this, it’s just further confirmation that we need to adapt and evolve — otherwise, we’re going to keep having billion-dollar disasters. If you look at the charts of these disasters, the slope just keeps going up to the right. Regardless of what’s happening on the climate side, there’s a human component to infrastructure, and if we can help provide some insight on where to deploy capital, we might be able to put a lid on the number of billion-dollar disasters, and that’s a step in the right direction.
Q: How do you frame the climate change argument to clients in hopes of steering them away from multi-billion dollar unforced errors?
It doesn’t necessarily have to be a climate argument. A lot of these things just make good financial sense. Obviously, if you buy into the idea that climate will continue to be volatile and unpredictable, scenarios help us understand the range of possibilities and entire distribution. But we’re not insisting every event is a climate change trigger. What’s driving valuations on a day-to-day basis are the factors that have always driven valuations, which is why this platform makes sense for anyone who considers themselves patient capital.
Q. What are the biggest shortcomings in the scientific field when it comes to tying climate models back to physical risk?
The market hasn’t cared about these things in the past. If you consider traditional risk modeling, they’re looking at what’s happened over decades — cycles of 10, 20, 30 years — and try to tie the climate trigger to the response, typically damages or losses associated with those events. And what happens is you have an event, the market reacts, and then they move on. But the broader financial trigger isn’t rooted in these singular events. They might be correlated, but there’s no causal link. If you build your approach based on misconstrued data, you’re going to over-fit models, make false assumptions, and have “right” answers for the wrong reasons. And that’s dangerous.
Our approach is to take the baseline valuations based on everything else we know is important — job growth, quality-of-living, things that matter on a day-to-day basis — and then layering in the climate. We’re not building climate models; we’re building future scenarios in which climate is an amplifier. That really resonates with our customers, and I think it’s the most exciting thing about our platform.
Q. If Climate Alpha isn’t building its own models, then what does it contribute in terms of understanding geographic and economic risks?
Think of a hurricane — when one starts rolling across the Atlantic, it has what’s called the “cone of uncertainty.” At the start, it could make landfall anywhere from South Carolina to the Gulf of Mexico. But after each run of each model, those error bands start to decrease and the cone gets a little tighter. So, within 48- or 72 hours, we have a very good idea of where it’s going to go — maybe not down to the house where the eye is going to make landfall, but close.
We look at climate models the same way. There’s no need to build our own — there are groups that all they do all day, every day, is build and refine biophysical climate models. We’ll take them as an ensemble and use them as our guide. Then, we can start to ask, “Which patterns in this region are repeatable?” and weight them accordingly. No one who builds climate models will tell you their model is perfect; they’re guides, and as they improve in accuracy, we can start to zero in on what’s happening in two-to-three years instead of decades-long scales.
Q. And how does Climate Alpha improve in turn?
It’s a never-ending cycle of iteration, data acquisition, and evaluation. As long as we ingest data in the right formats, these simulations can run on their own in the background, and eventually we’ll start to refine and focus on areas where they’re performing well and worse. Then we’ll know which data sets need improvement, where the gaps are, and what methods are appropriate. The more data we can ingest, cleanse, refine, and incorporate into our platform, the more it will increase our confidence in the models. But we’ll still need human eyes to check it. AI can help synthesize and distill these volumes of data, but ultimately we’ll still need to have a combination of human and machine.
Q. Climate Alpha is a startup; how do you envision a fully mature version of both the company and platform?
Whenever you have a new product or business, everyone wants to know what your TAM [Total Addressable Market] is right away. And I think we’re in an area where we don’t even know what the TAM is, because it’s so vast. After all, the total value of global land, property, and real estate assets is estimated to be more than $350 trillion. Climate Alpha can be one thing for investors, but it can also be a tool for governments, or a tool for homebuyers looking for a climate component alongside schools, walkability, and quality of life. We shouldn’t discount the importance of this for governments or NGOs when they’re making policy decisions on where to deploy capital over 10-, 20-, 30-, or even 40-year time horizon. We may have started with climate — it’s in the name — but this could ultimately become a guide for almost any financial decision someone could make.