How can we evaluate the costs and benefits of conservation to guide decision making towards the best outcomes for nature conservation while meeting human needs?
Prof Andrew Balmford’s research work has received international attention, shaping discussions about how to best manage landscapes. Scientific American named him one of the world’s top 50 visionaries building a better world.
He is Professor of Conservation Science at the Department of Zoology at the University of Cambridge and a Fellow of the Royal Society.
Interview by James Miller - student environmental activist, film-maker. Text lightly edited for clarity.
Tell me about the origins of 4C
I had been thinking for a long time that conservation was missing a trick by not helping create a more robust approach to estimating the carbon benefits of safeguarding and restoring natural habitats. We’ve been keeping offsetting at arm’s length, because it’s so important to cut down on emissions.
Then this opportunity came around [to explore that in more detail] when the University announced that it wanted to offset its flights, and set up a working group to figure out how to best go about doing it. I put in a proposal to fund Tom Swinfield to develop methods for that, and to bring a lot of different projects on one page with directly comparable metrics, really for the first time. We spent more than a year trying to devise the best methods we could to characterise the carbon but also biodiversity, livelihoods and social justice impacts of projects. Sometimes that involved improving methods that were already out there in the literature, but also where necessary inventing our own methods because we didn’t think what was out there was good enough.
In the course of that, we began talking with carbon economists at the University of Exeter and LSE, and with computer scientists like Anil and Keshav. They were very excited, and in due course that led to them raising the funding for 4C and bringing the possibility to roll this out more widely, which is really thrilling.
You mentioned some of the methodologies you’ve been developing. What areas do you think you’ve developed really well?
I think we’ve done a good job of bringing in state-of-the-art additionality methods. [When doing these projects, it’s really important to know that the money spent on them is conferring additional value – in some cases it might be that a patch of forest wasn’t going to be cut down anyway, for example]. The current system of assessing additionality in order to issue carbon credits is largely narrative-based - ‘we think our area is a bit like how this other plot of land was ten years ago, so let’s just assume that site's trajectory over the past ten years describes what would have happened to our project’. There are far better methods out there involving the creation of statistically-derived counterfactuals. In these we identify places which are demonstrably similar to the project area in their vulnerability to threats, and have shown similar deforestation tracks up until the point of the project. We then look to see if there’s any subsequent divergence in forest cover between those [matched ‘counterfactual’ samples and the project area. This is all done in an automated fashion using algorithms applied to satellite imagery].
What methodologies are still in development?
The area where we’ve done the most conceptual work, which we are still yet to finish but which I think is really exciting, is how to deal with the permanence issue. We know that nature-based solutions are not permanent - carbon comes out of trees and soil as well as going into them. Until that aspect is dealt with quantitatively, robustly and transparently, there are two dangers: firstly, that you can read a project document that says ‘trust us, our intervention is going to be pretty permanent, treat it as a permanent ton’ [and that doesn’t end up being the truth]. And secondly, the reaction to that from some hard-nosed folk is going to be ‘We don’t really believe you, so instead we’re going to pay a lot more for a technological solution that sinks it into a hole in the ground because we know that’s permanent’. Consequently, there’s a lot of missed opportunity – there is movement of demand towards solutions that are more costly and don’t bring a lot of the biodiversity and livelihood co-benefits that nature-based solutions can.
The key insight there is to say ‘OK, this carbon credit is impermanent, but until [all that additionally sequestered carbon is] released back into the atmosphere, it is still conferring a benefit to people. How can we estimate what that benefit is?’ Carbon economics thinking provides a framework for that with via the idea of the Social Cost of Carbon, which is our best estimate of the present day value of the cost that a ton of carbon put into the atmosphere now imposes through time on subsequent generations. If we can estimate when the CO2 that has gone into our credit is going to be released, we can estimate the cost of that release, and deduct it from the benefit that we are gaining now [through the impermanent storage]. The net is the value of the solution, accounting for its impermanence.
That’s easy if you know exactly when the CO2 is going to be lost. Of course we don’t, but what we can do instead is build probabilistic models of when those emissions are likely to happen based on what we see happening to forests when projects end. If we can construct a really conservative estimate of that, we limit the risks by putting forward a bad but plausible scenario. Buyers can buy additional credits so they can be confident that the overall value of the impermanent credits they’re buying is the same as that of a fully permanent credit. Our conservatism enhances the credibility of the scheme, with buyers being assured they are buying the equivalent of a permanent ton.
You say you’re coming up with a really conservative estimate, which is great for building trust in the system, but is there also a risk of inflating the price of nature-based credits so much that they fail to be competitive with other forms of offset?
Yes, that is a risk. On the one hand, when you run the numbers, the upfront cost tends to be so much lower for nature-based solutions that even when you include very conservative adjustments for their impermanence, they’re still much cheaper than the technological solutions. But if you were to be completely conservative, and say ‘My guess is that this is going to go up in smoke the instant that a project stops’, that would probably devalue the sequestered carbon too much – there needs to be enough money reaching the project otherwise the opportunity cost of undertaking it won’t be met. There’s definitely a tradeoff to navigate there.
Is there any way of improving that thinking? Or are those conservative probabilistic models where this line of development will have to end, leaving in a degree of uncertainty?
In fact, there’s some really clever stuff that Prof Keshav has brought to the project through his understanding of queuing theory for this. For a project, if, in due course, it turns out that less of the past carbon gain went up in smoke than expected under our conservative release schedule, then you can compensate for that by issuing another bit of credit for the better-than-anticipated performance of the project to date. And what that does, uniquely, is it provides an incentive after a credit has been issued for its continued safeguarding by the people responsible for it. At the moment the most common way of dealing with impermanence offers no such thing — it pays some carbon into a so-called buffer pool against which can be drawn down if releases happen, but there’s no benefit to come back to future generations if they you look after that patch of forest, say, in perpetuity. Our proposed system reduces the intergenerational inequity of that situation where you place a burden on future generations not to burn carbon somebody else has paid for. We’re really pleased with this correction idea, but there’s still some algorithm work to do.
Are there areas where you still have much further to go with these methodologies?
Perhaps the biggest gap is about the risk of leakage, which is where foregone production resulting from projects shifts somewhere else, where it has its own carbon impact. We currently adopt the standard in the conservation business which is constructing a ‘leakage zone’ [of a few km radius], around the project. We’ll see whether deforestation accelerates there, and if that’s the case we’ll deduct that from what we estimate the additionality within the project site to be.
That’s sort of OK, and we will probably calculate it as well as anyone else in this space because we won’t just look at the leakage zone. Instead we will match the leakage zone to its own counterfactual (as we do for the project), and estimate leakage from the difference between what happens in the leakage zone and in its own counterfactual. But in an increasingly interconnected world, the chance that leakage will necessarily be displaced just to within 5 or 10km of a project becomes less and less likely. Particularly where it’s a readily-traded commodity it’s possible that the displaced production pops up on the other side of the world. Trying to detect and adjust for that is very difficult, but we have some lines of thinking that we’re developing.
I can imagine it’s almost impossible to track where production is displaced to, and the carbon impact therefore associated. Do you think that’s an insurmountable barrier?
I don’t know, we need to do some more thinking about it. There’s a fantastic leakage expert moving to Cambridge at the start of next year, Rachael Garrett, a professor of Geography. She works on these problems and understands them properly — better than most people in the world — so putting those ideas to her and seeing what she thinks will be fun.
Addressing this problem goes beyond offsetting. I don’t think that conservation very often considers opportunity cost and displaced production, whether that’s with deforestation, habitat restoration or creating a protected area. We congratulate ourselves on impact by looking just locally, and we don’t consider that risk of offshore displacement. However, if you introduce a more robust approach to factoring-in the risk of leakage it would incentivise interventions to concern themselves not just with retaining and restoring natural value in a place but also ensuring that local production is not diminished. It would make projects think much more robustly about local livelihoods and larger-scale consequences of local actions.
It's really exciting to see that not only will this work streamline the process of getting projects up and running, but might also help to change the way that people undertake those projects for the better.
I hope so. If many more of those projects were to take those things seriously and integrate them into the heart of what they’re doing, right from the start, we think that would increase the additionality of the projects (because they’re better addressing the driver that is causing the problem in the first place), it would reduce leakage, it would help with livelihoods, and it would serve to increase permanence — if production hasn’t gone down, or people are connected to markets where they can get better prices for their goods, you may have helped alleviate the driver over the long run.
Is there a risk of incentivising projects that are easy wins for carbon but might be detrimental to nature?
Yes, I think that’s right. At the moment we don’t have anything like a mature market for biodiversity offsets, although there’s lots of noises that that might emerge in the next few years, which would be great. What we’re trying to do for now is to develop a globally robust quantitative metric of biodiversity impacts of a land use change project, like a habitat restoration or avoided deforestation project. In the short term, it might be that we express the results in discrete categories that people can understand (e.g. this is a class A project for biodiversity, this is a B, this is a C) and then interested buyers can filter projects based on how important biodiversity is to them. Those in category A might command a higher price than those B rated projects. We’ll have the same categorical ratings for social justice and local livelihoods.
Why is the timing right for Cambridge to develop solutions?
I think now is the moment because the wider world has woken up to the climate crisis, albeit half-way through the 12th hour, and is waking up quite quickly to the extinction crisis — so there is real recognition that these things have to happen, and rapidly growing commitment to do something about it. And within that space, avoided deforestation is one of the most extraordinary opportunities to achieve things very quickly and cost effectively in a way that meets multiple objectives at the same time. There is nothing else that comes close to it.
And yet there is this central problem of the lack of credibility (or perceived lack of credibility) of avoided deforestation as a climate change strategy. So we have both a real opportunity and a real problem to be tackled, while it just so happens that a lot of the pieces are in place here and elsewhere to do something about it right now - along with some very motivated people.
I guess that shows the importance of a place like the Cambridge Conservation Initiative that brings so many different researchers from different backgrounds together.
Yes, absolutely. A lot of people fought hard for having something like CCI. As a biologist, it’s clear in conservation that biology can only provide some of the answers and some of the understanding. It’s vital we work with colleagues across a whole load of other disciplines, and of course with conservation practitioners, and with business. That’s why this place was founded!