iRisk Conference: Making Decisions in an Uncertain World

COVID-19, war, and economic crisis. In this period of great uncertainty, decisions must still be made. How do agents and policymakers deal with uncertainty when pursuing individual goals and social objectives? Recognizing the increasing importance of linking theory and practice, the iRisk conference, held on 8 July 2022, provided a platform for dialogues between industry, policymaking, and academic experts. The conference was held at the IÉSEG School of Management and organized in partnership with the AXA Research Fund.

This event marks the opening of the IÉSEG’s Risk Research Center (iRisk) and is the latest in the IÉSEG’s series of “inspiring conferences” targeting academics, decision-makers, and practitioners. After a welcome address from Caroline Roussel of the IÉSEG School of Management, panelists were introduced by moderator Thibault Lieurade, Business Editor for The Conversation. The panel included Gilles Moëc, AXA Group Chief Economist and AXA IM Head of Research; Vicky Pollard, Acting Head of the “Foresight, Economic Analysis & Modelling” unit at the Directorate-General for Climate Action of the European Commission; Christian Gollier, Professor of Economics at the Toulouse School of Economics, and Mathias Dewatripont, Professor of Economics at the Université Libre de Bruxelles.

The conference is available on replay below.

Key takeaways

• The European Commission’s “Fit for 55” package includes legislation on thermal cars and a revision of the EU Emissions Trading System (ETS). Uncertainty surrounding war and energy markets has boosted the political will to get things moving, yet, there is uncertainty about what a fully decarbonized future will look like.
• Uncertainties about the COVID-19 virus, health measures, and closures are closely tied to political economy issues. By underestimating uncertainties, countries such as France made suboptimal decisions about vaccination strategies.
• Belief in the market and central banks, as well as massive fiscal support, minimized macro-uncertainty during the pandemic. The current situation is characterized by empirical and theoretical uncertainty, and the focus is on providing internally consistent scenarios rather than “predicting the future.”
• Integrated models used by the Intergovernmental Panel on Climate Change (IPCC) help set policy frameworks, but decision-makers need to better understand uncertainty and feed this into decision-making. Many countries do not include risk as an ingredient in policy evaluation.

Decision-making under risk and uncertainty

Thibault Lieurade (TL): Much has happened at the European Commission in recent weeks, such as the 2035 ban on thermal cars. What elements of uncertainty were taken into account in these decisions?

Vicky Pollard (VP): The last weeks have been a mid-point in the co-decision on a big package of legislation, unofficially termed the “Fit for 55” package. This package comes as part of the implementation of the European Green Deal, under which we have a climate neutrality target of net-zero CO2 emissions by 2050. Within the 55 package is legislation on cars, where the aim is to get to zero CO2 emissions by 2035. There is more agreement than we would have expected a year ago, which is a sign of the political will to get things moving. This has been helped by the current situation—we didn’t expect to be in a situation of war or to have such uncertainty in the energy markets.

The package also includes a revision of the EU ETS to extend carbon pricing to other sectors to increase the incentive action. This is a key piece of legislation because it covers our largest emitters and provides an element of certainty, with a declining cap over time that goes down to zero for those emitters and allows the market to set a price for carbon—this is a very important signal. There is a proposal for a new ETS for buildings and transport, and while the Parliament and Council agree that something should go ahead, there is uncertainty about what happens to revenues. These policies create certainty in terms of emission reduction pathways, but they also create important revenues, the use of which are key for transition in terms of investment in new technologies and support for more vulnerable people, such as large-scale insulation of houses. This means the carbon price can act without putting people under pressure, especially those who have the least capacity to reduce their exposure to the price.

TL: How do you react to this, Christian?

Christian Gollier (CG): Even in recent months, there has been a lack of will of the elected people to support actions that could be disliked by electors in Europe. The impressive 55 package is politically courageous, and the objective of zero CO2 emissions by 2035 is an incredible challenge. We know that action will trigger sacrifices for the people, but politicians don’t want to appear as following actions that could impose this sacrifice. So, there emerges a false “utopia” of happy energy transmission, and we rarely face the political issues associated with that.

Second, climate change—where risk and uncertainty are everywhere—is the best playground for decision theorists. What should we do in the face of this enormous uncertainty in climate change intensity, sensitivity, and technology? We don’t know what the economy will look like fully decarbonized 28 years from now. Faced with this uncertainty, the political situation becomes interesting—politicians want to provide a clear vision with no uncertainty, but how we get there is another question.

For example, there is a consensus to ban thermal engine cars by 2035 in Europe, but nobody knows whether it will be politically acceptable 13 years from now. What will be the price of electric cars or the price of gasoline? No one knows. If we stop purchasing oil and natural gas from oil-rich countries such as Saudi Arabia and Russia, these countries will quickly sell their reserve, and the price will go down. What will the consumer’s attitude be when they are forced to purchase electric cars at a time when gas prices are as low as 1 euro/liter at the pump?

TL: Mathias, let’s hear about another crisis. You were part of a group of experts monitoring the evolution of the COVID-19 crisis for the Belgian government. What did you have no answer to when a decision was necessary, and what lessons did you learn from this period? 

Mathias Dewatripont (MD): To build on what Christian said, people don’t like uncertainty. They want perspectives, and ideally rosy ones. I think one of the challenges of experts and politicians is to help people understand that the future is uncertain. For COVID, we first thought no mask was needed at a 1.5-meter distance, but 3 months later learned more and had to revise things. With these kinds of uncertainties, people start to become worried and conspiracist.

Even with “smart” lockdowns, there are winners and losers, and that’s when political economics comes into it. For example, shops could remain open and hairdressers not, which leads to a problem termed “prisoners dilemma” by economists: you’re not against your competitors or other sectors being closed, but you want to be open so you can take advantage of that. There are also always those who play the good guys for pure pandering reasons, such as those who say, “I’m in favor of hairdressers”, and are happy to let the Minister of Health be the bad guy.

Economists have also been important. Raj Shetty and others at Harvard reported that the health–economy trade-off everyone is talking about is not a true trade-off. However, this kind of robust and logical idea is hard to get through in the media, which instead gives voice to people who are losing out. So it’s important to give financial compensation to those people.

TL: Mathias, tell us more about your article in Le Monde about the vaccine strategy, which you said could have been better in France.

MD: I think everyone agrees that vaccine development has been a success. The US’s warp-speed operation gave a good example of industrial policy, and France also successfully adopted a smart strategy in terms of industrial policy by diversification and dual sourcing. But vaccines don’t save lives, vaccinations do, and we needed to convince people to get vaccinated. One good strategy was that the vaccine was first made available to vulnerable people and volunteers, as there weren’t enough for everyone. People realized that the vaccines had few adverse effects, and the number of people against vaccines dropped.

However, as in most countries, France said that vaccination would be optional; moreover, they said we would reach herd immunity. This was not the case, as they later learned when came the alpha, delta, and omicron variants. Governments had to start incentivizing vaccination, such as adopting the vaccine pass, and people felt betrayed. It would have been smarter to say, “we will start with vulnerable people and volunteers, and then see what happens.” They underestimated the uncertainty and the ability of the virus to become more contagious.

TL: Gilles, there’s been lots of mistakes in forecasting. We didn’t realize that the rebound of growth would be so quick or that inflation would be so strong and long-lasting. What have you learned from this crisis?

Gilles Moëc (GM): The developed economies managed to minimize uncertainty in an incredible manner at the beginning of the pandemic from a financial and economic point of view. On the financial market, there are two possible reactions to massive uncertainty such as the pandemic: do nothing at all, or sell everything remotely risky, which is what happened in March 2020. The only thing that saved us was the absolute belief in the market and that central banks could save our skin, and they did. Another beacon of stability was massive fiscal support. Incredibly, there was no internal debate on this during the pandemic, and this reduced the quantum of macro-uncertainty in the system to a large extent.

In the current environment, we have to deal with two sources of uncertainty. One is empirical uncertainty. We didn’t know COVID would happen and often had to tell our clients and stakeholders, “I don’t know”. A lot of what’s happening right now in the business cycle is a legacy of the pandemic, and the synchronized reopening across the world has triggered a lot of the current inflation spike. We’re also still dependent on whether China will see another flare-up in the next few months because this would further impact the global supply.

The other source is a creeping theoretical uncertainty. In 2020, there was a massive consensus that we knew what governments and central banks would do, and today we don’t know. Clearly, central banks are less happy to maintain an accommodative monetary policy, and we don’t know how far they’re willing to go to kill inflation.

Today, we’ve moved away from the idea that we can predict the future and the focus is on scenarios. My job as a private sector forecaster is to provide clients and decision-makers with a set of reasonable, internally consistent scenarios. We don’t know what’s going to happen—whether Russia will turn the tap off tomorrow, for instance—but can at least present a set of logical sequenced consequences of a number of events that will help policymakers and investors to prepare.

TL: You rely on these scenarios in policy-making, Vicky. How would you react to this?

VP: Yes, there have been discussions about the uncertainty of the green transition. We’re sure that climate change will happen and that the policy we put in place is a necessity. Now, we’re looking at the implications of that transition for different sectors. The European Commission invests a lot in integrated assessment models used by the IPCC that help set our policy framework, but also the policy for everybody else globally. This is also important in terms of getting consensus on the level of ambition and types of action. We’re doing a lot of work on many policies, including those in the 55 packages, for which we rely on a suite of models. Things move quickly, so we don’t always have time to update the baselines, but we try to have a consistent baseline and use scenarios that refer to a common baseline across all the policy measures so that we can at least show a direction and severity of possible impacts.

Unfortunately, climate change will keep leading to more extreme impacts. The IPCC’s work shows that we must reduce emissions substantially now and take ambitious action before 2030. Delaying will only mean greater challenges later on, and effects are seen not only in vulnerable parts of the world but also in Europe, as we’ve seen in Belgium and Germany with the floods and forest fires in France. This also impacts the economy in terms of lost productive capacity and public finances as governments become the last resort.

TL: Christian, you wrote Le climat après la fin du mois, published in 2019. Is the population’s reaction to policy-making the biggest point of uncertainty?

CG: Yes. It’s an exciting problem that everybody wants to solve, but often people point the finger at others: “it’s not the consumer, it’s the producer”, “it’s not the producer, it’s bankers”, “it’s not the US it’s China”, and so on. We’ve been aware of the climate change problem since at least the 1992’s Earth Summit conference in Rio, at which point there were 22 gigatons of CO2 emitted per year. Today, we’re emitting 36 gigatons, so we’ve almost doubled emissions rather than reducing them.

Mathias, Vikki, and I pointed to the same problem, which is that politicians are there to provide some certainty. Politicians have been helped by experts in climate change and COVID-19, which conveyed the idea that their models could predict what will happen. However, none of the models in March 2020 included uncertainty; even though the uncertainty was enormous—the R0, contagiousness, mortality—we knew almost nothing. And yet experts provided recommendations to politicians, assuming that there was no uncertainty. As economists, decision-makers, and theorists, we should try to communicate to policymakers that we are “not sure".

TL: Turning to another crisis, Mathias, do you think the ability to deal with uncertainty has improved since the financial crisis of 2008?

MD: Definitely. The 2008 crisis came after years of over-optimism about the resilience of the financial markets with very little capital in it and no real regulation of liquidity, so the system was working on very few buffers and with macro-imbalances, which is typically when a crisis starts. Gilles mentioned that it’s hard today to make predictions because of the uncertainty about policy. The 2008 financial crisis started in a weak system but was triggered by an unanticipated change in policy. Then began this problem of “strategic complementarities”—if everybody runs to the bank and you risk not being insured, you run to the bank too. Bailouts were massive, and the system clearly needed to be made more resilient, with more capital, better liquidity, and macro-prudential regulation. So far, it has been pretty resilient despite the serious shocks of COVID-19 and Ukraine, but there are potential clouds on the horizon with competition from FinTech and so on.

TL: At that time, there was some criticism about economists sticking too much to models, which meant they could not see uncertain things. Do you think there is still this model fetishism in the financial world?

GM: Many private and public sector economists have been quite humble in the face of the crisis in 2020/2021, and now Ukraine, and we had to innovate in the kind of data we use. I remember the scramble at the beginning of the pandemic for real-time indicators because the usual monthly indicators would come in too late and too infrequently. The Google Mobility Index helped us to gauge the quantum of pain in the system at the beginning, and INSEE provided relatively good real-time estimates of the shock.

I do see a risk in this huge quantum of data and the huge computation capacity that has become available; this may make it easy to design what is called blind models, which basically involves picking the model that provides the best fit to the data at the current moment. This is dangerous because if you estimate 1000 models, one will obviously work, but usually by chance. Without theoretical sequencing or a chain of inferences, you may come up with models that work for 3 or 6 months. It looks quantified, so it will convince many people, but there needs to be a story behind this. Storytelling to me remains central, and this is what economic theory provides.

TL: Christian, do you share this take on models and the possible danger of this huge amount of data?

CG: Dependence on one single model to explain the world is a dead end. France is the only country where risk is an ingredient in policy evaluation. Almost everywhere else, we assume pure risk neutrality, and the estimated flow of expected net benefits is used to determine whether a public investment is good or bad. In France, I and others convinced the state to introduce a discounting system where we give a bonus to investment projects with an insurance benefit, such as investing in ICUs. However, we have yet to convince people that risk and uncertainty are important when making policy recommendations, particularly in dimensions in which risks are so central, such as climate change and COVID-19.

TL: Vicky, do you feel that risk is underestimated in policymaking?

VP: We, of course, perform risk assessment, but it’s probably not the right sort of risk on the right scale—it’s challenging to know where exactly we should be focusing. That’s why the work that’s happening in the community here is important to us. This discussion helps me to draw on experience from very a much broader area than I would normally look at. We also need to get better at feeding this experience into our decision-making.

Q&A Session

Q: Can more be done about the distrust in science we’re seeing?

CG: This is increasingly important to understand and is also a political issue. In the US, the proportion of democrats who believe we should do more for climate change is four times larger than the proportion among republicans. In Europe, we don’t see this politically motivated bias in belief, yet this is probably not true for COVID-19 and vaccine efficiency. We need to talk with anthropologists, psychologists, and political scientists to try to understand this, as democracy is at stake.

Q: Could it be that people are used to science delivering goods that make life better? Maybe they reject science when it asks them to do something against their interests

CG: Behind this issue is this concept of rational expectation. We know that with optimal expectation, people bias their beliefs to better justify their way of life, and should probably look at that too.

MD: Typically, vaccine-hesitant people are less educated, less wealthy, and have less trust in the government. I think there’s more to this than just a distrust of science, and it may be a symptom of a different problem. If we dig a bit deeper into what’s wrong with our societies and manage to fix that, I think this would also fix vaccine hesitancy.

VP: The political question concerning the split in the US is a question of where people feel they belong, and so we need to address people in terms of their own values. Some of the practical reactions in Europe we have now include the EU Climate Pact, which focuses on climate ambassadors getting a broader range of people to explain the issue. The IPCC has similar initiatives to present information to a wider range of people, and not necessarily by government experts.

Q: How does the European Commission persuade the public to support policies, especially when confronted with the potential backlash if too much sacrifice is demanded?

VP: We’re working on better describing the impacts of our policies at member state and regional levels, as well as for different types of households and income groups. It’s also politically part of the debate: when the proposals are being debated by the Commission, our work goes to the body of commissioners who are appointed by member states. There’s been a big green vote in the Parliament, which is very vocal about maintaining the level of ambition. A less visible element is that there’s a lot of lobbying, which might also reflect the views of organized parts of society. I think politicians were really concerned about the impacts on vulnerable people, costs of living, how people perceive the transition, and the restrictions on lifestyle—the Social Climate Fund is a reaction to that and is an EU-level instrument that ensures all member states have access to funds to address the needs of the vulnerable.

August 2022