Radical uncertainty and long-term investment decision making
This story is based on concepts developed in a book I recently directed and co-authored about corporate decarbonisation (in French). In this series of articles, we examine the consequences of climate change on financial decision-making.
We are only slowly awakening to the dramatic cascade effects climate change is having on our economies. Taking a step back from the now daily inventory of catastrophes and rampant transformations that unfold under our eyes, one of the key consequences is that climate change has compromised our ability to project ourselves in the future.
The ability to forecast and ponder different scenarios, to allocate probabilities to potential reasonable outcomes is a key economic tool that has allowed us to plan big projects, from infrastructures to space stations, from factories to global supply chains. Key to this reasoning chain is the ability to use past statistical data to infer future probabilities.
However, climate change disrupts this ability as past data is losing its representativeness. When once-in-a-1000 years events occur multiple times in a few years, it means that the statistical data on which your probability of occurrence estimate is built is no longer a meaningful indicator. What is true of tropical storms applies to most natural phenomenon such as precipitation levels, droughts, wildfires, pandemic diseases…
This is however only the tip of the iceberg. The evolution of the climate itself is a source of uncertainty but there are others. Our limited scientific understanding of the consequences and their exact unfolding is another one. But as we move forward in time in our forecasting exercises, the main source of uncertainty is the way our societies are going to react to increasing environmental pressure. Are we collectively going to step up our efforts to achieve carbon neutrality? Are we headed for a word of climate-related instability characterised by very limited cooperation? These are the real structural questions that fuel uncertainty and drastically limit our ability to forecast the future.
Radical uncertainty or the “unknown unknown”
Because collective decision making on a global scale is impossible to predict over the long term, we cannot assign probabilities to various outcomes. We are thus heading into an era of radical uncertainty, in which we do not know how the Earth system is working, we do not know the limits to our understanding of this system and we do not know how this system will interact with human activities — activities that have a feedback loop on the Earth system through GHG emissions and resource consumption.
This raises significant issues when an investor, a pension fund or a government is trying to make long-term investment decisions (i.e. over a 20 years or more horizon). These long-term decisions are central to our economies as they pave the way for the future, enabling vast improvements in living conditions (infrastructures), the financing of pensions for vast swath of the population (pension funds and public pension systems), the financing of States, public and private borrowers (long-term bonds), etc.
Scenarios as a way forward
So how do we proceed in this uncertain environment?
The way forward is to build and prepare for competing alternative scenarios. These scenarios describe potential state of the world in the future and are mutually exclusive so they can cover collectively as much ground as possible. As it is impossible to establish which scenario is more likely, they are not assigned probabilities. However, the investment decision has to be robust, whichever scenario may actually unfold.
This approach is developing significantly under the impulsion of the Network for Greening the Financial System (NGFS) that has produced a coherent set of scenarios for the future and made them publicly available through its portal. The NGFS is a group of Central Banks and Supervisors who pooled resources on a voluntary basis to provide thought leadership, tools and insights to help tackle the consequences of climate change.
This approach is beneficial in the sense that it provides decision-makers with a glimpse into different potential futures. It has however also limitations, the first and main one being that the scenarios described, although the most realistic ones given our current knowledge and situation, are only a few possibilities among an unlimited universe of potential outcomes.
Scenarios are a departure from NPV analysis
Using scenario-based robust decision-making is also a strong departure from the traditional Net Present Value (NPV) analysis that underpins most of traditional financial decision-making. In this approach, scenario analysis is limited to stressing variables within a set business plan with fixed rules. This deterministic approach to decision-making strongly undervalues the risk of systematically disruptive change, i.e. change that would affect the rules by which the business plan is operating. To give an illustration of this, consider the COVID epidemic. No financial model in the world took this risk into account and no model could take it into account as it completely disrupted the underlying hypothesis, such as continuity in contracts, absence of major disruptions in global supply chains, availability of workers, etc. These are factors that cannot reasonably be factored in.
The NPV approach loses its internal consistency if one considers that such basic rules can and will very likely be disrupted by climate change in the future. However, it is still the dominating language being used for financial decision-making and actors departing from this practice will come up with valuations that are not in line with market values (either strongly above for assets mitigating / adapting to climate change or strongly below for exposed assets). In other words, they incur the risk of losing significant amounts of money over long time periods before events eventually prove them right. Not many financial institutions, private actors or even governments can afford to wait for more than twenty years to prove that their investment decision was the right one.
A quick classical example of this as a form of conclusion: should a municipality invest an extra hundred million euros to reinforce the dams controlling the flow of the local river? It is extremely complicated to forecast accurately the risk in 50 years. The NPV approach will focus on a core hypothesis of discounted potential damages, stressed to more extreme values. The scenario analysis will look at the conditions under which such an event could occur and derive conclusions to make the decision robust in each scenario (i.e. the dam is strong enough, no matter the scenario considered).
Other innovative approach exist, such as adaptative decision-making, that we will cover in future publications.
