Adaptative decision-making as an answer to radical climate uncertainty
In previous entries, we explained radical uncertainty linked to climate change and how scenario analysis could be an entry point in overcoming that difficulty (here). We went on to discuss robust decision-making techniques and the role that Artificial Intelligence could assume in this direction (here). In the present article, we will discuss another decision-making technique, adaptative decision-making pathways. These articles are based on a book I recently published with my coauthors on the decarbonisation of companies.
Scenario analysis and robust decision-making techniques share a common thread in that they are providing management with potential futures and examine the consequences of a decision made today in these futures. The underlying assumptions in both cases is that the decision today is irrevocable and therefore cannot be modified in the future. An example would be the decision to build a new nuclear power plant or a factory.
Adaptative decision-making departs from this hypothesis. When confronted with radical uncertainty about the future, having set decisions is by itself a factor of risk as adaptation is the key to overcoming a changing business environment. The goal then becomes to change the way the decision is formulated to factor-in adaptation possibilities down the road.
Planning for the worst has a cost
The seminal example for this is the sizing of a dam, a road or of a factory. These are long-term decisions that require the people in charge to take a stand on what the world will look like in multiple decades. Should they get it wrong, the whole economic equilibrium of the project could be compromised, or, in the case of a dam, the safety of the population leaving downstream could be endangered.
An approach based on scenario and/or robustness analysis would tend to maximise the effort in order to provide as much safety as possible, i.e. all negative scenarios are covered. This approach is however challenging as it can lead to significant resource over- or under-consumption. The dam is too high, the factory too big or too small. In economic terms, aiming for robustness in a significant number of scenarios leads to a misallocation of resources, both in terms of capital and labour. These resources could have been usefully mobilised elsewhere and there is thus a significant opportunity cost to planning and preparing for all outcomes. It is a bit as if uncertainty was levying a tax on economic actors.
Smaller decisions spread over pathways
In order to minize this cost, adaptative decision-making breaks down the decision into individual sub-decisions that can be spread over time. It identifies ex ante convergence points in time, when these sub-decisions have to be made. This approach provides decision-makers with different pathways for their decisions. Taking the factory-sizing example, we can identify multiple decision routes. One route would be to keep the current factory and do nothing, pushing back the investment later in time. Another antagonist route would be to invest now in the biggest possible factory, ramping up production over time with a tool that is already built. Between these two extremes, multiple alternatives can be identified: building only the walls and roof of the biggest factory and only one production line, keeping space for extensions later on, etc. The way the decision is formulated is thus no longer a simple GO / NO GO at a definite point in time but multiple smaller decisions that can be chained over a longer period, allowing for the inclusion of new information.
The strategy analyst can then describe multiple pathways within all these futures with decision nodes spread over years. The company may choose to do nothing for two years and then build the biggest factory. Or invest in a single line now and increase capacity in five years. The trajectories have specific risks and costs profiles and generate different potential results. These can be compared and management may choose between them without committing to a firm, irreversible decision.
The adaptative pathways decision-making process is an intellectually prolific way to rationalise decisions in an uncertain environment and it certainly helps to reduce or frame the cost of planning for the worst. However, it remains ill-suited for potentially dangerous activities that can be faced with low-probability, high impact events (i.e. so-called fat tail event). Indeed the adaptative pathways tend to reduce the incentives to invest in the maximum level of safety or to push back in time these decisions.
In the end, all these techniques are only here to provide assistance in decision-making. An incompressible subjective part remains, that boils down to how management and society value risk taking in the face of radical uncertainty.
