Glass half empty or half full?

https://www.linkedin.com/pulse/glass-half-empty-full-maxine-frerk/

One of the questions raised by Ofgem in its Consultation on the RIIO price control framework was about how to handle uncertainty and ensure that customers don’t end up paying for investment that turns out not to be needed. In its Decision this question was left unanswered, to be picked up in the sector specific strategies that it will be consulting on later this year.

That feels like a missed opportunity as this is an important question that has the potential to cast a long shadow over the forthcoming price controls. While Ofgem’s focus was implicitly on electricity distribution where they see the use of flexibility services as providing the answer, it is also a critical issue for gas distribution with the uncertainty that exists around the future of gas.

To help my own thinking on this question I’ve been looking back at “real options” theory and other tools for decision making under uncertainty and have produced a working paper which is available from my website www.gridedgepolicy.com.

The key conclusion is that what is needed is a “toolkit” approach rather than relying on one particular methodology, and to draw on the insights that methods such as real options assessment provide even if the methodology itself might not be practical to employ in its full glory.

When thinking about existential risks and questions of whether government policy will be met it can be hard to attach probabilities to different scenarios. That is the reason why National Grid in developing its Network Options Assessment (NOA) doesn’t attempt to put probabilities on its Future Energy Scenarios. Instead it uses a technique known as least worst regrets which tries to avoid taking a glass half full or glass half empty view of the world but looks instead at minimising “regret” – the sense of loss that one feels when you make what turns out to have been a bad decision given the way events unfold.

While it appears an objective and rigorous methodology it is recognised (and demonstrated in the paper) that the inclusion of an additional option in the assessment can affect the outcome even if the additional option isn’t the preferred one. That feels counter intuitive and certainly hard to explain to your average stakeholder. Moreover, while regret theory has a basis in individual decision making (and behavioural theory) that does not make it a sound basis for regulatory decision making – we shouldn’t be concerned here with the regret of the decision maker but of maximising the expected consumer benefit.

The answer then is to be wary of mechanistic calculations and to try to bring in more of the insight from real options assessment. Some of this thinking is beginning to be reflected in industry discussions with frequent reference to no or low regret decisions and the value in keeping options open. However the real options literature would also point to the importance of investment in learning, of unpacking decisions to create flexibility and of being ready to quit and cut one’s losses if circumstances change (which can be difficult for politicians and regulators and doesn’t mean the original decision was “wrong”).

And when it comes to RIIO some of the decisions that need to be taken around individual investment projects at the distribution level don’t need a full-blown scenario assessment and can draw on probabilities of different levels of load on the system and simple decision trees. ENWL have used this sort of thinking and the concept of “trigger” levels of demand to look at reinforcement in the network.

As another practical example, if you are digging a hole to repair or upgrade the network then it may well be worth spending a bit extra now to future proof that investment rather than running the risk of having to go back and dig another hole. How you make that decision depends on your assessment of likely future outcomes and the cost of the extra work.

So while the language that tends to be used is around no or low regret solutions, one of the challenges for Ofgem is that creating real option value can increase costs in the short run even if it is better value for customers in the long term. When Ofgem comes to benchmark the different approaches taken by companies for RIIO2 it will need to find a way to factor in where real option value has been created. That requires a new way of thinking about efficiency and is why Ofgem needs to start to address these questions sooner rather than later.