Our viewpoint

Modelling storage:
what went wrong?

Regardless of your position on storage, it is fairly uncontroversial to admit we have recently experienced a bubble in storage investment that has somewhat tempered. As the market settles, introspection is required and we must ask ourselves what happened. Here, I focus on the modelling behind many of these decisions.

A lack of fundamental, dynamic modelling 

There was a window where investing in batteries seemed too good to be true, driven in part through embedded benefits, lucrative frequency response contracts and 96% deratings in the capacity market. Projecting these revenues forwards made light work of sourcing financing for these projects, and so the market went all in on this relatively new asset. Now, the story is quite different. Many of the embedded benefits have been removed through charging reforms, a glut of competition in the frequency response market has all but killed the margins available, and short duration battery deratings have been dramatically reduced, in some cases to below 20%. What went wrong?

In part – a lack of fundamental analysis. Rather than modelling the intrinsic value to the system provided by fast frequency response, capacity or distributed generation, these revenues were simply projected forwards. Forecasts then banked these revenues over long term horizons, rather than treating them as fleeting opportunities that the market had not yet had time to correct for. Save for a few exceptions such as contractual obligations, all modelled revenue sources should be responsive to the value they provide to the system at the time, not an assumption.

Projecting revenues has another fatal flaw - it misses the crucial aspect of cannibalisation. That is, in a competitive market, opportunities are present to all participants. So if there are lucrative opportunities, competitors will take decisions that eat into that profit share until it reaches an equilibrium level. Again, competitor decisions must be reactive to the market they find themselves in. A fundamentals-driven approach may not have given investors the answer they wanted at the time, but may have saved some misinformed, expensive decisions.

Looking forward

The market is regularly in a cycle of abrupt changes and corrections – from the CM suspension, to embedded benefits reviews, to Brexit. When considering a long term forecast or decision, significant care should be taken in how to model these changes.

If you believe that the revenues your investment currently receives are not reflective of the intrinsic value offered to the system, then be wary of projecting these revenues forward. Competition and market corrections are likely to erode these revenues over time. Conversely, if the asset does offer intrinsic value to the system, the market will continue to reward this value in the long term, even if the precise mechanisms change. To fully characterise these dynamics, an approach rooted in market fundamentals is required.  A fundamentals approach can identify where long-term value lies, and where an investment really is too good to be true.