Skip to content

Publication · Working paper

Looking ahead: renewable infrastructure risk

A diagnosis of the 2022 price crisis and a set of regulatory proposals to assign each risk to the agent that can actually manage it.

Download the document PDF · 1.8 MB · in Spanish

In under a decade Chile connected wind and solar projects at a rate of roughly 400 MW/year and 650 MW/year respectively, and in 2022 alone some 2 GW of photovoltaics came into operation against a peak demand only slightly above 11 GW. That year the market showed high volatility between day and night, widespread zero prices in summer, and significant decoupling between the north and the country's load centre. Nine companies requested a transitory change to the wholesale market's price formation system.

The paper settles one question in that debate: whether the problem is the marginalist model or its local implementation. The answer is the latter. An expansion of intermittent generation with zero variable cost does not change the fundamentals of efficient electricity market design; what it does is raise the importance of improving implementation and of allocating risk properly.

The merit order is not the marginal cost

Prices are formed administratively by ordering units by variable cost at a reference bus. Calling that the system's marginal cost is, in the paper's words, an abuse of language. Three shortcomings: it does not capture short-term opportunity costs —critical for storage, which has no directly auditable costs—, it cannot represent non-linearities or the intertemporal relationships of dispatch, and it ignores the variability of a highly renewable system because it uses a single representative node and only three time blocks per day. The result is not just a wrong price: it is also an operation that is not cost-efficient, with rising side payments from supramarginal units running at technical minimum.

Five risks and who should bear them

  1. Price risk. Move at least to a binding schedule, and preferably to a bid-based day-ahead market.
  2. Decoupling risk. Zone the energy requirements of regulated-customer auctions and bring penalty factors closer to reality.
  3. Curtailment risk. Curtailment is endemic to deeply decarbonized systems; reassess the banked ("embancada") operation of thermal units and systematize hot-start conditions.
  4. Systemic-cost risk. Allow these costs to be passed through in auctions and move from hourly to daily coverage of losses for supramarginal units.
  5. Regulatory risk. Stability does not mean freezing the rules, but that changes point to a clear public policy objective.