In just a few years, Chile's power system changed faster than its contracts. The SIC 2013/03 tender was bid at a time of high marginal costs and today, indexed, averages 135.6 US$/MWh; the reference price of recent tenders is around 60. Meanwhile, between 3 and 10 November the marginal cost at Alto Jahuel averaged 4.7 US$/MWh, and batteries already deliver more than 80% of peak energy.
The thirteenth transitory article of Law 21,833 responds to that gap. It empowers the National Energy Commission (CNE), once and until December 2027, to design a mechanism that lowers the current price paid by regulated customers by modifying the price, term and conditions of the contracts, under three principles: legal certainty, voluntary participation and economic efficiency. Almost everything operational —what the cut is measured against, how it is negotiated, who settles disputes— is left to the regulation.
The tariff is already going to fall
It is worth starting with a fact that gets lost in the debate. The average price of regulated energy will fall from about 95 US$/MWh in 2027 to about 60 in 2034, simply because the expensive contracts expire and that energy is tendered again. Renegotiating does not change that outcome: it brings it forward. The question is how much bringing it forward costs, and who pays.
Extra term is a loan
Let us think about how a voluntary agreement is closed. The existing contract is the disagreement point: neither party will accept ending up worse off than under it. If the generator cuts its price today and receives more years of contract in return, the premium on the extended period has to pay that cut back. At the generator's discount rate the deal is zero-sum: no value is created, payments are simply moved forward in time.
Put differently, lowering the bill today in exchange for a longer term means borrowing from future consumers. The extension keeps the tariff from bouncing back when the contract reaches its original expiry, but on its own it adds nothing.
The discount rate decides who wins
If the deal is zero-sum at the generator's rate, whether the customer wins or loses depends on the rate used to evaluate it. For the 20 contracts of the 2013/03 tender, with the generator discounting at 10% real, the result is clear. At the 5.5% social rate the customer loses in every combination of price and extension simulated: between 446 and 1,249 million dollars in present value. At 10% the result is zero, by construction. Only a customer discounting at 15% comes out ahead.
The surplus lies in risk
Where is the value, then? In the risks that the old contracts allocated poorly and that tender rules have since corrected. The 2013/03 tender sells a single 24-hour price, admits six indexation formulas, bundles system costs into the price and was bid at a single node. Recent tenders pass those costs through to the customer, zone the supply and limit indexation to gas and CPI. Each of those changes is a possible trade-off: reallocating system costs (about 9 US$/MWh), zoning in the face of price separation between nodes of about 35, changing the indexation formula, setting a billing floor or moving to time-of- day block prices. None of them is free, because each one shifts one risk and creates another; but, unlike extra term, they can leave both parties better off.
What the regulation must settle
In our view, the regulation has to make the three principles of the law verifiable. First, credible voluntariness: not taking part cannot carry consequences in future tenders or before the CNE. Second, an efficiency test: each agreement is approved only if it leaves customers better off in present value, with the counterfactual and the discount rate fixed before the call. Finally, certainty: complete rules published before the process opens, a common menu of trade-offs awarded by present-value benefit, and a public technical report by the CNE on every agreement.
The timetable is tight. The regulation is due the same month the 2026/01 tender is awarded, and whatever signal it sends will seep into those bids: reopening contracts can raise the risk premium of the tenders to come. That is why the underlying question is not whether the bill can be lowered, but what cut justifies doing so. A real cut requires transferring risk; extra term, on its own, creates no value.