DECISION THESIS

A storage configuration may benefit the system and be unfinanceable under current remuneration, or be financeable and provide little benefit. System benefit is the net reduction in the system’s operating and unserved-energy costs against an explicit counterfactual. Capturable revenue is what the investor receives for energy, capacity, services, and contracts. The economic evaluation determines whether the configuration benefits the system; the financial evaluation, whether it

Mexico publishes a 3-hour reference, Colombia delivery windows, and the Dominican Republic a minimum of 4; the documents examined do not show how they were derived. The Mexican 2026 call for projects gives 3 hours as a "reference" and allows different proposals. In Colombia, the products in Resolución 40208 (Resolution 40208) set windows of 08:00–18:00 and 18:00–22:00, with no minimum battery duration. The Dominican Republic requires 4 hours for batteries in a renewable tende

In PNNL’s reference database, which uses 2023 values for the United States, duration changes the cost ranking of storage. Levelized cost per MWh discharged, at 1,000 MW. At 4 hours: lithium USD 174/MWh, pumped hydro storage 186, compressed air 130. At 10 hours: lithium 168, pumped hydro storage 97, compressed air 91. At 24 hours, lithium costs 1.76 times as much as at 10. The table combines technology, operating assumptions, and site, so it does not say which technology wins

At Crucero, in northern Chile, the arbitrage margin of a hypothetical 4-hour battery, with perfect foresight, fell in 12 months. It went from 137 to 105 USD/kW per year. In that period, system storage power rose from about 1.2 to 3.2 GW during 2025. These data cannot isolate how much of the decline is attributable to batteries. The average price across all hours also fell, from 51 to 43 USD/MWh. The transitional regime that recognizes capacity for batteries lasts 10 years fro

Even when regulatory capital is ample, concentration could constrain loss-absorbing capacity first. A large loan would consume capital, liquidity, and loss-absorbing capacity that would then be out of reach for the rest of the portfolio. If the transfer rate covers only funding and interest rate risk, that consumption could be left out of the price. The bank could charge its largest borrower as it charges a small one. With a Capital Allocation Decision System, the aim would b

If credit deteriorates, it would be worth checking whether expected losses on new cohorts exceed those built into the price. Pricing and origination could incorporate that deterioration with a lag. The current recovery cost could differ from the one used when the loans were priced. Whoever decides origination would need to check whether they receive that cost or whether recovery is managed only as an operation. With internal data, they could compare expected losses on new coh

In the Analytical Transition examples, optimization appears in 5 of the 6 functions. That assignment is our own proposal. To choose among alternatives, the decision-maker would first have to write down what is maximized or minimized and which constraints apply. The analytical family would be deduced from that problem. In Capital Allocation, finite capital would be distributed among alternatives. In Capital Projects, the configuration of a physical system would be decided. Opt

If the person who approves the use case starts with the decision, that person could set the decision owner and its criterion before choosing what to do with the data. A use case that originates in the data could produce an indicator, a dashboard, or a forecast. One that originates in the decision would seek to change a criterion and record the result. Both paths could improve a decision. A use case that originates in the formulation could also stall at a dashboard if its impl

If the premises arrive set from the top and the areas contribute no binding constraints, the plan may allocate a figure instead of discovering it. Targets, budget, and ambition would form a second-order system, whose output would be the objective function of the other systems. A process in which everyone participates and no real constraint can move the result would be integrated and would lack the capacity to decide. Participating is different from deciding. Its reopening tri

If the capital charge is set once a year, origination that deviates may make it inadequate before its next update. With the capital charge given as a number, the credit area would price loans without resolving each transaction jointly with the balance-sheet core. If a month's origination materially changes marginal capital consumption, by the third week the charge may no longer be the right one. The problem may lie in the cadence, in the methodology, or in both. Under that as

The transfer price would coordinate rates and funding. Where it leaves a resource out, or charges for it with a static average, the areas could go back to deciding on their own. In its matched-maturity form, the transfer price assigns each transaction an internal rate that carries the bank's funding cost and is set at origination. The margin would thus be protected from the interest rate and funding risk that the price transfers. That would be the case in which coordination b

In the PEG-5, more than half of the capacity awarded was solar; for new plants, the PEG-6 admits only geothermal and natural gas. In the PEG-5, the solar component added up to 701 of the 1,258 MW of maximum guaranteed capacity awarded for 15 years. Of 960 MW of maximum natural gas offers, 15 were awarded. For the PEG-6, the Comisión Nacional de Energía Eléctrica (CNEE) required firm energy with a capacity factor of at least 90% and energy offered in all hours. For new plants

In 2025, Chile tendered the energy for its regulated customers in 3 time blocks and awarded all 3 at USD 64.5/MWh. The rules of tender 2025/01 separated night, from 23:00 to 07:59; solar, from 08:00 to 17:59; and evening peak, from 18:00 to 22:59. The Comisión Nacional de Energía (CNE) received 708 offers, 3.9 times the energy requested. It awarded 100% to 1 generator, at USD 64.5/MWh in each block, on the lowest present value of the offered prices and their indexation. Price

In Colombia, the firm energy auction awarded 99% of new capacity to solar in 2024, and 55.9% to thermal in 2026. The February 2024 auction, for 2027–2028, awarded 4.44 GW of solar. The May 2026 auction, for 2029–2030, awarded 4.07 GW of new capacity: 55.9% thermal and 38% solar. Measured over all firm energy awarded, rather than over new capacity, solar contributes 7.7%. Between the 2 auctions, the rules, prices, and hydrological context also changed. Attributing the change i

Firmness mechanisms exist in the 7 countries reviewed. Contracting energy and covering critical hours should be evaluated together. Those mechanisms — capacity, backup, replacement — vary from country to country. In Peru, in the May 2025 to April 2026 tariff, the reserve, capacity, and reliability charges add up to PEN 10.1 per kW per month. The renewable premium adds up to 7.1. Those charges pay for system backup, and their attribution to renewables remains unproven. The pro