Solar Auction Under Curtailment and Currency Risk
A government runs a reverse auction for 600 MW of solar PV, split into six 100 MW lots, with a ceiling price of $48/MWh and a 20-year PPA. The auction clears at $35-42/MWh. The winning developers must connect to a 220 kV network where the utility reports 18-month delays in bay construction. The grid code allows curtailment without compensation, and first-year curtailment in the region has averaged 9%. Local content rules require 30% domestic content, but the only module factory can supply 15% of the required volume and costs 12% more than imports. The PPA is in local currency, indexed 50% to CPI, and the currency has depreciated 25% against the dollar since the auction. One winning bidder is now seeking to renegotiate; the ministry asks whether to enforce the contract, compensate for curtailment, or redesign the auction.
1. Assess the bankability of a 100 MW winning project at $38/MWh. Use a capacity factor of 24%, capex of $70 million, 70% debt at 8% over 15 years, and annual O&M of $1.2 million. Quantify the effects of 9% curtailment and an 18-month connection delay.
Read the worked answer
Energy = 100 MW × 8,760 h × 0.24 = 210,240 MWh. Revenue at $38/MWh = $7.99 million. O&M = $1.2 million. Debt = 70% × $70 million = $49 million; the 15-year annuity at 8% is about 0.1168, so annual debt service ≈ $5.72 million. Cash flow available for debt service before curtailment ≈ $7.99 - $1.2 = $6.79 million; DSCR ≈ 1.19. With 9% curtailment, revenue falls to $7.27 million and CFADS to $6.07 million; DSCR ≈ 1.06, which is thin but possibly financeable. An 18-month delay is worse: lost revenue is roughly 1.5 × $7.99 = $12.0 million, and interest during construction on $49 million at 8% for 1.5 years is about $5.9 million. If capitalized, capex rises toward $88 million and debt toward $61.6 million, annual debt service to about $7.2 million; CFADS at $6.79 million gives DSCR ≈ 0.94. The project is not bankable without risk mitigation; the auction price is below the cost of the unallocated risks.
2. Recommend how to allocate curtailment, grid delay, currency, and local content risks. Which should the government bear?
Read the worked answer
Allocate risks to the party best able to manage them. Curtailment: government/utility should compensate deemed generation above a 3-5% threshold, capped, and only for system-directed curtailment; otherwise developers price a 9% haircut into bids. Grid delay: connection agreement with milestones, liquidated damages, and a PPA extension; the utility controls bay construction. Currency: index the tariff more fully to CPI or to a hard-currency benchmark, or require a hedge; 50% CPI with 25% depreciation leaves roughly 12-15% real revenue loss. Local content: do not enforce 30% when domestic capacity is 15%; phase in with a clear roadmap, allow a price preference, and avoid retroactive penalties. Government should bear curtailment and grid delay because they are systemic; developers should bear construction, performance and O&M; currency risk should be shared through indexation. Enforcing the original contract without changes may win legally but will chill future auctions and raise the next ceiling price.
3. Would a feed-in tariff with a fixed price have produced a better outcome? Consider cost, risk, and deployment speed.
Read the worked answer
An auction can be better than a feed-in tariff because it discovers price and caps the budget, but only if the risk allocation is credible. Here the auction cleared at $35-42/MWh, below the cost of the unallocated risks, so the result is not bankable and will either be renegotiated or defaulted. A well-designed auction with connection guarantees, curtailment compensation and indexation could deliver both lower prices and bankability. A fixed feed-in tariff would transfer more risk to consumers and may overpay, but it would deploy faster. The best answer is not to abandon auctions; it is to fix the auction design and use a two-sided contract for difference or a tariff indexation mechanism. If the ministry needs rapid deployment and lacks credible institutions, a feed-in tariff with a regression mechanism may be the second-best choice.