A retirement date belongs in the contract, not the pledge
Southeast Asian governments and generators have announced dates for leaving coal. Few of those dates appear in the agreements that actually keep plants running, and until they do, early retirement is an intention that nobody has yet been paid to deliver.
Across Southeast Asia, coal retirement has acquired a calendar. Governments have published years by which coal will leave their power systems or stop growing, and several generators, including mine, have published years of their own. International partners have offered concessional finance for early closure. Conferences discuss the subject as though the main question were how soon.
I think the main question is who has signed. A coal plant in this region rarely belongs to a market in which an owner can simply stop running it when margins fall. It belongs to a set of long agreements, most of them with a state, and those agreements, not any pledge, decide when it closes. A retirement date that has not been written into them remains a hope. I say this as someone who once helped finance coal plants and is now helping to close one.
Four agreements hold a coal plant open
The first is the power purchase agreement. Most coal plants built in the region since the 1990s sell to a single state buyer under contracts of twenty-five years or more. The buyer pays a capacity charge for the plant being available and an energy charge that passes through the cost of fuel. The capacity charge repays the debt and the equity; it is the reason the plant could be financed at all. Closing early means the buyer stops paying that charge, and the owner loses the income meant to repay its investment.
The second is the debt. Project loans are sized to the contract and usually amortise until a few years before it ends. Lenders will not accept a shorter life unless they are repaid or refinanced, and refinancing a coal plant grows harder every year as banks adopt coal policies of their own.
The third is the fuel contract. Coal supply agreements carry minimum volumes and sometimes links to a mine owned by an affiliate. In Indonesia, where domestic coal supply obligations shape what generators buy and at what price, the mine and the plant often sit inside the same political economy.
The fourth is the grid. A large coal unit is part of a system operator's plan for reserve margin and voltage support. Remove it early without replacement at or near the same connection and the operator will, quite properly, object.
Our Johor station shows how these fit together. When we began negotiating in 2022 it had eleven years left on its contract, US$310 million of debt, a coal supply arrangement running to 2030 and a role in supporting the southern end of the Peninsular grid. Closing it at the end of 2029 required each of those four positions to move. It took two years. Nothing in a pledge would have moved any of them.
Pay for the replacement and the people, not the owner
Once the agreements are visible, the uncomfortable question follows: who should be paid to close a plant early, and how much?
The instinctive answer from owners is that they should be compensated for lost earnings. I understand the instinct; I owe a duty to our shareholders. But I think it is the wrong basis for public or concessional money. If transition finance simply buys out an owner's remaining profits, it rewards the plants that were most profitable and spends scarce capital on a transfer rather than on anything new. It also teaches every owner to wait for an offer.
A better principle is that concessional money should pay for three things: replacement capacity, the workforce and communities around the plant, and a lower cost of finance that makes a shorter life bearable. Owners should expect to absorb some loss. In our case the early date cost about US$140 million of net present value on our own model. A concessional refinancing recovered about US$35 million and avoiding a major overhaul about US$60 million. Our shareholders bore the remaining US$45 million. Kelawari Holdings Berhad, our state-linked anchor shareholder, accepted that loss after long discussion, because the alternative, a plant that might later be forced shut with no replacement and no plan for its people, carried a larger and less controllable cost.
I am wary of one popular answer to the remaining gap: carbon credits generated by the early closure itself. Methods for such credits are being developed in the region, with serious work behind them. But they are still being tested, buyers have not shown what they will pay, and a credit sold against emissions avoided by closing a plant that a government would have retired anyway raises an obvious question about whether anything additional was achieved. We counted no such credits in the Johor case. If they arrive, they will be welcome. A retirement that depends on them is a retirement that has not yet been financed.
The workforce deserves more than a line in the paper. Our station employed 410 people, many of them second-generation power workers from nearby towns. We committed to offer every permanent employee a new role, retraining or voluntary retirement. That commitment cost money, and it was also what made the agreement legitimate in the eyes of the government and our anchor shareholder. A transition that leaves a town worse off will be resisted, and in this region the resistance reaches a ministry long before it reaches a court.
Questions to ask before announcing a date
Many boards in our industry are now asked to approve retirement pledges, often for a sustainability report or an investor day. I would encourage directors to ask five questions before approving any date.
Has the buyer agreed in writing to the capacity charges it will stop paying and to the replacement it expects? Have the lenders said what they need, and is there a refinancing that works without assuming concessional money nobody has yet committed? What happens to the fuel contracts and to any affiliated mine? What will the system operator need at the connection point, and can the company compete to provide it? What has the workforce been told, and what will each person be offered?
If any answer is no, the date is an intention. That may still be worth disclosing, but it should be described as an intention, with the conditions that would turn it into a commitment. Our own disclosure about our East Java plant is written that way. Its contract runs to 2042, its buyer depends on government support, and no concessional package is yet on the table. We have said we are studying early retirement. We have not given a date, because we cannot yet name everyone who would have to sign.
The region needs coal plants to close sooner than their contracts allow, and much of the money to do it will be public or concessional. That money will be spent well only if the work begins with the contracts: who is paid, for what, and in return for which change. Pledges set a direction. Signatures close plants.