Indonesia's JETP: A coal retirement conundrum

Summary

Indonesia’s $20 billion Just Energy Transition Partnership (JETP) is as ambitious as it is vague

If mobilised, JETP cash could close half of Indonesia’s coal power capacity up to 10 years early

Total capacity for closure varies between 14.6 GW and 21.7 GW depending on criteria used for selection

Knock-on impacts of each coal closure selection are unknowable without systems-level analysis

TransitionZero’s forthcoming energy system model, FEO, will empower decision-makers by leveraging plant-level data

Indonesia, the world’s biggest coal exporter and a major domestic consumer of the emissions-intensive fuel, is about to embark upon what is arguably the most ambitious coal retirement plan in the world. With a potential $20 billion of climate finance to be mobilised through the country’s newly minted Just Energy Transition Partnership (JETP), Indonesia could soon find itself in a position to shutter more than half of its operational coal-fired power plants up to 10 years early.

That’s the headline finding of analysis run by TransitionZero using its open-data Coal Asset Transition (CAT) tool. CAT brings together asset-level operational data on Indonesia’s entire coal fleet and calculates the cost of buying out the remaining power purchase agreements (PPAs) underpinning each unit.

CAT calculates the long-term profitability, remaining asset life and levelised cost of clean power replacement (plus battery storage) to allow high-level screening of coal plants for retirement. It also estimates the social cost of water stress, air pollution and overall climate externality cost, and quantifies the job losses associated with the closure of each unit.

Download CAT Indonesia

DownloadDownload

Known unknowns

The high-level Indonesia JETP announcement at the G20 summit in Bali in November pledged to mobilise $20 billion over the next three to five years, split evenly between international partners (co-led by the USA and Japan, plus the UK and EU countries) and private financiers coordinated by the Glasgow Financial Alliance for Net Zero (GFANZ).

The funds will be used to accelerate coal retirements and renewables deployment so that renewable energy comprises at least 34% of all power generation by 2030, up from one-fifth in 2020. The ambition is large but commitments are vague, with many details still to be ironed out.

For example, the global banks providing private finance are yet to be confirmed and the incentive for their involvement is unclear. Will they expect voluntary carbon credits to support financial flows into early coal retirements? If so, how can the credibility of measurement, reporting, and verification (MRV) be guaranteed, given there is currently no approved methodology for accreditation?

There are also ‘known unknowns’ around how deals will be structured. It seems reasonable to expect them to mirror Asian Development Bank (ADB)’s recent memorandum of understanding (MoU) to retire the 660MW Cirebon-1 coal-fired power plant in West Java. This is the flagship deal under the ADB’s Energy Transition Mechanism (ETM), which the bank expects to be “one of the key delivery mechanisms” of Indonesia’s JETP.

The Cirebon-1 MoU proposes the plant owner retains ownership and will be compensated for foregone profits from early retirement using a new, lower-interest concession loan. This avoids the need to acquire the coal asset and get the plant on the books of the financial institution. But coal plants don't typically hold a 40–50-year loan, so there is work to do to ensure all parties are fairly compensated.

Last but not least, the mechanism for disbursing JETP funds is not clear, partly because the precise involvement of each arm of the Indonesian government is still being clarified. Each ministry is expected to have its own priorities, which will feed into the coal plant prioritisation process.

Competing priorities

Assuming these uncertainties can be resolved and the full $20 billion JETP fund is mobilised, the next question becomes: what criteria will be used to decide which plants will be picked for early retirement? There are many metrics that can be used: PPA buyout costs, air pollution emissions, water use stress, job losses, plant profitability, among others.

Financial incentives vary depending on plant ownership. State utility Perusahaan Listrik Negara (PLN) will probably want to retire the most profitable third-party independent power producers (IPPs) first since these are likely to have more expensive PPAs and contribute to higher system costs. But IPPs will want to get paid to retire their “at risk” plants first. These may be the least profitable units, or those that face risk of early shutdown for other reasons (such as local protests).

TransitionZero’s CAT tool arms decision-makers with the data needed to filter coal units based on their chosen priorities. The tool’s PPA buyout calculation can be used to produce a list of units that could be paid down using the available funds, ranked in order of greatest impact in their chosen priority area. This exercise yields vastly different results depending on the criteria used for selection.

Based on TransitionZero’s in-house methodology, when grid-connected¹ coal plants are ranked according to long-term profitability, with the least profitable units retired first, the $20 billion JETP fund could shutter 21.7 GW of capacity early (see slide 1, above). Ranking units according to abatement cost (slide 2), with the plants offering the cheapest emissions savings retired first, delivers a similar result.

But prioritising plants with the greatest local air pollution and water stress impacts – with highest impact units retired first – produces an entirely different mix. The average cost of buyout is higher for these units; observe how some of the highest cost per-megawatt units move into the $20 billion pot in slides 3 and 4. This eats up more of the budget, reducing the overall retired capacity under the $20 billion cap (14.7 GW and 14.6 GW, respectively).

If all the above risk metrics are given equal weighting (slide 5), the ranking shifts back to a more ‘economically rational’ selection with a higher overall retired capacity (21.3 GW).

Interestingly, the ownership split between PLN and third-party IPPs shifts between the selections. Under the ‘profitability’, ‘abatement’ and ‘equally weighted’ scenarios, the units are fairly evenly distributed between the two. But under ‘air pollution’ and ‘water stress’, the split is 65:35% and 68:32% in favour of closing PLN-owned units.

This implies that the state utility’s coal plants have higher externalities, which may be explained by the differing incentives and priorities between PLN and IPPs. IPPs are risk averse in their locations, preferring to operate in clearly risk-free locations. This leaves PLN, which is motivated by necessity, to move into more ‘risk-prone’ locations as the provider of last resort.

Pricing in the externalities

Why is it more expensive to buy out coal plants with greater air pollution and water stress impacts? Fundamentally, these are externalities that are not priced into the operational cost of each unit. If plant profitability explicitly reflected these impacts, optimising by profitability would give greater priority to these environmental risk factors.

Also, plants in the air and water selections are underpinned by more expensive PPAs and/or are younger installations with more operational life left in them. In the local air pollution selection, the average asset remaining life is 18.3 years (compared to 16.2 years in both the profitability and abatement cost plant selections).

In the water stress selection, the average tariff (PPA) estimated for each unit is $61.5/MWh (compared to $56.2/MWh and $56.5/MWh in the profitability and abatement cost selections, respectively). It is possible that these plants are located in more remote locations, making them more expensive to build and operate, thus requiring higher PPA revenues.

In both cases, these factors push up the per-MW cost of buyout, which means financiers get less decarbonisation ‘bang’ for their $20 billion JETP buck. The trade-off is that plants located near densely populated areas, and which have the most severe impacts on human health and wellbeing, are retired early.

Geographical distribution

There is one constant among the different selection scenarios: geographic distribution of plants between Indonesia’s many island electricity networks. The vast majority of units prioritised in all scenarios are located on Indonesia’s largest grid, Java-Bali.

The reasons for this are twofold: Java-Bali has by far the largest concentration of coal-fired power plants (24.2 GW) and a very high reserve margin of 59%, which is unsustainably high. Peak load on Java-Bali was 28 GW in 2021, which means there is a large amount of underutilised capacity. This presents an opportunity to retire many dirtier, higher cost units early without impacting grid stability.

Prioritising selection by ‘air pollution’ and ‘water stress’ metrics produces a modest shift in plant locations towards the smaller Sumatra and Kalimantan islands, which are home to 3.6 GW and 1.8 GW of thermal coal capacity, respectively. Reserve margins are lower on these grids at 35% and 45%, albeit still high by international standards. The trade-off here is that these islands are home to many coal mines, which could disproportionately impact upstream jobs.

Unintended consequences

The question of priorities is only the tip of the iceberg. JETP decision-makers will also need to understand the consequences of their choices. Ministers and state utility PLN will probably want to know how much of the retired coal capacity would need to be replaced with newer, cleaner and cheaper generators – and where, physically, new units should be located to maximise cost efficiencies. And they will probably need a clear view of how the emissions intensity of each grid changes as plants are taken offline and replaced.

A rudimentary analysis of these variables can be carried out using CAT’s plant-level data. For example, when prioritising profitability or abatement cost, these early retirement selections result in total plant level emissions savings of 1,165 million tonnes of CO2 over a period of up to 10 years. This is comparable to the global aviation industry’s pre-Covid peak in emissions. The local air pollution and water selections offer plant-level savings of 811 and 857 MtCO2, respectively, over the same timeframe – slightly more than the aviation sector’s post-pandemic carbon footprint.

These are significant savings that, if realised, could help to achieve the JETP’s headline target of peaking power sector emissions at 290 MtCO2 by 2030, down from a 2030 baseline of 357 MtCO2. Indonesia’s energy sector emissions were around 600 MtCO2 in 2021, of which electricity accounted for roughly 40%.

But emissions savings figures derived from plant-level data do not reflect how each power system will change over time as capacity is retired. What if coal retirements incentivise dispatch by other underutilised coal capacity on the same network? If that capacity has a higher emissions intensity than the retired generators, the impact could be a net increase in emissions.

Spend it wisely

Similarly, choosing how much retired coal capacity to replace with renewables (with or without storage) is tricky without an understanding of system-level impacts. Indonesia’s coal fleet is chronically underutilised, resulting in very high reserve margins across its largest grids. Not all retirements will require like-for-like capacity replacements, but some will in order to ensure grid stability.

The $20 billion JETP budget is intended to fund two overarching activities: early coal retirements, and accelerating renewable energy deployment so that renewables comprise at least 34% of all power generation by 2030. The money can’t be spent twice, so it must be split cost-effectively between the two.

Replacing every coal unit with equivalent solar PV capacity depletes the JETP budget. And when that capacity is backed up by storage capacity, that $20 billion starts to disappear very quickly. This underscores the importance of system-level analysis to ensure cost-effective allocation of JETP capital.

Political considerations

But it is not just about spending climate finance wisely. Perhaps the greatest challenge to accelerating coal retirements is the need to mitigate the impact of plant closures on jobs, employment and local communities. The political sensitivities of the energy transition are centred primarily on job losses in traditional energy sectors, particularly so in resource-rich emerging markets such as Indonesia.

As previously explored by TransitionZero, power sector decarbonisation is likely to yield net job gains at the power plant level thanks to jobs created from the deployment of renewables and storage. But there is no way of knowing whether these new clean assets will be built near the communities most affected by coal closures without systems-level analysis of land use change.

Political considerations might incentivise decision-makers to encourage the co-location of renewables build-out at or near retired coal plants in order to facilitate the transfer of workers into clean jobs and mitigate job losses. This could be a worthwhile approach that delivers net cost benefits, but the only way to be sure is to model system-wide impacts of doing so.

Systems problem, meet systems solution

CAT is enormously valuable when assessing coal plant retirement options. But plant-level data can only ever offer a snapshot of a power system at one moment in time. Change begets change, and every intervention will have non-linear knock-on impacts. Understanding these requires temporal data analysis.

This is why TransitionZero is developing a global energy system model that will produce net-zero scenarios for 165 countries. This platform is already being utilised to support the Climate Prosperity Plans of the Vulnerable Twenty (V20) Group of Ministers. The model is being rolled out in stages, starting with Indonesia in 2023.

MB will allow users to undertake project and regional level analysis in near-real-time, ensuring the costs and benefits of energy planning decisions – such as coal plant retirements – are understood and communicated in hours rather than months, as currently provided by bespoke closed-source models and datasets. If CAT is the ‘fuel’, then MB is the dynamic decision-making ‘engine’ that will leverage the underlying power of granular plant-level datasets.

Retiring coal assets in coal-dependent countries is a delicate and high-stakes process that requires careful planning and stakeholder buy-in. To foster consensus and confidence in the process, decisions must be collaborative, inclusive, transparent and based on credible assumptions. They cannot be taken ‘blind’ to systems-level consequences.

With momentum building around Just Energy Transition Partnerships, it is vital that the first JETPs are a success. International financiers and local stakeholders alike must have faith that large capital sums mobilised through these platforms are being allocated to maximum effect. Next-generation energy systems analytical capabilities will be key to underpinning investor confidence in JETPs as a viable weapon in the climate finance armoury.

Notes:

1. This analysis excludes captive plants because these are more challenging and complex to replace. Indonesia has 9.2 GW of captive thermal coal plants in operation at smelters and other facilities. Excluding these plants leaves a pool of 30.6 GW of grid-connected coal power capacity, spread across 118 generating units, considered as eligible for early retirement in this analysis. More than half (76 units) are less than 10 years old, with unit capacities ranging from 30 MW to 1 GW. Many plants are under-utilised and as a result suffer from low profitability, offering low-hanging fruit for early retirement.