Back to InsightsUpdate
May 23, 2023
How Indonesia can get ahead of the net-zero curve
Closing old coal plants is a win-win for emissions and costs according to our modelling
Contributors
Summary
There is no zero-sum trade off between cutting emissions and costs: Indonesia could save money and cut carbon by prioritising early coal plant closures
More than 21 GW of coal plants could be retired early without impacting system costs or grid reliability
Early coal retirement saves Indonesia $2 billion and avoids 1.3 gigatonne of CO2 – equivalent to boosting Indonesia’s GDP by 0.15% and simultaneously halting global shipping industry emissions for two years
Compared to business-as-usual, the cost of cutting carbon is negative: Indonesia could save $2 for every tonne of avoided CO2
Solar PV dominates the Indonesian energy mix from 2040 under all scenarios modelled by TransitionZero, regardless of emissions targets
Our tools put unparalleled systems modelling capabilities in the hands of Indonesian energy planners at a crucial moment in the development of the country’s Just Energy Transition Partnership (JETP)
Indonesia could get ahead of the net-zero curve by accelerating the retirement of its least profitable old coal-fired power plants. Doing so would cost less than business as usual, and would lower emissions more quickly over the next five years than relying solely on the country’s distant 2060 greenhouse gas target to decarbonise the power sector.
These are the headline findings from the first application of TransitionZero’s energy system modelling platform. We've created the highest resolution, open-access system model ever made publicly available for the Indonesian market, with nodes representing all 34 of the archipelago’s island provinces underpinned by rich asset-level data¹.
TransitionZero previously calculated it would cost $37 billion to buy out or renegotiate power purchase agreements (PPAs) financing Indonesia’s existing coal fleet (see Financing Indonesia's coal phase-out: Coal Asset Transition Tool). When the $20 billion Just Energy Transition Partnership (JETP) for Indonesia was unveiled at the G20 in November last year, TransitionZero’s analysis revealed that this budget could close more than half (21.7 GW) of the country’s coal capacity up to ten years early by targeting the least profitable plants first (see Indonesia’s coal retirement conundrum).
The questions that arose from that analysis were: what is the most cost-effective way of replacing lost coal capacity while ensuring adequate reserve margins? And how would these interventions impact the cost of operating the system? Answering these questions required a high-resolution system model. The findings from our first runs are the subject of this article.
Setting the scene
TransitionZero calculated the lowest cost generation mix in Indonesia’s coal-heavy electricity system under four different sets of assumptions and constraints. The model optimised capacity additions and dispatch at each node in the network to minimise system operating costs every year out to 2050. It also optimised for interconnection between nodes and added transmission capacity as required. The following scenarios were produced for these first model runs.
The Early Coal Retirement scenario removes 21.7 GW of coal capacity up to 10 years early. The plants were ranked by long-term profitability using asset-level data from TransitionZero’s Coal Asset Transition (CAT) tool. This is one of five early coal retirement scenarios explored in previous analysis, which concluded that closing the least profitable assets first would deliver the greatest capacity ‘bang’ from Indonesia’s $20 billion Just Energy Transition Partnership (JETP)². Read more in Indonesia’s coal retirement conundrum.
Getting paid to cut carbon
The most striking conclusion is that early coal retirement is a win-win: there is no zero-sum trade-off to be made between cutting emissions and saving money. Compared to Current Policies (baseline), the Early Coal Retirement scenario avoids 1.3 gigatonnes of CO2 and saves Indonesia $2 billion in the process. This is equivalent to simultaneously halting global shipping industry emissions for two years and boosting Indonesia’s GDP by 0.15%. Expressed another way, Indonesia could save $2 for every tonne of CO2 avoided if JETP capital is allocated cost-effectively.
These figures were calculated as follows. Under the Current Policies (baseline) scenario, the total cost of running the power system between now and 2050 comes in at $1.004 trillion. Total emissions from electricity generation over that period are 8,105 million tonnes of CO2 (MtCO2).
Taking that as our baseline, the model reveals that the total system cost under the Early Coal Retirement + PPA scenario is slightly higher at $1.022 trillion but emissions are one-fifth lower, at 6,970 MtCO2. This equates to a cost of abatement of $16 per tonne of avoided CO2.
This abatement figure might appear low compared to Western markets such as the European Union, where carbon is currently priced at ~€89/t (US$96/t). But it is considerably more than the Indonesian energy system is expecting to absorb. Indonesia launched a carbon trading mechanism in February that is expected to price CO2 between $2 and $18/t. Separately, the government is mulling a carbon tax of IDR 30,000/t (~$2/t) that would be charged only on marginal emissions above and beyond a baseline level.
Fortunately, the $16/tCO2 abatement figure does not tell the whole story. This captures the cost of expanding and running the power system (capital investments, fuel costs, operating expenses etc.) without the 21.7 GW of retired coal capacity, plus the cost of buying out the power purchase agreements (PPAs) underpinning those old coal plants. It assumes that the PPA buyout cost is borne entirely by the system and recovered from consumers.
It is highly likely that PPA buyout costs will be covered at least partially by the JETP, which is being created for this purpose. It may also be used for other purposes. If the $20 billion buyout cost could be covered entirely by the JETP and thus removed from system costs under the Early Coal Retirement scenario, the carbon abatement cost would drop into negative territory.
In this scenario, total system emissions are unchanged at 6,970 MtCO2 and system costs are $20 billion lower at $1.002 trillion. This equates to a negative cost of abatement of -$2/t of avoided CO2 – meaning Indonesia can cut emissions and save money in the process.
This finding illustrates the impact that JETP allocation can have on decarbonisation costs. To make it a reality, international donors and financiers would need to offer soft loans and grants through the JETP and direct these funds entirely towards early coal retirement.
Neither of these are a foregone conclusion. The notion that international banks and financial institutions will pick up the tab on PPA buyout costs is reasonable considering the stated purpose of the JETP. But the JETP framework is broad in scope, and there are unresolved discussions around the use of commercial loans to bankroll JETP activities. The government will be keen to minimise loan interest rates and maximise the portion of funds coming in the form of grants.
Stealing a march on net zero
The model runs reveal the importance of combining coal closures with emissions targets. Under the Early Coal Retirement scenario, power sector emissions drop rapidly in the first five years as those 21.7 GW of capacity start to come offline – more rapidly than under the Net Zero by 2060 scenario. The emissions savings from coal closures peak at 26 MtCO2 below net zero in 2027. Cumulatively over the 2023-28 period, Early Coal Retirement avoids 87 MtCO2 more than optimising Indonesia’s power sector to achieve net zero by 2060.
The main reason for this is Indonesia’s existing coal commitments. There are currently 13 GW of new coal under construction in Indonesia, and the model assumes that all of these plants get built in all scenarios. The difference in Early Coal Retirement is that plant closures outpace planned coal additions over the first five years. This scenario does not include any emission reduction constraint so coal capacity built in 2028-2030 can be used for firm power all the way until 2050, supplementing solar PV deployment as that gets cheaper.
None of these scenarios allow for new coal from 2030 onwards. But in the Net Zero by 2060 scenario there is no incentive to build more coal pre-2030 above and beyond the 13 GW already in the pipeline, because the emission constraint will prevent it from being used. So the model optimises capacity buildout accordingly³.
Limited funds, squandered gains
So far, FEO has revealed that Indonesia could get ahead of the net zero curve by using JETP cash to accelerate coal plant closures before 2030. But it also reveals the shortcomings of relying on the JETP to do all the heavy lifting on reducing emissions.
In short, Indonesia could squander the emissions head-start afforded by the JETP because the $20 billion budget can only go so far. The ‘beyond net zero’ savings of the Early Coal Retirement scenario quickly disappear after the JETP budget is exhausted in 2028 for the reasons outlined above.
A bigger budget for coal PPA buyouts would retire more capacity early and prevent Indonesia from reverting to a net increase in coal after the early retirement funds are spent. This underscores the need for donor countries to not only make good on existing climate finance promises, but also to go above and beyond with more substantial commitments.
Currently, the lack of financing commitments from wealthy nations means that closing the deal on existing JETP promises will be hard enough, let alone negotiating a bigger package. There are competing priorities and jurisdictions all vying for a slice of the limited climate finance available. This needs to change.
The main conclusion to draw here is that Indonesia needs the following: a bold and ambitious overarching energy transition plan targeted towards early coal retirement, a bigger JETP budget from donors without onerous repayment terms and a concrete commitment to its 2060 net-zero target. Crucially, any long-term emissions target must be underpinned by interim carbon budget milestone targets to improve near-term accountability and ensure the power sector stays on track.
The cost of clinging to coal
Regardless of emissions, there are good reasons for Indonesia to optimise its grid around accelerated coal closures. The government is paying to keep an excessive amount of spare generation capacity available across the archipelagic grid network for only very occasional use.
Capacity payments are made to independent power producers (IPPs) to cover their fixed operations and maintenance (O&M) costs and initial capital outlay. These payments keep plants available, even though very high reserve margins indicate that these assets are surplus to requirements.
Reserve margins measure the amount of spare capacity available above peak demand, and are a proxy for overcapacity. Southeast Asia power planners typically target 25-30% reserve margin levels. The average across Indonesia’s five main grid networks is 49%. In Java-Bali, Indonesia’s largest and most populous island province, reserve margins are almost 60%.
The problem is particularly acute on Java-Bali’s Banten subregional network, where reserve margins are 127% more than peak demand. Banten is home to 8.8 GW of coal plant capacity, of which 7.3 GW would be prioritised for early closure under Early Coal Retirement. Our model was configured to reduce reserve margins to 35% by 2030 to drive out excess capacity and improve system efficiency while ensuring a safe generation margin to meet peak demand in all scenarios.
Pivot to PV power
The Indonesian power mix will undergo a radical transformation as coal capacity is removed. Indonesia has been slow to embrace solar, but that will change if the country opts for a cost-optimised pathway for its energy system.
As solar costs fall, PV starts to dominate Indonesia’s generation mix after 2030 in all scenarios modelled by TransitionZero. Utility-scale arrays overtake coal capacity in 2040, supported initially by an expansion of new gas and nuclear capacity. Battery storage does not play a significant role until the 2040s, when costs fall below the marginal cost of dispatchable backup⁴.
Solar PV capacity in 2030 ranges from 15 GW in Current Policies (baseline) to 21 GW (in Net Zero by 2060). It then undergoes a tenfold increase in 20 years to hit between 170 GW (in Least Cost) and 210 GW (in Net Zero by 2060) in the year 2050. This is a blistering ramp-up from the current installed solar base of ~170 MW.
Solar PV generation exceeds coal in all scenarios, including Least Cost, a theoretically unconstrained scenario optimised purely around system costs that allows new coal plants to be built post-2030 regardless of the emissions impact. By 2050, solar PV generation exceeds coal by 23% under Least Cost, and by more than 10x under Net Zero by 2060.
Nuclear power starts to penetrate the mix from 2040 in all TransitionZero scenarios except for Least Cost (unconstrained). By 2050, Indonesia will have between 12 GW and 21 GW of reactors in operation providing between 705 GWh and 1,247 GWh per annum (7% to 12% share of generation).
There are remarkably few differences in the 2050 power mix under each scenario modelled by TransitionZero. In all cases, Indonesia embraces solar PV as the least-cost power source, backed up by significant amounts of battery storage and some combined cycle gas turbine (CCGT) capacity.
Location, location, location
The unrivalled cost-competitiveness of solar energy by mid-century guarantees it a central role in power generation, regardless of emissions targets. The only real question is by how much solar will come to dominate Indonesia’s energy mix – and where.
Our model's high spatial resolution provides some insight into the locations where solar generating capacity will be most cost effective. Coal is today the cheapest power source at most nodes on the network where plant utilisation is around 70-80%. Where coal plants are used less than this, solar energy is already cost-competitive.
Anticipated cost reductions will see PV capacity at nodes with the highest solar irradiance levels begin to undercut even the most highly utilised coal fired power plants from the mid-2030s onwards. The model anticipates this and deploys the majority of new solar capacity at these nodes first.
These findings are conservative because the model assumes that Indonesia will continue to subsidise domestic use of thermal coal for power generation at $70/t, and the government is considering reviewing this policy. The potential for significant coal plant fuel cost increases would merit further sensitivity analysis in our model.
Optimising the JETP debate
Indonesia is a carbon-intensive power market riddled with system-level inefficiencies that make it ripe for optimisation. Our modelling reveals the extent of the cost and emissions savings available to policymakers and regulators, should they choose to redesign the power mix around least-cost decarbonisation principles. The financial benefits of early coal retirement present a compelling case for allocating capital into targeted PPA buyouts as a no-regrets near-term option while the terms of longer-term decarbonisation finance are fleshed out.
This first application of TransitionZero's granular, open-access modelling is a momentous contribution to the debate around coal in Indonesia. The availability of power system modelling at unparalleled spatial and temporal resolutions, combined with unprecedented data granularity, marks a paradigmatic shift in the knowledge base available to decision-makers at this critical moment for the country. The Just Energy Transition Partnership Secretariat have formed working groups that need to complete a road map for early coal retirements and a comprehensive investment plan (CIP) by August this year.
In future, we will allow users to reproduce the scenario analysis undertaken here, and to tweak assumptions as required. The system-level insights it generates will allow the public policy conversation to move beyond the one-dimensional comparison of power sources according to levelised cost of energy (LCOE). While LCOE can be a useful metric for the purpose of financial modelling, its value to energy planning decisions is limited unless accompanied by an understanding of long-term system-level impacts.
This dimension, which TransitionZero's modelling platform can explore at greater depth than other publicly available power systems models, will become a crucial metric for energy planners contemplating the wholesale reorientation of power grids around intermittent zero marginal cost power sources. The fact that solar emerges as the dominant power source regardless of emissions criteria should be appreciated in this context.
And this is only the start. Further insights will be forthcoming on the cost of Indonesia achieving net-zero emissions by 2050, and the role of interconnectors in enabling a cost-efficient transition away from coal across the archipelago.
TransitionZero is expanding its platform to cover the Association of Southeast Asian Nations to explore how power trading and cooperation between ASEAN countries could break down barriers to lowest-cost decarbonisation pathways at the regional level. Looking further ahead, it will be developed to have global coverage and a user-friendly interface that will make cutting-edge systems modelling capabilities more accessible than ever.
Notes
1. Indonesia now has 38 provinces after administrative changes came into effect at the end of 2022. Data availability is limited from the four additional provinces, so our modelling still considers Indonesia as 34 provinces. This discrepancy does not affect the findings of the analysis.
2. Plant profitability is calculated by deducting the long-run marginal cost of generation (LRMC) from the plant’s current tariff price per unit of electricity generated, which is typically set out in a power purchase agreement (PPA). See TransitionZero’s Coal Asset Transition (CAT) open data tool for more details on how these figures are calculated. For more information on assumptions included in each scenario, please refer to the accompanying analysts’ slide deck.
3. From 2030, coal buildout is constrained in the model and emissions fall in both the Net Zero by 2060 and Early Coal Retirement scenarios. In reality, it is highly unlikely that banks would finance any newbuild coal-fired capacity in Indonesia post-2028, particularly if they participated in the JETP fundraising round that bought out up to 21.7 GW of underperforming asset PPAs in the five years prior to that. In fact, coal expansion projects are already running into funding-related obstacles. A recent example was a proposed 2 GW expansion of the Indramayu power station in West Java, which appears to have been cancelled following Japan’s decision to stop financing coal projects in Indonesia due to perceived reputational risks.
4. Battery energy storage costs fall from ~$350/kWh in 2020 to ~$150/kWh by 2050, per NREL projections.
