The Abatement Cost Curve: Which Measures to Do First

Imagine two proposals land on your desk in the same week. One is a RM40,000 LED retrofit for the warehouse. The other is a RM2.8 million biomass boiler to replace your gas-fired one. They have nothing in common — different sizes, different fuels, different lifespans. How do you decide which comes first?

The marginal abatement cost curve — everyone says MACC — is the one tool that puts both on the same ruler. That ruler is ringgit per tonne of carbon dioxide avoided. Once every measure is priced in RM/tCO2, you can line up your entire shopping list cheapest-first and see, at a glance, where your money does the most good. And here is the part that surprises most building owners: a large chunk of that list has a negative price. You get paid to do it.

The one number: marginal abatement cost

The marginal abatement cost (MAC) of a measure is simple:

$$\text{MAC} = \frac{\text{Annualised incremental cost (RM/yr)}}{\text{Annual CO}_2 \text{ avoided (tCO}_2\text{/yr)}}$$

"Incremental cost" means the extra cost of doing the measure versus doing nothing: the annualised capital outlay, plus any change in running costs, minus the energy you stop buying. When the energy savings outweigh the annualised capital, the numerator goes negative — and so does the MAC.

A negative MAC is the whole game. It means the measure saves money and cuts carbon at the same time. These are the "no-regret" measures: you would do them even if climate change were not a concern, purely because they are profitable. On a MACC they sit below the zero line, and they should be first in the queue every time.

Annualising fairly, so lifespans don't lie

You cannot compare a 5-year LED driver with a 20-year solar array by their sticker price. A one-off RM40,000 must be spread across the years it serves. We do that with the capital recovery factor (CRF) from Part 12 (Life-Cycle Costing) — the same factor that turns a lump sum into an equivalent annual payment at your discount rate:

$$\text{Annualised Capex} = \text{Capex} \times \text{CRF}$$

At this course's canon discount rate of 8%, the CRF depends only on the project life:

| Project life | CRF @ 8% |
|---|---|
| 5 years | 0.2505 |
| 10 years | 0.1490 |
| 15 years | 0.1168 |
| 20 years | 0.1019 |

So RM40,000 of LEDs living 10 years costs \(40{,}000 \times 0.1490 = \text{RM}5{,}960\) per year — an honest annual figure you can set against the annual savings. Get this step right and short-life, cheap measures stop looking artificially expensive next to long-life, pricey ones.

A worked shortlist: four Malaysian measures

Let's price a real shopping list for a mid-sized Malaysian factory. Canon assumptions: electricity RM0.45/kWh, natural gas RM38/GJ, biomass RM35/GJ, discount rate 8%, grid emission factor 0.585 kg CO2/kWh, and natural gas 0.0561 tCO2/GJ. Biomass is counted as carbon-neutral (the carbon it releases was absorbed while growing). If the RM0.45/kWh feels abstract, our kW vs kWh explainer sets out exactly what you are buying.

Measure 1 — Boiler combustion tune-up. Capex RM5,000, life 5 years. It trims 100 GJ of gas a year.
- Annualised capex: \(5{,}000 \times 0.2505 = \text{RM}1{,}252.50\)
- Fuel saved: \(100 \times 38 = \text{RM}3{,}800\)/yr; CO2 avoided: \(100 \times 0.0561 = 5.61\) tCO2/yr
- Net annual cost: \(1{,}252.50 - 3{,}800 = -\text{RM}2{,}547.50\)
- MAC = \(-2{,}547.50 / 5.61 = -\text{RM}454/\text{tCO}_2\)

Measure 2 — LED retrofit. Capex RM40,000, life 10 years, saves 27,700 kWh/yr.
- Annualised capex: \(40{,}000 \times 0.1490 = \text{RM}5{,}960\)
- Electricity saved: \(27{,}700 \times 0.45 = \text{RM}12{,}465\)/yr; CO2 avoided: \(27{,}700 \times 0.585 / 1{,}000 = 16.2\) tCO2/yr
- Net annual cost: \(5{,}960 - 12{,}465 = -\text{RM}6{,}505\)
- MAC = \(-6{,}505 / 16.2 = -\text{RM}402/\text{tCO}_2\)

Measure 3 — 100 kWp rooftop solar PV. Capex RM381,000 (about RM3.81/Wp), financed over 10 years, generates 140,000 kWh/yr.
- Annualised capex: \(381{,}000 \times 0.1490 = \text{RM}56{,}769\)
- Electricity saved: \(140{,}000 \times 0.45 = \text{RM}63{,}000\)/yr; CO2 avoided: \(140{,}000 \times 0.585 / 1{,}000 = 81.9\) tCO2/yr
- Net annual cost: \(56{,}769 - 63{,}000 = -\text{RM}6{,}231\)
- MAC = \(-6{,}231 / 81.9 = -\text{RM}76/\text{tCO}_2\)

Measure 4 — Biomass boiler replacing the gas boiler, displacing 10,000 GJ/yr of gas. Capex RM2,820,000, life 15 years. Biomass is cheaper per GJ but the boiler runs a little less efficiently (11,000 GJ of biomass input) and adds fuel-handling labour.
- Annualised capex: \(2{,}820{,}000 \times 0.1168 = \text{RM}329{,}376\)
- Biomass fuel: \(11{,}000 \times 35 = \text{RM}385{,}000\); gas avoided: \(10{,}000 \times 38 = \text{RM}380{,}000\); fuel delta +RM5,000
- Extra O&M: +RM60,000/yr
- CO2 avoided: \(10{,}000 \times 0.0561 = 561\) tCO2/yr (biomass counted neutral)
- Net annual cost: \(329{,}376 + 5{,}000 + 60{,}000 = \text{RM}394{,}376\)
- MAC = \(394{,}376 / 561 = +\text{RM}703/\text{tCO}_2\)

Reading the curve

Now stack them cheapest-first and you have your MACC:

| Rank | Measure | MAC (RM/tCO2) | Abates (tCO2/yr) |
|---|---|---|---|
| 1 | Boiler tune-up | −454 | 5.61 |
| 2 | LED retrofit | −402 | 16.2 |
| 3 | Solar PV | −76 | 81.9 |
| 4 | Biomass boiler | +703 | 561 |

Picture it as a bar chart. Each bar's width is the tonnes that measure avoids each year; its height (or depth, below zero) is RM per tonne. Read left to right, the first three bars hang below the zero line — that's 103.7 tonnes of carbon a year abated while the plant pockets RM15,284. This is the "free money" region. Efficiency and controls measures — tune-ups, LEDs, and the like — almost always land here.

The fourth bar shoots up above the line to +RM703. Fuel-switching and carbon capture typically do. Biomass cuts far more carbon than the first three combined (561 tonnes), but it costs you RM703 for every tonne. So when do you pull that trigger?

Draw a horizontal line at your carbon price — whether that's an internal price you've set, a future carbon tax, or the market price of an offset. Every bar below that line is cheaper than the alternative and worth doing yourself. Every bar above it is where internal abatement stops and buying offsets begins. If offsets cost RM150/tCO2, you do the first three measures without hesitation and you do not build the biomass boiler yet — buying 561 tonnes of offset at RM150 is cheaper than abating them at RM703. Raise the carbon price to RM750, or add a government mandate, and the biomass boiler suddenly makes sense. The curve doesn't decide for you; it shows you exactly where the crossover sits. (This is also why Part 1's 5R hierarchy puts offset dead last — you exhaust the cheap real cuts before you pay anyone else to cut for you.)

The limits — read these before you present the curve

A MACC is seductive precisely because it's so tidy. Respect three limits or a sharp CFO will unpick it in the meeting:

- It ignores project interactions. The bars are drawn as if each measure is independent, but they aren't. Install the LEDs and they dump less heat into the space, so a separate chiller-optimisation project now has less waste heat to remove — its savings, and therefore its own attractive MAC, shrink. Do both and the total saving is less than the sum of the two bars. Sequence-dependent savings are the MACC's biggest blind spot.
- It is only as good as its assumptions. Every bar rests on an electricity price, a fuel price, a discount rate, and an emission factor. Nudge electricity from RM0.45 to RM0.55 and every electrical measure slides further into "free money." Change the discount rate and every CRF — and every capital-heavy bar — moves. A MACC is a snapshot at today's prices, not a law of nature.
- It does not replace NPV or life-cycle costing. The MAC collapses a whole multi-year cash flow into one number per measure. That's perfect for ranking, useless for deciding. Once the curve hands you a shortlist, you still run the full NPV and life-cycle cost from Part 12 on each survivor before you sign anything. Think of the MACC as the bouncer at the door, not the investment committee.

Where the free money hides

The genuinely useful message of a MACC is that the left-hand side — the profitable side — is usually longer than anyone expects. The catch is that those negative-cost measures are also the least visible. A biomass boiler is a board-level project; a badly tuned burner or a bank of overlit corridors is invisible until something measures it.

That is the quiet argument for continuous monitoring. When you meter energy, indoor air, water, and chilled water across a building, the small, cheap, negative-cost wins surface first — the compressor short-cycling at 2 a.m., the lighting running in an empty wing, the boiler drifting off its optimal air-fuel ratio. Cobler's CobiNeural platform is built to reveal exactly those no-regret items before you spend a ringgit on the expensive end of the curve. (And your maximum-demand calculator will tell you what those kW cuts are worth on the tariff.) You can't rank what you can't see, and a MACC drawn on guesses ranks guesses.

The takeaway

The marginal abatement cost curve puts every measure — a RM5,000 tune-up and a RM2.8 million boiler — on one ruler: ringgit per tonne of CO2 avoided. Annualise each measure's capital fairly with the CRF, work out its net annual cost, divide by the tonnes it saves, and sort cheapest-first. The measures below zero are free money: do them today, no carbon price required. The measures above zero wait for a carbon price or a mandate to clear the bar. Just remember the curve is a shortlist, not a verdict — it tells you the order, then NPV and life-cycle costing tell you the answer.

Next up — Part 14: Sensitivity Analysis and the Tornado Diagram, where we stop trusting a single number and ask which assumption, if it's wrong, hurts the most.