District Heating Acquisition: Coal Exit, Co-Ownership, and a 2036 Bullet
A European infrastructure fund is bidding for a district heating network serving 210,000 households. Heat supply is 58% from a 190MW coal CHP unit co-owned with an industrial partner, 27% waste-to-energy, and 15% gas peakers. The coal unit's free ETS allowances end in 2026, and national law requires coal exit by 2030. The vendor's model assumes a flat €75/t carbon price, no coal-exit capex, and a 2032 sale at 11x EBITDA. The fund's own case includes €310m of capex for a river-water heat pump, electric boilers and 40MW of biomass, commissioned in 2031, and heat demand falling 8% by 2035 after efficiency retrofits. The purchase is 65% debt with a 2036 bullet and a 1.35x DSCR covenant. The industrial partner holds a right of first refusal over the CHP unit and needs process steam from it. Current EBITDA is €120m. The investment committee must decide whether to bid and at what price.
1. Identify the material flaws in the vendor's model and indicate the direction and rough size of each correction.
Read the worked answer
The vendor model is not a base case; it is a no-transition case with a transition multiple.
1. Flat €75/t carbon. EU ETS prices are not flat and the market has traded well above this level. On the coal unit's emissions - a 190MW coal CHP running for heat could emit on the order of 0.6–0.9 Mt CO2 a year - each €25/t of understatement is €15–22m a year of EBITDA. Direction: overstates EBITDA.
2. Free allocation ending in 2026. This is a known, dated step-down with no offsetting cost in the vendor model. It is a cost increase from year one of ownership. Direction: overstates EBITDA materially.
3. No coal-exit capex. The €310m replacement programme is not optional; it is the price of legal compliance by 2030. On an EV of roughly €1.3bn at 11x, that capex is over 20% of value and must be funded. Direction: overstates equity value, and creates a financing problem.
4. Heat demand down 8% by 2035. Not modelled. Regulated heat tariffs often cap pass-through, so volume loss flows straight to EBITDA. Direction: overstates revenue.
5. 11x exit multiple in 2032. This cuts both ways. A 2032 asset still carrying coal exposure before the 2030 exit is complete will attract a restricted buyer universe and a discount. Conversely, a post-transition, low-carbon, regulated network could command a premium multiple. The vendor gets the benefit of the second without paying for the first - that is the real error.
Net effect. Directionally, the vendor EV is overstated by the NPV of mandated capex plus the carbon and free-allocation step-ups, partially offset by any green-multiple benefit realised post-2031.
2. How should the fund handle the co-owned coal unit and the industrial partner's right of first refusal?
Read the worked answer
This is the deal's central governance risk: the fund may be buying a legal obligation it cannot control the means to satisfy.
The conflict. The industrial partner needs process steam and holds a right of first refusal over the CHP unit. It therefore has an incentive to keep the coal unit running, or at minimum to extract a price for agreeing to closure. The fund, by contrast, faces a 2030 legal deadline and a lender that will price the coal exposure. Two owners with opposite time horizons on an asset with a hard regulatory deadline is a classic deadlock structure.
What to require before bidding.
- Read the shareholder agreement for deadlock, put/call, and cost-sharing provisions on major capex and closure decisions. If the fund cannot compel a 2030 exit, it cannot underwrite the transition case.
- Pre-agree the coal-exit schedule and the replacement steam solution with the industrial partner as a condition of closing, not as a post-closing workstream.
- Quantify the partner's likely claim - the value of its right of first refusal and of steam continuity - as a deduction from the bid, not a hope.
- Obtain a vendor indemnity for any pre-closing regulatory non-compliance at the CHP.
Bid implication. Price the scenario in which the fund must buy out the partner's stake or pay for replacement steam. If that scenario breaks the return, the answer is to walk, or to bid only with a structure that transfers the coal unit out of the acquired perimeter entirely. Buying an asset whose most important decision requires a counterparty's consent is not infrastructure investing; it is an option written to the industrial partner.
3. Given the 65% LTV, the 2036 bullet, and the 1.35x DSCR covenant, how should the fund structure the financing and the exit?
Read the worked answer
The financing and the exit are the same problem: the asset must be refinanceable and sellable after the coal exposure is gone, not before.
Capex funding. €310m of capex against €120m of EBITDA cannot be absorbed by the operating cash flow while servicing 65% LTV debt. If it is debt-funded, DSCR breaks through the 1.35x covenant during construction; if it is equity-funded, the fund's return falls sharply and the bullet becomes harder to refinance. Structure an explicit capex reserve or equity commitment sized to the full €310m plus contingency, and negotiate covenant headroom or a covenant holiday for the construction window. Do not assume the lender will waive.
Covenant design. Push for a DSCR measured on a trailing basis with a defined treatment of construction capex, and consider a sustainability-linked margin ratchet tied to gCO2/kWh of heat delivered - it aligns the lender with the transition and gives the fund a pricing benefit for delivering it.
The 2036 bullet. Refinancing risk is concentrated precisely where the market is least willing to lend: a heating network with residual carbon exposure and a recent capex overrun. Either extend tenor with amortisation, or plan the refinancing after 2031 commissioning when the asset is demonstrably low-carbon.
Exit. A 2032 sale, before the replacement plant has a track record, hands the buyer the construction and commissioning risk and invites a discount. Better economics usually come from holding through 2032–2033, demonstrating actual heat costs and carbon intensity, and selling a regulated, taxonomy-aligned asset to a buyer pool that pays a premium for exactly that. Model both exits and bid on the worse one.