European Utility DCF: Carbon in the Cash Flows or a WACC Premium?

Cash flows, with only a narrow and disclosed role for the discount rate.

Why cash flows. Carbon prices are a cost that varies by asset, fuel, and year. A single WACC add-on applies the same penalty to the nuclear and hydro units, which benefit from high carbon prices, as to the lignite units, which are hurt. It also makes the valuation unauditable and incomparable with consensus, which is precisely what the PM needs to judge cheapness.

What WACC is for. The discount rate should capture undiversifiable, systematic risk. Most carbon price risk here is company-specific, forecastable, and hedgeable through dispatch and hedging - it belongs in the forecast. There is a genuine systematic component: the risk that policy tightens faster than the market expects, hitting all carbon-exposed assets at once. A modest premium for that is defensible, but it must be small, separately justified, and disclosed - not a 150bp blanket.

Practical method.

1. Build a unit-level dispatch model with a carbon price path (not a flat rate), pass-through assumptions, and declining lignite load factors.
2. Include the free-allocation phase-out as an explicit cost line through 2030.
3. Show the valuation sensitivity per €10/t on enterprise value - this is what answers the PM's question, because the €60 vs €85 gap is worth far more than any WACC tweak.
4. Treat terminal value carefully: 68% of EV means the terminal carbon assumption dominates. Do not assume a flat €85 forever without stating why; model the post-2035 merit order.
5. Add a documented, small policy-risk premium only after the above, and show value both with and without it.

The 150bp proposal is not simpler - it is a way of hiding an assumption that the PM will need to defend to an investment committee.

Free allocation is a subsidy. Its removal is a cost increase, but not a one-for-one hit.

Size the gross cost. If the lignite units currently receive, say, 60% of their allowances free and that falls to zero by 2030, the ultimate annual cost is roughly 14 Mt × €85 × 60%, on the order of €700m a year at full exposure. Even at half that, it dwarfs most line items in the model. Compare this directly against consensus: a €25/t difference on 14 Mt with 60% exposure is several hundred million euros a year.

Then model pass-through. Under marginal pricing, the carbon cost of the marginal plant is embedded in the power price, and free allowances carry an opportunity cost - historically, utilities recovered much of the cost this way. Three things erode that recovery here:

1. As renewables enter the merit order, lignite sets the price less often.
2. The units that lose free allocation are increasingly inframarginal, so they absorb the cost rather than passing it on.
3. Regulated or contracted volumes cap pass-through in some markets.

So model three quantities together: cash cost of uncovered emissions, running hours, and the residual asset value of the lignite fleet. A utility that gets the carbon price right but holds load factors flat will overvalue the asset badly.

Do not forget closure costs. Mine reclamation, severance, and just-transition obligations are real cash outflows that appear only in a credible phase-out scenario. They are the reason a 'run it to 2038' case often looks worse than an early closure.

Presentation. Show a bridge from consensus: carbon price gap, free allocation treatment, load factor, and terminal assumption. That tells the PM whether the stock is cheap or whether consensus is simply using a different, defensible carbon path.

Separate two questions: is the bond legitimately green, and does it move the valuation?

Eligibility. Grid upgrades are clearly eligible - they are core enabling infrastructure. New gas-fired backup is contested. Enabling-gas criteria in recognised taxonomies are narrow: an emissions threshold around 100 g CO2e/kWh, hydrogen readiness or blending commitments by a set date, capacity-factor limits, and a demonstrated role in displacing coal rather than extending it. If the units do not meet those conditions, they should not sit in the green pool. The critical distinction is whether the gas is backup for a renewables-heavy system or a life-extension for the lignite fleet - the framework language should make that verifiable, not aspirational.

Recommendation. Ring-fence the use of proceeds to grid and renewables; move gas to a separate transition or sustainability bond with a stated pathway and a cap; obtain a second-party opinion that addresses the gas question directly; commit to allocation and impact reporting.

Valuation effect. Small and easily overstated. The greenium in European utility issuance is typically a few basis points; capitalising it into the DCF moves value by an amount well below the carbon-price sensitivity. The larger effects are demand from dedicated buyers and the reporting discipline the framework imposes.

Risk effect - this is where it matters. A mislabelled use-of-proceeds is one of the fastest ways to convert a low-probability legal issue into a high-probability headline, and it contaminates the company's wider transition narrative. Treat framework quality as a governance input to the position size, not as a line item in the cash flow.