Flood exclusion and CRREM stranding in a logistics portfolio
Jens advises a Danish pension fund on a €1.4bn European logistics portfolio: 22 assets in the Netherlands, northern Germany and Belgium. Flood mapping places six assets in high-risk zones, and from 2026 the insurer excludes stormwater flooding at the two Rotterdam sites after raising premiums 40%. CRREM analysis shows the portfolio's 1.5°C stranding year at 2031 against 2041 on a 2°C pathway, driven by gas heating and 13 assets rated below EPC C. Rents have held, but two tenants now demand green lease clauses and one anchor tenant has moved to a new BREEAM Excellent building. The fund has a 2040 net-zero commitment and reports SFDR principal adverse indicators. The board asks whether to sell the two Rotterdam assets now, retrofit them, or hold and insure. Cap rates for brown assets have widened about 60bp relative to green assets over 18 months, although transaction evidence is thin because volumes are low.
1. How should Jens separate and quantify physical risk and transition risk, and where should each enter the valuation?
Read the worked answer
- Separate the two risks before valuing them. Physical risk is hazard x exposure x vulnerability: flood depth-damage curves, expected annual loss, deductibles and exclusions, business-interruption and reinstatement costs, and the effect on lettability. Transition risk is the carbon and energy performance of the asset against a pathway: CRREM carbon intensity versus the 1.5°C and 2°C budgets, EPC regulation, heating system, tenant demand and retrofit capex.
- Put both into cash flows first. Insurance premiums, uninsured-loss provisions, void periods, rent incentives, retrofit capex and green-lease costs belong in the cash-flow forecast. Only then ask whether the market is pricing a systematic premium into cap rates.
- Avoid double counting. If the cash flow already carries the 40% insurance increase and a retrofit provision, adding a further 60bp cap-rate penalty for the same risk charges the investor twice. The brown-versus-green cap-rate widening is partly market pricing of transition risk, but with thin transaction evidence it should be a sensitivity, not a fact.
- Scenario DCF. Run Net Zero 2050 and Current Policies scenarios with different carbon prices, EPC compliance dates, insurance availability and rents; report a value range. CRREM is a carbon model, not a valuation model, and its stranding year depends on grid decarbonisation and energy-intensity assumptions.
- Data quality. Check whether energy data is actual or modelled, whether it covers landlord or tenant-controlled consumption, and whether floor areas are defined consistently across the 22 assets.
2. Evaluate the three options for the two Rotterdam assets, and state what evidence would change your recommendation.
Read the worked answer
- Sell now. Crystallises the cap-rate loss and may require a discount because the buyer pool narrows once the insurance exclusion is known. It removes tail risk and retrofit cost, but gives up green reversion and may be criticised as selling the problem to a less informed buyer.
- Retrofit. Requires capex estimates, permits, tenant disruption and lease timelines. The split incentive is decisive: if leases are full-repairing and tenants pay energy bills, the landlord captures little of the benefit. A green lease with cost-sharing, a rent uplift or a longer term changes the maths. Evidence for a green premium suffers from selection bias, so use matched transactions rather than headline averages.
- Hold and insure. The insurer has excluded stormwater flooding at the two sites, so the fund retains a tail risk that may be uninsurable or only partly covered by parametric insurance. For a pension fund that may be inconsistent with prudent risk management even if expected annual loss is small.
What would change the recommendation: a credible flood defence or public investment, a tenant willing to sign a 10-year green lease, a retrofit quote with payback inside the hold period, or a credible buyer at a small discount. Without those, selling or ring-fencing the two assets is defensible; a portfolio-wide retrofit programme should be assessed separately on its own payback and carbon impact.
3. What must the fund disclose about these assets, and where are the data-quality traps?
Read the worked answer
- PAI and real-estate metrics. Use the real-estate GHG intensity indicator and state whether emissions are landlord-controlled or tenant-controlled and actual or estimated. Coverage matters: if 20% of floor area lacks energy data, the reported intensity is not comparable with peers.
- EU Taxonomy. A renovation is aligned only if it delivers the required improvement in energy performance, meets DNSH (including water and circularity) and respects minimum safeguards. Do not claim alignment for a refurbishment that merely replaces a boiler.
- Physical-risk disclosure. Under TCFD, ISSB S2 and CSRD, describe flood exposure, the insurance exclusion and mitigation. An insurance exclusion is a material risk factor, not a disclosure footnote.
- Traps. Modelled energy data presented as actual; CRREM results quoted without pathway assumptions; floor-area definitions that differ between assets; estimated taxonomy alignment presented alongside reported figures without labelling; and limited assurance that covers only selected datapoints. Each of these should be disclosed, not smoothed over, because clients and auditors will compare the fund's numbers with the underlying asset data.