Infrastructure Risk-Adjusted Returns: Why Financial Models Get It Wrong

Infrastructure has delivered impressive risk-adjusted performance over the past decade, with Sharpe ratios between 1.05 and 1.58 - roughly double that of public equities. Yet a critical problem undermines these headline figures: a growing disconnect between modelled returns and actual performance, driven by assumptions that no longer reflect a changing physical environment.

The Core Issue

Standard infrastructure valuations rely on historical yield data, linear degradation curves, and stable operating cost projections. These foundational assumptions fail when climate conditions shift materially. Solar assets underperform P50 forecasts by approximately 5%, while wind portfolios lag expectations by 5-10% in some regions. The causes - equipment failures, accelerated degradation, inverter shutdowns during extreme heat - are systematic, not anomalous.

As one infrastructure CIO noted: “The underwriting was over optimistic. People that have been in the industry for long enough noticed that 10 to 5 years ago.”

Gross vs. Net Loss: An Insurance Blind Spot

Infrastructure investors struggle to distinguish between gross average annual loss (total expected damage) and net loss (remaining exposure after insurance). This gap is widening dangerously as insurers withdraw from climate-exposed categories and premiums rise 18-30% annually.

A physical risk modeller at a major European bank highlighted the practical challenge: data access barriers prevent accurate net exposure calculation due to confidentiality regulations, forcing investors to either overstate risk using gross figures or assume insurance as a permanent backstop - both flawed approaches.

Climate Risk Remains Qualitative

Despite conducting climate risk screening during due diligence, most infrastructure funds treat findings as qualitative assessments rather than financial inputs. Climate data generates risk flags and heatmaps but rarely adjusts discount rates, operating expenditure projections, or exit multiples.

A portfolio manager explained plainly: “Those are clearly interesting data points. From a valuation, exit or acquisition scenario they will be treated as qualitative.”

This parallel processing - where climate analysis runs separately from financial modelling - leaves risk systematically unpriced in deal valuations.

The P50 Distribution Problem

While P50 forecasts represent median expected output, the distribution around that median is shifting under climate change. Research found that “the spread from P10 to P90 can show 3 to 4% degradation difference over the asset lifetime,” even when median values appear stable.

Infrastructure returns are asymmetric: modest upside from favorable conditions versus severe downside from structural yield erosion. A portfolio with tight P10-P90 variance differs fundamentally from one with wide distribution, yet standard Sharpe ratio calculations flatten this distinction.

Climate-Adjusted Valuation Framework

Rather than debating discount rate adjustments, practitioners should directly modify cash flow assumptions:

• Model how operating expenditure trends upward under increased extreme weather frequency
• Replace historical yield averages with forward-looking climate risk data
• Quantify adaptation CapEx with defined payback periods
• Employ Decoupled Net Present Value methodology separating risk from time discounting

Research shows resilience projects deliver average returns of 27% over ten years, yet only 8% of investment appraisals currently estimate full monetized resilience benefits.

Insurance Market Implications

Insurers absorbed over $137 billion in weather-related losses in 2024, triggering premium increases, coverage restrictions, and geographic market exits. Rising insurance costs directly erode operating cash flows while coverage gaps transfer risk to asset owners.

Assets becoming uninsurable face liquidity discounts at exit. This trajectory threatens the assumption of stable insurance costs embedded in most infrastructure financial models covering 10-20 year holding periods.

The infrastructure sector’s competitive advantage will accrue to funds integrating climate risk into both cash flow numerators and risk denominators, pricing physical exposures more accurately than competitors relying on climate-agnostic models.