Real Assets & Infrastructure Report 2018 — Buying a Cash Flow, Not a Sector

Institutional capital flowed into infrastructure and real assets substantially through this period, and the thesis is worth taking seriously because for the right assets it is genuinely sound.

The claim: infrastructure assets — utilities, toll roads, airports, pipelines, communications towers — produce stable, long-dated, inflation-linked cash flows from essential services with high barriers to entry. That profile matches the liabilities of pension funds and insurers unusually well: long-dated, inflation-sensitive obligations.

The match is real, and it is the strongest argument in the asset class's favour. It is a liability-matching argument, not a return argument, and conflating the two is where most of the analytical trouble in this asset class begins.

The difficulty is that the thesis applies to a much narrower set of assets than the label covers.

What actually produces the promised profile is the contract structure, not the physical asset. A regulated utility with a tariff set by a regulator to provide a defined return on capital has stable cash flow because a regulator has determined it should. A toll road whose revenue depends on traffic volume has cash flow that depends on economic activity and on competing routes.

Both are infrastructure. They are entirely different investments, and the difference is in the contract, not the concrete.

As competition for genuinely contracted assets intensified and prices rose, capital extended toward assets with more demand risk — financed on infrastructure terms, at infrastructure valuations, with infrastructure leverage, while carrying materially different risk.

The contract is the asset

This distinction is the substance of infrastructure analysis, and it is worth setting out as a taxonomy because the categories behave differently in every scenario.

Regulated assets. A regulator sets prices to allow a defined return on invested capital. Revenue is largely independent of demand within a range. The primary risk is regulatory — that the framework changes at a review. This is the closest thing to a bond-like profile in the asset class.

Availability-based contracted assets. The owner is paid for making an asset available, regardless of use. A hospital built under a long-term government contract is paid whether or not it is at capacity. Risk is counterparty risk — will the government continue to pay — plus operational risk of failing to meet availability standards.

Volume-contracted assets. The owner is paid per unit of use, under a long-term contract with a defined counterparty. A pipeline with take-or-pay contracts is paid a minimum whether or not volume flows. Risk is counterparty credit and contract renewal.

Merchant assets. Revenue depends on market prices and volumes with no contractual protection. A power plant selling into a wholesale market, or an airport dependent on passenger numbers. Risk is demand and price risk — which is equity risk, not infrastructure risk.

A regulated water utility and a merchant power plant are both "infrastructure". One has bond-like cash flows and the other has equity-like ones. The category tells you nothing; the contract tells you everything.

Where the assessment usually goes wrong: a fund describing itself as core infrastructure may hold assets across this spectrum. The reported return and volatility of the fund blend them. An investor allocating on the strength of the stable-cash-flow thesis may be holding substantial merchant risk without having assessed it — and the category label actively obscures the distinction rather than revealing it.

Inflation linkage is a contract term

Inflation protection is central to the institutional case, since the liabilities being matched are inflation-sensitive. It is also more variable than the category-level claim suggests.

Where linkage is genuine:

• Explicit contractual escalation, where a tariff or payment rises with a stated index at a stated frequency. This is the strongest form and it is verifiable by reading the contract.
• Regulated return frameworks that account for inflation in setting allowed returns, either directly or through indexation of the asset base.
• Pricing power in an asset where the owner can raise prices with inflation because customers have no alternative.

Where it is weaker than claimed:

• Caps and collars. Many escalation clauses cap the annual increase. In low inflation this never binds; in high inflation it binds precisely when the protection is most needed.
• Index mismatch. A contract linked to consumer price inflation held by an investor whose liabilities track wage inflation is only partially matched.
• Timing lags. Annual resets mean a year of inflation is absorbed before the adjustment arrives.
• Cost inflation offset. Revenue may rise with inflation while costs — labour, materials, energy — rise faster, so margins compress even with revenue linkage.
• Rate sensitivity working against it. Long-dated cash flows are highly rate-sensitive. If inflation prompts higher rates, the discount rate rises, and the valuation effect may exceed the cash flow benefit.

That final point is the one that surprised allocators in 2022. An asset can have genuine inflation-linked cash flows and still fall in value when inflation rises, if the rate response dominates the cash flow response. Inflation linkage protects the cash flow, not the valuation, and the two are different things.

The practical guidance: inflation linkage claims should be assessed contract by contract, with the mechanism, index, frequency, cap and lag identified. Category-level statements are not reliable, and the difference between two assets in the same fund can be larger than the difference between asset classes.

Duration matching is the real argument

The strongest case for institutional infrastructure ownership is not return. It is liability matching, and it deserves to be stated on its own terms rather than blended with return arguments.

The problem it solves. A pension fund has obligations extending decades. Matching them requires assets with similarly long duration. Long-dated government bonds do this but, in the environment described in the 2016 global report, yielded too little to meet the required return.

Why infrastructure fits. A regulated utility or a long-term contracted asset produces cash flows over decades, with a degree of inflation linkage, at a return above government bonds. That is close to what the liability requires.

Why illiquidity matters less here than elsewhere. For most investors, illiquidity is a cost — as the 2015 private equity report describes. For an investor with genuinely long-dated liabilities and no need to sell, it is closer to a non-issue. They are being paid a premium for accepting a constraint that does not bind them.

This is the strongest structural argument in the asset class, and it is worth distinguishing from the weaker version frequently offered. The weak version is that infrastructure offers attractive risk-adjusted returns, which invites the appraisal-smoothing objection the 2015 private equity report raises — infrastructure valuations are appraisal-based and the same statistical bias applies.

The strong version is that infrastructure matches a specific liability profile better than the alternatives, and that the match is worth something independent of the return statistics. That argument survives the smoothing objection entirely, because it does not rest on volatility measurement.

Category drift and the energy transition

Two developments changed the asset class in this period, and both pushed in the direction of more risk under the same label.

Category drift. As competition for genuinely contracted assets raised prices, capital extended toward assets with more demand risk. Data centres, communications towers, fibre networks, student accommodation and various operating businesses with physical assets were financed on infrastructure terms.

Some of these have genuinely infrastructure-like characteristics — long-term contracts with creditworthy counterparties, high barriers to entry, essential-service economics. Others are operating businesses with real estate attached.

The risk is not that these are bad investments. It is that assets with demand risk were financed with infrastructure leverage at infrastructure valuations. Leverage sized for a contracted cash flow is too much leverage for a merchant one, and the mismatch only becomes visible when demand disappoints.

The energy transition introduced a structurally different risk profile at scale. Renewable generation has attractive features — low marginal cost, long asset life, policy support — and one that differs fundamentally from traditional utility economics: much of it carries merchant price risk.

A regulated utility earns a return set by a regulator. A renewable generator selling into a wholesale market earns the market price. Where a long-term offtake contract exists, the profile resembles contracted infrastructure. Where it does not — or where the contract expires with decades of asset life remaining — the profile is merchant.

Two further features distinguish it:

• Capital intensity is front-loaded. Nearly all the cost is incurred before any revenue, so the return depends heavily on assumptions about prices decades out.
• Technology risk exists, which traditional infrastructure largely lacks. Improving efficiency in new equipment can undercut the economics of existing assets — an unusual risk for a category chosen for its stability.

Neither is an argument against the investments. They are an argument that the analysis required differs from traditional infrastructure analysis, and that applying the traditional framework — stable, contracted, inflation-linked — to a merchant renewable asset produces a materially wrong assessment.

Why leverage sized for a contract fails on a merchant asset

The category drift described above produced a specific and predictable failure mode, and it is worth setting out because it is the mechanism by which infrastructure investments go wrong.

How leverage is sized on a contracted asset. A regulated utility or an availability-based contract produces cash flow that is predictable within a narrow range. Lenders will advance a high multiple of that cash flow, because the probability of it falling short is low. High leverage is appropriate and it raises equity returns substantially.

What happens when the same leverage is applied to a merchant asset. A merchant asset's cash flow depends on market prices and volumes. The expected cash flow may be the same; the distribution around it is far wider.

The consequences follow directly:

• The debt service coverage that looked comfortable at the expected case is thin in the adverse case. Leverage sized to a narrow distribution is over-sized for a wide one, even at the same expected cash flow.
• The equity is wiped out faster. Higher leverage means a smaller equity cushion, so a given shortfall in cash flow consumes proportionally more of it.
• Refinancing risk compounds it. A merchant asset facing a refinancing in weak market conditions may find lenders unwilling to advance the same multiple, requiring an equity injection at exactly the wrong moment.

Why the mismatch occurs rather than being obvious. The asset is labelled infrastructure. Infrastructure carries high leverage. Lenders and equity investors alike apply the category's conventions rather than examining the specific cash flow distribution. The label does the sizing, and the label is wrong.

The check is straightforward and requires reading the contract: what proportion of the cash flow is contracted, for how long, with what counterparty, and what happens when the contract expires. An asset with fifteen years of contracted revenue and a thirty-year life is contracted for half its life and merchant for the other half — which is a materially different investment from one contracted throughout.

Leverage should be sized to the distribution of the cash flow, not to the category of the asset. The two coincide for genuinely contracted infrastructure and diverge sharply for everything that has been financed as though it were.

What an allocator could act on

Read the contract, not the sector. Regulated, availability-based, volume-contracted and merchant are four different risk profiles sharing one label. A fund describing itself as core infrastructure may hold assets across the whole spectrum, and its reported return and volatility blend them.

Assess inflation linkage asset by asset. Identify the mechanism, the index, the frequency, the cap and the lag. Caps bind precisely when protection is most needed. Index mismatch — consumer price linkage against wage-linked liabilities — leaves a residual exposure. Category-level claims about inflation protection are not reliable.

Distinguish cash flow protection from valuation protection. An asset with genuine inflation-linked cash flows can still fall in value when inflation rises, if the discount rate response exceeds the cash flow benefit. This surprised allocators in 2022 and it is arithmetic: long-dated cash flows are highly rate-sensitive, and inflation frequently brings higher rates.

Make the duration-matching argument on its own terms. The strongest institutional case for infrastructure is that it matches a specific liability profile — long-dated, inflation-sensitive — better than the alternatives. That argument survives the appraisal-smoothing objection entirely, because it does not rest on volatility measurement. The weaker version, that infrastructure offers attractive risk-adjusted returns, does not.

Treat illiquidity as near-costless only where it genuinely is. For an investor with long-dated liabilities and no need to sell, the illiquidity premium is compensation for a constraint that does not bind. For anyone else it is compensation for a real cost that arrives in the worst conditions, per the 2015 private equity report.

Assess renewable generation for merchant exposure and technology risk. Front-loaded capital means the return depends on price assumptions decades out. Improving efficiency in new equipment can undercut existing assets — an unusual risk for a category chosen for stability, and one the traditional infrastructure framework has no place for.

What 2018 established for real assets

• Contract structure, not asset type, determines the risk profile, and the category label obscures rather than reveals it.
• Inflation linkage is a contract term requiring asset-by-asset assessment, with caps, lags and index mismatch as the common failure points.
• Inflation protection covers cash flow, not valuation — a distinction that became expensive in 2022.
• Duration matching is the strongest institutional argument, and it survives the appraisal-smoothing objection that return-based arguments do not.
• Category drift and the energy transition introduced merchant risk under an infrastructure label, financed with infrastructure leverage.