What Network Averages Hide About Health Provider Capital Risk

The Capital Question a Network Average Cannot Answer

Capital allocation for a health treaty usually starts with an expected loss ratio, and that loss ratio usually starts with a network discount average. The problem is that the average was never built to answer a capital question, it was built to summarize a pricing relationship for reporting purposes. When actuaries and finance teams treat it as a reliable capital input, they inherit all the variation the average was designed to smooth over.

Why Does Hidden Provider Price Variation Raise Capital Allocation Questions?

Because capital held against a treaty is sized to cover expected losses plus a margin for adverse deviation, and that margin depends on how much the actual claims experience could plausibly differ from the average. If the true underlying provider price distribution is far wider than the blended average suggests, the capital margin calculated from that average is systematically too thin. That is not a rounding error, it is a structural mismatch between what the capital model assumes and what the claims data would show if examined at the provider level instead of the network level.

What Does This Look Like in Practice?

A treaty priced off a network average showing, for example, a consistent 35 percent discount off billed charges looks stable and low-risk on paper. But that stability is an illusion if the underlying billed charges themselves vary by a factor of two or three between providers in the same network, which research on hospital pricing has repeatedly found to be the case. The discount percentage staying flat tells a capital model nothing about whether the dollar claims cost is also staying flat.

How Does Undetected Provider Price Variation Affect Combined Ratio?

It shows up as loss ratio deterioration that was never priced for in the treaty terms. A portfolio that drifts toward higher-cost providers, whether through member choice, referral patterns, or network changes, pays more per claim without any corresponding change to premium or reserving assumptions. That deterioration often gets misread as general medical trend acceleration, when the real driver is narrower: a shift in which providers are actually being used within a network that looks unchanged on paper.

What Capital Allocation Question Should Actuaries Be Asking?

Whether reserves and capital held against a health treaty reflect the full percentile spread of actual provider pricing, not just the midpoint average discount rate reported by the network or third-party administrator. The RAND Hospital Price Transparency Study found commercial hospital prices averaging 254 percent of Medicare, with system-level figures ranging from roughly 150 percent to over 400 percent, a spread wide enough that a capital model anchored only to the 254 percent midpoint will misprice both ends of the distribution. An actuary asking this question needs percentile-level data, not just the aggregate average, to size a defensible capital margin.

How Large Is the Potential Per-Case Capital Exposure?

Large enough to matter at the individual claim level, not just in aggregate. Analysis from Serif Health on percent-of-billed contract structures found that some arrangements implied $150,000 to over $300,000 in overpayment per case relative to a fixed fee schedule benchmark, and that percent-of-billed contracts were costlier than fixed fee schedules in over half of cases reviewed, with the disparity exceeding 70 percent of cases at some health systems. For a high-volume health portfolio, that per-case exposure compounds quickly, and it is exactly the kind of tail risk a blended average discount figure is structurally unable to surface.

Should Capital Models Treat Network Discount Rates as a Stable Input?

No, because billed charges are a moving baseline rather than a fixed reference point. A stable discount percentage can still produce rising dollar costs if the underlying chargemaster rates increase or the provider mix shifts toward higher-baseline facilities, both of which happen routinely and neither of which shows up if a capital model only tracks the discount rate itself. Capital models that hold the discount percentage constant while ignoring the billed-charge baseline are effectively assuming away the exact risk this whole problem is about.

How Should Reinsurers Reprice Capital Allocation for This Risk?

By segmenting capital charges by provider price percentile rather than relying on a single blended network assumption. A treaty concentrated in a higher-cost provider segment, even within a network carrying a favorable average discount, should carry proportionally more capital support than one concentrated in lower-cost providers. This kind of segmentation is the natural next step after the diagnosis covered in the executive risk inside health provider inflation hidden by averages, moving from identifying the blind spot to actually pricing around it.

What Role Does Claims Leakage Play in This Capital Picture?

It compounds the provider price variation problem with a second, often harder-to-detect layer of cost. Leakage from coding errors, billable-but-unnecessary services, or outright fraud adds capital strain on top of legitimate but unpriced provider price dispersion, and the two sources of cost are frequently confused with each other in claims analysis. The scale of this compounding risk is explored further in claims leakage in high-volume health portfolios, which addresses the operational side of the same underpricing problem covered here from the capital angle. Claims Leakage Impact AI Agent can help separate genuine provider price variation from leakage-driven cost, giving actuaries a cleaner signal for capital modeling purposes.

Who Should Own This Capital Allocation Review?

The chief actuary and chief financial officer jointly, since closing this gap requires both actuarial percentile analysis and capital modeling expertise working together rather than in sequence. Neither function alone typically has the full picture, actuarial teams may have access to provider-level claims data without the capital modeling mandate to reprice against it, while finance teams may own the capital model without direct visibility into provider-level claims patterns.

How Should Reinsurance Treaty Wording Address This Risk?

By building provider-level pricing review directly into the treaty's rate adjustment and audit rights language, rather than leaving it as an informal expectation. Treaty wording can specify that either party may request percentile-level provider pricing data at renewal, and that material shifts in provider mix toward higher-cost facilities trigger a defined rate review mechanism rather than waiting for the next full renewal cycle to address a drift that has already been accumulating. That kind of clause converts an informal capital modeling concern into an enforceable contractual right, which matters considerably when a cedant is reluctant to share granular claims data voluntarily.

What Audit Rights Are Worth Negotiating For?

The right to request provider-level percentile pricing and claims mix data on demand, not just at scheduled renewal points, along with a defined response timeframe so the request cannot be delayed indefinitely. Reinsurers that negotiate this language upfront avoid the far harder position of trying to obtain the same data reactively, after a capital allocation review has already flagged a problem that the cedant has little incentive to help investigate quickly.

How Does This Affect Reinsurer Selection by Cedants?

It cuts both ways, and sophisticated cedants are starting to notice. A cedant evaluating reinsurance partners can reasonably ask whether a reinsurer's pricing model accounts for provider-level percentile variation or relies solely on blended network averages, since the latter suggests either underpriced capacity that will need repricing later, or a margin buffer padded in anticipation of that same uncertainty. Reinsurers that can demonstrate a percentile-based capital allocation process have a genuine differentiator in that conversation, offering pricing that is both more competitive on well-managed networks and more accurately reserved on higher-risk ones.

What Should a Cedant Ask a Prospective Reinsurer?

Whether the reinsurer's underwriting process requests provider-level claims and pricing data, or accepts a single network discount figure as sufficient input. A reinsurer unable or unwilling to answer that question specifically is signaling that its capital allocation is built on the same blended average this whole analysis is warning against, which is useful information for a cedant choosing between competing quotes.

How Should This Be Reflected in Reserve Adequacy Testing?

As a specific sensitivity scenario, not folded silently into a general adverse deviation margin. Reserve adequacy testing should include a scenario where claims mix shifts toward higher-cost providers within an unchanged network, isolating that specific driver from general medical trend so actuaries can see how much of any reserve shortfall is attributable to provider concentration versus broader cost inflation.

Capital allocated against a network average is capital allocated against a number that was never built to represent the risk it is being asked to cover. Closing that gap does not require abandoning network averages entirely, it requires treating them as a starting point for percentile analysis rather than a finished capital input.

Sources

• RAND Hospital Price Transparency Study, Round 5.0 (via The Alliance)
• Serif Health: Why Discounts Aren't Enough