PwC's 9% Group Medical Cost Trend for 2027 and the Rate Filing Benchmark Problem

PwC's Health Research Institute published "Medical Cost Trend: Behind the Numbers 2027" in June 2026, drawing on surveys and interviews with pricing actuaries at 27 U.S. health plans covering more than 103 million employer-sponsored members and 8 million individual ACA marketplace enrollees. The headline finding: a 9% composite group medical cost trend for plan year 2027, the highest since 2010 and the fifth consecutive year in which group trend has held above historical norms. The more consequential data point for pricing actuaries arrives in what PwC did alongside that projection. The firm simultaneously revised the 2026 group trend upward from 8.5% to 9.0% and the individual market trend from 7.5% to 8.5%, a mid-year acknowledgment that experience deterioration on 2025 and 2026 accident months is worse than the prior estimate assumed. When a respected external benchmark revises itself upward between publication cycles, it tells you something about how the development is running. The question for a plan filing 2027 rates this month is how much of that 9% belongs in their own selection.

Connecticut's 30-day public comment period for preliminary 2027 ACA rate filings opened June 5, 2026, with four carriers averaging a 15.7% individual market request. Nevada's SERFF deadline fell June 17. The rate filing windows are closing now, which means trend selection decisions for plan year 2027 are being made against this PwC publication in real time. What follows is a working methodology for converting the PwC benchmark into a defensible, component-level plan-specific trend selection documented to the standard required by ASOP No. 25 and ASOP No. 8.

PwC 2027 Group Medical Cost Trend Projection, Highest Since 2010

5th Year

Consecutive Year Group Trend Has Held Above Historical Norms

85%+

Share of Plan Actuaries Expecting 2027 Pharmacy Trend to Outpace Overall Medical Trend

What the 9% Actually Measures

The PwC composite is a weighted average across 27 plans, 103 million group members, and five distinct inflator categories. Weighted averages obscure what matters to a plan-specific filing. The five inflators PwC named carry very different magnitudes depending on network structure and benefit design: AI-enabled provider documentation and coding optimization, GLP-1 and oncology drug spending, behavioral health utilization growth, provider reimbursement pressure from No Surprises Act disputes, and general unit cost inflation from fee schedule renegotiations. Plans with closed physician networks and robust prior authorization programs see the AI coding effect very differently than open-access PPOs with weaker code-edit programs. A plan that carved GLP-1s to a separate pharmacy benefit manager with a dedicated rebate contract sees pharmacy trend differently than one running a blended formulary. The 9% is a starting point, not a selection.

The pharmacy signal inside the 9% is worth isolating. More than 85% of surveyed plan actuaries expect 2027 pharmacy trend to outpace the overall medical composite, driven primarily by GLP-1 prescriptions. Between December 2024 and December 2025, GLP-1 prescription volumes nearly doubled across the commercially insured population, and the oral formulations entering market in 2026 at price points below $200 monthly will accelerate adoption further. The implication for trend regression is that a plan running a log-linear fit on 36 months of total pharmacy PMPM data without isolating GLP-1 claims is fitting a curve that changes slope partway through the estimation window. The statistical trend from that regression will understate the forward pharmacy run rate unless the acceleration is accounted for explicitly.

Step One: Building the Plan-Specific Experience Trend

The credibility exercise starts with a clean experience trend estimate. Using 24 to 36 months of allowed charge data adjusted for benefit plan changes and enrollment mix shifts, the actuary runs a log-linear regression of per-member per-month (PMPM) costs against time. The regression produces an annualized exponential trend rate, typically expressed as e to the power of the slope coefficient minus one. Before fitting, three data quality checks matter: continuity of the claim runout basis (switching from paid-through to incurred-through mid-series distorts the slope), removal of large claimants above a given threshold if the plan is crediting stop-loss, and confirmation that the enrollment composition has not shifted systematically toward a higher-morbidity population during the window (a plan that expanded Medicaid-adjacent enrollment in 2024 will show trend acceleration that reflects mix rather than underlying cost inflation).

The regression period selection involves judgment. Weighting recent quarters more heavily via weighted least squares reflects the PwC revision signal: if 2026 development is adverse and you are running an equal-weight regression over 36 months, the last four to six quarters are being smoothed against three years of lower-trend data. For a 2027 filing, a 24-month window with exponential weighting toward the most recent quarters is defensible and will capture the mid-year deterioration PwC's revision implied. The resulting plan-specific trend estimate should be documented with its R-squared, the residual pattern, and the rationale for the chosen period and weighting scheme.

Suppose the plan's fitted annual trend comes in at 6.5%, compared to PwC's 9%. The actuary now faces the credibility question directly: how much of the difference is plan-specific good fortune versus sampling variance in a small or mid-size book?

Step Two: Buhlmann-Straub Credibility Weighting

The standard framework for blending plan-specific and industry experience in health insurance pricing is Buhlmann-Straub credibility, which assigns weight to the plan's own data based on its volume relative to the within-plan versus between-plan variance ratio. The formula is Z = n / (n + k), where n is the plan's exposure in member-months and k is the variance parameter estimated from a multi-plan dataset. The credibility-weighted selection is then:

Selected Trend = Z × (Plan Trend) + (1 - Z) × (Industry Benchmark)

For a 20,000-member group with two years of data, n equals roughly 480,000 member-months. If the between-plan variance in the reference dataset (the PwC 27-plan survey population, or a peer-group database the actuary maintains) implies a k of approximately 860,000, then Z equals 480,000 / (480,000 + 860,000), or roughly 0.36. The complement, 0.64, is external benchmark weight. Applying the blend: 0.36 times 6.5% plus 0.64 times 9.0% equals 8.1%, a selection materially above the plan's own experience but grounded in the reality that a 20,000-member book is not large enough to distinguish genuine low-trend performance from random favorable experience.

The k parameter is the critical actuarial judgment in this framework. It is estimated from the variance structure of a multi-plan dataset, not from the plan's own data alone. Actuaries who use the PwC 27-plan population as the reference should note that PwC does not publish the between-plan variance directly; it must be estimated from the dispersion of plan-level trend reports across the survey, or from a proprietary peer database. The Milliman Health Cost Guidelines and the SOA Getzen model both provide inputs that can inform the k calibration. Document the k value used, the source dataset it comes from, and the sensitivity of the final selection to plausible alternative k values. This is what ASOP No. 25 Section 3.6 requires when external data enters the credibility blend.

Step Three: Component-Level Disaggregation

A blended 8.1% applied uniformly across inpatient, outpatient, professional, and pharmacy services will be wrong in every category. The 2027 cost structure is unusually dispersed by service line: pharmacy trend for plans with significant GLP-1 utilization is likely running 12% to 15% gross of rebates, while inpatient trend is more plausibly 5% to 7%, reflecting stable utilization and moderate fee schedule pressure. Loading a uniform 8.1% overstates inpatient and understates pharmacy. For a plan where pharmacy already represents 30% of total allowed costs, that misallocation compounds directly into benefit leverage errors and pricing inadequacy on the pharmacy component.

The disaggregation process runs the same credibility framework at the service-category level, using component benchmarks from PwC's supplemental tables, the 2026 Milliman Medical Index, or the actuary's own peer database. The composite check is that the weighted sum of component trend selections, using the plan's current service-category PMPM distribution as weights, should reconcile to the 8.1% composite. If it does not, the component selections are inconsistent with the plan's own cost mix.

For pharmacy, the sub-component structure matters further. GLP-1 trend should be isolated and projected separately using the plan's NDC-level claims data. As described in our GLP-1 trend factor analysis, the logistic adoption curve approach better captures the non-linear acceleration in GLP-1 spending than a linear trend regression. The remaining specialty pharmacy trend (oncology biologics, rare disease agents) and traditional drug trend (generics, branded non-specialty) carry different slope characteristics and different rebate offsets. A composite pharmacy trend selection that does not distinguish these three sub-components will be internally incoherent in any year where GLP-1 adoption is still in the steep part of the S-curve.

The AI Coding Normalization Problem

Among the five inflators PwC identified, the AI-enabled provider documentation and coding effect demands the most careful treatment in trend regression. Roughly 70% of surveyed plans ranked AI coding tools in their top three inflators, and approximately 20% identified it as the single largest driver. The mechanism: revenue cycle management vendors have deployed machine learning tools that capture greater diagnostic specificity and optimize evaluation-and-management (E&M) code level selection, increasing average allowed amounts per claim without a proportionate increase in care intensity. For plans where provider networks adopted these tools during the trend estimation window, the historical regression captures a one-time code-level migration as a permanent trend rate.

The test for this effect uses E&M level distribution data. Pull the plan's allowed claims for professional services and examine the distribution across CPT codes 99213, 99214, and 99215 (the three most common office visit levels) before and after the adoption period. If the share of 99215 visits rose from 18% to 28% over 18 months while encounter frequency held flat and the plan's contracted fee schedule increased only 3%, the 10-percentage-point code-level migration represents a coding precision effect, not an acuity shift. A coding normalization factor, applied to the professional service PMPM before fitting the trend regression, removes this one-time step change from the slope estimate and prevents the regression from treating a structural shift as a forward-looking trend.

The normalization factor is calculated as the ratio of the pre-adoption E&M cost distribution to the post-adoption distribution, both applied to the plan's member count. In the example above, if the shift from 99213 to 99215 increases average allowed per encounter by 8%, the normalization factor applied to post-adoption professional claims is approximately 0.926 (1 divided by 1.08). The actuary should document the adoption period, the pre- and post-period E&M distributions, and the magnitude of the normalization adjustment. This is a judgmental modification to the historical data, and ASOP No. 8 Section 3.4 requires explicit disclosure of such adjustments in the rate filing actuarial memorandum.

The Individual Market Overlay: Morbidity Loading Above the Stated Trend

The PwC 8.5% individual market trend explicitly excludes adverse selection from the expiration of enhanced premium tax credits. The Congressional Budget Office estimated that ePTC expiration adds 4.3% to gross benchmark premiums in 2026 and 7.7% in 2027, and the Georgetown Center on Health Insurance Reforms' early state filing analysis suggests a second consecutive year of double-digit marketplace premium increases is materializing as carriers price in the enrollment composition shift. Individual market pricing actuaries should not treat the PwC 8.5% as a complete trend input; they need to add an explicit morbidity loading above that figure to capture the expected change in the risk pool as subsidy-sensitive enrollees exit.

The morbidity loading calculation requires an assumption about the elasticity of enrollment with respect to net premium. Plans with historical data from the 2018 to 2019 ePTC reduction can estimate a plan-specific elasticity, though the 2026 environment differs in several respects, including a larger subsidy-sensitive population and a tighter carrier network market in many states. Wakely's morbidity shift analysis, discussed in more detail in our Wakely morbidity methodology article, provides an external framework for estimating morbidity adjustment factors in the range of 2.9% to 6.5% above the baseline trend, depending on enrollment composition assumptions. The additive structure is critical: the PwC trend addresses cost per member, and the morbidity loading addresses the change in who the member is.

No Surprises Act IDR Volume: A Separate OON Trend Component

The 2026 clarification of the NSA independent dispute resolution payment methodology generated a sharp increase in IDR dispute volumes, and PwC's survey identified this as a named inflator contributing to the 9% composite. The mechanism is direct: as IDR arbitrators apply the qualifying payment amount framework, the effective out-of-network cost allowances embedded in plan designs are being ratcheted upward for high-dollar services where out-of-network billing has been most prevalent, particularly emergency medicine and anesthesiology. Plans that do not track out-of-network cost trend separately from in-network medical trend are at risk of understating the OON component in their prospective allowance.

The recommended treatment is to run OON trend as a discrete component using plan-specific cost-sharing data, not to embed OON pressure into the medical trend composite. Pull OON claims by facility type and professional specialty, and fit a separate trend for emergency facility, non-emergency facility, and professional OON services. For plans in markets where IDR adoption has been highest, the OON trend may be running 15% to 25%, well above the in-network composite. Blending it into a single medical trend line obscures the exposure concentration and makes it impossible to separate benefit design effects from utilization and unit cost effects in subsequent rate audits.

ASOP Documentation Requirements for the Final Selection

ASOP No. 25 (Credibility Procedures) and ASOP No. 8 (Regulatory Filings for Health Benefits) together define what the actuarial memorandum must contain when an external benchmark enters the trend selection. The required elements are specific: the source of the external benchmark and the population it represents (PwC, 27 plans, 103 million group members, June 2026 publication); the credibility methodology and its inputs (Buhlmann-Straub, the k parameter value and its derivation, the n value for the plan's own data); the component-level trend selections and their relationship to the composite; any judgmental adjustments to the historical data, including the AI coding normalization factor and its basis; and the basis for the behavioral health trend selection given MHPAEA parity enforcement constraints.

The behavioral health disclosure warrants a paragraph of its own in the memorandum. Behavioral health utilization has grown 62% since 2018 as MHPAEA parity enforcement expands reimbursable services, per PwC's survey. For a 2027 filing, the actuary must assess whether the plan's behavioral health trend is still in the acceleration phase or approaching a saturation plateau, because the two scenarios call for materially different forward selections. A plan that adopted telehealth parity in 2022 and saw utilization spike 40% in the following 12 months may now be running a lower incremental behavioral health trend as the post-adoption level stabilizes. A plan that has not yet implemented parity for the full range of MHPAEA-covered services may still be in the acceleration phase. The memorandum must document the factual basis for the actuary's placement on this saturation curve, not just the resulting trend assumption.

State regulators processing 2027 filings this summer will see a range of trend submissions across carriers in the same market, some anchoring to PwC's 9%, some to lower plan-specific estimates. Regulators are increasingly requesting service-category documentation rather than accepting a composite figure. A filing that arrives with a disaggregated trend build, a documented credibility blend with sourced k and n inputs, explicit treatment of the AI coding effect, and a MHPAEA utilization rationale will withstand review. One that simply states "trend assumption: 8.5%, consistent with industry benchmarks" will not.

Why This Matters for 2027 Filings

The mid-year upward revision to 2026 trend is the most actionable signal in the PwC release. When a survey covering 103 million members revises the current year's trend upward by 50 basis points at mid-year, the development pattern on 2025 and 2026 accident months is running adverse to initial expectations. Actuaries anchoring their experience period on a full 24 or 36 months of prior data without recency weighting are fitting a slope to a period that opens lower and closes higher, and the regression will understate the forward run rate. The fix is not complicated: weight recent quarters more heavily in the regression, or shorten the estimation window to 18 to 24 months while documenting the rationale for the truncation. The PwC revision is the external evidence that justifies the judgment.

The 9% composite is the highest group medical cost trend projection in 17 years, and it is not driven by a single cyclical factor that will normalize in one year. GLP-1 adoption continues its logistic curve. Provider coding optimization tools are not being unwound. MHPAEA parity enforcement is expanding, not contracting. No Surprises Act IDR volume will stay elevated as the payment methodology becomes better understood. Actuaries who build 2027 selections on a mean-reversion assumption toward a historical 5% to 6% norm are pricing against the cost drivers that are actually running in the underlying data. The Buhlmann-Straub credibility blend, with appropriate weight on the PwC benchmark and proper component-level disaggregation, produces a selection that can be defended to regulators, presented to clients, and revisited with the next year's data without embarrassment about what was embedded in the prior filing.