Arizona Medicaid ex parte renewal share, August 2026

Ex parte share measures how much of the renewal burden the state carries instead of the beneficiary. It trends with data-matching infrastructure, which improves in vendor-release steps rather than smoothly.

Latest 91.5 (02/26); mean monthly change over trailing window 0.61pp; damped six-month projection = 93.3.

Half-weight trend continuation with volatility-scaled uncertainty; renewal cohorts differ month to month, so the interval is wider than the point trend alone would suggest.

The resolver is a state row, not a national weighted average: Arizona's original first-publication August 2026 reporting-period row in the CMS eligibility processing dataset. The target is the share of completed renewals processed ex parte, reported in percent and rounded to one decimal.

Base-rate/reference-class anchor: the most relevant outside view is Arizona's own post-unwinding first-print run from July 2025 through February 2026. It shifted from the mid-80s into the low-90s, with the latest three observed points centered near 92.9 percent, so a persistence-plus-small-trend prior dominates a broad national average.

Level, momentum, and mechanism: the level is high and plausibly reflects mature data matching rather than a temporary spike, but the February dip from 93.9 to 91.5 argues against extrapolating rapidly toward 100. The upper bound is mechanically capped, while downside comes from manual-heavy cohorts or data-source interruptions.

Prior/update/interval: prior model is Arizona latest-value persistence with a damped local trend, using the available inspected first-print sample of July 2025, September 2025, November 2025, January 2026, and February 2026. Prior catalog baseline was 93.3, but the arithmetic starts from the latest observed 91.5; I add +1.0 pp for the November-January high plateau, +0.8 pp for the broader July-February improvement, and 0.0 pp for capped upside. The 80% interval uses realized inspected first-print moves up to 6.8 pp, widened for six-month cohort and missing-month uncertainty, then judgmentally capped at 99.0 because 100 percent is the mechanical maximum.

Counter-consideration: upside outside the interval is hard because it would require Arizona to print essentially universal ex parte renewals near 100 percent. Downside outside the interval would require a renewal cohort with unusually poor electronic match rates, a state eligibility-system problem, or a CMS reporting break that pulls the share below the mid-80s.

Point calculation: latest observed 91.5 percent plus +1.0 pp for the November-January high plateau and +0.8 pp for the broader July-February improvement gives 93.3 percent. Interval calculation: lower bound 93.3 - 7.4 = 85.9; upper bound is 93.3 + 5.7 = 99.0 after applying the upper-tail cap and rounding to one decimal.

Resolution-date note: the official CMS page verified the monthly release vehicle and current June 26, 2026 update, but did not expose a future dated August 2026 state-row placeholder. I keep the forecast tied to the canonical ledger date 2026-12-15 and bind resolution to the first official CMS dataset print.

Review disposition: accepted the resolver critique by naming the original O row and the numerator/denominator concept in the rule; accepted the timestamp critique by replacing runAt with the actual UTC command output, which falls on June 28 UTC while the local review date is June 27; accepted the interval critique by labeling the historical sample as the available inspected first-print points and noting missing-month uncertainty.