Date: 2026-08-11. Extends Phase 0 anchors 2–3; adjudicates H2. Programs: Combined Title II / Title XVI / concurrent disability determinations as published in SSA ODSSI workload tables (stage rates are decision-level, not program-split).
Search log (PRISMA-lite)
| Field | Record |
|---|---|
| Date | 2026-08-11 |
| Strings | ALJ allowance variance case mix GAO-18-37; SSA hearing average processing time 270; Appeals Council remand rate; federal court disability remand; GAO Social Security disability backlog hearings; OIG major management challenges hearings pending |
| Sources searched | SSA ODSSI FY2024 Workload Data PDF; SSA OIG Major Management Challenges FY2024; SSA APM 2024 hearings line; SSA appeals public data (A04 AC remands; A06 court remands); GAO-18-37; SSA OIG ALJ outlier audits (A-12-17-50220) |
| Inclusion | Stage-resolved SSA workload / performance series; GAO statistical models with case-mix controls; AC and federal-court remand tables from SSA |
| Exclusion | Law-firm “approval rate” marketing pages; raw ALJ league tables without case-mix; conflation of hearing allowance with initial allowance |
1. Stage-resolved waits and allowances (Phase 0 base, deepened)
FY2024 disability decisions (ODSSI, prepared 2025-01-30):
| Stage | Decisions | Allow | Deny / other |
|---|---|---|---|
| Initial | 2,086,885 | 38% | Deny 62% |
| Reconsideration | 495,700 | 16% | Deny 84% |
| ALJ hearing | 289,492 | 51% | Dismiss 33% / Deny 16% |
| Appeals Council (disability subset in ODSSI) | 45,641 | Allow 1% | Remand 16% / Deny 80% / Dismiss 3% |
| Federal court | 15,753 | Allow 1% | Remand 63% / Deny 32% / Dismiss 4% |
SAOR national initial allowance 38.3% corroborates the ODSSI initial rate.
Processing times and pending (OIG MMC FY2024 + APM):
| Metric | FY2023 | FY2024 | SSA goal |
|---|---|---|---|
| Hearing APT (annual avg) | 450 days | 342 days | 270 days |
| Hearing APT (Sep 2024 monthly) | — | 307 days | 270 |
| Pending hearings | ~322k | ~262k (−19% YoY) | — |
| Initial APT | 218 days | 231 days | — |
| Pending initials | — | ~1.18M (OIG MMC) | — |
Hearing APT has been above the 270-day administrative goal for a multi-year span (at least FY2023–FY2024 in the cited OIG/APM series; Phase 0 already locked the FY2024 verification year). Improvement is real; clearance of the goal is not.
Hearing allowance (51%) is not an independent draw from the initial pool — it is selection through denial and appeal. That fact is load-bearing for H7’s steelman and for any “ALJs are soft” claim that skips case mix (§2.2).
Confidence: strong — primary SSA workload / OIG performance series. KC2 remains unfired.
2. ALJ variance after case mix (H2 flagship leg)
GAO-18-37 (Dec 2017; FY2007–2015 adult disability hearing decisions):
- After holding constant claimant, judge, hearing-office, and other modeled factors, GAO estimated the allowance rate for a typical claim could still vary by as much as 46 percentage points across judges (5th–95th percentile range).
- Hearing-office residual range was smaller (~19 pp) than judge-level residual range.
- Residual judge variance fell ~5 pp over 2007–2015; SSA attributed the narrowing to quality assurance and training.
- Representation nearly tripled allowance probability relative to unrepresented claimants in the same model — a case-processing factor, not a medical listing change.
- Backlog share in the hearing office was not associated with allowance rates in GAO’s specification.
SSA OIG follow-on work on high/low-allowance ALJs (A-12-17-50220) documents that outliers often track office demographics and decisional independence, and that focused reviews move some outliers — consistent with residual discretion, not with a claim that case mix fully explains dispersion.
No equal-quality public study located in this pass overturns GAO’s post-controls dispersion finding. Microdata opacity for a fresh 2020s re-estimate is real; it does not erase the published case-mix result H2 requires (≥1 major study).
Confidence: strong on the existence of large post-controls dispersion in the GAO window; moderate on the exact current-year residual magnitude (study vintage FY2007–2015).
3. Appeals Council and federal court remands
Appeals Council remands as % of all AC request-for-review dispositions (SSA A04 public data):
| FY | AC remands | All AC RR dispositions | Remand % |
|---|---|---|---|
| 2022 (52-wk) | 13,888 | 86,395 | 16.08% |
| 2023 | 12,111 | 79,764 | 15.18% |
| 2024 | 13,330 | 73,834 | 18.05% |
| 2025 | 13,062 | 84,045 | 15.54% |
ODSSI’s disability-only AC slice for FY2024 shows remand 16% of 45,641 — same order; different universe (disability vs all AC RR).
Federal court (ODSSI FY2024): of 15,753 decisions, 63% remanded. SSA A06 court-remand receipts: FY2023 10,467; FY2024 10,030; FY2025 9,457 (not the same construct as ODSSI’s decision-year table — cases can pend across years).
Remand volume at AC and court is a quality/capacity pressure on the hearing layer: work returns to ALJs rather than resolving as final denials. It is not by itself a fraud rate.
Confidence: strong on published remand percentages; moderate on interpreting remands as pure “error” vs. incomplete records / new evidence.
4. Backlog as a capacity object (GAO / OIG)
GAO’s hearings-critical-case work (e.g. GAO-22-104191) and SSA OIG’s Major Management Challenges treat hearings pending and APT as standing management risks even as pending counts fall. The binding constraint for O2 is not whether the backlog is “at its worst ever” — it is whether waits remain outside the agency’s own standard while residual ALJ inconsistency survives case-mix controls. Both remain true in the verification window.
Initial-level pending (~1.18M) and rising initial APT (218→231) show capacity pressure is not only an ALJ phenomenon. DDS is the federalist hinge (state agencies, federal rules/funding).
5. H2 adjudication
H2 — Adjudication capacity binds.
| Criterion | Result |
|---|---|
| Hearing waits exceed a stated administrative standard in SSA’s own reports for a multi-year span | Met. APT 450 (FY2023) and 342 (FY2024) vs goal 270. |
| ALJ allowance dispersion remains large after published case-mix controls in ≥1 major study | Met. GAO-18-37: up to 46 pp residual judge range. |
Verdict: SUPPORTED. Backlog/timeliness and residual decisional variance jointly constrain access and integrity more tightly than a claim that the statutory disability definition alone is the binding object. Definition-tightening architectures that ignore DDS/ALJ capacity inherit O2/O5 penalties.
Confidence: strong.
Implications
- Architecture #1 (adjudication capacity surge) and #2 (ALJ consistency controls) stay first-class for O2/O5.
- “ALJs are soft” talking points that cite the 51% hearing allow rate without case-mix fail this workstream.
- AC/court remand rates are supporting evidence of quality friction, not a substitute for the GAO variance test.