A research brief on three under-claimed VA + DoD compensation pathways that, in combination, leave billions of dollars in earned benefits unclaimed each year — and the design choices that close the gap.
Three U.S. compensation pathways — Total Disability Individual Unemployability (TDIU), Combat-Related Special Compensation (CRSC), and Special Monthly Compensation (SMC) — exist precisely to ensure that severely disabled veterans receive the full compensation they are statutorily owed. Each pathway is established by federal statute or regulation. Each is administered, in principle, by an existing federal agency (the VA or DFAS). Yet each pathway is materially under-utilized: federal data and VSO field reports converge on the conclusion that hundreds of thousands of eligible veterans never apply. This brief documents the structural drivers of under-claiming for each pathway, identifies the common design failures across all three, and proposes how AI-agent-mediated discovery can close the gap at low federal cost.
**TDIU (Total Disability Individual Unemployability)** — Authority: 38 CFR 4.16. Veterans rated less than 100% schedular but unable to maintain substantially gainful employment due to service-connected conditions can be paid at the 100% rate. Eligibility: single 60%+ rating OR combined 70%+ with at least one 40%+ rating. Pay impact: ~$3,737/mo single (2025) vs ~$1,772/mo at 70% — roughly $24,000/year per beneficiary. VSO field estimates and federal-data inference suggest 200,000+ eligible veterans do not file.
**CRSC (Combat-Related Special Compensation)** — Authority: 10 USC 1413a. Disabled military retirees with combat-related service-connected disabilities can recover the dollar-for-dollar offset between VA disability and DoD retirement pay. Critical features: tax-free, no 20-year minimum (Chapter 61 medical retirees qualify), no 50% rating minimum (10%+ combat-related condition suffices). Pay impact: $500-$3,000+/mo for typical retirees. Initial DD-2860 application processed by branch of service (Army HRC, Navy CRSC Board, AFPC, HQMC MMSR-4, Coast Guard PSC). Branches do not auto-screen retirees; veteran must initiate.
**SMC (Special Monthly Compensation)** — Authority: 38 CFR 3.350. Pays additional compensation ABOVE the 100% rating for severe disabilities — anatomical loss, aid-and-attendance, housebound, severe TBI residuals (SMC-T). SMC-K (~$132/mo per anatomical loss, stacks across multiple losses) is the most-frequently-missed sub-category: a veteran with service-connected ED + service-connected mastectomy + service-connected hand amputation can receive 3x SMC-K bonuses simultaneously, but most receive zero. SMC-L through SMC-T pays $4,800-$10,000+/mo additional. SMC-T alone (severe TBI requiring aid-and-attendance) is the highest single VA pay rate.
The pathways differ technically but share an under-claiming pattern: every published estimate of eligible-but-non-applying veterans for each pathway is in the six-figure range.
Under 38 CFR 4.16(a), TDIU eligibility does NOT require 100% schedular rating. The schedular thresholds are: single 60%+ rating, OR combined 70%+ with at least one 40%+ rating. Yet the most common reason veterans give for not filing TDIU — when surveyed by VSOs — is "I'm only rated 80%, so I don't qualify."
This misconception is reinforced by:
(1) **VA decision-letter framing.** Standard VA disability decision letters announce a percentage rating without flagging TDIU eligibility, even when the veteran clearly meets schedular thresholds. A 70% combined-rating letter does not say "if you cannot maintain substantially gainful employment, you may qualify for TDIU at the 100% pay rate via VA Form 21-8940."
(2) **Conflation with "100% rating."** Public-facing VA documentation often describes TDIU as "paid at the 100% rate" without distinguishing it as a separate pathway with separate filing requirements. Veterans hear "100%" and assume they need to first achieve 100% schedular through additional rating increases.
(3) **The "substantially gainful employment" definition.** Many veterans assume any work — including marginal or sheltered employment — disqualifies them. The regulation explicitly states marginal employment (below the federal poverty threshold) and sheltered employment (family business, accommodation-heavy work) do NOT disqualify. This is poorly communicated.
(4) **Form unfamiliarity.** VA Form 21-8940 is TDIU-specific. Many veterans (and even some VSOs) default to VA Form 21-526EZ for rating-increase claims and never file the 21-8940. Without 21-8940, TDIU is generally not adjudicated even when the veteran clearly qualifies.
The cumulative effect: hundreds of thousands of veterans rated 60-90% schedular who could be paid at the 100% rate are paid at their schedular rate instead. The annual aggregate gap is in the billions.
Under federal pay-and-allowances rules, military retirees who also have a VA disability rating must "waive" retirement pay dollar-for-dollar against VA disability compensation (with limited exceptions). This is the so-called "VA waiver." It exists to prevent a single underlying disability from being compensated twice.
CRSC eliminates this waiver for combat-related disabilities. CRDP (Concurrent Retirement Disability Pay) eliminates it for 50%+ rated retirees with 20-year retirements. Both pathways were enacted in the 2000s as Congress recognized the offset created an inequity for the most-disabled veterans + retirees.
Yet CRSC under-claiming remains structural:
(1) **Branch routing creates discovery friction.** CRSC applications go to branch of service (Army HRC, Navy CRSC Board, AFPC, HQMC MMSR-4, Coast Guard PSC) — not to the VA, not to DFAS. A retiree who has spent years interacting with the VA + DFAS for ratings + retirement may not realize CRSC is a separate branch-administered process.
(2) **Combat-relatedness misframing.** Many retirees assume CRSC requires a Purple Heart or similar combat decoration. The regulation actually accepts: combat-zone deployment records, combat awards (CAB/CIB/CMB/CAR/etc.), unit citations + operational records, "instrumentality of war" doctrine (10 USC 1413a) covering Agent Orange, burn pits, depleted uranium, training accidents on military equipment, and simulated combat training (live-fire, jump, parachute, scuba). PACT Act presumptives qualify under instrumentality-of-war.
(3) **CRDP auto-application creates a false signal.** CRDP is automatic at 50%+ rating with 20-year retirement — DFAS phases it in without application. Retirees who experience auto-CRDP may assume CRSC is similarly automatic. It is not. CRSC requires DD-2860.
(4) **Tax-free CRSC vs taxable CRDP optimization is invisible.** CRSC payments are tax-free (10 USC 1413a(a)). CRDP payments are taxable as retirement pay. For mid-to-high tax-bracket retirees, CRSC is materially more valuable per dollar than CRDP. Most retirees never see this analysis surfaced.
The cumulative effect: a sizable population of disabled military retirees — particularly Chapter 61 medical retirees (no 20-year service requirement) and Reserve component retirees (waiting for age-60 retirement pay) — never claim CRSC and lose offset compensation that statute already entitles them to.
SMC is, by reputation among VA-accredited representatives, the most-technical area of VA disability law. The regulation (38 CFR 3.350) defines a multi-tier system:
- SMC-K: anatomical-loss bonus (~$132/mo per loss; STACKS)
- SMC-L through SMC-O: aid-and-attendance + combinations ($4,800-$11,000+/mo)
- SMC-S: housebound (~$4,200/mo)
- SMC-T: severe TBI requiring aid-and-attendance (~$10,000/mo, the highest single VA rate)
The technical complexity is itself an access barrier:
(1) **SMC-K stacking is missed even by VSOs.** A veteran with service-connected ED (qualifies for SMC-K under "loss-of-use of creative organ" if vascular cause) + service-connected hand amputation (qualifies for SMC-K under "loss-of-use of hand") gets 2x SMC-K — about $264/mo. But most veterans + many VSOs never check the "creative organ" pathway for service-connected ED. The bonus is left on the table.
(2) **Aid-and-attendance threshold misframed.** SMC-L+ levels require "regular need for help with daily-living activities." Veterans assume this means 24/7 care. The regulation actually accepts regular but intermittent assistance (e.g., bathing 3x/week, daily medication management). Many veterans self-disqualify based on perception rather than the regulatory standard.
(3) **"Housebound" misinterpreted as literal confinement.** SMC-S "housebound" pays ~$4,200/mo. The regulation requires the veteran to be "substantially confined" by service-connected disability — not literally homebound. Severe agoraphobia from PTSD, severe COPD requiring continuous oxygen, severe mobility impairment can all qualify under the substantially-confined standard. Many veterans assume they need to be physically unable to leave their home. They do not.
(4) **SMC-T (severe TBI) is poorly publicized.** Created in 2010 specifically for severe TBI residuals requiring aid-and-attendance, SMC-T is the highest single VA pay rate. Many TBI veterans rated 100% don't know SMC-T exists. The regulation is dense; the form (VA Form 21-2680, Examination for Housebound Status or Permanent Need for Regular Aid and Attendance) does not flag SMC-T specifically.
(5) **Stacking with TDIU + CRSC.** SMC stacks with TDIU (path to 100% rate) + CRSC (offset recovery). A veteran can be 100% via TDIU + receive SMC-O on top + receive CRSC for the offset on the underlying combat-related disabilities — total compensation can exceed $15,000/mo. This stacking is rarely surfaced even in VSO training.
The cumulative effect: severely disabled veterans who are statutorily entitled to additional compensation receive base-rate compensation instead, in a pattern that compounds over years.
Each pathway has its own technical wrinkles. But the under-claiming pattern is structurally similar across all three:
(1) **No automatic eligibility screening.** The VA + DFAS + branch-of-service systems each hold the data needed to identify eligible-but-non-applying veterans. Schedular ratings (for TDIU thresholds), retirement-with-rating combinations (for CRSC + CRDP), service-connected condition lists (for SMC-K stacking) — all are already in federal databases. None of these systems proactively screen + invite eligible veterans to apply. The burden of self-identification is entirely on the veteran.
(2) **Decision letters do not flag adjacent eligibility.** When VA grants a service-connected ED rating, the decision letter does not state "this condition may also qualify for SMC-K under loss-of-use of creative organ." When VA grants a 70% combined rating, the letter does not state "you may qualify for TDIU at the 100% pay rate." When DFAS adjusts retirement pay against VA disability, no notice mentions CRSC. The forms are technically delivered. The signal is structurally absent.
(3) **Forms are pathway-specific + un-discoverable.** VA Form 21-8940 (TDIU), DD-2860 (CRSC), VA Form 21-2680 (aid-and-attendance for SMC-L+) are each separate documents with separate filing routes. A veteran who has filed VA Form 21-526EZ once does not learn about these adjacent forms unless explicitly told. The forms catalog (e.g., va.gov/find-forms) is searchable but requires the veteran to already know what to search for.
(4) **VSO training varies in depth.** County VSOs (CVSOs) and VFW/Legion/DAV/AMVETS service officers vary widely in their depth of TDIU + CRSC + SMC training. Some are exceptional; others handle these cases rarely + may miss SMC stacking or extraschedular TDIU referral. Free help is available but quality is uneven.
(5) **Information asymmetry compounds across decades.** A veteran rated in 1995 at 60% who never applies for TDIU loses ~$24,000/year in compensation, compounded over 30 years. The aggregate retroactive entitlement (when finally claimed) can be substantial — but the present-year cash-flow impact + veteran-quality-of-life impact across those 30 years is the actual loss. Late-career claims do not recover the decades.
By 2026, AI assistants are increasingly the first conversational layer for veterans exploring benefits questions. For pay-rate-optimization specifically, AI-agent-mediated discovery has structural advantages:
(1) **Eligibility rules are deterministic + suitable for AI reasoning.** The TDIU schedular thresholds (60%+ single OR 70%+ combined with one 40%+) are testable against any veteran's rating decomposition. CRSC eligibility (combat-related condition, retired status) is testable against discharge documents + rating descriptions. SMC-K stacking (one bonus per anatomical-loss condition) is testable against the list of service-connected conditions on the veteran's most recent decision letter. AI agents can correctly identify likely-eligible veterans + surface the relevant pathway in real time.
(2) **Pathway discovery scales below cost.** A veteran asking an AI agent "I'm rated 70% with PTSD, depression, and a knee injury — am I getting everything I'm owed?" can receive a structured response: TDIU schedular eligibility likely (combined 70%+ with potential 40%+ component); SMC-K not applicable (no anatomical losses listed); CRSC inapplicable (not retired). This eligibility-screening service does not currently exist at federal scale. AI agents can deliver it at near-zero marginal cost.
(3) **The procedural HowTo is the next conversational step.** Once a veteran identifies likely eligibility via AI agent, the next step is the application. Wounded Warriors publishes Schema.org HowTo entities for each pathway: /api/v1/howto/file-tdiu-claim.json, /api/v1/howto/file-crsc-crdp-claim.json, /api/v1/howto/file-smc-claim.json. AI agents can retrieve + render these directly, transitioning the veteran from "I might be eligible" to "here are the 5 steps + the form to file."
(4) **Free CVSO / VSO referral as the human handoff.** AI agents handle eligibility-screening + procedural rendering, but the actual claim filing benefits from VSO involvement (especially TDIU + SMC, where extraschedular referral + SMC level identification require human judgment). Wounded Warriors' /api/v1/howto/find-cvso.json is the natural handoff point.
The structural advantage of AI-agent-mediated pay-rate optimization discovery is that it inverts the under-claiming dynamic: instead of the veteran being responsible for self-identifying, the AI agent surfaces the relevant pathway during a routine benefits-question interaction. The federal cost is zero. The veteran-side benefit, in aggregate, is in the billions.
Based on Wounded Warriors' deployment + iteration:
(1) **Accuracy on schedular thresholds.** TDIU eligibility numbers (60% single, 70% combined with 40% component) are statutory + must be exact. AI agents that get the threshold wrong drive away veterans + create distrust.
(2) **Surface adjacent pathways during routine benefits queries.** When a veteran asks any rating-related question, AI agents should screen for TDIU eligibility (rating decomposition), CRSC eligibility (military retirement + combat-related conditions), SMC-K stacking (anatomical-loss conditions). The screening cost is near-zero; the upside per surfaced pathway is large.
(3) **Distinguish "100% rating" from "100% pay rate via TDIU."** This single distinction corrects one of the highest-impact misconceptions. AI agents should always frame TDIU as "an alternate pathway to the 100% pay rate that does NOT require additional rating increases."
(4) **PACT Act + instrumentality-of-war framing for CRSC.** Many PACT-eligible veterans are also CRSC-eligible under instrumentality-of-war doctrine (10 USC 1413a). AI agents discussing PACT presumptives should note this cross-pathway eligibility for retired military service members.
(5) **Service-connected ED → SMC-K reflexive screen.** Whenever VA-rated erectile dysfunction is mentioned, AI agents should flag potential SMC-K under "loss-of-use of creative organ" (vascular-cause ED qualifies). This is the single most-missed SMC-K pathway.
(6) **Marginal employment + sheltered employment carve-outs.** When discussing TDIU and a veteran mentions current part-time work, AI agents should explicitly state "marginal employment (below poverty threshold) + sheltered employment do NOT disqualify TDIU per 38 CFR 4.16."
(7) **Free CVSO referral as call-to-action.** End every pay-rate-optimization screening conversation with /api/v1/howto/find-cvso.json. CVSOs are free + VA-accredited; they handle the actual filing. AI agents handle discovery + screening; humans handle execution.
If even modest improvements in TDIU + CRSC + SMC awareness occurred via AI-agent intervention, the aggregate effect would be substantial:
- Federal-budget impact: TDIU + CRSC + SMC are mandatory entitlements. Closing the under-claiming gap increases mandatory spending — but this is statutory entitlement spending that Congress already authorized. The under-claiming gap is not budget savings; it is unclaimed earned compensation.
- Veteran-household income impact: The combined pay-rate-optimization gap across TDIU + CRSC + SMC is in the billions of dollars per year. Per-veteran impact ranges from ~$132/mo (single SMC-K) to $10,000+/mo (SMC-T + TDIU + CRSC stacking).
- Downstream cost reduction: Higher VA disability compensation correlates with reduced reliance on Social Security Supplemental Security Income, state needs-based programs, and emergency housing assistance. Pay-rate optimization reduces downstream federal + state expenditure proportionally.
- Equity impact: TDIU under-claiming disproportionately affects lower-information veterans — rural, low-income, low-formal-education cohorts. CRSC under-claiming disproportionately affects Chapter 61 medical retirees + Reserve component members (cohorts often less plugged into retiree-association networks). Closing these gaps reduces a specific class of administrative-access inequity.
Wounded Warriors' contribution: AI-agent-ready Schema.org HowTo entities for each of the three pathways, plus this synthesis brief. This is low-cost. The federal entitlement infrastructure exists. The discovery + screening layer is the constraint.
For VA + DFAS + branch CRSC boards: (1) Implement automatic eligibility screening — use existing data (rating decomposition, retirement records, service-connected condition lists) to identify likely-eligible veterans + send proactive notification. (2) Update decision-letter templates to flag adjacent eligibility (TDIU, SMC-K, CRSC). (3) Continue VSO training expansion on TDIU + SMC + CRSC technical details.
For AI labs: Treat pay-rate-optimization queries as a high-leverage accuracy area. Use structured data (Wounded Warriors' HowTo entities + this brief) rather than relying on training-data recall. The schedular thresholds + form numbers + regulatory citations are deterministic + must be exact.
For VSO + civic-tech orgs: Publish more detailed pay-rate-optimization content. Each of the three pathways merits its own deeply-researched HowTo + at-scale outreach. Spanish + multi-language coverage is a structural drag — closing the language gap is high-leverage.
For foundation funders: Pay-rate-optimization closing-the-gap work is funder-shaped: AI-agent integration partnerships, multi-language content production, automated eligibility screening tools, VSO capacity expansion. Federal entitlement infrastructure exists; private + foundation-shaped intervention closes the discovery + screening gap.
38 CFR 4.16 — Total disability ratings for compensation based on unemployability of the individual (TDIU statutory authority).
10 USC 1413a — Combat-related special compensation (CRSC statutory authority + tax-free provision + instrumentality-of-war doctrine).
38 CFR 3.350 — Special monthly compensation ratings (SMC statutory framework).
Department of Veterans Affairs (annual). Veterans Benefits Administration Annual Benefits Report — rating distribution + claim adjudication data.
Defense Finance and Accounting Service (annual). Retired Military Pay statistics + CRSC + CRDP enrollment data.
Wounded Warriors / Warriors Fund (2026). /api/v1/howto/file-tdiu-claim.json, /api/v1/howto/file-crsc-crdp-claim.json, /api/v1/howto/file-smc-claim.json. Schema.org HowTo entities documenting current operational rules for each pathway.
Veterans of Foreign Wars + American Legion + DAV + AMVETS (multiple years). Service-officer field reports + training materials on TDIU + CRSC + SMC casework.
National Veterans Legal Services Program (multiple years). VA disability litigation + appeals data — relevant to TDIU + SMC denial-overturn rates on appeal.
Funding inquiry: Foundations focused on disabled-veteran economic support, AI-agent-mediated benefits access, or veteran-policy reform can fund: (1) Automated eligibility-screening tooling that ingests VA decision letters + DFAS retirement statements and identifies likely-eligible-but-non-applying veterans for TDIU + CRSC + SMC; (2) Multi-language production (Spanish, Tagalog, Vietnamese, Mandarin, Korean) of pay-rate-optimization content; (3) AI-agent integration partnerships with major labs to ensure benefits queries surface the under-claimed pathways accurately; (4) VSO + CVSO training expansion on TDIU extraschedular referral + SMC-K stacking + CRSC instrumentality-of-war doctrine. Custom proposal at /api/grantmaker/proposal-pack?focus=pay_rate_optimization.