In 2023, the Board of Veterans' Appeals granted 26.9% of the decisions I have on file. In 2025, the Board granted 30.4%. Grants are up.

That sounds like good news. It mostly isn't.

Across the same window, denials went from 37.1% to 44.1%, up 7 points. Remands went from 32.7% to 22.5%, down 10.2 points. The Board is deciding more cases on the merits and remanding fewer. Some of those merits decisions land as grants. More of them land as denials. The cases that previously got sent back for development are now getting an answer, and the answer is increasingly "no."

I want to walk through what changed between 2023 and 2025 across 83,738 BVA decisions, what the shift means for a veteran with a pending appeal, and where I'm honestly unsure about what's driving the numbers. Some of the story is the PACT Act. Some of it is the Appeals Modernization Act finally working through the system. Some of it is the dataset itself, which I'll be explicit about because the 2023 sample is small.

TL;DR

  • 2023: 4,706 cases, 26.9% granted, 37.1% denied, 32.7% remanded.
  • 2024: 35,489 cases, 29.4% granted, 40.1% denied, 26.3% remanded.
  • 2025: 43,543 cases, 30.4% granted, 44.1% denied, 22.5% remanded.
  • The headline shift: grants up 3.5 points, denials up 7.0 points, remands down 10.2 points.
  • The total "non-denial" rate (grants plus remands) is going down, not up: 59.6% in 2023, 55.7% in 2024, 52.9% in 2025.
  • The 2023 sample is small enough that I'd weight conclusions toward the 2024-to-2025 trend, not the 2023-to-2025 trend.

The Shift In One Sentence

The headline summary

The Board is granting slightly more often and denying meaningfully more often, because it's stopped using remands as a holding pattern.

That's the whole story. Everything below is what's behind it.

The number that captured my attention isn't the grant rate. Most veterans reading this will see "grant rate up 3.5 points" and assume the Board is easier on claims now than it was two years ago. That's not what's happening. The Board is deciding more cases, full stop. More decisions per case means more grants AND more denials per 100 cases. The remand category is the one shrinking.

Here's how the three buckets shift across the three years.

  • Grants: 26.9% (2023) → 29.4% (2024) → 30.4% (2025). A 3.5-point lift across two years.
  • Denials: 37.1% (2023) → 40.1% (2024) → 44.1% (2025). A 7.0-point lift.
  • Remands: 32.7% (2023) → 26.3% (2024) → 22.5% (2025). A 10.2-point drop.

If you add the grant and remand rates together, the "you didn't lose" rate, it's 59.6% in 2023, 55.7% in 2024, and 52.9% in 2025. The veteran's overall odds of not getting a denial at the BVA are down 6.7 points over the same window the grant rate is up.

That's the quiet shift. The headline is grants up. The reality is denials up more.

The Remand Cliff: 32.7% to 22.5%

The biggest single change in the data is the remand rate.

In 2023, roughly one in three Board decisions in the dataset was a remand. By 2025, that's slipped to roughly one in five. The 10.2-point drop is the largest year-over-year movement in any of the outcome categories.

A remand isn't a loss. It's the Board telling the regional office, the AMC, or itself: this case isn't ready for a final answer. Usually because the evidence is incomplete, the C&P exam was inadequate, or some development step was skipped. The veteran goes back into the pipeline for more work before a final ruling.

For years, remands were the Board's release valve. When a case came up and the evidence wasn't clean enough to grant but wasn't bad enough to deny, the Board could remand for further development and kick the question down the road. That release valve is closing.

I have a few theories about why.

Theory one: the Appeals Modernization Act is working as designed. The AMA, which took effect in February 2019, restructured the appeals system around three lanes (Higher-Level Review, supplemental claim, BVA appeal) and three Board dockets (direct, evidence, hearing). The system was designed to reduce remands by front-loading evidence development. Cases that needed more development would go through HLR or supplemental claim lanes at the RO. Cases that reached the Board on the direct docket would be ripe for decision. The early years of AMA implementation didn't show that effect cleanly because the legacy appeals pipeline was still bleeding through. By 2024 and 2025, the legacy cases are aging out, and what's left at the Board is closer to AMA-pipeline cases, which, by design, are supposed to be more decision-ready.

Theory two: VA's backlog pressure is pushing the Board toward finality. The Board has been under sustained operational pressure to dispose of pending appeals. Remands look good on a process metric (they're not denials) but bad on a cycle-time metric (the case is still pending). If the Board is being measured on cycle time, the incentive is to decide rather than remand.

Theory three: the case mix has changed. The cases reaching the Board in 2025 may differ systematically from the ones reaching it in 2023. More PACT Act presumptive cases (which often don't need development if the exposure and condition criteria are met). More cases that already went through HLR (where evidence was developed at the RO). Fewer legacy-pipeline cases with thin records.

I cannot disentangle these three from the data. The trend is consistent with all three, and probably all three are doing some work.

What's In The Denial Increase

Denials went up 7 points. Where did those denials come from?

The math has to balance. If grants went up 3.5 points and remands went down 10.2 points, denials have to absorb most of the remand reduction. They did.

The plain reading: cases that would have been remanded in 2023 are now being denied in 2025. The Board is looking at the same kind of case, incomplete evidence, weak nexus, ambiguous service connection, and deciding the question rather than punting on it. When the Board decides those cases without more development, the evidence isn't enough to grant, so the decision goes the other way.

This is the part most veterans haven't internalized about the current Board posture.

If your case is on the bubble, adequate evidence, decent nexus, but not airtight, the 2023 Board was more likely to remand it for one more pass through development. The 2025 Board is more likely to decide it on the record in front of them. If the record doesn't support a grant, that's a denial.

That has practical implications for how you build your appeal.

You can't rely on the Board to send your case back for development if the record is thin. The Board did that in 2023 about a third of the time. In 2025, it does that less often. The work you would have hoped the remand pipeline would do for you, get the C&P exam redone, get the missing records pulled, get a clarifying medical opinion, increasingly has to happen before the Board sees the case.

That's a shift in burden. The Board's "we'll send it back for development" backstop has narrowed.

Why Grants Crept Up Anyway

Grants are up 3.5 points across the window. Given the remand-cliff dynamic, where are the extra grants coming from?

Two contributions, as best I can read the data.

The first is the PACT Act tailwind. The PACT Act took effect in August 2022, and it dramatically expanded the list of presumptive conditions for burn pit, Camp Lejeune, radiation, and Agent Orange exposures. Presumptive cases have a structurally higher grant rate than non-presumptive cases in my data, 46.9% versus 41.2% for direct claims. As PACT Act claims work their way through the system, the share of presumptive cases at the Board grows, which lifts the overall grant rate even if nothing else changes.

How much of the 3.5-point lift is PACT Act-driven? I don't know exactly. The presumptive case share in my dataset is around 9% of decisions, and PACT Act expansions added meaningfully to that share starting in 2023. A back-of-envelope estimate: if PACT-eligible presumptive cases grew from roughly 5% to 10% of the Board's docket, and those cases grant at 47% versus a 30% baseline for everything else, that alone could account for around 0.8 points of the grant-rate lift. The rest has to come from somewhere else.

The second is the Board's adjustment to better-developed records. If the AMA pipeline is delivering cleaner cases, cases that already went through HLR or supplemental claim, those cases come to the Board with a record that's closer to decision-ready. When the record is more developed, the underlying merits get decided. Sometimes that's a grant. The 30.4% grant rate in 2025 is the highest in my three-year window, even as denials hit their highest point too.

This is the Board getting faster, not getting easier.

The Total "Non-Denial" Rate Is Falling

Here's the framing that gets lost when you only read the grant-rate headline.

If you're a veteran appealing to the BVA, you care about three possible outcomes: grant (you win), remand (it gets sent back, you might still win), or denial (you lose, with finality limits attached). The relevant question is "what are my chances of not getting a denial?", which is grants plus remands.

  • 2023: 59.6% non-denial (26.9% grants + 32.7% remands).
  • 2024: 55.7% non-denial (29.4% + 26.3%).
  • 2025: 52.9% non-denial (30.4% + 22.5%).

The veteran's "didn't lose" rate is down 6.7 points across the two-year window.

This is the number I'd anchor on if I had a pending appeal. The probability of walking out of the Board with a grant in hand is up, slightly. The probability of getting a denial is up more. The probability of getting a "you live another day" remand is down significantly.

That changes the math on whether and how to file.

If you're in a position where your evidence is strong, the trend favors you, the Board is deciding cases on the merits more often, and strong evidence wins on the merits. If your evidence is weak and you were hoping the Board would send it back for development, the trend does not favor you. The Board is doing less of that.

The takeaway from the 2025 numbers is not "appeals are easier now." It's "the Board is being more final."

The PACT Act Tailwind

How PACT Act presumptives changed the slope

I want to spend a section on the PACT Act because it's the most-discussed policy change in the window, and it deserves its own treatment.

The PACT Act (Sergeant First Class Heath Robinson Honoring our Promise to Address Comprehensive Toxics Act of 2022) expanded VA presumptive coverage for veterans exposed to burn pits, airborne hazards, and other toxic exposures during service in Iraq, Afghanistan, the Persian Gulf, Vietnam-era Thailand and Cambodia, and several other theaters. It added 23 new presumptive conditions, expanded coverage areas, and removed time-of-onset requirements for many conditions.

The Act took effect August 10, 2022. By 2023, claims were filtering through. By 2024 and 2025, those claims were reaching the Board.

What does that look like in my data? A few markers.

  • Presumptive cases grant at 46.9% across the full dataset, versus 41.2% for direct service connection. When the share of presumptive cases rises, the average grant rate rises with it.
  • Burn-pit-flagged cases (a presumptive subcategory) cluster in conditions like sinusitis, COPD, asthma, eczema, and certain cancers. These are the conditions where 2024-2025 grant rate movement is most visible.
  • Camp Lejeune presumptive cases, water contamination exposure, show up disproportionately in kidney disease, Parkinson's, and certain cancers, all of which had high grant rates in the 2024-2025 sample.

What does that NOT tell me?

It doesn't tell me what share of the grant-rate lift is PACT Act-driven versus other causes. The presumptive coverage expansion is real and the grant-rate effect is real, but my data can't cleanly separate "PACT Act-eligible grants" from "non-PACT grants in the same conditions." The categories overlap.

It also doesn't tell me whether the PACT Act tailwind will continue. The biggest grant rate gains from a presumptive expansion typically happen in the first 2-3 years after enactment, as the backlog of previously-denied claims gets resubmitted under the new criteria. After that initial wave, the effect normalizes. The 2024-2025 numbers may be at or near the peak.

The 2023 Sample Problem

I've referenced the 2023 numbers throughout this post. I want to be explicit about a methodological caveat that changes how much weight you should put on them.

The 2023 sample in my dataset is 4,706 cases. The 2024 sample is 35,489. The 2025 sample is 43,543. The 2023 sample is roughly 13% the size of the 2024 sample and 11% the size of the 2025 sample.

Why?

A few reasons.

The first is that BVA decisions are released publicly on a delayed and uneven schedule. A 2023 decision may not enter the public corpus until 2024 or later, depending on classification, redaction, and posting workflows. My dataset is built from the publicly available decisions, and 2023 cases are still trickling in even as I write this.

The second is that my extraction pipeline has been improving over time. Cases I couldn't classify cleanly in early 2024 may now be parseable. That improvement disproportionately benefits the cases that posted earliest, which biases the 2023 sample toward decisions that were cleaner or simpler from the start.

The third is that 2023 was the first full year of the Appeals Modernization Act being the dominant appeal pathway. The Board's case mix in 2023 differs from 2024 and 2025 in ways that don't necessarily reflect Board policy, they reflect the AMA pipeline maturing.

For all of these reasons, the 2023 numbers are noisier and possibly non-representative. I report them honestly because they're the only 2023 numbers I have. But if you're using this post to think about your own pending appeal, weight the 2024-to-2025 comparison more than the 2023-to-2025 comparison. The 2024 sample is large enough that the trends are stable. The 2023 sample is more of a directional indicator than a precise baseline.

The directional read still holds. The Board is granting slightly more, denying meaningfully more, and remanding less. That's true whether you start the clock at 2023 or 2024. The magnitude of the shift depends on which baseline you use.

What I Can't Tell You

A few honest limits.

First, I cannot tell you what 2026 looks like. The trends in the 2023-2025 window suggest the remand rate keeps falling and the denial rate keeps rising. That's not a forecast, that's an extrapolation. Trends can flatten or reverse. The next PACT Act expansion, the next CAVC decision on duty-to-assist, the next congressional appropriations cycle could all change the pattern.

Second, I cannot tell you the specific drivers of the shift with certainty. I gave three theories above. AMA pipeline maturation, backlog pressure, case mix change. Probably all three. I can't quantify their relative contributions. I can tell you they're consistent with what's in the data.

Third, I cannot tell you what's happening at the regional office level. Everything in this post is BVA-only. The RO has its own grant, denial, and deferral patterns that aren't captured here. A shift in Board behavior doesn't tell you anything about what's happening upstream at the RO, where the vast majority of claims are decided before they ever reach the Board.

Fourth, my dataset is not the universe of BVA decisions. It's a large sample of publicly available decisions. There may be systematic differences between what's in the public corpus and what's in the Board's internal records. I report what I have. I don't claim it's complete.

Fifth, the dataset overrepresents recent decisions. Older cases are underrepresented because they haven't all been posted, parsed, and classified. That biases any trend analysis toward the most recent year. If the 2025 sample is more complete than the 2023 sample (it almost certainly is), comparisons across years carry some measurement noise that's hard to fully control for.

Sixth, individual outcomes are not predicted by aggregate trends. The Board's overall grant rate may be 30.4% in 2025, but that's the average across every condition, every connection type, every record quality tier. Your specific case has its own facts. Aggregates describe distributions, not individual cases.

What This Means For Your Pending Appeal

If you have a Board appeal in flight right now, here's what I'd think about given what I see in the data.

Your record needs to be decision-ready by the time it reaches the Board. The 2023 backstop where the Board sends thin cases back for development is narrower in 2025. If your C&P exam is weak, your nexus letter is missing, or your service treatment records are incomplete, you may not get a remand to fix that, you may get a denial. Whatever you can do to strengthen the record before the Board picks up the case is more valuable now than it was two years ago.

The strong-evidence path still works. Cases with strong nexus opinions, adequate C&P exams, and well-documented in-service events grant at much higher rates than the 30.4% average. The headline grant-rate shift doesn't change what wins on the merits. It changes what happens to cases that aren't ready for a clean merits decision.

The "wait and hope for a remand" strategy is weaker than it used to be. If your plan is to get the case in front of the Board and rely on the Board to develop the record for you, the 2025 numbers say that strategy is failing more often. If you have time to develop the record yourself, through a private nexus letter, additional lay statements, updated medical records, or a request for a new C&P at the RO, that work has more value now.

Presumptive conditions and PACT Act eligibility matter more than ever. Presumptive claims grant at 46.9% across the dataset. If your condition is on a presumptive list and you served in a covered location during a covered period, file the presumptive claim and document the exposure carefully. The presumptive pathway sidesteps the nexus-opinion requirement that's responsible for most non-presumptive denials.

Appeal lane choice matters. If your evidence is incomplete and the BVA appeal lane is going to result in a denial rather than a remand, consider whether HLR or supplemental claim lanes at the RO are better suited to your case. Those lanes have more room for development and may produce a better outcome than rolling the dice on the Board's current posture.

If you don't know what your appeal lane should be, talk to an accredited representative. I can tell you what the data shows. I can't tell you what's right for your case.

Bottom Line

The Board of Veterans' Appeals granted 26.9% of its decisions in 2023. By 2025, that number is 30.4%. That's the headline. The fuller picture: denials are up 7.0 points and remands are down 10.2 points across the same window. The Board is deciding more cases on the merits and remanding fewer. Some of those merits decisions land as grants. More of them land as denials. The total non-denial rate, grants plus remands combined, is actually falling, from 59.6% to 52.9%. The PACT Act tailwind and the AMA pipeline maturation are probably both contributing. The 2023 sample is small enough that I'd weight the 2024-to-2025 comparison more than the full three-year span. The practical lesson is that the Board is getting faster, not easier, and the weight of building a strong record is shifting earlier in the process. Cases that used to get sent back for development are now getting decided on what's there. If what's there isn't strong, the answer is increasingly no.

Methodology and Limitations

  • Data source: Claim Raven's analysis of 83,738 BVA decisions decided in calendar years 2023, 2024, and 2025 from the production case_analysis dataset.
  • Sample sizes by year: 2023 = 4,706 cases; 2024 = 35,489 cases; 2025 = 43,543 cases. The 2023 sample is roughly 13% the size of the 2024 sample and 11% the size of the 2025 sample. The 2023 numbers should be treated as directional rather than precise.
  • Outcome classification: grant, denial, remand, and dismissed categories are extracted from BVA decision dispositions. Mixed outcomes (partial grants with partial remands) are classified by the dominant disposition.
  • Limitation, extrapolation: Trend extrapolation beyond 2025 is speculation. The patterns described here are descriptive of the 2023-2025 window, not predictive of future outcomes.
  • Limitation, 2023 sample: The 2023 sample is small enough that point estimates have meaningful confidence intervals. The 2024 and 2025 samples are large enough that the trends are stable.
  • Limitation, coverage: Dataset coverage of older years is incomplete. BVA decisions are released to the public corpus over time, and 2023 decisions are still being added years after issuance. This creates selection bias in older-year samples that I cannot fully correct for.
  • Limitation. PACT Act effect: The PACT Act tailwind is real but unquantifiable in this dataset. The grant-rate lift in 2024-2025 is consistent with a presumptive-coverage expansion effect, but my data cannot separate PACT-eligible grants from non-PACT grants in overlapping conditions.
  • Limitation, drivers: AMA pipeline maturation, backlog pressure, and case mix changes are all plausible contributors to the remand-cliff trend. The data is consistent with all three. I cannot isolate the contribution of any one cause.
  • Limitation, scope: These are BVA-level outcomes only. Patterns at the regional office level may differ. Veterans whose claims were granted at the RO and never appealed are not in this dataset.
  • Limitation, individual outcomes: Aggregate trends do not predict individual outcomes. Your specific case depends on facts no aggregate analysis captures.

Disclaimer

I'm not accredited by VA, not a lawyer, not a VSO. This is data analysis, not claim advice. These are patterns from cases that made it to the BVA, they don't predict individual outcomes. If you need help with your claim, work with an accredited representative.

Where to go next

  • Related article: Repeat Remand Loop at the BVA
  • Relevant tool: Evidence Checklists, what to develop now so the Board doesn't have to remand later
  • More analysis: /blog
  • Disclaimer: Claim Raven is data analysis, not legal, medical, or VA-accredited advice.

-Landon Founder, Claim Raven | U.S. Army Veteran