The verified 90% VA disability rate for a veteran alone is $2,362.30 a month in 2026. A 2027 increase is not yet established by the official sources checked for this article. At a purely illustrative 3.0%, that starting amount would become approximately $2,433.17. But an annual dollar increase and a move from 90% to 100% are two different questions, and the rounded “90%” on a benefit record cannot tell you how close the underlying rating is to the next level.
Evidence checked October 5, 2026. Published amounts are 2026 monthly rates, effective December 1, 2025. Future amounts are explicitly conditional arithmetic.
The hidden distinction is large enough to change the answer to an apparently simple calculation. Under VA’s combined-ratings table, underlying combined values of 85, 90 and 94 all finish at a displayed 90%. Add the same next 10% rating in three correctly ordered examples, and the resulting table values are 87, 91 and 95. The first two still round to 90%. The third rounds to 100%.
That does not mean a veteran can choose a new rating or that a calculator can determine what VA should award. It means the rounded headline number throws away information. To understand a 2027 payment forecast, first keep the existing payment basis intact. To understand a possible rating change, recover the individual ratings and the applicable combination rules instead of adding ten points to a displayed total.
The confirmed amount, and the part still unknown
VA’s current compensation table gives $2,362.30 for a veteran alone at 90%, $2,559.30 for a veteran with a spouse and no dependent parents or children, and $2,704.30 for a veteran with a spouse and one child but no dependent parents. Each is a complete basic monthly amount for the stated household category.
The same table lists $98 for each additional child under 18, $317 for each additional child over 18 in a qualifying school program, and $181 for a spouse receiving Aid and Attendance. Those are separate specified additions. They are not a substitute for selecting the correct first-child basic row or confirming that the dependent or addition is recognized in the award.
No official 2027 dollar row was established in the evidence reviewed on October 5. SSA’s latest COLA page describes the 2.8% increase for 2026. That percentage has already been reflected in the current year’s schedule; it is not evidence that the next increase will also be 2.8%.
The introduced versions of S. 4487 and H.R. 8552 propose a veterans’ compensation adjustment effective December 1, 2026, linked to the Social Security percentage. The official status records reviewed did not establish enactment. A proposed date, a forecast and a published compensation schedule are different forms of evidence. A precise dollar figure cannot make an unfinished process final.
Under the 3.0% illustration, $2,362.30 multiplied by 1.03 is $2,433.169, rounded here to $2,433.17. The monthly difference is $70.87. Twelve identical hypothetical months would be $850.44 more, but that annualized figure says nothing about an individual effective date, retroactive award or bank deposit. It is simply twelve times the rounded monthly example.
The 90% payment is not nine-tenths of the 100% payment
There is a second arithmetic trap before the combined-ratings rules even enter. Disability percentages select rows in a compensation schedule. They are not percentages of one universal maximum payment. The 2026 veteran-alone 100% amount is $3,938.58. Ninety percent of that would be about $3,544.72, which is not the published 90% rate.
The actual $2,362.30 amount is about 60% of the ordinary veteran-alone 100% amount. The difference between the two published rows is $1,576.28 a month. It is therefore especially misleading to describe 90% as “only ten percent less pay.” The numerical gap in the rating label and the dollar gap in the statutory schedule are different relationships.
A common assumed COLA does not close that schedule gap. With a 3.0% arithmetic illustration, the two ordinary veteran-alone amounts would be approximately $2,433.17 and $4,056.74. Their difference would be $1,623.57. Both rise proportionally; the larger starting amount gains more dollars. None of that changes a disability rating.
For the limited spouse-only 90% example, applying the same assumption to $2,559.30 gives approximately $2,636.08, a $76.78 change. This is an illustration applied to an existing whole published row. Future dependent schedules may use statutory component calculations and rounding that differ from multiplying that row. The official new table, when issued, controls.
Both household examples exclude children and dependent parents.
Static comparison: veteran-alone 90% pays $2,362.30 and 100% pays $3,938.58 in 2026, a $1,576.28 gap. At an illustrative 3.0%, the amounts become $2,433.17 and $4,056.74. These are separate published categories, not points on a proportional rating-to-dollar formula.
What the displayed 90% hides
VA’s explanation of combined ratings describes a whole-person approach. Separate service-connected disability ratings are combined rather than simply added. The table’s result is then rounded to the nearest ten at the end of the process. Values ending in five through nine round up; values ending in one through four round down.
The governing 38 CFR 4.25 table makes the hidden range visible. A final combined value of 85 rounds to 90. So do 86, 87, 88 and 89. Values from 90 through 94 also finish at 90. The threshold for rounding to 100 is 95. There are ten possible integer combined values beneath the same displayed 90% label.
The word “unrounded” needs care here. It means before the final conversion to a multiple of ten. It does not mean a secretly precise medical measurement or a long decimal calculated by an online formula. VA’s table already produces integer combined values at intermediate steps. Replacing the table with a continuous spreadsheet formula, carrying decimal places indefinitely, can produce a different result.
Nor can the underlying number be recovered by working backward from 90 alone. It is a many-to-one mapping. The same outer label can contain different rating combinations, and the combination path matters. A benefit summary that displays 90 does not establish whether the applicable table calculation ended at 85, 90, 94 or another value in that interval.
The three illustrated records below are deliberately fictional, contain no medical conditions, and use exact cells from the regulation. A 70% and a 50% rating combine to 85. An 80% and a 50% rating combine to 90. A 90% and a 40% rating combine to 94. All three finish at 90% if these are the only ratings and no special combining rule applies.
Static paths: 70 and 50 combine to 85; 80 and 50 combine to 90; 90 and 40 combine to 94. All three final values round to 90%. “Unrounded” here means before the final rounding to a multiple of ten, not an unrounded medical percentage.
The same next 10% can lead to three different results
Take the first illustrated record, 70 and 50. Its combined value is 85. A further 10% rating, entered as the next lower rating, leads to table value 87. The final rounded result remains 90%. The arithmetic does not say the additional condition is unimportant; it says that this combination does not cross the next payment-rating boundary.
For the second record, 80 and 50 combine to 90 before final rounding. The next 10 leads to 91, which again rounds to 90%. This is the unusual case in which the underlying integer and the displayed rating happen to look the same at the starting point. It must not be used as a universal model for everyone whose final rating reads 90%.
The third record starts from 90 and 40, combining to 94. The next 10 leads to 95, which rounds to 100%. The difference is the starting table value. The same next rating has taken three records to 87, 91 and 95, with different final outcomes. Adding 10 directly to each displayed 90 would wrongly give all three the same answer.
The interactive table window allows an additional 10, 20, 30 or 40 in these presets. Those choices remain no greater than the existing lower rating, so the stated highest-to-lowest order is preserved. This constraint is intentional. An arbitrary new higher rating would require placing all individual ratings back into order and recomputing the entire sequence, not attaching a new number to a convenient subtotal.
For example, with an additional 40 the three table results become 91, 94 and 96. Their rounded results are 90, 90 and 100. The visual therefore shows both the intermediate value and the final display. It does not interpolate a result between table cells or pretend that the width of a graphic is another form of VA evidence.
Select a cell to change both the starting example and the additional rating. Entries are combined values before final rounding.
| Start | +0 | +10 | +20 | +30 | +40 |
|---|---|---|---|---|---|
| 85 | |||||
| 90 | |||||
| 94 |
Upper marker: starting table value. Lower marker: after the selected additional rating. Positions show integers from the table, not predictions between them.
Read the complete selected example
Static 10% comparison: 85→87→90% displayed; 90→91→90% displayed; 94→95→100% displayed. With additional 40%, the underlying results are 91, 94 and 96. No bilateral factor or individual adjudication is modeled.
There is a reason the source trail for these scenes points to the regulation’s table rather than copying an entire online calculator. In the versions checked for this article, VA’s general explanatory page displayed 91 at the intersection of an underlying 86 and an additional 80. The table in 38 CFR 4.25 displayed 97 at that intersection. The discrepancy is in that published explanatory cell; it is not evidence that an individual award was calculated incorrectly.
The three interactive records do not use that disputed cell. Their starting combinations and all fifteen selectable outcomes were checked against the regulation. This matters because importing an official-looking table without reading its entries can preserve a transcription problem inside an otherwise polished tool. A source logo and an accurate general explanation do not independently verify every number in the underlying data.
This feature also avoids turning the familiar “remaining efficiency” explanation into a replacement calculation. That explanation helps show why 70 and 50 do not add to 120, but the selected exact table cells are the evidence for the displayed results. The output does not generate intermediate decimal percentages or assign meanings to positions between the numbered cells.
Where the example stops being an award calculation
Real records can be more complicated than three descending ratings. The bilateral-factor rule, for example, can affect the treatment of qualifying disabilities of paired extremities before the final combination. This feature does not apply it. It also does not determine which conditions are service connected, which evaluations are assignable, whether symptoms overlap or whether other rating limits apply.
Those exclusions are reasons to use the full decision and the applicable rules, rather than confidence-building fine print beneath an unrestricted calculator. The controls contain predefined mathematical examples, not a form for entering diagnoses or an invitation to pursue a condition with a particular target percentage. An increase in compensation is a legal and evidentiary determination, not a score to manufacture.
Readers also should not infer the next available step solely from distance to 95. The official table operates on recognized individual ratings in the proper sequence. A numerical threshold describes an arithmetic boundary; it does not establish that an additional rating of a particular size is supported or that a pending claim will be decided in any particular way.
There is a further limit to the equation “displayed 90% equals the 90% pay row.” VA’s individual unemployability benefit can pay compensation at the 100% rate while the underlying disability rating remains lower. A person already paid at the total rate through IU should not use $2,362.30 as their ordinary payment baseline merely because the rating summary displays 90%.
Special monthly compensation can change the relevant schedule as well, and dependent categories change the applicable amount within a schedule. The financial comparison here is therefore explicitly limited to ordinary compensation at the stated rating, with either no dependents or a spouse and no dependent parents or children. Selecting an example does not identify the structure of anyone’s real award.
Use the payment record and rating decision for different questions
If the question is “What might the annual increase add?”, begin with the current awarded payment structure and the official year’s table. Confirm the dependent category and any total-rate or SMC treatment. Keep those facts fixed while testing an assumed percentage. Do not let a calculator quietly substitute the veteran-alone rate for an award that uses a different row.
If the question is “Could my combined rating change?”, begin with the individual ratings and the decision explaining them. The rounded summary alone is insufficient. Check the actual combination sequence, any special rules and the effective dates. A new decision can change more than one component at once, which is another reason not to reverse-engineer an award from a single deposit or a displayed percentage.
When the official 2027 table becomes available, replace the hypothetical dollars with its published amounts. That update answers the annual-rate question. It does not change the arithmetic rule that puts combined values from 85 through 94 into the 90% band, or establish a new evaluation for any disability. A COLA can increase dollars while every rating in the record remains exactly as it was.
The distinction also helps interpret a larger-than-expected payment. A rating decision, revised dependent award, retroactive period or other adjustment could be involved. Comparing a deposit with the feature’s estimate can identify a question to investigate, but not the cause. Use the decision dates and payment explanations rather than treating every dollar difference as the annual increase.
The useful bottom line for a 90% award
For the ordinary veteran-alone example, $2,362.30 is confirmed; $2,433.17 at 3.0% is conditional. The illustrative increase is $70.87 per month. A different recognized household or payment structure needs its own starting amount, and the official future schedule must replace the illustration when released.
For a rating-change question, the important missing number is the combined value before final rounding. The same printed 90 can hide materially different positions inside the band. The three exact-table examples show why an identical next 10% can leave two records at 90 and move another to 100, without implying that any of those fictional records describes a particular veteran.
Keeping those two questions separate is more useful than one oversized forecast. One concerns the price adjustment to a payment schedule. The other concerns the combination of recognized disabilities under a table. Knowing which question you are asking prevents the most consequential arithmetic error before it reaches a household budget.
The VA COLA status guide tracks the annual adjustment process. For the total-rate payment structure, the 100% disability guide explains SMC, IU and permanence.
Reviewed by Donna Fuscaldo. The publisher confirms ongoing editorial review. Prepared with AI-assisted research, writing and design.
Sources and calculations checked October 5, 2026. This is document analysis, not firsthand reporting or an individual benefits determination. No new interviews were conducted.