Is 3% a good raise in 2026? Depends on CPI
My dad got a 3% raise every year from 1998 until he retired in 2019, and for most of those years he came out ahead of inflation. Inflation averaged 2.1% from 1998 to 2019 (BLS CPI-U data). 3% raises meant he was netting about 0.9% real wage growth every year. Compounded over two decades, that's a real 20% purchasing-power gain on top of his salary, not counting promotions.
That math broke in 2021. Then 2022 incinerated it. Then it got weird in 2023 and has been getting back to normal since. The short version: 3% used to be a raise. Then it was a pay cut. Now it's back to almost breaking even.
So is 3% a good raise in 2026? The answer is a math problem. Let's do it.
The one number that decides it
CPI-U year over year, the Consumer Price Index for Urban consumers, is the benchmark most employers and economists use for inflation. The BLS April 10, 2026 release showed March 2026 CPI-U at 3.3% YoY.
That's the number you subtract from your nominal raise to get your real raise. Simple formula:
real_raise_pct = nominal_raise_pct - cpi_yoy_pct
Your 3% raise in a 3.3% inflation world: -0.3%. You make "more money" but can buy less with it than you could last year. That's a pay cut measured in goods.
Run the numbers through the pay raise calculator and it'll show you the real raise alongside the nominal one. The default CPI anchor is 3.3%; you can override for scenario modeling.
Where 3% came from as the "default"
Mercer, WillisTowersWatson, and SHRM have been publishing annual comp survey data since the 1990s. The merit increase median hovered around 3% from 2000 to 2020, give or take a few tenths. During that time, inflation averaged 2 to 2.5%. A 3% raise genuinely beat inflation.
Then 2021 hit. CPI-U went from a quiet 1.7% in early 2021 to peaking at 9.1% in June 2022. Companies hadn't built a playbook for that. Most kept the 3% merit pool and called it a "COLA." Which it wasn't. It was a 6% real pay cut being presented as a raise.
SHRM's 2023 and 2024 surveys showed merit bumps pushing up to 3.8% to 4.2%, but still tracking below actual CPI. For 2026, most companies I know about have budgeted 3.5 to 4.0%. In a 3.3% CPI world that puts the average employee at a very thin real-wage-growth margin.
The math in both directions
Here's a grid. Your nominal raise minus 2026 CPI of 3.3% gives your real raise.
- 2% nominal = -1.3% real (you're losing ground)
- 3% nominal = -0.3% real (barely treading water)
- 3.5% nominal = +0.2% real (marginal growth)
- 4% nominal = +0.7% real (modest growth)
- 5% nominal = +1.7% real (decent growth)
- 7% nominal = +3.7% real (strong growth, typical promotion bump)
- 10% nominal = +6.7% real (job change territory)
The job-change premium has historically been 10 to 20% of gross salary, which makes external moves the fastest way to stay meaningfully ahead of inflation. A 2023 Pew Research study found that workers who changed jobs from early 2022 to early 2023 saw median real wage growth of 9.7%, while those who stayed saw -1.7%. The job-change gap may be narrowing in 2026 as hiring cools, but historically it's 2 to 3x the stay bump in real terms.
What employers tell themselves
I've been in the management side of this conversation. Here's what usually happens at the comp planning table.
HR presents a merit budget (let's say 3.5%). The exec team looks at projected revenue, cost of attrition, market trends. Someone says "last year's pool was 3.2, we should probably bump a little." Nobody pulls up a CPI chart. Nobody does the real-wage math. The merit pool becomes a budget item, not an inflation-adjustment item.
Then the pool gets distributed. Top performers get 5 to 7%. Middle get 3 to 4%. Bottom get 1 to 2% or nothing. The average is the budget number. The median is often lower.
If you're a middle performer and you get 3% in a 3.3% CPI year, you are not getting punished; you're getting the default. The only way out of the default is to either move up a performance band internally (which means a bigger merit pool share) or move externally (which means paying the job-change premium).
The COLA semantic trick
Some employers label the 3% merit raise as a "Cost of Living Adjustment." This is technically false if it doesn't match CPI. A true COLA indexes to CPI-U or CPI-W. Social Security does a proper COLA every year (2026 SSA COLA was 2.5% based on Q3 2025 CPI-W). Federal employees get a COLA-like adjustment. Most private-sector "COLAs" are a marketing term for the merit pool.
If your HR letter says "COLA" and gives you less than CPI, push back. More in the COLA fight.
What to do about it
If you got 3% this year:
- Check CPI-U at the BLS release. If it's below 3%, you're fine. If it's 3.0 to 3.5%, you're treading water.
- Run the calculator to see the after-tax impact. 3% gross is 2% to 2.3% net depending on state and bracket.
- Plan for a negotiation next cycle using the 3-email playbook.
- If your role has a higher market rate than your comp, consider external interviews. You don't have to accept; you just have to know your market value.
For the bracket math on whether a bigger raise would push you into a new bracket (spoiler: it wouldn't cost you), see bracket creep explained.
The honest answer
In 2026, a 3% raise is a 0.3% real pay cut. That's a small haircut, not a crisis, but it is not a raise by any economic measure. If your employer is framing it as a raise, they're using the headline number, not the math.
4% this year beats inflation by 0.7%. 5% is a decent result. 7% is a promotion-tier bump. Benchmark against CPI, not against last year's 3%.
Run your own numbers in the calculator. Read the guide for the full tax-plus-inflation walk-through. And if you're being offered 3% and thinking about whether to push back, yes, push back. Worst case they say no.