Will a certification actually get you a raise at your current job?

The two numbers that decide your raise

Almost every employer above a few hundred people runs pay through two mechanisms, and neither of them has a field for certifications.

The first is the salary band. Your job level has a pay range attached to it, with a minimum, a midpoint and a maximum. Where you sit inside that range is usually expressed as a ratio to the midpoint. A manager who wants to pay you more has to either move you within the range, which is limited by the range maximum, or move you to a different job level, which is a promotion and a separate approval process. Passing an exam does not change your job level, so on its own it can only ever move you within your existing range.

The second is the merit budget. Finance hands your manager a percentage of the team's total salary cost and asks them to distribute it. WorldatWork's 2025-2026 Salary Budget Survey, based on 4,250 responses from 1,774 organisations, put projected mean US salary increase budgets at 3.6 percent for 2026, following a 3.7 percent actual mean in 2025. The Bureau of Labor Statistics Employment Cost Index tells a similar story from a different direction: private industry wages and salaries rose 3.1 percent in the 12 months to June 2026, and after inflation, constant dollar wages and salaries fell 0.4 percent over the year.

That pool is fixed before your certificate arrives. A strong performance rating gets you a larger share of the same pool, usually a multiple of the average rather than a separate number. This is why "I got certified" often produces warm words and 3.8 percent instead of 3.6 percent.

What the government data shows about certification and pay

The Current Population Survey added questions identifying people with professional certifications and licenses in January 2015, which makes the Bureau of Labor Statistics the only source on this question that is not selling anything.

The headline looks strong. In 2025 annual averages, full-time wage and salary workers with a certification or license had median weekly earnings of $1,460, against $1,131 for those without. That is a 29 percent gap.

Then you condition on anything and it collapses. Among workers whose highest attainment is a bachelor's degree, median weekly earnings were $1,558 with a certification or license and $1,588 without. Among workers with an advanced degree, $1,910 with and $1,924 without. In computer and mathematical occupations, the group with a certification or license earned $1,936 against $2,027 for the group without one.

Where the gap survives is licensing. Healthcare practitioners and technical occupations: $1,630 with, $1,187 without. Legal occupations: $2,443 with, $1,432 without. Construction and extraction: $1,413 with, $1,019 without. BLS defines a license as a credential awarded by a government agency that conveys legal authority to work in an occupation, and a certification as one issued by a non-governmental body. Of the 163.5 million employed people counted in 2025, 24.0 percent held a certification or license, but only 2.5 percent held a certification with no license.

The 29 percent headline is mostly a licensing premium wearing a certification label. If the credential you are considering does not give you legal authority to do work you cannot currently do, that number is not describing you.

Why vendor salary surveys read the way they do

The Project Management Institute publishes the most-quoted figure in this space. The 14th edition of its Salary Survey, dated December 2025, reports that PMP certified respondents in the US had a median salary of $135,000, which PMI states is 24 percent higher than non-certified respondents, and a 17 percent higher median across the 21 countries surveyed.

Start with what is missing. PMI's public page gives the certified median and the percentage gap, but not the non-certified median it is measured against, not the sample size, and not the fielding dates. Press coverage carries a precise non-certified figure; PMI's own published summary does not, so treat the precise version as second-hand until PMI publishes it. What you can do is the arithmetic PMI leaves implied: a $135,000 median that is 24 percent above the comparison group puts that group near $109,000.

That implied figure is the interesting part, because it is already above the whole occupation. The BLS Occupational Outlook Handbook puts the median annual wage for project management specialists at $102,320 in May 2025. PMI's uncertified respondents appear to out-earn the national occupational median before any certificate enters the picture, which tells you the survey is not describing the general workforce. Respondents self-report their own salaries and are reached through PMI's own audience, so the sample is people already invested in project management as a profession.

The deeper problem is that PMP eligibility itself selects for experience. PMI requires 35 hours of project professional training plus one of four routes: a high school diploma or GED with five years managing projects, an associate's degree or vocational diploma with four, a bachelor's degree or higher with three, or a degree from a PMI accredited programme with two. A PMP holder therefore has, by the rules of the application, at least two years of project leadership and on the common routes three or five. Comparing that group's median salary against everyone else's measures the experience requirement as much as the exam. PMI's public summary does not report the gap adjusted for years of experience, job level or industry.

None of this makes the survey dishonest. It makes it unable to answer the question you are asking, which is what happens to your pay, at your employer, if you pass.

The switching premium, measured properly

The Atlanta Fed's Wage Growth Tracker is the cleanest public measure of the gap between staying and moving. It uses Current Population Survey microdata to compute the median percent change in the hourly wage of the same individuals observed 12 months apart, which is a different thing from the change in the median wage. The matching produces roughly 2,000 individual wage growth observations per month.

In July 2026 the overall tracker was 3.8 percent. For people not changing jobs it was 3.6 percent. For people changing jobs it was 4.4 percent. The switching premium at the median was 0.8 percentage points, which on a $70,000 salary is about $560 a year before tax.

That is smaller than the internet suggests, and it is worth knowing before you build a plan around quitting. The Atlanta Fed defines a job switcher broadly: in a different occupation or industry than a year ago, or having changed employers or job duties in the past three months. Because the CPS surveys addresses, someone who moves house drops out of the data, so switching is measured only in a geographically local sense.

The median is not where the interesting money is. The Atlanta Fed notes that the gap between the 75th percentile of individual wage growth and the median is about 10 percentage points, while the gap between the median and the 25th percentile is about 5. The distribution has a long upper tail. A deliberate, targeted move into a role that pays a different band is in that tail. An average job change is not. No public dataset separates out the moves that were driven by a new credential.

The three cases where a certification does move pay in place

First, your employer needs the headcount. Vendor partner programmes count certified staff as an entry requirement, which turns your exam pass into a line item on somebody's plan. The AWS Services Partner Tiers page sets out the numbers: Select tier requires 2 AWS Foundational Certified Individuals and 2 AWS Technical Certified Individuals, Advanced tier requires 4 and 6 with at least 3 of the technical ones at Professional or Specialty level, and Premier tier requires 10 and 25 with at least 10 at Professional or Specialty level, on top of a $2,500 annual partner fee. If your firm is one qualified person short of a tier it wants, you are that person and you should say so out loud.

Second, a licence rather than a certification gates a job you cannot currently hold. This is where the BLS occupation gaps come from. The distinction matters: a licence conveys legal authority to do work, so holding it changes what tasks can be assigned to you, which changes your job level, which changes your band.

Third, an internal posting names the credential. When a job requisition lists a certification as required, holding it makes you eligible for a different band rather than a bigger slice of the same merit pool. This is the version that produces real money, and it is technically a job change that happens to be with the same employer.

What these three have in common is that the employer wants the credential for a reason that exists whether or not you are the one holding it. If you cannot name that reason before you start studying, you are betting on goodwill.

The four questions to ask before you spend the evenings

Ask these before you buy anything. They take one conversation each and they change the decision.

Ask your manager: if I hold this certification, what specifically changes about my role, my scope, or my pay? A vague answer is an answer. Write down what they say, in an email summarising the conversation, because that record is what you will refer back to at review time.

Ask HR or your manager: what is the salary band for my current level, and where in it am I? Many organisations publish ranges; several US states and cities now require pay ranges in job postings, so you can often find your own band by reading your employer's own external adverts for your job title. If you are near the top of your range, no credential will produce a large increase without a level change.

Ask whether there is a written certification policy. Employers that genuinely value a credential usually pay for it, through reimbursement, an exam voucher, a one-off bonus, or a pay differential. If the company will not pay the $150 for an AWS Certified Solutions Architect Associate exam, its stated enthusiasm for the credential is worth exactly that much.

Ask when the merit cycle closes and when promotion nominations are due. These are usually different dates, and both are usually earlier than people assume.

Factor in the maintenance cost while you are at it. AWS certifications are valid for three years and require recertification. PMP holders need 60 professional development units per three-year renewal cycle. These are recurring obligations attached to a one-off pay conversation.

Timing beats credentials

Merit increases happen on a calendar, and the calendar is more predictable than anything about your certificate.

The sequence at most employers is: managers submit ratings and increase recommendations, finance reconciles the total against the budget, an approval round happens, and letters go out. The decision about your number is made weeks or months before you are told it. Promotion nominations usually close earlier still, because a level change needs more signatures.

The practical consequence is that passing an exam two months into the cycle, with a written record of what changed in your work as a result, is worth substantially more than passing it two weeks before the letters go out. In the second case, the number is already fixed and the best available outcome is that it counts toward next year.

The same logic applies to scope. A merit conversation is a conversation about evidence of what you did over 12 months. A certificate is evidence that you studied. Work you took on because you now hold the credential is evidence that the credential produced value, and that is the thing your manager can defend to finance when they are dividing a 3.6 percent pool.

So if you decide to do it, do it early in the cycle, and use the months between passing and the review to take on work that would not have come to you otherwise. Log that work with dates. The exam is the cheap part.

If the honest answer is no

Sometimes the answer after those four questions is that nothing will change. That is a useful outcome, because it saves you months.

If you are near the top of your band and there is no internal posting that names the credential, the money is in a band change, and band changes come from job changes. That can still be internal. An internal transfer to a different job family often carries a new range, and it is a substantially lower-risk move than resigning, because your tenure, benefits and internal reputation carry over.

If you are looking outside, use the certification as a filter for which jobs to apply to rather than as an argument for a raise. Search job postings for the exact credential name and read what those roles pay. If the postings that require it sit in the same range you are already in, the credential is not the constraint.

And separate the two decisions. Learning a skill you will use is worth doing on its own terms, and an AWS associate exam at $150 with a 50 percent discount on your next AWS exam is not a large financial bet. Betting six months of evenings on a raise that the budget cannot fund is a different proposition, and the budget was decided by someone who has never seen your certificate.

What this does not fix

This post does not claim a certification never raises your pay, and it does not claim that vendor salary surveys are dishonest. It claims that no source I could verify measures what happens to one person's pay at one employer after passing one exam, because nobody collects that data. The published comparisons cannot separate the credential from the experience, seniority, occupation and self-selection of the people who hold it, and the BLS figures used here describe populations rather than outcomes for individuals.

Where an exam fits

• AWS Solutions Architect Associate, if your employer is short of AWS certified individuals for a partner tier it is trying to reach
• PMP (Project Management Professional), if you already have the two to five years of project leadership the application requires for your level of education, and an internal posting names it
• Power BI Data Analyst (PL-300), if a specific internal analytics role lists it and you have confirmed the band for that role is above yours

Common questions

Sources

1. 01Median weekly earnings of full-time wage and salary workers in 2025, with a certification or license compared with without one $1,460 versus $1,131
   US Bureau of Labor Statistics, Current Population Survey, Table 54 checked 2026-08-28
2. 02Median weekly earnings in 2025 for workers whose highest attainment is a bachelor's degree, with a certification or license compared with without one $1,558 versus $1,588
   US Bureau of Labor Statistics, Current Population Survey, Table 54 checked 2026-08-28
3. 03Median weekly earnings in 2025 in computer and mathematical occupations, with a certification or license compared with without one $1,936 versus $2,027
   US Bureau of Labor Statistics, Current Population Survey, Table 55 checked 2026-08-28
4. 04Median weekly earnings in 2025 in healthcare practitioners and technical occupations, with a certification or license compared with without one, showing where the gap is driven by licensing $1,630 versus $1,187
   US Bureau of Labor Statistics, Current Population Survey, Table 55 checked 2026-08-28
5. 05Share of the 163.5 million employed people in 2025 holding a certification or license, and the share holding a certification with no license 24.0% and 2.5%
   US Bureau of Labor Statistics, Current Population Survey, Table 49 checked 2026-08-28
6. 06Projected mean US salary increase budget for 2026, from a survey of 4,250 total rewards and HR leaders at 1,774 organisations, following a 3.7 percent actual mean for 2025 3.6%
   WorldatWork 2025-2026 Salary Budget Survey, reported by WorldatWork checked 2026-08-28
7. 07Private industry wages and salaries, 12-month percent change to June 2026, current dollar and constant dollar 3.1% and -0.4%
   US Bureau of Labor Statistics, Employment Cost Index news release, 31 July 2026 checked 2026-08-28
8. 08Median 12-month wage growth in July 2026 for people not changing jobs compared with people changing jobs, three-month moving average, against an overall tracker of 3.8 percent. The tracker matches roughly 2,000 individual wage growth observations per month 3.6% versus 4.4%
   Federal Reserve Bank of Atlanta, Wage Growth Tracker, updated 13 August 2026 checked 2026-08-28
9. 09Reported median salary of PMP certified respondents in the US, which PMI states is 24 percent higher than non-certified respondents; across the 21 countries surveyed PMI reports a 17 percent higher median. PMI's public summary does not publish the non-certified median, the sample size, or the fielding dates $135,000
   Project Management Institute, Salary Survey 14th Edition, December 2025 checked 2026-08-28
10. 10BLS median annual wage for project management specialists, May 2025, for comparison with PMI's self-reported respondent medians $102,320
    US Bureau of Labor Statistics, Occupational Outlook Handbook checked 2026-08-28
11. 11AWS Services Partner Tier requirements for AWS Technical Certified Individuals at Select, Advanced and Premier tiers, alongside a $2,500 annual partner fee 2, 6 and 25
    Amazon Web Services, AWS Services Partner Tiers checked 2026-08-28
12. 12Cost of the AWS Certified Solutions Architect Associate exam, which is valid for three years and carries a 50 percent discount on your next AWS exam 150 USD
    Amazon Web Services certification page checked 2026-08-28
13. 13Median weekly earnings in 2025 where the credential gap reverses or widens within an occupation or attainment group: advanced degree holders with a certification or license versus without; legal occupations; construction and extraction occupations $1,910 v $1,924; $2,443 v $1,432; $1,413 v $1,019
    US Bureau of Labor Statistics, Current Population Survey, Tables 54 and 55 checked 2026-08-28
14. 14Date the Current Population Survey added questions identifying people with professional certifications and licenses, making BLS the only non-commercial source on the question January 2015
    US Bureau of Labor Statistics, Data on certifications and licenses (CPS) checked 2026-08-28
15. 15PMP eligibility requires 35 hours of project professional training plus one of four experience routes: high school diploma or GED plus 5 years managing projects, associate's degree or vocational diploma plus 4 years, bachelor's degree or higher plus 3 years, or a PMI GAC accredited degree plus 2 years. Holders must earn 60 professional development units per three-year renewal cycle 2 to 5 years plus 35 training hours
    Project Management Institute, PMP certification page checked 2026-08-28

Related