The meeting is in late January. Your manager opens with the number, which is 3.4%, and follows it with the sentence that is meant to make it land well: the pool was 3.2%, and they fought to get you above it. You say thank you. There is nothing else to say, because by the time the two of you sat down the decision was three months old.

Most advice on asking for a raise treats it as a negotiation in that room. For the annual increase, it mostly is not one. The budget is set centrally, it is flat, and the spread between an average performer and a strong one inside it is small. The larger movements in pay come through other mechanisms, each with its own rules, evidence and calendar. The useful conversation is the one that happens earlier and asks a different question: not “can I have more?” but “which door is open, and what does it take to go through it?”

The pool is small, flat, and set before you sit down

US employers surveyed by Mercer in July 2026 plan average merit increases of 3.2% for 2027, and total salary increases of 3.5% once promotions and other adjustments are included [1]. WTW, surveying 1,650 organizations, finds much the same: budgets of 3.4% expected for 2027, against 3.5% actually delivered in 2026 [2]. These are plans, not outcomes, but they have barely moved in two years.

They also buy less than they appear to. Real average hourly earnings for private-sector employees in the United States fell 0.2% in the year to July 2026 [3]. An average merit increase has recently done little more than keep pace with prices.

And the pool is fixed early. As the article on promotion decisions sets out, the money behind a January decision is shaped in the second half of the year before. A request made in the review meeting arrives after most of the arithmetic is finished.

Inside the pool, the ceiling is low

How much of that pool can a strong performer claim? Less than most people assume. In merit-cycle data from the first half of 2026, employees rated above expectations who were not promoted received a median raise of 4.7%, against 3.3% for those who met expectations [4]. Mercer puts the typical merit premium for top performers at about 2% over colleagues who meet expectations [5]. And about a third of organizations, 32%, plan to give a uniform across-the-board increase in 2027 [6], which leaves less of the budget to follow the rating.

A point and a half is worth having, and it compounds. A well-evidenced self-review is how you earn it. But it is the smallest of the available moves, and it is the only one most people ask about.

Four other doors

Promotion. In Mercer’s late-2025 planning survey, employers budgeted an average increase of 8.7% for promotions in 2026, against 3.2% for merit [7]. A change of level is worth nearly three years of merit increases at once. It is also decided by a different process, with different evidence, and usually on a different timetable.

An off-cycle or market adjustment. Nearly two-thirds of organizations in Mercer’s 2026 survey, 64%, say they have made off-cycle salary adjustments or will make them in 2027 [1]. These exist to fix a specific problem: a salary that has fallen behind the market or behind newer colleagues. Employers know the problem is real. In Robert Half’s latest salary guide, 57% of managers say they are offering more than planned to attract new hires [8], and when the firm surveyed executives during the 2022 hiring boom, 56% had seen pay gaps open between new hires and longer-serving staff, with 62% of those saying they were raising existing employees’ pay to close them [9].

Whether this door is open to you depends on your role. Payscale’s 2026 analysis finds that where new hires hold a market advantage they earn on average 3.6% more than tenured colleagues in the same job, while in roles where experience is what counts, tenured employees earn 6.1% more than new hires [10]. Find out which describes your job before you argue that you are underpaid.

The market itself. You can now check. New York Fed economists report that the share of US online job postings carrying pay information rose from an average of 15% before 2018 to roughly 53% since January 2024, and that ten states and the District of Columbia had adopted pay-range posting laws by late 2025, though about 24% of covered ads still omitted the required range [11]. Five current postings for your role, level and region are a reasonable starting point for a market case that five years ago would have required a consultant. Posted ranges are often wide and describe hiring bands, so treat them as a guide to where the market sits, not as proof of what you should be paid.

An outside offer. This door has narrowed. In ADP’s payroll data for September 2026, base pay rose 4.8% over the year for people who changed jobs and 3.0% for those who stayed [12]. That gap peaked at 8.2 percentage points in April 2022 and has sat at about two points since August 2025 [13]. The Atlanta Fed’s wage tracker shows a similar picture: 4.7% for switchers and 3.4% for stayers in September 2026 [14]. Leaving still pays more than staying, on average, but by much less than the folklore of 2022 suggests.

When an outside offer is used as leverage, it often ends in a renegotiation. A study of 772 German firms and nearly 10,000 workers found that most firms are willing to bargain individually over pay, and that most such episodes end with the worker turning the outside offer down and staying [15]. The practical rule is an old one: only bring an offer you would accept.

What asking does, and does not, achieve

It is worth being clear about the base rate. In a study of about 4,600 Australian workers across more than 800 employers, men who asked for a raise received one 20% of the time and women 16% of the time, although women asked as often as men [16]. Asking, by itself, usually does not work.

It also carries a cost that falls unevenly. In laboratory experiments, evaluators penalized candidates who initiated a negotiation over pay; male evaluators penalized women more than men, and female evaluators penalized everyone who negotiated [17].

Transparency has a less obvious effect. Research on US state laws protecting workers’ right to discuss pay found that wages fell by about 2% on average afterward: once everyone can compare, an employer has a credible reason to refuse any one person a special deal, since it would have to be extended to the rest [18]. That finding concerns laws about discussing pay, not about posting ranges, but the logic is worth taking seriously. A request framed as a personal exception is easy to decline. A request framed as a correction to the employer’s own pay band, or to a level the role already operates at, asks the organization to apply its system and not to break it.

Matching the door to your situation

The doors are not interchangeable. Each fits a particular situation, and choosing the wrong one is the most common reason a well-prepared request fails.

  • Your scope has outgrown your level. You are doing work the level above normally does. This is a promotion or re-levelling case. The evidence is next-level work already delivered, and the timing is whenever nominations are gathered, which is usually well before reviews.
  • Your pay sits below the range for your own role. Posted ranges for the same job, level and region are higher than your salary. This is a market adjustment. The evidence is external, the request can be made at any point in the year, and the answer often depends on whether money was set aside for adjustments when the budget was built.
  • Newer colleagues are paid more for the same job. This is compression. Make the case from published ranges and your own record, not from what individual colleagues earn, and ask whether the organization runs a periodic review for it.
  • You are paid fairly for your level and rated well. Then merit is your door this year, and the larger question is what it would take to reach the next level.
  • No internal door is open and the market is well above you. The outside option is now the real one. Treat it as a decision about where you want to work, not as a tactic.

Whichever applies, ask one further question: what else is adjustable? Where base salary is fixed for the cycle, an employer may still have latitude on a one-time payment, on equity, or on the scope and title that set up the next increase. You will not know unless you ask what the options are before you ask for a number.

The conversation, in order

Taken together, the evidence points to a sequence. Most of it happens before review season.

  • Ask about the mechanism first. Which of these exist here: merit, promotion, off-cycle adjustment, re-levelling? Who approves each, and when is each decided? This is a process question, and managers generally answer it.
  • Choose one door. A promotion case, a market adjustment and a merit argument need different evidence. Mixing them weakens all three.
  • Bring two exhibits. The first is your record: three to five outcomes with numbers and dates. The second is the market: current posted ranges for your role, level and region, or the internal band if it is published.
  • Name the number and the basis. “I am asking to be considered for an adjustment to this figure, based on these ranges and this scope” can be acted on. “I feel I am underpaid” cannot.
  • Ask what would have to be true. If the answer is not yet, ask what evidence or scope is missing and when the question will next be decided. Write both down.
  • If the answer is the budget, ask for a date. An off-cycle review at mid-year, agreed now, is worth more than sympathy in January.

Persona Map is built to hold the first exhibit: a private workspace where outcomes and feedback are recorded as they happen, so the case does not have to be rebuilt each cycle. The rest is timing and choice of door. The annual increase is a number your manager is handed. The larger moves are decisions someone makes, earlier in the year, on evidence. Put your effort where a decision is still open.

Sources

  1. Mercer. “US Employers Stick With Moderate Pay Increases for 2027.” QuickPulse US Compensation Planning Survey, 2026. Link
  2. WTW. “Employers Shift from Bigger Pay Budgets to Smarter Pay Strategies, WTW Finds.” Salary Budget Planning Report, July 15, 2026. Link
  3. US Bureau of Labor Statistics. The Economics Daily: real average hourly earnings, July 2025 to July 2026. Link
  4. Pave. Merit increase benchmarks by performance rating, first half of 2026. Link
  5. Mercer. “Transforming Merit: From Flawed to Fair.” Link
  6. Payscale. 2026–2027 Salary Budget Survey, press release. 2026. Link
  7. Mercer. “Most US Employers Plan to Keep 2026 Salary Increases Flat.” QuickPulse US Compensation Planning Survey, 2025. Link
  8. Robert Half. 2027 Salary Guide, press release. October 1, 2026. Link
  9. Robert Half. “56% of U.S. Companies Have Experienced Pay Compression in the Last 12 Months.” April 8, 2022. Link
  10. Payscale. 2026 Flight Risk Report. Link
  11. Federal Reserve Bank of New York, Liberty Street Economics. “Do Employers Comply with Pay Transparency Requirements in Job Postings?” October 2025. Link
  12. ADP. National Employment Report, September 2026. September 30, 2026. Link
  13. ADP Research. “Pay Matchups: No Bricks or Airballs This Season.” Main Street Macro. Link
  14. Federal Reserve Bank of Atlanta. Wage Growth Tracker, September 2026 reading. Link
  15. Sydnee Caldwell, Ingrid Haegele & Jörg Heining. “Bargaining and Inequality in the Labor Market.” NBER Working Paper 33396, 2025. Link
  16. University of Warwick. News release on Artz, Goodall & Oswald, “Do Women Ask?” September 2016. Link
  17. Hannah Riley Bowles, Linda Babcock & Lei Lai. “Social Incentives for Gender Differences in the Propensity to Initiate Negotiations: Sometimes It Does Hurt to Ask.” Organizational Behavior and Human Decision Processes 103(1), 2007. Link
  18. Zoë Cullen & Bobak Pakzad-Hurson. “Equilibrium Effects of Pay Transparency.” NBER Working Paper 28903; Econometrica 91(3), 2023. Link