Salary Budgets Have Stalled at 3.5%, and Pay Strategy Is All That’s Left to Change

Every September, HR leaders walk into budget season hoping this is the year the raise pool grows. Three major 2026 surveys have now said the same thing: it is not. The money for 2027 looks almost exactly like the money for 2026, and what has changed is how much pressure sits on the decisions made inside that fixed number.

Three Surveys, One Number

WorldatWork's 2026-2027 Salary Budget Survey, drawn from 1,799 organizations, projects a mean U.S. salary increase budget of 3.6% for 2027. That matches the 2026 actual and marks the fourth consecutive year without a higher annual projection. Content director Sue Holloway described budgets as having reached "a new equilibrium," with employers budgeting carefully while still recognizing the need to invest in pay to attract and retain critical talent.

Marsh's July QuickPulse survey of 1,001 U.S. organizations, published August 31, arrives at nearly the same place: a 3.2% average base merit increase and a 3.5% average total increase once promotions and other adjustments are counted. The promotion rate in the plan is 8.4%. In its own words, the firm found that "compensation dollars are tight," with 57% of employers expecting economic conditions to affect compensation plans.

The spread underneath the average is where the tension lives. According to HR Brew's coverage of the Marsh data, high tech employers are budgeting 3.8% merit raises while consumer goods employers are budgeting 2.9%. The same piece quotes Marsh's Mark Bowling saying the high water mark was 2023 and budgets have moderated since. Roughly a point of merit separates the best-funded sector from the least, and neither is expecting a reversal soon.

The Average Hides Who Is Really Spending

WorldatWork also flagged a statistical wrinkle that matters for anyone benchmarking against the headline. In 20 of the 24 countries it surveyed, the mean budget exceeded the median, which indicates that a small group of organizations with substantial budgets pulls the average up while the typical employer budgets more conservatively. If your finance partner cites 3.6% as "the market," the median employer may be sitting a little below it.

That gap changes the conversation with leadership. A flat market average is easy to defend. A market average inflated by a few big spenders is harder to apply to a mid-sized company that has to hold onto the same people those big spenders are trying to hire, which is why the same surveys show employers leaning on tools other than the annual pool.

Off-Cycle Pay and Automation Are Quietly Doing the Work

Marsh found 64% of employers now make off-cycle pay adjustments and 70% use some form of compensation automation. When the annual increase cannot move, targeted fixes become the pressure valve, and software is how teams manage the volume. AI is spreading into the work too: 53% use it for market pricing and benchmarking, 50% for salary recommendations, and 49% for job matching and leveling. Yet only 1% of the surveyed employers say they are at an advanced stage of AI transformation.

Put those numbers together and a pattern appears. Pay decisions are getting faster, more individualized, and more automated, while the organizational maturity to govern them is still early. John Legg of Marsh told HR Brew, "This cannot be, 'put it on autopilot, let it run.'" His point was about differentiated awards: performance ratings alone are not enough, and the other factors that justify a larger raise have to be considered and documented. A model that recommends a number cannot explain it to the employee who receives it.

HR Brew also cited Lightcast data showing AI talent commanding a 28% compensation premium. A fixed pool plus a premium for scarce skills is how pay compression and perceived unfairness grow, which is exactly the problem our earlier look at AI skills premiums going to new hires described from the employee side.

• Benchmark against the median, not just the mean. Ask your survey vendor for medians and sector cuts before you tell leadership what "the market" is doing.
• Write down the differentiation rule. If top performers, critical skills, and compression fixes all get more than the average, decide in advance what the criteria are and who approves exceptions.
• Track off-cycle adjustments as a program. Log reason, amount, and recipient so a rolling stream of fixes can be audited for equity rather than discovered in the next pay analysis.
• Treat AI pay recommendations as drafts. Require a human reviewer who can articulate the reasoning behind any number the tool suggests.
• Prepare managers before the letters go out. A smaller pool means more difficult conversations, and managers need talking points that explain the why, not only the what.
• Measure the result. Compare retention and promotion outcomes for differentiated versus average raises after the cycle, so next year's plan rests on evidence instead of habit.

A budget that stays at 3.5% for four years stops being a planning variable and becomes a constraint. The employers who do well inside it will not be the ones who found more money. They will be the ones who can show, person by person, why the money went where it did.