Compensation and benefits trends for the future of work

Compensation and benefits are being rebuilt around three forces: pay transparency written into law, benefit costs climbing faster than salaries, and pay attached to proven skills instead of job titles. US salary increase budgets have settled at 3.4% for 2026, while employers expect health benefit costs per employee to rise 6.5%. The money is moving, just not into base pay.

That gap is the story of the next pay cycle. A team can hold raises flat, watch the cost of employing each person rise anyway, and still lose a candidate to an employer whose job ad simply showed a number. Most compensation plans in use right now were built for a decade when none of those three things were true.

TL;DR

• Raises are flat and benefits are not. Salary budgets sit at 3.4% while health benefit cost per employee is set to rise 6.5%, the steepest jump since 2010.
• Benefits are already about a third of what an employee costs, so treating pay as "salary plus some extras" leaves out nearly a third of what the employer actually spends.
• Pay transparency stopped being a policy choice. The EU deadline has passed and US state rules keep widening, so ranges get published whether a company is ready or not.
• Publishing a range exposes internal pay gaps before it attracts anyone. Audit first, publish second.
• Skills, not titles, are becoming the unit of pay, which only works if a company can actually measure the skill it claims to be paying for.

What is the future of compensation?

The future of compensation is a package with three moving parts instead of one: a published pay range tied to a defined role, a benefits mix the employee chooses from, and progression linked to demonstrated skills rather than time served or a title change. Base salary stays the anchor. It stops being the whole conversation.

The numbers already say so. In private industry, benefit costs averaged $14.01 per hour worked and made up 30.1% of what employers spend on compensation, against $32.60 per hour for wages and salaries, according to the US Bureau of Labor Statistics employer cost data for March 2026. Nearly a third of the cost of employing someone never appears on their payslip as salary.

Most pay conversations ignore that third entirely. A candidate compares two offers on base salary, picks the higher one, and never learns that the other package carried better cover, a real learning budget, and an employer pension contribution worth thousands. That is not a candidate failure. It is an employer communication failure, and it is getting more expensive every year the benefits share grows.

The shift worth understanding is a shift in what the word compensation covers. It used to mean the number in the contract. It now means the total cost of keeping someone, most of which is negotiated once and then never explained again.

Top trends in compensation and benefits for 2026

Five changes are doing most of the work this cycle. They are connected: flat salary budgets push employers toward benefits, rising benefit costs push them toward personalization, and transparency law forces both into the open.

Understanding current trends in employee compensation

Start with the salary line, because it sets everything else. US salary increase budgets for 2026 are expected to stay at 3.4%, the same as the actual figure for 2025, based on a survey of 1,876 US organizations by WTW on salary budget planning. Within that, 62% of employers made no change to their projected budgets, 21% cut them, and only 6% raised them.

Now the other side. Total health benefit cost per employee is expected to rise 6.5% on average in 2026, the biggest increase since 2010, in an analysis of more than 1,700 US employers by Mercer on health benefit costs. The detail underneath that headline matters more than the headline: 59% of those employers are already making cost-cutting changes to their plans, and without those changes the average increase would have been close to 9%.

Read those two figures together and the squeeze is obvious. The cost of employing a person is rising roughly twice as fast as the raise that person receives, and a good share of the difference is being absorbed by quietly trimming the plan. Employees experience this as a flat raise and a worse benefits year at the same time, which reads as a pay cut even when total spend per head went up.

A pattern we keep seeing in hiring conversations is that nobody ever explains the trade. The plan change lands in an email from a broker in November, the raise lands in March, and no one connects the two. The cheapest fix in compensation right now is not more money. It is a one-page total reward statement per employee, showing base pay, employer benefit contribution, and what the plan change actually cost or saved them. It costs a payroll export and an afternoon.

The caveat: a total reward statement backfires if the underlying package is genuinely weak. Showing someone a big number made mostly of statutory minimums invites the comparison a company would rather avoid. Fix the package first if it is thin, then tell the story.

Which companies are shaping the future of pay?

The companies shaping the future of pay are not the famous employers whose perks get written up. Pay practice is being set by three groups: regulators writing transparency into law, employers who publish ranges before they are forced to, and the firms repricing scarce skills fast enough that title-based bands cannot keep up. Most organizations end up following one of the three rather than setting the pace themselves.

The regulators moved first and hardest. Directive (EU) 2023/970 on pay transparency and equal pay required member states to have national law in force by 7 June 2026. It obliges employers to give applicants pay information before an interview, bans asking candidates about their pay history, gives workers the right to information on average pay levels for the same work broken down by sex, and adds gender pay gap reporting duties for larger employers. The European Commission on equal pay sets out the wider enforcement picture.

Here is the part most trend articles miss. That deadline has already passed, and national implementation is uneven, so a company hiring across several European countries is now working against a patchwork rather than one rule. The practical consequence is that a single European job ad template no longer works. Pay disclosure has to be set per country, and the safest default for a multi-country employer is to write to the strictest rule it is exposed to rather than the weakest.

The second group is employers who publish ranges voluntarily, usually because they hire remotely and got tired of losing candidates at the offer stage. They shape pay in a quieter way: once a competitor publishes a range for a role, every candidate for that role has a reference point, and every recruiter negotiating without one is negotiating blind.

The third group is the one that will matter longest. Employers expect 39% of workers' core skills to change by 2030, according to the World Economic Forum skills outlook, drawn from a survey of over 1,000 employers. The same research projects 170 million new roles created and 92 million displaced by 2030, a net gain of 78 million jobs, in the Future of Jobs findings. A pay band written against a job title assumes the job holds still. Nearly four in ten of its skills will not.

So the honest answer to who is shaping pay: law sets the floor, early publishers set the market reference, and skill scarcity sets the ceiling. A company that watches only competitors' salary surveys is reading the slowest of the three signals.

Which future of work benefits matter most?

The benefits that move decisions are the ones that solve a problem the employee currently pays for out of pocket or out of their own time. Health cover, genuine schedule flexibility, a learning budget with time attached, caregiving support, and financial help such as an employer pension contribution or debt support. Perks that decorate an office do not survive a remote workforce.

Ranking them generically is where most advice falls down, because the ranking is a function of who works there. A workforce with a median age of 26 and one with a median age of 44 want almost opposite things from the same budget. The first will take the learning budget and the flexible hours. The second will take the family cover and the pension contribution every time.

This is the actual case for personalization, and it is a cost argument before it is a wellbeing one. Every dollar spent on a benefit nobody uses is a dollar that could have gone into base pay, where at least it is visible. With benefits already running at 30.1% of employer compensation cost, a plan with poor take-up is not a small inefficiency.

Three moves worth making before the next renewal:

1. Pull utilization data per benefit. As a rule of thumb, anything under about 20% take-up is a candidate for cutting or replacing, unless it is catastrophic cover that exists precisely because it is rarely used.
2. Ask, but ask about tradeoffs rather than wishes. "Which two of these would you keep if we could only fund two" produces usable answers. "What benefits would you like" produces a wish list.
3. Convert the savings into either base pay or one benefit people actually asked for, and say explicitly that is what happened. Silent reallocation reads as a cut.

One tradeoff to name honestly: choice has an admin cost, and a flexible plan run badly is worse than a simple plan run well. A ten-person team does not need a benefits marketplace. It needs one good plan and a clear explanation of it. Complexity earns its place somewhere north of about fifty employees, and even then only if someone owns it.

How should pay transparency change your job ads?

Publishing a range changes hiring in one immediate way: the negotiation now starts from a number the company chose, in public, instead of a number a candidate guesses. That is an advantage, but only for employers who did the internal work first. For everyone else the range is a disclosure of their own inconsistency.

Run it in this order:

1. Audit before you publish. Pull current pay for everyone in the role, sorted. The outliers will be obvious and at least one of them will be uncomfortable.
2. Fix what the audit finds, or budget to fix it. A published range that sits above what several current employees earn is a conversation that will happen whether it is planned or not.
3. Set the range against the role, not the person. Define the level, its responsibilities, and the skills it requires, then price that.
4. Publish a real range. A band spanning 60,000 to 160,000 tells a candidate nothing and signals that the level was never defined.
5. Say what moves someone through it. A range without progression criteria creates the exact question it was supposed to answer.
6. Check the rule per country and per state, and keep the strictest version as the template default.

The part teams underestimate is step two. Transparency is not primarily a recruiting change, it is an internal equity change that happens to arrive through a job ad. Companies that treat it as a job-ad formatting task get the internal conversation anyway, just without preparation. If the compliance side is unfamiliar, start with the compensation and benefits law obligations a team should already be tracking.

How do you pay a team spread across countries?

There are two honest models and a company has to choose one on purpose. Location-based pay sets the band by where a person lives. Location-independent pay sets one band per role and pays it everywhere. Most distributed teams land between the two, which only works when the rule is written down and applied the same way every time.

Location-based pay is the cheaper option and the easier one to defend with local market data. It also produces the conversation nobody enjoys: two people doing identical work, reviewed by the same manager, paid differently because one of them lives somewhere expensive. It punishes anyone who relocates, and in a remote team people do relocate.

Location-independent pay solves that and creates a different problem. One global band is simple to explain and a genuine recruiting signal, but in lower-cost markets it means paying well above what local competitors pay, which is fine until the budget meets the hiring plan. Companies that adopt it usually cap how many roles it applies to.

The practical middle is a small number of geographic tiers rather than a rate for every city. Three tiers is usually enough. More than that and nobody can explain the model, which defeats the purpose of having one.

Currency and statutory benefits are where cross-border pay quietly goes wrong. Employer social contributions, mandated leave, notice periods and healthcare obligations differ by country, so paying two people the same salary does not mean giving them the same compensation. Since roughly a third of employer compensation cost sits outside wages in the first place, a package compared on salary alone can be out by a wide margin in either direction. Compare total employer cost per hire, not the number in the contract.

Transparency law tightens this further. A company advertising one role across several European countries now has to satisfy the strictest disclosure rule it touches, and the national rules do not match. The workable default is to write the ad to the strictest exposure and keep a per-country note for the rest, rather than maintaining a separate template per market and hoping the right one gets used.

Two rules worth setting before any of this is published. First, name the tier in the job ad, so a candidate knows which band they are being measured against instead of discovering it at offer stage. Second, do not cut someone's pay retroactively when they move to a cheaper location. Grandfather existing employees and apply the new rule to new hires. A retroactive cut saves a modest amount of payroll and costs an amount of trust that is much harder to rebuild, and the people most likely to leave over it are the ones who were mobile enough to move in the first place.

The caveat here is scale. A team of twelve with people in three countries does not need a tiering model, it needs consistency and a written note explaining how each offer was set. Formal geographic bands start earning their keep somewhere around fifty to a hundred people, once no single person can hold every pay decision in their head.

How do you pay for skills instead of job titles?

Skills-based pay means the band is set by the capability a role requires and the person is placed in it by demonstrated skill, not by tenure or by a title someone negotiated three years ago. It is the clearest fix for a market where nearly 40% of core skills are expected to turn over by 2030. It is also the trend most often described and least often implemented, for one reason: paying for a skill requires measuring it, and most companies cannot.

That measurement gap is what the Testlify Competency-to-Evidence Matrix is for. It maps a role to the competencies that actually matter, connects each competency to measurable evidence through assessments, work simulations, structured interviews, references and reviewer feedback, and gives every competency a weight and a benchmark. The output is a defensible answer to "what is this person paid for", built from evidence rather than from a job title inherited from an org chart.

Applied to pay, it works like this. Picture a hypothetical 120-person marketing agency with eight people carrying some version of an account role, all paid within a band nobody can quite justify. The matrix asks what an account role at this agency genuinely requires: client communication, budget control, analytics reading, and now the ability to direct AI tools without shipping their mistakes to a client. Each of those gets an evidence source and a weight. Assess against it once, and the band stops being a negotiation and starts being a reading.

The uncomfortable part is that the reading will not match the current pay order. It rarely does. That is the point, and it is also why this work needs a budget line for corrections before it starts, not after. Running the assessment and then declining to act on it is worse than never running it, because now the gap is documented.

Pro tip: pilot skills-based pay on one role family before touching the whole company. One family of six to ten people is enough to expose whether the competencies were defined well, and small enough that correcting two salaries is a rounding error rather than a board conversation. Companies that roll it out everywhere at once usually discover their competency definitions were vague only after they have repriced two hundred people against them.

Two caveats worth stating. Skills-based pay is a poor fit for roles where output is genuinely uniform and regulated, where a rate card already does the job better. And it needs re-measuring, because a skill priced as scarce in 2026 may be ordinary by 2029. Build the review cadence in or the new bands go stale exactly the way the old ones did.

Teams redesigning bands from scratch usually need the vocabulary first. Two starting points: the main types of pay and benefits, and the case for why compensation and benefits drive retention.

What is the fastest way to price a skill?

Measure it on one role family, against the same evidence for every person in it, before the next review cycle. That is the whole answer. Every trend on this page runs into the same requirement, which is knowing what a person can actually do, consistently enough to attach a number to it.

Testlify measures role-relevant skills with validated assessments, work simulations, coding and cognitive tests, and structured reviewer scoring, so a pay band can be set against evidence rather than a title. Browse the skills test library to see how a role maps to measurable competencies, or book a live walkthrough and bring one role family you already suspect is mispriced.

Key takeaways

• The raise is not where the money went. Salary budgets sit at 3.4% while health benefit cost per employee rises 6.5%. Total spend per head is growing faster than pay, so retention plans built only around the annual increase are arguing about the smaller number.
• A third of compensation is invisible. Benefits are 30.1% of employer compensation cost in private industry. If a company never states that figure per employee, it is spending nearly a third of its people budget on something candidates do not weigh in an offer comparison.
• Transparency is an internal audit wearing a recruiting costume. The EU deadline of 7 June 2026 has passed and US state rules keep widening. The hard part is not writing the range, it is what the range reveals about current pay, so audit before publishing rather than after.
• Personalization is a cost lever first. Spend on a benefit with low take-up is spend that could have been base pay. Pull utilization data before the next renewal, cut what nobody uses, and say out loud where the money went.
• Skills go stale faster than titles do. With 39% of core skills expected to change by 2030, any band written against a job title is depreciating from the day it is signed off. Price the capability and set a review cadence.
• Measurement is the bottleneck, not philosophy. Most teams agree with skills-based pay and stall at proving the skill. Fix the evidence problem on one role family first, with a correction budget approved in advance.
• Explain the package or lose the credit for it. A one-page total reward statement is the cheapest retention move most employers have available this year, because it needs a payroll export rather than a budget approval.

FAQs