Merit vs Cost-of-Living Raises for Veterinary Staff

Start by naming what each kind of raise responds to, because the blending is what can make a raise cycle feel arbitrary to the people it pays.

The three tools, side by side:

Each fails differently too.

A cost-of-living-only practice pays its weakest and strongest performers the same, and nothing in it tracks pay to performance.

A merit-only practice asks one number to do two jobs, and a technician whose raise comes in below the index move — 3.4% in the latest CPI-U reading — has lost ground to the index without a word being said.

Equal-pay law sits under all three.

The federal Equal Pay Act bars paying employees of one sex less than the opposite sex in the same establishment for equal work requiring equal skill, effort and responsibility under similar working conditions — except under a seniority system, a merit system, a production-based system, or a differential based on a factor other than sex.

A documented merit system is therefore one of the law's named justifications for paying two employees doing equal work differently.

Colorado keeps a different list: its Equal Pay for Equal Work Act bars paying an employee of one sex (or sex combined with another protected status) less than another sex for substantially similar work, and such a differential is lawful only if based on seniority, merit, production, geographic location, job-related education, training or experience, or necessary travel — with each factor applied reasonably and accounting for the entire differential.

Your matrix and review records are the evidence that the factor is real.

The reference number BLS publishes for cost of living is the Consumer Price Index for All Urban Consumers (CPI-U), and the 12-month change is the figure cost-of-living formulas hang on.

As of this writing, the latest reading is the all-items CPI-U rise of 3.4% over the 12 months ending August 2026, before seasonal adjustment, in the release BLS published on September 11, 2026.

Excluding food and energy, it rose 2.4% over the same period.

BLS scheduled the September 2026 CPI release for Wednesday, October 14, 2026 — check for a newer figure before you set this year's number, because your cycle will use a different month.

Three decisions turn that reading into a policy.

Pick the measure: the all-items figure (3.4% for that window) or the ex-food-and-energy figure (2.4%) — both are BLS-published 12-month changes for the same period, and switching between them year to year makes your own history meaningless.

Pick the window: a 12-month change ending in a month whose figure is published before you announce.

Pick the response: match the index move in full, match it in part, or treat it as the floor under the merit layer — that choice is pure budget, and our research found no veterinary-specific benchmark that sets it for you.

Two cautions.

The CPI-U is a national index for all urban consumers: it does not measure prices in your town or on any employee's own bills, so treat it as the shared reference for the conversation, not a read on anyone's household.

And resist announcing an automatic formula — "we match CPI every year" — unless you can fund it in a year when the index moves the way it just did, because staff will reasonably hold you to what you announce.

Work the arithmetic in one direction: pool first, then split.

Take the payroll you intend to raise, decide the percentage you can fund, and treat the result as one pool with three layers inside it — the cost-of-living floor for everyone, the merit layer that differentiates, and the market fixes that re-price roles which have drifted.

Our research found no sourced veterinary-specific raise percentage to copy — any figure presented as the typical vet raise has nothing sourced behind it.

The percentage comes out of your margins, your payroll and your own turnover picture.

Place people before you split anything.

The federal figures are BLS OEWS, whose May 2025 national wages read:

Read the table with its caveats.

The middle column is the median — quote it as a median, never as an average.

The assistant series combines veterinary assistants with laboratory animal caretakers, so it is not a clean read on clinic assistants alone; the receptionist series covers receptionists in every industry, not just veterinary; and the animal caretaker series is mostly boarding, grooming, kennel and zoo workers, with only a minority in clinic-type settings.

Lean harder on the technician and veterinarian rows, and treat the other three as context.

Series discipline matters as much as placement.

The federal figures above are BLS OEWS survey wages from May 2025; VeterinaryHires' own posting data is the other series you will see.

On VeterinaryHires, employers hiring veterinary technicians posted a median of $24.50 an hour — an employer median across 354 employers and 670 listings stating hourly pay, as of October 6, 2026 (the Q4 2026 release).

That figure is advertised pay on live postings, not what anyone earned — it excludes bonuses and production pay unless a posting folds them into the range — and the two series answer different questions.

Pick the one that matches the decision, and label whichever you use the same way every year.

The placement rule: anyone below the 25th percentile for their role is a market-fix candidate before merit enters the picture, anyone above the 75th probably needs a conversation about rungs and scope rather than a bigger percentage, and the middle is where the merit layer does the work.

For current salary detail on the two roles you will re-price most often, see the veterinarian salary guide and the vet tech salary guide.

A merit matrix is a small grid that converts a review outcome into a raise percentage before anyone talks about money.

Review ratings run down one axis, position in the band runs across the other, and the cell where they meet holds the percentage that person receives from the merit layer.

Fill the cells with percentages you choose — our research found no sourced figure that pins them — and the matrix does its real work afterward, by making the same rating mean the same money for everyone.

Hold two rules when you build it.

Decide the cells before the reviews are written, not after: a matrix filled in afterward is a justification, not a system.

And keep the review honest, because a matrix converts whatever the ratings measure into pay differences — if the ratings are interchangeable, the matrix is converting noise, and a merit-system defense is only as good as what the ratings actually measure.

The documentation is the defense.

Three of the federal Equal Pay Act's named justifications are systems — a seniority system, a merit system, a production-based system; the fourth is a differential based on a factor other than sex.

Your matrix and your review records are what show the merit system is real rather than asserted.

In Colorado the requirement runs further: there, each factor behind a sex-based differential for substantially similar work — merit included — must be applied reasonably and account for the entire differential, which in practice means your records have to explain the whole gap, not part of it.

One compression guardrail.

When a market fix lifts a newer hire above a longer-tenured employee doing the same work, the temptation is to shrink the higher salary back down.

If the two are of opposite sexes doing equal work, check the gap against the federal Equal Pay Act's named exceptions first: under that Act, an employer that is paying an unlawful wage differential may not cure it by reducing the wage rate of any employee.

In that case the correction runs upward: the fix for an unlawful sex-based gap is a raise for the lower-paid employee, not a cut for the higher-paid one.

And if your reviews are not yet strong enough to carry money, design around it: run the cost-of-living floor and the market fixes this cycle, say plainly that merit money is paused while the review instrument is rebuilt, and fix the reviews first.

A merit layer built on rubber-stamp ratings costs more trust than it buys.

Some pay events do not wait for the annual cycle.

When a technician passes a credentialing exam — the VTNE is the classic case — their qualifications change, and pricing that change only at the next review date tells them the credential does not matter to you.

Handle it as its own event, with its own money, decided when the credential lands.

What the credential lets the technician do legally varies by state, so confirm scope with your state veterinary board before you price the new rung.

What the rung pays is your design decision against your band.

And if the credential should move the person up a level rather than add money to the same level, define the levels first: a tech career ladder gives you the rungs, and the credential raise then prices the new rung instead of inflating the old one.

Budget credential raises outside the annual pool.

If they come out of the merit layer, every credential earned mid-year quietly eats the money your reviews were supposed to distribute, and the matrix stops meaning anything.

A separate credential line keeps both honest.

On timing: if the increase is effective on the date the credential passed but lands in a later paycheck, the catch-up amount is a retroactive pay increase.

IRS Publication 15 (2026) lists retroactive pay increases among supplemental wages — run it through payroll as such, and confirm the withholding mechanics with your payroll provider when you set the effective date, not after the check has gone out.

Each kind of raise carries a different message, and blending them wastes all three.

The cost-of-living message names the index: prices moved by a measured amount, and the floor of the pay scale moved with it.

The merit message names what the review showed and where the person sits in the matrix.

The market message names the role's rate, not the person — it is not a reward, and presenting it as one creates an expectation the next cycle cannot meet.

When one number quietly does three jobs, employees assign it the meaning they prefer — which is why the three messages are worth keeping separate.

Say the mechanics you can defend: which inflation measure you referenced and for which 12 months, where the review bands sat, and the first paycheck the new rate appears in.

People do not need the spreadsheet, but a raise with no stated reason reads as either charity or an error, and neither reading helps you.

Freezes need more communication, not less.

If the pool cannot fund a cycle, say so early, say what caused it, and name the date you will revisit the question — then revisit it when you said you would.

A silent freeze is read as permanent, and the first people it costs you are the ones with the most outside options — while the market benchmarks you will eventually re-price against keep republishing whether your payroll does or not.

For the surrounding employer guides — the role salary guides, the bonus and agreement mechanics, and the rest — the veterinary hiring hub is the index.