Geographic pay differentials: how to set location-based pay for a remote team
A geographic pay differential adjusts salary based on where an employee works, because the same role costs a different amount in different labor markets. Most companies apply it as a multiplier against a national band: 1.15 times the midpoint in the Bay Area, 1.00 in a typical metro, 0.90 in a lower-cost market. The decision that matters is not the multiplier. It is whether you price off cost of labor or cost of living, and those are not the same thing.
This became a live question for companies that never planned to be distributed. You hire one engineer in Denver, one in Austin, one in Raleigh, and suddenly you own a multi-market compensation problem with no comp analyst. Here is how the zones work, where the numbers come from, and what tends to go wrong.
Cost of labor, not cost of living
This is the single most consequential choice on the page, and it is where most small companies go wrong.
- Cost of living measures what it costs an employee to live in a place: rent, groceries, gas. It describes your employee's spending.
- Cost of labor measures what employers in that place actually pay for the role. It describes your competition.
You are not reimbursing someone's rent. You are trying to win a candidate against the other companies bidding for them, and those companies pay the local market rate. So the correct input is cost of labor.
The two diverge more than people expect. Plenty of mid-sized metros have modest housing costs and a thin supply of senior specialists, so the cost of labor for a niche role sits well above what a cost-of-living index would suggest. Price that role off a cost-of-living calculator and your offer loses, and you will never learn why, because candidates do not send you a reason. The reverse happens too: expensive cities with deep talent pools can price certain roles lower than their rent index implies.
Cost-of-living indexes are easy to find and free, which is exactly why they get used. Cost of labor takes actual wage data by occupation and metro. The US Bureau of Labor Statistics publishes it through the Occupational Employment and Wage Statistics program, broken out by metropolitan area, and it is the most defensible free starting point in the country.
How pay zones work
The standard structure: build one national band per role and level, then define a small number of zones, each with a multiplier applied to that band. A zone 0.90 multiplier on a $100,000 national midpoint gives a $90,000 local midpoint, and the minimum and maximum move with it.
| Model | How it works | Best for | The cost |
|---|---|---|---|
| National single rate | One band per role, same everywhere in the US. | Small teams, senior-heavy orgs, anyone who hires nationally and values simplicity. | You overpay in cheap markets and still lose candidates in expensive ones. |
| Zone multipliers | National band times a zone factor, typically 2 to 4 zones. | Most companies under 200 people with people in several metros. | You maintain the zone map and defend the boundaries. |
| Metro-specific bands | A separate band per market, built from local data. | Large employers with real density in each market. | Heavy upkeep. Overkill below a few hundred people. |
| Anchor market | Price everyone against one chosen reference metro. | Companies whose talent competition is concentrated in one hub. | Expensive if the anchor is a high-cost market and most staff are not there. |
Zones only earn their keep when the gap between them is big enough to matter. A zone that differs from its neighbor by 4 percent is administration without benefit, and it invites arguments about which side of a line someone's suburb falls on. Aim for gaps of roughly 10 percent or more, and keep the count low: two to four zones covers almost every company under 200 people.
Building your zones, step by step
- List where your people actually are. Not where you might hire one day. Real current locations, plus the markets you are actively recruiting in this year.
- Pull cost-of-labor data per metro. For each significant role, get the local wage percentiles from BLS OES or a benchmarking source, and record the date. You are looking for the same occupation across your metros.
- Index each metro against your national band. Divide the local midpoint by the national midpoint. That quotient is the raw differential for that market.
- Cluster into zones. Group metros whose raw differentials land close together, and round to a clean multiplier. Do not create a zone for every city. Three tidy zones beat nine precise ones nobody can administer.
- Write down the rules. Which zone a new hire lands in, what happens when someone moves, how often you refresh. If you skip this step, every relocation becomes a one-off negotiation and your differentials stop being a system.
- Refresh annually. Differentials compress and stretch as markets move. A zone map built in 2022 does not describe 2026.
What happens when someone moves
This is the question that will land on your desk, so decide it before it does. Three defensible options:
- Hold the salary. The person keeps their number and sits high in their new zone's band. Simple, generous, and it means two people doing the same job in the same city are paid differently because of where they used to live.
- Adjust at the next review. No immediate cut. Future increases are smaller or paused until the salary fits the new zone. Least disruptive, and it takes a while.
- Adjust on the move. Salary re-prices to the new zone at the relocation date. Internally consistent, and a real pay cut for someone who did nothing wrong. Expect it to be unpopular and to occasionally cost you the person.
None of these is correct in the abstract. What is not defensible is deciding case by case, because that is how you end up applying a cut to one person and not another with no written reason, which is the fact pattern that turns a comp decision into a legal problem.
Is location-based pay legal?
Yes, in the US. Paying different salaries by work location is lawful when the differential reflects the local labor market and you apply it consistently. Under the federal Equal Pay Act, a pay difference between employees doing substantially equal work has to rest on seniority, merit, quantity or quality of production, or a factor other than sex. A documented geographic differential is a factor other than sex. An undocumented one, applied to some people and not others, is just a pay gap with a story attached.
Several state equal pay laws set a higher bar than the federal one and require the factor to be job related and consistent with business necessity. That is not a reason to avoid differentials. It is a reason to be able to produce the zone map, the data source, the date, and the rule you applied. In practice, the companies that get in trouble are not the ones with aggressive multipliers. They are the ones that cannot explain why two people in the same zone are paid differently.
Geography also collides with posting rules. If you advertise a remote role that someone in Colorado, California, New York, Washington or Illinois could perform, you are likely covered by that state's disclosure law and you have to post a good faith range. With zones, the honest approach is either to post the full span across the zones you would hire in, and say so, or to post the range for a named zone. What you cannot do is post a $60,000 to $300,000 range and call the width a geography problem. Our pay transparency guide breaks the rules down state by state, and job posting salary ranges covers how to turn a band into a range you can publish.
Three ways this goes wrong
Zones built from cost-of-living indexes. Covered above, and it is the most common one. You will lose candidates in markets where housing is cheap and talent is scarce.
Differentials that never get refreshed. The gap between expensive and cheap metros is not a constant. It compressed noticeably when remote hiring went mainstream, as employers everywhere started competing for the same people. A zone map you have not touched in three years is describing a market that no longer exists.
Precision nobody can administer. Nine zones and a spreadsheet with a lookup by county is a system that works until the person who built it leaves. It also creates an endless supply of boundary disputes, because someone always lives forty minutes outside the metro you priced. Fewer zones, applied consistently, beat more zones applied approximately. If you are already stretching a small team across sourcing and screening candidates in five markets at once, the last thing the process needs is a comp model that requires a specialist to operate.
Frequently asked questions
Should you pay remote employees based on location?
It depends on where you hire. Location-based pay lets you compete in expensive markets without raising pay everywhere, and most companies under 200 people use it. National pay is simpler and easier to explain, but costs more in cheap markets and underpays in expensive ones. Either works. Applying neither consistently does not.
How do you calculate a geographic pay differential?
Divide the local market midpoint for the role by your national midpoint. If a role pays a $115,000 median in the target metro and your national band midpoint is $100,000, the raw differential is 1.15. Round to a clean zone multiplier, and use cost-of-labor wage data by occupation and metro rather than a cost-of-living index.
How many pay zones should a company have?
Two to four for most companies under 200 employees. Zones only justify their upkeep when the gap between them is around 10 percent or more and you have enough people in each to matter. More zones buy precision you will not administer well and boundary arguments you do not need.
Do you have to cut pay if an employee moves to a cheaper city?
No. You can hold the salary, adjust at the next review, or adjust on the move. Many companies hold, because a pay cut for someone who did nothing wrong is a bad trade for the savings. Whichever you choose, put it in the relocation policy before anyone asks, and apply it to everyone the same way.
Where can I get geographic pay data for free?
The US Bureau of Labor Statistics publishes Occupational Employment and Wage Statistics by metropolitan area, with wage percentiles per occupation. It is free, public, and the most defensible no-cost starting point. Its limits are lag and coarse occupation categories, so it works better for established roles than for fast-moving specialties.
Where to start
If you have people in more than one market and no zone map, do the cheap version first: pull local wage data for your three or four biggest roles in each metro you employ people in, index them against your national bands, and see whether the spread is even big enough to bother with. Sometimes it is 5 percent and the answer is to pay one national rate and move on. When the spread is real, build the fewest zones that describe it, write down the rules, and refresh next year.
The prerequisite is a national band worth indexing against. Our guide to how to create salary bands covers building one, compa ratio explains how to measure people against it once zones are applied, and compensation philosophy is where the geography rule should be written down.