"We Pay Local Rates" Means the Company Captured Your Arbitrage

Two engineers do identical remote work. One in San Francisco, one in Bengaluru. Under location-based pay, the second can be paid two to three times less, purely on home address. This is the visible edge of a much wider pattern, where the same job pays up to 4x more depending on where you sit. The work is the same. The output ships to the same customers at the same price. The gap doesn't vanish. The policy just decides who keeps it: you, or the company's margin.

That's the part nobody says out loud. "We pay local rates" sounds like fairness. It's a cost decision dressed as one. The value of your output prices globally. Your salary prices locally. The spread between those two numbers is real money, and a localization policy is the mechanism that routes it away from your paycheck. Once you see the structure, you stop arguing about whether it's fair and start arguing about the number.

What is location-based pay, really?

It's a formula where your salary is a function of where you live, and the company's revenue is a function of what you ship. Those are two different inputs. The policy treats them as one.

Look at how the most transparent companies actually compute it. GitLab's formula starts from the San Francisco 50th-percentile benchmark, then multiplies by a location-derived factor, a level factor, an experience factor, and a country factor. Buffer runs the same shape: SF benchmark times a cost-of-living multiplier, where a high-cost city gets 100%, an average one 85%, and a low-cost one 75%.

Notice what's in the formula and what isn't. Where you live is in it. What your work is worth to the customer is not. The SF benchmark is the ceiling, and your address is the discount applied to it. The company isn't measuring your value differently in Bengaluru. It's measuring your value the same and paying you less for it.

Where does the gap actually go?

Walk the arithmetic. A US software engineer averages around $140,000 a year; the same role in India averages around $45,000. A company hiring remotely in India instead of the US saves roughly 68% on equivalent-role payroll.

The output is identical. It ships to the same customers, gets sold at the same price, generates the same revenue. So the $95,000 difference between $45K and $140K doesn't evaporate. It flows somewhere. It flows up, into the company's margin.

The structure is the same wherever location pay is used. A company's product sells globally at the same price regardless of who built it, while the salary it pays for that work is discounted by the builder's address. The policy is the difference between the price the company collects for the work and the price it pays for it.

Think of it like commodity grain. Wheat grown in Kansas and wheat grown in France sell at the same global price. The farmer's input costs vary enormously by location. The output price doesn't. Remote work has the same structure: your code, your analysis, your writing prices globally; your salary prices locally. The company is the grain elevator. It buys at local cost and sells at global price.

Is the company "capturing" it, or did I agree to it?

Both, usually. And this is where you have to be honest with yourself.

The macro evidence says workers have already given up real money for flexibility. An NBER working paper found the shift to remote work produced a cumulative wage-growth moderation of 2.0 percentage points over two years, and that the amenity value of remote work lowered labor's share of national income by 1.1 percentage points. In plain terms: people accepted slower raises in exchange for working from home, and employers pocketed the difference. It tracks with what the broader data shows about remote work paying the same while promoting you slower. That's the macro version of the same trade you're making in a single offer.

Workers do put a real cash value on flexibility, which is why the NBER trade above held across millions of jobs. So if you willingly take $92K remote over $100K in-office, that's a rational choice. The company didn't extract it from you. You traded it.

The honest reframe: the policy captures the arbitrage when you let the default stand. The in-office SF salary was never automatically your baseline. But neither is the localized lowball. The number is set in the negotiation, not by the map. If you accept the framing without naming the gap, you're the one making the choice for them.

Cost of labor or cost of living, which is the company using?

This is the single most important thing to read in an offer, and most people miss it. There are two different benchmarks hiding behind the same phrase "local rates," and only one of them is defensible.

Cost of living is what it costs you to survive in your city: rent, groceries, transport. Cost of labor is what it costs the company to hire someone with your skills in that market. They diverge, often sharply, because what locals pay for housing and what employers pay for talent in the same city are set by different forces. ERI's own data on how companies set geographic pay shows the serious ones lean on labor-market evidence, not your grocery bill: in ERI's survey of geographic pay practices, only about 4% of organizations rely on cost-of-living data alone, while most blend salary-survey and geographic pay-differential data.

PostHog states the honest version explicitly: "Location factors are based on cost of market, not cost of living." Their floor is 0.6 globally, 0.8 in the US.

What they benchmark to	What it means for you	Defensible?
Cost of labor (local market for your skill)	They pay what your role commands in your city	Yes. Argue you're above that local market.
Cost of living (your rent and groceries)	They pay what they think you need to survive	No. Your survival cost is not your value.
SF benchmark minus an address discount	They admit your value, then dock it for geography	This is the one to push on hardest.

If the offer is benchmarked to your local market, the exploitation framing weakens: they're paying the going rate where you are. If it's benchmarked to your living costs, or to SF-minus-a-discount, the gap is pure spread, and that's negotiable. The flip side is worth holding too, because a high-cost-city salary is often smaller than your hometown offer once rent eats the difference, so a localized number is not automatically the worse deal.

How do I negotiate against localization?

You name the gap and argue value over address. Three moves, in order.

First, benchmark the role at market, not your geography. Pull the actual US-remote or global-remote number from Levels.fyi, Glassdoor, and live job postings for the same title and level. You need a defensible figure before you open your mouth.

Second, reframe from your address to your output. The company's formula starts at the SF benchmark and discounts it. Your counter starts at the value you produce and asks why the discount is so steep. The NBER finding is your edge here: companies already know the spread exists. You're not accusing them of anything. You're naming a number they've already modeled.

Third, if base won't move, move the rest. Equity, a signing bonus, an accelerated review cycle, a higher band on the next promotion. The base might be policy-locked; the total package rarely is.

Weak: A senior engineer in Warsaw takes $65K from a US startup because "that's good for Poland." The same role posts on Levels.fyi at $130K for US-remote. The company captures a 2x spread on identical output, and the engineer never knew the ceiling existed.

Strong: Same engineer benchmarks the US-remote comp, cites it in writing, and counters at $95K. "I understand you localize. Here's the global-remote market rate for this role. I'm asking for a number between your offer and that, because the output you're buying is the same." The company still saves $35K. The engineer doubles the local rate. Both sides win more than the default would have given either.

You won't win every time. But you will never win if you don't play.

When should I just take the localized offer?

When the absolute number changes your life more than the spread offends you. This is the trade-off, named plainly.

For someone early in their career in a lower-cost market, a foreign company paying two to three times the local rate, even after a location discount, is often a large absolute gain over any local employer. Negotiating hard against localization can kill that offer for a principle that doesn't pay your rent. The spread is real, and so is the fact that the localized number might still be the best money available to you right now.

There's also a genuine cost on the company's side worth acknowledging. Some location adjustment reflects real tax, compliance, and employer-of-record fees, not pure margin. And location-agnostic pay has its own distortion. Basecamp pays every US employee in the same role and level the same, with target pay set at the top 5% of its Chicago market rates, and reports an average tenure of five years in an industry where companies are lucky to keep people two. Jason Fried's line is "Pay for people's skills, not their address." Clean. But pay a Bengaluru engineer full SF rates and they earn 3x their local peers, which warps the local market and breeds resentment. Fair to the individual and fair to the market can pull in opposite directions.

The point isn't that localization is always theft. It's that you should know which number you're being paid against, and decide with your eyes open.

What to do now

Before you accept any location-adjusted offer, run three checks.

1. Find the ceiling. Pull the US-remote or global-remote rate for your exact role and level. Levels.fyi, Glassdoor, live postings. Write the number down.
2. Find the floor. What does your role pay at market in your own city? That's the honest local-labor benchmark, the one a defensible policy uses.
3. Name the gap. The distance between those two is the arbitrage. Decide consciously whether you negotiate it, take the offer anyway, or walk. All three are valid. Accepting by default is the only losing move.

The spread exists in every remote hire. The only open question is which side of it you're on, and whether you decided or just received.

Want to know your actual market number before you counter, by role, level, and location? Message Praxy on WhatsApp. I'll help you find the ceiling, the floor, and the line you should hold.