The VPN affordability index

One advertised price is 467 different prices once income is the unit

Every VPN quotes a single figure. The figure is global, the customers are not, and the gap between those two facts is not small — it is the largest variable in the whole transaction. A subscription that a Swiss household covers with 7.4 hours of average earnings takes 145 days in Burundi. That is a factor of 467, and it is invisible on every pricing page in the industry, including ours.

The unit here is deliberately blunt. We take the World Bank’s GNI per capita, divide it by 365 to get a day of average income, and ask how many of those days one year of the subscription costs. Days of income travel across currencies without an exchange-rate argument, and they are hard to misread: a number above 30 means more than a month of the average person’s entire output, before food or rent.

The bottom of the table is worth reading beside the top, because it is where almost all VPN marketing is written. In Switzerland, Norway, Luxembourg and the rest of the high-income group, the annual cost of a VPN rounds to a rounding error — well under a day of average earnings, which is why the category can be sold on convenience. The product is identical. The decision is not remotely the same decision.

65 of 172 countries pay more than a week of average income for a year of privacy

The headline extremes are easy to dismiss as outliers. The middle of the distribution is the part that actually reframes the category. The global median across the 172 countries with an income figure is 4.4 days — meaning that for a country sitting exactly in the middle of the world, a VPN subscription costs the best part of a working week’s average output.

That is not a catastrophic number. It is, however, roughly ten times what the same subscription costs the median reader of an English-language VPN review, and it is the reason that “just get a VPN” lands differently depending on where it is read. 20 countries are past a full month of average income. In those places the annual cost of a privacy tool is a genuine budget line, competing with things that are not optional.

• Days of average income: the annual subscription cost divided by GNI per capita, multiplied by 365. The unit the whole page is built on.
• GNI per capita, Atlas method: the World Bank’s income measure, smoothed over three years of exchange rates to stop a currency swing from moving a country up the table for no real reason.
• Held-constant price: one price applied to every country — the median ongoing annual cost across the subscriptions we track, $95.19. It does not vary by country in this index, deliberately.
• Confirmed block: an OONI measurement that positively identified blocking, as distinct from an anomaly, which is a connection that failed for reasons that may or may not be censorship.

Where a platform is confirmed blocked, the tool costs more — by a margin too small to lean on

The obvious hypothesis is that a VPN costs most where it is most needed. It is a good line and it is only partly true, so it is worth showing exactly how far the data carries it. Joining this index to our monthly censorship measurements for the same window gives 6 countries with at least one platform confirmed blocked at national scale. Their median cost is 6.0 days, against 4.4 days for the 164 countries with no confirmed block.

The direction supports the hypothesis. The sample size does not support leaning on it. Six countries is a handful, the difference between the two medians is under two days, and the ranking inside the blocking group is driven almost entirely by income rather than by how much is blocked — the country blocking the most platforms in this window is not the country where the subscription costs the most.

Two further countries returned confirmed blocks in this window but were flagged by our pull as possibly measured on a filtered local network — a school, office or hotel connection rather than a national policy. They are excluded from the group above rather than named as censoring countries, because a handful of measurements from one filtered network looks identical to a national block until someone checks. Naming a country as a censor on that evidence would be the single easiest mistake to make with this dataset, and the hardest to walk back.

Practical rule

A price is only meaningful next to an income. Before repeating that a VPN costs “a few dollars a month”, check whose month you mean — the same subscription is 7.4 hours of average earnings in Switzerland and 145 days in Burundi.

The income figures are oldest exactly where the question matters most

12 of the 172 countries here have no income observation from the past two years. This is not a rounding problem to be quietly averaged away — the countries with the oldest statistics are largely the ones in or recently out of conflict, which is to say precisely the population this piece is about. A figure carried forward from 2015 is presented as such below rather than dressed up as current.

A related gap: 9 of the 181 countries on our measurement axis returned no income figure at all and are absent from every table here. Most are small territories and dependencies. One, Taiwan, is absent because of how the source organises its country list rather than because the data does not exist. Absence from this index is absence of a World Bank observation, nothing more.

What this measures, and the three things it does not

The method is deliberately simple, because the finding does not need a complicated one. Each month we take the median ongoing annual cost across the VPN subscriptions we track — read from real checkout pages by our own daily capture, 51 snapshots in the current window — convert to US dollars at the European Central Bank reference rate, and divide by each country’s GNI per capita.

• It does not measure price discrimination. One price is held constant for every country. We do not yet measure per-country checkout, so this index shows the affordability of a given price, not whether providers charge different countries differently. Those are different studies and conflating them would be the obvious way to overclaim.
• It does not measure what any individual pays. The figure is a median across tracked subscriptions, not a specific product. Introductory offers are substantially cheaper than the ongoing rate — the median first-year cost across the same set is $34.73, against $95.19 to renew — so a first-year buyer faces a smaller version of every number here, and the same buyer faces the full version twelve months later.
• It does not describe a typical household. GNI per capita is a national average across the whole population. In unequal countries the median earner is well below it, so for most people in most of these countries the real figure is higher than the one printed here.

No provider is named anywhere in this index, and that is a methodological choice rather than a courtesy. The question is what a VPN costs relative to income, which is a property of the market and of the country, not of any one company. Publishing a per-brand ranking here would turn a policy measurement into a comparison table, and the comparison table is not the finding.

Why an affiliate site publishing critical pricing data is worth reading carefully

vpnrank.io earns commission when a reader buys a VPN through our links. That is worth stating plainly on a page about VPN pricing, because it is the obvious reason to distrust it. The incentive runs towards making subscriptions look cheap and necessary, and this index does the opposite: it says that for 65 of 172 countries the annual cost is more than a week of average income, and that 20 are past a month.

The defence against the incentive is that everything here is re-derivable. The income side is a single public World Bank request that anyone can repeat. The price side comes from our own daily checkout capture, which also underpins the VPN price index and its published methodology. The blocking side is OONI’s open data, joined the same way as in our outage-versus-blocking study. If a figure here is wrong, it is wrong in a way that can be caught.

One practical consequence worth naming, since the data points at it directly: in the countries at the top of this table, the honest answer to “which VPN should I buy” is often that a paid annual subscription is not a proportionate expense. Proton VPN’s free tier exists and is not time-limited, and every provider we track offers a 30-day money-back window that makes a genuine trial possible without committing a month of income. Neither of those is a recommendation to buy anything.

How this index is kept current, and when it changes

The two halves move at different speeds and it would be misleading to imply otherwise. The price side is re-derived every month from the current checkout capture, so the constant at the top of this page changes as the market does. The income side updates roughly annually, when the World Bank refreshes the indicator — the current edition uses data last updated 2026-07-13. A month where only the price moved will shift every country by the same proportion and change none of the ranking.

The blocking join is re-pulled monthly alongside our censorship index, so the 6 countries in that section are specific to August 2026 and should be expected to change. Every edition states the window it covers, and previous editions are not silently overwritten.