BigMacIndex.app vs The Economist: which number should you use?

The Economist runs the survey. We don't, and we've never claimed to. What we publish is the six months in between, the countries their file folds together, and the markets where there is no Big Mac to price at all.

That is the whole answer, and if it's what you came for you can stop here. The rest of this page is the arithmetic behind it — including the four jobs where the right move is to close this tab and go read theirs.

What the upstream file actually contains

Every Economist-tier price on this site comes out of one public CSV. It is not scraped, not inferred, and not ours: big-mac-full-index-jul-26.csv, published 29 July 2026, 2,056 country-date records running 1 April 2000 to 1 July 2026. Anyone can download it straight from them.

Four facts fall out of that file, and all four are checkable in about a minute.

• 45 editions, the first dated 1 April 2000 — 26 years of one sandwich. The index itself is older than the file; it has been running since 1986.
• Two a year. The last ten gaps between editions run 181 to 184 days, and every edition since 1 July 2009 — 34 in a row — is dated either 1 January or 1 July.
• 54 markets in the July 2026 edition — and one of those rows is Euro area, standing in for the entire currency union.
• A GDP-adjusted series ships with it. The adj_price and GDP_bigmac columns are the version that corrects for the fact that poor countries are supposed to look cheap.

Take that last one seriously, because it cuts against us. The GDP-adjusted index is the better economics, it has been The Economist's own headline argument since 2011, and the number this site puts in front of you is the raw one. We ship the adjusted columns in the dataset and explain the difference in the methodology, but we don't lead with it. If your question is "is this currency cheap relative to how rich the country is", the raw index is the wrong tool and ours is the raw index.

The six months in between

Here is the gap we actually fill. Between the January 2026 and July 2026 editions — 181 days — 21 of the 54 markets did not change their local menu price by a single unit of currency. Not a cent, not a baht, not a forint.

Currencies were not so patient. Over the same 181 days the median currency in the file moved 1.6% against the dollar, and 10 of 54 moved more than 5%. So for those 21 markets, every cent of movement in the published dollar price came from the exchange rate. Nothing about the burger changed.

I live in Australia, so this one I can check by walking down the road. The burger has been 8.50 AUD in both editions. The index reported $5.69 in January 2026 and $5.95 in July 2026. Same sandwich, same price on the board, 4.5% apart in dollars. Both numbers were right on the day they were published — and one of them was on the chart for 181 days.

That stickiness is not an accident of one edition. Thailand has repriced roughly once in five years while Argentina has done it eleven times over a comparable stretch, and we went through the whole series counting. Menu boards move in steps. Currencies move continuously. A semi-annual publication has to pick one moment and live with it.

So we convert continuously instead. The Live Rates toggle on the front page takes the same local prices and applies the current market rate, and the calculator will do the same arithmetic for any pair you name. What we are not doing is re-surveying prices — a live rate on a stale menu price is still a stale menu price, which is exactly why the two dates get reported separately.

The countries the file folds together

One row says Euro area. In all 45 editions, DEU, FRA, ITA, ESP, NLD appear exactly zero times. A whole currency union on one line: 6.19 EUR, $7.08.

We price 5 of them separately, from franchise menus and receipts rather than from the survey:

Spain to France is a spread of 17.6% inside a single currency, and the aggregate sits above 3 of the 5. Averaging that into one PPP observation throws away the most interesting thing in it — that the euro is mispriced in opposite directions for its northern and southern members. The long version, with the Bruegel real-exchange-rate work behind it, is in why the eurozone aggregate is a fiction.

A caveat we'd rather say out loud than bury: these 5 rows are not from The Economist's synchronised survey. They were priced on different days, from different kinds of evidence, and they carry a confidence score for exactly that reason. Dropping them into a cross-country league table alongside surveyed prices, as if all the rows were the same kind of thing, is the mistake the confidence scoring exists to prevent. The scoring rules are written out in full.

The markets with no Big Mac

9 markets on our map have no McDonald's — some lost it, some never had one. The survey cannot price what isn't sold, so it doesn't, and the map has holes. We fill them with the closest local substitute, draw them in green so they're never mistaken for survey data, and keep them out of the rankings. Brands, products, prices and sources are all on the burger index page.

Where you should use theirs, not ours

Four cases, and none of them is close.

1. Anything you have to cite. The series is theirs, the methodology is theirs, and the file is free. Cite big-mac-full-index-jul-26.csv. Our data page carries a ready-made APA and BibTeX entry that points at both, in that order.
2. Income-adjusted comparisons. Their adj_price column, not our front page.
3. Anything where the sample has to be one survey week. Our editorial and substitute rows are the whole point of this site and also the reason we are not a drop-in replacement for a synchronised sample.
4. Licensing. The underlying series is The Economist's and we claim no rights in it. If you want to redistribute the original, that's a question for them, not for us.

Which number for which job

What a twice-yearly file can't do

The survey is a table. Most of what people actually want from it is a question the table doesn't answer directly, so the rest of this site is that same arithmetic pointed at particular questions:

• The calculator — how far a given amount goes somewhere else, in burgers rather than in a rate.
• The world map — 54 surveyed markets plus 9 substitutes, coloured by price, switchable to live rates.
• 76 country pages — one price, its sources, its confidence, and how it got there. Try Japan or Switzerland.
• CSV and JSON downloads and a documented, key-free API — including our editorial and substitute rows, which the upstream file doesn't have.
• The app pricing advisor — the one use of this data that pays for itself, and the reason this site exists at all.
• 47 written pieces, from the full country-by-country breakdown to the case against trusting any of it too far.

The part where being small helps

A reader in Egypt emailed to say our figure was E£125 and his receipt said E£170. He was right, it's on the country page now, and the whole exchange is written up. A reader in Pas-de-Calais sent a till receipt from a commune of 11,000 that prices its Big Mac above the entire euro area. And a claim about a McDonald's in 1970s Iran sat on this site until someone asked for the source, there wasn't one, and we published the correction against ourselves.

None of that is a competitive advantage over The Economist. It's just what a one-person site can do that an institution can't: answer the email, change the number the same week, and say in public that it was wrong. If a price here doesn't match what you paid, the contact page lists exactly what to include so it can be checked.

They have published this thing for forty years and they give the file away for nothing. That is the part most comparison pages leave out — there is no paywall to route around here, and no rival dataset. If the survey stopped tomorrow, this site would have nothing left to convert six months later. We're downstream, and we'd rather say so on a page than have you find out.