Why PPP Theory Fails: The Hidden Limits of the Big Mac Index

I love the Big Mac Index, but I’d be lying if I said it’s the full story. I run bigmacindex.app as a side project and pull The Economist’s monthly CSV myself, and the more time I spend with the data, the more I see how often the index gets cited as if it were a verdict. It isn’t. It’s a starting point. The 38% Swiss premium and the 59% Taiwanese discount in the January 2026 dataset describe a real gap between two prices, but they don’t describe what most readers think they describe — the “true” value of a currency.

Purchasing power parity (PPP) is the deeper theory the Big Mac Index is trying to illustrate. The textbook version goes: if a tradable basket of goods is identical across countries, prices should equalise at market exchange rates, and any deviation reveals an over- or undervalued currency. The Big Mac is a single-item version of that basket — a clever, accessible, but ultimately leaky proxy.

This article walks through seven structural limits of the Big Mac Index, with concrete country examples from the January 2026 data. At the end I’ll be honest about the limits of my own dataset on bigmacindex.app — what I can and cannot claim. And I’ll close with how I think the index should actually be used.

Limit 1: Most of the Price Is Labor, Not Exchange Rate

The biggest single line item in a Big Mac’s local price is not the patty — it’s the wage of the person who flipped it, the rent on the store where you ate it, and the salary of the manager. None of those costs cross borders. None of them adjust with the exchange rate. They are local prices for local services, and that means a Big Mac is overwhelmingly a wage indicator, not a tradable goods indicator.

Switzerland makes this point starkly. As of January 2026, a Big Mac in Switzerland costs CHF 7.20, which converts to USD 7.99 at an exchange rate of CHF 0.90 per USD — a 38% premium over the US benchmark. The standard PPP reading is “the franc is 38% overvalued.” But Swiss minimum wages and rents are roughly that much higher than the US equivalent. The “overvaluation” is mostly the price of Swiss labor, embedded in the burger.

This isn’t a fringe point — it’s the central critique of the Big Mac Index in academic economics. Most non-tradable inputs (labor, real estate, local services, energy delivery to the store) don’t arbitrage across borders, so they don’t enforce PPP. Countries with higher wages will have more expensive burgers, full stop, regardless of whether their currencies are “correctly” priced in financial markets.

Norway tells the same story. NOK 75 per burger, USD 6.67, +15.3% vs. the US. Norway’s minimum wage and Norway’s cost-of-living scale are both well above US levels. The burger is responding to Norwegian wage levels, not Norwegian financial market valuation of the krone.

The Economist itself publishes a “GDP-adjusted” version of the index that tries to control for income levels, and it’s noticeably more credible than the raw index for this reason.

Limit 2: The Product Isn’t Actually the Same Product

PPP requires the basket of goods to be identical across countries. The Big Mac Index advertises that the burger is identical worldwide. It isn’t.

The cleanest example is India. McDonald’s India doesn’t sell beef for religious reasons, so the “Big Mac” in the index for India is actually the Maharaja Mac, made with chicken (and sometimes paneer). As of January 2026, the Maharaja Mac costs INR 226, or USD 2.62 — 54.8% below the US Big Mac. Comparing INR 226 of chicken to USD 5.79 of beef and calling the rupee “undervalued” by 55% is a stretch. We’re comparing two different products.

Other variants:

Israel — the Big Mac is made without cheese in some kosher locations, with kosher-process beef in others. Same name, different supply chain.
Saudi Arabia, UAE, Lebanon — halal beef, different supply chain economics from US/European.
Indonesia, Malaysia — halal, sometimes different bun specifications.
Vietnam — quietly different sandwich size in some stores; The Economist tries to standardise to a global spec, but the local product still varies.

These differences are not trivial. Beef vs. chicken can be a 40-50% input cost difference. The index treats them as the same.

Limit 3: Government-Managed Exchange Rates Break the Math

The Big Mac Index assumes the exchange rate is a market price set by supply and demand. For a meaningful chunk of the world, it isn’t.

Argentina is the headline case. As of January 2026 the Argentine Big Mac costs ARS 7,300 (USD 6.95 at the official rate of ARS 1,050 per USD) — 20% more expensive than the US, ranking #2 in the world. But ask anyone in Buenos Aires what they actually pay for a dollar and they’ll quote you the blue dollar rate, the MEP rate, or the contado con liquidación rate, which over the past few years have run anywhere from 30% to 90% above the official rate. At the blue rate, the same Big Mac would dollar-convert to roughly USD 3.50 to 4.50 — a totally different story. The “Argentine burger is more expensive than the American one” is, in lived reality, often false. The index reports the official rate because that’s what’s publishable.

Venezuela has the same problem at a larger scale. The official Big Mac price in the January 2026 data is VES 252 at a rate of 56.58 per USD, giving USD 4.45. Parallel-market rates in Venezuela have run multiples of the official rate at various points. Whatever the index reports, locals are not transacting at that rate.

Lebanon — at LBP 480,000 and an official-ish rate of 89,550 per USD, the published Big Mac dollar-price is USD 5.36. Lebanon in 2024-2026 has been operating with three or four parallel exchange rates simultaneously. Pick a different rate, get a different burger price.

China publishes a Big Mac at CNY 25.50 / USD 3.52 / -39.2% in the January 2026 dataset. The yuan is not strictly pegged but is heavily managed by the PBOC inside a narrow daily band. The “implied” PPP rate from the burger is not what the yuan would trade at in a free float.

Hong Kong dollar is explicitly pegged to the USD in a narrow band. The HKD 24 / USD 3.08 / -46.8% reading in January 2026 is more a story about local prices not adjusting than about currency valuation.

PPP theory assumes exchange rates are flexible. In about a quarter of the index, they aren’t.

Limit 4: Taxes, Tariffs, and Regulatory Cost Are Invisible in the Headline

The headline Big Mac price includes whatever local taxes and tariffs apply. The index doesn’t separate them out. That distorts comparisons substantially.

A few specifics:

Brazil — Brazil has one of the most complex sales-tax structures in the world (ICMS, PIS, COFINS, ISS, often stacked). Local prices include all of it. The Brazilian Big Mac costs BRL 23.90 (USD 4.03, -30.5% vs. US) in January 2026. Some material fraction of that BRL 23.90 is tax that doesn’t exist in the US comparison. Strip tax-for-tax and Brazil would look cheaper still.
European Union — VAT on restaurant meals ranges from ~5% (Luxembourg) to 25% (Denmark, Sweden). The Euro area Big Mac at EUR 5.67 (USD 5.95) is reported with VAT included; the US Big Mac at USD 5.79 is reported without sales tax in most state averages.
Beef tariffs — Countries with protective tariffs on imported beef (Japan, Korea, India for the small beef market that exists) carry that tariff in the burger price.

The Economist’s index does not attempt to correct for these. Two countries that look 10% apart in dollar terms might be at parity after tax-and-tariff adjustment.

Limit 5: McDonald’s Brand Positioning Varies by Market

In the US, McDonald’s is fast food — the inexpensive option you eat in a hurry. In several other markets, McDonald’s is positioned as a premium aspirational brand.

China — McDonald’s is a middle-class restaurant chain. Pricing reflects that brand positioning, not “cheap food.”
India — Same story, more pronounced. McDonald’s India is a sit-down, middle-class outing, not a bottom-of-the-pyramid product.
Vietnam (VND 76,000 / USD 3.03) — A Big Mac in Vietnam is roughly equivalent to a half-day’s wage for many workers. It is positioned as a Western treat, not a default lunch.
Russia (pre-2022) — Before McDonald’s exit, the brand was a quasi-aspirational Western symbol; pricing reflected that.

When McDonald’s prices for a “treat” segment, the local burger price is set by what the local middle class will pay for an outing — not by what raw input costs imply. The index reads this as “currency undervalued,” but the burger price was never trying to be at PPP in the first place.

The opposite is also true. In the US, McDonald’s is engaged in a fairly aggressive price war with Wendy’s, Burger King, and Chick-fil-A. US Big Mac pricing is at the low end of what the market will bear. So the US benchmark of USD 5.79 is itself unusually competitive, which makes every other country look more expensive than it would against a “natural monopoly” McDonald’s.

Limit 6: Supply Chains and Local Sourcing Move the Price

The Big Mac is not made the same way everywhere. McDonald’s runs a regional supply chain model — bread bakery in one country, beef supplier in another, sauce in a third. Local sourcing rules mean inputs are not on a single global cost curve.

Australia (AUD 7.75 / USD 4.87) — Australia is a major beef exporter, but McDonald’s Australia uses largely Australian beef anyway because of import/processing rules. The local beef is more expensive than what US McDonald’s pays for US beef. The burger price reflects that.
Japan (JPY 480 / USD 3.11) — Japan’s beef supply is partially imported, partially Wagyu-adjacent domestic. The supply chain is its own creature, priced on its own dynamics.
South Africa (ZAR 51.90 / USD 2.78) — Local sourcing, local labor; the burger price is a fairly clean read of local input costs, which explains the deep discount.
Norway, Sweden, Denmark — Cold-climate beef supply, expensive logistics, narrow domestic market. All push the burger price up independent of any currency story.

The point: even setting wages aside, the traded inputs of a Big Mac aren’t really traded globally at one price. Each region has its own supply graph, and the index’s PPP reading is contaminated by those graph differences.

Limit 7: One Burger Is Not a Basket

This is the deepest limit, and it deserves a moment.

Purchasing power parity is a claim about baskets of goods — broadly diversified consumption baskets that approximate what households actually spend money on. The Big Mac is one item. One. Its price is sensitive to beef cycles, McDonald’s franchise economics, and brand positioning in ways that a real consumption basket would average out.

The IMF and OECD publish proper PPP indices using thousands of price observations across hundreds of categories — food, rent, transportation, healthcare, education, electronics, services. Those indices are far more credible as PPP measurements, and they often disagree with the Big Mac Index.

For example, the OECD’s PPP rate for Switzerland in 2025 suggested the franc was about 15-20% overvalued vs. the USD — meaningful, but well below the Big Mac Index’s 38% reading. The OECD’s basket includes things like Swiss healthcare and Swiss rents that don’t enter a burger. The IMF’s PPP for India implies the rupee is 60-65% undervalued vs. USD — more than the Big Mac Index’s -54.8% reading.

So even within the family of PPP measures, the Big Mac is an outlier. It’s the most accessible, the easiest to update, and the easiest to communicate. It is not the most rigorous.

What I Can and Cannot Claim About My Own Dataset

I want to be honest about what bigmacindex.app actually is and isn’t.

What it is: A monthly mirror of The Economist’s published Big Mac Index, with country pages, charts, and plain-English explainers. I pull the raw CSV from The Economist’s GitHub release, normalise the country list, add some metadata (flags, regions, slugs), and publish. The underlying numbers are not my work — they are The Economist’s data, redistributed with attribution.

What it isn’t:

Not original price collection. I do not have a network of correspondents pricing burgers in each country. The Economist does that work; I redisplay it.
Not a real PPP index. I’m reusing the Big Mac Index, with all the limits above intact.
Not a currency forecast. The “implied” or “fair value” exchange rates that come out of this index are illustrative, not predictive.
Not comprehensive. The dataset covers 54 countries as of January 2026. About 140 countries are not in it.

When I write things like “the yuan is 39% undervalued” on a country page, I mean “the Big Mac Index’s specific arithmetic puts the yuan 39% below USD-PPP parity at January 2026 prices.” That’s a much narrower claim than “the yuan is really worth 39% more.” The narrow claim is the one I’m willing to defend.

How to Use the Big Mac Index Without Misusing It

Despite all of the above, the Big Mac Index is still useful. The trick is to treat it as a question generator, not an answer.

A short field guide:

Treat large gaps as questions, not facts. “Why is the Indian Big Mac 55% cheaper than the American one?” is a better use of the index than “the rupee is 55% undervalued.” The first prompts inquiry; the second pretends to a precision the index doesn’t have.
Use it for relative changes more than absolute levels. A country whose Big Mac dollar-price has moved from -30% to -45% over five years is telling you something real about currency softening, even if the absolute -45% reading is contaminated by all the limits above.
Cross-check against OECD or IMF PPP rates. If the Big Mac and the OECD basket agree, your conclusion is robust. If they disagree by 20 points, the burger is leading you astray.
Mind the regime. For countries with managed or multiple exchange rates (Argentina, Venezuela, Lebanon, China to a lesser extent), the index’s reading is mechanical arithmetic on a published rate, not a behavioural read on currency value.
Read it alongside wage data. The Economist’s GDP-adjusted version of the index handles a lot of the labor-cost critique. The raw version doesn’t.
Recognise it as a teaching tool first. The Big Mac Index has survived for 40 years not because it is accurate, but because it is communicable. It makes PPP intuitive in a way that the OECD PPP report does not. That’s a legitimate function — just don’t mistake intuition for precision.

Closing

Despite all this, the Big Mac Index is still the most accessible way to think about exchange rates. The Economist invented something genuinely useful in 1986 when they published the first version — a way to talk about currency valuation without an econ PhD. Forty years on, it’s still the best way to make PPP feel like something you can hold in your hand.

But hold it gently. Use it as a starting point, not an answer.

If you want to see the underlying data, every country page on bigmacindex.app links back to the raw CSV. If you spot an error, the email is [email protected]. I’d rather get a correction than ship a clean-looking number that’s quietly wrong.

Data: The Economist Big Mac Index, January 2026 release. Methodology: see /about#methodology. Author: Robert. W, independent developer and editor of bigmacindex.app.