Daniel sent us this one — he's asking about how we actually compare purchasing power across different countries, what the real methodologies are beneath the attention-grabbing indexes, and whether using the dollar as a common yardstick even makes sense when currencies are bouncing around for reasons that have nothing to do with what things cost on the ground. It's a question that gets at something genuinely fundamental.
It's one of those questions where the moment you start poking at it, you realize how much of what we treat as solid economic data is actually held together with assumptions and duct tape.
The duct tape of macroeconomics. There's a dissertation in that.
There probably is. So the core problem here is straightforward to state and hard to solve. If you want to compare living standards between two countries, you can't just convert salaries at the market exchange rate and call it a day. A thousand dollars in Istanbul buys a very different life than a thousand dollars in Tel Aviv. And the ratio between those two isn't stable — it shifts month to month based on things that have nothing to do with whether a family can afford rent.
The lira could halve against the dollar because of monetary policy decisions in Ankara, but the price of bread in the local bakery doesn't instantly halve in dollar terms. The exchange rate can move for reasons that are completely decoupled from the actual experience of buying things.
And this is why economists invented purchasing power parity — PPP. The basic idea is beautifully simple. You define a basket of goods and services, you price it in each country in the local currency, and then you calculate the exchange rate that would make that basket cost the same in both places. That's your PPP exchange rate. It's the rate that equalizes purchasing power.
Instead of asking "what's the market exchange rate," you're asking "what exchange rate would make life cost the same.
And the gap between the market rate and the PPP rate tells you something real. If the market rate is much weaker than PPP would suggest, the currency is undervalued in purchasing power terms — meaning a visitor with dollars feels rich, but locals earning the local currency are getting less than their productivity would suggest they should.
I imagine this is where the Big Mac Index comes in — as the version of this that you can explain at a dinner party without putting anyone to sleep.
The Big Mac Index was launched by The Economist in nineteen eighty-six, and it was a joke that turned out to be useful. The idea is that a Big Mac is a standardized product — same ingredients, same preparation, available in about a hundred and twenty countries. So you price a Big Mac in local currency, convert at the market exchange rate, and see where it's overpriced or underpriced relative to the US dollar price. If a Big Mac costs the equivalent of seven dollars in Switzerland but five dollars in the US, the Swiss franc is overvalued.
The Middle East equivalent, as the prompt mentioned, is the falafel index. Which I have to say, as someone who lives in Jerusalem, feels much more locally relevant than a burger I've never actually eaten.
The falafel index has been tracked informally for years — there was an actual academic paper on it out of the Hebrew University, and various journalists update it periodically. Same logic: a falafel pita is a standardized street food across the region, so comparing its price in Tel Aviv, Cairo, Amman, and Beirut tells you something about relative costs.
Though I'd argue the standardization breaks down fast. The falafel I get in my neighborhood has pickled mango and fried eggplant and about seven salads. You go somewhere else and it's three balls in a dry pita and a single sad pickle.
That's actually one of the fundamental critiques of all these single-product indexes. How standardized is the product really? A Big Mac in India is a chicken Maharaja Mac because beef isn't served. The recipe varies. Even within the same country, the price of a Big Mac differs between Manhattan and rural Alabama. And the index doesn't account for local substitutes — in countries where people don't eat burgers, the Big Mac is a niche luxury item, not a staple.
The Big Mac Index is the glockenspiel of economic measurement — attention-grabbing, fun, and not what you'd use if you actually needed precision.
That's a perfect description. But the thing is, it's not useless. Multiple studies have shown that the Big Mac Index actually does a decent job of predicting where exchange rates will move over the medium term. Currencies that are deeply undervalued on the Big Mac measure do tend to appreciate over time. It's a rough proxy that captures something real.
Which brings us to the serious version of this. If the Big Mac Index is the toy, what's the actual tool?
The International Comparison Program, or ICP, run by the World Bank. This is the big one. They do this in cycles — the most recent full cycle was the twenty twenty-one round, with results published in twenty twenty-four. They coordinate price collection across about a hundred and ninety countries, covering thousands of specific products. And I mean specific — they don't just price "rice," they price a particular grade of rice in a particular package size from a particular type of outlet. The level of detail is staggering.
How do they handle the problem that people in different countries consume completely different things? A basket that makes sense in Germany doesn't make sense in Ghana.
This is the hardest methodological problem in the whole field, and there are multiple approaches. The simplest is to use a common basket — price the same set of goods everywhere. But as you say, that doesn't reflect actual consumption patterns. If people in one country eat mostly rice and people in another eat mostly wheat, forcing both into the same basket distorts things.
It's measuring how much it costs to live like an American in Ghana, not how much it costs to live in Ghana.
So the ICP uses a more sophisticated approach. They create regional baskets that reflect actual consumption in each region — Africa, Asia, Latin America, and so on — and then they link the regions using a subset of products that are common across regions. It's called the ring country approach. A handful of countries participate in multiple regional comparisons and serve as bridges.
You're not comparing Ghana directly to Germany. You're comparing Ghana to a regional average, which is linked to another regional average, and so on until you reach Germany. Each link introduces a little bit of error.
Every link compounds the uncertainty. The World Bank publishes these numbers with error margins, and for some country pairs the uncertainty is substantial — plus or minus ten, fifteen percent. Which means that when you see a ranking that says country A has twelve percent higher PPP-adjusted GDP per capita than country B, that difference might not be statistically meaningful.
This is the part that never makes it into the headlines. Someone publishes a ranking and suddenly it's "we're the fifteenth richest country in the world" or "we've fallen to twenty-third," and the underlying data might not support distinctions that fine.
This is before we even get to the problem of within-country variation. The ICP produces national averages, but as the prompt pointed out, there's no single cost of living within a country. The difference between London and rural Wales is enormous. The difference between Manhattan and Mississippi is enormous. The difference between Tel Aviv and a development town in the Negev is enormous.
Which means that even if the national PPP number is methodologically sound, it might not describe the experience of anyone in particular. It's an average of situations that are radically different.
There's been some movement toward subnational PPP. The US Bureau of Economic Analysis publishes regional price parities for states and metro areas. The OECD has done work on this for member countries. But it's expensive and difficult, and most of the world doesn't have it.
Let me bring us back to the currency question, because I think it's the most interesting part of the prompt. The prompt asks: if we're benchmarking everything against the dollar, and currencies are experiencing volatility against the dollar for reasons that have nothing to do with cost of living, is the dollar even a fair yardstick?
This is such a good question. And the honest answer is that the dollar introduces distortions, but we use it because the alternatives are worse. Let me walk through why.
When you're comparing purchasing power across countries, you need some unit of account. You could use any currency in principle, but the dollar is the global reserve currency, it's relatively stable compared to most alternatives, and there's a vast infrastructure of financial data denominated in dollars. So the ICP reports results in what they call "international dollars" — which is a theoretical currency that has the same purchasing power as the US dollar in the United States in a given base year.
An international dollar isn't a real thing that exists. It's a statistical construct.
It's a pure statistical construct. And here's where it gets tricky. If the dollar strengthens against other currencies due to, say, Federal Reserve interest rate hikes, that changes the market exchange rate but shouldn't change the PPP calculation — because PPP is based on actual prices in local currency, not on exchange rates. The whole point of PPP is to strip out exchange rate noise.
In theory, PPP is immune to the problem the prompt is worried about.
In theory, yes. In practice, there's a complication. Global commodity prices are often set in dollars. If the dollar strengthens, oil and wheat and copper become more expensive in local currency terms, even if nothing changed in the local economy. This feeds through to domestic prices over time. So exchange rate movements do eventually affect PPP — just with a lag and indirectly.
The insulation isn't perfect. The dollar seeps in through the commodity channel.
And there's another issue, which is that when the ICP benchmarks its results to the dollar, it's using a specific year's price level in the US as the reference point. If US inflation is significantly different from global inflation, the reference point drifts. The twenty twenty-one ICP round pegged everything to twenty twenty-one US prices. By twenty twenty-six, the actual purchasing power of a dollar in the US has changed, but the international dollar is still defined by twenty twenty-one.
The yardstick itself is stretching or shrinking over time, and we're pretending it's fixed.
And this is why comparing PPP numbers across different ICP rounds requires careful adjustment. You can't just take the twenty eleven number and the twenty twenty-one number and subtract them.
Okay, so let's get practical. The prompt mentions wanting to compare two families — one in Turkey, one in Israel — and understand their actual purchasing power. Walk me through how you'd actually do that in a way that's more meaningful than just converting salaries at the spot rate.
The first and most important step is to use PPP exchange rates rather than market rates. So you'd take the family's income in local currency, divide by the PPP conversion factor for that country, and you get income in international dollars. That tells you what that income could buy if prices were the same as in the US.
Where do you actually find these conversion factors?
The World Bank publishes them. The OECD publishes them for member countries. For twenty twenty-three, the PPP conversion factor for Israel was about three point seven shekels per international dollar. For Turkey, it was about nine point three lira per international dollar. Those numbers shift annually. The market exchange rates in the same period were very different — the lira was trading at something like twenty-seven to the dollar, which is about three times weaker than its PPP rate would suggest in purchasing power terms.
Which means a Turkish family earning lira and spending lira is actually much better off than the raw dollar conversion would make them look. Their purchasing power is higher than the exchange rate suggests.
And this is the single most important insight that PPP gives you. The market exchange rate systematically understates the living standards of people in poorer countries, because non-traded goods and services — housing, haircuts, restaurant meals, local transportation — are cheaper in those countries. PPP captures that. Market exchange rates don't.
If you're comparing Israel and Turkey, the raw income gap in dollar terms is going to look much larger than the actual gap in what families can afford.
Israel's nominal GDP per capita in twenty twenty-three was about fifty-five thousand dollars. Turkey's was about thirteen thousand. That's a four-to-one ratio. But on a PPP basis, Israel's GDP per capita was about fifty thousand international dollars, and Turkey's was about thirty-four thousand. The gap shrinks dramatically — from four-to-one to about one point five to one. That's a completely different picture of relative living standards.
That's not a small adjustment. That's the difference between "these countries are in completely different universes" and "one is somewhat better off than the other.
It has enormous real-world implications. International organizations use PPP to determine eligibility for aid programs. The UN uses it for human development index calculations. The IMF uses it for quota allocations. If you use market exchange rates instead, you dramatically understate the economic weight and living standards of developing countries.
Let's talk about housing specifically, because the prompt brings it up and it's where a lot of these comparisons get especially tricky. The metric mentioned is the number of average salaries required for a down payment. Is that a useful way to do international comparisons?
It's useful but limited. The "salaries for a down payment" metric is intuitive — it answers the question "how many years of saving does it take to buy a house?" — but it runs into all the same problems we've been discussing. What's an average salary? Is it mean or median? Does it include the informal sector? And what's a typical down payment? In some countries it's twenty percent, in others it's thirty or forty, in some it's five.
The housing itself is completely different. A "typical apartment" in Tokyo, in Istanbul, in Tel Aviv, in suburban Dallas — these are not the same thing.
Not even close. Size, construction quality, location relative to employment centers, amenities — all different. The OECD has done some work on comparing housing affordability across countries, and they use a measure called "price-to-income ratio" which is the median house price divided by median household income. But even that requires careful normalization for dwelling size and quality.
There's a company called Numbeo that does crowd-sourced cost of living comparisons. I've seen their numbers floating around in online debates. How seriously should we take those?
Numbeo is fascinating and problematic in equal measure. It's the world's largest crowd-sourced cost-of-living database — they have data on thousands of cities, contributed by users who enter prices for everything from a loaf of bread to a month of preschool. The coverage is incredible, and for major cities the sample sizes are decent. But the methodology has real issues.
Self-selected contributors, no verification of prices, no standardized product definitions.
All of that. There was a paper in twenty twenty-three that compared Numbeo data to official statistics and found that Numbeo tends to overstate the cost of living in expensive cities and understate it in cheap ones — basically, the errors amplify the differences. Which makes intuitive sense. If you're contributing to Numbeo, you're probably someone who notices prices and has opinions about them, and you're more likely to report the prices that strike you as noteworthy.
The people who are outraged by the price of milk are the ones entering the data.
That said, for cities where official statistics are sparse or outdated, Numbeo is often the best available data. It's just important to understand its limitations.
We've got the World Bank's ICP as the gold standard, the Big Mac and falafel indexes as the attention-getters, Numbeo as the crowd-sourced upstart. What else is out there that people actually use?
There's the UBS Prices and Earnings report, which used to be published every three years — UBS would send researchers to price a standardized basket in about seventy cities worldwide, and they'd also compare wages across a set of common professions. It was incredibly detailed. They'd tell you how many minutes of work it takes to buy a Big Mac, an iPhone, a loaf of bread, across cities. Unfortunately, UBS discontinued the report after the twenty eighteen edition.
That's a real loss. Those "minutes of work to buy X" metrics are so much more intuitive than abstract index numbers.
They really are. There's something viscerally understandable about "a bus driver in Zurich works twelve minutes to buy a Big Mac, while a bus driver in Lagos works three hours." It connects the abstraction of price levels to the concrete experience of labor.
Deutsche Bank does something similar, don't they? I remember seeing a "Mapping the World's Prices" report.
Deutsche Bank publishes an annual report called "Mapping the World's Prices" that covers about fifty cities. It's less rigorous than the UBS report was, but it's updated annually and it covers things that matter to expatriates and travelers — gym memberships, cinema tickets, dating costs, that kind of thing. It's more lifestyle-focused than the ICP.
The dating cost index. Finally, an economic indicator I can relate to.
I believe a date in Zurich runs about two hundred dollars. In Mumbai, about forty.
I'm not sure if that says more about Zurich or about dating expectations.
And that's the perennial problem with all of these comparisons. The basket of goods and services that defines a "date" in Zurich is not the same as the basket that defines a "date" in Mumbai. The activity, the venue, the expectations — they're culturally specific.
This is the problem that runs through everything we're talking about. The deeper you go, the more you realize that the very concept of a "standard of living" is culturally embedded. What counts as a necessity, what counts as a luxury, what counts as normal — these are not universal.
Yet we need to compare. Aid organizations need to know where poverty is concentrated. Companies need to know where to set wages for remote workers. Governments need to know how their citizens' living standards compare to peers. The demand for these comparisons is real and pressing, even if the methodology is inherently imperfect.
Let's talk about the remote work angle, because I think it's where a lot of this hits the ground for our listeners. If a company is hiring globally and paying location-based salaries, they're implicitly making a purchasing power judgment. They're saying "this salary in country X is equivalent to that salary in country Y.
Most companies are doing this badly. The typical approach is to take a US salary and apply a geographic differential based on some combination of market rates and cost-of-living data. But the data they're using is often outdated, or it's national-level when the hire is in a specific city, or it doesn't account for the fact that a remote worker's consumption basket is different from a local office worker's.
A remote worker in a small Turkish city earning dollars is not living the same economic life as a local earning lira. Their housing market might be different, their consumption patterns might include more imported goods, their savings might be in foreign currency.
This is what the prompt was getting at with the currency volatility point, I think. If you're a remote worker paid in dollars but living in a country with a weakening currency, your purchasing power is actually increasing over time in local terms, even if your dollar salary stays flat. The exchange rate movement is a windfall. But if the company adjusts your salary based on the exchange rate, they might be taking away a purchasing power gain that reflects monetary policy, not a change in your value or in local prices.
A fair location-based pay system would use PPP, not exchange rates.
But PPP data updates slowly — the World Bank cycle takes years. Exchange rates update every second. Companies tend to use what's available in real time, even if it's the wrong measure.
Which creates a systematic bias against workers in countries with weak currencies. They're being paid based on the exchange rate, which understates their actual cost of living relative to the PPP rate.
The workers in those countries often know this intuitively. They feel that the purchasing power adjustment isn't keeping up with what things actually cost. They're right, but the data infrastructure to prove it doesn't update fast enough to make the case.
If you're a worker trying to negotiate a fair location-adjusted salary, what should you actually bring to the table?
I'd bring three things. First, the World Bank PPP conversion factor for your country, which shows the gap between the market rate and the purchasing power rate. Second, city-level cost-of-living data from a source like Numbeo or Mercer — recognizing the limitations but using it to show that national averages don't capture your situation. Third, specific prices for the things that dominate your budget — rent, utilities, transportation, food — with receipts or listings. The specific beats the abstract every time.
On the flip side, if you're an employer trying to do this fairly, what's the right approach?
There's no perfect approach, but the better ones use a combination of PPP for the base adjustment, local market salary data for the specific role to ensure competitiveness, and periodic reviews that aren't tied to exchange rate movements. The worst approach is to set a salary once in dollars and then adjust it quarterly based on the spot exchange rate. That transfers all the currency risk to the employee and creates constant uncertainty.
The prompt also mentions the average number of salaries for a down payment. Are there any organizations that do this metric well across countries?
The OECD Affordable Housing Database has a measure called "price-to-income ratio" that's roughly this. But the most widely cited cross-country housing affordability data comes from the Demographia International Housing Affordability Survey, which covers about a hundred metropolitan areas across eight countries. They use a "median multiple" — median house price divided by median household income. They rate markets as affordable, moderately unaffordable, seriously unaffordable, or severely unaffordable.
Where do most major cities land these days?
In the twenty twenty-four edition, Hong Kong was the least affordable, with a median multiple of about sixteen — meaning the median house costs sixteen times the median household income. Sydney, Vancouver, and San Jose were all above eleven. Tel Aviv was in the severely unaffordable category. For reference, Demographia considers anything above five to be severely unaffordable.
Sixteen times income. That is staggering.
It means that a median family in Hong Kong, saving at a reasonable rate, would need decades to buy a median home — and that's assuming prices don't rise further. In practice, they either get family help, buy much smaller than median, or rent indefinitely.
Which connects to the point the prompt made about intergenerational wealth transfer. When housing costs sixteen times income, the only way to buy is to have parents who can help with the down payment. The market becomes a mechanism for transferring advantage from one generation to the next.
This is where PPP comparisons become politically charged. If you look at Hong Kong's GDP per capita on a PPP basis, it's among the highest in the world — around seventy thousand international dollars. But that doesn't capture the fact that a huge portion of that income goes to housing, leaving less for everything else. PPP measures what things cost, but it doesn't measure the share of income that's absorbed by necessities.
Two countries could have the same PPP-adjusted income, but in one, housing takes thirty percent of income and in the other it takes fifty percent. The actual disposable income after necessities is completely different.
There's a concept called "residual income" that some researchers use for exactly this reason. You take after-tax income, subtract the cost of a minimum acceptable basket of housing, food, transportation, and healthcare, and what's left is residual income. That's what's actually available for savings, leisure, and everything else. Two places can have identical PPP-adjusted incomes but completely different residual incomes because of housing costs.
I'd guess residual income is almost never reported in official statistics.
It requires detailed data on both incomes and local costs for specific necessities. Some academic researchers compute it, and a few housing advocacy groups use it, but it's not part of the standard statistical repertoire.
To summarize where we are: PPP is the best tool we have for comparing purchasing power across countries, it's methodologically sophisticated but has real limitations, it strips out exchange rate noise in theory but not perfectly in practice, and it doesn't capture the distribution of costs within a country or the share of income absorbed by necessities. The simpler indexes like the Big Mac are useful for intuition but not for serious analysis. And the data infrastructure for doing this well at the city level or for specific household types barely exists.
That's a fair summary. And I'd add one more thing: even with all those limitations, PPP is enormously valuable. Before the ICP and similar efforts, we simply had no systematic way to compare living standards across countries. The fact that the tools are imperfect doesn't mean we should discard them. It means we should use them with appropriate humility and understand what they can and can't tell us.
The scientific approach: all models are wrong, some are useful.
And PPP is one of the more useful wrong models in economics.
If someone comes to you and says "I want to understand whether my quality of life would be better in country A or country B," what's your practical advice?
Don't rely on any single number. Look at PPP-adjusted income for a starting point. Then look at specific costs for the things that matter to you — rent in the neighborhoods you'd actually live in, food prices for what you actually eat, transportation costs for how you'd actually get around. Then look at taxes and what they buy you — healthcare, education, public services. Then look at the intangibles — safety, pollution, social connection, climate. The economic data gets you part of the way, but quality of life is broader than purchasing power.
I'd add: talk to people who actually live there. Preferably people who've also lived where you're coming from, so they can translate between the two contexts. A number can tell you that rent is forty percent cheaper, but it can't tell you whether the landlord will fix the boiler or whether the neighbors will be friendly.
The qualitative stuff matters enormously, and it's completely invisible in the statistics. There's a reason people sometimes choose to live in places with lower PPP-adjusted incomes — because the things that make life good aren't all priced.
Which is maybe the deepest limitation of all. Purchasing power parity measures what things cost. It doesn't measure what things are worth.
That's beautifully put.
I have my moments.
And now: Hilbert's daily fun fact.
Hilbert: In the nineteen seventies, Japanese meteorologists studying katabatic wind patterns on Hokkaido documented their findings in a series of hand-drawn wind maps so precise that local fishermen began consulting them to predict where driftwood would wash ashore after storms.
There's a niche.
I don't know what to do with that information, but I'm glad I have it.
Where does all this leave us? I think the prompt was asking for the basics of how these comparisons work, and I think we've covered that ground. But the thing that sticks with me is how much of our economic discourse treats these numbers as gospel when they're really best-effort approximations. The next time I see a headline about a country's ranking on some cost-of-living index, I'm going to hear your voice saying "plus or minus fifteen percent.
That's probably healthy. A little methodological skepticism goes a long way. The numbers are useful, they're the best we have, and they should be taken seriously — but not literally.
That seems like a good place to land. If you enjoyed this, please leave us a review wherever you listen, or find all our episodes at myweirdprompts.Our producer is Hilbert Flumingtop. This has been My Weird Prompts. I'm Corn.
I'm Herman Poppleberry. See you next time.
