Video walkthrough · Thailand · Four cities

Best place to buy a condo in Thailand? I ranked all four cities.

Chiang Mai first, Phuket second, Pattaya third, Bangkok fourth. That is the honest order on the evidence I hold, and I will defend every line of it below.

Then I am going to argue that the ranking barely matters. The distance between the best and worst building inside a single city is several times the distance between the cities themselves. The question everybody asks is the small one; the question almost nobody asks is the large one.

The same ranking, on video

Nothing on this page depends on the video. Everything in it is written out above and below.

The episode publishes shortly. Everything it covers is written out in full below.

How were the four cities ranked?

On two things a buyer can check: what it costs to get in, and what the file says the stock actually returns after the costs of holding it. Not on how a city feels to spend a fortnight in.

Those two are deliberately kept apart below. An entry price and a net figure are different claims with different evidence behind them, and stacking them in one row invites a reader to treat the pair as a promise. They are not a promise. They are two descriptions of a market.

Why is Bangkok bottom of the list?

Because it has the highest entry price of the four and the lowest published net median of the four. The median building in my Bangkok file is about two million baht more than the equivalent thing in Chiang Mai, and the middle of the Bangkok stock returns less after costs than any other city here.

That is the whole case against it, and it is not a case against the city. Bangkok does one thing better than anywhere else on this board, and it is the thing people discover last: the exit.

In my data a Bangkok seller is competing with roughly ten other units inside their own building. In Pattaya it is nearly thirty. On the day you want out, that is the number that decides whether you set the price or the stack sets it for you.

Is Pattaya worth it if it is the cheapest way in?

It is genuinely the cheapest entry on this list, and the published net median sits above Bangkok's. On those two facts alone it would rank higher than third.

Two things hold it at B, and neither is about the place itself.

The first is the exit again. My Pattaya file averages close to thirty units listed in the same building at the same time. You are not competing with a city, you are competing with twenty-nine near-identical boxes that share your view and your lobby. The only lever anyone in that stack holds is price.

The second is evidence. Of 147 buildings examined, only 51 carried rent evidence from enough independent places to publish a figure. That is the heaviest cut of any city file I have built. It is not a judgement about Pattaya; it is a statement about how many places a Pattaya rent can be checked against.

Phuket has the best numbers, so why is it only A?

Because of one assumption sitting underneath every net figure on this page: twelve paid months.

Phuket does not rent twelve months. It has a season. My own Phuket research says so on the page itself — every net figure there is gross of seasonality. So the honest version of Phuket's published median is lower than the published median, and I cannot tell you precisely how much lower. That uncertainty is the reason it is not top, not the number.

Why is Chiang Mai top?

Disclosure first, because it changes how you should read this section: I own two condos in Chiang Mai. Weigh the ranking accordingly. I am based in Amsterdam and do the fieldwork here.

Chiang Mai's published net median is level with Phuket's, on a comparable evidence base, at the lowest entry price of the four. That is the same income evidence as Phuket for less money, and better income evidence than Bangkok for considerably less money.

And it rents all twelve months. The tenants are long-stay people signing year leases, not a season. Phuket's figure needs an asterisk about the months it is empty. This one does not, and that asterisk is worth more than a tenth of a point on a median.

Why is the city ranking the smallest decision you make?

Because every figure in that second table is a median, and a median is the middle of everything that was for sale.

Which makes it the number you get if you do nothing in particular. Walk into a city, buy something ordinary at an ordinary price, rent it out the ordinary way, and you land on the median. That is what a median is. Nobody intends to be average, but average is the default, and the default is what arrives unless something specific stops it.

So the useful question was never which city has the better middle. It is how far above the middle you can get inside one city, and what that takes.

Two units, one city, and the gap between them

I can show that, because I have done it twice in the same city, with the same method — and it worked far better one time than the other.

Same city. Same method. Same buyer. Eight point one five, and five. The reason is not luck, it is two ratios: I paid about a third more capital for a unit that rents for about a sixth less, and the maintenance on the second eats roughly fifteen per cent of its rent where the first eats five. Close to three times the drag, in the same city.

One more thing about the first unit, because it is the part people skip: it came out third on a full scan of the city. Not third because I liked it third best — the top two sold before I could get an offer in. The list existed before the unit did. That order is the product; the apartment was just what survived it.

How big is the spread inside a single city?

Larger than the gap between the cities, in every city I have published.

In the Pattaya file the weakest building I published sits under two per cent net and the strongest clears 7.64% — two buildings clear seven. In Chiang Mai the range runs from under two to over six. Bangkok is the tightest of the four and it still close to doubles from one end to the other.

Now set that against the distance between the cities themselves. Best city to middle city is about one and a quarter points. Inside Pattaya alone it is over five.

What does this ranking not tell you?

It does not tell you whether a specific building has foreign quota left, what its sinking fund looks like, or whether its juristic person has a record worth reading. A city median cannot answer any of those, and those are the questions that decide what your unit actually does.

It also carries every limit of its sources. Pattaya's published figures rest on a heavily cut evidence base. Phuket's rest on an assumption about months that the island does not keep. Both are stated on the pages they come from, and both are the reason a median is a starting point rather than an answer.

What do the tier letters actually mean here?

They are a shorthand for one question: how much does this city give back for what it asks, once the evidence behind the number is taken into account?

S is the city where the published income evidence is strong, the entry price is the lowest of the group, and there is no seasonal asterisk hanging over the figure. A is a city whose headline number is level with the top but rests on an assumption the market does not keep all year. B is a city that is cheap to enter and hard to leave, on the thinnest evidence base of the four. C is the city that costs the most and returns the least in the middle of its stock, while offering the cleanest exit.

None of those letters is a verdict on whether a city is a good place to live, visit or retire to. They are about one narrow thing: what the file says happens to money.

Why does the exit matter as much as the income?

Because the income figure assumes you keep the unit and the exit figure decides what happens when you do not. Most of the analysis a buyer reads is about the first and almost none of it is about the second.

The measure I use is simple and checkable: how many units are listed for sale in the same building at the same time. It is a count of the people you are competing with who own something functionally identical to yours — same building, same view, often the same floor plan.

In my Bangkok data that number is around ten. In Pattaya it is close to thirty. When a buyer walks into a stack of thirty near-identical units, nothing distinguishes yours except the price, and the only direction that moves is down. A strong income figure on a unit you cannot leave is a weaker position than it looks on a spreadsheet.

Why do most buildings get cut before a figure is published?

Because a rent that cannot be corroborated is an advertisement, not evidence. The asking rent on a listing is what someone hopes to receive. What matters is what the building actually achieves, and that requires finding the same building described by more than one independent source and seeing whether the numbers agree.

Most buildings fail that. In Pattaya, of 147 buildings examined only 51 carried enough independent evidence to publish. Across the four cities, roughly 270 buildings survived from a starting set many times larger.

That cut is the reason the building-level figures are worth more than the city medians, and it is also the reason there are so few of them. A file of 33,030 listings becomes a few hundred buildings you can say something defensible about. The rest are not wrong; they are simply unverified, and publishing an unverified rent as though it were a measurement is the specific failure this whole method exists to avoid.

What does 33,030 actually count?

Distinct listings collected across the four cities. Not units screened, not units delivered to a client, and not rows in a spreadsheet — those are three different measures and conflating them is the most common way a property number turns into a false claim.

A collected listing is something that was advertised and captured. A screened listing has been through the full check. A delivered row is what a specific client received for a specific brief. The same underlying work produces all three, and they are never the same size. When you see a large round number attached to a property claim, the useful question is which of the three it is, and the honest answer is usually that nobody said.

So: 33,030 is what was collected. The building-level figures rest on a far smaller set, roughly 270 buildings, because that is how many survived rent corroboration. Both numbers are on this page precisely because one without the other is misleading.

What has a net figure already paid for?

Everything that comes off the rent before the money is yours, which is where most advertised figures quietly stop.

On my first unit that is one month of agent fee to place the tenant, the building's annual maintenance, and a combined figure for property tax and a repair buffer. On the second there is no agent line because I let it directly — but the building maintenance is two and a half times larger, and that single difference does more damage than the agent fee ever did.

That is the part worth carrying away from the two case studies. The rent is the number people compare. The costs of holding are the number that decides the outcome, and they vary far more between buildings than rents do. A building with strong rent and heavy maintenance can easily finish behind a cheaper building with light maintenance, in the same city, on the same street.

How much does the season really change Phuket?

Enough that I will not publish a corrected figure, which is an uncomfortable thing to write on a page full of figures.

Every net number on this page is built on twelve paid months. In a market with a season, some of those months are not paid. To restate Phuket's median honestly I would need a reliable occupancy figure at building level — how many months a typical unit in a specific building is actually tenanted — and that is not something I can corroborate to the standard the rest of this work is held to.

I could estimate it. An estimate presented beside corroborated figures would look like one of them, and within a week it would be quoted back as though it were measured. So the asterisk stays, Phuket sits second rather than first, and the gap it would need to close is stated rather than guessed. That is the whole reason the ranking has a city with the joint-best headline number in second place.

What would change the ranking?

Better evidence, mostly. If Pattaya's rent evidence improved to the point where a majority of its buildings could be corroborated rather than a third, its position would be argued on the numbers rather than on the size of the gap in them.

If a reliable occupancy figure existed for Phuket — how many months a typical unit in a given building is actually paid for — its published median could be restated honestly instead of carrying an asterisk, and it might well take the top position. The number is not the problem. The missing denominator is.

And if Bangkok prices moved relative to its rents, the arithmetic that puts it fourth would change with them. None of these letters is permanent. They describe a file at a moment, which is why every figure on this page carries the size of the evidence base beside it.

How do you use this on your own shortlist?

Pick the city on the things a median cannot capture — where you want to be, how long you intend to hold, whether you need twelve months of tenancy or can live with a season. Then spend the real effort one level down, on the building.

At building level the questions are specific and answerable: is there foreign quota left in this particular building, what does the juristic person's record look like, what does the sinking fund actually hold, what do comparable units in the same building genuinely rent for, and how many of them are for sale right now. Those five answers move the outcome several times more than the choice between Chiang Mai and Phuket.

That is the whole argument of the episode, and the reason the tier list is presented and then immediately discounted. The ranking is the part that is easy to have an opinion about. The building is the part that decides what happens.

Why did this ranking take so long to publish?

Because the honest answer is to a slightly different question than the one people ask, and I did not want that to read as dodging it.

The question that arrives is “which city”. It is a good question for deciding where to spend your time. It is a weak question for deciding what happens to your money, because the answer moves the outcome by about a quarter of what the next decision moves it by. Answering it alone would have been useful and slightly dishonest at the same time.

Refusing to answer it would have been worse. People asking which city are not making a mistake; they are starting at the obvious end. So the order here is deliberate: commit to a ranking, defend it with the figures, and then show exactly how much of the result it accounts for.

The two units are in the episode for the same reason. One of them worked and one of them did not, and they were bought in the same city by the same person using the same process. If the city were the decisive variable, that could not happen. It happened, the line items are published, and the difference between them is larger than the difference between any two cities on the board.

That is the argument. The tier list is real, defensible, and the least important thing on this page.

Where to read the underlying work

The city-level pages carry the area breakdowns behind each ranking: Chiang Mai, Phuket, Pattaya and Bangkok. The gap between what a listing advertises and what it keeps is set out in the net yield gap.