The seven dangerous days

Every April, Thailand does something I have never seen another country do.

It announces in advance that a large number of people are about to die, gives the period a name, and publishes a running total each morning.

เจ็ดวันอันตราย — the Seven Dangerous Days. In 2026 it ran from 10 to 16 April, and the final figures were 1,242 accidents, 1,200 injured, and 242 dead.

The daily tally went 20, 50, 24, 51, 30, 23, 17. Anyone in Thailand could follow it in the morning news, the way you’d follow a score.

What visitors see, and what it is

Foreign coverage of Songkran is water. Water pistols on Silom, a soaked backpacker, a photograph of somebody laughing with a bucket.

That is real and it is a genuinely joyful thing and I don’t want to be sour about it.

But the water fight is the surface of the largest scheduled human movement in the Thai year.

Buses carried 859,247 passengers across 47,596 trips between 9 and 15 April, with a single-day record of over 150,000 on 11 April. Rail carried over 1.1 million, and on 17 April alone — the return — 1,227,707 people. Public transport passed 10 million trips in the first four days, with reporting putting total movement across the period around 20 million.

The majority went to Isan, then the North, then the South.

That is not a festival travel pattern. That is a country reassembling itself.

Why it exists

Article 65 argued that Thailand is a capital with a country attached — 17.4 million people in the Bangkok Metropolitan Region, about a quarter of the population, more than ten times the second city.

Songkran is what that fact looks like when it moves.

Article 37 described the remittance economy: 67% of elderly Thais have a migrant child, and around 30% of rural household income arrives from somewhere else. Article 67 described katanyu, the obligation that makes the relationship non-optional.

The money goes home every month. The people go home in April.

Which means the journey isn’t leisure and isn’t really discretionary. It is the annual, physical, in-person discharge of a duty that the rest of the year is performed by bank transfer. You go because your mother is there and you have not seen her since last April.

Where the deaths actually come from

The breakdown is the part that changed how I think about this.

Motorcycles: 64.55% of crashes. Speeding: 40.65%. Cutting in: 25.20%. And 87.80% of crashes happened on straight roads.

Straight roads are not dangerous. Straight roads are the easy part.

So the picture assembling itself out of those numbers is not one of treacherous conditions or difficult driving. It is a person on a motorbike, on an ordinary rural road, at the end of a long journey, during a week when the whole country is drinking.

These are not accidents caused by the roads. They are caused by the journey existing.

Bangkok recorded the highest provincial toll at 21. Ten provinces recorded none at all.

The uncomfortable conclusion

I want to state this carefully because it’s the argument of the piece and it can be misread as callous, which is the opposite of the intention.

The Songkran road toll is a cost of urban primacy.

If a quarter of the population did not live several hundred kilometres from their parents, the journey would not need to be made, and a large share of those 242 deaths would not have occurred. The deaths are not caused by the festival. They are caused by the distance between where Thais earn and where their families are — which is the same distance that produces the remittance economy, the ageing farm workforce, and the ฿600 pension that adult children top up.

Every consequence of primacy this site has documented has been financial or demographic. This one is a number of funerals, and it arrives on the same seven days every year, and the state knows it well enough to book the name in advance.

Article 76 argued that Thailand’s structural problems all require planning horizons longer than its institutions have had. Decentralisation is on that list. So is this.

What’s actually working

And I should be equally careful to say that the trend is good, because a piece that only reports the toll is not reporting honestly.

The 2026 figures were down against the three-year average — accidents by around 35%, injuries by around 37%, and deaths by roughly 5 to 10% depending on which comparison is used.

Accidents falling 35% while deaths fall 5% is worth pausing on. That gap usually means the remaining crashes are the severe ones — helmetless motorcycle riders at speed, which is exactly what the vehicle and cause breakdown describes. Enforcement and checkpoints appear to be removing the survivable collisions faster than the fatal ones.

Which points at where the remaining gains are: helmets and alcohol, on rural roads, on motorbikes. Not signage, not road engineering, not the straight roads that 87.80% of the crashes happened on.

Ten provinces reaching zero is a genuine achievement and worth studying rather than filing as luck.

If you’re travelling in April

Book ahead and travel outside the peak days. The 11th and the return weekend are the crushes.

If you ride, wear the helmet. Two-thirds of the fatal crashes involve motorcycles and this is the single highest-value decision available to you.

Don’t drive at night on rural roads during the week, and be aware that other people have been drinking even if you haven’t.

And expect Bangkok to be empty. The city is genuinely lovely for four days a year, which is the compensation for everyone you know being four hundred kilometres away.

Common misconceptions

“Songkran is a water festival.” It’s the Thai new year, and functionally it’s the annual national return home. The water is one part of it.

“The deaths are caused by drunk water fights.” The fatal crashes are overwhelmingly motorcycles on ordinary roads, most on straight sections, during the travel days.

“It’s the worst road safety week in the world.” It’s a severe week in a country with a serious year-round road safety problem. The concentration is what makes it visible.

“Nothing is being done.” Accidents fell around 35% against the three-year average in 2026 and ten provinces recorded zero deaths.

“Thais travel for the holiday.” Most travel because their family lives somewhere else — which is a different thing and the reason the pattern is so predictable.