Songkran road deaths: what the 2026 count records

Thailand’s Songkran road-safety campaign counted 242 deaths between 10 and 16 April 2026. The same official seven-day record reports 1,242 road accidents and 1,200 people injured and admitted to hospital.

Those are the numbers worth knowing. They are also more limited than the annual headlines sometimes make them sound.

They are the Road Safety Operation Centre’s count for a defined seven-day high-enforcement campaign. They are not Thailand’s whole April death toll, an annual road-safety total, or a measurement of every risk a traveller faces during Songkran. A count can be important without pretending to be a complete theory of the roads.

An earlier version of this page went the other way: it printed a daily series that did not reconcile with its own final total and treated partial cause and vehicle data as though it described the whole campaign. Those claims have been removed, not made more dramatic.

The published 2026 count

The government says these figures were lower than the equivalent Songkran campaign in 2025: 1,538 accidents, 1,495 injured people and 253 deaths. That is a fall of 296 accidents (19.25%), 295 injured people (19.73%) and 11 deaths (4.35%).

The direction is good. The number 242 is still not a success story in the ordinary sense of the phrase. A percentage decrease is a comparison with another bad week, not a certificate of safety.

What is being counted

The campaign period is called เจ็ดวันอันตราย — the “seven dangerous days”. Government agencies coordinate traffic management, enforcement and public-safety work around Songkran, then publish daily and cumulative updates during the designated period.

The final release distinguishes three things that reporting too often blends together:

• road accidents, the campaign’s recorded incidents;
• people injured and admitted to hospital; and
• deaths during that seven-day window.

Those categories should not be casually converted into one another. In particular, the injury figure is not a count of every scrape or outpatient visit, and the article does not estimate fatalities outside the campaign from it.

The Ministry of Public Health’s final notice adds two age-group observations: people aged 24 or under were 22.3% of those killed, while older people were 16.9%. It is a useful warning about who appears in this count. It does not tell us each group’s exposure to driving, riding or travelling, so it is not a rate of risk by age.

Enforcement is a different dataset

Police recorded 435,387 offences across ten priority traffic categories during the same 10–16 April period. The five largest reported categories were:

These are enforcement records, not causes of the 1,242 accidents. A large number of helmet offences does not prove that helmet use caused a particular share of crashes; a speed citation is not a finding about the cause of an individual fatality. The two datasets describe different parts of the same week.

That distinction is not pedantry. It is how you avoid announcing that a campaign “worked” because a number moved when the official record has not isolated the campaign’s causal effect. The police can accurately report enforcement; the Road Safety Operation Centre can accurately report its campaign count. Neither statement, on its own, proves why the count changed from 2025.

What the province headlines can — and cannot — say

The final government notice identifies Phrae as having the highest cumulative accident and injury totals, and Bangkok as having the highest cumulative death total. Police separately named ten provinces with no recorded deaths in the campaign: Nakhon Phanom, Bueng Kan, Pattani, Phang Nga, Rayong, Satun, Samut Songkhram, Sing Buri, Nong Bua Lamphu and Mae Hong Son.

That is a noteworthy public disclosure. It is not a league table of road safety. The releases do not provide, in the material checked for this article, comparable exposure measures such as traffic volume, population, distance travelled or a consistent provincial rate. A province at zero is not automatically safer than one with deaths, and a province with a larger total is not automatically less competent. The number has to be read with its denominator, which the headline does not supply.

Drink-driving cases are not the whole drink-driving story

The Department of Probation reported 6,180 cases entering its process during the campaign: 5,806 drink-driving cases, 366 drug-driving cases, five reckless-driving cases and three street-racing cases. It says the drink-driving total was 294 cases, or 4.82%, lower than during Songkran 2025.

That is a measure of cases in the probation process, not a count of every driver who had alcohol, every police stop, or every accident involving alcohol. It is relevant to enforcement. It cannot, without more data, tell us the share of the 242 deaths caused by drink-driving.

What a careful reader can say

Songkran 2026’s seven-day campaign count was lower than the equivalent 2025 count in all three published headline categories. The reduction is real in those records. It is also too little evidence to declare a policy victory, explain a national road-safety trend or put a precise risk number on an individual journey.

The hard fact remains hard enough: in one defined week, 242 people died in reported road incidents. For anyone travelling, the useful response is prosaic rather than mystical — wear the relevant safety equipment, do not drive after drinking, and allow more time than the itinerary says you need. Roads are not improved by pretending a festival headline is fate.

For the larger question of why Songkran moves so many people at once, and what that mass travel says about Thailand’s geography and families, read The seven dangerous days. This page is deliberately the ledger, not the essay.