Nothing lasts long enough to build

Thailand has had twenty constitutions and charters since 1932. The current one, promulgated in 2017, is the twentieth.

The average lifespan of the previous nineteen was four years and three months.

Sit with that for a second, because the comparisons are unflattering in a way that’s genuinely useful. Four years and three months is shorter than a British parliament. It’s shorter than a typical mortgage fix. It’s about the length of a degree with a year abroad. It is less than the time it takes to train a specialist doctor, which will matter shortly.

Why it happens, mechanically

The cycle is documented and it is almost elegantly circular.

Removing a government by force violates the constitution in effect. So the constitution in effect is abolished. A drafting committee is then appointed, and a new constitution is written.

Thailand has had thirteen coups since 1932 by the most commonly cited count — sources vary between twelve and thirteen depending on how certain events are classified, and more if you include attempts that failed. Between 1932 and 1991 there were ten, of which eight produced a new constitution.

So the rewriting isn’t incidental to the political cycle. It is the mechanism by which the cycle completes itself.

And that produces the structural problem, which has nothing to do with any particular government and everything to do with the design.

Each constitution is drafted by the winner of the most recent dispute. Which means the document that is supposed to be the neutral referee is, every time, written by one of the players. A referee appointed by the team that just won the match is not a referee, whatever is printed on the shirt, and no amount of drafting quality fixes that.

What a constitution is actually for

The entire point of a constitution — the only thing that distinguishes it from ordinary legislation — is that it is harder to change.

That’s it. That’s the whole technology. A constitution is a set of rules deliberately placed beyond the reach of whoever currently holds power, so that long-term arrangements can be built on top of them without being renegotiated every time the government changes.

A document replaced on average every four years and three months is not doing that job. It has the formatting of a constitution and the function of a policy paper.

You can see the loss of confidence in the text itself. The 2017 constitution has 279 articles, and permanent Thai constitutions have grown steadily longer — the recent ones running well past two hundred. Post-coup interim charters, by contrast, often ran to a few dozen.

Length is a proxy for distrust. You write everything down in exhaustive detail when you do not trust the institutions to interpret a general principle in good faith. A short constitution is a statement of confidence in the people who will apply it. A 279-article one is a statement of something else.

The bit that connects to everything else on this site

Here is why I think this is the second structural article, alongside article 65.

Every serious problem this publication has documented requires a planning horizon longer than four years and three months.

Article 38’s ฿600 monthly pension. Building a genuine old-age income system is a thirty-year project — contributions accumulate, entitlements mature, the fiscal cost lands decades after the political cost.

Article 67’s arithmetic. Katanyu carrying an ageing population with fertility at 1.18 is a generational problem that needs a generational response.

Article 73’s doctor shortage. The training pipeline for a specialist is over a decade. Longer than the average constitution.

Article 65’s decentralisation. The standard prescription for urban primacy is sustained investment in secondary cities over decades, against the immediate returns of investing in Bangkok.

Article 33’s ageing farmers, article 40’s migrant labour dependency, article 09’s fertility rate. All of them slow, all of them expensive early and rewarding late, all of them requiring an institution to keep going after the people who started it have gone.

None of these are things a country can do on a four-year cycle, and it is not because Thai policymakers lack the ideas. The ideas are all published, in Thai, in considerable detail. What’s missing is a structure that outlives the argument.

Article 61 asked why so many public functions here ended up delivered by private companies, and concluded that they filled gaps where the state didn’t build. This is a substantial part of why the state didn’t build. 7-Eleven has had a more stable operating framework than the constitution has, which is a strange sentence and I believe it’s true.

The fair counter-argument

I don’t want to overstate this, because the raw average flatters my argument and I should say so.

The nineteen-constitution average is dragged down by short interim charters — brief documents issued to bridge a period, never intended to last. Counting them alongside the substantial ones inflates the instability.

And the recent rhythm is slower. The last three permanent constitutions arrived in 1997, 2007 and 2017 — roughly a decade apart. The current one is now nine years old, more than double the historical average, and by the standards of the series it has been durable.

Whether that represents genuine institutional settling or a longer swing of the same pendulum is a live question and I’m not going to pretend to answer it. But a fair reading is that Thailand’s constitutional churn has slowed, and an article that only quotes the four-year figure is telling you something true in a misleading way.

The structural point survives the correction, though. Even a ten-year cycle is short against a pension system, a training pipeline or a decentralisation programme, and the deeper issue isn’t the interval. It’s that the document is authored by whoever prevailed most recently, which is the thing that stops it becoming a settlement rather than a scoreline.

Common misconceptions

“Thailand is politically unstable.” Thailand’s governments change in the ordinary way. It’s the constitutional layer that turns over, which is unusual and much more consequential.

“Twenty constitutions means twenty revolutions.” Many were short interim charters. The count is real and the drama is less than it sounds.

“Nothing has changed since 1932.” The recent cycle is roughly decadal rather than four-yearly, which is a meaningful difference.

“It’s a cultural attitude to rules.” It’s a documented procedural mechanism: removing a government breaches the constitution, so the constitution is replaced.

“A better-drafted constitution would fix it.” The 2017 text runs to 279 articles. Drafting quality isn’t the binding constraint.