NESDC raised Thailand's growth forecast on Monday to between 2.0 and 2.5%, and lifted the 2026 export growth projection from 9.6% to 15.1%. Budget season is coming soon, so that number is about to be reflected in a lot of plans for next year.

We were benchmarking off the wrong number

The example I remember is a business I ran here that used to set its top line growth targets off GDP and inflation projections. The problem I saw in Thailand is that it didn't reflect how fast construction was moving, because a lot of that was tied to financing for the projects, based on the speed of condo unit order bookings, or commercial leases getting signed. There were years the market was running well ahead of what they were talking about, and sometimes well behind. All of this was tied to mostly political issues in the country as well as what happened during COVID.

What we were trying to use as the benchmark instead was the data we had from the general contractors, and our own survey of the projects actually running in and around the major cities, combining that with our historical win rate, and the additional market share we believed we could get. It looked nothing like what the GDP forecast numbers were showing us.

I went back and checked whether we had a point

I went back and checked whether we had a point. The clean data all sits after my time there, but the pattern hasn't changed. Over eleven years the government's first forecast for the year never left a band between 2.8 and 4.1%, while construction ran anywhere from minus 3 to +17%. In 2015 the forecast was 4 and construction grew 17%. In 2017 the economy beat its own forecast at 4.2% and construction still shrank 3%. Three of those eleven years the forecast called for growth and construction contracted.

ELEVEN YEARS, PERCENT 0 Government first forecast 2.8 to 4.1 Actual construction growth -3 +17 Three of the eleven years the forecast called for growth and construction contracted. Source: the eleven-year comparison in the article. Forecast band 1.3 points wide, construction moved across 20.

The problem was the process methodology being applied by headquarters to all regions comparatively as the benchmark, which unfortunately was never built to describe our market. We lobbied with data to try to modify the process, but in the end, we were overruled. This was a global standard HQ process that they weren't willing to change.