Sri Lanka's labour depth is now its real scale ceiling
Originally published in Sunday Observer on 2026-08-16.
Read Original Article on Sunday ObserverCore Argument
On August 3 an official of the National Chamber of Exporters, a body representing around 700 companies, said publicly what members had been saying privately for a year: they cannot find or keep workers at any skill level, least of all in the specialised trades that export manufacturing and agriculture depend on. The Chamber described firms running at half the workforce they need. At that level a factory does not scale. It survives the month. The column argues this is not a hiring problem dressed up as a national one. Sri Lanka spent three years rebuilding order books and market access, and the binding constraint has moved. Capital is no longer the first thing to run out. People are. Over 143,000 Sri Lankans left for overseas employment during the year against a State placement target of 250,000, so the country is exporting the capability its export sector is short of. The numbers do not describe a shortage of people. Unemployment ran at 3.7 percent in the first quarter, but labour force participation was 49.2 percent, meaning barely half the working-age population is economically engaged. Youth unemployment stood at 16.1 percent, higher for young women, against more than 400,000 open private sector vacancies. That is a failure of architecture, not of supply. In boat building, a designated priority sector, attrition in fiberglass moulding and marine electrical roles has been reported near 80 percent. The three standard responses each buy a quarter and expand nothing. Wage rises move scarce technicians between competitors without adding one to the national pool. Short training programmes disconnected from any order book do not complete. A foreign worker window is a legitimate instrument but not a substitute for domestic skills formation. The break point sits at SME scale, where an order requiring a second shift hits the wall at roughly fifty to a hundred specialised roles, precisely the transition from small business to scaled business.
What I'd Revise Now
The Singapore comparison is the part most often read wrongly, including in how I framed it. Readers take the lesson to be the technical institutes. It was the levy. A calibrated, sector-by-sector price on foreign labour is what kept firms investing in productivity while the institutes did their slower work. Copy the training capacity without the price mechanism and you get subsidised access to cheap labour and no productivity pressure at all. I would lead with the levy now. The fifty-to-a-hundred threshold has held up, but I would add a qualifier: it is the number of specialised roles, not headcount, and firms consistently misjudge which of their roles are genuinely specialised. Most discover the constraint is narrower and deeper than they assumed, two or three functions rather than a department. The connection I did not draw is to import intensity. A firm deferring capital investment because of the currency squeeze and a firm deferring capacity because it cannot staff a second shift are making the same decision for different stated reasons, and the outcome is identical. Of the five moves, the one that has proven most reliable is training against a named contract with a delivery date. Programmes with a customer attached complete. Programmes without one do not.
Key Takeaways
- Attach a rupee value of orders at risk to your staffing gap; it becomes a board decision only when it carries a revenue number
- Redesign the process step consuming the scarcest skill before recruiting against it
- Tie every training programme to a named contract and delivery date; untethered programmes do not complete
- Fund shared technical pipelines with peer firms rather than competing over a pool nobody refills
- The break point is roughly fifty to a hundred specialised roles, not total headcount