The FT reports from Longhui county in Hunan that China's 300mn-strong migrant labour force is increasingly staying put — a quiet reversal of four decades of rural-to-urban movement. A state-run labour agency in Longhui says that since 2023 job listings have fallen while the number of jobseekers has risen; the same recruiter says the county's sports-shoe factories now employ 200–300 people, down from close to 3,000 at peak, blaming the US–China trade dispute. A construction worker interviewed at a mah-jong parlour says monthly pay on Guangdong building sites has dropped from Rmb 15,000–16,000 at peak to Rmb 7,000–8,000 today. A marble tiler says a Dongguan project on pause could pay Rmb 20,000 a month, versus around Rmb 4,000 locally: "Whatever the pay is, I will take it."
The macro explanation is the combination of a multi-year property and construction slump with a structural shift away from labour-intensive manufacturing. Ernan Cui at Gavekal Dragonomics says urban labour markets are weakening "across construction, manufacturing, and services", and that new jobs are increasingly concentrated in AI-related sectors — creating a structural skills mismatch with displaced older workers. Jenny Chan at Hong Kong Polytechnic University says workers in their late forties through early sixties are being caught out by industrial upgrading and fall back on farming and odd jobs. Inter-provincial migration has been declining since 2015. Wages for those who still move are higher than for intra-provincial migrants, but in 2024 they grew more slowly.
The politics are starting to show. In November, China's Ministry of Agriculture and Rural Affairs held a work conference that explicitly flagged workers' "large-scale returning and staying in the countryside". Beijing has issued guidelines to boost jobs for migrants, including transport subsidies for inter-provincial travel. A local recruiter frames the risk bluntly: "Practically every village now has a number of stranded workers. In terms of social order, it's not great." Beijing does not publish annual data on returnees, but analysts believe the trend has accelerated since 2023. Source: Financial Times, 18 April 2026, William Langley (with additional reporting by Tina Hu).
The story in one line. China’s 300mn migrant workforce — the single largest labour flow in modern history — is slowing down: an ageing cohort is being priced out of AI-driven urban hiring, older industrial hubs are shrinking, and Beijing has begun flagging the risk to “social order” in internal meetings.
Key numbers
- Migrant workers in China: ~300 million.
- Inter-provincial migration: declining every year since 2015.
- Longhui county sports-shoe factories: 200–300 jobs now vs close to 3,000 at peak — blamed on US–China trade dispute.
- Guangdong construction wages: Rmb 15,000–16,000 at peak, 7,000–8,000 today (peer interviewee).
- Dongguan vs Longhui wage gap: Rmb 20,000/month urban, Rmb 4,000/month local.
- Since 2023: job listings falling, jobseeker count rising (state-run Longhui agency).
- November 2025: Ministry of Agriculture and Rural Affairs work conference warns about “large-scale returning and staying in the countryside”.
- Lunar new year 2026 public holiday: ended 23 February; many older migrants were still home weeks later.
Why it matters
This is the demand-side signature of China’s two-speed economy. The urban coastal machine that consumed Hunan labour for thirty years — electronics, apparel, construction, light manufacturing — is shrinking at the very moment the new urban growth engine (AI, advanced manufacturing, services) demands skills the displaced workers do not have. Ernan Cui’s observation that net hiring is declining “across construction, manufacturing, and services” is the key sentence: this is not a single-sector adjustment, it is a generalised slowdown in private-sector hiring. The 2015 turning point in inter-provincial migration suggests the reversal is structural, not cyclical; the US–China trade dispute has accelerated it but did not cause it.
The politics are the real signal. Beijing’s preferred internal framing — migrants “stranded” in villages, “social order” — is the same register used during the 2009 rural-return panic after the global financial crisis. The November work conference confirms the Communist Party sees this as a national stability file, not a local labour-market story. The policy response so far (transport subsidies, local job fairs) is small relative to the 300mn-person flow; the next step would be a larger fiscal transfer to rural and hinterland provinces, or a meaningful push on services-sector job creation.
Takeaway
The bearish read for 2026 is straightforward: weakening migrant-worker demand translates into weaker consumption in interior provinces and lower construction activity, both of which weigh on nominal GDP and on Beijing’s attempts to rebalance toward consumption. Watch for (1) further signals from the Politburo on rural employment, (2) data from coastal exporters on labour demand once Q2 trade numbers land, and (3) whether interior-province retail spending diverges further from Tier-1 city data. The AI/manufacturing upskilling gap is the structural story; in the short term, the US–China trade dispute remains the swing factor.
Source: Financial Times, 18 April 2026, William Langley (with additional reporting by Tina Hu).