The pollution that follows the ballot: how immigration politics distorts industrial enforcement

At a glance, the 2026 working paper from KAIST and Singapore Management University reads like environmental economics boilerplate. Drill into the regression tables, and a more awkward finding emerges: in South Korean districts where immigration becomes politically salient, regulated industrial polluters emit more of the pollutants that the same regulators are supposed to police.

The result is not about immigrants themselves. It is about what happens to the administrative state when a fast-moving political issue pulls oxygen away from a slow-moving one. Hazardous-air-pollutant releases from facilities in "high-salience" districts rose relative to counterparts elsewhere, after the researchers accounted for plant size, sector mix, baseline emissions and the usual seasonal controls. The authors frame this as a shift in enforcement intensity, not a change in factory behaviour per se — but the public-health arithmetic is the same either way.

What the researchers actually measured

The team, led by Narae Lee of KAIST's School of Business and Technology Management and Heli Wang of Singapore Management University, paired district-level data on the Korean National Assembly's immigration-related agenda items with the Toxic Release Inventory-style disclosures that South Korean facilities are required to file. The dependent variable is emissions of regulated hazardous air pollutants — not total output, not jobs, not revenues. The independent variable is a constructed "issue salience" measure built from legislative questions, committee hearings and parliamentary records tagged to immigration.

The headline coefficient is not the only thing worth noting. The effect appears concentrated among facilities whose emissions are most sensitive to local inspector behaviour — exactly the slice of the economy where discretion matters and where a stretched bureaucracy shows its seams first. Plants that are subject to continuous-emissions monitoring, by contrast, do not move much. The implication is uncomfortable for environmental regulators who treat monitoring technology as a substitute for staffing.

The political-economy reading

There is a long literature — older than any single country's debate — on how politicians allocate the scarce resource of bureaucratic attention. Immigration policy in South Korea in 2026 sits inside a broader regional pattern: labour inflows have accelerated as the country's working-age population contracts, and parliamentary interest has followed. The Korean political system, like most, has finite bandwidth. When one file rises, others do not get the same number of phone calls.

The structural argument is straightforward. Environmental enforcement is a localised, labour-intensive public good. It depends on inspectors showing up, on complaints being logged, on follow-up letters being written and, occasionally, on fines being levied that hold up in administrative court. None of that scales automatically with the size of the regulated sector. When a politically charged issue captures agenda space, the political returns to visible immigration enforcement rise, and the political returns to quiet environmental enforcement — work whose beneficiaries are diffuse and whose costs are concentrated — fall.

This is not a story unique to Korea. The same pattern appears in studies of US environmental enforcement under divided government, in Brazilian state-level inspections during election cycles, and in Chinese prefectural data on pollution control around politically sensitive moments. The KAIST–SMU paper's contribution is to show that the channel operates even in a high-capacity bureaucracy with relatively strong rule of law — a useful counter to the assumption that better institutions inoculate a state against the trade-off.

What the paper does not yet settle

Three caveats deserve space. First, salience is endogenous: districts that already pollute heavily may attract more immigration-related political activity for reasons unrelated to immigration itself. The authors attempt controls for plant vintage, ownership and pre-treatment emissions trends, but the standard concerns about omitted variables in observational environmental work do not vanish.

Second, the emissions measure captures reported releases, not ambient air quality. If plants also adjust their disclosure behaviour when they expect less follow-up, the finding understates the true enforcement gap. The authors flag this directly and call for future work linking the salience measure to independent air-monitoring data.

Third, the paper is one study in one country over one window. South Korea's regulatory architecture is unusually transparent by regional standards, which makes the data usable but also limits how far the result travels. The authors are careful not to over-claim; the reader should be too.

Why the finding travels

If the KAIST–SMU result holds up under replication, it changes how industrial policy and immigration policy should be drafted in the same room. Ministries that run environmental inspectorates rarely coordinate with ministries that handle migration; both report upward through different chains. The paper implies that the cost of treating them as separate portfolios is paid in air that nearby residents breathe.

For industrial economies in East Asia and Europe facing the same demographic arithmetic — shrinking working-age populations, rising reliance on foreign labour, and aging plant stocks that need either retrofitting or replacement — the practical lesson is unglamorous: enforcement budgets need explicit counter-cyclical protection against agenda capture. That is a fiscal commitment, not a slogan.

The uncomfortable corollary is that the public does not see environmental non-enforcement the way it sees a visible immigration raid. The harm is delayed, diffuse and contested. That asymmetry is precisely why the political system under-supplies it — and why a paper measuring the gap is, in its own quiet way, a small act of accountability.

This publication framed the KAIST–SMU working paper as a study of bureaucratic attention under political pressure, emphasising the enforcement channel rather than immigration itself; coverage in Korean business outlets has tended to foreground the demographic backdrop, which the underlying data does not directly test.