Do Publication Metrics Distort Research Effort? Bunching Evidence from Thailand’s 2019 Higher-education Reforms

Many universities in middle-income countries lack the peer-review infrastructure to assess research quality directly and instead tie financial rewards to publication in journals classified as Q1 under the SCImago Journal Rank system. By converting continuously varying journal quality into discrete institutional categories, these systems create sharp incentive discontinuities at quartile boundaries. In a preregistered study, we apply a bunching estimator — a method from public finance that detects excess concentration of observations around institutionally salient thresholds — to 149,402 Scopus-indexed publications from Thailand over 2016–2025, exploiting Thailand's 2019 higher-education reform as a source of temporal variation. We find no significant bunching before 2019 but substantial excess concentration immediately above the Q1 boundary afterwards — a pattern not observed in Singapore, whose publication environment is not organised around explicit quartile-based financial rewards. The post-reform excess mass corresponds to roughly 1,575 additional publications over 2020–2025, implying an estimated 39 million THB (approximately US$1.1 million) in cumulative institutional expenditure. These findings indicate that discrete quartile-based reward systems redirect research effort towards threshold optimisation rather than research quality, and that replacing binary quartile rewards with continuous percentile-based incentives would better align institutional evaluation with scientific output.