Linguistic Mismatch and Learning Productivity: Evidence from Mother Tongue-based Education in the Philippines

The study examines how linguistic mismatch between students’ home languages and the language of instruction shapes learning productivity in linguistically diverse societies. A framework is developed in which linguistic distance acts as a multiplicative tax on educational production, and welfare-theoretic conditions for optimal language choice are derived. The framework is applied to Philippine elementary schools during the 2009–2024 mother tongue-based multilingual education (MTB-MLE) reform, exploiting policy-induced variation in linguistic alignment using difference-in-differences and shift-share instrumental variables designs. The Foster-Greer-Thorbecke family of deprivation measures, traditionally used to quantify income poverty, is extended to network settings in order to capture spatial variations in the degree of linguistic mismatch and estimate substantial school-based linguistic deprivation under a counterfactual pre-reform bilingual language regime, concentrated in non-Tagalog-predominant regions. Each unit of mismatch is found to reduce achievement by 35-57 percent of potential, depending on the curricular subject, yielding productivity losses comparable in magnitude to canonical estimates of spatial misallocation in firm production. Under a counterfactual pre-reform bilingual regime, an average student would have operated at 27–45 percent below learning potential, with losses concentrated in linguistically diverse regions. MTB-MLE reduced these losses by roughly two-thirds nationally, though systematic deviations from optimal language choice leave substantial efficiency gains unrealized. The reform also increased student continuation rates by 9-12 percentage points but temporarily strained classroom capacity, reflecting general-equilibrium resource pressures when demand-side constraints are relaxed. Overall, the results demonstrate that language policy is a first-order determinant of educational efficiency in multilingual settings and that measuring the quality of linguistic matching, rather than policy adoption alone, is central to evaluating language-in-education reforms. Comments to this paper are welcome within 60 days from the date of posting. Email publications@pids.gov.ph.