Colonial policy, ecological transformations, and agricultural “improvement”: comparing agricultural yields and expansion in the Spanish and U.S. Philippines, 1870–1925 CE

The study investigates how colonial policies and global market integration transformed land use, production, and yields of major Philippine crops under late Spanish (circa 1870–1898) and early U.S. (1902–1925) rule. It addresses whether colonial ambitions of agricultural “improvement” (i.e., increasing production per hectare/yield by reducing perceived “waste”) were realized, and how these changes varied regionally. The Philippines, with diverse ecologies and socio-political contexts across 7100+ islands and two colonial regimes, provides a critical case to quantify landscape transformation. Motivated by the need to translate archival records into quantitative datasets usable in ecological sciences, the paper aims to characterise national and provincial trends in land use, production, and yields, and to situate these within broader Anthropocene debates concerning tropical agro-ecosystem change. The study’s importance lies in bridging historical archives with Earth system science, enabling more precise reconstructions of human-environment interactions and evaluating the efficacy of colonial agricultural policies.

The paper situates its inquiry within scholarship on colonial environmental change and global commodity expansion, highlighting deforestation, biotic exchange, and reconfigured land-use systems under European imperialism in the tropics. Prior works describe the rise of statistical governance and factory-inspired notions of productivity underpinning colonial notions of “improvement,” and the spread of “scientific agriculture.” Philippine-focused historiography documents the expansion of export crops (sugar, abacá, tobacco, copra) and infrastructure during both Spanish and U.S. regimes, with U.S. policies often framed as better financed and more systematically geared toward agricultural reform (irrigation, roads, railroads, land redistribution, agricultural banks, and extension). Existing literature suggests late 19th-century export growth was driven primarily by area expansion rather than increased yields, and notes the uneven geographic and socio-political impacts of agricultural change across provinces. However, comprehensive, quantified provincial comparisons of land use, output, and yields across the two periods have been limited, especially due to data quality and consistency issues for the Spanish period.

Data sources: The study compiles quantitative agricultural data from the U.S.-era Censuses of 1903 and 1918 (HathiTrust) and the Joseph Ralston Hayden (JRH) Papers (Division of Farm Statistics, Bureau of Agriculture) covering national data for 1903, 1910, 1915, and 1920–1925, and provincial data for 1925. Spanish-period information includes the 1876 civic census by Cavada and Spanish-era manuals, with export data for 1870–1898 reproduced in the 1903 Census. Data preparation: All tables were transcribed into Excel and released as supplementary datasets. To minimize inconsistencies in the 1918 Census, only comparative tables aligning 1903 and 1918 provincial data were used. Crops analyzed had to be reported at three timepoints (1903, 1918, 1925), convertible to kilograms, and exceed 5000 ha (except coffee and cacao). Excluded crops include bananas, cotton, mandarins, oranges, cassava, ube, gabi, mango, maguey, lumbang, peanuts, castor bean, kapok, pomelo, pineapple, papaya, pili nuts, lanzones, tugui, rubber; kamote was excluded due to under-reporting in 1903. Production units reported as hectolitres/litres were converted to kilograms using factors from JRH papers. For multi-product crops, only solid products in kg were used: palay (rice), granulated sugar (sugarcane), abacá fibre, tobacco, corn, copra (coconut), cacao, coffee. Boundary harmonisation followed the 1918 comparative tables; sub-province and city totals were combined into provinces as needed. Mountain Province was excluded entirely; 1903 data were missing for Batanes, Lanao, Bukidnon, Agusan. Minor boundary shifts in several Luzon and Mindanao provinces remained unresolved but did not preclude comparison. Analysis: For national and provincial scales, productive yields (kg/ha) were computed from land use and production. Average annual growth rates of land area and production were calculated for 1903–1918 and 1918–1925. Provincial production-vs-land-use plots with linear regressions assessed yield trends across timepoints. Spanish-period quantitative comparisons relied on export data (sugar, abacá, tobacco, coffee, indigo, tintarrón) and qualitative analysis of farmers’ manuals (Spanish and U.S.) to infer practices and likely yield dynamics. Map-making used a georeferenced 1929 John Bach map (QGIS, 2nd order polynomial) to visualize provincial land use. Qualitative sources: Thematic analysis of Farmers’ Bulletin volumes and Spanish manuals compared paradigms, problem framings, and prescriptions (fertilizers, mechanization, crop choice, irrigation, plantation vs smallholder dynamics) across regimes.

- National land use: Reported cultivated land expanded from ~1.2 million ha (1903) to ~3.6–3.7 million ha (1921–1925), tripling by 1921. The six dominant crops by area were rice, sugarcane, corn, coconut trees, abacá, and tobacco. From 1903 to 1925: rice +191% area, corn +218%, coconut +233%, sugarcane +119%, abacá +384%, tobacco +128%. Coffee and cacao occupied <0.2% of land and declined in area (coffee −7%, cacao −60%). - National production (all eight products in kg increased, 1903–1925): palay +298%, corn +377%, copra +746%, sugar +292%, abacá +170%, tobacco +146%, coffee +1270%, cacao +142%. - 1918 anomalies: 1918 was exceptional for several products, showing peak palay, corn, and abacá yields and troughs in sugar and cacao yields, likely due to disease (e.g., sugarcane mosaic), World War I market conditions, natural hazards, or data collection differences. - National yields (excluding 1918 clarifies trends): • Palay: yield increased ~37%; post-1920 yields ~1052–1164 kg/ha. • Copra: yield increased ~166%; from 289 kg/ha (1903) to ~767–903 kg/ha (1920–1925), largely reflecting a shift from nut to copra production rather than agronomic change. • Coffee: yield increased ~1365% (83 kg/ha to ~1261–1323 kg/ha by 1920–1925), largely recovery from blight rather than policy effect. • Cacao: yield increased ~509% (from ~130 kg/ha to ~775–859 kg/ha by early 1920s). • Corn: no sustained increase; fluctuated and was ~1% lower in 1925 than 1903. • Tobacco and abacá: yields largely stable within narrow bands (tobacco ~500–601 kg/ha, abacá ~306–407 kg/ha outside brief dips). • Sugar: volatile with disease impacts; no clear sustained yield increase before a high in 1925 (2953 kg/ha) after earlier ranges of ~1897–2507 kg/ha. - Provincial patterns: All provinces saw increases in land use and production, but growth rates were uneven. Mindanao provinces (e.g., Davao) had the highest average annual land growth (e.g., Davao ~43% for 1903–1918) from small bases; Luzon, Panay, and Cebu had the largest total areas but lower growth rates. Increasing provincial specialization meant national trends could be skewed by a few provinces. Example: Pangasinan in 1918 produced ~594 million kg of palay from ~122,000 ha (yield ~4872 kg/ha), constituting ~33% of national palay output and greatly elevating the national yield for that year. - Spanish-period exports (1870–1898): Exports (sugar, abacá, tobacco, coffee, indigo, tintarrón) generally rose to 1890s; sugar and abacá peaked in 1895 (~341 and ~107 million kg, respectively), while coffee collapsed due to blight. Historical evidence indicates late Spanish export growth was driven largely by area expansion rather than yield gains. - Manuals and policy continuity: Spanish and U.S. manuals shared paradigms (scientific agriculture, anti-shifting cultivation, fertilizer promotion, mechanization where possible). Despite greater U.S. investments (irrigation, banks, mills, education), most crop yields did not meaningfully increase, suggesting policies mainly expanded area and market integration rather than improving per-hectare productivity.

The findings demonstrate that colonial aspirations to increase agricultural yields—the core metric of “improvement”—were largely unrealized. While total cultivated area and production grew dramatically, yields for most major products remained flat or only modestly improved. The exceptions (palay, copra, coffee, cacao) reflect crop-specific factors: irrigation expansion for wet rice; accounting and product-shift effects for copra; and disease recovery for coffee, with cacao showing gains whose mechanisms remain uncertain. Provincial analyses reveal strong linear production–area relationships and localized exceptional years that can distort national trends, emphasizing the need to interpret national statistics through fine-grained regional dynamics. These results indicate that colonial interventions—transport infrastructure, credit, land redistribution, and agricultural extension—primarily facilitated horizontal expansion rather than intensification. The ecological implications include large-scale conversion of forests and wetlands, shortened fallow periods, erosion, and biodiversity loss. Divergent regional outcomes (e.g., Negros Occidental’s plantation sugar vs. Pampanga’s smallholder mix) underscore the roles of labor regimes, land ownership, and local socio-political geographies in mediating environmental impacts. The work links quantified historical land-use changes to ecological processes, providing a foundation for cross-disciplinary research on how colonial-era agricultural strategies shaped technosphere–biosphere interactions and Anthropocene trajectories in tropical settings.

The study provides the first integrated, quantitative national and provincial assessment of Philippine agricultural land use, production, and yields from 1903 to 1925, contextualized with late Spanish-era exports and agronomic manuals. It shows that while cultivated area and output expanded substantially under both colonial regimes, yields for most major crops did not meaningfully rise, calling into question the efficacy of colonial strategies aimed at per-hectare productivity gains. Notable yield increases in palay and cacao, and apparent gains in copra and coffee, stem from irrigation expansion, accounting/product shifts, and disease recovery rather than broad-based intensification. The paper contributes a digitized, harmonized dataset suitable for ecological modeling and historical–environmental synthesis, highlighting pronounced regional specialization and variability in outcomes. Future research should: (1) undertake comparative studies across the Global South to evaluate colonial agricultural policies and environmental legacies; (2) integrate historic land-use modeling with ecological proxies to trace local-scale impacts (deforestation, wetland fragmentation, erosion, biodiversity change); (3) focus on rice-producing regions to explain palay yield increases; and (4) link labor regimes and land tenure to ecological transformations, refining our understanding of diverse Anthropocene pathways.

- Spanish-period quantification: Lack of directly comparable, reliable Spanish-era land-use and production data precludes direct yield comparisons; reliance on export records and qualitative manuals limits inference. - Data quality and harmonization: Early censuses contain inconsistencies; 1918 data may reflect methodological differences. Unit conversions (hectolitres/litres to kg) and product selection (solid-form outputs) can introduce uncertainty. - Spatial issues: Mountain Province excluded due to boundary changes; 1903 data missing for Batanes, Lanao, Bukidnon, Agusan. Minor unresolved boundary shifts in several provinces may affect provincial comparability. - Crop coverage: Several crops excluded due to missing years, insufficient area, or non-convertible units; kamote underreported in 1903. - Anomalous years and external shocks: 1918 yields influenced by disease, WWI market conditions, natural hazards; provincial extremes (e.g., Pangasinan palay) skew national figures. - Causality and attribution: Observed yield changes cannot be conclusively attributed to specific policies or practices (e.g., fertilizer adoption, mechanization) due to limited direct evidence of implementation and confounding factors (disease cycles, weather variability).