Understanding and predicting socioeconomic determinants of deforestation in Vietnam's Central Annamites Landscape (CAL): Pilot study implementing a spatial econometric approach

The forests of the Greater Mekong Subregion, consisting of Myanmar, Thailand, Cambodia, Laos and Vietnam, are under high pressure from economic development and exploitation of natural resources, including but not limited to land concession, smallholder plantations and commercial agriculture, agroforestry development, mining, and road infrastructure development. While these threats are well-known, the magnitude and dynamics of their individual and interacting effects on forest cover are not fully understood. This pilot study aims to apply existing, publicly available macro, micro, and socioeconomic data in addition to remote sensing forest cover data to explore economic determinants of deforestation in the Greater Mekong, using the case of the Central Annamites Landscape (CAL) ecoregion of Vietnam. A longitudinal panel was constructed for 2000-2017, containing 1,658 observations for 144 variables across 95 Tier 2 (district) administrative units in CAL from 2000-2017 for modeling macroeconomic and socioeconomic conditions against annual tree cover loss aggregated to the Tier 2 administrative unit. The first phase of the study used tiered spatial regression analysis to correlatively identify which commodities, economic development activities, and social conditions have historically had the greatest effect on forest cover by magnitude in CAL. Based on first phase results, we selected a subset of these determinants for scenario modelling to predict possible deforestation outcomes given certain economic scenarios. The results, among others, indicate that, when spatially collocated, high poverty rates and smaller scale agricultural land conversion are key immediate determinants of deforestation. This therefore provides evidence to support programs targeting rubber, acacia harvesting, artisanal mining, and land conversion for cash crop plantations. Education is also key immediate socioeconomic factor, as poverty rate is consistently associated with tree cover loss and increased educational attainment is consistently associated with reduced tree cover loss, at approximately 12 ha per percentage increase in secondary school graduation. On the macro level, economic growth in China and Vietnam are correlatively associated with tree cover loss, as are rising trade in Myanmar and Laos. This study provides a methodological contribution to the current academic literature identifying socioeconomic dimensions of deforestation through spatial econometric analysis and scenario modelling at the landscape level. For practitioner work, this pilot provides a model or tool for strategic planning of conservation interventions in light of economic conditions and factors.