Published online by Cambridge University Press: 24 May 2017
Tropical deforestation has been one of the most significant global environmental changes of recent decades and has spurred a substantial academic literature. Despite this, the empirical literature provides inconsistent evidence on the role of basic economic determinants of deforestation such as income and population. The authors suggest that one explanation for this inconsistency is a tendency to use incomplete or incorrect dynamic regression specifications. They derive three plausible dynamic empirical model specifications and compare their performance both with each other and with an ad hoc model typical of those used in the literature. Statistical tests support the use of more completely specified models which allow for dynamic adjustment of forest cover.