Economic dynamics and forest clearing: A spatial econometric analysis for Indonesia

This paper uses a large panel database to investigate the determinants of forest clearing in Indonesian kabupatens since 2005. Our study incorporates short-run changes in prices and demands for palm oil and wood products, as well as the exchange rate, the real interest rate, land-use zoning, forest protection, the estimated opportunity cost of forested land, the quality of local governance, the poverty rate, population density, the availability of communications infrastructure, transport cost, local rainfall and terrain slope.The econometric results highlight the role of dynamic economic factors in forest clearing: significant roles for lagged changes in all the short-run economic variables-product prices, demands, the exchange rate and the real interest rate-as well as communications infrastructure, some types of commercial zoning, rainfall and terrain slope. We find no significance for the other variables, although our measure of protected-area status is too general for strong conclusions to be drawn about the empirical relationship between forest clearing and protection. It is also possible that the insignificance of transport cost in our results is due to collinearity with our proxy for communications cost, and/or the offsetting effects of transport cost on the direct profitability of forest clearing (negative) and the effectiveness of local forest regulation (positive).Our results strongly support a model of forest clearing as an investment that is highly sensitive to expectations about future forest product prices and demands, as well as changes in the cost of capital (indexed by the real interest rate), the relative cost of local inputs (indexed by the exchange rate), and the cost of land clearing (indexed by local precipitation). By implication, the opportunity cost of forested land fluctuates widely with changes in international markets and decisions by Indonesia's financial authorities about the exchange and interest rates. Our results suggest that forest conservation programs are unlikely to succeed if they ignore such powerful forces.

► This paper uses a panel database to investigate the determinants of forest clearing in Indonesian kabupatens since 2005. ► High-resolution remote-sensing database of monthly forest clearing has been used. ► The model linked forest clearing to expected revenues from-and expected cost of-clearing or holding a parcel of land. ► The model was estimated using new spatial panel estimation methods that allow for temporal and spatial autocorrelation. ► Results highlight the role of international markets and Indonesia's exchange and interest rates in forest clearing.