Improving forest management for sustainable livelihoods: Measuring the performance of Malawi's forest co-management program

John Mazunda, Purdue University

Abstract

This thesis investigates the impacts of Malawi’s forest co-management program (FCM) on rural households’ behavior, their decisions and economic and environmental outcomes. We evaluate the contribution of the FCM to household income and to protection of forest resources. We also evaluate whether household maize sufficiency influences participation in the program. The study uses panel data collected in 2002 and 2009 from 218 households in Kasungu and Machinga Districts of Malawi. The program was implemented through forest user groups (FUGS) that were responsible for a set of activities related to setting, implementing, and enforcing rules governing forest management and use. To study the impacts of the FCM, a two stage regression approach is used. In the first stage, a multinomial logistic model (MNL) is estimated to identify characteristics of program participants. Generalized propensity scores are predicted following estimation of the MNL model. In the second stage, an ordinary least square regression model is used in which the inverse of the propensity scores serve as sampling weights to estimate the effects of the FCM on household forest and household total incomes. A difference-in-difference model (DID) is also used to measure the medium term effect of the FCM on household forest income. A weighted Probit model is used to identify factors correlated with observed decisions to clear forests. A weighted Tobit model is further employed to assess the effects of program participation on forest clearing. Results suggest that participation in the FCM program decreased household income derived from forests. The program had a negligible short term effect on overall household income and a similarly small and insignificant medium term effect on income derived from the sale of forest products. Maize sufficient households were found to be more likely to participate in the program. Results suggest that the program was successful at lowering the rates and extent of forest clearing. Households that participated in the FCM program were less likely to clear forests compared to non-participants. FCM participants cleared 0.25 hectares less forest land per household for agricultural expansion compared to their non-participating counterparts.

Degree

M.S.

Advisors

Shively, Purdue University.

Subject Area

Forestry|Agricultural economics|Sustainability

COinS