This study examines the welfare effects of community plantations in Ethiopia via Contingent Valuation. Both single-bounded and double-bounded survey methods were considered, and, with respect to double-bounded methods, the potential for anomalous response behaviour was also taken into account. The results generally confirm that there are statistically significant welfare benefits to be derived from community forestry; however, the range of the estimated benefits is large. After controlling for anomalous response behaviour, the range of estimated benefits narrows, and our preferred estimates place the welfare gain between Ethiopian Birr (ETB) 20.14 and 30.41 per household, which is much lower than the estimated benefits without controlling for anomalous preference responses.

We use inverse probability weighting to examine the effects of a unique two-pronged common-property forestry program in the Gimbo district of Ethiopia, which includes Joint Forestry Management and improved nontimber forest product marketing efforts. The program was found to have affected household access to agricultural land, and, thus, reduced livestock holdings, due to program strictures. Furthermore, despite those reductions, there is evidence that the program had economically significant effects on other activities. Households were able to increase their earnings from nontimber forest products, partly due to an increased labor allocation toward non-timber forest product collection Keywords: community forestry, treatment effects, matching and Ethiopia 1 We are grateful to Volkswagen Foundation/ZIAF for financing the postdoc fellowship of the first author. We are also grateful to EfD (EDRI/EEPFE) SIDA for generously availing the data used for analysis of this paper, and to Economic Research Southern Africa for their financial support. This research is that of the authors, and does not reflect the views of the supporting agencies. We acknowledge that we have received quite valuable comments from Abebe Damte, Chitalu Chama and Gauthier Tshiswaka-Kashalala. The usual declaimer applies. 2 ACKNOWLEDGEMENTSWe are grateful to Volkswagen Foundation/ZIAF for financing the postdoc fellowship of the first author, and Economic Research Southern Africa for its support of this research. We are also grateful to EfD (EDRI/EEPFE) SIDA for generously availing the data used for analysis of this paper. We acknowledge that we have received valuable comments from Abebe Damte, Chitalu Chama-Chiliba and Gauthier Tshiswaka-Kashalala. Finally, we thank the editor and three anonymous reviewers for their suggestions to improve our presentation and analysis.

Extant literature on Joint Forest Management (JFM) impact evaluation has concluded that it generally does not provide sufficient incentives to justify the costs that forest use restrictions impose on local people. However, there is a dearth of evidence concerning whether alternative JFM intervention with improved market linkages for non-timber forest products has similar implications. In this study, we evaluated the income and distributive effects of a JFM program in Ethiopia in which additional support was provided for improved market linkages for non-timber forest products (NTFPs). Exploiting exogenous variation in customary rights across eligible groups of communities that participate in JFM programs, as well as using heteroskedasticitybased instrumentations, we identified the income and distributive effects of the program. Our analysis shows that the program has raised the income of the households who chose to participate by approximately 400 Ethiopian Birr or 26% of per capita expenditure; that result was robust to various specifications. We also found that this effect is largely driven by marketing incentives to use non-timber forest products. However, we found that the program's benefit is * We thank associate editor, Professor Arild Angelsen, and the reviewer for their comments. We are grateful to Volkswagen Foundation and University of Johannesburg for funding postdoc fellowships of the first author and Economic Research Southern Africa for its support of this research. We also gratefully acknowledge generous financial support of EfD (EDRI/ EEPFE) SIDA for collection of the data used for analysis of this paper. 1 biased toward the upper end of the income distribution, a result that points to the inequalityreinforcing effects of the program.

Through the implementation of a choice experiment valuation exercise, this study set out to identify the set of community plantation attributes that impact the welfare of potential community forestry program participants. We employed a combination of choice models to evaluate the preferences, welfare impacts and choice elasticities associated with alternative community forestry programs, allowing for different assumptions regarding heterogeneity. In line with economic theory, increased participation costs reduced the demand for community forestry, while increases in expected productivity raised the demand. With respect to preferences for the other alternatives considered: type of forest, area enclosure and type of land upon which the forest was to be situated, the results point to significant differences in preferences across the study population, suggesting that programs should be tailored to the communities in which the program is to be implemented.Keywords: community forestry, choice experiment, conditional logit, random parameters logit and latent class model J.E.L Classification: Q23, Q28, Q51* We would like to thank Abebe Damte, Chitalu Chama, Alemu Mekonnen, Gauthier TshiswakaKashalala and University of Pretoria seminar participants for their helpful comments. We would also like to thank two anonymous reviewers for their contributions in improving this research. We would also like to thank Economic Research Southern Africa for their support of this research. However, all errors remain the responsibility of the authors.