A COMPARATIVE ANALYSIS OF RANDOM UTILITY MODEL AND CONTINGENT VALUATION ESTIMATES OF NON-TIMBER FOREST PRODUCTS’ INCOME IN SOUTH NANDI, KENYA

Abstract

Non-timber forest products (NTFPs) provide a variety of economic benefits to forest-adjacent communities in rural economies. However, forest products exhibit non-competing consumption and lack market-driven excludability, subjecting them to open access degradation. To identify market inefficiencies and propose better policies, this study compares revealed and stated estimates of non-timber forest products’ incomes in South Nandi, Nandi County of western Kenya. Using survey data of 224 households located within a four km radius of South Nandi forest, income diversification strategies are examined by applying the inverse Simpson index of diversity. Regression analysis is used to evaluate the factors that influence forest income dependency. Convergence and criterion validity tests are employed to compare the economic values obtained from random utility model and contingent valuation estimates. The benefits of NTFPs in the rural household welfare, and the local economy of South Nandi are assessed using compensating variation and Gini coefficient analyses. The results show that forest-fringe households use off-farm and forest portfolio of activities in constructing income diversification programs. The income index of diversity measure shows the farm income stream as financially secure and stable. The NTFP income had a strongly equalizing effect (Gini reduction of 9%) contributing 24% of the household income while farm and off-farm incomes contribute 54% and 22%, respectively. The compensating variation analysis demonstrates that when disposable income reduces marginally, households with low education fall back on the local environmental resources than those with high education and the households are not willing to use off-farm incomes to compensate for a reduction in NTFP incomes. The findings of Random utility model (RUM) and Contingent valuation method (CVM) preferences reveal significant correlation (0.611, p<0.01) and confirm similarity of the theoretical construct of the two approaches. These methods, however, yield significantly different willingness to pay mean values (CVM had KSh. 4,033 and RUM KSh. 68,261, p<0.05). CVM construct validity tests reveal a systematic relationship between key socio-economic variables (farmland size, forest proximity, off-farm income, number of household cattle) and the households’ willingness to pay (WTP), which demonstrate that contingent bids are consistent with theoretical expectations. Hence, the difference in the households’ stated and revealed preferences amounts cannot be adequately explained on the basis of economic theory. The main reasons of the difference could be that the CVM bids were capped by the low forest permit fees as demonstrated by the strong correlation between Maximum WTP values and NTFPs user fees charges (0.602, p<0.01). Lack of pricing of most forest products and the weak enforcement of institutional arrangements governing use of forest resources affect availability of environmental markets leading to undervaluation in the contingent market. The study concludes that the RUM and CVM preferences converge in the predictors for the WTP bids but diverge in the valuation estimates. The RUM is more effective in valuing NTFPs, but the CVM bid function give qualitative information that can be used to support design and planning of forestry programs that incorporate improved rural incomes for sustainable management of forest resources. The revealed and stated preferences’ convergence analyses can be used to reconcile and determine economic values of environmental commodities as well as understand underlying drivers of the rational economic decisions in both the expressed and contingent markets. Restricting access of the rural poor to NTFPs might have effect of increasing income inequalities with substantial impacts on households’ welfare. Reliable compensating welfare analysis can guide policy makers in determining compensation measures in situations where forest resources become unavailable to the forest dependent households. The study recommends the establishment of an independently managed Forest Trust Fund at the county level to finance regulation and management of South Nandi forest.