Multinomial logit regression models help describe novel spatiotemporal dimensions in environmental policy enrollment patterns

JOSH HILLER; NUNEZ REGUEIRO MAURICIO MANUEL

San Diego

Congreso; Joint Mathematics Meetings; 2018

American Mathematical Society and Mathematical Association of America

Over the last 50 years, payment for ecosystem services schemes (PES) have been lauded as a market-based solution to simultaneously curtail deforestation while lifting local populations out of poverty. Nowhere in the world are these two goals more evident than in the Argentine Forest Law. This law compensates landowners for enrollment of their land in conservation-type projects on a scale seldom seen in the rest of the world. However, PES programs often fail to conserve sites under strong deforestation pressures and allocate limited financial resources to protecting sites that would likely be conserved in the absence of PES - a problem called adverse selection. Here, we analyze adverse selection patterns of enrollment in this program. We do this by utilizing multinomial logistic regression models in a novel way and examine several socio-economic variables. Among the most striking conclusions is that all large parcels of enrolled land show evidence of spatial or temporal adverse selection. Furthermore, conservation-type projects are less effective in avoiding adverse selection than more profitable approaches such as non-timber forest product production and silviculture. (Received September 26, 2017