Vincent de Leijster

131 Chronosequence analysis of economic performance of agroforestry coffee farms in Colombia 6 6.3.2 Costs, benefits and net revenues and the factors that explain them Tree management and spatial arrangement We found that monoculture coffee farms had higher coffee gross and net revenues than agroforestry farms, and specifically higher than agroforestry farms with dispersed trees (Figure 6-3; appendix Figure A6-6). On the other hand, total actual net revenues and potential net revenues did not significantly differ between agroforestry and monoculture farms, nor between farms with different tree spatial arrangements. The value of non- marketed products and services (potential additional benefits) was significantly higher for agroforestry farms than for monoculture farms (R 2 = 0.073, F=11.5, P=9.1*10 -4 ), specifically for agroforestry farms with dispersed trees (R 2 =0.067, F=4.1, P=0.008). In the ‘realistic scenario’ we found that potential value of non-marketed timber and carbon was €83 ± 72 ha -1 y -1 , for the ‘minimum scenario’ this was €39 ± 34 ha -1 y -1 and in the ‘maximum scenario’ €198 ± 34 ha -1 y -1 (appendix section 6). We found that in all scenarios the combined value of non-marketed timber and carbon was lower than for either non-marketed tree-fruits (€272 ± 677 ha -1 y -1 ) and non-marketed Musa fruits (€542 ± 635 ha -1 y -1 ). Economic performance explained We found that pest control intensity, fertilization intensity, and farm size also significantly improved coffee gross revenues (Table 6-4). We also found a parabolic relationship between coffee gross revenues and time since pruning coffee plants (Table 6-4). Furthermore, we found that coffee sold to coffee associations resulted in higher coffee gross revenues compared to private intermediaries, and farmers also had higher realized net revenues than those farmers selling to the national coffee cooperative or to private intermediaries (Table 6-4 and appendix Figure A6-6). Further, we found that there were lower tree densities on larger farms (r=- 0.27, P<0.01) and at higher altitudes (r=-0.0.21, P<0.05), and that higher tree densities corresponded to longer time since pruning coffee plants (r=0.23, P<0.05). We found that pest control intensity was negatively related to altitude (r=-0.21, P<0.05) and positively related to time since pruning coffee plants (r=0.24, P<0.05). Further, we found a weak relationship between intermediary choice and canopy characteristics (Appendix Table A6-4). We found that gross coffee revenue was best explained by pest control intensity and altitude (R 2 =0.15, F=8.4, P=4.4*10 -5 ; Table 6-5), while net coffee revenue was best explained by altitude and farm size (R 2 =0.04, F=3.8, P=0.03) and actual net revenue was best explained by intermediary choice (R 2 =0.04, F=3.2, P=0.04). We found that the aggregated