Vincent de Leijster

136 Chapter 6 and A6-7). The coffee associations directly sold coffee to clients in high income countries, which, in combination with higher coffee quality, enabled them to pay higher prices ( personal communication with the management of Aso Cuchilla del San Juan, Aso Tatamá , Aso Jardín, Aso Cristalina, and Flor de Apía ). Most of the coffee associations had their own extension service providers and organized workshops, which may have given direction towards a more diversified farming approach among their members. A study in the Mexican coffee sector showed that the farmer’s choice of intermediary purchaser may affect farm operations, for example resulting in more labor to meet the intermediaries’ requirements (Milford, 2014). This choice may also affect the price, since different intermediaries may sell to different clients, have longer or shorter chains, and work with certificates or not (Milford, 2014; Millard, 2011) Our results show that there is variation in economic performance of coffee farms, which could be due to a few confounding factors and limitations of the study design. First, we applied space-for-time substitution, which has been used in a few studies to analyze the developments of economic and biophysical factors in agroforestry systems (Jain and Singh, 2000; Kattwinkel et al., 2011; Nijmeijer and Harmand, 2019). While the approach is effective in identifying strong trends, it is less sensitive to subtle changes over time as it is more influenced by variation in the landscape matrix and variability in farm characteristics (Rácz et al., 2013). Second, the potential net revenues were not analyzed over time, since the calculations of some potential revenues were based on tree or crop densities (tree fruits and Musa fruits), which did not develop over time (De Leijster et al. 2021). Finally, it is important to note that we studied flexible management costs, and did not include capital costs (Jezeer et al., 2018). The net revenues in our study are therefore not a true representation of the effective income of farmers. Agroforestry can be implemented to rehabilitate ecosystem services in coffee plantations, whiteout trade-in of net revenues. To maximize ecological and economic benefits further, dispersed arrangement of trees is recommended, then alley agroforestry and finally living fence agroforestry. Co-products (tree fruits, timber and carbon) provide an important contribution to net revenues and to improve net revenues of agroforestry farms furthermore co-products should be brought to market. Finally, choice of intermediary relates to the economic performance of farms, which can be acted upon by individual farmers. Programs to promote marketing of tree products and services may be organized in collaboration with farmers associations, to use this untapped potential for improving the economic performance of coffee agroforestry farms.