Another significant variable, and one that is negatively related to farmers’ earnings, was whether the CSA is organized with a direct connection between the farmer and the members, meaning that a farm partner personally handles communication with the members. This direct connection is significantly, negatively related to earnings (at the 10 percent level). Personally knowing members well serves the goal of building social relationships, but it also enhances farmers’ sense of obligation. In contrast, CSAs that hire someone to handle member relations are more likely to have higher farmers’ earnings. Having a person in this position shields the farmer from a large workload, can decrease the sense of obligation that can be detrimental to the farmer making a living, and/or serve as a third party to look after the farmer's economic interests even if the farmer neglects them.
The interviews revealed that the obligation and loyalty felt by the farmer toward her or his members often negates the original CSA equity relationship. Many farmers mentioned feeling pressure to provide a “normal” share, regardless of what happens during the growing season. Many farmers conform to this pressure by supplementing their boxes with produce purchased from nearby farmers, often when production is lean or not diverse (13 percent did so all the time and 44 percent did so sometimes). They feel a responsibility to provide their members with what they think members expect, which many farmers described as consistency—of both quantity and diversity from week to week. In this way, the social embeddedness of the CSA relationship perversely acts to negate the reciprocal nature of the equity relationship by prompting farmers to engage in preemptive self-exploitation. CSAs that were run like this, which was common in the study region, share the bounty in times of plenty, but are not reciprocal when production is low or homogeneous—the farmers are not sharing the risks of production, but rather taking a self-inflicted economic hit, one hidden from members since it is not communicated as such. The regression model supports this interpretation that supplementing boxes means lower earnings. The “share_indicator” variable stands for CSA farmers purchasing from other farms to supplement their CSA boxes, and was coded categorically (with 0 = never, 0.5 = sometimes, and 1 = always, meaning that the higher the number, the more likely it is that farmers are spending their own money to supplement their boxes). This variable is negatively related to earnings, although the relationship is not significant at the 10 percent level.