Using village data from Tanzania, we test whether gifts and loans between households are voluntary while correcting for mis-reporting by the giving and receiving households. Two maintained assumptions underlie our analysis: answers to a question on who people would turn to for help are good proxies for willingness to link; and, conditional on regressors, the probability of reporting a gift or loan is independent between giving and receiving households. Building on these assumptions, we develop a new estimation methodology that corrects for response bias. Our testing strategy is based on the idea that, if lending and gift giving are voluntary, then both households should want to rely on each other for help. We find only weak evidence to support bilateral link formation. We do, however, find reasonably strong evidence to support unilateral link formation. Results suggest that if a household wishes to enter in a reciprocal relationship with someone who is sufficiently close socially and geographically, it can do so unilaterally.
