Rural mobile-money agent density and women's petty-trade capitalisation in Ugandan sub-counties: an instrumental-variable estimation of proximity thresholds

Rural Ugandan women who trade in petty goods depend on mobile-money agents to convert working capital into cash and to hold value between market days, yet the density of agent networks varies sharply across sub-counties. The argument treats agent density as an endogenous feature of sub-county economies: agents locate where trade is already monetised, so naive comparisons of capitalisation across high- and low-density sub-counties confound the effect of proximity with the effect of market activity. It specifies a threshold model in which capitalisation responds to density only above a minimum number of agents per trading population, and it derives the conditions under which the instrument identifies that threshold rather than an average effect. The contribution is a design that makes the proximity threshold estimable with sub-county administrative data, together with an explicit statement of the assumptions under which the resulting estimates would be interpretable. The paper does not report empirical estimates; it specifies the data, the instrument, the threshold estimator and the falsification tests that a subsequent study would require.