Maximizing Benefits for Women: A Charitable Donation Allocation Problem

Charitable foundations should endeavor to allocate their limited resources to best serve their constituents. However, few foundations use mathematical programming techniques despite overwhelming evidence of their superiority at selecting projects that yield higher levels of total benefits. The Fund for Women, a Delaware foundation that makes grants to programs serving women, is a notable exception to this pattern as they have begun using a novel “Hybrid Selection Model” that combines both binary linear programming and the heuristic rank-based model. Using data from the foundation, this study shows how the rank-based selection model that was previously used by this group, and currently in use by most foundations, yields lower levels of aggregate benefits compared to binary linear programming or goal programming. Using historical data from 2010, this research shows that a Hybrid model would have selected the top three ‘signature’ projects can maintain an above average project benefits while also securing a 180% improvement in the number of projects funded, 66% improvement in the number of women served, and a 139% improvement in total benefits achieved. The Fund for Women incorporated the Hybrid model in their selection process in 2012 and this paper describes the benefits achieved and the challenges with adopting this approach in a foundation context, including educating and achieving consensus amongst the selection committee and individual member’s project selection preferences that were outside of the initial model’s objective function.