Assessing the Targeting Efficiency of Clean Water Subsidies Under Uncertainty

Abstract

Access to affordable clean water remains a critical challenge for many communities in the United States, particularly those with limited fiscal capacity. This study evaluates the risk-return efficiency of Clean Water State Revolving Fund principal forgiveness in Tennessee from 2019 to 2022 using a Modern Portfolio Theory framework. The analysis explicitly incorporates uncertainty arising from year-to-year changes in community economic and demographic conditions and from potential shifts in how economic distress is weighted in state program priorities. Expected return is defined as the alignment with economic distress (represented by the inverse of a county’s affordability score), so higher returns correspond to greater socioeconomic need, while risk is measured as the standard deviation of portfolio returns across uncertainty scenarios. These measures are used to construct an efficient frontier representing optimal subsidy targeting. Results show that historical allocations consistently fell below the efficient frontier, with the efficiency gap widening over time as portfolio risk increased more rapidly than expected returns. Although alignment with economic distress improved modestly, these gains were accompanied by greater variability in allocation performance, indicating declining reliability in need targeting. The findings demonstrate that substantial improvements in risk–return efficiency and alignment with economic distress were theoretically achievable without increasing exposure to outcome uncertainty, particularly by reallocating funds toward counties with lower affordability scores. Because the unconstrained frontier represents an infeasible theoretical upper bound, these results indicate the direction and scale of potential gains rather than specific targets. Overall, uncertainty-aware frameworks help refine eligibility rules and direct limited public resources toward communities facing the greatest economic strain.