Benchmarking Infrastructure Using Public Investment Efficiency Frontiers

With limited financing options, increasing investment efficiency will be a critical avenue to building infrastructure for many countries, particularly in the context of post-pandemic recovery and rising debt emanating from higher energy costs and other pressures. Estimating investment efficiency, however, presents many methodological pitfalls. Using various methods--stochastic frontier analysis, data envelopment analysis (DEA), and bootstrapped DEA-this paper estimates efficiency scores for a wide range of countries employing metrics of infrastructure quantity and utilization. We find that efficiency scores are relatively robust across methodologies and data used. A considerable efficiency gap exists: Removing all inefficiencies could increase infrastructure output by 55 percent overall, when averaging across 12 estimation approaches-in particular, by 45 percent for advanced economies, 54 percent for emerging countries, and 65 percent for low income countries. Infrastructure output would increase by a still-sizeable 30 percent if instead of eliminating all efficiency, countries achieved the efficiency level of their income group's 90th percentile