Money versus procedures — Evidence from an energy efficiency assistance program

In many countries, governments have put in place targeted programs intended to support energy efficiency investments by low-income households, but have encountered low take-up even when subsidies are high. Using evidence from a large energy efficiency assistance program, we demonstrate that seemingly small procedural changes can substantially improve take-up and that these changes have effects comparable to significantly raising subsidies. Observing 77,305 durable goods purchase decisions in a refrigerator replacement program, our RD design exploits two quasi-exogenous temporal discontinuities in voucher value and procedures. Despite seeming disadvantageous, the procedural changes actually raise replacement rates among the target demographic of low-income households, an effect roughly equivalent to raising voucher values by 35 Euro. These results suggest that even under fixed budgets, the performance of energy efficiency assistance programs can be improved through empirically guided procedural design.