Working Paper: NBER ID: w16653
Authors: Severin Borenstein; Lucas W. Davis
Abstract: Residential natural gas customers in the United States face volumetric charges for natural gas that average about 30% more than marginal cost. The large markup on natural gas - which is used to cover the fixed infrastructure and operating costs of the local distribution companies - is widely recognized to be inefficient. Nonetheless, attempts to reduce volumetric charges, and cover the revenue shortfall through increased fixed monthly fees, have faced opposition based on the belief that current rate schedules have desirable distributional consequences. We evaluate this claim empirically using nationally-representative household-level data. We find that natural gas consumption is weakly correlated with household income, so current rate schedules are only mildly progressive. Under current rate schedules, high-volume customers pay a disproportionately large share of fixed costs, but these exhibit a weak correlation with high-income households. The correlation is somewhat weaker still when we consider alternative indicators of household financial stress, such as poverty status or number of children in the household. We show, for example, that poor households with multiple children would receive lower bills on average under marginal cost pricing. We present evidence that one cause of the weak redistributional impact of the current pricing policy is that the poor tend to live in less energy efficient homes.
Keywords: natural gas; pricing; equity; efficiency; tariffs
JEL Codes: D42; L50; L95; Q48; Q54
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| current volumetric charges are marked up by approximately 30% above marginal cost (L97) | inefficiencies and deadweight loss (H21) |
| transitioning to marginal cost pricing (D40) | average increase of $44 in annual bills for lowest income quintile (I14) |
| transitioning to marginal cost pricing (D40) | average decrease of $58 in annual bills for highest income quintile (D12) |
| transitioning to marginal cost pricing (D40) | net loss for low-income households (G59) |
| transitioning to marginal cost pricing (D40) | savings for higher-income households (D14) |
| households with children (D19) | lower bills under marginal cost pricing (D40) |
| weak correlation between natural gas consumption and household income (D19) | current pricing structures do not effectively redistribute costs (D49) |