1. DO AMERICANS CONSUME TOO LITTLE NATURAL GAS?An Empirical Test of Marginal Cost Pricing. By Lucas W. Davis and Erich Muehlegger. Key words :Efficient pricing, marginal pricing, natural gas.

2. Introduction The paper performs an empirical test of marginal cost pricing in the natural gas distribution market in the U.S during period 1989-2008. A standard result in regulation is that efficiency requires that P=MC. This eliminates welfare losses e.g in natural monopoly. The paper applies the standard natural monopoly framework to natural gas distribution in US. Unregulated Natural Monopoly>Sub optimal level of output (P>MC) (deadweight Loss) Individually and jointly, for all 50 states the paper rejects the null hypothesis of marginal cost pricing. In practice, most distribution co’s charge prices approx. equal to average cost, Incl. amortization of capital expenditure. Current pricing system yields annual welfare losses of $2.6 billion compared to MC pricing.

3. Background Natural gas market(US) consists of 3 players: Gas producers. Interstate pipeline Operators. LDCs (local Distribution Cos). LDCs purchase gas at the ‘city gate’ and distribute it through a distribution network to 4 classes of customers: residential, commercial, industrial and electric power. LDC’s costs; Main fixed costs= installation and maintenance of the pipeline network. Others ; installing + maintenace of gas meters, processing bills, and taking customer calls.(these costs depends on total no. of customers) Marginal customer cost-cost of additional customer to the network/substantial. Marginal commodity cost- cost of providing an additional unit of gas/cost of the natural gas(city gate price) Note: the marginal cost of distributing gas through the local network is zero.

4. Contd. LDC’S are regulated by state utility which set tariff for each class of customers using ROR. ∑Revenue(pi) = sB + expenses where Revenue(pi) =revenues generated from customer class i given the price schedule pi, s = allowed rate of return, and B =rate base, the total value of the firms investment i.e capital investments. Others are Ramsey pricing; P i - MC i / P i = λ /h i Two part tariff pricing= fixed fee + per unit charge. DATA. To perform the test of marginal cost pricing, they assembled a 20-year panel of natural gas sales and prices at the state-level (1989-2008).City gate prices play an important role in determining Marginal pricing of natural gas for LDC’s. Month level abstracts are used to measure dead weight loss.

5. Results. First, for each state,month, and customer class, calculate “net revenue”NR=TR collected by LDC’s net of commodity cost. Test if NR is a function of natural gas consumption. NR t = α 0 + α 1 q t + u t, t = 1, 2,..., 12 Monthly net revenue from residential sales per customer, NR t, is regressed on monthly gas consumption per customer, q t. - The constant α 0 is the average amount paid in fixed fees - α 1 is the average per unit markup over the city gate price. The test provided strong evidence of departures from marginal cost pricing in all the 50 states. Average mark-up was over 40% per unit of gas. Total Deadweight from non-marginal pricing given that demand elasticity's is inelastic for all three sectors was obtained to be 2.6Bn dollars and it exceeds 1.2bn dollars p.a.

6. Possible explanations. Profit maximization by LDC’s. One possible explanation for the current price schedules lies in the incentives created by rate-of-return regulation. S>r. Distributional considerations. With low connection fees the existing rates imply that within customer classes high-demand customers pay a large share of fixed costs. Environmental Externalities. e.g. implementation of taxes that equate to marginal damages.

7. Solutions. Ownership Structure and Efficiency. An alternative approach for lowering per-unit prices is public provision. Energy utilities in the United States operate under one of two types of institutional arrangements: (1) privately-owned companies regulated by state public utility commissions, (2) publicly-owned or “municipally owned” companies that are directly under public control. In both cases, the public sector controls rate setting. Privately-owned companies recoup fixed costs from end users of natural gas. Municipally-owned companies recoup fixed costs through government subsidies, thereby shifting the burden of fixed costs from natural gas consumers to taxpayers. While subsidies increase the welfare of natural gas users, these gains are offset by tax distortions in other parts of the economy.

8. Concluding remarks. The study rejects marginal cost pricing. Departures from marginal cost pricing are most severe for residential and commercial customers with markups averaging 45% and 42%. For conservative estimates of the price elasticity of demand, these distortions impose large aggregate welfare losses compared to marginal cost pricing. The current system with low fixed fees and high per unit prices implies that there are too many natural gas customers, each consuming too little natural gas. Then public intervention in these markets should proceed with extreme caution because a carbon tax or cap-and-trade program would further reduce consumption below the socially efficient level, exacerbating the welfare losses. Fixed fee should be increased to recoup lost revenue. Differential rate across customers classes.