��

Charles J. Untiet, Antitrust Divison, US Department of Justice1

��

��

A firm is a natural monopoly if production costs are minimized by a single firm doing all the production in its market. The firm could be a single product or multi-product firm. If the average cost curve of a single product firm declines until it intersects the market demand curve, the firm would be a natural monopoly. But declining average costs (or economies of scale) at all output levels is not a necessary condition for a single-product natural monopoly. The firm's average cost curve could eventually begin to rise without losing its natural monopoly character, as long as market production costs are still minimized by a single firm. (As we will see later, this distinction can be important.) A multi-product natural monopoly also exhibits economies of scope; a single firm will minimize production costs of the multiple products (i.e., the industry cost function is "subadditive"). These products could be highly related; e.g., gas pipeline transport to industrial users versus gas pipeline transport for residential use.

In a natural monopoly, market production is efficient: production costs are minimized by a single firm. But the allocation of the natural monopoly's output is often not efficient, since natural monopoly often conveys market power: the ability to raise price above competitive levels.

��

Market Definition

Market definition plays an important role in defining a natural monopoly. A natural monopoly minimizes production costs by producing all the market's output: this implies that the market has been defined. The methodology for defining markets differs according to its purpose. Market definition for ascertaining whether a firm is a natural monopoly ?and thus a candidate for the imposition of regulation ?is different from market definition for determining whether one can safely deregulate a regulated (or state-owned) firm.

In delineating the relevant market for the purpose of imposing regulation on a natural monopoly, one fears that current prices are too high. The market definition question in this case is whether the firm could supply the entire market at a lower price and still recover its costs.

For example, consider an unregulated firm that operates an oil refinery that is the only source of petroleum on an island. This might appear at first to be a natural monopoly. But suppose the refinery exports into the world market, and the potential of arbitrage prevents it from price-discriminating against consumers on the island. The relevant market here is the world market. The refinery is not a candidate for natural monopoly regulation since it cannot supply the entire world market at lower price. If the island government attempted to depress local prices by regulation, the refinery might respond by exporting its entire output, This refinery is not a good candidate for the imposition of regulation.

Not only geographic markets but also product markets can be drawn too narrowly. A monopoly railroad would not be a natural monopoly in the carriage of a particular product if it could not carry traffic that a cut in its rates attracts from trucks. A monopoly newspaper may not have the capacity to handle advertising business that a cut in its rates attracts from other advertising media.

The deregulation of a currently regulated (or state-owned) firm requires a different market definition exercise. In deregulation analysis, one asks whether deregulation would result in prices above existing market values. The key question here is whether the monopolist can raise price upon deregulation without sustaining an unprofitable loss of business. If the monopolist faces competition from other regions and other products ?e.g., the monopoly railroad faces competition from trucks, the monopoly newspaper faces competition from radio and television, or the local monopoly oil refinery faces competition from imports ?then the monopolist may be unable to raise price, and continued regulation would not be necessary.

Two comments are in order regarding the relevant benchmark price for deregulation. There could be a distinction between the price and market value of a given supply of output if existing regulation depresses price below market clearing levels. Also, as discussed below, regulation of multi-product firms may keep certain prices too low and others too high.

The definition of a natural monopoly and its relevant geographic market thus creates links between the regulation of natural monopoly and free trade. If the natural monopoly's market extends beyond governmental borders, free trade will promote production efficiency by realizing the firm's economies of scale. This efficiency gain may have a political consequence, however. With the removal of trade barriers in a natural monopoly industry, one firm will survive, while less efficient firms in other regions will be forced to exit. In industries that are not natural monopolies, the removal of trade barriers will tend to obviate the regulation of local 嚙練onopolies?by reducing their market power. This can occur in two ways; either by exposing the local firms to competition from imports, or by encouraging them to reduce price in order to export.

��

�. Regulatory Solutions for Natural Monopolies

��

Pricing

As all economics students know, the socially optimal pricing strategy is to set price equal to marginal cost. At a price higher than marginal cost, society would be better off producing more goods, since the cost of the incremental production is less than its market value. Likewise, at price lower than marginal cost, society should reduce production. Unfortunately, if the natural monopoly exhibits economies of scale �� as many natural monopolies do �� marginal cost pricing will not allow the firm to cover its cost. With increasing returns to scale throughout the relevant range, marginal cost is less than average cost. A simple example is a total cost function C = F + cQ, where C is total cost, F is fixed cost, c is marginal cost, and Q is output. Setting the price at marginal cost c would not allow the firm to recover its fixed cost F.

��

Nonlinear Pricing

One way to address the need for the firm to recover its fixed costs F is to set a two-part tariff: a rate with a fixed fee and a variable fee. If the firm had N customers, it might start by setting the per capita fixed free at F/N, and the variable fee at marginal cost c. The problem with this strategy is that there may be some marginal customers whose consumer surplus is less than F/N, and they drop out of the market. These customers receive no net benefit and do not contribute to fixed costs. In many cases, the best way to deal with this problem is for the firm to offer a multi-part tariff which has a lower initial fixed fee and has variable fees that decline with quantity purchased. These multi-part tariffs ?also known as declining block tariffs ?have often been used by public utilities.

��

If the natural monopolist exhibits increasing returns to scale, marginal cost pricing will not allow it to recover its cost. If lump-sum transfers, two-part tariffs, and other more complex tariffs are infeasible ?in other words, the firm is to charge simply a constant per unit (or linear) price ?then average cost pricing is optimal. But what is the best way to apply average cost pricing in a multi-product setting? The optimal set of average cost prices in a multi-product setting is called Ramsey prices. In a simplified case of independent consumer demands, Ramsey prices imply that the relative markup over marginal cost (p_i-c_i)/p_i for product i is inversely related to the demand elasticity for product i. In other words, one should charge high markups for products with very inelastic demand, and low (or zero) markups for products with elastic demands. As a result, the markups have a minimal effect on quantities demanded: the quantities demanded most resemble those resulting from marginal cost pricing. There is no need for the regulator to impose 嚙瘤ully Allocated Costs? or some other arbitrary markup on the multiple products.

��

There are two common forms of economic regulation: 嚙磋ate of Return?regulation, or the regulation of profits, and 嚙瞑rice Cap?regulation, or the direct regulation of prices. In rate of return regulation, the regulated firm files tariff rates that will permit it to receive no more than a regulated rate of return on its capital stock or 嚙緝ate base? Rate of return regulation was the traditional method of regulation of gas, electric, and telephone utilities in the United States.

Rate of return regulation attempts to promote allocative efficiency in that it seeks to set price equal to cost. Unfortunately, rate of return regulation does a poor job in promoting productive efficiency, since increases in allowed costs will be covered eventually by increases in tariff rates. If a regulated firm can pass along its costs, it can waste resources, not produce maximum output given its use of inputs. But even if the firm was not wasting resources, it still may choose to substitute capital inefficiently for other inputs if the allowed rate of return is above competitive market levels. If the allowed rate of return is too low, firms may not make useful investments. Even though rate of return regulation recognizes the right of firms to recover their capital, certain risky investments may not be undertaken since the regulator may not allow the firm a higher risk-adjusted rate of return.

Rate of return regulation has been associated with inefficiently inflexible pricing regulations. For example, arbitrary cost allocations under ?Fully Allocated Cost?pricing will prevent efficiency gains from Ramsey pricing. Using an 嚙瞌riginal Cost?rate base instead of adjusting the rate base upward for inflation can inefficiently 嚙篆rontload?the time-path of capital recovery; i.e., shift the return on an investment towards the early years of the project. While such inflexible rules are not necessary for rate of return regulation, they may be an inevitable consequence in practice, since rate of return regulation demands careful accounting of profit, and engenders disputes among various customer classes.

Price Cap regulation is the direct regulation of price. It tends to break the link between the cost of the firm and its regulated price. This yields a great benefit; it encourages the firm to reduce costs.

Under price cap regulation the firm's maximum price might be allowed to increase with an inflation index. If greater production efficiency or innovation is expected, the regulator might set the growth in the price cap to be some amount less than the inflation index. The maximum price cannot be dependent on the firm's own costs, otherwise the firm would lose its incentive to reduce costs. Instead, the price cap could be based on a general inflation index, or an index of industry prices or costs. A general inflation index would be universally accepted and readily observed by consumers, but it may not reflect cost trends in the particular regulated industry. Industry indices might better reflect industry cost trends, but they might be subject to manipulation by the industry.

Price caps might permit greater pricing flexibility: this could be an advantage in regulating a firm that has market power in some markets and faces competition in others. Price caps could be limited to the noncompetitive markets, allowing flexible pricing in the competitive markets. Even in noncompetitive markets, price caps could allow firms to cut prices during downward shifts in demand. Downward flexibility is more difficult under rate of return regulation, since a price cut in one market may permit a price rise in another.

While the independence of the price cap and the firm's cost promotes cost reduction, it also has disadvantages. Price caps could discourage investment since there could be less guarantee that the regulator will allow the firm to recover its investment costs. If the regulated product has a quality component, price caps may reduce quality. If the inflation index used to adjust the price cap greatly exceeds the growth rate of the firm's costs, prices can soar, reducing allocative efficiency. If the inflation index is greatly less than the growth rate of the firm's costs, the firm's average variable cost could rise above the price cap, forcing the firm to shut down.

In the real world, price cap regulation might not yield maximum production efficiency to the extent that the price cap is not completely independent of the firm's cost. If a firm reduces its cost, the regulator may lower the price cap. But to the extent the regulatory lag under price caps is longer than under cost based schemes, there still may be an incentive to reduce costs under price cap regulation.

��

A natural monopoly industry generally would not require regulation if entry were relatively easy. There are two factors that tend to make entry easy: the relative absence of sunk costs and/or the presence of entities that can sponsor entry by committing a large portion of the market volume to the entrant. Newspaper delivery may be a natural monopoly since production costs may be minimized by a single carrier delivering papers to each house in the community. But newspaper delivery has relatively little sunk costs. While a carrier may need to invest in a vehicle, the investment is not a sunk cost �� there are other alternative uses for the vehicle. Thus entry is "easy:" if entry turns out to be unprofitable, a carrier can take its vehicle elsewhere and exit the market. The ease of entry will tend to make delivery rates competitive: if the incumbent carrier raises its rates, a rival can enter and undercut the incumbent, without incurring significant costs. This is an example where free, unregulated entry may yield competitive pricing, making economic regulation unnecessary.

There are, however, theoretical examples where entry into natural monopoly industries can be inefficient, warranting entry regulation. This is the situation of a non-sustainable natural monopoly: i.e., there is no set of prices that would both cover the costs of a natural monopolist and discourage entry. The following is a simple example. Consider a power generation plant on an island with 3 identical communities. Suppose the total cost of supplying power is $300 for any one community, $400 for any two communities, and $650 for all three communities. This is a natural monopoly, but one of those special natural monopolies in which average cost first falls with increased output, but then begins to rise. The power plant must receive an average of $650/3 = $217 from each community to cover its cost. But two communities can make themselves better off by signing a contract committing themselves to the construction of a new plant, serving only the two communities at an average cost of $200 per community. While the entry makes the two communities better off, it makes the island collectively worse off, and the island government might consider entry regulation.

Entry regulation has been used in the United States to restrict output inefficiently in competitive industries such as taxicab service. Entry regulation, whether by direct prohibition or in the form of import restrictions, can protect local firms that are not true natural monopolies from competition. Such entry regulation results in inefficiently high coal consumer prices and acts as a barrier to free trade. Thus entry regulation is a very dangerous tool that governments should use very carefully. In the case of a sustainable natural monopoly, the threat of inefficient entry can be avoided simply by adopting 嚙編ubsidy-free?pricing. For example, consider a power plant on a second island with only 2 communities. Suppose the communities are equal in population, but one is wealthier than the other. (For simplicity, assume that power is a basic necessity, the demand for which is invariant with respect to price or income.) As before, assume the cost of service is $300 for any one community, and $400 for the two communities. Thus, the power plant is a natural monopoly. But suppose, in the interest of equity, the island regulator imposes a regulated price of $310 on the wealthy community, and $90 on the poor community. This pricing is not 嚙編ubsidy-free? the wealthy community is paying more than the $300 嚙編tand-alone cost?of serving itself, and the poor community is paying less than the $100 嚙箠ncremental cost?of serving a second community. The wealthy community could make itself better off by signing a contract committing itself to the construction of a new plant, serving only itself at the 嚙編tand alone cost?of $300. This entry is inefficient since the power plant is a natural monopoly. But entry regulation is unnecessary here; inefficient entry can be avoided simply by setting regulated prices at levels between incremental cost and stand-alone cost.

��

In some cases negotiation between buyers and sellers can obviate the regulation of natural monopolies. The bargainers may have a collective interest in ensuring the socially efficient level of output. Consider a natural monopoly railroad that is the only outbound means of transportation from a coal producing region. Suppose the railroad carries the coal to a competitive downstream market; i.e., its volume of carriage cannot affect the downstream price of coal. The natural monopoly railroad could still have monopoly market power against the local coal producers: the railroad could raise its rates, thus depressing mine prices and volumes to inefficient levels. But the coal producers and the railroad ?as a coalition ?have a collective interest in maximizing their joint profit. As long as the coalition is not able to pass along market power pricing to other persons located further downstream or upstream, the coalition will completely internalize the market power inefficiency. Thus the coalition's incentive to maximize joint profit translates to an incentive to maximize efficiency. The railroad and coal producers have a collective interest to agree to a price equal to marginal cost on the incremental traffic, thus transporting the optimal amount of coal. While bargaining entails its own costs, these costs may be less than the cost of regulation as long as the product is simple and the number of significant producers is small. Generally such bargaining would occur prior to the construction of the railroad in order to reduce risk, but the same collective interest would exist after the construction. Of course, bargaining could effect a significant transfer of wealth, but the outcome could still be efficient.

When the cost of bargaining is high, it is not a good solution for inefficiencies from natural monopoly pricing. For example, it would be infeasible for a local electric utility to negotiate with each individual residential customer to maximize joint benefit. This suggests that when a natural monopoly has market power over atomistic consumers, regulation may be the preferred option.

��

Certain transportation industries may feature competition that obviates regulation of natural monopolies. Many transportation facilities (e.g., railroad lines, oil and gas pipelines) exhibit economies of scale. Ex ante, it is often cost-minimizing for a single transport line to operate between two points, making the line a natural monopoly. Over time, however, as the economy grows and patterns in regional demand and supply shift, a dense network of transport facilities may emerge. The network might feature multiple transmission between points A and B. But even if such multiple parallel lines do not exist, source competition can render regulation unnecessary.

To illustrate, consider a natural gas pipeline operating from origin A to destination B. Natural gas is a homogeneous commodity: a consumer at destination B does not care if the gas it consumes comes from origin A or from origin X via an alternative pipeline, as long as the delivered prices are comparable. Likewise, a producer at origin A does not care if the gas it produces goes to destination B or to destination Y via an alternative pipeline, as long as the FOB prices are comparable. If gas from X is consumed at B and if gas from A is shipped to Y, these alternative pipelines are providing source competition for our natural monopoly pipeline at both origin A and destination B.

If a pipeline from origin A to destination B faces source competition at both origin A and destination B, the pipeline's rate from A to B may not require regulation, even if the pipeline is the only transport facility between A and B. Rates to an intermediate point between A and B may still require regulation, however. Source competition can take other forms. For example, a pipeline carrying refined petroleum products to destination B can compete at B with alternative transport modes to B, such as water transportation, or with petroleum products from a local oil refinery production facility near B.

��

�. Leveraging Natural Monopoly into Related Markets

An important regulatory issue is the activities of a natural monopoly in related markets. For example, related activities of a natural gas pipeline would include the production and marketing of natural gas. There may be certain efficiencies generated from the natural monopoly's operations in these areas. At the same time, these activities pose a potential risk of regulatory evasion. If the activities are not themselves regulated, a natural monopoly could evade regulation by vertically integrating into the activities and denying others access to the natural monopoly facility. The regulated rate of a pipeline does not really matter if the pipeline has a monopoly (because of the access denial) on the unregulated sale of natural gas.

Until recently, natural gas pipelines in the U.S. could deny access to other shippers. U.S. regulators addressed the problem of regulatory evasion by regulating the wellhead price of gas. The results were unfortunate. Gas production is not a natural monopoly; it is instead subject to increasing marginal cost. Regulation of the price of wellhead gas created shortages as the demand for gas greatly exceeded supply at the regulated price. This caused inefficient substitution to alternative fuels. U.S. regulators responded by relaxing controls off certain higher-cost gas. But this wasted resources that were devoted to producing such high-cost gas, where lower cost gas could have been profitably produced had all the price controls been lifted.

An alternative, less drastic, strategy to regulate vertically integrated pipelines would have been to limit the allowed price of gas at the wellhead to comparable prices paid by other pipelines. This modest approach could work if the pipeline's origin market for gas were reasonably competitive due to the presence of alternative pipelines and other outlets for gas.

In recent years U.S. regulators have dealt with the regulatory evasion problem by mandating open access to natural gas pipelines. Natural gas pipelines must offer transportation services to unaffiliated shippers, although they may also continue to sell gas to customers.

Open access is only one approach to prevent market power inefficiencies from the move to adjacent markets. Open access does not eliminate regulators?concerns over regulatory evasion. One must still regulate the terms at which access is obtained. Open access has the advantage of reducing the scope of regulation. If competitive open access reduces transportation and marketing costs, society benefits from both lower prices in the adjacent market and the prevention of regulatory evasion. In some industry contexts, however, open access may not be desirable if there are significant efficiencies of vertical integration between the natural monopoly activity and the adjacent market. If the efficiencies of vertical integration are large enough, regulation of the monopoly's activity in the adjacent market may be preferable to competition in the adjacent market.

��

�. Conclusion

This is a brief summary of some aspects of the economics of natural monopoly and its regulation. The economics of natural monopoly is a complex subject with a highly technical literature. But economists should realize that regulation is essentially a political process in which economic efficiency is often a secondary goal. Regulatory initiatives generally create winners and losers, and a political consensus is generally required before they will be adopted.