Cathal Guiomard
The historic link between subsidies and restriction of competition is an ancient one. In 1690 Louis XIV officially adopted and began to subsidise the ... Comédie Française. He gave it the exclusive right to perform all plays by authors no longer living. Many independent ventures were formed, but as they expanded their offerings, the Comédie complained and induced the Crown to impose various restrictions on them: Some could let their characters faint or bleed, but not die. Some could let their performers do only acrobatics or pantomime. Two small establishments ... were forced to hang a gauze curtain between actors and audience. One amateur theatre ... could open only at 7, over an hour after the others.
(Kahn, 1989, Vol. 2, p. 4)
In this paper, I review briefly the recent evolution of the academic perspective on optimal economic regulation by reference to selected representative publications from the 1970s to the 1990s. Other papers in the present volume (for instance, Forsyth and McKenzie-Williams) indicate the extent to which recent reviews of airport regulatory policy incorporate the evolving academic perspective.
Lying behind, and perhaps partly prompting, academic research into economic regulation is the gradual worldwide liberalisation of such economic sectors as aviation, telecoms, electricity, gas, and postal services. A thorough analysis of the technological and economic reasons for this liberalisation of network industries is offered in Newbery (1999). As part of this liberalisation process, many sectoral economic regulators have been created. The academic research agenda addresses the question of how these agencies should go about their work.2 Over the past 15 years or so, the focus of the academic analysis of economic regulation has substantially changed direction. Some of the main changes are briefly surveyed below.
In 1970, Alfred Kahn, in a text that remains a classic in the field, set out an exhaustive description and analysis of the regulation of US public utilities, as it was practised at the time and as he thought it could be improved upon
He laid the greatest possible stress on marginal cost as the basis of efficient pricing. In the reprint of this book, which appeared some 20 years later, Kahn pointed to a particular instance where the marginal-cost principle is still flouted in the US: the charging for landing 'slots' at congested airports according to the weight of the aircraft, rather than by auctioning or with reference to variable marginal costs (Kahn, 1989, p. xxii).
Kahn considered in detail the case of natural monopoly, defining it as:
an industry in which the economies of scale - that is, the tendency for average costs to decrease the larger the producing firm - are continuous up to the point that one company supplies the entire demand. [Natural monopoly] is a reason, also, why competition is not supposed to work well in these industries
(Kahn, Vol. 1, p. 123 - 4)
Spulber, (1989), offers an equivalent though perhaps clearer definition: under natural monopoly a single firm is able to serve the entire market at less cost than could two or more firms, p. 3).
The existence - or otherwise - of natural monopoly conditions is plainly something to be determined empirically.
This definition is more precise than the rale of thumb sometimes applied, namely, industries with large fixed costs. Properly defined, 'fixed costs' are a short-run concept; they are the costs of an input in fixed supply. However, in the longer term, all inputs are variable. Therefore, an industry may be characterised in the short run by fixed costs, but when all inputs are variable, returns to scale may nonetheless be constant or decreasing. For example, agriculture is an industry with very high fixed costs in the short run, but without such economies of scale as to make agriculture production by a single farm the most efficient way to organise production.
Kahn goes on to draw out the implication of economies of scale for a (socially) optimal economic organisation. Such a company's marginal costs will be lower than average costs and therefore a contradiction arises between two fundamental rules for a social optimum:
There are only two ways out of this contradiction: either marginal-cost pricing for all buyers, or commercial break-even, has to be sacrificed. If break-even is abandoned, a company can survive only if it receives a subsidy from taxpayers. In practice, this has not been the route policymakers have mostly opted to go.
Alternatively, marginal-cost pricing must be sacrificed. As Kahn argues, the justification for marginal cost pricing lies entirely in the elasticity of demand: consumers buy more, or less, than the optimum according to the price charged. Then why not charge customers (in practice: classes of customers) according to their different willingness to pay? This would allow total revenues to cover total costs while permitting a level of output closer to the optimum through seeking to avoid turning away any class of buyer (up to the optimum output) by charging them no more than they are willing to pay for that output.
Of course, such price discrimination is one area at least where there could be tensions between the approaches of economic regulators and of competition authorities. Price discrimination is open to abuse, particularly where buyers have no choice of supplier. Furthermore, if different prices are charged to customers who are themselves commercial rivals - for example, competing transport modes competition in their industry may be impaired, requiring a trade-off between the importance of different inefficiencies to be evaluated.
When is price discrimination justified? In theory, the answer is when three tests are met:3
Did price discrimination in regulated US industries generally meet these conditions? Kahn comments on the case of US aviation (before 1970) as follows.
Consider for example promotional airfares ... As long as reduced fares are for standby service or travel on off-peak days or seasons they are not discriminatory; they merely reflect (perhaps even inadequately) the lower marginal costs of travel that makes no marginal demand on capacity. To the extent that they are offered to select groups - young people under 21, wives and children accompanying husbands - they are discriminatory but the discrimination may be justified because (1) of the presence of short-run decreasing costs at times when capacity is in excess and (2) the demand by these classes of travellers may well be more elastic than by single unaccompanied adults, many of them travelling on business. But to the extent that the purpose of such fares, or of the subsidies paid to the airlines that serve smaller towns or the discriminatory low rates on short-hop trips, is merely to 'promote air travel' or the habit of using airplanes, without reference to whether the travel promoted is on-peak or off-peak, it seems to lack economic justification, in view of the apparent absence of economies of scale in the air transport business in the relevant range.
(Kahn, 1989, Vol 1, p. 149-150).
Even in the relatively narrow set of circumstances m which he considers economic regulation to be justified, Kahn does not underestimate the practical difficulties in applying the principle of marginal-cost pricing. At least the following set of issues falls to be considered by an economic regulator:
There is an evident challenge in seeking to carry out this large set of difficult tasks. But, in addition, Kahn is concerned about the possible organisational culture 3f regulators' offices. He argues that, at least historically, the 'regulatory mentality' has suffered from an anti-competitive bias. The result is that regulation has often not supplemented (inadequate) competition but in fact has supplanted it.4 This leads to a general regulatory dilemma.
An important shortcoming of the approach to economic regulation just discussed is that it makes formidable data and computational demands on regulators. But, far from having perfect information, economic regulators may know relatively little about
In the 1980s, the alternative concept of regulation by 'price caps' was devised and implemented in the UK. This approach arose out of the prospect of substantial deregulation of utility businesses in the UK, the corresponding requirement to set up a regulatory regime, and the known informational asymmetry between the regulator and the regulated entity.
The logic of price-cap regulation has recently been assessed for the UK Civil Aviation Authority by Professor David Sibley (Sibley, 2000).
Sibley notes that, in the simplest version of price caps, the regulator does nothing except impose a cap on the regulated firm's prices and enforce adherence to that cap. The cap typically declines from one year to the next at a given percentage rate X. The theory assumes that the firm has superior information about demand and costs, as compared to the regulator. Using this superior information, the firm simply acts to maximize profits, discounted into the future, subject to the price cap.
Price-cap regulation mimics a competitive market in which the producer cannot influence the price and can increase profits only by reducing costs. As long as the cap is set at the economically correct level, this creates powerful incentives for the firm to behave efficiently. In any given time period, the firm gains by trimming its costs and reducing waste. Furthermore, because the market value of the firm is the present value of the future stream of its profits, the firm also gains from efficient investment behaviour over time. Price caps also allow the firm to price discriminate so that economic efficiency may rise as well, as the firm offers prices that are determined by consumer tastes.
However, Sibley stresses that for these attractive properties of price cap regulation to hold, there is one key requirement: the regulator must be able credibly to pre-commit to refrain from opportunistic behavior towards the firm. Suppose that the firm has been operating under a system other than price caps and is put under a price cap scheme. The regulator may have formed some understanding of the firm's costs and marketing abilities under the old system, which will be shown to be false under the new one.
If the firm does indeed respond to the incentives offered by the price cap plan by reducing costs or pricing more efficiently than before price caps, the regulator, observing the new, lower costs and more efficient prices, may wish to respond by requiring the firm to return some or all of the resulting extra profit. In a price cap system, a common way to accomplish this is to increase the X factor, even after the regulator has committed to a particular level of X. Unless a regulator credibly commits to refrain from such behaviour, the firm's incentive to pursue efficiencies is considerably reduced.
An example, discussed in Forsyth in this volume, was the 2002 decision of the Australian authorities to lift airport price-cap regulation (and replace it with price monitoring), only five years after its introduction, because, Forsyth argues, price-cap regulation against a difficult economic background led to serious profit volatility and the risk of firm failures.
There exists a large academic literature on optimal regulatory incentive schemes in the presence of informational issues such as adverse selection and moral hazard.5 Although the schemes described in this literature are generally too complex to implement directly, they all have one element in common: optimal regulation with informational asymmetries involves some degree of earnings sharing. Thus, appending at least a small amount of earnings sharing to a price cap scheme increases economic efficiency.
As the discussion of commitment above makes clear, despite good intentions, in practice regulators often do not honor initial commitments to a regulatory scheme. There will always arise political pressures to increase X if the firm is making large profits, even if those profits are due to no more than responding to incentives put in place by the regulator. Similarly, if the firm's earnings are unexpectedly low for too long, there will be pressures to modify the scheme so as to increase earnings.
Sibley proposes a potential response to this problem: periodic adjustment of the X factor according to a formula that is known in advance to the firm, and is not opportunistic.
From today's perspective, one of the striking features of the original approach to the analysis of economic regulation is that the regulatory process itself generally lies outside the analysis. Economic regulators arrive from out of the blue to correct the imperfections of the market without themselves being analysed as entities pursuing their own institutional or other objectives within the same world that the other actors also form part of. This is part of the reason for Kahn's comment that:
The ... major limitation of any mere exposition of the normative rules of economic behaviour and performance [is] that they tell us absolutely nothing about how to achieve these results ... by what kinds of institutional arrangements can we obtain the maximum assurance ... that the goals will in fact be achieved?
That shortcoming is being made good, Spulber (1989) offers a thorough (graduate-level) survey. He argues that a full analysis of regulation needs to take account of the following drawbacks to regulation.
Regulation may entail high administrative costs. The formulation and implementation of regulation may entail policies that not only are redistributive in nature but that interfere with allocative efficiency. Regulation of markets is subject to all of the problems that attend economic planning. The regulator operates with imperfect information. Regulatory policies may be second best, reflecting rules of thumb, administrative procedures, and institutional constraints. Even well-intentioned public policies can create incentives for inefficient private behaviour.
(Spulber, 1989, p.65)
In the Spulber approach, a three-part 'market failure test' is used to evaluate the need for economic regulation (or, equally, competition policy). First, the market in question must be shown to depart from the first-best equilibrium. Second, it must be shown that regulation - facing the same institutional, technological and informational constraints (that likely are the cause of the market failure) - can alleviate or correct the market failure. Third, the potential benefits of a regulatory remedy must justify market intervention, taking account of the costs and any distortions of economic regulation.
Spulber presents his analysis using the contestable markets framework. His argument for doing so is that this amounts to the framework of perfect competition but without the need to assume that there are a large number of firms that can compete only by varying output (but not price). In particular, contestabilty allows for natural monopoly. Spulber argues that this is the best basis for the study of significant scale economies because, if contestable markets perform well under scale economies, then natural monopoly cannot justify regulation. But if natural monopoly and departures from contestability - such as sunk costs or other barriers to entry - lead to market failure, regulatory policy may be desirable. However, it must then focus on both the natural monopoly and the non-contestability features of the market failure.
For example, Spulber concludes that under natural monopoly conditions, if public policy seeks 'zero-profit subsidy-free' prices,6 then contestability without regulation is sufficient to produce these prices, since only by setting zero-profit subsidy-free prices can the incumbent make the natural monopoly sustainable.
However, subsidy-free prices may not be welfare maximising. For example, Ramsey prices may involve cross-subsidies so are not sustainable. This leads to a trade-off between the welfare gains of Ramsey pricing (plus regulatory costs) and the welfare attainable under subsidy-free prices (without regulatory costs).
Spulber analyses a series of different (increasingly information-intensive) price-setting policies that regulators might set - including cost-based prices, break-even prices, price discrimination, and perfect price discrimination - but concludes that for the contestable market case, the most economically efficient approach is an auction for the right to supply the market in question. Where there are barriers to entry (and hence non-contestability) the franchise cannot be used repeatedly so traditional forms of economic regulation may be required. A market solution for sunk costs is, of course, the long-term contract. Spulber evaluates economic regulation - itself a form, he maintains, of implicit contract - against privately negotiated and enforced contracts. Regulation may reduce the cost of negotiation (in a market with many agents) and the cost of contract enforcement.
Spulber's analysis is at a very high level of abstraction and is some way from offering a set of practical tools for economic regulation, though these may well be developed. Indeed, one of the book's key conclusions - that franchise contracts be substituted for the apparatus of formal economic regulation - could be challenged by one of the lessons being drawn from the bankruptcy of the UK company Railtrack.
In October 2001, a columnist with the Financial Times' argued that Railtrack followed 'the 1990s orthodoxy' (of concentrating on its 'core competence' of managing the rail network) and outsourcing much else including track maintenance to contractors. The approach, he argued, suffered from three flaws: first, even well-specified contracts cannot replicate the operations of integrated businesses; second, while information technology has lowered the costs of information exchange it has left the costs of trust-building and alignment of interests as high as ever; finally, while contracts may work well against a stable economic and political background, after a crisis, such as a rail accident, they may become very unstable.
The logic of sectoral economic regulation is strongest - at least from the viewpoint of economic efficiency8 - in an industry marked by significant economies of scale. The existence of a natural monopoly can justify a certain level of economic regulation to seek to protect consumer welfare. But how is such a worthwhile aim to be achieved?
Marginal-cost pricing remains sound in principle but difficult to implement with precision because sufficient information may not, in general and in practice, be available. In addition, once the government agency is made endogenous to the analysis of regulation, the theoretical case for economic regulation is weakened and novel contract-based forms of regulation may have theoretical attractions to earlier approaches. However, inter alia, the UK Railtrack collapse of 2001 has suggested that contract-based regulation, at the present stage of knowledge, may also be difficult to put into practice.
At a less technical level, it is helpful to recall that the objective set for the UK Office of Fair Trading is "to make markets work well for consumers". Carefuly attention to developments in the theory of optimal economic regulation can help regulatory agencies to pursue the interests of consumers successfully.
1I am grateful for helpful discussions with, and comments from, my colleagues at the Commission for Aviation Regulation (CAR) during the preparation of this paper; also from participants at the 6th Air Transport Research Society Conference in Seattle, July 2002, and especially from Professor Hans-Martin Niemeier of the Hamburg Institute of International Economics and Professor Peter Forsyth of the Department of Economics of Monash University in Australia. The views expressed in the paper are personal and do not necessarily reflect the opinions or positions of the Irish Commission for Aviation Regulation.
2These new institutions have become topics of sometimes-heated discussion. Following She first report of the Irish CAR on airport pricing, a columnist with an Irish daily newspaper remarked: "Have you heard about regulators? It rhymes with terminators and they're like a sort of new gunslinger in town." Fergus Finlay, Irish Examiner newspaper, 1st September 2001.
3 An alternative to price discrimination would be two-(or multi-)part tariffs.
4 As illustrated by the quote on the cover page.
5 Some of this literature is surveyed in Laffont and Tirole (1994).
6 In a multi-product firm, subsidy-free prices are those such that the revenue generated by any subset of output levels does not exceed the stand-alone cost of producing those output levels.
7 Peter Martin, "Lessons from Railtrack", Financial Times, 9 October 2001, p. 21.
8 For example, the Aviation Reguiation Act 2001 charges the Irish CAR with the task of setting maximum airport charges in such a way as to seek to promote "the operation and development of cost-effective airports that meet the requirements of users".
Kahn, Alfred E (1989) The Economics of Regulation (reprint), MIT Press, Cambridge MA, published in two volumes in 1970 and 1971, by Wiley.
Laffont, Jean-Jacques and Jean Tirole (1994) A Theory of Incentives in Procurement and Regulation, MIT Press, Cambridge MA.
Newbery, David M (1999), Privatization, Restructuring and Regulation of Network Utilities, MIT Press, Cambridge MA.
Sibley, David (2000) "Economic Analysis of the CAA's Initial Incentive Regulation Proposal for the National Air Traffic Services", available at www.caaerg.co.uk/erg/.
Spulber, Daniel F. (1989) Regulation and Markets, MIT Press, Cambridge MA.