Go back to John Quiggin's 1998 articles

Go back to John Quiggin's home page

Quiggin, J. (1998), 'The premature burial of natural monopoly: Telecommunications reform in Australia', Agenda 5(4), 427­440.

Date created: 20 April 1998

Last modified: 20 April 1998

Maintained by: John Quiggin

Although much of the present policy framework was embodied in the Telecommunications Act 1991, the views of governments and telecommunications regulators have evolved over time. Moreover, there has been a complex interaction between telecommunications-specific policy and the general framework of National Competition Policy) embodied in the Competition Policy Reform Act 1995 and reflected in the regulatory role of the Australian Competition and Consumer Council (ACCC). Hence, although it is possible to describe current policies reasonably accurately, it is often difficult to distinguish between outcomes that are the result of deliberate design and those that are unintended consequences of interactions between decisions.

The central distinguishing element of the policy framework that has emerged in Australia has been the belief that in order to make the Australian telecommunications system fully competitive, competitors should be encourage to build complete telecommunications network to compete with that of Telstra. This approach to reform is sometimes referred to as facilities-based competition. However, this term is used to describe a large class of policy approaches. In this paper, the approach adopted in Australia will be referred to as the policy of network duplication.

Reform of the Australian telecommunications system has not produced the benefits predicted by those who advocated it. In particular, Telstra has retained its market dominance despite the dissipation of billions of dollars in the development of duplicate infrastructure networks. The rise and fall of the Super League football competition and the decision of the ACCC to bankrupt one firm (Australis) in order to save another (Optus) are illustrations of the failure of telecommunications policy to produce economically rational outcomes.

The purpose of this paper is to argue that the main deficiency in Australian telecommunications policy has been the failure to take adequate account of natural monopoly. Alternative institutional structures, designed to exploit economies of scale and scope could have yielded both a more competitive market structure and lower industry costs.

The paper is organised as follows. Section 1 deals with the concept of natural monopoly as it applies to the telecommunications industry. Section 2 describes the process of telecommunications reform. Section 3 deals with some of the outcomes of reform. It is argued that outcomes have fallen short of expectations. Section 4 presents some alternative policy frameworks which, it is argued, would have yielded superior policy outcomes. Section 5 deals with corporatisation, privatisation and regulation of Telstra. In Section 6, policy options for the future are considered briefly. Finally, some concluding comments are offered.

1. Natural monopoly and the telecommunications industry

The concept of natural monopoly is central to an understanding of policy issues associated with telecommunications reform. The concept was not given a theoretically satisfactory definition until the late 1970s (Baumol 1977; Baumol, Panzar and Willig 1982). However, the basic idea may be traced back to the 19th century debates on the regulation of railways. If the production of a given good or service displays strictly increasing returns to scale, it may be provided at least cost by a single enterprise. Governments faced with the problem of regulating railways soon realised that allowing unrestricted construction of railways by competing firms would lead to wasteful duplication and an inefficient network. A railway between two points was a 'natural monopoly'.

Baumol and his co-workers clarified the understanding of natural monopoly in the case of firms producing multiple outputs. A particularly important contribution has been the idea of sub-additivity. A cost function is sub-additive when any given total output can be produced at least as cheaply by a single firm as by two or more separate firms. An industry in which the cost function is sub-additive may therefore be regarded as a natural monopoly. Baumol, Panzar and Willig distinguish between economies of scale, that is, cost savings arising from producing larger volumes of a given bundle of goods and services, and economies of scope, that is, cost savings arising from producing more than one good or service in the same enterprise. In broad terms, an enterprise that displays economies of scope and scale in the production of a basket of goods may be regarded as a natural monopoly.

Under conditions of natural monopoly, unregulated markets are unlikely to yield a satisfactory outcome. Initially, unrestricted competition is likely to lead to inefficient duplication. For example, in the 19th century, competing railway companies laid parallel tracks in an effort to improve their relative market position. Overinvestment on strategically important routes is likely to be matched by underinvestment in areas where one or other of the firms has an effective monopoly. Eventually, however, one firm will either merge with its competitors or drive them out of business and the industry will become an actual as well as a natural monopoly. This process of strategic duplication and merger is now taking place in the Australian telecommunications industries, although strategies are complicated by the need to take account of the possible reactions of regulators.

The distributional consequences of unregulated natural monopoly are not clear. For consumers, the initial phase of wasteful competition will yield significant benefits. Firms battling for market share will offer low prices in the hope of being the single survivor, or at least of being able to demand favourable terms in any merger or takeover. However, when the monopoly state is reached, the benefits previously given to consumers will be clawed back in the form of monopoly profits. Assuming full rationality, firms entering the industry will expect ex ante to receive normal returns to capital. The supernormal profits earned by the surviving monopolist will be counterbalanced by the losses incurred by the firms that fail. From a broader social viewpoint, what matters is that unrestricted competition implies the dissipation of all the social surplus associated with the existence of the natural monopoly industry.

Cost subadditivity in telecommunications

In the case of network goods such as telecommunications, a finer disaggregation of the possible sources of subadditivity is required. Consider a telecommunications enterprise offering local telephone services in a number of districts, long-distance telephone services between those districts and other services such as pay-TV and data communications. Several potential sources of cost subadditivity may be distinguished, including potential cost savings from:

(i) serving more customers in any one of the markets mentioned above;

(ii) serving multiple local markets;

(iii) providing local and long-distance services jointly; and

(iv) providing telephone, data and pay-TV services jointly.

In the Baumol, Panzar and Willig framework, (i) represents economies of scale and (ii)-(iv) are different forms of economies of scope. However, it is common in the analysis of network goods to use the term 'economies of density' to refer to (i), 'economies of scale' to refer to (ii) and 'economies of scope' to refer to (iii) and (iv).

In addition to these sources of cost savings, it is necessary to take account of network economies which arise with communications services because the service becomes more valuable to any given consumer as the number of other consumers connected to the service increases. Network economies may be viewed as consumption externalities opposite in nature to congestion externalities. They are associated with a tendency to natural monopoly, since the monopolist can internalise the externality.

Unfortunately, the empirical evidence on the extent and nature of economies of scale and scope in telecommunications, surveyed by the Industry Commission (1997), is ambiguous. Part of the problem is that monopolies do not normally generate data sets that make econometric estimation of cost functions feasible. Some observers (Shin and Ying 1992, Spiller and Cardilli 1997) have concluded that telecommunications monopolies are 'unnatural', but others, such as Röller (1990) and Gabel and Kennet (1994) reach the opposite conclusion. Application of Stigler's (1958) 'survivor' principle supports the latter view. Despite deregulation in a great many countries, with many different policy regimes, incumbent firms have remained dominant in most markets, and in nearly all local telephony markets (Harrison and Fisse 1997).

For the purposes of analysis in this paper, it will be assumed that economies of scale, that is economies of type (i), apply to local services of all kinds, but not to long-distance services. No assumption will be made about the presence or absence of economies of types (ii) to (iv). The problem for policymakers is to design a framework which minimises restrictions on competition while ensuring that economies of scale are captured and that, where economies of scope are present, they can be exploited.

Policy response to natural monopoly

Until recently the most common response to the problem of natural monopoly was public ownership of monopoly enterprises. The main exception to this pattern was the United States, where regulated private monopolies, with cross-subsidies and entry barriers similar to those of public enterprises elsewhere, were preferred. Since the early 1980s, however, it has increasingly been assumed that market discipline is adequate even in situations of natural monopoly.

A theoretical development supporting this view was the idea of contestable monopoly, developed by Baumol, Panzar and Willig who stressed the importance of potential entry. The contestable monopoly analysis represented an extension of older models of potential entry such as Bain's (1956) 'limit pricing' model. Baumol, Panzar and Willig supplemented the standard consideration of barriers to entry by consideration of barriers to exit, that is, of sunk costs. Baumol, Panzar and Willig showed that if there were no barriers to entry or exit, the only sustainable price vector for a natural monopoly would be the Pareto-optimal vector of Ramsey prices.

This theoretical point was taken by many policymakers to be an all-purpose argument for laissez-faire. As Baumol and Willig (1986, p. 10) state:

...before anyone can legitimately use the analysis to infer that virtue reigns in some economic sector and that interference is therefore unwarranted, that person must first provide evidence that the arena in question is, in fact, highly contestable. The economy of reality is composed of sectors which vary widely in the degree to which they approximate the attributes of contestability. Thus, the conclusion that perfectly contestable markets require no intervention claims little more than the possibility (which remains to be proven, case by case) that some markets in reality may automatically perform in a very acceptable manner despite the limited number of firms that inhabit them.

Unfortunately, these strictures have been ignored by many participants in the Australian policy debate. The term 'contestable' has been used loosely to refer to any market in which restrictions on entry have been relaxed or removed, without any attempt to examine the presence or absence of sunk costs, and hence the likelihood of an outcome free from monopoly pricing problems. The sloppy usage of terms like 'contestability' reflects a more general acceptance, on a basis of faith rather than economic analysis or empirical evidence, of a belief that competition, or merely the threat of competition,will always and everywhere generate socially optimal outcomes.

Sloppy use of contestability theory has been supplemented by wishful thinking about technology, drawing on the claims of writers such as Toffler (1980) of a 'third wave' of post-industrial society in which technology is inherently biased towards small-scale production, and therefore towards competitive outcomes. Oddly enough, the paradigm chosen by Toffler is the microcomputer industry, in which more than 80 per cent of the market is controlled by a single operating system (Microsoft Windows) using central processing unit (CPU) chips supplied by a single firm (Intel). Many industry participants argue that alternative operating systems and CPU chips are technically superior, but that the market is inherently conducive to monopoly.

The natural monopoly characteristics of the computer industry arise from a number of sources including

(i) economies of scale in research and development of operating systems and CPU chips

(ii) cost savings in writing and marketing software for a single operating system and chip instruction set

(iii) network economies for consumers who use compatible systems

Disregarding the inconsistency of claiming that a monopolistic industry will generate a competitive future, advocates of the proposition that technology has rendered the concept of natural monopoly obsolete focus on evidence of convergence between the previously separate computer, television and telecommunications industries. Although most of these industries have historically been either oligopolies or natural monopolies, it is asserted that convergence between them will result in a competitive free-for-all in which regulation will be unnecessary, or perhaps impossible.

The central theme of this paper is that there are no grounds for supposing that the concept of natural monopoly will be any less relevant in the past than it has been in the future. Technological change may allow competition in previously monopolistic industries, but it may also generate new sources of natural monopoly.

2. The process of telecommunications reform

As in most other OECD countries, Australian telecommunications services were provided, until the 1980s, by a publicly owned monopoly. Significant reforms had taken place during the 1970s with the replacement of the former Postmaster-General's Department, under which telecommunications and postal services had been organised on public service lines, by two profit-making statutory corporations, Telecom and Australia Post. These reforms were highly successful in raising productivity and profitability (Industries Assistance Commission 1989) although critics argued that this improvement was due to the fact that the enterprises had started from a very low base and that even after these gains, the enterprises were well below world best practice (Swan 1990, Bureau of Industry Economics 1992). As the reforms introduced in the 1970s were directed towards strengthening the public monopoly model, rather than towards market competition, they will not be considered here.

The first move towards free-market reform came with the Davidson Report (Committee of Inquiry into Telecommunications Services in Australia 1982). The Davidson Report recommended that private networks be permitted to interconnect with the public network and resell excess capacity, and to provide terminal equipment and certain telecommunications services. Telecom unions strongly opposed the Report and it was initially rejected.

By the late 1980s however, the Labor government had been converted to a belief in competition as the main force for economic efficiency. This was reflected in the 1988 Statement issued by Gareth Evans, then Minister for Communications which indicated the government's desire for more extensive competition in the telecommunications sector. One component of reform was the corporatisation of public telecommunications enterprises. Under the Australian Telecommunications Corporation Act 1989, Telecom was required to follow 'clearly defined commercial principles and objectives'. In 1992, Telecom and OTC were merged to form the Australian and Overseas Telecommunications Corporation, subsequently renamed Telstra, a corporation in which the government initially held all the shares.

A new regulatory framework for the telecommunications market was introduced through the Telecommunications Act 1991. Under this Act, a single competitor for Telstra was admitted to the general telephony market. The successful tenderer was called Optus. As part of the entry conditions, Optus was required to purchase the Aussat satellite system, paying the government more than its market value. The premium was, in effect, a license fee (Maddock 1992).

Under the duopoly regime Optus was permitted to compete with Telstra in all areas of telephone service. The duopoly regime was intended as an interim measure on the way to more open competition in 1997 (Brown 1996). A new body, originally called the Australian Telecommunications Authority (Austel), and subsequently renamed the Australian Communications Authority, was established to regulate prices and conditions of service. Austel was also required to establish the access rules which determined how much long distance carriers would pay for the access to Telstra's local network that was needed to complete calls.

It was not immediately clear whether the government favoured network duplication, or whether it was hoped that the mere threat of entry would force Telstra to set prices for access to its network at a level consistent with competitive outcomes. It gradually became apparent, however, that the government had adopted a policy of network duplication.

The first clear indication of the government's intentions came with the decision to establish three separate networks for digital mobile telephony. To induce Optus and Vodafone to establish digital networks, the government promised to phase out the existing analog network by 2000. The irony of promoting competition in one service by prohibiting the provision of another appears to have escaped those responsible for this decision.

The natural conclusion of reforms leading to a competitive telecommunications market would be the withdrawal of government from the telecommunications industry. Although Richardson (1994, p. 306) asserts that this proposal was favoured by Paul Keating in 1991, the Labor government did not adopt it. Following the Coalition's election victory in 1996, one-third of the public shareholding in Telstra was sold through a public float. Legislation for the complete privatisation of Telstra was introduced in 1998, but was defeated in the Senate. In response to opposition from members of the National Party, the proposal for complete privatisation was withdrawn by the government in July 1998 and replaced with a proposal for a reduction of the public shareholding to 51 per cent of the total.

3. The outcomes of reform

Despite radical changes in telecommunications policy, changes in the telecommunications market, particularly for services to residential users, have been modest. Moreover, the limited increase in competition that has occurred has been achieved at the cost of substantial dissipation of resources in strategic investments. Some of the more obviously unsatisfactory outcomes of reform are that:

… Telstra has remained officially dominant in all markets except those for mobiles and international calls, and its monopoly of local telephony has remained intact,

… Public control over Telstra's actions has been reduced through corporatisation and partial privatisation.

… Telstra's quality of service has declined (Australian Communications Authority 1998) even though technological improvements should have led to an increase in reliability

… Price levels continue to be determined by regulation rather than competition. In most years, Telstra has reduced prices by exactly the amount required to meet the price caps imposed under regulation (Austel various years, Albon 1998).

… After racing to construct parallel hybrid-fibre cable networks for pay-TV services, Optus and Telstra halted their rollouts in 1997 leaving two incomplete, but largely overlapping, networks.

… Similar parallel networks for digital mobile phone telephony have failed to match the coverage of the existing analog mobile phone network

… The analog mobile phone network is to be compulsorily shut down.

… To provide content for pay-TV networks, duplicate rugby league competitions were operated in 1997, providing entertainment services considerably less valuable than that of a unified competition, but at much greater cost.

… Average prices have fallen, but only at the same rate as before reform took place.

Most of these outcomes are evident to any observer. However, since it has often been claimed that competition has resulted in lower prices, it may be worth examining this issue in more detail. The price cap imposed on Telstra from 1991 to 1996 required a 4.5 per cent annual reduction in real prices, which is very similar to the rate of reduction achieved over preceding decades (Arena, Bahtsevanoglou and Branton 1992). As Austel (various years) and Spiller and Cardilli (1997) observe, Telstra met the price cap exactly in most years. However, the pattern of price reduction changed from 1991. Prior to 1991 standard charges were reduced steadily across the board. After 1991, price reductions were delivered primarily through discounts targeted at those consumers most likely to switch to Optus. Since 1997, the rate of reduction required under the price cap for Telstra's general bundle of services has been increased to 7.5 per cent. However, this increase is due to the increased importance of mobile phone services for which the rate of cost reduction is faster than for wires-based local and long-distance services. The rate of reduction required for the latter subset of the bundle remains around 4.5 per cent. Albon (1998, pp. 323-4) concludes that 'some productivity growth appears not to have flowed through to excessively higher profits and maintenance of higher costs rather than price reductions'

Some of the unsatisfactory outcomes of reform may be put down to bad luck or bad management. An arguable example of bad luck is the technical difficulties which have plagued Optus' attempts to enter the market for local telephony. Similarly, it may be argued that the creation of Super League as a rival for the Australian Rugby League was merely a bad commercial decision analogous to, but less successful than, the creation of the World Series Cricket competition in the 1970s.

However, there are some consistent features underlying all the policy failures listed above. First, in each case, they result, in large measure, from a failure to take account of the importance of natural monopoly in telecommunications networks. Policy-makers who wanted to promote competition accepted claims, based more on wishful thinking than on serious analysis, that technological developments had rendered the concept of natural monopoly obsolete, and proceeded to act on the assumption that a competitive market could be legislated into existence.

Second, because of the duopoly or monopoly character of the main markets in which they operate, Telstra and Optus are in a position to pass on the costs of unprofitable ventures to consumers. The absence of price benefits from reform, noted above, suggests that the cost savings associated with the corporatisation of Telstra and the introduction of private sector competition have been used to finance strategic investments.

Finally, the difficulties involved in managing the industry through regulation, rather than public ownership have been underestimated. For example, competition policy is designed to deal with problems of monopoly power in a particular industry. The problems of defining the scope of an industry, distinguishing 'predatory' from merely 'vigorous' competition , and so on have been the subject of considerable study. But the study of multimarket interactions, such as those considered by Bulow, Geanakoplos, and Klemperer (1985) is much less well developed. Both the Super League and Australis cases involved multimarket interactions. In the case of Super League breaches of contract that were at least arguably intended to enhance monopoly power in telecommunications and pay-TV markets were found to be justified because the contracts in question were seen as reducing competition in the rugby league market. Effects on other markets were ignored. By contrast, in the Australis case, the interests of the shareholders in the Australis pay-TV venture were sacrificed to promote the (arguably illusory) prospect of competition in local telephone markets. It is impossible to avoid decisions of this kind in dealing with multimarket enterprises having natural monopoly characteristics. But regulators and courts are not well equipped to make such decisions, which can rarely be made through the application of prespecified rules.

4. Policy alternatives

Policy choices in telecommunications, such as the extent to which dominant enterprises should be broken up, often involve trade-offs between competition and the exploitation of economies of scale and scope. In the current debate, competition is assumed to be highly desirable, but the grounds for this assumption are not always clear. Leaving aside ideological beliefs, two main arguments may be put forward in favour of competition.

First, under the standard assumptions, competitively determined prices are equal to marginal costs and are therefore consistent with Pareto-optimality while monopoly prices are set above marginal costs and therefore generate deadweight losses. Where economies of scale and scope are present, this simple analysis does not apply and it is necessary to consider two-part pricing schemes and Ramsey prices (Baumol, Panzar and Willig 1982). In the absence of contestability, there is no guarantee that reductions in barriers to entry will yield a welfare improvement. Moreover, telecommunications prices will, for the foreseeable future, be determined by a complex interaction between regulation, strategic pricing and arbitrage opportunities. It is therefore not appropriate to assume either that a regulated monopoly will adopt the price vector that maximises monopoly rents or that frameworks designed to enhance competition will lead to the adoption of Pareto-optimal prices.

In practice, it seems likely that policies favoring increased competition and lower barriers to entry will lead to a reduction in usage charges and an increase in access charges, a combination often called 'rebalancing'. Albon (1988, 1991) argues that rebalancing, by bringing usage charges closer to marginal costs, will generate substantial welfare gains. However, Quiggin (1997a) observes that access charges are similar in their effects to poll taxes, and generate similar adverse equity effects. The welfare costs of tax and welfare measures sufficient to offset these equity effects may exceed the welfare gains from rebalancing.

The second argument often used in support of more competitive markets is the claim that competition will yield 'dynamic' improvements in technical efficiency . This claim has little basis in economic theory and little empirical support. Arguments that monopolies are subject to X-inefficiency (Leibenstein 1966) and rent-seeking (Krueger 1974) have been supported by little more than anecdotal evidence. The debate on monopoly and innovation, arising from Schumpeter's (1961) contention that monopolies are more innovative because of their greater capacity to capture returns to innovation, remains unresolved. Against Schumpeter's argument, it may be observed that if technological innovations, including innovations in customer service, are firm-specific, the removal of barriers to entry will maximise the probability that the most efficient firm will prevail. This point is discussed further in Quiggin (1998a).

The absence of a clear rationale for competition leads to confusion in policy design, since policies that maximise opportunities for entry will not in general, be the same as policies that encourage price competition. Policymakers in Australia appear to have been most concerned about opportunities for entry. The next section of this paper will, therefore, be concerned with ways in which the objective of increasing opportunities for entry could have been achieved at a lower cost in terms of reduced technical efficiency than under the policy of network duplication.

The crucial requirement in achieving technical efficiency is that the gains from natural monopoly should be exploited, rather than dissipated through unnecessary duplication of facilities. I will assume that duplication should be encouraged in the provision of long-distance telephone services and avoided in the physical provision of digital and analog mobile telephone networks, local copper-wire telephone networks and local optical-fibre or hybrid fibre-cable networks for telephony and pay-TV. This specification of objectives leaves open the boundaries of the natural monopoly sector , but is sufficient to suggest alternatives to the policy of network duplication.

A better duopoly

The effect of network duplication has been to maintain Telstra's monopoly of local telephony, while setting up an effective duopoly in most other segments of the market. I begin by specifying a policy framework that would have yielded greater technical efficiency and at least as much competition in all segments of the market as the policy of network duplication.

The first step, as with the policy of network duplication, would have been the admission of a second long-distance carrier. However, the transaction should have been made more transparent by unbundling the sale of Aussat and requiring a competitive tender for the right to enter. The second step would be the replication, for digital mobile telephony, of the common carrier arrangement already in place for analog. That is, a single physical network would be constructed, with both long-distance carriers having access on equal terms. The third step would have been the construction by Telstra of an optical fibre or hybrid fibre-cable network. Telstra would be excluded from the provision of pay-TV or other content, and the network would be a common carrier for content providers. Finally, the restrictions on entry by value-added resellers would have been removed immediately, rather than in June 1997.

This framework would have yielded significant efficiency gains by avoiding wasteful duplication in the provision of physical networks, and minimising the dissipation of resources in attempts to secure monopoly control of pay-TV content. It would also have provided more competition than the policy of network duplication. First, consumers would have had a continuing choice between analog and digital mobile telephone networks, at least until the digital network was clearly superior in coverage and service. Second, by paying for a single connection to the cable network, consumers would have had a wide choice of content providers. Third, with the removal of the need to provide 'infant industry' protection to Optus in the forlorn hope of building up a full-scale competitor for Telstra, competition from value-added resellers could have been allowed five years earlier.

Competition between technologies

Existing policy has focused on competition between enterprises, at the cost of making highly prescriptive decisions about technology, such as the compulsory phaseout of analog mobiles. An alternative would have been to encourage competition between technologies. Analog and digital mobile services are close substitutes and are also a substitute for wires-based local and long-distance services. Competition could have been enhanced relative to the 'better duopoly' model, by selling the analog network and the right to develop a digital network to separate firms that were not engaged in the provision of wires-based services. On the other hand, some potential economies of scope would have been foregone.

Because of the trade-off between competition and economies of scope, it is difficult to say whether the 'competition between technologies' model would have been superior to the 'better duopoly' model. Once again, however, both competition and technical efficiency would be greater than under the policy of network duplication.

A more radical extension of the 'competition between technologies' model would have created a new public or private enterprise to provide cable telephone and pay-TV services in competition with Telstra's copper wire network. Such a policy would have involved a number of difficulties. Telstra would have an incentive to forestall entry, for example by strategic investment in technologies such as Asymmetric Digital Subscriber lines (ADSL) in areas where cable rollout was likely to take place early. This is the opposite pattern to that required for a cost-minimising solution which would involve upgrading the existing copper wire network precisely in those places where a cable rollout was uneconomical. However, the waste involved in strategic investments of this kind would be less than that observed under the policy of network duplication.

A breakup of Telstra

A third possibility would be a breakup of Telstra into a set of local monopolies and a long-distance enterprise competing with new entrants, similar to the breakup of the Bell monopoly in the United States in 1976. Within any given market, the effect would be fairly similar to that of the improved duopoly model, except that it would no longer be possible to choose an integrated local and long-distance carrier. In effect, some economies of scope would be sacrificed to achieve more even competition between long-distance carriers.

The idea that natural monopolies should, where possible, be disaggregated into regional components has been popular since the Bell breakup, though the underlying reasoning is not clear. The possibility of 'benchmark competition' has been mentioned but the difficulties in defining comparable benchmarks are formidable (Quiggin 1997b). The fact that horizontal reintegration (through mergers) of US local telephone services commenced as soon as regulatory barriers to merger were relaxed suggests that breakups involve a loss of economies of scale and scope.

An assessment

All the alternatives considered above are superior to the policy of network duplication on the criteria of competition and technical efficiency. As compared to public monopoly, the policy of network duplication has yielded a modest increase in competition at a high cost in terms of technical inefficiency. Maintenance of the public monopoly would have yielded an outcome superior to that actually achieved.

Comparisons among the three options and between the options for change and the maintenance of the public monopoly are less clear-cut. My tentative judgement is that the improved duopoly model would have yielded significantly increased competition at only a modest cost in terms of technical efficiency. In view of the clear desire of policymakers for more competition, this would probably have been the optimal choice.

5. Corporatisation, privatisation and regulation of Telstra

In all the alternatives considered above, Telstra would remain the dominant supplier of local telephone services, just as it is under the policy of network duplication. Hence, regulation of access prices and final consumer prices would be necessary. The policy issues arising from the need for continued control of the local telephone monopoly are largely independent of the framework adopted for the remainder of the telecommunications system.

The central policy issue is the appropriate organisational form for the local monopoly. Possibilities range from a statutory corporation like the old Telecom Australia through a corporatised government business enterprise like Telstra (prior to partial privatisation), to a privately owned corporation. For each possible organisational structure, it is necessary to consider an appropriate regulatory framework. In general, control through public ownership is a substitute for regulation, so that the closer the ownership is to full privatisation, the more elaborate the system of regulation required to achieve any given divergence from the unregulated monopoly outcome. Thus, the choices of organisational form and of regulatory framework must be made jointly.

The central issue in regulation of monopolies is that of setting maximum prices. For short periods, systems based on a historical starting point, such as CPI-X regulation, may be adequate. However, in the long term, such regulatory systems inevitably reduce to rate-of-return regulation. The problem for a regulator is to set the rate of return high enough to induce appropriate investment, but low enough to approximate the outcome of a competitive market or a perfectly contestable natural monopoly. The result will be a repeated game between the monopoly and the regulator (acting as an agent for consumers), in which both sides face substantial risk. Welfare losses are associated with this risk and with the dissipation of resources in rent-seeking attempts to influence regulatory outcomes, either through lobbying or through strategic investment and pricing decisions.

The risk associated with the regulation game are, in large measure, internalised if the monopoly is publicly owned, since the government acts as the representative of both consumers and taxpayers, the owners of the enterprise. A gain to consumers from low prices is balanced by a corresponding loss to taxpayers. Although these two groups are not identical, there is a great deal of overlap between them. Expenditure on rent-seeking would be minimised, and formal or informal side-payments could beused to harmonise the interests of consumers and taxpayers.

In considering the desirability or otherwise of privatisation, it is necessary to balance any efficiency gains arising from privatisation against the cost of regulatory risk. The critical question is whether privatisation passes the present value test, that is, whether, with an appropriate choice of discount rate, the present value of future income under public ownership exceeds the price that private buyers are willing to pay.

The issue is complicated by the unresolved debate over the relative capacity of governments and private capital markets to bear the risk associated with systematic fluctuations in aggregate output (Hathaway 1997, Quiggin 1997c, Grant and Quiggin 1998), and hence whether the discount rate should include a large 'equity premium' (Mehra and Prescott 1985). However, rate-of-return regulation effectively insulates monopolies from systematic risk, and, as argued above, regulatory risk is internalised through public ownership, so that no risk premium is appropriate. Hence, the appropriate discount rate is the rate of interest on government bonds, and the present value test may be expressed in flow terms as a comparison between the interest savings that would be made if asset sales are used to repay debt and the income foregone through privatisation.

The partial privatisation of Telstra clearly fails the present value test. ollowing the passage of the Telstra (Dilution of Public Ownership) Act 1996, one third of the public shareholding in Telstra, One third of the government's shareholding, consisting of approximately 4.3 million shares, was sold at an average price of $3.40 yielding sale proceeds of $14 billion. At a nominal interest rate of 6 per cent, and ignoring the diversion of $1 billion to fund electoral commitments, the associated interest saving is $840 million per year.

Income foregone through privatisation comprises dividends, retained earnings and imputation credits unclaimed by the government. For 1997-98, the first year after privatisation, total profits were $3.0 billion of which $1.8 billion were paid as fully franked dividends and the remaining $1.2 billion retained (Telstra 1998b). The value accruing to the private shareholders and foregone by the public consisted of $600 million in dividends, approximately $240 million in franking credits and $400 million in retained earnings, for a total of $1240 million. Thus there was a net loss to the public of $400 million.

Unless Telstra's earnings fail to grow in nominal terms, the loss to the public will rise over time. The probability that Telstra's earnings will decline in nominal terms is minimal, regardless of whether the enterprise is fully privatised. Although a variety of projections of the expected value of future earnings can be made, the probability that the public sector will realise a profit from the partial privatisation of Telstra is close to zero.

The relatively favourable regulatory and market outcomes experienced by Telstra over the past two years have led to an increase in the market value of Telstra shares. However, this value still embodies a substantial discount associated with regulatory risk. It follows that further losses in public sector net wealth, and therefore in social welfare, can be expected to result from full privatisation (Quiggin 1998b).

The abandonment of proposals for full privatisation raises new difficulties. A partially privatised government business enterprise is an unsatisfactory hybrid. Its managers are subject simultaneously to political accountability, for example through Freedom of Information legislation, and to the fiduciary obligation to put the interests of shareholders ahead of all other concerns (Quiggin 1996). Each system of accountability is compromised by the presence of the other. For example, claims of commercial confidentiality have beenj used to override the principle of Freedom of Information. The present mixed status of Telstra is unsatisfactory. If full privatisation is not to take place in the near future, full public ownership should be restored.

6. Unscrambling the egg

Given the mistakes that have already been made, where do we go from here? There may still be an opportunity to reorganise the Australian telecommunications industry to take account of the implications of its natural monopoly characteristics. The central element of any such reorganisation would be the restoration of full public ownership of Telstra, financed by a mixture of additional debt and the sale of those components of Telstra for which there is no natural monopoly justification for public ownership. Telstra would act as a common carrier system for cable TV and local telephony allowing consumers access to all content providers regardless of the carrier supplying their access to the cable network. This would permit the divestiture of Telstra's interest in the content provider Foxtel. Telstra could also be required to divest its digital mobile telephony business, while being allowed to continue offering analog services beyond the year 2000.

The feasibility of a reorganisation of this kind is constrained by the fact that the government has given industry participants contractual guarantees that policy will not be changed. Most notably, although it is generally recognised that the decision to close down the analog mobile network was a mistake, the Labor government set this mistake in concrete through its contracts with the digital carriers, Optus and Vodafone. (By contrast, commitments made to ordinary electors, even when enshrined in L-A-W, have no binding status).

The difficulties of fixing past mistakes in telecommunications policy is an example of a more general problem. In many cases where the provision of public services is handed over to the private sector, the profitability of the private provider depends primarily on political decisions. Hence, the private provider must demand either a substantial risk premium or a guarantee against future adverse politcal decisions.

In most cases, politicians eager to secure a politically attractive deal have found it easier to provide the guarantees and leave their successors to deal with the consequences. The builders of toll roads have received guarantees against the provision of adequate public transport in the areas served by their roads. State gambling enterprises have been sold with a guarantee that no competition will be allowed. In some cases, blanket indemnities against government action of any kind have been offered. Binding the hands of future governments in this way is not only antidemocratic but also bad policy. If such guarantees are required to attract private providers of services, it is best to keep those services in the public sector.

Concluding comments

Australia, like New Zealand and the United Kingdom has undergone radical microeconomic reform over the past fifteen to twenty years. Yet, as the debated recorded in Industry Commission (1998) shows, there is no clear evidence that this reform has resulted in an increase in the sustainable rate of economic growth. An analysis of Australian telecommunications policy suggests a partial explanation for this unsatisfactory outcome. Much of the potential benefit to be derived from reform of the telecommunications industry has been dissipated in wasteful and technically unnecessary investment in duplicate networks. This is the natural result of policies based on a naive enthusiasm for competition and wishful thinking about the death of natural monopoly.

References

Arena, A.,Bahtsevanoglou, J., and Branton, S. (1992), The importance of telecommunications in Australia's economic development and the likely effects of telecommunications reform, Occasional Paper Economics No. 1, Australian Telecommunications Authority (AUSTEL), Melbourne.

Austel (1993), Competitive safeguards and carrier performance, 1992-93, AGPS, Canberra, Report to the Minister for Telecommunications.

Bain, J. (1956), Barriers to New Competition, Harvard University Press, Cambridge, Mass.

Baumol, W. (1977), 'On the proper cost tests for natural monopoly in a multi-product industry', American Economic Review 67(5), 809-22.

Baumol, W. (1982), 'Contestable markets: an uprising in the theory of industry structure', American Economic Review 72(1), 1-15.

Baumol, W., Panzar, J. and Willig, R. (1982), Contestable Markets and the Theory of Industry Structure, Harcourt Brace Jovanovich, New York.

Baumol, W., Panzar, J. C. and Willig, R. D. (1983), 'Contestable markets: an uprising in the theory of industry structure: reply', American Economic Review 73(3), 491-96.

Baumol, W. and Willig, R. (1986), 'Contestability: developments since the book', Oxford Economic Papers 38(Supplement), 9-36.

Brown, A. (1996), Reform and regulation of Australian telecommunications, Working Paper No. 5, Griffith University, School of Economics.

Committee of Inquiry into Telecommunications Services in Australia (1982), Report (Davidson Report), AGPS, Canberra.

Hathaway, N. (1997), 'Privatisation and the cost of capital', Agenda 4(1), 1-10.

Krueger, A. (1974), 'The political economy of the rent-seeking society', American Economic Review 64, 291-303.

Leibenstein, H. (1966), 'Allocative efficiency vs X-efficiency', American Economic Review 56, 392-45.

Mehra, R. and Prescott, E. C. (1985), 'The equity premium: a puzzle', Journal of Monetary Economics 15(2), 145-61.

Quiggin, J. (1997a), 'The equity premium and the government cost of capital: a response to Neville Hathaway', Agenda 4(4), 475-88.

Quiggin, J. (1997b), 'Why have the returns to microeconomic reform been so disappointing ?', paper presented at , Centre for Applied Economic Research, University of New South Wales.

Quiggin, J. (1998), The proposed privatisation of Telstra: an assessment, , Submission to Senate Environment, Recreation, Communications and the Arts Legislation Committee.

Schumpeter, J. (1961), The theory of economic development, Oxford University Press, London.

Shin, R. T. and Ying, J. S. (1992), 'Unnatural monopolies in local telephones', Rand Journal of Economics 23(2), 171-83.

Spiller, P. and Cardilli, C. (1997), 'The frontier of telecommunications deregulation: small countries leading the pack', Journal of Economic Perspectives 11(4), 127-38.

Stigler, G. J. (1958), 'The economies of scale', Journal of Law and Economics 1(October), 54-71.

Toffler, A. (1980), The Third Wave, Collins, London.

Go back to John Quiggin's home page

Go back to John Quiggin's 1998 articles