You are on page 1of 24

Efficient Investment Pricing Rules and Access Regulation

*
by

Joshua S. Gans and Philip L. Williams
Melbourne Business School University of Melbourne 3 July, 1998

This paper reviews recent advances in regulatory theory concerning the effect of access pricing regulation on incentives to invest in infrastructure. We demonstrate that regulation has a dual role of ensuring that investment costs are themselves shared by multiple users of a facility and that potential providers compete to invest in a timely manner. Our paper, therefore, provides a rationale for using fixed access charges to allocate investment costs so as to ensure timely investment and competition.

*

Parts of this paper previously appeared in Gans and Williams (1998a). We thank David Briggs and Stephen King for helpful comments. Responsibility for all views expressed in this paper lies with us. All correspondence to: Joshua Gans, Melbourne Business School, 200 Leicester Street, Carlton Victoria 3053; E-mail: J.Gans@mbs.unimelb.edu.au; Fax: (03) 9349 8133.

2 Previous discussions of the role of Part IIIA of the Trade Practices Act have focussed on the optimal use of essential facilities. From an economist’s viewpoint such facilities have a proportionately high level of sunk costs relative to operating costs. Hence, once a facility is built, social efficiency dictates that the facility be used intensively. In particular, if usage prices – or the prices of final goods that use the facility as a key input – are not close to short-run marginal (or avoidable) cost, usage will be less than socially optimal. Of particular concern in trade practices law is, of course, the role of competition in facilitating optimal usage. If a facility is owned by a vertically integrated producer, there is concern that individual facilities may be restricted in their use. This is a traditional

problem of monopoly power. In this case, the firm who owns the facility restricts access to it by other firms, thereby creating an entry barrier allowing it to effectively restrict access to it by final consumers. Part IIIA is designed to facilitate optimal use by encouraging facility owners to give access to firms who might use the infrastructure as an input. This has the advantage of avoiding costly duplication of the facility as well as reducing entry barriers to downstream competitors. Moreover, economists have a good notion of the access pricing rule that should be set to encourage such use and competition – access usage charges should by set equal to short-run marginal cost. Any higher and a facility will be under utilised. Any lower and the facility will be used too much. Issues of asymmetric information and

operating cost reductions tend to mean that in practice this benchmark is not reached.1 However, the benchmark remains an important pricing goal. The problem with this focus, however, is that it neglects the dynamic effects of access regulation. That is, while economists have established principles to guide pricing behaviour for a facility that has been built, only recently has attention been turned to the critical issue of incentives to build the facility in the first place. After all, to the extent that static access regulation diminishes the use of monopoly power over the asset, it also

3 potentially reduces profits that can be earned by a private investor. Hence, there is a

potential concern that such regulation might deter investment that would otherwise be socially optimal. After all, no competition can easily be preferred to no production at all. In order to see this problem clearly, imagine that the Trade Practices Act included lawnmowers as an “essential facility,” in this case, for the production of neat gardens. The Smith family are considering purchasing a lawnmower. However, before they do this they notice that their neighbours, the Jones family, have a nice new lawnmower. The Smiths propose to the Joneses that perhaps they could borrow their mower for one day a week. This they argue would not inconvenience the Joneses who use the lawnmower themselves for one day each week. Of course, the Smiths will compensate the Joneses for fuel used and physical depreciation caused. This offer is, of course, consistent with static notions of pricing efficiency. That is, given that the lawnmower exists and is not fully utilised by the Joneses (i.e., there is excess capacity), if the Smiths are will to bear the costs of their usage, it is socially efficient for them to be granted access to the Jones’ mower. To the extent that there is a legal stipulation for the Joneses to grant the Smiths access, so much the better. The problem, however, is that the Joneses were considering purchasing an electric weeder. All other things equal they might have decided to go ahead with the purchase, even if they did not expect to rent it out for some days of the week. Seemingly the prospect of renting would only enhance the Joneses benefits from purchasing the weeder. However, the Jones family are sophisticated thinkers. They reason that it might be better to see if someone else on the street purchases the weeder first. That household would bear the capital costs of the weeder while the Joneses could simply rent it out for one day a week. Under a proposal such as that of Smith for the mower, Jones would only have to pay for the operating expenses of the weeder – a negligible amount relative to the purchase costs.
1

For a discussion of such deviations see Armstrong et.al. (1994) and King and Maddock (1996).

4 Notice that Jones’ reasoning would not change if access seekers, such as Smith for the mower, were forced to contribute towards capital according to use. To see this,

suppose that Smith also was forced to pay Jones for one seventh of the capital costs. A potential investor, such as Jones, would still be better off waiting for another household to purchase the asset. In this case, that provider would have to bear most of the capital costs associated with the necessarily idleness that accompanies mowers and weeders. As

providers of an asset are not compensated for idleness that arises in such lumpy investors, under such access regulation they are better off being a seeker rather than a provider. For Smith and Jones’ street, access regulation based on simple cost recovery rules, while encouraging efficient utilisation of assets, discourages efficient investment. Even purchases that might have been individually optimal are delayed. Access regulation that does not respect the incentives to invest encourages a problem of free riding among potential providers. For these situations, instead of optimal usage, the access regime can potentially discourage provision and hence, any usage at all. The parable of the lawnmower is what motivates this paper. Notice that there are no intrinsic issues of competition between Smith and Jones here – although we guess Smith could just be trying to keep up with the Joneses and Jones trying to get ahead of the Smiths. Instead, there is simply an issue of how to encourage optimal usage of an investment without discouraging investing itself. Recent economic theory demonstrates that there are some principles that can resolve this difficulty.

I.

Access Pricing and Investment
This paper provides a discussion of the relationship between access regulation and

investment in essential facilities. Regulation of access terms and prices affects the return a facility provider can expect to receive on its investment. Hence, expectations of the nature of regulation affect investment incentives. A regulator, therefore, has the power to

influence investment indirectly by pre-committing to an access-pricing regime. However,

5 uncertainty and inappropriate signals can potentially have an adverse influence on investment. So the regulator must take care when stating regulatory policy and applying regulatory instruments. In Australia, Part IIIA of the Trade Practices Act (1974), gives the Australian Competition and Consumer Commission (ACCC) the power to determine prices of access to essential facilities. These powers are constrained somewhat, however. For example:
s44W(1): The Commission must not make a determination that would have any of the following effects ... (d) resulting in the third party becoming the owner (or one of the owners) of any part of the facility, or of extensions of the facility, without the consent of the provider, (e) requiring the provider to bear some or all of the costs of extending the facility (or maintaining extensions of the facility).

This indicates that the ACCC does not have absolute power to impose any form of pricing. In this paper, we wish to review how regulation affects investment in infrastructure that involves sizable sunk costs. In particular, we suppose that the infrastructure produces its services by way of a natural monopoly technology. In addition to the sunk costs of its provision, there are constant marginal costs involved in its use. Moreover, the

infrastructure effectively has unlimited capacity. It is not certain that the ACCC has the power to regulate access to such infrastructure. The Trade Practices Act requires that in order for a service to be declared it must be “uneconomic for anyone to develop another facility to provide the service.” (s44G(b)) For a natural monopoly technology, a single provider is economically efficient. However, it may be commercially viable to duplicate the facility allowing an entrant to bypass an incumbent. Here, we will cast a broad net concerning the powers of the ACCC in this regard and assume that it can regulate the pricing of access to all facilities with a natural monopoly technology. As we demonstrate below, regulation can improve social outcomes in this instance. As noted in the lawnmower parable, regulation serves to affect investment indirectly. By committing to an access pricing structure that it will implement whenever access is sought, the ACCC influences expectations. Here we focus purely on the dynamic

6 aspects of regulation and hence, assume that the regulator has complete information so that the optimal form of pricing takes to the form of a two part tariff (King and Maddock, 1996, Chapter 5). The first part of the tariff sets the usage charge equal to short-run marginal cost. Note that under incomplete information this is not necessarily the preferred solution (Laffont and Tirole, 1993; Armstrong, Cowan and Vickers, 1994; Vickers, 1995; and Armstrong, Doyle and Vickers, 1996). However, we abstract from such concerns to focus on dynamic issues. The second part of the two part tariff sets a fixed charge for access. Under complete information, the key regulatory choice that influences investment is, therefore, the choice of the fixed charge. The determination of fixed charges has always been a contentious issue in regulation. In the past, its choice has been seen as arbitrary -- essentially, redistributing income from access seeker to provider -- without any real efficiency consequences. However, from the point of view of market participants, the level of fixed charges is a contentious issue. This is because providers realise that it affects the overall return on their investments and access seekers realise it influences their incentives to enter markets and compete with incumbents. We contend that the use of fixed charges can have a key role in determining investment incentives. To this end, whenever we refer to access charges below it will concern only fixed charges. To have an effect on investment, regulation must modify incentives. A higher fixed access charge raises the incentives of firms to invest. However, it cannot be too high or seekers might have an incentive to duplicate the facility. For providers and seekers, therefore, a formula that determines their expected access charge will form a critical part of their perceived returns from their actions. This paper will demonstrate that appropriate regulatory policies can have a beneficial effect on investment incentives. We demonstrate this by comparing regulation to what occurs without regulation. In the absence of regulation, access providers are torn between their incentive to maximise the use of the infrastructure and a desire to limit

7 competition downstream. After all, intensive use of a natural technology facility reduces long-run average costs for all users. However, allowing your downstream competitors to have access to the facility reduces your ability to earn monopoly rents downstream (Rey and Tirole, 1996). In general, in an unregulated environment, providers will limit optimal use of the facility so as to limit profit-reducing competition downstream. Below we distinguish between two important cases. The first concerns a situation in which use of the investment is non-rival. In this case, alternative users of the

infrastructure do not compete directly downstream. This might include access to facilities such as towns, ports and rail lines for mines that are world price-takers or situations in which there is high product differentiation downstream. In this situation, providers do not fear the rent dissipation caused by downstream competition and hence, have an incentive to optimise the use of the facility. Even here regulation can have the beneficial role of allowing sunk investment costs to form part of access pricing -- something that would not occur in an unregulated environment. Hence, access pricing can beneficially accelerate investment decisions to the benefit of all. In contrast when the use of the investment is rival, the regulatory issues are more complex. When there is some competition downstream, a vertically integrated provider may restrict access so as to maintain monopoly rents. Here the regulator has a dilemma. By forcing the provider to grant access, it reduces the returns a provider can earn on investment. So if the expected access charge is too low, investment might be deterred. However, it is too high, alternative providers might race to investment. In this case, investment might take place with an inefficiently inferior technology.

II.

The Non-Rival Case
There are extensive deposits of iron ore in Western Australia’s Pilbara region.

Most of these were discovered in the early 1960s. These deposits vary in quality and in location. There were no existing facilities to exploit these resources. That is, there were no

8 towns, ports, railways and the like. These were developed by pioneer investors in the face of considerable uncertainty regarding their actual future use. These facilities involve natural monopoly technologies. Investment in them

continues. Users of the facilities are unlikely to compete directly with one another. If a potential user of one of these facilities came to the NCC seeking access or the ACCC for access terms, what should these regulators do? How should it determine any fixed access charges? Can it have any favourable role here? The answer is yes and this can be demonstrated by the following simple example.

Investment Costs Suppose that today (at date 0), a potential provider is considering making an investment in, say, a rail line. The current cost of that investment would be $100.

However, if it waited five years, those costs will fall to $77.88 and if it waited ten years, they would fall to $60.65. This is because of technological progress that results in a continuously falling cost of production of rail lines and engines. In present value terms, therefore, the rate of cost reduction is even greater. These costs are illustrated in Table 1 assuming an annual interest rate of 5%. Table 1: Illustrative Infrastructure Costs Time of Construction Present Value at Time 0 Current Cost Stand-Alone Choices The firm can earn total profits (over time) of $100 from the use of the infrastructure, regardless of when it invests. With this, what are its incentives to invest in the infrastructure on its own? That is, what happens if it does not believe there will be any 0 $100 $100 5 $60.65 $77.88 10 $36.79 $60.65

9 other users of the facility. Its stand-alone returns are listed in Table 2. Those figures tell us that the firm will earn the greatest return (in present value terms) by waiting ten years to invest. When it is the only user of the facility, it pays the firm to wait until the costs of investment are less prohibitive. Table 2: Stand-Alone Incentives Time of Construction Present Value of Profits Net Present Value of Investment 0 $100 $0 5 $77.88 $17.23 10 $60.65 $23.86

Social Optimum in the Presence of Multiple Users Suppose now that another user of the rail exists. While the provider knows about the potential existence of such a user, they cannot get together prior to the investment to form a joint venture. Instead, that user will seek access at some time after the rail line is constructed. Suppose that this user can earn total profits from the use of the rail of $50. Recall that in the non-rival case, allowing the user to have access does not diminish the profit opportunities of the provider. From a social point of view, it is now optimal for the rail line to be built in five years. This can be seen from the calculations in Table 3. Table 3: Social Incentives Time of Construction Present Value of Profits Net Present Value of Investment 0 $150 $50 5 10

$116.82 $90.97 $56.17 $54.18

10 No Regulation Case However, it is unlikely this outcome will be achieved when there is no regulation. Consider what happens in this case. First, the provider chooses when to invest. Then the seeker chooses when to seek access.2 Finally, the seeker and provider negotiate over the access charge. The provider’s choice in the first stage is contingent on its expectations of when the seeker will seek access and the outcome of resulting negotiations. In economics, negotiations can only be over variables that can be altered at the time of negotiation. As the investment has already taken place at this time, the infrastructure costs themselves cannot play a role in negotiations. They are sunk. This weakens the provider’s position in negotiations, lowering the access price it receives. If the provider wishes to terminate access negotiations, it can only lose. By denying the seeker the use of the infrastructure it can gain nothing and lose whatever access charge it might receive. For the seeker, if it terminates access negotiations, it might be able to duplicate the facility and still earn some profits. Table 4 depicts the seeker’s bypass incentives. Notice that the seeker never earns a positive profit by bypassing the provider’s investment. In general, this will not be the case, so that option will provide the seeker some bargaining power. Table 4: By-Pass Incentives Time of Construction Present Value of Profits Net Present Value of Investment 0 $50 -$50 5 $38.94 -$21.71 10 $30.32 -$6.47

When the by-pass option is not attractive to the seeker, both parties gain nothing by terminating negotiations. However, if an agreement is reached for a charge of p, the provider receives p above the amount it would otherwise earn, while the seeker receives

11 $50 - p. In total they jointly create $50 in additional profits by allowing the seeker to use the facility. Thus, under many bargaining games (e.g., Nash, Shapley or Rubinstein -- see Gans and Williams, 1997), the resulting agreement would have p = $25. Note that with this expected price, it is always in the interests of the seeker to seek access as soon as possible, i.e., as soon as the investment takes place, as it earns $25 sooner rather than later. The provider builds this and the expected price into its timing calculations. Its new anticipated returns from investment are depicted in Table 5. In this example, while the provider earns more from investing at each date (as compared with the stand-alone case), it still chooses to provide the infrastructure in ten years. So, in the absence of regulation, investment is delayed relative to the social optimum (Gans and Williams, 1998). Table 5: Incentives under No Regulation Time of Construction Present Value of Profits Net Present Value of Investment 0 $125 $25 5 $97.35 $36.70 10 $75.81 $39.02

The potential inefficiency arising in the no regulation environment comes from the fact that investment costs are sunk. Therefore, these costs do not enter into any

negotiations over access. Hence, as the provider only claims a proportion of seeker value in those negotiations, it does not internalise the full social benefits of the infrastructure when making its timing choice. Indeed, its private incentive is less than the social incentive, leading to delayed investment.

2

Note that the seeker here is not a potential investor. We consider the “free riding” and other implications of this below.

12 The Potential for “Racing” The above analysis makes a critical assumption: that the seeker cannot, for some reason, actually become the provider itself and sell access to the other firm. This might be because the seeker does not have good access to capital. It is worth considering what occurs if this assumption is relaxed. In this situation, the “small” firm can invest first and pre-empt the “large” firm. In this case, the tables on negotiations are turned. However, if the “larger” firm is the access seeker rather than the provider, it does have a credible bypass option. To see this, suppose that the infrastructure is built in ten years. The large firm could ensure itself a profit in cash terms at that time of $100 minus $60.65 (the current cost of investment in ten years). Unless that firm receives at least $39.35 after paying for access it will find it profitable to by-pass the “small” provider. Gans and Williams (1998) demonstrate that it is reasonable to expect the access charge in such situations to be the minimum of half the seeker’s profits and the investment costs incurred in by-passing the provider. That is, p = min[$50,$60.65] which, in this case, will equal $50. Alternative specifications, such as pure Nash bargaining, yield

slightly different results but with the same qualitative implications. As it turns out, in this example, despite the by-pass option being positive for the seeker, it never binds so p = $50 regardless of the time chosen by the “small” firm. Table 6 depicts the “small” firm’s incentives to invest given this expected access charge. Table 6: “Small” Firm Incentives under No Regulation Time of Construction Present Value of Profits Net Present Value of Investment 0 $100 $0 5 $77.88 $17.23 10 $60.65 $23.86

13 The key point here is that there is the potential for both firms to race in order to be the first to provide the infrastructure. By doing so, that firm avoids having to pay the other for access and, moreover, receives an access payment. In the unregulated case, each firm, regardless of their size, can expect to receive an additional $75 (= $50 + $25) by becoming the provider rather than the seeker. This prize raises their investment incentives by giving them a reason to pre-empt the other. However, this prize must be traded off against the actual investment costs. Providers have to pay these while seekers do not. To see this, consider the “large” firm’s decision. It knows that if it does not invest in five years, then the “small” firm will surely invest in ten years time. In present value terms, the difference is $6.38. Hence, the “large” firm is potentially better off by preempting the other firm. Similarly, the “small” firm using the same reasoning will gain $2.07 by investing early. Both would be individually better off if they could invest later, but the “race” between them speeds up their choices. However, this advantage only goes so far. Pre-empting the other firm at time turns out to be too costly for both firms. What will happen in equilibrium? Gans and Williams (1998) demonstrate that the equilibrium under racing speeds up investment relative to the situation in which only one firm can possibly be the provider. While this might align the timing choice closer to the social optimum it could actually do too much. A firm might invest with an inefficiently costly technology in order to become the provider. Hence, investment might take place too soon. So while, in the absence of regulation, racing might occasionally help, it could also lead to overinvestment.

Optimal Use of Regulation In Australia, access regulation is not necessarily mandatory. Instead, Part IIIA of the Trade Practices Act sets up a system of “regulation by negotiation.” That is, when infrastructure is provided, a seeker first approaches the provider to privately negotiate access terms. If they fail to reach an agreement, the seeker can either by-pass the facility or

14 approach the NCC seeking a declaration. If a declaration is granted, the parties may negotiate a price or the parties might approach the ACCC to impose a regulated access regime. The point is that regulation and expected regulatory prices play a key role in the outcomes negotiated by the provider and seeker. The seeker can either by-pass the facility or seek a regulated solution. The provider can also go to regulation by refusing the grant access. It is more difficult for a provider to prevent by-pass. For the purposes of this paper, we shall simplify the regulatory process by assuming that when regulation is a possibility, by whatever means, the stated regulatory price is actually imposed. For regulation to be most effective, therefore, pricing policy must be stated prior to access being sought and indeed, prior to investment being made. That is, it is important for the ACCC to establish pricing guidelines that are more precise than those it has currently adopted. While, as mentioned earlier, any regulatory price should have a usage charge set equal to short-run marginal cost, what should the fixed access charge be? One possibility is for the regulator to use the access charge to allocate the investment costs between the provider and the seeker. That is, if these costs are C, then p = αC where α < 1. Notice that this brings those costs into the price -- a situation that did not arise in the unregulated environment. When considering investment cost allocation, the first question to ask is: what basis should be used to value those costs? For instance, if infrastructure was built in five years for a cost of $77.88 and access is sought at that time, it is clear what the value of the infrastructure is. However, if access is only sought in ten years, the issue is more

contentious. One could use a historical cost methodology which sets the asset valuation at the actual costs incurred less depreciation. Since there is no depreciation in our example, this would be $77.88. Alternatively, one could use a replacement cost methodology that values the investment at the amount it would cost to replace the investment at the latest

15 technology. In this case, replacement cost would be $60.65.3 Gans and Williams (1998) demonstrate that any optimal regulatory policy should provide incentives for an access seeker to do so as soon as possible. If it is paying a share of historical costs, a seeker will choose to seek access at the earliest possible date, as delay does not alter the amount it pays. On the other hand, under replacement cost, a seeker might have an incentive to delay so as to reduce the charge it has to pay. Nonetheless, it will be demonstrated below that it is possible to find a cost sharing rule that overcomes this difficulty. Gans and Williams (1998) demonstrate that by setting α equal to the relative share of the seeker in total profits then, regardless of the method of asset valuation, investment takes place at the socially optimal time and access is sought immediately. Recall that, in our example, we want investment to take place and access to be sought in five years time. If α = 1/3 = ($50/($50+$100)) when the seeker is “small” and 2/3 when it is “large,” then this occurs. Note that if investment takes place in five years, under historical cost, the seeker always seeks immediately. Under replacement cost, by seeking access in five years rather than ten years a “small” seeker saves $0.64 in present value terms while a “large” seeker saves $1.31. So access is sought immediately regardless of which firm is the provider. Given this, a provider now expects to receive an access payment immediately. In effect, a “large” provider expects to incur 2/3 of the investment costs whenever it invests, while a “small” provider expects to incur 1/3 of those costs. Indeed, each firm is now indifferent between being a provider and being a seeker. So there is no incentive to preempt the other. Regulation removes any adverse incentives associated with this. What happens to actual timing under these conditions. Table 7 depicts the relevant payoff calculations for the “large” and “small” providers. Notice that both find it optimal to choose to invest in five rather than ten years. In equilibrium, one of them will do so. Notice also that each firm earns more in present value terms than it would have under the

3

See King (1996) for an excellent discussion of issues in asset valuation.

16 no regulation case. This is because investment is taking place at a time that maximises their joint returns. Table 7: Incentives under Regulation Present Value of Profits Large Firm Small Firm 0 $33 $17 5 $37.44 $18.72 10 $36.13 $18.06

Using a cost sharing rule based on relative profits, the regulator can ensure that infrastructure is provided and used at the socially optimal time. It does this by removing pre-emption incentives and aligning individual firm timing choices with the social optimum. Regulation allows the actual investment costs to be shared -- something that cannot occur without regulation unless firms enter into a joint agreement prior to investment taking place. However, if this could actually occur then there would be no access issue to be concerned about. This point is worth emphasising. The efficient investment price is the same price that the firms would agree upon if they were to enter into a joint venture. In that venture the public good aspects of the infrastructure are internalised. So the venture would choose to invest at the socially optimal date as we have assumed there are no other beneficiaries. Gans and Williams (1998) demonstrate that this is exactly the same as a Lindahl equilibrium in public good problems. That is, if we asked -- as did Lindahl (1919) -- if the firms were to nominate a time to invest, under what cost allocation rule, would they agree to the same date? It turns out that the cost allocation rule based on relative profitability would achieve this outcome. Therefore, the role of regulation in a non-rival environment is to achieve an outcome that mimics what would occur if the relevant parties had been able to form a joint venture agreement.

17

III. The Rival Case
The access issues that are perhaps more common are where alternative users of a facility compete in some downstream market. In some cases, this competition might be weak. For instance, a seeker might produce a product differentiated from the provider’s. Hence, the profits of the provider may only be slightly affected. Here the provider might still grant access in the absence of regulation because industry profits are enhanced by so doing. In many cases, however, the competitive effects are more extreme. The provider might be relinquishing a monopoly position by granting access. As monopoly profits

generally exceed total industry profits under more competitive conditions, the provider will not grant access unless required to. If industry profits were the only concern, then it would be best to allow a monopoly if that is profitable and to use the optimal pricing discussed in the previous section when multiple firms maximise industry profits. However, in general, we are concerned about competition because it can reduce deadweight losses. Hence, consumer surplus under competition is greater and, moreover, the increment to consumer surplus more than offsets any losses in industry profits. So, in general, access regulation is designed to allow for competition where it would not arise in an unregulated environment. As identified earlier, however, if industry participants are to provide the infrastructure they will only take into account their individual profits when determining when to invest. Timely competition will, therefore, reduce those incentives. To see this suppose that firm A receives a profit of $100 if it competes with another firm B and $200 if it is a monopolist. Suppose that firm B earns $50 under competition and $150 if it is able to act as a monopolist. Note that even if the provider were to receive a payment equal to the profits of its rival under competition the most they could earn would be $150. However, a monopolist could earn more than this. Firm A, for instance, could earn $200. Hence, if it was guaranteed a monopoly firm, A would invest earlier than if it expected competition.

18 This problem arises in many contexts. Take, for instance, the issue of local number portability (LNP) in telecommunications. To provide LNP one must invest in digital switching technology and other hardware. However, LNP also has a competitive impact by allowing consumers to easily switch between telecommunications firms. This increases the intensity of competition those firms face and hence, reduces industry profits. An

entrant is more likely to desire LNP than an incumbent but often it falls on the incumbent to make the investment. Without compensation, however, the incumbent will wish to delay the introduction of LNP. This can be solved by requiring the entrant to bear costs greater than would be freely negotiated; but for the regulator such burdens might reduce the possibility of entry itself. How should the regulator manage the tension between providing investment incentives and allowing for timely competition? In Gans (1998), it is demonstrated that appropriate access pricing can be used to create competition for the provision of infrastructure. Where in Section I, regulation removed pre-emption incentives. Here regulation can manage them. In particular, an appropriate access pricing formula can ensure that infrastructure is provided at the socially optimal date (including consumer and producer surplus) despite competition being expected immediately. And all this without the use of government subsidies. How is this optimal “race” achieved? An important requirement is that there are at least two potential providers of the infrastructure. Moreover, it must be the case that, when a single facility is produced, but there is competition among firms, all firms earn a positive economic profit: that is, the cost of investment at the socially optimal date is less than total industry profits under competition at that date. Therefore, competition among firms cannot be too intense.4 Given this feasibility condition, the access charge itself determines the “prize” in competition for infrastructure provision. Recall that the difference in profits between providing infrastructure and seeking access to it is simply the sum of access

19 charges that each firm would expect to pay. Therefore, by carefully choosing the basis upon which access prices are charged the regulator can ensure that firm race to provide the infrastructure at the socially optimal time. This is despite the expectation of immediate competition. The racing feature that accompanies lumpy investments in infrastructure means that regulation can be used to provide appropriate incentives for both timely investment and timely competition. Moreover, in Gans (1998), the optimal pricing formula is derived for the case of a duopoly. The access price is simply the sum of the investment cost sharing rule described in the previous section and a proportion of the marginal social benefit of providing the infrastructure sooner rather than later. Essentially, the access charge is a function of the (replacement or historical) cost of the investment with factors adjusting for the rate of technological progress and the flow of social surplus. Effectively, the share of costs paid by the provider falls relative to the seeker over time. If this rate is aligned properly, it can be such that the equilibrium in the “racing” game between market rivals coincides precisely with the socially optimal date. No participant can gain anything from pre-empting their rivals at this point. And to delay beyond this point is to invite preemption and a higher cost share. Practically this suggests that, in order to encourage optimal competition for provision, access seekers should pay a charge reflecting a greater share of investment costs if they do not seek access immediately or are late into the industry. This gives providers a sufficient bonus to ensure they invest in a timely manner. Thus, for issues such as LNP, this suggests that entrants should contribute relatively more (taking into account their size and profitability) than the incumbent provider.5 To do otherwise would send a poor signal
4

This requires some form of Cournot quantity competition in oligopoly or substantial product differentiation among firms. 5 In its draft determination on local number portability, the ACCC has appeared to require the provider (Telstra) to bear all of the investment costs of the infrastructure. The seeker (Optus) will only bear costs associated with customer switching. In doing this the ACCC appears to have sent a signal to others in telecommunications that providers will bear all infrastructure costs and hence, will generally be better off waiting for others to invest.

20 to incumbent providers: diminishing their incentives to invest in new infrastructure (Gans and King, 1998).

I V . Conclusions
The introduction of Part IIIA to the Trade Practices Act is the most important change to Australia’s antitrust law since the introduction of the statute in 1974. The

example that has been explored in this paper suggests certain lessons as to how the access regime of Part IIIA should be implemented and it raises a question about the drafting of the new legislation. The first lesson concerns the value of certainty in the regulatory regime. The ACCC and the Tribunal should move as quickly as possible to develop pricing rules for arbitrations under Part IIIA. The example presented above shows that the principles that the ACCC and the Tribunal eventually adopt for pricing arbitrations under Part IIIA will constrain the range of pricing outcomes within which the parties can bargain for access. Such principles can only have important implications for the timing of major infrastructure investments if firms are aware that they will be imposed before they actually invest. Indeed, the pricing principles may influence whether or not projected infrastructure projects are judged to be viable; and uncertainty could deter otherwise socially optimal investment. Of course, the ACCC and the Tribunal must assess each application on its merits. Nevertheless, early cases before the ACCC and the Tribunal will be critical in influencing subsequent decisions by firms as to the timing and viability of major infrastructure investment projects. The second lesson concerns the allocation of investment costs among users. While it might be tempting for the ACCC and the Tribunal to treat any given investment as sunk, any regime that allocates investment costs sends important signals to future investors. The ACCC and the Tribunal should ensure that the access prices compensate investors for a

21 portion of the replacement value of the assets to whose services access is sought. Replacement cost causes potential providers to consider in an appropriate way the possibilities of others seeking access and it induces potential providers to take this into account in their timing decisions. This paper has explored the effects on investment incentives of fixing the compensation for investment expenditure in a standard multi-part pricing scheme. The

paper has not explored financial accounting historical cost as a basis for pricing because it is obvious that such a basis would have adverse effects on incentives to invest -- and, indeed, on incentives to seek access. If the mere elapse of time were to decrease the price of access, the access seeker would have an incentive to delay access; and this delay by the access seeker could create an incentive for any potential access provider to delay investment. The historical cost that was explored in the paper is defined according to economic principles -- it is a cost that makes no allowance for the technical progress that has occurred between the time of investment and the time at which access is sought. The analysis shows that even this (economic) notion of historical cost only has the efficiency properties of a replacement cost measure when, because access is sought immediately after investment, historical cost happens to equal replacement cost. In general, however, replacement cost is the preferred method of valuing the compensation for investment expenditure in a multi-part tariff. A third lesson from the above analysis goes to the method by which the value of the asset should be divided between the access provider and the access seeker. The analysis suggests that capital costs should be apportioned according to the relative economic profit that is expected to accrue to the provider and the seeker. We say expected profit because the capital contribution must be levied as a lump sum when access is sought -- and

22 must not vary with post-access decisions or else those decisions may be distorted by the capital charge. 6 Finally, our analysis suggests that the criteria for declaration under Part I I I A may be too narrow in excluding from declaration services provided by facilities that would be economical for any other person to duplicate. Section 44H(4)(6) states that the designated minister cannot declare a service unless he or she is satisfied “... that it would be uneconomical for anyone to develop another facility to provide the service.” Under Part IIIA, unless the service is declared, the ACCC and the Tribunal have no influence on bargaining over price between the access provider and the access seeker in this case. The analysis presented above suggests that, providing the ACCC and the Tribunal develop appropriate principles to govern pricing, the influence of these constraints on pricing can promote efficiency in the timing of investments. Although the words

“uneconomical for anyone to develop another facility” have yet to be tested, it may well be that they will be found to correspond closely to the distinction between small and large investments in the model. If this were the case, the analysis suggests that it may be better to amend s44H(4)(6). The reason is that, in the example used, price regulation can form a socially useful function by removing the incentive for the access provider to bring forward an investment (in a wasteful way) to gain a pre-emptive advantage. The problem with s44H(4)(6) is that it seems to refer to the private profitability of an investment rather than the social desirability of duplication. An alternative would be to re-draft s44H(4)(6) so as to deny access unless the facility were a natural monopoly -- that is, unless the service would be most efficiently provided by having only one facility. In the case of mines, it may well be that the mining leases owned by both the access provider and the access seeker are expected to generate operating surpluses that would justify investment in parallel railway lines to take ore to the coast. Nevertheless, the railway line may be a

6

Note that this implies that access seekers should contribute towards the capital costs of any excess capacity that is part of the infrastructure.

23 natural monopoly in that duplication would be inefficient. In this case, the mining

companies may well negotiate access independently of Part IIIA. However, re-drafting s44H(4)(6) to include such a case would ensure that the negotiations over price take place with the framework established by the ACCC and the Tribunal. Providing the ACCC and the Tribunal develop appropriate rules they could enhance the socially efficiency of private decisions to invest in infrastructure.

24

References
Armstrong, M., S. Cowan and J. Vickers (1994), Regulatory Reform: Economic Analysis and British Experience, MIT Press: Cambridge (MA). Armstrong, M., C. Doyle and J. Vickers (1996), “The Access Pricing Problem: A Synthesis,” Journal of Industrial Economics, 44 (2), pp.131-150. Gans, J.S. (1998), “Regulating Private Infrastructure Investment: Optimal Pricing for Access to Essential Facilities,” mimeo., Melbourne. Gans, J.S. and S.P. King (1998), “When Being First Doesn’t Pay,” Australian Financial Review, Friday 30th January, p.32. Gans, J.S. and P.L. Williams (1998a), “A Primer on Access Regulation and Investment,” in M. Arblaster & M. Jamison (eds.), Infrastructure Regulation and Market Reform: Principles and Practice, ACCC/PURC, Melbourne. Gans, J.S. and P.L. Williams (1998b), “Access Regulation and the Timing of Infrastructure Investment,” Working Paper, No.13, Melbourne Business School. Katz, M.L. and C. Shapiro (1987), “R&D Rivalry with Licensing or Imitation,” American Economic Review, 77 (3), pp.402-420. King, S.P. (1996), “Asset Valuation and Access to Essential Facilities Under Part IIIA of the Trade Practices Act 1974 (Cth), in M. Richardson (ed.), Deregulation of Public Utilities: Current Issues and Perspectives, Centre for Corporate Law and Securites Regulation, Melbourne, pp.94-116. King, S.P. and R. Maddock (1996), Unlocking the Infrastructure, Allen & Unwin: Sydney. King, S.P. and R. Maddock (1997), “Regulation by Negotiation: A Strategic Analysis of Australia’s Essential Facility Access Regulation,” mimeo., ANU. Laffont, J.-J. and J. Tirole (1993), A Theory of Incentives in Procurement and Regulation, MIT Press, Cambridge (MA). Lindahl, E. (1919), Die Gerechtigkeit der Besteuering, Gleerup: Lund, Part I, Chapter 4, “Positive Losung,” trans. by E. Henderson and reprinted as “Just Taxation -- A Positive Solution,” in Classics in the Theory of Public Finance, R.A. Musgrave and A.T. Peacock (eds.), Macmillan: London, 1958. Rey, P. and J. Tirole (1996), “A Primer on Foreclosure,” Handbook of Industrial Organization, Vol.3, Elsevier, Amsterdam, (forthcoming). Vickers, J. (1995), “Competition and Regulation in Vertically Related Markets,” Review of Economic Studies, 62 (1), pp.1-18.