Abstract:

This article considers the vertical structure of the
telecommunications industry and examines the welfare effects of
privatization on the public enterprise, where the public enterprise
supplies the essential network service of the upstream market and
competes against private, independent firms in a mixed downstream
market. It is shown that the cost advantage of the independent rivals
improves welfare postprivatization. Several relevant and current policy
issues on the process of privatization in the telecommunications
industry are also discussed. (JEL L50, D42)

Subject:

Privatization
(Laws, regulations and rules)

Privatization (Economic aspects)

Privatization (Analysis)

Telecommunications services industry (Laws, regulations and rules)

Telecommunications services industry (Economic aspects)

Communications industry (Laws, regulations and rules)

Communications industry (Economic aspects)

Privatization (Economic aspects)

Privatization (Analysis)

Telecommunications services industry (Laws, regulations and rules)

Telecommunications services industry (Economic aspects)

Communications industry (Laws, regulations and rules)

Communications industry (Economic aspects)

Author:

Lee, Sang-Ho

Pub Date:

04/01/2006

Publication:

Name: Contemporary Economic Policy Publisher: Western Economic Association International Audience: Academic; Trade Format: Magazine/Journal Subject: Business; Economics Copyright: COPYRIGHT 2006 Western Economic Association International ISSN: 1074-3529

Issue:

Date: April, 2006 Source Volume: 24 Source Issue: 2

Topic:

Event Code: 930 Government regulation; 940 Government regulation (cont); 980 Legal issues & crime Advertising Code: 94 Legal/Government Regulation Computer Subject: Telecommunications services industry; Government regulation

Product:

Product Code: 4800000 Communications NAICS Code: 513 Broadcasting and Telecommunications SIC Code: 4810 Telephone Communication

Geographic:

Geographic Scope: United States Geographic Code: 1USA United States

Accession Number:

144921676

Full Text:

I. INTRODUCTION

For decades, there has been a monopoly in telecommunications services controlled by a stable firm, a public enterprise in most of the world. The absence of competition was motivated by the existence of large fixed costs in several parts of the network, whose duplication was neither privately profitable nor socially desirable.

However, the telecommunications industry has been changing rapidly. Due to technological improvement and the political movement toward market liberalization, there is growing interest in substituting competition and privatization for regulation and nationalization in the telecommunications industry. The poor economic and financial performance of many public enterprises and the cases of successful privatization have been used as arguments for privatization and competition. In this worldwide policy-transition situation, policy makers need to know under what conditions privatization of the public enterprise will increase welfare.

In the economic literature there is conventional knowledge on the benefits and costs of privatization. For example, Laffont and Tirole (1993, pp. 637-59) and Viscusi et al. (1995, pp. 453-74) discussed the important regulatory policy issues of public enterprises. Vickers and Yarrow (1988, 1991) and Megginson and Netter (2001) used empirical research to assess the effects of privatization as a public policy. They concluded that enterprises operating under public ownership will be less efficient than their private sector counterparts.

However, standard economic theory is not particularly helpful in understanding the welfare effects of privatization in the telecommunications industry. Little attention has been devoted to incentives in publicly owned firms, even though a "natural monopoly" has been widely used as an argument for regulatory subjects.

In the context of vertical integration, on the other hand, market closure has for a long time been the prime policy concern. For example, Economides (1998), Mandy (2000), Weisman and Kang (2001), and Hackner (2003) have examined several academic debates on market foreclosure and discrimination in the telecommunications industry.

The structure of the telecommunications industry is characterized by an upstream monopolist who supplies an input essential to the competitive downstream firms that are vertically integrated. In other words, the monopolist provides the upstream services and the downstream services as well, and also competes with several firms in the competitive downstream markets. Thus the downstream market structure consists of a mixed market. Because a public enterprise competes with independent, profit-maximizing downstream firms in the telecommunications industry, game-theoretic analysis on privatization policy in vertically mixed market should be examined.

Theoretic research on privatization in a mixed market has been widely studied. De Fraja and Delbono (1989, 1990) showed that welfare might be higher when a public enterprise is a profit-maximizer rather than a welfare-maximizer in an imperfect competition market model. However, they did not consider the privatization effects on improving productivity. Matsumura (1998) and Lee and Hwang (2003) considered the possibility of partial privatization in a Cournot duopoly model and showed that it is optimal for the government to sell some but not all of its shares in public enterprises when there exist production-efficiency effects of partial privatization.

This article considers the telecommunications industry with a vertical market structure to investigate the welfare effects of privatization in a mixed downstream market where public enterprise competes with independent private firms. It is shown that the cost advantage of the independent rivals improves welfare postprivatization. It is also shown that a privatization policy in the telecommunications industry will be well justified if the competition in the downstream market is fierce and the access charge is regulated at certain level.

It then extends the basic model to discuss the policy implications of privatization, which include the structural separation between upstream and downstream firms, the role of leadership in public enterprise, the objectives of the public enterprise and its managerial incentives in agency relationship, the possibility of partial privatization, and the strategic choice of access charge by the privatized monopolist.

The structure of this article is as follows: In section II, the author constructs a simple Cournot model in a downstream market using linear market demand and constant marginal costs, analyze two market equilibria between pre- and postprivatization, and compare the welfare results. The analysis finds the degree of cost advantage of the independent rivals improves welfare postprivatization. Several extensions with abstracted forms and policy implications on the issues of privatization are discussed in section III. The final section gives concluding remarks.

II. THE BASIC MODEL AND ANALYSIS

Consider a monopolist providing both an essential upstream and downstream services. There are assumed to be k - 1 identical, independent downstream rivals of the monopolist, where k [greater than or equal to] 2. The author assumes that downstream firms provide a homogenous product, and the inverse demand of a downstream market is a linear form of P(Q) = a - bQ; where Q = [q.sup.m] + [[summation].sub.j=1.sup.k-1][q.sub.j.sup.n] is the market demand, which the monopolist provides with essential facilities in an upstream market, [q.sup.m] and [q.sub.j.sup.n] represent the outputs of the monopolist and rivals in a downstream market, respectively. Then, the author denotes [q.sup.n] as the equilibrium output of the representative independent rival in a downstream market. The author assumes that the monopolist engages in Cournot competition against private firms in the downstream market.

Production of the downstream service is of the fixed-coefficient type. Each unit of the downstream output requires one unit of the upstream service and a complementary input that may be self-supplied by an independent rival. The marginal cost of the upstream service is c and the regulated input price (or access charge) of the downstream market is assumed to be fixed as r where r [greater than or equal to] c. (1)

The per unit cost of the complementary input is [s.sup.i], where [s.sup.i] [greater than or equal to] 0 and i = m, n denote the monopolist and the representative independent downstream rival, respectively. In addition, for the traceable and interior solutions throughout the analysis where both the monopolist and rivals produce the nonnegative outputs in the coexistence equilibrium, the author needs the following assumptions about [DELTA] [equivalent to] [s.sup.m] - [s.sup.n]:

A1. [DELTA] [greater than or equal to] 0.

A2. [DELTA] [greater than or equal to] r - c.

A3. [DELTA] [less than or equal to] (r - c) + (a - r - [s.sup.n])/k.

A1 indicates that there is an efficiency gap between the monopolist and independent rivals, and the independent firms are more cost efficient than the monopolist. Otherwise, the privatization policy always decreases the welfare in the model. (2) A2 represents that the efficiency gap is greater than the price-cost margin of the upstream monopolist, which is nonnegative. It ensures that the monopolist is not able to exclude the independent rivals from the downstream market. Thus, the rivals will produce nonnegative outputs in the model. Finally, A3 implies that the efficiency gap is large enough that the inefficient monopolist can produce the output in the downstream market. Otherwise, the monopolist will not provide the downstream service in the model.

Then, the profit functions for the monopolist and the independent rival are given by, respectively,

(1) [[pi].sup.m] = (r - c)[k-1.summation over (j=1)][q.sub.j.sup.n] + (P(Q) - c - [s.sup.m])[q.sup.m],

(2) [[pi].sub.j.sup.n] = (P(Q) - r - [s.sup.n])[q.sub.j.sup.n].

In the profit function of the monopolist, the first term comes from the access profit by selling access to the independent rivals, and the second term comes from the operating profits by selling downstream services in a downstream market.

The consumer surplus in the downstream market is then

(3) CS = [[integral].sub.0.sup.Q]P(v)dv - P(Q)Q,

and social welfare, which is defined as the simple sum of consumers surplus and firms' profits, is given by

(4) W = [[pi].sup.m] + [k-1.summation over (j=1)][[pi].sub.j.sup.n] + CS,

(5) = [[integral].sub.0.sup.Q]P(v)dv - (c + [s.sup.m])[q.sup.m] - (c + [s.sup.n])[k-1.summation over (j=1)][q.sub.j.sup.n]

The author assumes that the government maximizes social welfare while the private firms focus on profits in the following analysis. This implies that the monopolist will maximize social welfare in (4) when it takes the form of public enterprise preprivatization, whereas it will maximize its profit in (1) postprivatization. (3)

Now, the author will examine the equilibrium outcomes by the public enterprise preprivatization and the outcomes by the privatized monopolist postprivatization, respectively, and then compare the two equilibrium outcomes. First, the author considers the preprivatization case where the upstream monopolist takes the form of a public enterprise and produces its output to maximize the goal of the government, the social welfare function in (5). Then, the first-order condition for the public enterprise with interior solution yields

(6) P(Q) = c + [s.sup.m].

Notice that at the equilibrium the market price is exactly equal to the marginal cost of the monopolist, c + [s.sup.m], which is assumed to be larger than the marginal cost of the independent rivals, r + [s.sup.n]. This is so because the public enterprise is only able to control its output level, not that of the rival when it maximizes social welfare in the Cournot competition situation. (4) Notice also that the operating profit level of the public enterprise is always nonnegative at equilibrium if r [greater than or equal to] c. (5)

The first-order condition for the independent rival's profit in (2) with the interior solution is as follows:

(7) P(Q) = r + [s.sup.n] + b[q.sub.j.sup.n].

Then, from the profit function in (2), the author knows that the profit level of the rival firm in a downstream market is always positive at equilibrium.

Using the equilibrium output of the representative rivals and combining the equations in (6) and (7), one obtains the following Cournot-Nash equilibrium output levels for the public enterprise and private rival:

(8) [q.sub.N.sup.m] = (X - kY)/b,

(9) [q.sub.N.sup.n] = Y/b,

where X = a - r - [s.sup.n] [greater than or equal to] kY and Y = [DELTA] - (r - c) [greater than or equal to] 0 as assumed. Then, [q.sub.N.sup.m] [greater than or equal to] 0, [q.sub.N.sup.n] [greater than or equal to] 0, and [q.sub.N.sup.m] [greater than or equal to] [q.sub.N.sup.n] if X - kY [greater than or equal to] Y, [q.sub.N.sup.m] [less than or equal to] [q.sub.N.sup.n] if X - kY [less than or equal to] Y. It means that an inefficient public enterprise might produce greater output than the efficient independent rival firms preprivatization. This is because the public firm extends the output to where the market price is equal to its marginal cost even though it is inefficient. In sum, in preprivatization, the total output is [Q.sub.N] = (X - Y)/b and the equilibrium price is [P.sub.N] = C + [s.sup.m].

Next, the author considers the postprivatization case with the upstream monopolist and thus, the privatized monopolist maximizes its own profit function in (1). Then, the first-order condition for the privatized monopolist with interior solution yields

(10) P(Q) = c + [s.sup.m] + b[q.sup.m].

Then, from the profit functions in (1) and (2), the author knows that all firms in a downstream market have positive profits at equilibrium.

From the first-order condition of the individual rival in (7), one gets the following Cournot-Nash equilibrium output levels:

(11) [q.sub.p.sup.m](X - kY)/b(k + 1),

(12) [q.sub.p.sup.n](X + Y)/b(k + 1).

Then, at the postprivatization equilibrium the efficient independent firm produces greater output than the inefficient privatized monopolist: [q.sub.p.sup.n] [greater than or equal to] [q.sub.p.sup.m]. In sum, under the privatized monopolist situation, the total output is [Q.sub.P] = (kX - Y)/b(k + 1) and the equilibrium price is [P.sub.p] = a - (kX - Y)/(k + 1).

Now, the author compares the two equilibrium outcomes pre- and postprivatization. Notice that the output of the inefficient monopolist decreases postprivatization, whereas output of the efficient independent rival increases post-privatization: [q.sub.N.sup.m] > [q.sub.p.sup.m] and [q.sub.N.sup.n] < [q.sub.p.sup.n]. This yields the following two propositions. (6)

PROPOSITION 1. Suppose that the assumptions in Al, A2, and A3 hold. Then, all firms produce nonnegative output and earn nonnegative profits in the Cournot-Nash equilibrium in which we have:

(i) [Q.sub.N] > [Q.sub.P] and [P.sub.N] < [P.sub.p],

(ii) [[pi].sub.N.sup.m] < [[pi].sub.p.sup.m] and [[pi].sub.N.sup.n] < [[pi].sub.p.sup.n].

Proposition 1 indicates that (i) the consumer surplus is decreasing but (ii) the profits of both individual firms and total industry profit ([[pi].sup.m] + [k - 1][[pi].sup.n]) are increasing postprivatization. It means that it is beneficial for the firms, although it might be harmful for the public to implement a privatization policy. Hence, privatization might be harmful to social welfare when the privatized monopolist reduces its output level too much. Put differently, the inefficient monopolist might be beneficial to social welfare when the inefficient monopolist produces greater output that could not be produced by more efficient firms. As a result, only when the welfare-increasing effects of cost reduction from the private firms is greater than the welfare-decreasing effects of the reduction in total output from public enterprise, social welfare will increase postprivatization. Therefore, welfare will be increasing in psotprivatization if the superiority of the cost efficiency of the independent rivals outweighs the negative effects of the decrease in total output. Proposition 2 indicates the condition where social welfare is increasing postprivatization.

PROPOSITION 2. Define R [equivalent to] (X - kY)/2([k.sup.2] - 1) [greater than or equal to] 0. Then, one has [W.sub.N] [less than or equal to] [W.sub.P] if R - [DELTA] [less than or equal to] 0 or [W.sub.N] [greater than or equal to] [W.sub.P] if R - [DELTA] [greater than or equal to] 0.

Proposition 2 shows that the welfare change in postprivatization depends on the relative size of the cost efficiency gap between the monopolist and the rivals. A few remarks are in order.

First, one can see that R = 0 when [DELTA] = (r - c) + X/k and R is decreasing in [DELTA]. It implies that there exists a unique threshold level of the efficiency gap such that privatization improves the welfare, that is, [W.sub.N] [less than or equal to] [W.sub.P]. Specifically, R [less than or equal to] [DELTA] if [DELTA] [greater than or equal to] (X + k[r - c])/(sk[.sup.2] + k - 2). Thus, privatization is beneficial to society only when the efficiency gap between the monopolist and the rivals is large. In sum, the cost efficiency superiority of the independent rivals will be a necessary condition to improve welfare postprivatization.

Second, R is decreasing in k while R = (X - 2Y)/6 [greater than or equal to] 0 when k = 2. It implies that the welfare in postprivatization tends to be increasing as the number of downstream firms is increasing. Therefore, privatization will be beneficial to society only if the competition in the downstream market is sufficient. (7)

Finally, R is increasing in r. It implies that welfare postprivatization will be decreasing if the regulated access charge (the input price of a upstream service) is high. In particular, privatization directly benefits society only if r [less than or equal to] c + ([DELTA][2k[.sup.2] + k - 2] - X)/k. Thus, the necessary condition to increase social welfare is [DELTA] [greater than or equal to] X/(2[k.sup.2] + k - 2). Otherwise, privatization is always harmful to social welfare. Therefore, it is important for the regulatory agency to maintain the regulated access price at a certain level even post-privatization.

III. EXTENSIONS AND DISCUSSIONS

In the telecommunications industry, it is common for an upstream monopolist to supply an essential input to private firms operating in the downstream market. Furthermore, usually in developing areas, such as Asian, European, and Latin American countries, the upstream monopolist has taken a form of a public enterprise. Therefore, the downstream market has consisted of a mixed market where the public enterprise competes with independent private firms.

Recently, the government activated both privatization and competition policies in mixed markets. Now, it is a real concern of policy makers whether the privatization of the upstream public firm, which will induce the upstream monopolist to engage in profit-maximizing strategies, yields welfare-increasing outcomes in the market. In the simple model presented, the welfare implications of privatization depend primarily on the relative cost efficiency between upstream monopolist and downstream firms. However, there are many other important policy aspects that have been abstracted for reasons of tractability and simplicity. The article will extend the analysis into several theoretical issues and discuss some policy-relevant implications.

A. Vertical Separation and Competition Strategies

The author has confined the analysis into the vertical integration model where the upstream monopolist vertically integrates the downstream firm and competes against the other independent rival firms. However, as shown in Proposition 2, in certain conditions of the cost-efficiency gap, the public enterprise could not bring about greater social welfare compared to the case where it is a privatized firm. This is because the public enterprise in a downstream market produces greater output than that when it is privatized, that is, [q.sub.N.sup.m] > [q.sub.p.sup.m] even though it is inefficient compared to the rival firms.

From a different policy perspective, if the public enterprise can be separated into an independent upstream firm and independent downstream firm, and the upstream firm can be managed in a form of public enterprise while the downstream firm is privatized, the market outcomes will be the same as those of complete privatization of the public enterprise. In this case, without privatizing the public enterprise in both upstream and downstream markets together, the government can increase social welfare in the telecommunications industry.

However, it might be technologically or politically inefficient to separate the vertically integrated monopolist into two markets. (8) For example, if high technology can be applied to the connection (bundling) between the upstream and downstream services (through coutilizing facilities and human resources), then it will be optimal to integrate two markets vertically because technological economies and dynamic investments (e.g., stable supply of telecommunications services) could be realized. Politically, in addition, it might lead to huge social costs to separate the historical public utility into different firms if the government cannot make an agreement among employees and public citizens. Therefore, before separating into independent upstream and downstream firms, it is necessary both to check if there are technological or political linkages between two markets and to consider how the regulatory agency is able to treat these problems without high social costs.

For instance, the Korean government announced plans to privatize its electricity power utility (Korea Electric Power Corporation), which was a government-invested monopolist that supplied electric power in Korea. During 2001, the Korean electric power industry underwent major changes as its power generation unit was separated into six subsidiaries and the Korea Power Exchange was inaugurated. In addition, the power generation subsidiaries are supposed to be privatized, and it is preparing separate power distribution units. But there still remain many debates between the government and employees (or even the public). Therefore, it is expected that there are ongoing social costs in both the privatization and separation processes.

On the other hand, there is the other policy debate on privatization and separation in the telecommunications industry. One might argue that without separation between upstream and downstream firms, the welfare could be increased if the public enterprise acts as a Stackelberg leader. (9)

As an intermediate step toward dynamics, the public enterprise can make a strategic situation where it restricts its choice to gain strategic advantages. For example, if the public enterprise can act as a leader in a mixed market, it will be able to set [q.sup.m] = [q.sub.P.sup.m] to induce the rival firms to produce [q.sub.P.sup.n], which would give the same market outcomes as privatization of the monopolist.

However, the question remains whether the inefficient public enterprise can be a leader in the telecommunications market. If the public enterprise can commit its output level to the rival firms in a credible way and the rival firms take its commitment into consideration effectively, then the public enterprise can obtain the same social welfare level with that of postprivatization. But if the public enterprise lacks leadership and cannot commit its output level in advance, then the strategy of choosing [q.sup.m] is not a best response of the monopolist when the rival produces [q.sub.P.sup.n]. As has been shown, [q.sub.P.sup.m] is not the equilibrium output for the monopolist under a Cournot-Nash type competition, where the monopolist is in interactive strategic competition with rivals. Therefore, it is important for a public enterprise to find an effective way to hold and sustain the role of leadership in the mixed market structure. If Stackelberg leadership is not available to the public enterprise, then policy makers should consider privatization of the public enterprise as an alternative.

Finally, one might also consider the other competition pattern in the telecommunications market where downstream firms set fees competitively. (10) For example if the downstream firms compete in the form of Bertrand price competition, the market price will be determined at P(Q) = r + [s.sup.n] in (7) and thus, [q.sup.m] = 0 at equilibrium, irrespective of nationalization or privatization. It implies that the welfare level does not depend on the privatization policy as long as there exists price competition among the firms in a downstream market.

In sum, the welfare consequences of a privatization policy depend not only on the competition patterns of between upstream and downstream firms, that is, Cournot versus Stackelberg, but also on the competition patterns within downstream firms, that is, Cournot versus Bertrand.

B. Objectives of the Public Enterprise

In the basic model, it is worthwhile to reconsider the objective function of the public enterprise, which was defined as the total social welfare in (5). This comes from the assumption that the public enterprise maximizes social welfare, the objective of the government. In the ideally hypothesized environment where the government has complete information and absolute authority, the public enterprise will maximize social welfare. However, due to incomplete information or costly monitoring, the objective of public enterprise will differ from that of government. The theoretical treatment of the ownership effects deduced from the property rights and principal-agent perspectives should be considered to examine the efficacy of the incentive system that is designed to maximize the efforts of the agents.

One of the policy considerations to capture the interest of the government is that the public enterprise will be guaranteed to support its nonnegative operating profit. For example, if r < c, the operating profits level of public enterprise will be negative at equilibrium, that is, [[pi].sup.m] = (r - c)(k - 1)Y/b in (1). In this case, the government needs to subsidize a lump-sum subsidy to support the public enterprise in the market equilibrium. If not, the enterprise may not survive. To meet the nonnegative operating profit conditions, for instance, the government might transfer the public funds when the public enterprise takes loses in the business. Then, the objective of the government will be CS + (k - 1)[[pi].sup.n] + [[pi].sup.m] + (1 - [mu])T, where T is the lump-sum transfer to the firm and thus, [mu]T captures the cost of the extra distortions created elsewhere in the economy.

In the analysis, however, if the public enterprise maximizes social welfare, the equilibrium outcomes in this market are the same in Proposition 1, as long as the transfer has the form of a lump sum. However, the amount of transfer might be increasing postprivatization since the rivals' output level, for which the monopolist should provide the access service with a negative price margin, will be increasing, that is, [q.sub.N.sup.n] < [q.sub.P.sup.n]. Therefore, the welfare will likely be decreasing postprivatization.

In the past, in fact, governments in developing areas considered telecommunication services as basic public infrastructure, and thus tried to encourage the usage of telecommunications services by both setting the access price below its cost and subsidizing the operating deficits of the public enterprise. However, it became a policy concern that subsidization yielded not only social welfare loss but also social costs of public funds. In recent years, therefore, the government has changed its policy to set the access price at the exact cost level. (12)

The other regulatory possibility of the government is that the public enterprise might be regulated to maximize the consumer surplus only under the constraint of its nonnegative operating profit. (13) Then the objective of the public enterprise is to maximize a part of social welfare instead of total social welfare, where, without considering the rival's profit in (2), it maximizes the consumer surplus in (3) with its own profit constraint in (1). Specifically, the objective function of the public enterprise will be CS + [lambda][pi][.sup.m] where [lambda] [greater than or equal to] 0. Note that it is simply the sum of consumer surplus and monopolist's own profits when [lambda] = 1. Then, assuming the interior solution again, the first-order condition is as follows:

P(Q) = (c + [s.sup.m]) - b(Q - [lambda][q.sup.m])/[lambda].

Using the first-order condition for the representative independent rival in (7), one can get the following equation in the Cournot-Nash equilibrium:

[q.sup.n] = Y/b - (Q - [lambda][q.sup.m])/[~.[lambda]]

Thus, if 1 [less than or equal to] [lambda] [less than or equal to] Q/[q.sup.m], (14) the market price and the rival's output level will be lower than the outcomes in (6) and (9), where the public enterprise maximizes total social welfare. Consumer surplus and the inefficient public enterprise's output will be certainly increasing compared to (6) and (8), respectively. Therefore, social welfare will be certainly decreasing compared to the case where the public enterprise maximizes total social welfare instead of maximizing a part of social welfare. It implies that in the case that the public enterprise maximizes some parts of social welfare there is a greater possibility to increase social welfare of postprivatization.

On the other hand, based on the principal-agent theory, the author can also incorporate the existence of information asymmetry and the absence of competition mechanism in the public sector to explain the nontotal social welfare-maximizing strategies of public enterprise. That is, the realistic alternative for the objective of the public enterprise could be to consider private incentives in the public ownership environment. For example, it is possible for the manager of the public enterprise to incur self-interested and inefficient expenditures, such as waste in the form of goldplating, (accounting and managerial) cross-subsidies to the other businesses, excessive employee compensation, and so on. Such "wasteful" expenditures can arise from political reasons, lack of manager's incentives to economize, or principal-agent problems in general. (15) These wasteful expenditures return fringe benefits to the decision maker of the public enterprise.

It is also noteworthy that the most important policy aspect of privatization for the government is to induce the public enterprise to achieve cost efficiency by reducing its managerial inefficiency. (16) In other words, the policy makers tend to believe that an inefficient production cost level of the privatized public enterprise would be decreasing and finally, will be equal to the cost level of the independent private firm in postprivatization.

The author will assess this policy view by incorporating the managerial inefficiency of the public enterprise into the model. For simplicity, let [s.sup.m] = [s.sup.n] + e, where e(0 [less than or equal to] [e.bar] [less than or equal to] e [less than or equal to] [bar.e]) denotes the managerial inefficiency term in the objective function of the firm, such as wasteful expenditure, which is independent of the act of production. The author will assume that this expenditure will cause fringe benefit per output, g(e), where g' > 0, and cost (management compensation from the principal) per output, h(e), where h' < 0, for the manager of the firm and thus for the society as well. Denoting d(e) [equivalent to] g(e) - h(e), where d' > 0, the author then has the following profit function of the monopolist and social welfare:

[[pi].sup.m] = (r - c)[k-1.summation over (j=1)][q.sub.j.sup.n] + (P(Q) - c - [s.sup.n] - e)[q.sup.m],

W = ([[integral].sub.0.sup.Q]P(v)dv + d(e)[q.sup.m]) - (c + [s.sup.n] + e)[q.sup.m] - (c + [s.sup.n])[k-1.summation over (j=1)][q.sub.j.sup.n]

Thus, the manager of the public enterprise will maximize W + d(e)[q.sup.m], the sum of social welfare, the goal of the government, and its total fringe benefits, whereas the manager of the privatized firm will maximize [[pi].sup.m] + d(e)[q.sup.m], the sum of its profits and its total fringe benefits, respectively. Then, it is easily shown that in certain conditions for d', specifically 1/2 < d' < 1, the monopolist will set e = [e.bar] when it acts as a privatized firm in postprivatization, and e = [bar.e] when it acts as a public enterprise in preprivatization.

Then, one obtains the following two propositions.

PROPOSITION 3. Suppose that [s.sub.N.sup.m] > [s.sub.P.sup.m] = [s.sup.n] and [DELTA]' = [s.sub.N.sup.m] - [s.sup.n]. Then, at the Cournot-Nash equilibrium one has:

(i) If X - kY > [DELTA]', [q.sub.N.sup.m] > [q.sub.P.sup.m], [q.sub.N.sup.n] > [q.sub.P.sup.n], [Q.sub.N] > [Q.sub.P], and [P.sub.N] < [P.sub.p].

(ii) If X - kY < [DELTA]', [q.sub.N.sup.m] < [q.sub.P.sup.m], [q.sub.N.sup.n] < [q.sub.P.sup.n], [Q.sub.N] < [Q.sub.P], and [P.sub.N] > [P.sub.p].

Proposition 3(i) indicates that privatization might be harmful to social welfare when the privatized monopolist reduces its output level too much, which is the same policy implication in Proposition 1. But, Proposition 3(ii) shows that social welfare will be always increasing under post-privatization since the production of the inefficient firm is displaced with the production of more efficient firms and thus, the total output provided is increased.

PROPOSITION 4. Define R' [equivalent to] (X - kY - [DELTA]')[.sup.2]/2(kX - [k + 1]Y + [r - c]). Then, one has [W.sub.N] [less than or equal to] [W.sub.P] if R' - [DELTA]' [less than or equal to] 0 or [W.sub.N] [greater than or equal to] [W.sub.P] if R' - [DELTA]' [greater than or equal to] 0.

Proposition 4 shows that the welfare change postprivatization depends on the superiority of the cost efficiency between pre- and postprivatization. If X - kY < [DELTA]' in preprivatization, then R' < 0 and thus [W.sub.N] < [W.sub.P]. This result comes from the reduction of the wasteful expenditures postprivatization. However, if X - kY > [DELTA]' preprivatization, then the welfare effects of privatization in Proposition 2 applies. That is, privatization is beneficial to society when (1) the efficiency gap between the public enterprise and the rivals is large, (2) the competition in the downstream market is enough, and (3) the access charge is regulated not too high.

C. Partial Privatization

During privatization, the government may be able to manage the activities of the privatized firm by controlling its portion of shares. In other words, there is a possibility that full privatization, where the government sells all its shares in a public enterprise, is not fulfilled at once. Thus, the government might determine the degree of privatization instead of the extreme full privatization.

In the public economics literature, De Fraja and Delbono (1989) and De Fraja (1991) examined the efficiency of full privatization in an oligopoly market. Extending their works, Matsumura (1998) and Lee and Hwang (2003) considered the possibility of partial privatization and showed that it is optimal for the government to sell part (but not all) of its shares in public enterprises in the context of mixed duopoly.

To examine the possibility of partial privatization in the model, let the privatization parameter be [mu], which takes a value in a unit interval [0,1], and a higher value of [mu] indicates a higher degree of privatization. In particular, [mu] = 1 represents full privatization, and [mu] = 0 indicates no privatization. Then, for any given degree of privatization, the partially privatized monopolist will maximize the weighted average of the payoff to the government W and its own profit [[pi].sup.m].

Thus, one can specify the objective of the partially privatized firm as follows: (17)

(13) [U.sup.m] = (1 - [theta])W + [theta][[pi].sup.m].

Note that the fully privatized monopolist ([mu] = 1) will maximize its usual profit in (1), whereas the fully public enterprise ([mu] = 0) will maximize social welfare in (5), the sum of consumer surplus and industry profits.

The first-order condition for [U.sup.m] yields

(1 - [theta])[[[partial derivative]W]/[[partial derivative][q.sup.m]]] + [theta][[[partial derivative][[pi].sup.m]]/[[partial derivative][q.sup.m]]] = 0,

or, equivalently,

p(Q) = c + [s.sup.m] + [theta]b[q.sup.m].

Using the first-order condition of the rivals in (7), the Cournot-Nash equilibrium output levels are given by

(14) [^.q.sup.m] = (X - kY)/b(k[theta] + 1),

(15) [^.q.sup.n] = ([theta]X - Y)/b(k[theta] + 1).

The effects of the government's decision parameter [mu] on the equilibrium level of output can be easily shown as follows: [[partial derivative][^.q.sup.m]/[partial derivative][theta] < 0, [partial derivative][^.q.sup.n]/[partial derivative][theta] > 0, and [partial derivative][^.Q]/[partial derivative][theta] = [partial derivative][^.q.sup.m]/[partial derivative][theta] + (k - 1)[partial derivative][^.q.sup.n][partial derivative][theta] < 0. Thus, as the degree of privatization increases, the public enterprise's output decreases but the rival's output increases and totally, the equilibrium market output decreases monotonically.

To investigate the optimal degree of privatization of a public enterprise, the author needs to check the following condition for the optimum degree of privatization that maximizes the government's payoff, W:

(16) [dW]/[d[theta]] = [[[partial derivative][q.sup.m]]/[[partial derivative][theta]]][[[partial derivative]W]/[[partial derivative][q.sup.m]]] + [[[partial derivative][q.sup.n]]/[[partial derivative][theta]]][[[partial derivative]W]/[[partial derivative][q.sup.n]]],

Substitution of equilibrium conditions in (14) and (15) into this equation in (16) yields

[dW]/[d[theta]] = [[X - kY]/[b(k[theta] + 1)[.sup.2]]] x ([DELTA] - [theta](k - 1)(X - kY)/(k[theta] + 1))

It is noteworthy that dW/d[mu] takes a positive sign at [mu] = 0. Therefore, the no privatization of the public enterprise is not an optimum in this model. The author thus obtains the necessary conditions for the existence of a positive interior solution for [mu], which is that dW/d[mu] takes a negative sign at [theta] = 1, that is, [DELTA] < (k - 1)(X - kY)/(k + 1) or [DELTA] < (k - 1) (X + k[r - c])/(k + 1). It implies that the full privatization is socially beneficial only when the cost efficiency gap is sufficiently high, [DELTA] [greater than or equal to] (k - 1)(K + k [r - c])/(k + 1). Otherwise, the partial privatization is socially beneficial in the model.

D. Discrimination and Access Charge

When policy makers consider privatization policy, there is a concern about market foreclosure in which the profit-maximizing, privatized integrated firm may have an incentive to discriminate the access from the rival firms to the upstream market and thus earn a strategic gain by increasing the costs of downstream competitors.

In the regulatory economics literature of the telecommunications industry, there have been discussions on discriminating access charges from the private monopolist. Economides (1998), for example, claimed that the incentive to discriminate is independent of the relative efficiency of the monopolist and the independent rivals. But, Weisman and Kang (2001) showed that discrimination always arises in equilibrium when the monopolist is no less efficient than its rivals in the downstream market. Sibley and Weisman (1998) analyzed the interrelationship between the market size and the incentive of discrimination in the telecommunications industry. Empirically, using the data of the U.S. telecommunications market, Mandy (2000) tested the parameters that affect the monopolist's incentives to discriminate and concluded that there is an incentive to discriminate in the U.S. telecommunications market.

The author will examine the possibility of discriminating access charge under postprivatization and its welfare effects in the model.

First, from the market equilibrium outcomes in (8) and (9) of pre-privatization, and (10) and (11) of postprivatization, one has [partial derivative][p.sub.N.sup.m]/[partial derivative]r > [partial derivative][q.sub.P.sup.m]/[partial derivative]r > 0 and [partial derivative][p.sub.N.sup.n]/[partial derivative]r < [partial derivative][q.sub.P.sup.n]/[partial derivative]r < 0. It means that the higher access charge is preferable to the monopolist, whereas the lower access charge is beneficial to the rival firms. In addition, the output effects of access charge in the case of preprivatization is stronger than that of postprivatization. In this sense, it will be the main policy concern for the regulatory agency who seeks to make fair competition in the telecommunications industry, to set the access charge not too high in either pre- and postprivatization.

This concern can be also applied to social welfare. From the welfare function in (5), one obtains the following relationship at each equilibrium:

[partial derivative]W([q.sub.P.sup.m], [q.sub.P.sup.n])/[partial derivative]r < [partial derivative]W([q.sub.N.sup.m], [q.sub.N.sup.n])/[partial derivative]r < 0.

It indicates that the welfare effects of lowering access charge in postprivatization is stronger than that of pre-privatization. Thus, it will be socially beneficial to set the access charge as low as possible. For instance, under the assumption A3 in section II, the author has the following optimal access charge rate for the society:

r* = c - (a - k[DELTA] - c - [s.sup.n])/(k - 1).

Notice that r* = c when k goes to the infinity. Notice also that r* < c when (a - k[DELTA] - c - [s.sup.n])/(k - 1) > 0 or [DELTA] < (a - c - [s.sup.n])/k, that is, the cost efficiency gap is sufficiently small. (18)

On the other hand, if the monopolist can set the access price r by its own will postprivatization, it may strategically engage in discriminatory behavior in a downstream market by setting the higher r. Then, it might reduce the welfare-increasing effects of privatization because the efficient rivals would produce less output. However, as long as the cost efficiency of the rival firms is superior to the monopolist, the monopolist can earn higher profits from access service if the rivals increase their outputs postprivatization. Therefore, there is a trade-off for the monopolist between profits arisen from the increment of its own output when rivals decrease their outputs and profit arisen from the access of the rivals when they increase their outputs.

To investigate this, the author can find the following profit-maximizing level of access charge from the profit function of the monopolist in (1):

[r.sub.P] = c + ([k + 3]X - [k - 1]Y)/2(k + 1),

or

[r.sub.P] = (a + c - [s.sup.n])/2 - (k - 1)[DELTA]/2(k + 3).

A few remarks are in order. First, [r.sub.P] > c. It implies that the privatized monopolist will set a higher access charge then the possible access charge rate under preprivatization. This fact might call for the intervention of the regulatory agency to control the access charge rate in post-privatization.

Second, [partial derivative][r.sub.P]/[partial derivative]k < 0. It means that the monopolist's profit-maximizing access charge will be reduced when the government introduces more competitive rivals in a downstream market. Therefore, an open access policy of the upstream essential input is important in the telecommunications industry.

Finally, [partial derivative][r.sub.P]/[partial derivative][DELTA] < 0. If the efficiency superiority of the independent rivals prevails sufficiently, then it will be still socially beneficial for society to privatize the public enterprise in certain conditions even though the monopolist is able to discriminate the access charge postprivatization. Compared to the result of Proposition 2, however, welfare will be reduced under postprivatization. (19)

On balance, the regulatory agency needs to construct an appropriate access charge regulation for the privatized monopolist and implement open access policy for the rivals in a downstream market. (20)

IV. CONCLUSION

This article has considered a simple vertical network structure of the telecommunications industry where a public enterprise supplies the essential network service to downstream private firms while the public enterprise competes with the independent firms in a mixed downstream market. Then, the author examined the equilibrium outcomes between pre- and postprivatization and compared the welfare effects of privatization of the public enterprise.

The analysis has shown that the cost advantage of the independent rivals in a downstream market improves the welfare if the upstream public enterprise is privatized. It is also shown that competition policy or open access regulation including access charge regulation are required to improve social welfare postprivatization. Therefore, the recent trend of both privatization and competition policy in the telecommunications industry is well justified. The author has also discussed several policy debates on the process of privatization in the telecommunications industry. Current policy issues, including vertical separation between upstream firm and downstream firms, the role of leadership of the public enterprise, the objectives of the monopolist and its managerial incentives in agency relationship, the possibility of partial privatization, and the strategic choice of access charge of the privatized monopolist, have been examined with abstracted forms.

For future policy consideration, it is worthwhile to investigate the financial treatment in the process of privatization and/or the fundraising effects of privatization for the government. These policy-relevant concerns should be carefully investigated before the government implements the privatization process of the public enterprise in the telecommunications industry.

REFERENCES

Anderson, S. P., and M. Engers "Stackelberg versus Cournot Oligopoly Equilibrium." International Journal of Industrial Organization, 10, 1992, 127-35.

Baumol, W. J., and J. G. Sidak. Toward Competition in Local Telephony. Cambridge, Mass.: MIT Press, 1994.

Cook, P., and R. V. Fabella. "The Welfare and Political Economy Dimensions of Private versus State Enterprise." Manchester School, 70, 2002, 246-61.

De Fraja, G. "Efficiency and Privatization in Imperfectly Competitive Industries." Journal of Industrial Economics, 39, 1991,311-21.

De Fraja, G., and F. Delbono. "Alternative Strategies of a Public Enterprise in Oligopoly." Oxford Economic Papers, 41, 1989,302-11.

______. "Game Theoretic Models of Mixed Oligopoly." Journal of Economic Surveys, 4, 1990, 1-17.

Dowrick, S. "Von Stackelberg and Cournot Duopoly: Choosing Roles." RAND Journal of Economics, 17, 1986, 251-60.

Economides, N. "The Incentive for Non-Price Discrimination by an Input Monopolist." International Journal of Industrial Organization, 16, 1998, 271-84.

Hackner, J. "Vertical Integration and Competition Policy." Journal of Regulatory Economics, 24, 2003, 213-22.

Haskel, J., and A. Sanchis. "Privatization and X-Inefficiency: A Bargaining Approach." Journal of Industrial Economics, 43, 1995, 301-21.

Laffont, J.-J., and J. Tirole. A Theory of Incentives in Procurement and Regulation. Cambridge, Mass.: MIT Press, 1993.

______. Competition in Telecommunications. Cambridge, Mass.: MIT Press, 2000.

Lee, S. H. and H. S. Hwang. "Partial Ownership for the Public Firm and Competition." Japanese Economic Review, 54, 2003, 324-35.

Mandy, D. M. "Killing the Goose That May Have Laid the Golden Eggs: Only the Data Know Whether Sabotage Pays." Journal of Regulatory Economics, 17, 2000, 157-72.

Matsumura, T. "Partial Privatization in Mixed Duopoly." Journal of Public Economics, 70, 1998, 473-83.

Megginson, W. L., and J. M. Netter. "From State to Market: A Survey of Empirical Studies on Privatization." Journal of Economic Literature, 39, 2001, 321-89.

Panzar, J. C, and D. S. Sibley. "Optimal Two-part Tariffs for Inputs: The case of Imperfect Competition." Journal of Public Economics, 40, 1989, 237-49.

Sappington, D. E. M., and D. S. Sibley. "Regulatory Incentive Policies and Abuse." Journal of Regulatory Economics, 5, 1993, 131-41.

Sibley, D. S., and D. L. Weisman. "The Competitive Incentives of Vertically Integrated Local Exchange Carriers: An Economic and Policy Analysis." Journal of Policy Analysis and Management, 17, 1998, 74-93.

Vickers, J., and G. Yarrow. Privatization: An Economic Analysis. Cambridge, Mass.: MIT Press, 1988.

______. "Economic Perspectives on Privatization." Journal of Economic Perspectives, 5, 1991, 111-32.

Viscusi, K., J. Vernon, and J. Harrington. Economics of Regulation and Antitrust. Cambridge, Mass.: MIT Press, 1995.

Vives, X. Oligopoly Pricing: Old Ideas and New Tools. The MIT Press, 1999.

Weisman, D. L., and J. Kang, "Incentives for Discrimination when Upstream Monopolists Participates in Downstream Markets." Journal of Regulatory Economics, 20, 2001, 125-39.

SANG-HO LEE*

*This is a revision of a paper presented at the KASIO and KIEA fall conference (2003, Seoul), Chonnam National University (2003, Gwangju, Korea), Fudan University (2003, Shanghai), and Western Economic Association International 79th annual conference (2004, Vancouver). The author is grateful for the comments and suggestions made by Oz Shy, Yongyeop Sohn, Jongguk Park, Zhao Chen, Zonglai Kou, Margriet F. Caswell, Kristen Monaco, and two anonymous referees. This study was financially supported by Chonnam National University in the 2003 program.

Lee: Associate Professor of Economics, Chonnam National University, 300 Yongbong-dong, Bukgu, Gwangju, South Korea. Phone 82-62-530-1553, Fax 82-62-530-1559, E-mail sangho@chonnam.ac.kr

1. In the telecommunications industry, the input price of the downstream market is regulated by the government or the regulatory agency such as the Federal Trade Commission in United States. The rationale of access charge regulation can be found in Laffont and Tirole (2000). In the following analysis, the author assumes that access charge r is fixed both pre- and postprivatization. Section III will discuss on the welfare effects of changing access charge in postprivatization.

2. To focus on the welfare effects arisen from the cost difference, the author assumes that [s.sub.i] is independent of the ownership structure, that is, [s.sup.i] is fixed both pre- and post-privatization. Section III will incorporate the possibility of cost reduction of postprivatization where [s.sup.m] = [s.sup.n].

3. There might be a discrepancy between the goal of the government and that of the public enterprise. See Laffont and Tirole (1993) on this point. The author will review the principal-agent relationship in section III.

4. Section III will discuss about the possibility of leadership of public enterprise and commitment problem.

5. From the profit function in (1), the author knows that the operating profit level of a public enterprise depends on the price-cost margin of the upstream market. Section III will examine the objective of public enterprise under the possibility of negative operating profits and public transfer.

6. The straight calculation yields the outcomes of propositions and thus, the proofs of all propositions are omitted.

7. In a mixed model with identical and increasing marginal costs, De Fraja and Delbono (1989, 1990) showed that if the market is sufficiently competitive, then it is socially better for the public enterprise to maximize its own profit instead of maximizing welfare.

8. There might be another theoretical issues on the increase of transaction costs between two markets or/and double marginalization effects when the public enterprise is privatized at both markets and maximizes its own profits in the upstream market. On this point, see Viscusi et al. (1995).

9. The competition patterns of Cournot or Stackelberg depend on the strategic economic environments among firms. In the economics literature, therefore, many studies have considered the order of play in a game and compared the results of Cournot and Stackelberg. On this point, see Dowrick (1986), De Fraja and Delbono (1989, 1990), Anderson and Engers (1992), and Vives (1999), among others. For example, De Fraja and Delbono (1989, 1990) show that if the public enterprise is able to act as a leader to induce Stackelberg competition in a mixed market, it is possible to increase its social goal.

10. In general, the price competition model can be applied for the case of product differentiation where the firms sell different products to consumers. Laffont and Tirole (2000), for example, provide the model of network competition in the telecommunications market. In a product-differentiated market, however, the symmetry between Cournot quantity competition and Bertrand price competition can be established with the duality argument. See, for example, Vives (1999).

11. Cook and Fabella (2002) considered the model in which the state-owned enterprise maximizes an unspecified objective function and examined the theoretical treatment of the welfare and political economy dimensions of the choice between public ownership and privatization.

12. On the theoretical treatments on the access price regulation including ECPR (Efficient Component Price Rule) or LRIC (Long-Run Incremental Cost), see Baumol and Sidak (1994), Sibley and Weisman (1998), and Laffont and Tirole (2000).

13. The regulatory incentives of the government to take care of market price or consumers surplus will yield the Ramsey solutions. On this point, see Vickers and Yarrow (1988).

14. From the second-order condition, it is sufficient to hold that [lambda] [greater than or equal to] 1.

15. The managerial inefficiency is defined as "waste" and "abuse" in the literature on the agency model of regulatory economics. On modeling of the managerial inefficiency, for examples, see Vickers and Yarrow (1988), Sappington and Sibley (1993), Laffont and Tirole (1993), and Cook and Fabella (2002).

16. Haskel and Sanchis (1995) and Lee and Hwang (2003) claim that one of the goals of the privatization policy will be to reduce the managerial X-inefticicncy of the public enterprise and to obtain the dynamic efficiency of the privatized firm.

17. For the case that r < c, the nonnegative operating profit constraint should be taken into the consideration of the government. For example, if the government subsidizes a transfer to the firm, the firm's profit will be [[pi].sup.m] + T while social welfare will be W + (1 - [mu])T, where T is the lump-sum transfer paid to the firm and thus, [mu]T captures the cost of the extra distortions created elsewhere in the economy. Therefore, the partially privatized firm will maximize [^.U.sup.m] = (1 - [theta])(W + [1 - [mu]T] + [[theta]([[pi].sup.m] + T) = (1 - [theta])W + [theta][[pi].sup.m] + (1 - [mu] + [theta][mu])T = [U.sup.m] + (1 - [mu] + [theta][mu])T. If T is the lump-sum style, then the analysis will be exactly same to the following examination. On the other hand, if T is dependent of the negative size of the operating profit, [[pi].sup.m], the government should devise an appropriate subsidy scheme. Simply, if the government subsidizes the exact amount of the negative size of the operating profit, T = -[[pi].sup.m] and thus, the subsidized profit level of the firm will be zero. In this case, one can rewrite that [^.U.sup.m] = (1 - [theta])(W + [1 - [mu]]-[[pi].sup.m]) = (1 - [theta])(2 - [mu])([1 - [beta])W + [beta][-[[pi].sup.m]]) where [beta] = (1 - [mu])/(2 - [mu]). Therefore, the author can also analyze the welfare effects of the partial privatization with the similar process in the following examination.

18. Panzar and Sibley (1989) showed that a welfare-maximizing regulator can set the marginal price of an upstream input below its marginal cost to offset downstream market power and thereby induce downstream firms to produce the efficient level of outputs.

19. Numerical example: Let a = 10, b = 1, c = 0, k = 2, and [s.sup.m] = 5. Assuming that regulated access charge r = c under pre- and postprivatization, then, from the Proposition 2, one has [W.sub.P] < [W.sub.N] when [s.sup.n] is below about 4.29 or [DELTA] > 0.71. But if the privatized monopolist can set the access charge, [r.sub.P] = 4.5 - 0.42[s.sup.n]. Then the welfare is increasing under postprivatization only when [s.sup.n] is below about 3.62 or [DELTA] > 1.38.

20. In reality, even postprivatization, the government keeps the regulatory power to control the access charge by organizing independent regulatory institutions, such as OFTEL in the United Kingdom and FTC in Korea.

For decades, there has been a monopoly in telecommunications services controlled by a stable firm, a public enterprise in most of the world. The absence of competition was motivated by the existence of large fixed costs in several parts of the network, whose duplication was neither privately profitable nor socially desirable.

However, the telecommunications industry has been changing rapidly. Due to technological improvement and the political movement toward market liberalization, there is growing interest in substituting competition and privatization for regulation and nationalization in the telecommunications industry. The poor economic and financial performance of many public enterprises and the cases of successful privatization have been used as arguments for privatization and competition. In this worldwide policy-transition situation, policy makers need to know under what conditions privatization of the public enterprise will increase welfare.

In the economic literature there is conventional knowledge on the benefits and costs of privatization. For example, Laffont and Tirole (1993, pp. 637-59) and Viscusi et al. (1995, pp. 453-74) discussed the important regulatory policy issues of public enterprises. Vickers and Yarrow (1988, 1991) and Megginson and Netter (2001) used empirical research to assess the effects of privatization as a public policy. They concluded that enterprises operating under public ownership will be less efficient than their private sector counterparts.

However, standard economic theory is not particularly helpful in understanding the welfare effects of privatization in the telecommunications industry. Little attention has been devoted to incentives in publicly owned firms, even though a "natural monopoly" has been widely used as an argument for regulatory subjects.

In the context of vertical integration, on the other hand, market closure has for a long time been the prime policy concern. For example, Economides (1998), Mandy (2000), Weisman and Kang (2001), and Hackner (2003) have examined several academic debates on market foreclosure and discrimination in the telecommunications industry.

The structure of the telecommunications industry is characterized by an upstream monopolist who supplies an input essential to the competitive downstream firms that are vertically integrated. In other words, the monopolist provides the upstream services and the downstream services as well, and also competes with several firms in the competitive downstream markets. Thus the downstream market structure consists of a mixed market. Because a public enterprise competes with independent, profit-maximizing downstream firms in the telecommunications industry, game-theoretic analysis on privatization policy in vertically mixed market should be examined.

Theoretic research on privatization in a mixed market has been widely studied. De Fraja and Delbono (1989, 1990) showed that welfare might be higher when a public enterprise is a profit-maximizer rather than a welfare-maximizer in an imperfect competition market model. However, they did not consider the privatization effects on improving productivity. Matsumura (1998) and Lee and Hwang (2003) considered the possibility of partial privatization in a Cournot duopoly model and showed that it is optimal for the government to sell some but not all of its shares in public enterprises when there exist production-efficiency effects of partial privatization.

This article considers the telecommunications industry with a vertical market structure to investigate the welfare effects of privatization in a mixed downstream market where public enterprise competes with independent private firms. It is shown that the cost advantage of the independent rivals improves welfare postprivatization. It is also shown that a privatization policy in the telecommunications industry will be well justified if the competition in the downstream market is fierce and the access charge is regulated at certain level.

It then extends the basic model to discuss the policy implications of privatization, which include the structural separation between upstream and downstream firms, the role of leadership in public enterprise, the objectives of the public enterprise and its managerial incentives in agency relationship, the possibility of partial privatization, and the strategic choice of access charge by the privatized monopolist.

The structure of this article is as follows: In section II, the author constructs a simple Cournot model in a downstream market using linear market demand and constant marginal costs, analyze two market equilibria between pre- and postprivatization, and compare the welfare results. The analysis finds the degree of cost advantage of the independent rivals improves welfare postprivatization. Several extensions with abstracted forms and policy implications on the issues of privatization are discussed in section III. The final section gives concluding remarks.

II. THE BASIC MODEL AND ANALYSIS

Consider a monopolist providing both an essential upstream and downstream services. There are assumed to be k - 1 identical, independent downstream rivals of the monopolist, where k [greater than or equal to] 2. The author assumes that downstream firms provide a homogenous product, and the inverse demand of a downstream market is a linear form of P(Q) = a - bQ; where Q = [q.sup.m] + [[summation].sub.j=1.sup.k-1][q.sub.j.sup.n] is the market demand, which the monopolist provides with essential facilities in an upstream market, [q.sup.m] and [q.sub.j.sup.n] represent the outputs of the monopolist and rivals in a downstream market, respectively. Then, the author denotes [q.sup.n] as the equilibrium output of the representative independent rival in a downstream market. The author assumes that the monopolist engages in Cournot competition against private firms in the downstream market.

Production of the downstream service is of the fixed-coefficient type. Each unit of the downstream output requires one unit of the upstream service and a complementary input that may be self-supplied by an independent rival. The marginal cost of the upstream service is c and the regulated input price (or access charge) of the downstream market is assumed to be fixed as r where r [greater than or equal to] c. (1)

The per unit cost of the complementary input is [s.sup.i], where [s.sup.i] [greater than or equal to] 0 and i = m, n denote the monopolist and the representative independent downstream rival, respectively. In addition, for the traceable and interior solutions throughout the analysis where both the monopolist and rivals produce the nonnegative outputs in the coexistence equilibrium, the author needs the following assumptions about [DELTA] [equivalent to] [s.sup.m] - [s.sup.n]:

A1. [DELTA] [greater than or equal to] 0.

A2. [DELTA] [greater than or equal to] r - c.

A3. [DELTA] [less than or equal to] (r - c) + (a - r - [s.sup.n])/k.

A1 indicates that there is an efficiency gap between the monopolist and independent rivals, and the independent firms are more cost efficient than the monopolist. Otherwise, the privatization policy always decreases the welfare in the model. (2) A2 represents that the efficiency gap is greater than the price-cost margin of the upstream monopolist, which is nonnegative. It ensures that the monopolist is not able to exclude the independent rivals from the downstream market. Thus, the rivals will produce nonnegative outputs in the model. Finally, A3 implies that the efficiency gap is large enough that the inefficient monopolist can produce the output in the downstream market. Otherwise, the monopolist will not provide the downstream service in the model.

Then, the profit functions for the monopolist and the independent rival are given by, respectively,

(1) [[pi].sup.m] = (r - c)[k-1.summation over (j=1)][q.sub.j.sup.n] + (P(Q) - c - [s.sup.m])[q.sup.m],

(2) [[pi].sub.j.sup.n] = (P(Q) - r - [s.sup.n])[q.sub.j.sup.n].

In the profit function of the monopolist, the first term comes from the access profit by selling access to the independent rivals, and the second term comes from the operating profits by selling downstream services in a downstream market.

The consumer surplus in the downstream market is then

(3) CS = [[integral].sub.0.sup.Q]P(v)dv - P(Q)Q,

and social welfare, which is defined as the simple sum of consumers surplus and firms' profits, is given by

(4) W = [[pi].sup.m] + [k-1.summation over (j=1)][[pi].sub.j.sup.n] + CS,

(5) = [[integral].sub.0.sup.Q]P(v)dv - (c + [s.sup.m])[q.sup.m] - (c + [s.sup.n])[k-1.summation over (j=1)][q.sub.j.sup.n]

The author assumes that the government maximizes social welfare while the private firms focus on profits in the following analysis. This implies that the monopolist will maximize social welfare in (4) when it takes the form of public enterprise preprivatization, whereas it will maximize its profit in (1) postprivatization. (3)

Now, the author will examine the equilibrium outcomes by the public enterprise preprivatization and the outcomes by the privatized monopolist postprivatization, respectively, and then compare the two equilibrium outcomes. First, the author considers the preprivatization case where the upstream monopolist takes the form of a public enterprise and produces its output to maximize the goal of the government, the social welfare function in (5). Then, the first-order condition for the public enterprise with interior solution yields

(6) P(Q) = c + [s.sup.m].

Notice that at the equilibrium the market price is exactly equal to the marginal cost of the monopolist, c + [s.sup.m], which is assumed to be larger than the marginal cost of the independent rivals, r + [s.sup.n]. This is so because the public enterprise is only able to control its output level, not that of the rival when it maximizes social welfare in the Cournot competition situation. (4) Notice also that the operating profit level of the public enterprise is always nonnegative at equilibrium if r [greater than or equal to] c. (5)

The first-order condition for the independent rival's profit in (2) with the interior solution is as follows:

(7) P(Q) = r + [s.sup.n] + b[q.sub.j.sup.n].

Then, from the profit function in (2), the author knows that the profit level of the rival firm in a downstream market is always positive at equilibrium.

Using the equilibrium output of the representative rivals and combining the equations in (6) and (7), one obtains the following Cournot-Nash equilibrium output levels for the public enterprise and private rival:

(8) [q.sub.N.sup.m] = (X - kY)/b,

(9) [q.sub.N.sup.n] = Y/b,

where X = a - r - [s.sup.n] [greater than or equal to] kY and Y = [DELTA] - (r - c) [greater than or equal to] 0 as assumed. Then, [q.sub.N.sup.m] [greater than or equal to] 0, [q.sub.N.sup.n] [greater than or equal to] 0, and [q.sub.N.sup.m] [greater than or equal to] [q.sub.N.sup.n] if X - kY [greater than or equal to] Y, [q.sub.N.sup.m] [less than or equal to] [q.sub.N.sup.n] if X - kY [less than or equal to] Y. It means that an inefficient public enterprise might produce greater output than the efficient independent rival firms preprivatization. This is because the public firm extends the output to where the market price is equal to its marginal cost even though it is inefficient. In sum, in preprivatization, the total output is [Q.sub.N] = (X - Y)/b and the equilibrium price is [P.sub.N] = C + [s.sup.m].

Next, the author considers the postprivatization case with the upstream monopolist and thus, the privatized monopolist maximizes its own profit function in (1). Then, the first-order condition for the privatized monopolist with interior solution yields

(10) P(Q) = c + [s.sup.m] + b[q.sup.m].

Then, from the profit functions in (1) and (2), the author knows that all firms in a downstream market have positive profits at equilibrium.

From the first-order condition of the individual rival in (7), one gets the following Cournot-Nash equilibrium output levels:

(11) [q.sub.p.sup.m](X - kY)/b(k + 1),

(12) [q.sub.p.sup.n](X + Y)/b(k + 1).

Then, at the postprivatization equilibrium the efficient independent firm produces greater output than the inefficient privatized monopolist: [q.sub.p.sup.n] [greater than or equal to] [q.sub.p.sup.m]. In sum, under the privatized monopolist situation, the total output is [Q.sub.P] = (kX - Y)/b(k + 1) and the equilibrium price is [P.sub.p] = a - (kX - Y)/(k + 1).

Now, the author compares the two equilibrium outcomes pre- and postprivatization. Notice that the output of the inefficient monopolist decreases postprivatization, whereas output of the efficient independent rival increases post-privatization: [q.sub.N.sup.m] > [q.sub.p.sup.m] and [q.sub.N.sup.n] < [q.sub.p.sup.n]. This yields the following two propositions. (6)

PROPOSITION 1. Suppose that the assumptions in Al, A2, and A3 hold. Then, all firms produce nonnegative output and earn nonnegative profits in the Cournot-Nash equilibrium in which we have:

(i) [Q.sub.N] > [Q.sub.P] and [P.sub.N] < [P.sub.p],

(ii) [[pi].sub.N.sup.m] < [[pi].sub.p.sup.m] and [[pi].sub.N.sup.n] < [[pi].sub.p.sup.n].

Proposition 1 indicates that (i) the consumer surplus is decreasing but (ii) the profits of both individual firms and total industry profit ([[pi].sup.m] + [k - 1][[pi].sup.n]) are increasing postprivatization. It means that it is beneficial for the firms, although it might be harmful for the public to implement a privatization policy. Hence, privatization might be harmful to social welfare when the privatized monopolist reduces its output level too much. Put differently, the inefficient monopolist might be beneficial to social welfare when the inefficient monopolist produces greater output that could not be produced by more efficient firms. As a result, only when the welfare-increasing effects of cost reduction from the private firms is greater than the welfare-decreasing effects of the reduction in total output from public enterprise, social welfare will increase postprivatization. Therefore, welfare will be increasing in psotprivatization if the superiority of the cost efficiency of the independent rivals outweighs the negative effects of the decrease in total output. Proposition 2 indicates the condition where social welfare is increasing postprivatization.

PROPOSITION 2. Define R [equivalent to] (X - kY)/2([k.sup.2] - 1) [greater than or equal to] 0. Then, one has [W.sub.N] [less than or equal to] [W.sub.P] if R - [DELTA] [less than or equal to] 0 or [W.sub.N] [greater than or equal to] [W.sub.P] if R - [DELTA] [greater than or equal to] 0.

Proposition 2 shows that the welfare change in postprivatization depends on the relative size of the cost efficiency gap between the monopolist and the rivals. A few remarks are in order.

First, one can see that R = 0 when [DELTA] = (r - c) + X/k and R is decreasing in [DELTA]. It implies that there exists a unique threshold level of the efficiency gap such that privatization improves the welfare, that is, [W.sub.N] [less than or equal to] [W.sub.P]. Specifically, R [less than or equal to] [DELTA] if [DELTA] [greater than or equal to] (X + k[r - c])/(sk[.sup.2] + k - 2). Thus, privatization is beneficial to society only when the efficiency gap between the monopolist and the rivals is large. In sum, the cost efficiency superiority of the independent rivals will be a necessary condition to improve welfare postprivatization.

Second, R is decreasing in k while R = (X - 2Y)/6 [greater than or equal to] 0 when k = 2. It implies that the welfare in postprivatization tends to be increasing as the number of downstream firms is increasing. Therefore, privatization will be beneficial to society only if the competition in the downstream market is sufficient. (7)

Finally, R is increasing in r. It implies that welfare postprivatization will be decreasing if the regulated access charge (the input price of a upstream service) is high. In particular, privatization directly benefits society only if r [less than or equal to] c + ([DELTA][2k[.sup.2] + k - 2] - X)/k. Thus, the necessary condition to increase social welfare is [DELTA] [greater than or equal to] X/(2[k.sup.2] + k - 2). Otherwise, privatization is always harmful to social welfare. Therefore, it is important for the regulatory agency to maintain the regulated access price at a certain level even post-privatization.

III. EXTENSIONS AND DISCUSSIONS

In the telecommunications industry, it is common for an upstream monopolist to supply an essential input to private firms operating in the downstream market. Furthermore, usually in developing areas, such as Asian, European, and Latin American countries, the upstream monopolist has taken a form of a public enterprise. Therefore, the downstream market has consisted of a mixed market where the public enterprise competes with independent private firms.

Recently, the government activated both privatization and competition policies in mixed markets. Now, it is a real concern of policy makers whether the privatization of the upstream public firm, which will induce the upstream monopolist to engage in profit-maximizing strategies, yields welfare-increasing outcomes in the market. In the simple model presented, the welfare implications of privatization depend primarily on the relative cost efficiency between upstream monopolist and downstream firms. However, there are many other important policy aspects that have been abstracted for reasons of tractability and simplicity. The article will extend the analysis into several theoretical issues and discuss some policy-relevant implications.

A. Vertical Separation and Competition Strategies

The author has confined the analysis into the vertical integration model where the upstream monopolist vertically integrates the downstream firm and competes against the other independent rival firms. However, as shown in Proposition 2, in certain conditions of the cost-efficiency gap, the public enterprise could not bring about greater social welfare compared to the case where it is a privatized firm. This is because the public enterprise in a downstream market produces greater output than that when it is privatized, that is, [q.sub.N.sup.m] > [q.sub.p.sup.m] even though it is inefficient compared to the rival firms.

From a different policy perspective, if the public enterprise can be separated into an independent upstream firm and independent downstream firm, and the upstream firm can be managed in a form of public enterprise while the downstream firm is privatized, the market outcomes will be the same as those of complete privatization of the public enterprise. In this case, without privatizing the public enterprise in both upstream and downstream markets together, the government can increase social welfare in the telecommunications industry.

However, it might be technologically or politically inefficient to separate the vertically integrated monopolist into two markets. (8) For example, if high technology can be applied to the connection (bundling) between the upstream and downstream services (through coutilizing facilities and human resources), then it will be optimal to integrate two markets vertically because technological economies and dynamic investments (e.g., stable supply of telecommunications services) could be realized. Politically, in addition, it might lead to huge social costs to separate the historical public utility into different firms if the government cannot make an agreement among employees and public citizens. Therefore, before separating into independent upstream and downstream firms, it is necessary both to check if there are technological or political linkages between two markets and to consider how the regulatory agency is able to treat these problems without high social costs.

For instance, the Korean government announced plans to privatize its electricity power utility (Korea Electric Power Corporation), which was a government-invested monopolist that supplied electric power in Korea. During 2001, the Korean electric power industry underwent major changes as its power generation unit was separated into six subsidiaries and the Korea Power Exchange was inaugurated. In addition, the power generation subsidiaries are supposed to be privatized, and it is preparing separate power distribution units. But there still remain many debates between the government and employees (or even the public). Therefore, it is expected that there are ongoing social costs in both the privatization and separation processes.

On the other hand, there is the other policy debate on privatization and separation in the telecommunications industry. One might argue that without separation between upstream and downstream firms, the welfare could be increased if the public enterprise acts as a Stackelberg leader. (9)

As an intermediate step toward dynamics, the public enterprise can make a strategic situation where it restricts its choice to gain strategic advantages. For example, if the public enterprise can act as a leader in a mixed market, it will be able to set [q.sup.m] = [q.sub.P.sup.m] to induce the rival firms to produce [q.sub.P.sup.n], which would give the same market outcomes as privatization of the monopolist.

However, the question remains whether the inefficient public enterprise can be a leader in the telecommunications market. If the public enterprise can commit its output level to the rival firms in a credible way and the rival firms take its commitment into consideration effectively, then the public enterprise can obtain the same social welfare level with that of postprivatization. But if the public enterprise lacks leadership and cannot commit its output level in advance, then the strategy of choosing [q.sup.m] is not a best response of the monopolist when the rival produces [q.sub.P.sup.n]. As has been shown, [q.sub.P.sup.m] is not the equilibrium output for the monopolist under a Cournot-Nash type competition, where the monopolist is in interactive strategic competition with rivals. Therefore, it is important for a public enterprise to find an effective way to hold and sustain the role of leadership in the mixed market structure. If Stackelberg leadership is not available to the public enterprise, then policy makers should consider privatization of the public enterprise as an alternative.

Finally, one might also consider the other competition pattern in the telecommunications market where downstream firms set fees competitively. (10) For example if the downstream firms compete in the form of Bertrand price competition, the market price will be determined at P(Q) = r + [s.sup.n] in (7) and thus, [q.sup.m] = 0 at equilibrium, irrespective of nationalization or privatization. It implies that the welfare level does not depend on the privatization policy as long as there exists price competition among the firms in a downstream market.

In sum, the welfare consequences of a privatization policy depend not only on the competition patterns of between upstream and downstream firms, that is, Cournot versus Stackelberg, but also on the competition patterns within downstream firms, that is, Cournot versus Bertrand.

B. Objectives of the Public Enterprise

In the basic model, it is worthwhile to reconsider the objective function of the public enterprise, which was defined as the total social welfare in (5). This comes from the assumption that the public enterprise maximizes social welfare, the objective of the government. In the ideally hypothesized environment where the government has complete information and absolute authority, the public enterprise will maximize social welfare. However, due to incomplete information or costly monitoring, the objective of public enterprise will differ from that of government. The theoretical treatment of the ownership effects deduced from the property rights and principal-agent perspectives should be considered to examine the efficacy of the incentive system that is designed to maximize the efforts of the agents.

One of the policy considerations to capture the interest of the government is that the public enterprise will be guaranteed to support its nonnegative operating profit. For example, if r < c, the operating profits level of public enterprise will be negative at equilibrium, that is, [[pi].sup.m] = (r - c)(k - 1)Y/b in (1). In this case, the government needs to subsidize a lump-sum subsidy to support the public enterprise in the market equilibrium. If not, the enterprise may not survive. To meet the nonnegative operating profit conditions, for instance, the government might transfer the public funds when the public enterprise takes loses in the business. Then, the objective of the government will be CS + (k - 1)[[pi].sup.n] + [[pi].sup.m] + (1 - [mu])T, where T is the lump-sum transfer to the firm and thus, [mu]T captures the cost of the extra distortions created elsewhere in the economy.

In the analysis, however, if the public enterprise maximizes social welfare, the equilibrium outcomes in this market are the same in Proposition 1, as long as the transfer has the form of a lump sum. However, the amount of transfer might be increasing postprivatization since the rivals' output level, for which the monopolist should provide the access service with a negative price margin, will be increasing, that is, [q.sub.N.sup.n] < [q.sub.P.sup.n]. Therefore, the welfare will likely be decreasing postprivatization.

In the past, in fact, governments in developing areas considered telecommunication services as basic public infrastructure, and thus tried to encourage the usage of telecommunications services by both setting the access price below its cost and subsidizing the operating deficits of the public enterprise. However, it became a policy concern that subsidization yielded not only social welfare loss but also social costs of public funds. In recent years, therefore, the government has changed its policy to set the access price at the exact cost level. (12)

The other regulatory possibility of the government is that the public enterprise might be regulated to maximize the consumer surplus only under the constraint of its nonnegative operating profit. (13) Then the objective of the public enterprise is to maximize a part of social welfare instead of total social welfare, where, without considering the rival's profit in (2), it maximizes the consumer surplus in (3) with its own profit constraint in (1). Specifically, the objective function of the public enterprise will be CS + [lambda][pi][.sup.m] where [lambda] [greater than or equal to] 0. Note that it is simply the sum of consumer surplus and monopolist's own profits when [lambda] = 1. Then, assuming the interior solution again, the first-order condition is as follows:

P(Q) = (c + [s.sup.m]) - b(Q - [lambda][q.sup.m])/[lambda].

Using the first-order condition for the representative independent rival in (7), one can get the following equation in the Cournot-Nash equilibrium:

[q.sup.n] = Y/b - (Q - [lambda][q.sup.m])/[~.[lambda]]

Thus, if 1 [less than or equal to] [lambda] [less than or equal to] Q/[q.sup.m], (14) the market price and the rival's output level will be lower than the outcomes in (6) and (9), where the public enterprise maximizes total social welfare. Consumer surplus and the inefficient public enterprise's output will be certainly increasing compared to (6) and (8), respectively. Therefore, social welfare will be certainly decreasing compared to the case where the public enterprise maximizes total social welfare instead of maximizing a part of social welfare. It implies that in the case that the public enterprise maximizes some parts of social welfare there is a greater possibility to increase social welfare of postprivatization.

On the other hand, based on the principal-agent theory, the author can also incorporate the existence of information asymmetry and the absence of competition mechanism in the public sector to explain the nontotal social welfare-maximizing strategies of public enterprise. That is, the realistic alternative for the objective of the public enterprise could be to consider private incentives in the public ownership environment. For example, it is possible for the manager of the public enterprise to incur self-interested and inefficient expenditures, such as waste in the form of goldplating, (accounting and managerial) cross-subsidies to the other businesses, excessive employee compensation, and so on. Such "wasteful" expenditures can arise from political reasons, lack of manager's incentives to economize, or principal-agent problems in general. (15) These wasteful expenditures return fringe benefits to the decision maker of the public enterprise.

It is also noteworthy that the most important policy aspect of privatization for the government is to induce the public enterprise to achieve cost efficiency by reducing its managerial inefficiency. (16) In other words, the policy makers tend to believe that an inefficient production cost level of the privatized public enterprise would be decreasing and finally, will be equal to the cost level of the independent private firm in postprivatization.

The author will assess this policy view by incorporating the managerial inefficiency of the public enterprise into the model. For simplicity, let [s.sup.m] = [s.sup.n] + e, where e(0 [less than or equal to] [e.bar] [less than or equal to] e [less than or equal to] [bar.e]) denotes the managerial inefficiency term in the objective function of the firm, such as wasteful expenditure, which is independent of the act of production. The author will assume that this expenditure will cause fringe benefit per output, g(e), where g' > 0, and cost (management compensation from the principal) per output, h(e), where h' < 0, for the manager of the firm and thus for the society as well. Denoting d(e) [equivalent to] g(e) - h(e), where d' > 0, the author then has the following profit function of the monopolist and social welfare:

[[pi].sup.m] = (r - c)[k-1.summation over (j=1)][q.sub.j.sup.n] + (P(Q) - c - [s.sup.n] - e)[q.sup.m],

W = ([[integral].sub.0.sup.Q]P(v)dv + d(e)[q.sup.m]) - (c + [s.sup.n] + e)[q.sup.m] - (c + [s.sup.n])[k-1.summation over (j=1)][q.sub.j.sup.n]

Thus, the manager of the public enterprise will maximize W + d(e)[q.sup.m], the sum of social welfare, the goal of the government, and its total fringe benefits, whereas the manager of the privatized firm will maximize [[pi].sup.m] + d(e)[q.sup.m], the sum of its profits and its total fringe benefits, respectively. Then, it is easily shown that in certain conditions for d', specifically 1/2 < d' < 1, the monopolist will set e = [e.bar] when it acts as a privatized firm in postprivatization, and e = [bar.e] when it acts as a public enterprise in preprivatization.

Then, one obtains the following two propositions.

PROPOSITION 3. Suppose that [s.sub.N.sup.m] > [s.sub.P.sup.m] = [s.sup.n] and [DELTA]' = [s.sub.N.sup.m] - [s.sup.n]. Then, at the Cournot-Nash equilibrium one has:

(i) If X - kY > [DELTA]', [q.sub.N.sup.m] > [q.sub.P.sup.m], [q.sub.N.sup.n] > [q.sub.P.sup.n], [Q.sub.N] > [Q.sub.P], and [P.sub.N] < [P.sub.p].

(ii) If X - kY < [DELTA]', [q.sub.N.sup.m] < [q.sub.P.sup.m], [q.sub.N.sup.n] < [q.sub.P.sup.n], [Q.sub.N] < [Q.sub.P], and [P.sub.N] > [P.sub.p].

Proposition 3(i) indicates that privatization might be harmful to social welfare when the privatized monopolist reduces its output level too much, which is the same policy implication in Proposition 1. But, Proposition 3(ii) shows that social welfare will be always increasing under post-privatization since the production of the inefficient firm is displaced with the production of more efficient firms and thus, the total output provided is increased.

PROPOSITION 4. Define R' [equivalent to] (X - kY - [DELTA]')[.sup.2]/2(kX - [k + 1]Y + [r - c]). Then, one has [W.sub.N] [less than or equal to] [W.sub.P] if R' - [DELTA]' [less than or equal to] 0 or [W.sub.N] [greater than or equal to] [W.sub.P] if R' - [DELTA]' [greater than or equal to] 0.

Proposition 4 shows that the welfare change postprivatization depends on the superiority of the cost efficiency between pre- and postprivatization. If X - kY < [DELTA]' in preprivatization, then R' < 0 and thus [W.sub.N] < [W.sub.P]. This result comes from the reduction of the wasteful expenditures postprivatization. However, if X - kY > [DELTA]' preprivatization, then the welfare effects of privatization in Proposition 2 applies. That is, privatization is beneficial to society when (1) the efficiency gap between the public enterprise and the rivals is large, (2) the competition in the downstream market is enough, and (3) the access charge is regulated not too high.

C. Partial Privatization

During privatization, the government may be able to manage the activities of the privatized firm by controlling its portion of shares. In other words, there is a possibility that full privatization, where the government sells all its shares in a public enterprise, is not fulfilled at once. Thus, the government might determine the degree of privatization instead of the extreme full privatization.

In the public economics literature, De Fraja and Delbono (1989) and De Fraja (1991) examined the efficiency of full privatization in an oligopoly market. Extending their works, Matsumura (1998) and Lee and Hwang (2003) considered the possibility of partial privatization and showed that it is optimal for the government to sell part (but not all) of its shares in public enterprises in the context of mixed duopoly.

To examine the possibility of partial privatization in the model, let the privatization parameter be [mu], which takes a value in a unit interval [0,1], and a higher value of [mu] indicates a higher degree of privatization. In particular, [mu] = 1 represents full privatization, and [mu] = 0 indicates no privatization. Then, for any given degree of privatization, the partially privatized monopolist will maximize the weighted average of the payoff to the government W and its own profit [[pi].sup.m].

Thus, one can specify the objective of the partially privatized firm as follows: (17)

(13) [U.sup.m] = (1 - [theta])W + [theta][[pi].sup.m].

Note that the fully privatized monopolist ([mu] = 1) will maximize its usual profit in (1), whereas the fully public enterprise ([mu] = 0) will maximize social welfare in (5), the sum of consumer surplus and industry profits.

The first-order condition for [U.sup.m] yields

(1 - [theta])[[[partial derivative]W]/[[partial derivative][q.sup.m]]] + [theta][[[partial derivative][[pi].sup.m]]/[[partial derivative][q.sup.m]]] = 0,

or, equivalently,

p(Q) = c + [s.sup.m] + [theta]b[q.sup.m].

Using the first-order condition of the rivals in (7), the Cournot-Nash equilibrium output levels are given by

(14) [^.q.sup.m] = (X - kY)/b(k[theta] + 1),

(15) [^.q.sup.n] = ([theta]X - Y)/b(k[theta] + 1).

The effects of the government's decision parameter [mu] on the equilibrium level of output can be easily shown as follows: [[partial derivative][^.q.sup.m]/[partial derivative][theta] < 0, [partial derivative][^.q.sup.n]/[partial derivative][theta] > 0, and [partial derivative][^.Q]/[partial derivative][theta] = [partial derivative][^.q.sup.m]/[partial derivative][theta] + (k - 1)[partial derivative][^.q.sup.n][partial derivative][theta] < 0. Thus, as the degree of privatization increases, the public enterprise's output decreases but the rival's output increases and totally, the equilibrium market output decreases monotonically.

To investigate the optimal degree of privatization of a public enterprise, the author needs to check the following condition for the optimum degree of privatization that maximizes the government's payoff, W:

(16) [dW]/[d[theta]] = [[[partial derivative][q.sup.m]]/[[partial derivative][theta]]][[[partial derivative]W]/[[partial derivative][q.sup.m]]] + [[[partial derivative][q.sup.n]]/[[partial derivative][theta]]][[[partial derivative]W]/[[partial derivative][q.sup.n]]],

Substitution of equilibrium conditions in (14) and (15) into this equation in (16) yields

[dW]/[d[theta]] = [[X - kY]/[b(k[theta] + 1)[.sup.2]]] x ([DELTA] - [theta](k - 1)(X - kY)/(k[theta] + 1))

It is noteworthy that dW/d[mu] takes a positive sign at [mu] = 0. Therefore, the no privatization of the public enterprise is not an optimum in this model. The author thus obtains the necessary conditions for the existence of a positive interior solution for [mu], which is that dW/d[mu] takes a negative sign at [theta] = 1, that is, [DELTA] < (k - 1)(X - kY)/(k + 1) or [DELTA] < (k - 1) (X + k[r - c])/(k + 1). It implies that the full privatization is socially beneficial only when the cost efficiency gap is sufficiently high, [DELTA] [greater than or equal to] (k - 1)(K + k [r - c])/(k + 1). Otherwise, the partial privatization is socially beneficial in the model.

D. Discrimination and Access Charge

When policy makers consider privatization policy, there is a concern about market foreclosure in which the profit-maximizing, privatized integrated firm may have an incentive to discriminate the access from the rival firms to the upstream market and thus earn a strategic gain by increasing the costs of downstream competitors.

In the regulatory economics literature of the telecommunications industry, there have been discussions on discriminating access charges from the private monopolist. Economides (1998), for example, claimed that the incentive to discriminate is independent of the relative efficiency of the monopolist and the independent rivals. But, Weisman and Kang (2001) showed that discrimination always arises in equilibrium when the monopolist is no less efficient than its rivals in the downstream market. Sibley and Weisman (1998) analyzed the interrelationship between the market size and the incentive of discrimination in the telecommunications industry. Empirically, using the data of the U.S. telecommunications market, Mandy (2000) tested the parameters that affect the monopolist's incentives to discriminate and concluded that there is an incentive to discriminate in the U.S. telecommunications market.

The author will examine the possibility of discriminating access charge under postprivatization and its welfare effects in the model.

First, from the market equilibrium outcomes in (8) and (9) of pre-privatization, and (10) and (11) of postprivatization, one has [partial derivative][p.sub.N.sup.m]/[partial derivative]r > [partial derivative][q.sub.P.sup.m]/[partial derivative]r > 0 and [partial derivative][p.sub.N.sup.n]/[partial derivative]r < [partial derivative][q.sub.P.sup.n]/[partial derivative]r < 0. It means that the higher access charge is preferable to the monopolist, whereas the lower access charge is beneficial to the rival firms. In addition, the output effects of access charge in the case of preprivatization is stronger than that of postprivatization. In this sense, it will be the main policy concern for the regulatory agency who seeks to make fair competition in the telecommunications industry, to set the access charge not too high in either pre- and postprivatization.

This concern can be also applied to social welfare. From the welfare function in (5), one obtains the following relationship at each equilibrium:

[partial derivative]W([q.sub.P.sup.m], [q.sub.P.sup.n])/[partial derivative]r < [partial derivative]W([q.sub.N.sup.m], [q.sub.N.sup.n])/[partial derivative]r < 0.

It indicates that the welfare effects of lowering access charge in postprivatization is stronger than that of pre-privatization. Thus, it will be socially beneficial to set the access charge as low as possible. For instance, under the assumption A3 in section II, the author has the following optimal access charge rate for the society:

r* = c - (a - k[DELTA] - c - [s.sup.n])/(k - 1).

Notice that r* = c when k goes to the infinity. Notice also that r* < c when (a - k[DELTA] - c - [s.sup.n])/(k - 1) > 0 or [DELTA] < (a - c - [s.sup.n])/k, that is, the cost efficiency gap is sufficiently small. (18)

On the other hand, if the monopolist can set the access price r by its own will postprivatization, it may strategically engage in discriminatory behavior in a downstream market by setting the higher r. Then, it might reduce the welfare-increasing effects of privatization because the efficient rivals would produce less output. However, as long as the cost efficiency of the rival firms is superior to the monopolist, the monopolist can earn higher profits from access service if the rivals increase their outputs postprivatization. Therefore, there is a trade-off for the monopolist between profits arisen from the increment of its own output when rivals decrease their outputs and profit arisen from the access of the rivals when they increase their outputs.

To investigate this, the author can find the following profit-maximizing level of access charge from the profit function of the monopolist in (1):

[r.sub.P] = c + ([k + 3]X - [k - 1]Y)/2(k + 1),

or

[r.sub.P] = (a + c - [s.sup.n])/2 - (k - 1)[DELTA]/2(k + 3).

A few remarks are in order. First, [r.sub.P] > c. It implies that the privatized monopolist will set a higher access charge then the possible access charge rate under preprivatization. This fact might call for the intervention of the regulatory agency to control the access charge rate in post-privatization.

Second, [partial derivative][r.sub.P]/[partial derivative]k < 0. It means that the monopolist's profit-maximizing access charge will be reduced when the government introduces more competitive rivals in a downstream market. Therefore, an open access policy of the upstream essential input is important in the telecommunications industry.

Finally, [partial derivative][r.sub.P]/[partial derivative][DELTA] < 0. If the efficiency superiority of the independent rivals prevails sufficiently, then it will be still socially beneficial for society to privatize the public enterprise in certain conditions even though the monopolist is able to discriminate the access charge postprivatization. Compared to the result of Proposition 2, however, welfare will be reduced under postprivatization. (19)

On balance, the regulatory agency needs to construct an appropriate access charge regulation for the privatized monopolist and implement open access policy for the rivals in a downstream market. (20)

IV. CONCLUSION

This article has considered a simple vertical network structure of the telecommunications industry where a public enterprise supplies the essential network service to downstream private firms while the public enterprise competes with the independent firms in a mixed downstream market. Then, the author examined the equilibrium outcomes between pre- and postprivatization and compared the welfare effects of privatization of the public enterprise.

The analysis has shown that the cost advantage of the independent rivals in a downstream market improves the welfare if the upstream public enterprise is privatized. It is also shown that competition policy or open access regulation including access charge regulation are required to improve social welfare postprivatization. Therefore, the recent trend of both privatization and competition policy in the telecommunications industry is well justified. The author has also discussed several policy debates on the process of privatization in the telecommunications industry. Current policy issues, including vertical separation between upstream firm and downstream firms, the role of leadership of the public enterprise, the objectives of the monopolist and its managerial incentives in agency relationship, the possibility of partial privatization, and the strategic choice of access charge of the privatized monopolist, have been examined with abstracted forms.

For future policy consideration, it is worthwhile to investigate the financial treatment in the process of privatization and/or the fundraising effects of privatization for the government. These policy-relevant concerns should be carefully investigated before the government implements the privatization process of the public enterprise in the telecommunications industry.

REFERENCES

Anderson, S. P., and M. Engers "Stackelberg versus Cournot Oligopoly Equilibrium." International Journal of Industrial Organization, 10, 1992, 127-35.

Baumol, W. J., and J. G. Sidak. Toward Competition in Local Telephony. Cambridge, Mass.: MIT Press, 1994.

Cook, P., and R. V. Fabella. "The Welfare and Political Economy Dimensions of Private versus State Enterprise." Manchester School, 70, 2002, 246-61.

De Fraja, G. "Efficiency and Privatization in Imperfectly Competitive Industries." Journal of Industrial Economics, 39, 1991,311-21.

De Fraja, G., and F. Delbono. "Alternative Strategies of a Public Enterprise in Oligopoly." Oxford Economic Papers, 41, 1989,302-11.

______. "Game Theoretic Models of Mixed Oligopoly." Journal of Economic Surveys, 4, 1990, 1-17.

Dowrick, S. "Von Stackelberg and Cournot Duopoly: Choosing Roles." RAND Journal of Economics, 17, 1986, 251-60.

Economides, N. "The Incentive for Non-Price Discrimination by an Input Monopolist." International Journal of Industrial Organization, 16, 1998, 271-84.

Hackner, J. "Vertical Integration and Competition Policy." Journal of Regulatory Economics, 24, 2003, 213-22.

Haskel, J., and A. Sanchis. "Privatization and X-Inefficiency: A Bargaining Approach." Journal of Industrial Economics, 43, 1995, 301-21.

Laffont, J.-J., and J. Tirole. A Theory of Incentives in Procurement and Regulation. Cambridge, Mass.: MIT Press, 1993.

______. Competition in Telecommunications. Cambridge, Mass.: MIT Press, 2000.

Lee, S. H. and H. S. Hwang. "Partial Ownership for the Public Firm and Competition." Japanese Economic Review, 54, 2003, 324-35.

Mandy, D. M. "Killing the Goose That May Have Laid the Golden Eggs: Only the Data Know Whether Sabotage Pays." Journal of Regulatory Economics, 17, 2000, 157-72.

Matsumura, T. "Partial Privatization in Mixed Duopoly." Journal of Public Economics, 70, 1998, 473-83.

Megginson, W. L., and J. M. Netter. "From State to Market: A Survey of Empirical Studies on Privatization." Journal of Economic Literature, 39, 2001, 321-89.

Panzar, J. C, and D. S. Sibley. "Optimal Two-part Tariffs for Inputs: The case of Imperfect Competition." Journal of Public Economics, 40, 1989, 237-49.

Sappington, D. E. M., and D. S. Sibley. "Regulatory Incentive Policies and Abuse." Journal of Regulatory Economics, 5, 1993, 131-41.

Sibley, D. S., and D. L. Weisman. "The Competitive Incentives of Vertically Integrated Local Exchange Carriers: An Economic and Policy Analysis." Journal of Policy Analysis and Management, 17, 1998, 74-93.

Vickers, J., and G. Yarrow. Privatization: An Economic Analysis. Cambridge, Mass.: MIT Press, 1988.

______. "Economic Perspectives on Privatization." Journal of Economic Perspectives, 5, 1991, 111-32.

Viscusi, K., J. Vernon, and J. Harrington. Economics of Regulation and Antitrust. Cambridge, Mass.: MIT Press, 1995.

Vives, X. Oligopoly Pricing: Old Ideas and New Tools. The MIT Press, 1999.

Weisman, D. L., and J. Kang, "Incentives for Discrimination when Upstream Monopolists Participates in Downstream Markets." Journal of Regulatory Economics, 20, 2001, 125-39.

SANG-HO LEE*

*This is a revision of a paper presented at the KASIO and KIEA fall conference (2003, Seoul), Chonnam National University (2003, Gwangju, Korea), Fudan University (2003, Shanghai), and Western Economic Association International 79th annual conference (2004, Vancouver). The author is grateful for the comments and suggestions made by Oz Shy, Yongyeop Sohn, Jongguk Park, Zhao Chen, Zonglai Kou, Margriet F. Caswell, Kristen Monaco, and two anonymous referees. This study was financially supported by Chonnam National University in the 2003 program.

Lee: Associate Professor of Economics, Chonnam National University, 300 Yongbong-dong, Bukgu, Gwangju, South Korea. Phone 82-62-530-1553, Fax 82-62-530-1559, E-mail sangho@chonnam.ac.kr

1. In the telecommunications industry, the input price of the downstream market is regulated by the government or the regulatory agency such as the Federal Trade Commission in United States. The rationale of access charge regulation can be found in Laffont and Tirole (2000). In the following analysis, the author assumes that access charge r is fixed both pre- and postprivatization. Section III will discuss on the welfare effects of changing access charge in postprivatization.

2. To focus on the welfare effects arisen from the cost difference, the author assumes that [s.sub.i] is independent of the ownership structure, that is, [s.sup.i] is fixed both pre- and post-privatization. Section III will incorporate the possibility of cost reduction of postprivatization where [s.sup.m] = [s.sup.n].

3. There might be a discrepancy between the goal of the government and that of the public enterprise. See Laffont and Tirole (1993) on this point. The author will review the principal-agent relationship in section III.

4. Section III will discuss about the possibility of leadership of public enterprise and commitment problem.

5. From the profit function in (1), the author knows that the operating profit level of a public enterprise depends on the price-cost margin of the upstream market. Section III will examine the objective of public enterprise under the possibility of negative operating profits and public transfer.

6. The straight calculation yields the outcomes of propositions and thus, the proofs of all propositions are omitted.

7. In a mixed model with identical and increasing marginal costs, De Fraja and Delbono (1989, 1990) showed that if the market is sufficiently competitive, then it is socially better for the public enterprise to maximize its own profit instead of maximizing welfare.

8. There might be another theoretical issues on the increase of transaction costs between two markets or/and double marginalization effects when the public enterprise is privatized at both markets and maximizes its own profits in the upstream market. On this point, see Viscusi et al. (1995).

9. The competition patterns of Cournot or Stackelberg depend on the strategic economic environments among firms. In the economics literature, therefore, many studies have considered the order of play in a game and compared the results of Cournot and Stackelberg. On this point, see Dowrick (1986), De Fraja and Delbono (1989, 1990), Anderson and Engers (1992), and Vives (1999), among others. For example, De Fraja and Delbono (1989, 1990) show that if the public enterprise is able to act as a leader to induce Stackelberg competition in a mixed market, it is possible to increase its social goal.

10. In general, the price competition model can be applied for the case of product differentiation where the firms sell different products to consumers. Laffont and Tirole (2000), for example, provide the model of network competition in the telecommunications market. In a product-differentiated market, however, the symmetry between Cournot quantity competition and Bertrand price competition can be established with the duality argument. See, for example, Vives (1999).

11. Cook and Fabella (2002) considered the model in which the state-owned enterprise maximizes an unspecified objective function and examined the theoretical treatment of the welfare and political economy dimensions of the choice between public ownership and privatization.

12. On the theoretical treatments on the access price regulation including ECPR (Efficient Component Price Rule) or LRIC (Long-Run Incremental Cost), see Baumol and Sidak (1994), Sibley and Weisman (1998), and Laffont and Tirole (2000).

13. The regulatory incentives of the government to take care of market price or consumers surplus will yield the Ramsey solutions. On this point, see Vickers and Yarrow (1988).

14. From the second-order condition, it is sufficient to hold that [lambda] [greater than or equal to] 1.

15. The managerial inefficiency is defined as "waste" and "abuse" in the literature on the agency model of regulatory economics. On modeling of the managerial inefficiency, for examples, see Vickers and Yarrow (1988), Sappington and Sibley (1993), Laffont and Tirole (1993), and Cook and Fabella (2002).

16. Haskel and Sanchis (1995) and Lee and Hwang (2003) claim that one of the goals of the privatization policy will be to reduce the managerial X-inefticicncy of the public enterprise and to obtain the dynamic efficiency of the privatized firm.

17. For the case that r < c, the nonnegative operating profit constraint should be taken into the consideration of the government. For example, if the government subsidizes a transfer to the firm, the firm's profit will be [[pi].sup.m] + T while social welfare will be W + (1 - [mu])T, where T is the lump-sum transfer paid to the firm and thus, [mu]T captures the cost of the extra distortions created elsewhere in the economy. Therefore, the partially privatized firm will maximize [^.U.sup.m] = (1 - [theta])(W + [1 - [mu]T] + [[theta]([[pi].sup.m] + T) = (1 - [theta])W + [theta][[pi].sup.m] + (1 - [mu] + [theta][mu])T = [U.sup.m] + (1 - [mu] + [theta][mu])T. If T is the lump-sum style, then the analysis will be exactly same to the following examination. On the other hand, if T is dependent of the negative size of the operating profit, [[pi].sup.m], the government should devise an appropriate subsidy scheme. Simply, if the government subsidizes the exact amount of the negative size of the operating profit, T = -[[pi].sup.m] and thus, the subsidized profit level of the firm will be zero. In this case, one can rewrite that [^.U.sup.m] = (1 - [theta])(W + [1 - [mu]]-[[pi].sup.m]) = (1 - [theta])(2 - [mu])([1 - [beta])W + [beta][-[[pi].sup.m]]) where [beta] = (1 - [mu])/(2 - [mu]). Therefore, the author can also analyze the welfare effects of the partial privatization with the similar process in the following examination.

18. Panzar and Sibley (1989) showed that a welfare-maximizing regulator can set the marginal price of an upstream input below its marginal cost to offset downstream market power and thereby induce downstream firms to produce the efficient level of outputs.

19. Numerical example: Let a = 10, b = 1, c = 0, k = 2, and [s.sup.m] = 5. Assuming that regulated access charge r = c under pre- and postprivatization, then, from the Proposition 2, one has [W.sub.P] < [W.sub.N] when [s.sup.n] is below about 4.29 or [DELTA] > 0.71. But if the privatized monopolist can set the access charge, [r.sub.P] = 4.5 - 0.42[s.sup.n]. Then the welfare is increasing under postprivatization only when [s.sup.n] is below about 3.62 or [DELTA] > 1.38.

20. In reality, even postprivatization, the government keeps the regulatory power to control the access charge by organizing independent regulatory institutions, such as OFTEL in the United Kingdom and FTC in Korea.

Gale Copyright:

Copyright 2006 Gale, Cengage Learning. All rights
reserved.