I. Introduction

In a developing country, enacting a new competition law can be both a milestone in the process of market oriented economic reforms and a new device for rent-seeking activities. When properly designed and enforced, competition law may provide an institutional framework focused on promoting consumer welfare and productive efficiency. However, besides the usual challenges faced by competition agencies everywhere, such as capture by interest groups, asymmetric information and monopoly regulation, two additional problems are frequent in developing economies: the limited availability of accurate data on the country’s conditions of competition and the lack of experience in this type of law enforcement.

In Latin America, the magnitude of the task to be accomplished was well described by Ana Julia Jatar: "Owing to the long history of protectionism, which characterized the economic strategy of the region, promoting competition in these countries is a far more complicated task than simply implementing an antitrust law. It is a wider struggle to wrest market control from entrenched government-protected monopolists. An educational effort designed to change behavior, attitudes and culture has to be made: Consumers have to be persuaded that price liberalization is better than price controls, and producers have to be convinced that competition is the new name of the game regardless of how painful it is for them. The competition agency must also oppose government attempts to reverse the policies implemented (1993, p. 80)."

Due to this context of unfinished reforms, with few mechanisms to check the accountability of public policies, some authors have been questioning the relevance of an active competition policy (see Rodríguez and Williams, 1994, 1995; Rodríguez and Coate, 1996). According to them, priority should be given to competition advocacy that would challenge government-generated entry barriers and would enhance coherence among public policies. As Rodríguez and Coate argued: "... reformers should focus on creating government institutions to thwart rent-seeking before moving to create an active antitrust regime (1996, p. 322)."

A more compromising approach is held by Khemani and Dutz: "We examine and reject the view that the administration and enforcement of competition law itself must inevitably become a source of intervention in the market, corruption, misuse of bureaucratic power, or cause of market distortions. All of these risks can be dealt with through institutions that incorporate accountability, transparency, checks and balances, and clear rules and procedures. The design and implementation of competition law, and the mix of policy instruments and enforcement priorities must, however, reflect the institutional endowments and technical capacity of countries at different stages of economic development (1995, p. 31)."

Within the scope of the current governmental discussions about the creation of a hemispheric Free Trade Area of the Americas (FTAA), the working group on competition policy is examining these issues, with special attention to the informational needs of antitrust agencies. This paper contributes to this effort by analyzing the role of industrial statistics in the implementation of competition policy. A reliable data base describing the present conditions of competition at the national level is a key instrument not only for advocacy purposes, but also to meet the agenda pointed out by Jatar and the institutional goals mentioned by Khemani and Dutz. Section 2 introduces the economic concepts that are underlying the data base and indicates the policy messages that can be extracted from it. Section 3 presents the methodology for assembling and updating the information. Finally, a concluding remark is made in section 4.

2. Increasing returns, transaction costs and competitiveness

The data base described in section 3 provides a comprehensive view of the current conditions of competition at the national level. For each industry, conditions of competition are the result of an interplay among the following variables:

1. market size;
2. entry barriers;
3. rhythm of technical progress;
4. market power of incumbent firms; and
5. the set of market strategies that is feasible in that particular industry.

Before examining the quantitative and qualitative indicators that could grasp the evolution of the above variables, it is convenient to make a brief discussion on the use of that data base as an instrument of competition policy. From a governmental viewpoint, there are three important questions to be asked about the current conditions of competition:

1. Do they impose enough discipline on the established firms, thereby protecting the public interest, or leave open space for unfair practices?
2. Do they allow domestic producers to follow the international rhythm of technical progress?
3. Are the regulated industries meeting the international levels of productivity?

1. Internationally competitive industries that are not protected by permanent entry barriers or temporary measures of economic policy. In these industries, transient entry barriers may appear as a result of innovative efforts made by incumbent firms. The limit case of this type of industry would be named as sustainable and contestable by the contestability theory (see Baumol and others, 1982).
   
2. Internationally competitive industries that are protected by government-generated entry barriers. This class includes both those industries that are regulated for sound reasons, such as environment, public health and natural monopoly, and those receiving superfluous protection for any political reason. In the contestability jargon, they would be sustainable but incontestable.
   
3. Temporarily uncompetitive industries. This class includes those industries that could easily attain the international levels of efficiency through the removal of distortions created by domestic economic policies.
   
4. Long-lasting uncompetitive industries that could only reach international competitiveness after a deep restructuring process which would include the disappearance of some firms, mergers and joint-ventures, adoption of new technologies, etc.

The specific measures to be taken by the competition agency regarding industries of types B, C and D will not be discussed here. However, it should be noted that although the conditions of competition are ephemeral by definition, very often the competition agency faces situations that will only change after years of technological and institutional improvements framed by coherent public policies. Under these circumstances, the aforementioned typology provides a benchmark for the agency’s daily operations. It helps identify the priority areas for competition advocacy, displays average figures that are useful for checking data gathered at the firm level during the investigations conducted by the agency, and, more importantly, highlights the time dimension of the competition policy agenda.

The use of those economic indicators implies an analytical approach that is rigorous but modest, and is shared by authors such as Robert Solow (1985) and Oliver Williamson (1989). As Solow commented: "There is enough for us to do without pretending to a degree of completeness and precision which we cannot deliver. To my way of thinking, the true functions of analytical economics are best described informally: to organize our necessarily incomplete perceptions about the economy, to see connections that the untutored eye would miss, to tell plausible — sometimes even convincing — causal stories with the help of a few central principles, and to make rough quantitative judgments about the consequences of economic policy and other exogenous events (1985, p. 329)."

Yet, this is not "... a license for loose thinking. Logical rigor is just as important in this scheme of things as it is in the more self-consciously scientific one. The same goes for econometric depth and sophistication, maybe even more so (ibidem, p.329)."

Two basic concepts for analyzing the current conditions of competition are increasing returns and transaction costs. As the recent literature on endogenous growth has emphasized, every industry wherein research and development (R&D) are relevant inputs has increasing returns to scale. The reason is simple: Once the knowledge obtained through R&D has been transformed into a new production technique, the innovating firm gets an asset (i.e., a list of instructions for blending inputs) that "... can be used over and over again at no additional cost. Developing new and better instructions is equivalent to incurring a fixed cost. This property is taken to be the defining characteristic of technology (Romer, 1990, p. S72)." Since innovations are random events, the conditions of competition will necessarily be heterogeneous across industries, and monopolistic competition will be predominant among the fast growing industries.

This interplay between technical progress and competition poses an intricate challenge to the antitrust agency. As Baumol and Ordover (1992) explained: "... while monopoly is rightly recognized as an enemy of static efficiency, there are a number of reasons why it is suspected that its effects on intertemporal efficiency are not so clearly one-sided. Because both large firm size and the possession of market power can, in this view, be helpful to innovation and productivity growth, it is sometimes suggested that antitrust activity, as the enemy of market power and even of large firm size, can serve as an impediment to growth and, by enhancing its costs, as a source of intertemporal inefficiency. Furthermore, when antitrust rules create barriers to efficient interfirm cooperation in research and development and in the exploitation of the fruits of such activity, the adverse consequences from intertemporal efficiency are further exacerbated (p. 83)."

While the existence of increasing returns is sufficient to dismiss the naive approach to antitrust in which the main target to be watched is the large firm, the state of transaction costs provides important guidelines for regulatory action. Every contract between economic agents presupposes a set of activities for its implemention, such as information gathering, bargaining, monitoring and enforcing. All these activities generate transaction costs. For any firm, the relationship between production costs and transaction costs will indicate the convenient degree of vertical integration, i.e. its input/output ratio. Therefore, whenever the national levels of transaction costs were higher than the international standards, or the costs for incumbents in certain industries were lower than those for potential entrants, there will be room for competition policy. But, in most cases, due to the peculiarities of transaction costs, the only feasible action to be taken by the competition agency is the promotion of transparency, which is, in fact, the ultimate role played by a national data base on industrial organization: "Many seemingly inefficient practices of firms, and their market or contract outcomes, are revealed on closer scrutiny to be quite creditable attempts to cope with difficult transaction cost problems. Therefore one should not jump from the observation of an apparent inefficiency to the assertion of a remedy; one should examine the organization of firms and the institutions of transactions more closely to look for any mechanisms to cope with transaction costs, and ask how well they do their job (Dixit, 1996, p. 40)."

3. The national data base on industrial organization

The first, and most difficult, decision to be made when assembling a national data base on industrial organization is the selection of the aggregation level. In principle, the statistics should be highly detailed and arranged according to a standard nomenclature, such as the four-digit codes of the International Standard Industrial Classification (ISIC). This would facilitate comparative analysis and yield an accurate focus for assessing the conditions of competition across industries. However, the ISIC's four-digit code does not always match real markets. For instance, a capital goods producer may compete in a market that includes several groups of Division 38 (Metal Products, Machinery and Equipment), while the relevant market for a firm that only produces computer printers is just a fraction of Group 3825 (Manufacture of Office, Computing and Accounting Machinery). Besides, in large countries, an industry may be decentralized at the national level, but operate as regionally-segmented oligopolies.

There is no orthodox recipe for solving this problem. In practice, each country will start the data base with the aggregation level allowed by the existing national statistics. Afterwards, the structural characteristics of domestic industries and the new trends opened by technical change will naturally highlight the priority areas for improving the information system.

The second problem is that every economic index has methodological limitations. Some indicators are not available at the desired level of aggregation, or cover just some selected industries instead of the whole economy. Others are based on simplistic assumptions, or obtained through expensive research projects that are carried out only sporadically. But, despite their individual shortcomings, the set of indicators listed in box 1 can provide accurate answers for the questions raised in section 2. Two examples may illustrate how this is possible:

1. Foreign trade statistics such as import penetration, export shares and revealed comparative advantage (2) are not sufficient to tell whether a particular industry is internationally competitive. Domestic distortions, such as exchange rate appreciation and targeted incentives, may lead to artificial performances. However, any doubts about the competitiveness of that industry will disappear after a careful examination of its innovative efforts, the domestic prices, productivity indexes, and profitability rates.
   
2. The Lerner Index of market power, which is defined as the inverse (1¤e) of the price elasticity of demand, is based on the narrow assumption that incumbent firms produce a single homogeneous good (3). Nevertheless, when examined in conjunction with other indicators, such as capacity utilization, market shares and cost structures, that index gives a secure assessment of market power.

Box 1 summarizes the content of the data base on industrial organization, which is composed of three blocs. The first contains uniform statistics covering all industries, the second includes sectoral statistics which, in most countries, are available only for selected industries, and the third bloc lists the special studies that, in general, should be updated every five years.

3.1. Uniform statistics

With the exception of linkage indexes, all data included in groups A and B are annual time series classified by industry, covering the whole economy. Group A indicates the industry's size and the importance of its transactions with the rest of the economy, as defined by the Cella method for measuring forward and backward input/output linkages (see Cella, 1984). The linkage indexes also show whether the industry’s conditions of competition are influenced by the behavior of firms operating in other industries and vice versa. Thus, industries like pulp and paper, textiles and apparel, vehicles and auto-parts, steel and capital goods, etc. should be examined as members of industrial complexes that function as unified markets.

Direct estimates of linkage indexes will be available only for those countries which compile input-output tables and, in most cases, the aggregation level of these tables will be too broad for analyzing the competition process. However, since linkage indexes will be used here just as a complementary tool for ranking industries according to their macroeconomic impacts and for identifying industrial complexes, any rough approximation will do. One easy solution is to use indirect estimates based on input-output data from foreign countries. As Chenery and others argued several decades ago and the recent literature on convergence of productivity has reinforced, value added coefficients may vary across industries, but they are very similar across countries (see Chenery and Watanabe, 1958; Simpson and Tsukui, 1965; Baumol and others, 1994).

Group B reports some conventional indicators of international competitiveness: the presence of imported goods in the domestic market, the exporting performance of the domestic industry and the country's revealed comparative advantage. Every country has reliable data for calculating these indexes; the only minor difficulty is harmonizing the statistical nomenclatures on foreign trade and domestic production.

3.2. Sectoral statistics

Group C contains additional indicators of competitiveness: domestic prices, profitability and total factor productivity indexes. Although these indexes provide relevant information when presented as sectoral averages, they often need to be complemented by data that capture firms’ behavior. For instance, in industries where the dominant competition strategy is product differentiation, a raise of average prices may signify stronger competitiveness, if firms were launching new products. In the case of profitability indexes, there are at least three complementary methodologies for calculating them. The first is to estimate the profit share as a residue from the sector’s value added less the wage share. The second is to use the firms’ accounting figures, which may show profitability either as profit margin, i.e., the ratio of profits to sales revenue, or as profit rate, i.e., the rate of return on assets or equity. The third is the price-cost margin, i.e., the markup of price over long-run average cost, which can be obtained from studies on the cost structures of selected firms.

In its conventional version, the total factor productivity index (??) is defined as the ratio of an industry’s value added (v) to a weighted average of employment (l) and gross capital stock (k):

?? = v¤[al+(1-a)k], where a is the wage share. In this version, only sectoral data are necessary for measuring ??. More sophisticated versions, based on research at the firm level, may be required for a careful analysis of the technical change process. By using translog production functions, these versions distinguish intermediate inputs from the other production factors — labor and capital —

and consider gross output figures instead of value added (see Pinheiro, 1989; Bonelli, 1992).

Group D deals with capital formation, capacity utilization and employment. Investment and employment data, which were already included in group C, are used here for a different purpose, namely, for providing an additional feature of the industry’s size. The levels of capacity utilization show the state of business cycles and also serve as an indicator of entry barriers, although this latter role may yield mixed signals, as Gilbert (1989) observed.

Group E shows some conventional indicators of industry structure: number of incumbent firms, their size distribution and life expectancy. According to the approach adopted here, these are endogenous variables that result from the competition process. They should be interpreted within the parameters determined by investment rates, technical change, market size and entry barriers. Conversely, the process of capital formation and the diffusion of innovations are better explained when accompanied by data on the demography of existing enterprises.

3.3. Special studies

The five groups of special studies listed in box 1 are, in general, very expensive. They depend upon large samples of quantitative and qualitative data collected at the firm level, the follow-up of several types of government actions, the adoption of sophisticated methodologies and a radical attention to accuracy. In principle, they should cover the whole economy or at least all relevant industries in terms of size and growth potential. But, excepting entry barriers, the findings of these studies remain valid for several years, even in periods of intense structural change. This peculiarity turns them affordable for every country.

Group F presents a taxonomy of entry barriers that is convenient for competition policy purposes. It distinguishes natural and institutional long-lasting entry barriers from those transitory obstacles created by technological innovations and government policies. This data set is evidently crucial for analyzing the competition process. Without it, the other economic indicators in box 1 would give a description of the industrial system that could be useful, but insufficient, for attending to the needs of the antitrust agency. Therefore, it should be updated on a continual basis, registering the arrival of new restrictions, the destruction of others and the subtle changes of existing ones. In fact, a specific routine should be established for updating the information on government-generated entry barriers. All regulatory measures, incentives and protection mechanisms should be classified by policy areas such as environment, foreign trade, public heath, national security, industrial development, technical standards, public services, etc. These actions could also be arranged according to their effects in each industry, such as raising transaction costs, protecting incumbents from potential entrants, reinforcing the technological power of innovating firms, stimulating product differentiation and market segmentation and outright entry prohibitions.

Group G includes the industry’s R&D expenditures and patented innovations. As an instrument for describing conditions of competition, these indicators highlight three overlapping features: [a] the sources of the competitiveness levels displayed in groups B and C; [b] whether domestic firms meet the international standards of technological performance; [c] whether the domestic competition process differs significantly from the world pattern. It is a well known fact that Latin American industries invest very little in R&D. The data base in box 1 will improve the knowledge about the causes and consequences of this fact and may also suggest guidelines for policy action.

The special studies in group H review, once again, the investment data, this time from the perspective of the international insertion of domestic industries. Inward and outward foreign direct investment should be classified by country of origin/host country and type of operation, such as joint ventures, mergers and acquisitions, creation of new foreign subsidiaries and routine expenditures inside the parent-subsidiary network of transnational corporations. The presence of foreign capital in the national economy and the international activities of domestic firms are defining factors of the competition process. Like other aspects of the industrial system, they can be either beneficial or detrimental to the public interest. It all depends on the established rules.

Finally, groups I and J conclude the description of the competition process. They are focused on the issues of increasing returns, transaction costs and market power, for the reasons discussed in section 2. To a great extent, these studies control the quality of the whole data base, since their results ought to be consistent with all remaining indicators, particularly those related to competitiveness and entry barriers.

4. Conclusion

In the preceding sections, the competition process was examined only from the perspective of the antitrust agency. However, the data base described here is obviously useful to other government agencies, the private sector and the academic community. The participation of the private sector would be particularly important for improving the quality of the information system and for creating a common knowledge between the government and the business community. That data base contains several relevant guidelines for investment decisions, by highlighting fast growing

markets, opportunities for new alliances, the most profitable sectors, technological trends, protected industries, etc. To established producers, it provides a benchmark for interindustry comparisons and for assessing the efficiency of their current market strategies. Thus, the objectives of competition advocacy, promoting transparency and institutional strengthening would be better served with the diffusion of those indicators throughout the country.

References

Baumol, W., Nelson, R. and Wolff, E., eds. (1994), Convergence of Productivity, Oxford University Press.

Baumol W. and Ordover, J. (1992), "Antitrust: Source of Dynamic and Static Inefficiencies?" in Antitrust, Innovation, and Competitiveness, Thomas M. Jorde and David J. Teece (eds.), Oxford University Press.

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

Bonelli, R. (1992), " Growth and Productivity in Brazilian Manufacturing Industries", Journal of Development Economics, 39: 85-109.

Cella, G. (1984), "Input-Output Measurement of Interindustry Linkages", Oxford Bulletin of Economics and Statistics, 46 (1).

Chenery, H., and Watanabe, T. (1958), "An International Comparison of theStructure of Production", Econometrica, 26.

Curiel, C. , Ferrin, J. and Genel T. (1996), "Barreras a la Entrada Derivadas de los Distintos Canales de Intervención del Estado", Discussion Paper, Pro-Competencia, Caracas.

Dixit, A. (1996), The Making of Economic Policy: A Transaction-Cost Politics Perspective, The MIT Press.

Gilbert, R. (1989), "The Role of Potential Competition in Industrial Organization", Journal of Economic Perspectives, 3 (3).

Jatar, A. (1993), "Implementing Competition Policy in Recently Liberalized Economies: The Case of Venezuela", Fordham Corporate Law Institute Annual Report, New York.

Khemani, R. and Dutz, M. (1995), "The Instruments of Competition Policy and their Relevance for Economic Development", in Regulatory Policies and Reform: A Comparative Perspective, Claudio R. Frischtak (ed.), The World Bank, Washington, D.C.

Ordover, J. (1990), "Economic Foundations of Competition Policy", in Competition Policy in Europe and North America: Economic Issues and Institutions, W.S. Comanor and others, New York: Harwood Academic Publishers.

Pinheiro, A.C. (1989), An Inquiry into the Causes of Total Factor Productivity Growth in Developing Countries, phd Dissertation, University of California.

Rodríguez, A., and Coate, M. (1996), "Limits to Antitrust Policy for Reforming Economies", Houston Journal of International Law, 18 (2).

Rodríguez, A., and Williams, M. (1994), "The Effectiveness of Proposed Antitrust Programs for Developing Countries", The North Carolina Journal of International Law and Commercial Regualtion, 19: 209-232.

Rodríguez, A., and Williams, M. (1995), "Economic Liberalization and Antitrust in Mexico", Revista de Análisis Económico, Santiago: Forthcoming.

Romer, P. (1990), "Endogenous Technological Change", Journal of Political Economy, 98, S71- S102.

Simpson, D., and Tsukui, J. (1965) "The Fundamental Structure of Input-Output Tables: An International Comparison", Review of Economics and Statistics, 36, 1965.

Solow, R. (1985), "Economic History and Economics", American Economic Review, 75:328-331.

Williamson, O. (1989), "Transaction Cost Economics", in Handbook of Industrial Organization, R. Schmalense and R. Willig (eds.), North-Holland.

2 - Revealed comparative advantage is measured by the share of the industry's exports in the country's total exports divided by the equivalent share for the world economy.

3 - For a discusion on the limitations of Lerner Index, see Ordover (1990)