Although the virtuous consequences of competitive product and factor markets for enterprise efficiency are fundamental tenets in most economists' thinking, the empirical case for those beliefs is relatively weak. Most empirical studies of firm and industry performance have focused on some measure of profitability as the outcome variable, rather than efficiency, but there is little necessary connection between the two.

Previous Studies of Competition and Efficiency

The relatively few studies examining efficiency have mostly produced rather weak results with respect to the effects of competition, and they have been hampered by data and identification problems. Three types of problems occur in this type of research. First, some studies, while informative, face difficulties controlling for unobserved heterogeneity and the possible endogeneity of market structure. For instance, market structure may be endogenous (or determined within the structure of equations) in that better performance of some firms may lead to higher concentration, so that there is no necessary relationship between the two variables in a cross-section.

A problem with studies using panel data is that data availability has typically limited them to examining only small numbers of firms spread across a wide variety of industries. In the stable economies of the West, there have been few opportunities to analyze firms that have undergone substantial changes in their market environments, particularly large numbers of firms undergoing diverse experiences of such changes simultaneously.

A third problem with most of the studies is that the potential efficiency gains due to competition that may be observable by the researcher are rather small, insofar as most of the economies in which such studies are undertaken are dominated by "workably competitive" markets, so that the general environment may still exert a disciplining force, even if the particular conditions facing the firm do not. Natural monopolies in the West, after all, usually operate in competitive markets for managers, labor, and most other factors; they can avail themselves of the latest technologies, organizational innovations, and managerial techniques; their performance can be compared at least along some dimensions with neighboring competitive firms; and instances of gross malfeasance can be publicly evaluated and remedied through a democratic process. All of these factors would seem to go quite some distance toward mitigating inefficiencies associated with monopoly power.

New Evidence: The Russian Experiment

The situation in Russia stands in stark contrast. Russia's "product market structure" at the beginning of the transition was determined by decisions that had been made by Soviet central planners rather than by market forces. Those decisions were motivated to a considerable extent by the need to monitor and control the production units and by political considerations, which tended to skew choices of location, scale, and integration; economic efficiency was frequently secondary. Moreover, prices and wages were controlled, output quantities and types were set by plan, sales were virtually guaranteed, soft budget constraints prevented bankruptcy, and firms could grow (or shrink) only by order of the central authority. Although there is some debate over the extent to which the planning system resulted in monopolized or highly concentrated sectors, it is clear that relative to market economies, the Soviet system had very few small firms and very low rates of entry and exit.

This ossified industrial organization was suddenly liberalized on January 1, 1992, when prices, foreign trade, supply arrangements, labor mobility, and the entry of new businesses were simultaneously freed. Where before there was little or no effective competition in any product or factor markets, suddenly Russian firms were forced to compete with one another and with new foreign competitors, while demand became more fickle and budget constraints became harder. Thus Russia presents an unusual quasi-experiment in which we may test whether an exogenously determined market structure affects firm efficiency. We may also examine the time pattern of adjustment to the liberalization shock (how quickly the increased competition may have improved enterprise productivity).

The shock of liberalization was not only unexpected, but also hit a set of enterprises that had been operating far from the production frontier when the "big bang" policies took effect. The plant and equipment, labor force, managerial skills, organizational capital, and modes of operating of these firms were built up during a period when there was nothing approaching a competitive market environment, essentially no private ownership existed, and efficiency and profits were not primary objectives. Since competitive shock hit, therefore, Russian firms have had the potential to exhibit large changes as they restructure along many dimensions and reorient their objectives from the state toward the market. For the researcher, the situation holds out the possibility for observing substantial differences in behavior.

Some peculiarities of Russian geography suggest that it may be fruitful to investigate regional dimensions of market structure. By far the largest country in the world, Russia is characterized by widely scattered industries frequently separated by vast distances and connected by poor infrastructural networks. While the possibility that product market scope should be defined by region as well as industry has been considered in research on U.S. market structure, the distances and transport and communication costs appear to pose much greater obstacles to unifying markets in Russia.

The geographic argument also has implications for factor markets, in particular for labor. While in principle it is clear that monopsony power may create slack that permits inefficiency precisely analogous to the slack under monopoly power, there have been no empirical studies of the relation, possibly because Western economies show rather little labor market concentration. Concentrated labor markets?in the limit, one-company towns?are much more common in Russia, suggesting the possibility that monopsony power may indeed be a significant factor in cushioning firm behavior.

A small number of other studies have examined the impact of competition in Russia and other transitional economies. Our research has several distinct advantages over earlier work in terms of the size and coverage of the data set, the time span of observations on each firm, and the availability of a variety of measures of market structure and competition. The panel data set we use for estimation purposes is comprehensive, including 75 percent of total employment in Russian industry in 1992, the year of the liberalization shock, and covering the seven years from 1992 to 1998. Since we have nearly the entire population of medium-size and large industrial enterprises, including information on their exact locations and disaggregated five-digit industries, we can use much more precise measures of market structure than those available to other researchers in Russia or many other countries. The size and scope of the database, as well as the unusual characteristics of Russia, permit us to investigate several geographic dimensions of competitive pressure and to examine competition in local labor markets. Our data also contain useful controls and instruments and much better and more disaggregated information on import competition than was heretofore available. We use this rich database to trace out the impact of several dimensions of market competition and other factors on firm-level efficiency.

Testing the Competition Effect.

Our analysis embraces a variety of dimensions of competitive pressure, including not only domestic market structure at the national level but also import competition, regional variation of product markets, transportation infrastructure, and labor market competition. We study the time path of the impact of the competitive shock - the abrupt liberalization of prices, entry, imports, and labor mobility - in January 1992.

We first test for determinants of firm survival. We find no statistically significant effects of competition on survival. Demand shocks have the expected effects: industry and regional growth and relative price changes all increase the probability of survival. Initial conditions also matter, as higher initial profitability, exporting activity, and size increase the probability of survival, while military-industrial complex affiliation lowers it. Foreign joint venture and predominantly private ownership lower the probability of survival relative to other ownership types.

The production function results provide strong evidence that domestic product market competition and local labor market competition have strong positive effects on efficiency, while import competition has no statistically significant effect. Domestic product market concentration has an effect, however, only once we take into account the fact that the geographic scope of the market differs across industries. The impact of liberalization appears only gradually in the domestic product market, taking about four years to attain about 90 percent of the long-run value, but we find no such lag with respect to labor market competition. Domestic product market and labor market competition show increasing marginal effects.

Better and lower-cost transportation infrastructure; more complete implementation of economic reforms such as privatization, price liberalization, limitation of production subsidies; reductions in bureaucratic regulation, and greater control of the regional economy by criminal groups all are found to facilitate domestic product market competition. Economic reforms facilitate import competition. Better and lower-cost transportation infrastructure also improves firm efficiency, by making workers more mobile, thereby reducing firms' monopsony power in the labor market.

Our results suggest that downsizing firms have more difficulty maintaining productivity in industries and regions with more negative demand shocks. Positive price shocks appear to have cushioned firms from having to adjust. Initial conditions are significant as well: higher initial profitability and being an exporter raise productivity, while the extent of exporting activity and military-industrial complex affiliation lower it.

We find that joint ventures and predominantly privately owned firms outperform other firms, even after controlling for potential selection bias in the determination of ownership. There is some evidence of decline in the relative performance of joint ventures and predominantly privately owned firms in recent years, however.

Policy Implications

The results have implications for competition policy. Previous research has shown that Russia's industrial structure is not more concentrated on average at the national level than that of the United States or other market economies. Russia's vast size, poor transportation infrastructure, and political barriers could potentially segment markets geographically, however, creating a large number of regional monopolists. If markets are indeed segmented, the competition authority might wish to focus more on reducing geographical barriers to competition than on reducing concentration.

Our results provide strong evidence of geographic market segmentation. Specifically, they suggest that a policy of reducing national or regional concentration that does not pay close attention to the geographic scope of the particular firm's market would lead to no gain in productive efficiency. A reduction in geographic market segmentation?which could be accomplished through investment in transportation infrastructure and implementation of pro-competitive policies such as privatization, price liberalization, reduction of production subsidies and bureaucratic regulation, and elimination of interregional trade barriers?would lead to significant productivity gains, however, by making it easier for already existing potential competitors to start competing with one another. J. David Brown and John. S. Earle are economists at the Stockholm Institute for Transition Economics and Research Fellows of the William Davidson Institute.