The competitive role of import penetration.
The United States has run ever-increasing record trade deficits in the 1980's. For example, the 1984 deficit of $123.3 billion was almost twice the 1983 level. Most of the growth, some 92%, is accounted for by the manufacturing sector of the economy, which is the traditional focus of industrial organization research. While the deficit's rate of growth has fallen recently, 1987's deficit of $171 billion is still considered to reflect a fundamental problem by some policy makers.(1)
In response to this, the U.S. Department of Commerce proposed relaxing antitrust enforcement, predicting that the change would improve the competitiveness of domestic firms vis-a-vis their foreign counterparts. A major component of the proposed relaxation focuses on section 7 of the Clayton Act, which forbids mergers that substantially reduce competition. Along with the antitrust enforcement proposal, the U.S. Congress has considered a host of trade legislation that would address overall competitiveness of U.S. industries.(2)
Of key importance to this article, the antitrust revisions address mergers involving larger firms in the same industry, that is, mergers that would generate higher domestic concentration ratios. In determining the legality of mergers, judges would be required to consider the presence and impact of foreign competition on domestic efficiency. Also, the President and the International Trade Commission would designate import-injured industries where legal restrictions against merger activity would particularly be relaxed.
The proposals for antitrust relaxation are, at first glance, inconsistent with conclusions drawn from prior work on the relationship between import levels and domestic concentration. For example, Marvel finds that concentration is positively related to import penetration.3 However, finding a positive relationship between concentration and imports does not necessarily mean that a higher level of concentration will be associated with a larger incidence of import penetration.
The direction of causality is the reason for the uncertainty. Caves reports that smaller enterprises disproportionately lose market share in the face of import competition, which implies that higher import penetration will cause concentration ratios to be higher.(4) The causality question is obviously an important issue when examining the effects of import penetration on relative market shares of large and smaller firms.
This article reexamines the basic role played by import competition on industrial market structure. However, we take a significantly different approach from that used in prior studies. Instead of focusing on the question of whether or not import penetration tends to be higher in concentrated industries, we test the (possibly different) interaction effects of imports and three industry concentration groups on profitability. If monopoly rents attract imports and concentration is a proxy for those rents, then imports should have negative effects on the profits of firms in the concentrated industries. Alternately, if higher profits across larger firms result from efficient management, then imports should have no effect on the profits of larger firms in concentrated industries. If Caves' hypothesis is correct, then imports should have the largest impact on the profits of smaller firms.
Our results indicate that there is an important interaction effect between imports and market concentration. Consistent with the studies that show market concentration and imports are related, we find imports to have a significant effect only on the profits of firms in concentrated industries. However, we find that it is small firms, not large firms, in concentrated industries that are negatively affected by increased incidence of import penetration. The results are therefore consistent with Caves' finding that imports primarily affect the market shares of small firms.
The policy implications are clear. If past antitrust merger policy has in some way promoted a hothouse existence for smaller, higher cost firms in order to avoid higher domestic concentration ratios, that policy has also provided a window of opportunity for imports to enter the market. When faced with import competition, the weaker small firms will disappear, either by merger or bankruptcy.
II. A model of market structure and import penetration
As outlined above, there is ample precedence to support the notion that import penetration may have different effects in concentrated and unconcentrated industries. That possibility is one reason prior researchers have examined domestic concentration as a determinant of import penetration. The model developed here is based on a series of studies on the interaction of unionization and market structure.
Karier, and Hirsch and Connolly have tested the possible interaction of unionization and market structure in the determination of industry profits.(5) Karier used three unionization-market structure interaction variables as determinants of industry price-cost margins.(6) He found that unions have a significant, negative effect on the profits of high and moderately high concentrated industries but no statistically significant association with the profits of firms in low concentrated industries. Karier concluded from these results that monopoly returns are a principal source of union rents.
Hirsch and Connolly tested the Karier model with data from 367 Fortune 500 firms. In contrast to Karier's results, they found no interaction between market structure and unionization.(7) They concluded that union rents are generated from sources other than profits associated with market structure characteristics.
Based on the interaction model initially developed by Karier, we propose the following model for testing the possible interaction of imports and market structure in determining profit indices:
PCM = f(CR4, ASR, KSR, GR, UN, IMP*DH, IMP*DM, IMP*DL), where PCM is industry price-cost margins and is an index of profitability.(8) As with any proxies for the concept of economic profits, this one has its advantages and disadvantages. The principal advantage is that the measurement is constructed from plant-level observations and therefore does not have the contamination problem associated with firm-level profitability data. Diversified firms generate profits from more than one industry, which creates a problem when trying to relate industry profits and market shares using firm-level data. As to a disadvantage, the cost of capital and out-of-plant expenses (like advertising) are not deducted from revenues. This is partially corrected in our model by including capital-sales ratios (KSR) and advertising-sales ratios (ASR) on the right-hand side of the equation.(9)
Domestic concentration, CR4, is hypothesized to be a positive determinant of price-cost margins. Such a relationship may reflect either the monopoly power of large firms in concentrated industries or the ability of those firms to earn above-average profits because they are more efficient than their smaller rivals, or both of these factors.(10)
Even in competitive industries, industry growth can result in firms' temporarily earning above-average short-run profits. Price-cost margins should therefore be positively related to industry growth, GR. The inclusion of GR also causes CR4 to yield a steady-state reading on the link between profits and concentration.
Unionization, UN, is also included as a determinant of price-cost margins across industries. Although there is dispute as to the way unionization interacts with market structure, there is little doubt that unionization should be inversely related with profit indices across industries, all else being equal.
The remaining three variables in the model show the possible interaction of imports and market structure. DH, DM, and DL are three dummy variables that represent high, moderate, and low concentration groups in the economy.(11) The level of imports times the three dummy variables therefore represents the effect of import penetration in high, moderate, and low concentrated industries.
Three different price-cost margins were tested with this model, with the first being avenge industry price-cost margins. Because Caves has found that imports affect the sales of smaller firms disproportionately, the model was also tested using the average price-cost margins of the largest four firms in each industry and the margins of all firms smaller than the largest four firms. The choice of the largest four firms as representing "large" is due to the obvious reasons. First, it is the group of large firms for which the census provides enough information to allow calculation of a profit index. Second, the choice corresponds to the division used in four-firm concentration ratios. If the Caves' result (small firms disproportionately lose more in sales to imports than large firms) applies consistently in this model, imports will have the largest effect on the profits of small firms. Whether or not this effect will be influenced by market structure remains an empirical question.
III. The data and results
The models are tested with a 1977 data set composed of 327 four-digit manufacturing industries drawn from a total potential universe of 450 such industries. Deletions resulted from empty cells in the input-output tables from which the data were drawn and because of the absence of data in the Census of Manufacturers for other key variables.(12) Values for all price-cost margins, four-firm concentration ratios, capital intensity, and growth are taken from the 1977 Census of Manufacturers.(13) Advertising and import levels across industries are taken from input-output tables. Unionization is measured at the lowest aggregation level available, which is the three-digit level as reported by Kokkelenberg and Sockell.(14) A description of the variables and their means are reported in table 1.
Table 1 Variable Definitions and Means-327 Four-digit Manufacturing Industries Variable Definition Mean PCM Industry price-cost margin. Calculated as value- .278 added less payroll expense divided by value of shipments. PCM4 The price-cost margin of the four largest firms. .295 PCM<4 The price-cost margin of all firms smaller than the .264 four largest firms. CR4 Four-firm concentration ratio. .396 ASR Advertising-sales ratio. .011 KSR Capital-sales ratio. .286 GR Industry growth. Percentage change in sales, .722 1972 to 1977. UN Percentage of work force unionized. .362 IMP*DH Imports relative to domestic sales in industries .021 with concentration ratios greater or equal to 60%. IMP*DM Imports relative to domestic sales in industries .018 with concentration ratios less than 60% and greater than or equal to 40%. IMP*DL Imports relative to domestic sales in industries .012 with concentration ratios less than 40%. NOTE: Data to construct PCMs, CR4, KSR, and GR are from 1977 Census of Manufacturers, vol. II, parts 1, 2, 3. See tables entitled Historical Statistics for the Industry: 1977 and Earlier Years. Percentages of value-added, payroll, and value of shipments for the two groups of firms (four largest and all others) are taken from the Census volumes. See Share of Selected Items Accounted for by the 4 Largest Companies and Complementary Groups Ranked on Value of Shipments for Each Industry: 1977. Advertising and import levels are from the 1977 input-o utput tables. Unionization levels are taken from Kokkelenberg & Sockell, Union Membership in the United States, 1973-1981, 38 Indus. & Lab. Rel. Rev. 497 (July 1985).
The determinants of industry, large firm, and small firm price-cost margins are presented in table 2. Regression equation 1 shows the determinants of industry price-cost margin, PCM. Ignoring the import-market structure variables for the moment, we note all other variables have the expected signs and are significant. Industry profitability indices are positively and significantly related to industry concentration, advertising intensity, capital intensity, and growth.
[TABULAR DATA 2 OMITTED]
Now, consider the import-market structure interaction terms in the PCM regression. Of the three interaction terms, only IMP*DH is significant. Following Karier, the interpretation of this result is that imports only have a significant, negative effect on the profitability of firms in concentrated industries. The results are consistent with Marvel's finding that imports and market concentration are positively related. Marvel offers a monopoly rent explanation for the positive relationship between imports and domestic concentration: The lure of monopoly returns in concentrated industries attracts imports. That is also a possible explanation for our finding that imports are inversely related to profitability of firms in high concentrated industries, but not in moderate or low concentrated industries. But, as will be explained shortly, that explanation appears less plausible after considering the regression equations for large and small firms' profit indices.
The initial reason for our interest in the relationship among imports, market structure, and profitability of large and small firms was the finding by Caves that imports have a disproportionate effect on the market shares of large and small firms. Caves learned that it is small firms that lose the most when imports increase. The results in regression equations 2 and 3 address the issue of firm she and the competitive role of imports.(15) The determinants of large-firm price-cost margins, PCM4, are shown in regression equation 2. Imports are not significantly related to the profitability of large firms in any market structure. However, turning to the determinants of small-firm price-cost margins, PCM<4, imports are significantly related to the price-cost margins of small firms in both high and moderate concentrated industries.
The result in table 2 give new insight into the competitive role that imports play in the economy. Imports appear to compete primarily against small firms in concentrated industries. The result is consistent with the efficiency differential hypothesis developed by Demsetz: Import penetration takes advantage of the inefficiency of small firms relative to large firms in concentrated industries. This would explain why Caves found that imports affect the market shares of small firms more than large firms. Although the finding that imports and concentration are positively related is consistent with Marvel's results, the finding is not consistent with his monopoly rent explanation of this relationship. If that story were correct, then imports should have a negative effect on the profits of large firms in concentrated industries, perhaps in conjunction with negative effects on smaller firm profits.
IV. Final thoughts
In this article, we suggest there is an important interaction between import competition and market structure characteristics. An examination of data from 327 four-digit manufacturing industries leads us to believe that imports have their strongest competitive effect in concentrated industries. Importantly, it is small firms in these industries that are affected the most by import competition.
The results reconcile the findings of several prior studies on import penetration and market structure and have important implications for antitrust merger policy. Relaxing antitrust policy so that efficiency differentials across the size distribution of firms is minimized, reduces the ability of imports to enter domestic markets. On the other hand, an overly restrictive policy against horizontal mergers may spawn smaller less efficient domestic firms that can be displaced by foreign competitors. Put differently, our results suggest U.S. market structure will be determined largely by market forces in spite of merger policy. However, antitrust policy will determine global ownership patterns of the larger firms that compete effectively in the market. (1) The trade deficit data are taken from the U.S. Department of Commerce, International Trade Administration, United States Trade: Performance in 1984 and Outlook (June 1985) and United States Trade: Performance in 1987 and Outlook (June 1988). (2) See John C. Hilke & Philip B. Nelson, International Competition and the Trade Deficit (Federal Trade Commission, 1987). (3) See Marvel, Foreign Trade and Domestic Competition, 18 Econ. Inquiry 103 (January 1980). Rhoades finds only a tenuous positive relationship between concentration and import levels. Using 1967 and 1972 data, he found a positive relationship between the two variables in only two of eight tests. (See Rhoades, Wages, Concentration, and Import Penetration: An Analysis of the Interrelationships, 13 Atlantic Econ. J. 23 (July 1985).) In a Federal Trade Commission study, Hilke & Nelson, supra note 2, examine the relationship between various measures of imports and exports in a large multivariable model that includes Herfindahl indices and adjustments for minimum efficient scale (MES) of plants across 360 industries. They find no statistical association between their measure of concentration and import penetration for the years 1975, 1981, 1982, and 1983 and interpret the results to suggest that concentration is not related to comparative advantage (as related to import penetration) when adjustments are made for minimum efficient scale of plant. However, they interpret their MES variable to indicate that antitrust actions have limited U.S. firms from achieving the scale of plant needed to compete against their foreign counterparts. (4) Caves, Trade Exposure and Changing Structure of U.S. Manufacturing Industries, in International Competitiveness (A. Michael Spence & Heather A. Hazard, eds. 1988). (5) See Karier, Unions and Monopoly Profits, 67 Rev. Econ. & Statistics 34 (February 1985); and Hirsch & Connolly, Do Unions Capture Monopoly Profits? 38 Indus. & Lab. Rel. Rev. 497 (July 1985). (6) Karier used two-digit manufacturing data across states, which raises a problem. There are 20 two-digit industries in manufacturing. The market share of the four largest firms at this level of aggregation is not very meaningful. For that reason, most studies of market power employ data at the four-digit level of aggregation, which divides the manufacturing sector into 450 industries. Our study uses four-digit manufacturing data. (7) Essentially, they argue that the unionization-market structure effect vanishes when unionization-research and development interaction terms are added to the model. The fact that three-digit industry data and not firm-level unionization data, which are not available, were used in the study presents a problem. (8) Price-cost margins are defined as value of shipments less cost of materials and payroll, divided by value of shipments. (9) The validity of price-cost margins as indices of profitability has been questioned by Liebowitz. (See Liebowitz, What Do Census Price-Cost Margins Measure? 25 J. L. & Econ. 231 (October 1982)) Vie leave it to other researchers to test the robustness of our results by using other data sets and other profit indices. (10) See Demsetz, Industry Structure, Market Rivalry, and Public Policy, 16 J. L. & Econ. 1 (April 1973) and Clark, Davies & Waterson, The Profitability-Concentration Relation: Market Power or Efficiency? 32 J. Indus. Econ. 435 (1984). (11) If the four-firm concentration ratio is equal or greater than 60%, the industry is classified as highly concentrated. Moderate and low concentration are defined as 0.4 <= CR4 < 0.6, and CR4, < 0.4, respectively. The categories correspond closely to the divisions used by Karier and then by Hirsch and Connolly. (12) A list of the industries used in the study is available on request. (13) A more detailed explanation of data sources is provided in table 1. (14) Kokkelenberg & Sockell, Union Membership in the United States, 1973-1981, 38 Indus. & Lab. Rel. Rev. 497 (July 1985). Salinger, and Hirsch & Connolly use three-digit unionization data to measure unionization for each firm in their samples. (See Salinger, Tobin's q, Unionization, and the Concentration-Profits Relationship, 15 Rand J. Econ. 159 (1984) and Hirsch & Connolly, supra note 5.) Using three-digit unionization for four-digit industry observations, as in this sample, is more justifiable than matching three-digit industry and firm data. Consistent with this article, Domowitz, Hubbard, and Patterson merge four-digit industry data with three-digit unionization data in their work. (See Domowitz, Hubbard & Patterson, The Intertemporal Stability of the Concentration-Margins Relationship, 35 J. Indus. Econ. 13 (September 1986).) (15) There are five variables other than the import-market structure variables in the three regression equations. Advertising intensity, capital intensity, growth, and unionization perform similarly in industry-wide regressions and in large and small firm regressions. Only market concentration performs differently: It is significant only in the industry equation. This is consistent with prior studies and, as Scherer notes, is due to the greater weight accorded to the profits of large firms in concentrated industries relative to their counterparts in unconcentrated industries. (See F. M. Scherer, Industrial Market Structure and Economic Performance 282 (1980).)
|Printer friendly Cite/link Email Feedback|
|Author:||Chappell, William F.; Yandle, Bruce|
|Date:||Dec 22, 1992|
|Previous Article:||Bigness, junk, and bust-ups: end of the fourth merger wave?|
|Next Article:||An economic evaluation of federal antitrust activity in the manufacturing industries: 1980-1985.|