Strategic Adjustment of Price by Japanese and American Automobile Manufacturers

Kaye G. Husbands

• Publisher: Routledge
• Year: 2017
• ISBN:
  9781351664615
• Edition: 1
• Topic: Business
• Subtopic: Automobile Industry
• Index: Business

This book, originally published in 1993, develops for the US automobile industry a demand-supply model which incorporates both wholesale and retail sectors and which allows strategic pricing behaviour of US and Japanese producers to be internally determined and its effects on market behaviour and national welfare analyzed. It develops the framework for and presents the results of an econometric simulation of the transaction and wholesale prices, quantities demanded and produced, manufacturer's costs and factor demands. The impact of the Voluntary Export Restraint of 1981 on profits and consumer welfare are generated from the simulation results.

1.0    Introduction

In March 1981, the Japanese government agreed to U.S. demands for limits on the exports of Japanese passenger cars to the United States.¹ Domestically, proponents embraced the Japanese-U.S. Voluntary Export Restraint (VER) as a policy instrument which would stabilize profits and employment in the automobile industry. U.S. automobile producers perceived the VER as a means by which Japanese penetration of the then growing domestic small car market would be impeded. They also anticipated that strategic investment of revenues earned as a result of the VER would facilitate their efforts to gain a strong competitive position in domestic and world markets.

U.S producers’ expectations of significant gains from the imposition of the VER had some merit. For international trade theory predicts that in oligopolistic industries--such as the automobile industry--equilibrium prices and profits of foreign and domestic firms will be higher under a VER than under free trade. However, quantities sold of the domestic firms may increase or decrease with a VER, depending on the values of own- and cross-price elasticities of demand. Thus, it is possible for national welfare to increase only if domestic quantities are increased, average price-cost margins of domestic firms are sufficiently large and gains to foreign firms are sufficiently small. But some economists see the more likely result of the VER would be to show the losses of domestic consumers exceeding the gains of domestic producers and, therefore, a decline in the national welfare. (See Krishna, 1983/84; Harris, 1984; Itoh and Ono,1984; Ono,1984).

The VER’s effect on equilibrium prices and quantities and thus on consumer and producer surplus is, in part, the result of its influence on competition in the targeted industry. Krishna (1984) shows that a VER set at or slightly below the free trade level creates a more collusive market. Specifically, by assuming that a Bertrand-Nash equilibrium exists during free trade and that the domestic and foreign goods are close substitutes, Krishna shows that the profit functions and hence the price reaction functions of representative foreign and domestic firms are altered by the existence of a VER.²

Under these conditions, an equilibrium in pure strategies is no longer attainable. The new equilibrium is instead a mixed strategy equilibrium where the foreign firm sets a single price and the domestic firm randomizes over its set of profit maximizing prices in selecting its market price. Both firms earn higher profits, with the domestic manufacturer earning the profits of a price leader.³ Thus, the presence of a VER can effect a change in the manufacturers’ behavior.

The theory explains why higher prices accompany VERs and why foreign manufacturers favor VERs over other forms of trade restraints. However, the theory is limited in three ways. First, without estimates of demand elasticities, it cannot unambiguously predict the VER’s impact on quantities demanded of the domestic good or on national welfare. Second, it cannot predict the magnitudes by which prices or profits increase, or the magnitude of losses to consumers during a VER regime. Third, the assumed state of market behavior in the pre-VER period in part determines the expected behavioral outcome as well as the equilibrium prices and quantities under the VER. Thus, if the rivalry among automobile producers does not coincide with that posited by the theory, the theoretical model becomes an inadequate predictor of the VER effects for that industry. Empirical models are, therefore, used in this thesis to examine such issues.

Previous empirical studies of the restraints’ impact on the automobile industry all agree with the theoretical predictions that the VER hurts purchasers of automobiles and benefits both domestic and Japanese automobile manufacturers (see Feenstra (1984); Gomez-Ibanez, Leone and O’Connell (1983); Crandall (1984,1985); Tarr and Morkre (1984); Hickok (1985); Mannering and Winston (1987); Collyns and Dunaway (1987); Willig and Dutz (1987); Dinopoulos and Kreinin (1988)). All of these studies also find that average prices of domestic and foreign makes would have been lower had the VER not been imposed. Not surprisingly, there is no consensus about the restriction’s effect on quantities sold of domestically produced cars. Some researchers estimate that fewer domestic makes would have been sold without the VER, while others reach the opposite conclusion. This ambiguity in the overall findings could stem from the differences in methodologies applied as well as from the nonuniform nature of the data.

Although these empirical studies verify the theoretical predictions of Krishna and Harris referred to above, none has attempted to answer the following two important questions regarding the effects of the VER on pricing strategies and profits: (1) How was the nature of competition in the domestic car market affected by the VER? (2) How much of the profits (or losses) generated by domestic sales of Japanese cars accrued to domestic dealers and thus should be entered as an adjustment to U.S. national welfare?

Therefore, the purpose of this study is to develop for the domestic automobile industry a demand-supply model which incorporates both wholesale and retail sectors, and which allows strategic pricing behavior of U.S. and Japanese producers to be endogenously determined. It is assumed that similarities in technologies permit the aggregation of manufacturers in which the parent firm is located. Thus the analysis proceeds as though there were only two firms, U.S. and Japanese. The cars sold by U.S. producers are categorized as either small or large, while those sold by Japanese producers are categorized only as small cars.

A time series approach is used. The pre-VER period runs from the first quarter of 1973 to the first quarter of 1981, while the VER period runs from the second quarter of 1981 to the fourth quarter of 1986. The model is estimated for the entire period 1973:1 to 1986:4 and a simulation of the model is used to forecast retail (transaction) prices, wholesale prices and quantities for the VER period. These predicted values are assumed to be those which would have prevailed had the restraints not been imposed. They are used together with the actual values of these variables to calculate the VER’s impact on prices, quantities and profits of both manufacturers and dealers. In essence, it has been assumed that the VER was the predominant factor which did not exist during the historical period and which influenced market equilibrium during the VER regime.

The degree to which consumers perceive Japanese cars and U.S. cars as substitutes should be reflected in the model’s estimates of price flexibility⁴ and of demand elasticities. These estimated elasticities are expected to indicate that Japanese cars were better substitutes for U.S. small cars than for U.S. large cars. The type of market behavior which existed before and during the VER should be reflected in the model’s estimates of conjectural variations (CV) parameters. Although U.S. manufacturers are the dominant producers in the domestic market in terms of market share, Japanese producers are expected to be the price leaders in the small car market. No prior expectation is made with respect to the U.S. large car price response to a change in Japanese car prices. Comparison of the CV parameter estimates for the pre-VER and VER periods is expected to reveal a movement toward U.S. price leadership in the small car market as quantity restraints on Japanese cars prevent the Japanese from maintaining their leadership role. The consistency of the estimated conjectural variation parameters with actual market behavior is statistically tested.

For both Japanese and U.S. cars, the model’s simulation of the 1980’s domestic market without the influence of the VER is expected to yield lower retail and wholesale price than those which existed under the VER. There is no ex ante prediction, however, of the VER’s effect on quantities sold of U.S. cars or of its effects on profits earned by producers and domestic dealers. This is because these outcomes depend heavily on the model’s estimates of the demand slopes and of the conjectural variation parameters.

In Chapter 2 of this thesis, the previous empirical models which focused on assessing the VER’s impact on profitability in the domestic automobile market are reviewed. Two methodologies are distinguished among those models: those which rely on base-year data to predict either the VER or the non-VER equilibrium prices and quantities for the 1980’s, and those which use the ex post forecasting method to predict the non-VER equilibrium prices for comparison to actual prices for the 1980’s. The advantages and limitations of the two methodologies as well as specific assumptions and procedures used in the previous empirical models are discussed. The outcome of this critical review suggests the use of the ex post forecasting method for the present analysis and highlights the necessary innovations for obtaining more accurate estimates of the VER’s effects on domestic welfare.

Chapter 3 describes the demand-supply model which is developed for the current analysis. Since one purpose of this study is the assessment of the VER’s effects on profits of both dealers and manufacturers, the model includes consumer demand and dealer demand equations. These equations are used in simulation to predict transaction prices and quantities. To facilitate the estimation of the type of market behavior during the pre-VER and VER periods, manufacturer behavioral equations are also included in the model. These equations enable the model to estimate the CV parameters as well as to simulate wholesale prices without the VER. Since the manufacturer behavioral equations require estimates of marginal cost in order to identify the CV parameters, cost equations are included in the model. Input demand equations supply the necessary cross equation restrictions on the cost parameters. Marginal costs of Japanese and U.S. automobile manufacturers are thus analytically derived from the cost equations and are substituted into their respective manufacturer behavioral equations for estimation within the simultaneous system. These estimates of marginal costs are used in the determination of the VER’s effects on manufacturer profits.

The availability of data and methods used to convert the raw data into the required time series are described in Chapter 4. Aggregate time series data on transaction prices, wholesale prices and quantities were constructed from model or segment specific data. The procedures used to convert firm specific data on production costs to aggregates for Japanese and U.S. producers are also given in detail. The sources of these and other data such as demand shift variables and price deflators are also disclosed.

In Chapter 5, the results of the analysis are presented and compared to selected results of previous studies. Of all the findings, the most significant and, possibly, surprising are:

1.Japanese price leadership vis-à-vis U.S. small and large cars cannot be rejected for the pre-VER period.

2.There is evidence of movement toward Bertrand pricing or toward U.S. price leadership with the VER.

3.U.S. manufacturers actually sold fewer cars under the VER than they would have sold had the VER not been imposed.

4.U.S. automobile manufacturers and domestic automobile dealers win, while Japanese automobile manufacturers and domestic automobile consumers lose because of the VER, dealers gaining more than manufacturers. The total U.S. welfare effect is positive.

Estimates of price flexibility, of demand and of gasoline price elasticities are also presented in this Chapter.

Concluding remarks and suggestions for future research are set forth in Chapter 6. The Appendices contain regression results as well as lists of the automobiles used in creating the quantity, wholesale price and transaction price series.

---

¹The VER is sometimes referred to in the literature as the Voluntary Restraint Agreement (VRA). Beginning April 1, 1981, the Japanese announced their first quantity limits on car exports to be 1.68 million for the year ending March 31, 1982. The initial agreement stated that the VER would be in effect for the following two, possibly three years, but with upward adjustments to be made according to increases in domestic sales of U.S. cars. The Japanese have continued to announce their export limits and have set restraints for their fiscal year beginning April 1989 at 2.3 million.

²These results also hinge on the following assumptions. The industry has one domestic and one foreign firm that strategically set prices to maximize profits earned in the domestic market. The industry’s product is heterogeneous and domestic consumers perceive the variants of the product as imperfect substitutes. Rationing of the foreign good affects demand for the domestic good. There exists a numeraire good and thus income effects are assumed away. The Nash equilibrium concept is imposed and the reaction functions and the market equilibrium satisfy existence, uniqueness and stability conditions with and without the VER. The VER is set at or close to the free trade level.

³Harris (1984) also assumes that firms exhibit Bertrand-Nash behavior before the VER is imposed. He derives similar results to Krishna’s. However, he considers only the set of prices for which the VER is binding and therefore derives an equilibrium in pure strategies where the domestic firm becomes the price leader under the VER regime. In Krishna (1983) there is also a case where it is assumed that the domestic firm is a Stackelberg leader and specific demand functions are introduced. The equilibrium in pure strategies exists and is unique with or without the VER. In this case the qualitative results reported above hold for both prices and profits. Domestic outputs are shown to decrease because of the specified numerical values of slopes of the demand curves.

⁴The responsiveness of price to a change in quantity.

2.0 Previous Studies

2.1 OVERVIEW

This Chapter presents a historical review of empirical studies on the Japanese-U.S. VER and discusses the merits and limitations of the methodologies which have been used to study the VER’s efficacy. In keeping with the present analysis, only those studies which evaluate the VER’s effects on the profitability of firms in the domestic automobile market are reviewed. Studies which primarily assess the VER’s impact on consumer welfare or on product quality are therefore excluded from the discussion.⁵

The studies reviewed can be placed in one of two methodological categories:

(1) base-year simulation method [Gomez-Ibanez, Leone and O’Connell (1983), Tarr and Morkre (1984)], or

(2) ex post forecast method [Willig and Dutz (1987), Crandall (1985), Mannering and Winston (1987), Collyns and Dunaway (1987)].

The latter technique is distinguished from the former in that actual changes in the economic conditions of the market are reflected in the simulated values for the forecast period.

Sections 2.2 and 2.3, respectively, provide a critical examination of each methodology. Each section also contains a critique of the individual models. This is accomplished, wherever possible, by comparing and contrasting parallel procedures peculiar to those studies which share the same general methodology. Section 2.4 gives summary remarks.

2.2 BASE-YEAR SIMULATION METHOD (BYSM)

BYSM is a comparative statics exercise in which the base year solutions to a system of equations is compared to other particular solutions of the same system which are generated under alternative economic s...