DRAFT FOR STAFF USE ONLY A FRAMEWORK FOR ANALYSIS OF BILATERAL TRADE AND COUNTERTRADE: THE CASE OF THE CHjEA COUNTRIES ;- 2 ter ;,,urrel 1 CPD Discussion Paper No. 1986-16 March 1986 CPD Discussion Papers report on work in progress and are circulated for Bank staff use to stimulate discussion and comment. The views and interpretations are those of the authors. A Framework for Analysis of Bilateral Trade and Countertrade: The Case of the CMEA Countries Peter Murrell University of Maryland and Consultant, Bank Assistance Policy Division, CPD March 1986 I would like to thank Zdenek Drabek, Michael Marrese, Richard Michalski, Randi Ryterman, Martin Schrenk, Josef van Brabant, Jan Vanous, and Thomas Wolf for helpful advice in the writing of this paper. I alone am respon3ible for any opinions and errors that appear here. 307/cover/P .Murrell/MS/mcc/03.18.86 TABLE OF CONTENTS Page I. Introduction ................ ................................. 1 II. A Stylized Description of Some Important Features of Trade Within the CMEA ....................................... 8 a. Bilateralism ............................................. 9 b. The Transferable Ruble .................................... 9 c. Trade Denominated in Dollars .............................. 10 d. Prices in the CMEA ........................................ 11 e. 'Soft' and 'Hard' Goods ................................... 13 f. The Special Case of 'rrade with the USSR ...................1 g. Deficits ................................................. 15 III. On Errors Frequently Committed When Analyzing CMEA Trade .......16 a. Using The Official Ruble-Dollar Exchange Rate ..............17 b. Ignoring Ruble Trade ....................................... 19 c. Using the Same Exchange Rate in Transactions with All CMEA countries ......................................... 21 d. Using the Same Exchange Rate for all Types of Goods ........21 e. Indiscriminate Application of the Law of One Price .........22 f. Assuming an Accounting Dollar is a Real Dollar .............23 g. Forgetting the Special Character of Commercial Relations with t-he USSR ....................................24 h. Assuming that a Country's Terms-of-Trade are Independent of Trade Policy ............................... 25 IV. On the Measurement of CMEA Trade Flows and Non-convertible Currency Debts .................................26 a. The Shadow Price Approach.................................. 26 b. The Model and Notation .................................... 30 c. Valuing Trade Flows ........................................ 33 d. Valuing Debt ............................................... 37 e. The Importance of the Distinction Between Hard and Soft Goods ............................................. 40 f. Valuing Transferable Rubles Used in Transactions with the USSR ............................................ 43 307/contents/P.Murrell/MS/mcc/03.04.86 Page V. Valuing Trade Flows: What is Possible .......................... 47 a. Introduction ..................... ...... 47 b. The Information Available or Required ...................... 49 c. Reconstructing Trade Statistics: A Sophisticated Procedure ................................ 54 d. Updating Information on TR/$ Exchange Rates: An Ad Hoc Procedure ....................................... 56 e. Some Illustrative Calculations ............................ 57 f. Conclusions .............................................. 60 Appendix A The Relevance of the Above Arguments For Countertrade ..................................... 63 Appendix B - A Checklist of the Information Needed to Evaluate Foreign Trade Statistics and Foreign Trade Policy in the CMEA Countries .......................... ...... 66 Appendix C: Proofs of Propositions ...............................70 Footnotes .........................................................77 References ........................................................81 307/contents/P.Murrell/MS/mcc/03.04.86 ABSTRACT The theory of international trade has paid little attention to the issues that arise when bilateralism is an important element of a country's trade relations. Such a lacuina impedes the ability of the economic analyst to assess creditworthiness and to formulate policy advice for countries that conduct a significant share of trade within the framework of bilateral agree- ments. Moreover, the analyst also faces difficulties in merely evaluating the information embodied in trade statistics obtained from such countries. The present paper constitutes an initial attempt at formulating and solving some of the problems of economic analysis that arise when bilateralism and non- convertibility are important features of trade. The discussion focuses on the case of the CNEA countries. However, it should be emphasized that many of the paper's arguments are implicitly relevant to nations that participate heavily in countertrade. The investigation begins with a brief stylized description of the trade processes of CM4EA countries. The emphasis is on identifying the distinguishing features of CMEA trade behavior. Then the paper examines the most common errors of analysis that occur when one ignores the special charac- teristics of bilateral trade in general, and CMEA trade in particular. The ensuing theoretical discussion establishes results that can be used to analyze commercial policy under bilateralism and to construct econo- mically meaningful statistics from trade and debt data denominated in non- convertible currencies. The discussion suiggests that one should construct shadow prices of the transferable ruble in dollar terms. Shadow price formulae are developed and their application is considered in a variety of contexts. A further contribution of the theoretical analysis is that it helps to identify concepts that are important in analyzing the effects of bilateralism. The paper also examines the possibilities of applying the shadow price formulae in a practical setting. Previous empirical work on the valuation of the transferable ruble is discussed and possible sources of data are reviewed. Suggestions are made concerning different approaches to valuation, conducted at varying levels of sophistication. The paper concludes with an example of the application of an extremely simple method of revaluing transferable ruble trade flows. The analysis leads to the compilation of a checklist of information on institutional rules and practices needed for a realistic assessment of creditworthiness and trade performance of a country with a sizeable counter- trade share (Appendix B). 307/contents/P.Murrell/MS/mcc/03.18.86 Vladimir, a Russian dog-owner, told his friend Igor that he was planning to sell the animal. "How much?", asked Igor. "Two thousand roubles", replied Vladimir confidently. "You must be joking", said Igor, "that mongrel will be lucky to get two hundred roubles." A few days later, Vladimir met Igor again and told him that he had sold the dog -" and for two thousand roubles at that. Igor was astonished. "Nothing to it", said Vladimir. "I went into a pet shop and the man there happened to have two cats for sale at a thousand roubles each." The Economist, October 5,'1985 307/contents/P.Murrell/MS/mcc/03.04.86 I. Introduction 1.1 As one of its central elements, the theory of international trade and finance employs the assumption of multilateralism. Trade is assumed to be conducted in convertible currencies. Each country, therefore, faces a single balance-of-payments constraint. While this assumption may be viewed as a reasonable approximation for a majority of international transactions, in an important subset multilateralism does not hold. Trade within the Council for Mutual Economic Assistance (CMEA) is carried out effectively on the basis of bilateral balance. Each CMEA country, therefore, simultaneously faces a number of balance-of-payments constraints. Moreover, a degree of bilateralism can occur in trade between market economies to the extent that countertrade contracts require bilateral balancing of subsets of a country's trade. 1.2 There exists little theoretical analysis of the international trade behavior of countries facing multiple payments constraints. (In the Handbook of International Economics (Jones and Kenen, 1985), there is no mention of bilateralism and non-convertibility). This theoretical lacuna, of course, impedes our ability to understand and predict the behavior of CMEA countries. On a more practical level, one faces difficulties in merely evaluating the information embodied in trade statistics. For example, existing theory cannot be brought directly to bear orx a discussion of the following question: given that a particular country is running surpluses on some trade accounts and deficits on others, how can one decide whether the country's external financial relations are fundamentally out-of-balance or not? 1! 307/CMEA/P.Murrell/MS/mcc/03.18.86 -2- 1.3 The present paper constitutes an initial attempt at formulating and solving some of the problems of economic analysis that arise when bilateralism and non-convertibility are important features of trade. The discussion focuses on the case of the CMEA countries. Given the distinctive charac- teristics of those countries, it would not be possible to discuss countertrade simultaneously, without adding undue complexity. However, it should be emphasized that many of the paper's arguments are implicitly relevant to countertrade. That relevance is made explicit in Appendix-A. 1.4 Before turning to the detailed arguments of the paper, it is informative to state some of the main questions that arise when bilateral trade flows exist. These questions are posed in the context of a simple hypothetical exz^mple. In the example, there are two areas - 'West' and 'East'. In East-West trade, convertible currencies are used. Each Eastern country is faced with a single multilateral balance, denominated in $Is, that summarizes all Western transactions. Trade between Eastern countries is denominated in 'transferable rubles' (TR's) and conducted under conditions of strict bilateralism. TR's earned in one country cannot be used in another. Moreover, although an official exchange rate exists between the $ and TR, it is purely an arbitrary accounting rate. Comparisons of monietary aggregates made at that rate do not lead to accurate comparisons of the real magnitudes corresponding to those aggregates. 1.5 In our example, let us assume that the East comprises five countries, labelled A to E. The external financial statement for each country consists of five balances, one with the West and one with each of its neighbors. Consider a possible scenario summarized in the following statistics showing the current account balances of five countries: 307/CMEA/P.Murrell/MS/mcc/03.18.86 -3- Current Account Balances World Total West ($'s) East (TR's) Assuming Official TR A B C D E /$ rate ($'s) A -1000 * 0 0 0 0 -1000 B -500 0 * +250 0 -750 -1000 C 0 0 -250 * -250 -500 -1000 D 0 0 0 +250 * -1250 -iooo E 0 0 +750 +500 +1250 * +2500 1.6 We are primarily interested in analyzing A, B, C, and D (E being a large country able to balance trade easily by using its natural resource base.) The general question to be asked is the following: what can be said about the relative degree of severity of the payments problems faced by each nation? Standard international trade theory brings little to bear on this problem and, indeed, might invite misinterpretation of the data at hand. For example, in the case of A, one migh.t be tempted to ;xamine convertible- currency trade irt isolation, since such trade is Lhe usual focus of standard theory. But the severity of A's problems cannot be judged without taking into account the character of A's trade in the East. For example, even though A's deficit with the West is twice B's, A surely does not have twice B's problems, given the differences in the Eastern trade balances. 1.7 Because of the lack of any alternative in existing theory, one could be tempted to aggregate all of one country's balances usirng tiLe official exchange rate. (This temptation is often enhanced because official publica- tions frequently report all figures in one currenicy.) If the official exchange ratio were one-to-one, one would be forced to conclude that A, B, C, and D face balance-of-payments difficulties of equal gravity. There is no 307/CMEA/P.Murrell/MS/mcc/03.18.86 -4- economic justification, however, for using the official exchange rate. One cannot even aggregate the figures denominated in TR's because, for example, the 250TR that B 1has earned in C cannot be used to pay E. Moreover, a TR used in trade with one country may not have the same real value as a TR used elsewhere. But, possibly, an aggregation could be made at some exchange rate. Given the usefulness of aggregate statistics, one is led to ask the following question: (i) Is it reasonable to combine all bilateral and multilateral balances into one summary balance-of-payments figure, or must one always analyze separately as many balances as there are countries (or areas) among which financial fungiblity is restricted? 1.8 Obviously a summary trade balance would be extremely useful. For example, the creditworthiness analysis of A, B, C)jand D would be aided enormously. But the above example suggests why such aggregation may be difficult. The deficits of A, B, C, and D are with different partner countries. The ease of reducing these deficits must depend upon the characteristics of the different trading partners. Also, A's difficulties are of a different nature than those of B or C because A has only one imbalance to correct. The foregoing discussion suggests the following important questions: (ii) Which procedures are available (at differing levels of complexity) for creating aggregate measures of a country's balance-of-payments situation? (iii) How does the nature of a country's difficulties differ when its trade imbalances are concentrated in one region from the case in which smaller imbalances are present in all regions? Are there theoretical concepts that help in understanding and evaluating the difference? 307/CMEA/P.Murrell/MS/mcc/03.18.86 -5- 1.9 Often policy analysts require information on a country's payments (or debt) situation that goes beyond a mere balance. Such analysts employ a standard set of indicators (e.g., debt-service to export ratios) to measure the severity of a particular imbalance. When there is bilateralism, this set of indicators must be changed. The above example, again, illustrates the problem. Suppose that under a regime of multilateralism an analyst would compare the size of debt-service to the level of exports. Would this analyst compare A's $1,000 imbalance tG A's exports to the West or to some aggregate measure of total exports? How would one create this aggregate export measure? Thus, it is important to address the following question: (iv)How should one measure the total severity of a couintry's trade imbalances (or debt burdens), given that the imbalances can simultaneously occur at different levels in different regions? 1.10 In the discussion above, the questions of interest have been stated as they might occur to an economist interpreting payments data. The posing of those questions, howezer, naturally leads to the examination of one funda- mental issue. The underlying concern over a country's balance-of-payments problems is that, at some time, deficits must be rectified. Policy-oriented economists want to know how significant the ensuing adjustments in the economy will be. The best measure of the size of the adjustments is, of course, the share of the country's real resources that have to be removed from alternative uses to solve the payments problem. Hence, under bilateralism, in order to evaluate fully a country's payments (or debt) situation, one must be able to address the following question: For every nation with which a country trades, what share of a country's real resources would be needed to liquidate a payments imbalance (or debt) of a given size with this nation? An answer to 307/CMEA/P.Murrell/MS/mcc/03.18.86 -6- this question must inevitably lie behind any rigorous analysis of the meaning of balance-of-payments figures under bilateralism. 1.11 When considering the effects of bilateralism, the issues Chat arise go beyond the reinterpretation of payments and debt data. Given that trade is now conducted in several separate regions, the usual lessons that economic theory provides for the formulation of policy might be inappropriate. There- fore, this study also addresses the following question: (v) Which policies must be modified due to the existence of bilateralism and what extra consideratiqcns must now be introduced into policy analysis? 1.12 The above questions could be answered at many different levels of precision. At one level, a country analyst might make an educated guess at the shadow price of a TR in $ terms for a particular country; at the other extreme, one could generate results from a computable general equilibrium model. The investigation in this paper does not center on questions of the appropriate level of analysis. Rather, thte purpose is to identify the important general issues that must be faced at any level of analysis and to make suggestions on how thtese issues should be addressed. 1.13 The investigation proceeds in a number of steps. Section II contains a stylized description of the trade processes of CMEA.countries. The emphasis is on identifying the distinguishing features of CMEA trade behavior and relating those features to the issues raised in the above paragraphs. Section III identifies the errors that can be made when one ignores the special characteristics of bilateral trade in general and CMEA trade in particular. The purpose of that section is to alert policy analysts to the difficulties that bilateralism presents in an examination of trade and balance-of-payments 307/CMEA/P.Murrell/MS/mcc/03.18.86 -7- issues. 1.14 Section IV examines, in a theoretical manner, the ways in which one might construct aggregate statistics on trade and debt. In that section, it is argued that one should construct shadow prices of the TR and the $ in order to undertake such aggregatiGn. Shadow price formulae'are developed and their application is considered in a variety of contexts. In formulating the theoretical analysis, one naturally uses concepts derived from economic theory. Many of these are familiar economic concepts, but some are relevant only in the context of the CMEA - for example, the distinction between hard and soft goods. By conducting the theoretical analysis in a rigorous manner, one naturally identifies the concepts that are important in analyzing the effects of bilateralism. Such identification lays the ground for future research by identifying the areas in which future theoretical and empirical work must be undertaken. Appendix B summarizes some of the lessons of section IV in a manner that is particularly useful for the country analyst. It trans- lates the lessons into a checklist of questions for which one would need specific information, if one is to understand more fully the foreign trade behavior of a CMEA country. 1.15 Section V examines the possibilities of applying the shadow price formulae in a practical setting. That section reviews previous empirical work on the valuation of the TR in $ terms and discusses possible sourcos of data. Suggestions are made concerning different approaches to valuation, conducted at varying levels of sophistication. Section V concludes with an example of the application of an extremely simple method of revaluing TR trade flows. The example shows how much can be gained from investing a minimal amount of effort into the construction of more meaningful statistics. The 307/CMEA/P.Murrell/MS/mcc/03.18.86 -8- application is to Romanian data for the 1980's. It is shown that dramatic differences in conclusions can result when one values intra-CMEA data at approximate shadow exchange rates rather than official rates. II. A Stylized Description of Some Important Features of Trade Within the CN1EA 2.1 Trade within the CMEA is co .:ucted under a set of principles wholly different from those present in the international transactions of market economies. The,member countries of the CMEA accept its arrangements, perhaps grudgingly, and adapt policies accordingly. When conducting policy analysis of the centrally-planned economies, therefore, one must remain alert to the distinctive elements of trade relations within the CMEA. 2.2 The present section examines the distinguishing features of trade within the CMEA. The intention here is not to explain the reasons for the existence of these features. To do so would require an extended discussion of the political economy of the origins and development of the CMEA. 1/ Such discussion would aid little in the present exercise, both because of the limited extent of our knowledge of intra-Eastern European politics and because the basic arrangements affecting trade within the CMEA evince such stability that one can take such arrangements as fixed. Therefore, the focus in the following is not on analysis of cause and effect, but rather on description. 2.3 The ensuing sub-sections describe the important features of trade within the CMEA and explain some of their most immediate implications. Of course, the following description is a simplified image of a complex, poorly understood, reality. What follows serves as the briefest starting point for the analysis, rather than a detailed overview of CMEA institutions. 307/CMEA/P.M>Iurrell/MS/mcc/03.18.86 -9- II(a). Bilateralism 2.4 Trade within the CMEA is conducted on the basis of bilateral agreements between the central economic institutions of member countries. These agreements specify, in great detail, the prices and quantities of goods to be traded (van Brabant 1980, p. 123). The negotiations between each pair of countries take place separately. Every effort is made to ensure bilateral balance on a yearly basis. If an unintended imbalance occurs, the ensuing debt will usually be paid in the following year with an appropriate flow of goods, rather than a financial transaction. 2.5 Since the movement from bilateral balancing to multilateralism can result in welfare gains for all nations, one would expect to see some tendency for multilateralism to arise in the CMEA. This tendency has been observed, but only to a limited extent. Estimates of the proportion of trade conducted outside bilateral agreements vary because of the imprecision of available information, but all seem low. Hewett (1974, p. 15) estimates multilateral trade as 2% of all transactions, Brainard (1980, p. 123) 1.5%, and van Brabant (1977, p. 95) gives a range of 0.4% to 5%. As a good approximation, one can assume that all trade is bilateral. II(b). The Transferable Ruble 2.6 Trade flows between CMEA countries are usually denominated in transferable rubles (TR's). (See II(c) for a discussion of exceptions). The name of this currency reflects past hopes, rather than present realities: TR's are not convertible. They are only issued to the extent that one nation runs a temporary trade deficit with another; they can only be used in the surplus country by the deficit country and only to buy goods that the deficit country 307/CMEA/P.Murrell/MS/mcc/03.04.86 -1 0- specifies)! The official exchange rate of the TR against Western currency is fixed quite arbitrarily. Since this rate is not used as the basis for any -real exchanges, it has no economic content. (The TR/$ cross rates implicit in the exchange rates of local currencies with the TR and the $ vary across countries. For some nations, the TR/$ cross rates do have -8ome real content. Discussion of these rates is reserved for section V(b).) II(c). Trade Denominated in Dollars 2.7 A small share of intra-CMEA trade is conducted using prices set in dollars. It appears that such trade arises because two countries find it mutually advantageous to use prices different fromn those currently prevailing in TR-denominated CMEA trade. Estimates of the significance of such trade vary. Brainard (1980, p. 121) and Marer (1977, p. 546) set it as high as 10% of CMEA trade, while Marrese and Vanous (1983, p. 100) and van Brabant (1977, p. 354) estimate half this value. As this form of trade tends to threaten CMEA pricing rules, there has been pressure in recent years to reduce its significance. 2.8 Three points of clarification must be made about such trade flows. First, this trade is not necessarily conducted on a multilateral basis. A significant share of dollar-denominated intra-CMEA trade is balanced bilate- rally, separately from ruble trade. Second, when such trade is balanced bilaterally, the dollar prices in the transaction might not be equal to world market prices: the dollar appears in such transactions solely as a unit of account. Third, the existence of dollar-denominated intra-CMEA trade does not necessarily imply the possibility of earning a $ surplus that can be used to balance Western trade deficits. 307/CMEA/P.Murrell/MS/mcc/03.04.86 II(d). Prices in the CMEA 2.9 At the simplest level, one can identify two basic elements in the formation of TR prices. First, the CMEA has a set of rules that trade negotiators are supposed to follow. Elements of these rules include the use of "competitive" world prices and the importance of precedent in establishing a particular price. 4/ The price rules, however, are in no sense obligatory, although they do carry some force. The second element in price formation is the striking of bargains between each pair of CMEA countries. As Hewett (1974) has clearly shown, these negotiations produce results that can be at odds with the official pricing formulae. 2.10 Given that the CMEA rules are neither mandatory nor without ambiguity and given that prices are set in many separate bargains, it is not surprising that the level of CMEA prices cannot be predicted using any simple set of hypotheses. Nevertheless, there is agreement in the Western literature on a number of general properties of TR prices. These properties are as follows: 'i) Converted at the official exchange rate of the TR, average CMEA prices lie significantly above wcrld market prices. For the Soviet Union, the calculations of Marrese and Vanous (1983, p. 129) show TR price levels to be 35% higher than world levels on average during 1960-78. 5/ Moreover, within every broad commodity group similar results hold. 6/ (ii)The structure of relative prices within the CMEA is significantly different from that in world markets. As a rough approximation, one can say that the extent to which CMEA relative prices are higher than world relative prices is a function of the degree of domestic manufacturing content in a product (i.e., raw materials have a lower 307/CMEA/P.Murrell/MS/mcc/03.04.86 -12- relative price than manufactures). For example, in Soviet trade with Hungary in 1978, Marrese and Vanous (1983, pp. 129-182) estimate TR fuel prices, converted at the official exchange rate, to be 10% above world market prices, while the corresponding figure for machinery and equipment is 140%. There is no single well-accepted theory explaining CMEA relative prices. Explanations abound, ranging from customs union effects to the political motives of the Soviet Union. (iii)The bargaining over trade flows within the CMEA focuses primarily on quantities and does not seem to be constrained by some pre- existing set of prices. It appears that individual prices are sometimes changed, on a somewhat arbitrary basis, after quantity negotiations are complete, in order to ensure financial balance. Hence, prices-for individual goods may not be reliable indicators of the real trade-offs facing an economy. It should be emphasized7 however, that the foregoing point does not imply that CMEA prices have no informational content. On average, the prices of goods included in the same balance must approxi.nate the rates of exchange between those goods. However, the prices of goods that appear in different balances might be non-comparable. 7/ (iv)Given that trade between each pair of countries is bilaterally balanced and given that the focus in each set of separate bilateral negotiations is on quantities, not prices, it is not surprising that there can be large price disparities within the CMEA. The same good, produced in the same country, can have very different prices when exported to different countries. Brainard (1980, pp. 122-3) quotes a Hungarian estimate that prices can vary by as much as 20%. 307/CMEA/P.Murrell/MS/mcc/03.04.86 -13- Therefore, a transferable ruble may have different values when used in trade with different partners. In fact, the methodology implicit in Soviet foreign trade planning recognizes this point (Brainard (1976, p. 702)). The estimates of Marrese and Vanous (1983, p. 104) show, for SoViet trade, that average prices normally vary acrois trading partners by as much as 10%. (v) In trade relations, as in many areas of activity in centrally-planned economies, there are forces which lead to rigidities. The difficulty of changing complex bilateral trade agreements, the use of precedent to establish prices, and the excessive focus on past results in preparing new plans 8 lead to a price structure that is more rigid than in market economies. In fact, the official CMEA pricing formula, adopted in 1975, codifies this rigidity by employing a five- year moving average formula for calculating prices. One important result of the price rigidity is that rapid changes can occur in the real. value of a transferable ruble relative to the dollar. For example, a TR used by Hungary to buy goods from the Soviet Union increased In value by 67% relative to the dollar in 1974 (Marrese and Vanous 1983, p. 182). IIte). WSoft" Goods and "Hard" Goods 2.11 In the analysis of intra-CMEA trade, an important distinction,is always made between two types of goods - hard and soft. These terms are variously defined: soft goods are ones in excess supply in the CMEA region, given present pricing rules, and hard goods are in short supply; or hard goods are ones that can sell for convertible currencies; alternatively, soft goods 307/CMEA/P.Murrell/MS/mcc/03.04.86 are ones that do not meet Western quality and technological standards. In practice, goods are more likely to be soft the higher is their local manufac- turing content. Since economists use the terms hard and soft in various ways, the distinction between the terms might be too vague to define precisely. (Some possible definitions of "hard" and "soft" are discussed in section IV(e).) Nevertheless, the distinction does correspond to a very important reality on CMEA markets. 2.12 Regardless of the definition of the difference between hard and soft goods, one very important conclusion follows. Given the present structure of the CMEA and the resultant price rules, countries (with one very important exception - the USSR) are unwilling to trade-off a TR trade surplus in hard goods against a TR trade deficit in soft goods. Therefore, the smaller CMEA countries ensure that hard good trade and soft good trade balance separately (Hewett 1974, p. 113). 2/ Given separate balances for different types of goods, there is.no reason why a TR used in trading soft goods will have the same value as a TR used for hard goods. As a result of the existence of bilateralism and the distinction between hard and soft goods, CMEA countries face many balances of payments. There is a separate hard and soft good balance with each CMEA partner, as well as the normal multilateral convertible currency balance. II(f). The Special Case of Trade with the USSR 2.13 In trade between the small Eastern European countries, the separate balancing of hard and soft goods is viable, if not effidient, because relative resource endowments are sufficiently similar between countries. If the Soviet Union insisted on balancing hard good flows, little trade would result, given 307/CMEA/P.Murrell/MS/mcc/03.04.86 -15- the Soviet Union's obvious comparative advantage in raw materials. In order to maintain active trade relations with Eastern Europe, the Soviet Union accepts soft goods in exchange for hard ones. (In practice, this means manufactures for raw materials.) This fact is of the utmost importance because of the sheer size of Soviet trade flows. 2.14 Two important conclusions flow immediately from the special nature of Soviet trade. First, CMEA countries obtain terms of trade from the Soviet Union that are more beneficial than those with the rest of the world. 1"/ Given that the relative prices of soft goods are higher in the CMEA than on world markets, redirection of trade away from or towards the USSR has important consequences for a country's aggregate terms-of-trade. Second, in order to stop arbitrage by the smaller CMEA countries, the USSR must place constraints on the extent to which hard goods trade for soft. Hence, average terms of trade will be very different from marginal ones, which are a function of the constraints, The value of the TR will therefore depend on the exact nature of the restrictions on trade flows placed by the Soviet Union. 1I(g). Deticits 2.15 CMEA rules ensure that interest rates are low on debts incurred as a result of a trade imbalance. Not surprisingly therefore, extensive efforts are made to ensure that trade balances on a yearly basis. There is clear evidence that any imbalances that do arise are reversed in the following year (Vanous 1980, p. 190). Again, however, there is an important exception. When rapid changes in the terms of trade favored the Soviet Union in recent yearvs, some smaller CMEA countries were allowed to incur deficits in order to dampen the speed of the resultant adjustments. Loans to cover such deficits may not 307/CMEA/P.Murriell/MS/mcc/03.04.86 -16- necessarily have the same meaning as loans from the West because interest rates could be low, because such loans have been cancelled in the past (Marrese and Vanous (1983), p. 111), and because the true burden of debt depends upon whether payment must be made in hard or soft goods. III. On Errors Frequently Committed When Analyzing CMEA Trade 3.1 Both the complexity of procedures within the CMEA and the difficulty of obtaining the relevant statistical information make CMEA trade relations opaque to our understanding. In addition, economists steeped in the analysis of free markets find the behavior of centrally-planned economies resistant to their standard modes of thought. It is therefore unsurprising to find that there exists a tendency for economists to downplay, or to forget,.the distinguishkng features of CMEA trade when analyzing Eastern European economies. 3.2 The present section identifies some of the errors of analysis that might arise when due attention is not paid to the special nature of CMEA trade relations. Where possible, examples are cited and information that bears on judging the significance of the errors is given. The purpose here is not to criticize the works cited, but rather to give some practical content to the discussion and provide examples of places in which future analysis can be improved. 3.3 The following discussion should not be taken to imply that all problems identified can be corrected. Given the poor quality of statistical informa- tion on CMEA trade, the data might not exist to make the necessary amend- ments. However, even if the appropriate corrections cannot be made, identi- fication of possible sources of error is useful. Knowing the direction of 307/CMEA/P.Murrell/MS/mcc/03.04.8-6 -17- bias of an analysis is useful in judging the quality of policy conclusions derived therefrom. Thus, while one may not be able to conduct all analysis in the optimal manner, one may be able to improve the judgements that arise from existing analyses. 3.4 When ignoring the nature of CMEA trade, errors of judgement can arise in two ways. First, statistical information can be derived in a manner that makes value aggregates poor representations of real phenomena. Then, the data give misleading information on nast economic performance and on future possi- bilities. Second, partly as a result of the first, inappropriate policy advice might be formulated. The CMEA functions under a very different set of arrangements than do free markets. Policy analysis, therefore, cannot be premised on standard economic principles, without deeper consideration of their applicability in specific situations. Both using inadequate data and neglecting institutional peculiarities can lead to misjudgements concerning the nature of the possibilities open to an economy. In the examples that follow, both types of misjudgment's are considered. ILI(a). Using Transferable Ruble/Dollar Comparisons Made at the Official Exchange Rate 3.5 As noted above, the official exchange rate of the transferable ruble (TR) is arbitrary: it does not correspond to any real magnitude. Therefore, one should not use such a rate when converting CMEA trade flows and non- convertible currency debt into dollars. The resultant aggregates can be highly misleading. Hungary's foreign trade statistics show this in a dramatic way. 3.6 The World Bank Country Study, Hungary: Economic Developments and Reforms, (1984, p. 20) states that the share of Hungary's trade with socialist 307/CMEA/P.Murrell/MS/mcc/03.04.86 -18- countries declined from 64% in 1970 to 55% in 1981. This statement has obvious significance in any evaluation of Hungary's reforms. No information is given, however, on how TR and convertible-currency prices are converted into common units in order to make the statement. Given the arbitrary nature of some TR exchange rates, further analysis is necessary in order to determine whether the apparent redirection of trade is anything more than a statistical artifact. 12/ 3.7 Consider the following information from the Country Study (World Bank 1984, P. 174-5). Percentage Share of Socialist Countries in Hungarian Trade 1970 1975 1976 1980 Exports 63.3 70,0 57.8 52.1 Imports 62.7 64.5 52.3 48.2 Apparently a geographical reorientation in trade occurred in one year, 1975 to 1976. How could this have happened, especially at a time when CMEA fuel prices were rising very rapidly relative to world prices? 3.8 Although a conclusive answer to the above question would require more exact knowledge of the construction of Hungarian foreign-trade statistics, one can hazard an informed guess. In 1976, Hungary changed the rate used to convert TR trade and dollar trade into comparable units. The change, to a rate closer to a realistic value, 13/ implicitly devalued the TR by 45% relative to the dollar (Vanous 1981, p. 712). Recalculating the 1975 shares using the 1976 TR/$ rate, one obtains 56.5% for exports to socialist countries and 50% for imports. The implicit devaluation of the TR seems a more 307/CMEA/P.Murrell/MS/mcc/03.04.86 -1 9- plausible explanation for taie apparent change in the direction of trade than any argument that would focus on changes in real flows. 14/ 3.9 Three conclusions follow directly from the above discussion. First, when arbitrary exchange rases are used in constructing aggregate information, arbitrary results can follow. The decline ir the socialist countries' share of Hungary's trade may have been more apparent than real. Second, in order to interpret data on total trade flows, the researcher must know which exchange rate has been applied to trade within the CMEA. The third conclusion follows from noting that the exchange rate used in later years was more realistic than the official TR/$ rate used before 1976. As a consequence, the estimated share of socialist countries in Hungarian trade "dropped" by nearly one- fifth. The use of official TR/$ rates, therefore, leads to an overestimate of the importance of CMEA trade in a country's affairs. III(b). Ignoring Ruble Trade 3.10 Some researchers, sensitive to the differences between ruble and convertible-currency trade, adopt a tactic diametrically opposite to the one discussed above. Ruble trade is igncured when analyzing East-West inter- actions. Since problems of comparison are regarded as large, they are implicitly deemed insurmountable. Ignoring ruble trade, however, leads to two problems: the present economic picture is inadequately summarized and possible future trade-offs are masked. 3.11 The ignoring of ruble trade arises most notably in analyzing the burden of debt. Because dollar debt must be repaid in dollars, analysts are prone to compare the size of repayments to convertible-currency exports alone. Zoeter (1981, p. 730), for example, calculates two measures - debt and 307/CMEA/P .Murrell/MS/mcc/03 .04 .86 -20- debt-service as a proportion of hard-currency exports. In 1979, those figures show Romania with much less of a debt problem than the GDR. With hindsight, one knows that these figures did not faithfully represent the relative burden of debt on the two countries. But as the following table shows, hindsight was not the only way to gain a more realistic impression of the debt burden. 16/ GDR Romania hard currency debt/hard currency exports 2.39 1.31 hard currency debt/total exports 0.74 0.81 hard currency debt service/hard currency exports 0.55 0.24 hard currency debt sevice/total exports 0.17 0.15 The simple reason why one obtains a distorted picture when ruble trade is ignored is that Romania's trade is more oriented to the West than the GDR's. Using total exports in the calculations, one obtains .a different impression of the adjustment burdens on the economies that the payment of debt was to impose. (Ruble debt can be ignored in the present example because it was insignificant in 1979.) 3.12 Perhaps, at the present stage, it is necessary to counter one possible objection to the above argument. It may be claimed that trade with the CMEA and trade with the West are not substitutes. Therefore, adjustments in CMEA trade cannot help to ameliorate the burden of paying $ debts. Such an objection probably overstates the degree of rigidity of the Eastern European economies. It also understates the extent to which imports from both regions are substitutes in consumption, thus, allowing the maintenance of the level of imports from one region to reduce the relative burden of a decline in imports 307/CMEA/P.Murrell/MS/mce/03.04.86 -21- from another. III(c). Using the Same Exchange Rate in Transactions With All CMEA Countries 3.13 CMEA countries bilaterally balance trade with each CMEA partner. Because arbitrage is not possible between separate balances, the absolute level of TR-denominated prices can vary between the separate bilateral balances of a single country. Since there is no possibility of trade-off between the seperate imbalances, no errors or inefficiencies are implied by such price variations (apart from those already consequent on bilateralism itself). However, errors and inefficiencies can arise if such differences either become embodied in statistical summaries of a nation's performance or affect the structure of domestic prices. For example, price signals may convey incorrect messages about the goods in which a country has a comparative advantage. 3.14 Using a single TR exchange rate in constructing trade statistics leads to problems similar to those discussed in section III(a), albeit on a smaller scale. For example, a geographical reorientation in trade within the CMEA may appear as a change in the volume of trade . Thus, the economy's real performance is masked. Any analysis of future policy directions based on such statistics would be likely to produce incorrect conclusions. The magnitude of possible errors is indicated by the finding of Marrese and Vanous (1983, p. 104) that aggregate Soviet import prices had an average spread of 10% across CMEA countries in the 1970's. IlI(d). UTsing the Same Exchange Rate for All Types of Goods 3.15 Let us suppose that we have found a set of country-specific TR/$ 307/CMEA/P.Murrell/MS/mcc/03.04.86 -22- exchange rates that make average CMEA trade pricPis equal to average world market prices. However, manufactured goods are relatively overpriced in CMEA trade, compared to world markets. If world market prices are presumed to reflect real values, the use of the average TR/$ rate will lead to an over- statement of the share of manufactured goods in trade. As a consequence, spurious statistical results can arise. 3.16 Consider, for example, the Romanian machinery sector. According to official Romanian statistics, 17/ machinery exports to socialist countries rose by $680 million from 1982 to 1984, while those to non-socialist countries fell by $450 million. Aggregating these figures indicates that the machinery sector increased its exports. However, data from Marrese and Vanous (1983, pp. 58-59) show CMEA machinery prices to be more than double world prices when the official exchange rate is used and 50% higher when using an exchange rate that equates average TR prices to dollar ones. Revaluing Romania's exports appropriately, one would conclude that the machinery sector failed to improve its export performance. (See the results in Section V.) 3.17 The misinterpretation of data, pointed ouAt above, occurs because of the erroneous assumption that the same exchange rate must be applied in all transactions conducted in a single currency. This assumption can lead also to inappropriate policy advice, as the following shows. III(e). Indiscriminate Application of the Law of One Price 3.18 The most basic policy rules that economists employ are invariably derived in analyses that assume free markets. In such analyses, prices and quantities freely adjust to remove disequilibria. The CMEA, however, does not have such adjustment mechanisms. Therefore, indiscriminate application of the 307/CMEA/P.Murrell/MS/mcc/03.04.86 -23- usual policy rules must be questioned. 3.19 One such-rule involves the use of only one price for each economic entity. For example, an analyst might advise using a single exchange rate for the TR. Implicitly such advice is proffered when economists comment appro- vingly on the removal -of multiple exchange rate systems. The same premise probably underlies negative remarks on the existence of schemes designed to tax or subsidize specific participants in CMEA trade. 18/ 3.20 As has been made clear above, relative prices in the CMEA are not market clearing ones. There are quantity constraints on trade, for example. The transfer of TR's between countries is not possible. Therefore, the law of one price should not be applied to the TR because, in effect, a TR used in trade with a different country is a different economic entity. Furthermore, if hard goods trade and soft goods trade are balanced separately, a TR- used to buy hard goods has a different value than one used for soft. 3.21 One can conclude, therefore, that efficient economic policy in an Eastern European economy may require either multiple exchange rates or good- specific taxes or subsidies. In pointing out such a conclusion, the intention is not to minimize the extent to which a bureaucratic system of subsidies and taxes creates incentives to expend efforts on lobbying rather than on produc- tion. But such inefficiencies may be a necessary, and worthwhile, cost of providing appropriate incentives in export and import decisions. III(f). Assuming an Accounting Dollar is a Real Dollar 3.22 Some CMEA trade flows are denominated in dollars. A significant share of this dollar-denominated trade is bilaterally balanced. Therefore, the absolute levels of the dollar prices might have no significance: no 307/CMEA/P.Murrell/MS/mcc/03.04.86 -24- equilibriating forces produce correspondence between these prices and the ones on world markets. The real meaning of CMEA dollar prices is, therefore, unclear. 3.23 When analyzing the foreign trade position of Eastern European countries, careful attention is often paid to the distinction between "socialist" trade and "non-convertible-currency" trade (e.g., World Bank (1984, p. 19)). But this distinction may not carry as much force as is often thought. One cannot presume that CMEA trade denominated in convertible currencies is the same as multilateral trade. Whether this trade has characteristics closer to TR trade or to multilateral trade depends upon the extent to which the dollar-denominated flows between pairs of countries are balanced. Unfortunately, evidence on this point is not easily available. 3.24 Two basic differences Qccur between bilaterally-balanced dollar- denominated CMEA trade and multilateral trade. First, absolute prices in the former need not conform to those in the latter, due to the absence of equilib- rating forces. Care must therefore be taken in aggregating the two sets of trade flows. Second, dollar-detnominated CMEA trade usually cannot be used to settle imbalances in other regions. The existence of dollar trade flows within the CMEA does not necessarily imply an ability to generate dollar surpluses in that region. III(g). Forgetting the Special Character of Commercial Relations with the Soviet Union 3.25 The Soviet Union is a partner in over 50% of CMEA trade transactions (Vanous 1981), which means that Eastern European countries conduct, on average, more than one-quarter of their trade with the USSR. In addition, virtually all non-convertible currency debt is owed to the Soviet Union. The 307/CMEA/P.Murrell/MS/mcc/03.04.86 -25- sheer size of Soviet trade alone would argue for special attention. More important, however, is the fact that the commercial relations of the USSR are conducted in a manner different from those of the rest of the CMEA. 12- To ignore the special character of Soviet trade is to miss a crucial element in the analysis of the Eastern European economies. 3.26 Let us take, for example, the case of non-convertible currency debt. One would like to be able to value this debt in dollars (or domestic currency) in order to obtain some indication of its burden on the economy. But that burden is a function of the manner in which the debt must be repaid. Is the debt a claim held by the Soviet Union on hard goods or on soft ones? The answer to this question is an important determinaut of the way in which debt repayments will affect an economy. 3.27 The character of Soviet trade behavior and, in particular, the constraints that the Soviet Union places on its sale of hard goods, must affect the value to be placed on a TR in trade with that country. Systematic examination of the nature of these constraints is a precursor to any careful evaluation of the meaning of statistics on trade flows and debt. In section IV(f), a preliminary attempt at such an examination is given. III(h). Assuming that a Country's Terms of Trade are Indeperient of Trade Policy 3.28 In relations between market economies, trade policy focuses on domestic institutions and on the way in which domestic incentives should be structured to produce efficient results. For most countries, given the absence of market power, little is to be gained from considering ways in which the external environment can be manipulated to improve the terms of trade. The character of CMEA institutions, however, is such that a country's terms- 307/CMEA/P.Murrell/MS/mcc/03.04.86 -26- of-trade can be changed by domestic policies. 3.29 The prices offered by the Soviet Union to its CMEA partners are usually more favorable than those existing in trade with market economies. Thus, changes in a country's aggregate terms-of-trade may result solely from a regional shift in trade. In evaluating a country's performance, therefore, attempts should be made to consider the causes of changes in aggregate trade prices. One must separate truly exogenous changes from those caused by regional shifts in trade. Only by doing so, can one accurately evaluate the effect of international economic events on the domestic economy. 3.30 Consider also a policy question: how to affect the classification of goods into hard and soft categories in trade negotiations. Obtaining a classification of hard for a previously soft good may lead to large benefits. If a good is classified as hard when it can sell on Western markets, a country might find it profitable to subsidize Western exports on a good-by-good basis. Such a subsidy would increase the number of goods that can be sold in the CMEA as 'hard' and, therefore, improve the country's terms of trade in the ruble area while worsening $ terms of trade. (See Section IV.) One need not emphasize that such a conclusion runs counter to received doctrine on trade policy. IV. On the Measurement of CMEA Trade Flows and Non-Convertible-Currency Debts IV(a) The Shadow Price Approach 4.1 The focus of the present and following sections is on how to generate the information that can best summarize the foreign economic relations of CMEA countries. Of course, it is always possible to suggest new items of data that should be collected in order to improve accuracy. But that is not the 307/CMEA/P.Murrell/MS/mcc/03.04.86 -27- approach taken here. Rather, the amount of data available is taken as given, essentially dictated by the domestic authorities. The present examination centers on the ways in which one can improve the quality of information derived from that given set of data. 4.2 The problem at hand can be simply described. One has two sets of data on both trade flows and debts. 20/ One set is generated by bilateral agreements and denominated in transferable rubles, the other arises within a multilateral payments system and is given in dollars. 21/ Since there is no market in TR's, no equilibrium exchange rate exists at which one can compare the two sets. The natural approach is to use economic theory to derive the analog of such a rate and then to examine whether one can estimate that analog using currently available data. The theory is developed in the present section; the review of existing data sources is given in section V. 4.3 Before proceeding to the analysis, it is useful to justify why converting the two data sets into common units is both informative and possible. When there are choices to be made between alternative paths of action, it is informative to express all variables in the same units in order to know the trade-offs relevant to the choices. Consider an example. When similar goods are traded in both the CMEA and the West, debts can be paid by readjusting exports and imports in a variety of ways in both regions. Two simple scenarios for paying a $1 debt are as follows: reduce domestic consumption of goods that can be sold in the West and, consequently, increase $ exports; or, reduce domestic consumption of goods that can be sold in the CMEA in order to increase CMEA exports to pay for additional imports of, say, energy from the CMEA, consequently reducing energy imports bought with $'s. in order to assess these scenarios, one must be able to compare the domestic 307/CMEA/P.Murrell/MS/mcc/03.04.86 -28- opportunity costs of obtaining a unit of currency in each region and to compare the relati - amounts of energy that each unit of currency can buy. TR/$ conversion ratios are implicit in each of these comparisons. 4.4 In summary, one translates TR and dollar values into the same units for exactly the same reason as one converts other economic magnitudes (e.g., production or consumption figures) into common measures. Only by doing so, can one take stock of the economy's present situatiorn and evaluate the future trade-offs facing the economy. 4.5 Now let us turn to the question of whether it is possible to compare TR and $ figures in a meaningful way. The answer is based upon the observa- tion that trade is simply a means to enhance economic welfare, not an end in itself. Thus, one can view trade in the same way that one views any produc- tive activity. Expressing TR and dollar flows in common units is exactly equivalent to placing monetary valuations on goods produced by two different industries. One values goods by considering the relative contributions to economic welfare that can be produced by incremental units of each good. 22/ Si:ace currencies are not themselves objects of welfare, the value of a currency is the contribution to economic welfare that is derived from the goods that can be obtained with an incremental unit of the currency. This approach to valuing a currency is simply the shadow price technique. 4.6 Much of the economics literature on the calculation of shadow prices centers on the difficult methodological issues of valuation in the presence of economic distortions. Given the available information on the socialist econo- mies, the present discussion must have a different purpose. 23/ Here, the focus is on the most basic issues, gaining insights from simple economic models that embody elements of the stylized description of CMEA trade deve- 307/CMEA/P.Murrell/MS/mcc/03.04.86 -29- loped in Section II. The goal is not to show how to apply the most advanced techniques of shadow price calculation. Rather, the intention is to suggest ways in which comparisons of TR and dollar values can be constructed in a manner that significantly improves on present practice. 4.7 The ensuing discussion has a broader focus than solely the presenta- tion of shadow price formulae that can be used to reinterpret existing data. The shadow price approach is not only helpful as a means of valuation when data is available, it also provides a framework for organizing a more general analytical discussion. This approach helps to clarify which variables, para- meters, and relationships should be relevant when one is trying to understand a particular economic phenomenon. Therefore, as a natural consequence of the following analysis, conclusions are derived that are of wider relevance than simply providing a means of valuing trade flows. In oarticular, the discus- sion provides information in the following areas: (i) The inaccuracies of analysis, loosely described in section III, are systematically examined. The source of such inaccuracies is rigorously derived in order to better understand the problems that occur in present policy analysis. (ii) Theoretical constructs that are useful in examining CMEA trade flows are identified. The purpose in such identification is to become aware of the areas in which theoretical developments are required, if policy analysis of socialist economies is to be improved. (iii) Suggestions are made concerning the most important areas in which better data collection could improve the statistical analysis of CMEA trade flows. 307/CMEA/P.Murrell/MS/mcc/03.04.86 -30- IV(b). The Model and Notation 4.8 The details of the model underlying the present discussioi- are presented in Appendix C, together with the derivations of the main results. Here, the main features of the model are surnmarized and the results are presented in an heuristic fashion. In most cases, readers will find satisfactory justification of the results in intuitive arguments: often, the model is solely a formalization of common-sense notions. 4.9 The basic features of the model are as follows: (i) Domestic distortions are not taken into account. To do so would require modelling the internal behavior of centrally-planned economies. Given the lack of a well-articulated theory of these economies, such modelling is beyond the scope of the present exercise. Thus, it is assumed, for example, that domestic autho- rities freely choose internal prices: no pre-existing set of tariffs determines the relationship between foreign and domestic prices. (ii) It is assumed that trade and production patterns are consciously chosen in order to maximize economic welfare. (iii) Within the CMEA, trade with each single partner country is distinguished. In such a way, bilateral balancing is introduced into the model. (iv) Trade conducted on a multilateral basis is treated as trade with a single entity - the rest-of-the-world. (v) For CMEA countries other than the Soviet Union, it is assumed that hard goods trade and soft goods trade are balanced separately. (vi) Explicit attention is paid to the ways in which one might model tne 307/CMEA/P.Murrell/MS/mcc/03.04.86 -31- constraints on trade imposed by the Soviet Union. The effect of the constraints on the valuation of trade flows is examined. (vii) International trade prices are assumed to be fixed. Bargaining and market power are absent from the model. Thus, the results are relevant for a small Eastern European country. (viii) Solely to simplify the presentation, results on bilaterally- balanced dollar-denominated intra-CMEA trade are not derived within the model. However, one can easily derive conclusions about the valuation of such trade by extending the model's results in a straightforward manner. (The discussion of III(f) shows that the $'s used in bilaterally-balanced intra-CMEA trade must be treated as a third currency: one cannot assume that the prices used in such trade are simply related to either TR prices or to world-market prices. Then, all the results on the valuation of the TR would apply, mutatis mutandis, to the valuation of this third currency.) 4.10 Unfortunately, the notation that must be used to define the shadow price formulae is somewhat cumbersome. This is due to the necessity of taking into account a variety of different types of trade flows within one model, rather than from any intrinsic complexity in the analysis itself. In fact, in most cases, the shadow exchange rates can be defined in simple intuitive ways. 4.11 Each price or quantity variable has three subscripts, The first, whose general value is denoted by a "g", indicates the type of good. It is crucial to the analysis to distinguish three types of goods - thcse traded in convertible-currency markets (subscript w), those sold in the CMEA as hard goods (h), and those sold in the CMEA as soft goods (s). The second subscript, whose general value is a j, indicates the partner country in a 307/CMEA/P.Murrell/MS/mcc/03.04.86 -32- trade. When the trade is conducted on a multilateral basis, the identity of the partner country is of no consequence and the second subscript is a The third subscript identifies the traded good and has the general value i. To economize on notation, i i., always used for the goods index. However, it should be borne in mind that when two variables have the same third subscript, but different first subscripts (w, h, or s), different goods are indicated. 4.12 To simplify, the notation does not include an element for the "home" country that is being examined within the analysis. Thus, in the following, all variables are assumed to be defined with respect to one country - a typical small Eastern European country. The basic variables defining trade flows are: Pgji - foreign currency price of the ith good of type g exported to ^ountry j Xgji - quantity of the ith good of type g exported to country j Rgji - domestic price of the ith good of type g exported to country j Qgji - foreign currency price of the ith good of type f imported from country j Mgji - quantity of the ith good of type g imported from country j Sgji - domestic price of the ith good of type g imported from country j The subscripts have the following ranges: g = w, s, or h j = $, Bulgaria, Czechoslovakia, ..., USSR. where n gj is the number of different of goods of type g traded with country j. 4.13 In order to present the basic results in a concise manner, it is helpful to define the following indices: 307/CMEA/P.Murrell/MS/mcc/03.04.86 -33. E.= Z.P ..X(1 gj 1 gi. gil Z.R ..X .. 1 gj 1 gJi = foreign currency value of exports of type g to country j domestic value of exports of type g to country j M EM. = EiQgj. M J. (2) gj 1gigj 1 Z .M 1 gj i gj i - foreign currency value of imports of type g from country domestic value of import, of type g from country j Given that the convertible-currency area is defined as one "country" and that there is no distinction between hard and soft goods in that area, only two of the above indices contain $ figures: Ex and EM There is a multitude of indices containing TR figures, however: for each country j, one has EX E>it, EMI, and EM.. IV(c) Valuing Trade Flows 4.14 The first results focus on valuation when hard good and soft good trade are balanced separately and when there are no constraints on the level of trade. These results are most applicable to trade between small CMEA countries. The following establishes the irrelevance of the official TR/$ exchange rate. PROPOSITION 1: The shadow price of the TR is independent of the official TR/$ exchange rate. The source of this result is obvious. No transactions are undertaken at the official exchange rate and, therefore, that rate cannot influence real behavior. Given Proposition 1, one should be wary of any statistics constructed using official currency valuations. 24 307/CMEA/P.Murrell/MS/mcc/O3.18.86 -34- 4.15 The nex'. result establishes the basic formulae for calculation of shadow prices. (Questions of data availability are not examined here. They are reserved for Section V.) PROPOSITION 2: When domestic distortions can be ignored, trade of type g (= h or s) with country j should be converted into $Is using the following conversion factors: C. E= E orC.=E /E gj w$ gj gj W$ gj The intuition behind the proposition is simp]e. If domestic prices reflect social valuations, then the denominator of each "C" is the inverse of the social value of one TR. The numerator reflects the social value of one $ in a similar manner. The ratio, then, reflects the relative social value of the $ versus the TR. 4.16 In interpreting this proposition, one should bear in mind the following: x M ersgfi (i) If there are no distortions, Cgj = Cgj* If these differ signift- cantly, one would need to investigate the reasons for the difference before deciding which to use. (ii) In calculating the E's, it is not necessary to include all trade flows. (That is, the summations in formulae (1) and (2) need not vary over all i.) The index can be calculated using a representa- tive sample of goods. (iii) The previous point implies that if one can find a set of goods that are traded in both the CMEA and the West, one can simplify the calculations. Cgj and CMg would each become simply the ratio of the $ value of a market basket of this set of goods divided by the TR value of the same market basket. This is the approach followed 307/'CMEA/P.Murrell/MS/mcc/03.18.86 -35- by Marrese and Vanous (1983,1984) who value Soviet trade flows (i.e., the market basket) in both TR's and $'s. The problem with this approach is finding a set of goods that is traded in both regions. 25/ Marrese and Vanous (1983) finesse this problem by estimating quality differences between goods traded in the West and those sold within the CMEA. They then construct quality-adjusted prices and assume that such prices can account for all differences in the social values of goods traded in the two regions. Their 1984 study improves upon this methodology by using Polish and Hungarian trade data for a limited number of commodities that are traded in both the East and the West, to estimate the numerators of the E's. 4.17 The following are corollaries of' the previous proposition: PROPOSITION 3: When hard-good and soft-good trade are balanced separately, there is no necessary conne-ction between the shadow price of a TR used in hard goods and cre used in soft goods (i.e., X EXan M M in general, Ehj * Es; and Esj * Ehj PROPOSITION 4: When trade is bilaterally balanced, there is no necessary connection between the shadow prices of TR's used in trade with different countries (i.e., in general, .EX E x x EM M we E E E , Es k E and E,. X g whenj * k). Si sky si sk hj hk nj hk' These two propositions simply formalize the discussion of section III. They show that trade flows occurring in separate balances must be converted into $Is at different rates. 4.18 It is fitting to comment on the relationship of the above formulae to those existing in the economios literature. All that is new in the above 307/CMEA/P.Murrell/MS/mcc/03.18.86 -36- is the adaptation of existing results to the special circumstances of bilateralism and the separate balancing of different types of goods. The above results are much simpler, and more clear cut, than those in the litera- ture because no domestic distortions are taken into account. It is only in the presence of distortions that the various shadow price formulae existing in the lit,rature significantly differ. Thus, the present approach is consistent with most existing derivations of foreign exchange shadow prices. 26/ Significant in the present context is the consistency with Kornai's (1967, p. 306) methodology. 4.19 Propositions 3 and 4 lead to an important conclusioni relatling to the commercial policy of CMEA countries. These countries face an eriormously complex external environment. For example, there will be four implicit exchange rates in trade with eahn of the CMEA trading partners. Many seperate trade balances must be supervised simultaneously. To manage such a system is in itself an extremely complex administrative task. When advi5ing on commer- cial policy, analysts must remain sensitive to fact that their usual advice might be very difficult to implement in such a complex environment. Consider, for example, market research for export promotion. In a multilateral environ- ment, the exporter need only examine the foreign demand for the export. When such research is directed at a trading partner who insists on bilateralism, the demand for exports cannot be estimated independently of the trading partne-'s ability to produce goods suitable for exchange. Study of the market for exports becomes inseparable from analysis of the ability to benefit from imports, if gains fom trade are to be generated. 307/CMEA/P.Murrell/MS/mcc/03.04.86 IV(d). Valuing Debt 4.20 A debt constitutes a claim on a country's resources. Given the normal equilibration processes present in international trade, one does not usually need to specify which goods are subject to the claim. At the margin, any mode of debt payment will affect welfare in the same way. However, the real value of one TR of goods depends upon the goods in question. Thus, the real burden of a TR debt is unknown, unless further information is available. 4.21 The crux of the problem is that soft goods are relatively overvalued in trade: 27/ x x m M si chip sj S 0, with ai > 0 if the country would like to import more of hard good i. Thus -ei + Shui Ahu Qhui (A.13) and 1i mhuiai +i hui shui =hu Ei Qhui hui (A.14) Thus, if one or more constraints are effective 1 zi QhuiMhui M Ahu Ei MhuiShui hu Now, since A hu/Aw$ is the dollar value of one TR's worth of imports from the Soviet Union, the use of C ( = EM /E M > Ah/A ) will lead to an 'I'lu w$ hu huw overvaluation of the TR. Proposition 7 is proved. Proposition 8 is proved in an analogous manner. When the conditions assumed for Proposition 9 apply, the constraints (A.3) are no longer applicable in trade relations with the Soviet Union. They are replaced by an overall balance-of-trade constraint and two conditions on the structure of trade: 307/CMEA/P.Murrell/MS/mcc/03.04.86 -76- P X + P X - Q M ' Q M =0 (A.15) hu. hu. SU. SU. hu. hu. SU. CU. p X ~b(P x + p x )(A.16) hu. hu. hu. hu. su. su. I Qsu. sU. (Qsu. su. Qhu. hu.) (A.17) (In this formulation, there is no difference between the marginal proportion of hard goods in exports and the average proportion. Such a difference could be trivially introduced by adding a positive constant to the right-hand-side of (A.16). For soft goods, the same change could be made to (A.17). Such changes do not affect the results, because shadow prices depend only on marginal values. Using the Kuhn-Tucker theorem with Aa, Ax and Am as the Lagrange multipliers of (A.15), (A.16) and (A.17) respectively, one can derive the following necessary conditions: Du + A P + A P (1-b) 0 (A.18) hui a hui x hui Mhui Aa Qhui -me Qhui=° (h.l9) ax + . - AXb P .=0 (A.20) xa sui x Sul SUl A Q + A (1-e)Q . 0 (A.21) 3M a sui m sui Proposition 9 follows upon substituting the domestic prices for marginal social values, solving for Aa$ and realizing that Aa is the marginal social value of a TR. 307/CMEA/P .Murrell/MS/mcc/03. 04.86 -77- FOOTNOTES 1. Note that surplu:,x3 and deficits can arise under bilateralism when agreements ct-ver time periods not coterminous with the reporting periods of trade statistics. For example, deficits currently run by the Soviet Union will be settled over a number of years. 2. See van Brabant (1980) and Marrese and Vanous (1983) for differing views. 3.. See van Brabant (1977) for details. 4. See Hewett (1974) for a detailed summary of CMEA pricing rules and an analysis of their consequences. 5. For reasons given below, the average level of Soviet trade prices will be lower than those for the CMEA as a whole. 6. Marrese and Vanous (1983 p. 122) calcullate comparisons of dollar prices versus ruble prices for 5 commodity groups over a 19-year period. They find ruble prices higher in only 7% of the comparisons. 7. These points will be -ade in greater detail in SectUon IV. 8. For a detailed discussion of planning "from the achieved level"' see Birman (1978). 9. In fact, reality is somewhat more complicated than this stylized description. Trhe degree of softness can vary. As a result, it seems that trade negotiations often result in many separate balances. 10. Marrese and Vanous ('983) refer to the Soviet Union as giving Eastern Europe an "implicit trade subsidy". 11. Given CMEA prices, the re-exporting of Soviet raw materials to the West would be immensely profitable for Eastern Eurcpean countries. 12. To see how such an artifact can easily arise, consider, for example, the following scenario. No inflation in TR prices, dollar price inflation and the arbitrary exchange rate remains fixed. The share of socialist trade would fall. 13. See the discussion in Sections IV and V. 14. The conclusion cannot be made definitive here. To do so would require direct information on how the tables in the Country Study were consructed. 307/CMEA/P.Murrell/MS/mcc/03.04.86 -78- 15. Zoeter (1981) is used solely as an example. Marer (1981 p. 67), Matusek (1981 P. 99) and World Bank (1984 p. 129) all present similar types of data. 16. These figures have been derived in rough calculations based on data in Matusek (1981) and Marrese and Vanous (1983). They are for illustrative purposes only. 17. The data are taken from the World Bank's "Romania Economic Memorandum' (1985 p. 10). 18. I interpret certain statements in Hungary: Economic Developments and Reforms as examples of such policy advice. See pp.-33, 42, 61 cf World Bank (1984). 19. See Section II(f) for a brief discussion of the distinguishing charac- teristics of Soviet commercial relations. Brada (1985) provides a more detailed review of' current views. 20. In fact, one has many sets. TR's used in hard goods trade are different from those used in soft. Also, the $'s used in bilaterally-balanced intra-CMEA trade are different from those in multilateral trade. 21. For some countries, these data might have already been converted into domestic currency. If so, it is assumed that one knows the rates at which the conversion has been made. Then one can easily derive the. original TR and convertible currency figures. 22. One is assuming, of course, that economic welfare can be defined unamb'guously. This assumption is implicitly present in all practical policy analyses, in the meaningful construction of all aggregate measures of economic activity, and in a large proportion of theoretical welfare economics exercises. It seems judicious, therefore, not to question the assumption here. In a study particularly pertinent to the present exercise, Kornai (1967, p. 421) shows how a single welfare criterion can be used, even if non- comparable goals of economic policy are prezunt. 23. This is not to suggest that such methodological discussions are irrelevant here. Instead, given present knowledge on the CMEA, they could only be informative at a much later stage. 24. Of course, the official exchange rate could be equal to the shadow rate by coincidence, or even by design. As will be shown in Section V1, some CMEA countries are moving towards the use of official exchange rates that implicitly reflect shadow price calculations. 25. This is especially difficul.t for soft goods. 26. See Bacha and Taylor (1971) for a review of the literature. 307/CMEA/P.Murrell/MS/mcc/03.04.86 -79- 27. Additionally, one must know to which country the debt is owed. This is a lesser problem for two reasons. Most CMEA debt will be owed to the Soviet Union and the variation in real valuation of trade flows is less between co-untries than between hard and soft goods. 28. These calculations are extremely rougn. They also ignore some of the issues raised in section IV(f). Nevertheless, the approximate order of magnitude is probably reliable. 29. The use of the term "bias" does not imply that the data has been manipulated in any way. Rather, all that is implied is that data collection procedures in a country do not reflect the requirements of the particular policy analysis. 30. Here again, the data on dollar-denominated trade flows within the CMEA is ignored to avoid making the discussion too complex. 31. In the World Bank's Hungary study, such information was not given. 32. If the debt is given in local currencies, then the discussion in the previous paragraph will apply. 33. See Marer (1981) on Hungary and Michalski (1985) on Poland. van Brabant (1985 p 28) states that the methodology underlying the Romanian rate is unknown. 34. For Poland this date was early 1981 (van Brabant 1985 p. 28) and, at least in theory, the exchange rate changes if the cost of earning a unit of foreign currency changes by more than 5%. 35. This point applies to the 1984 study. 36. The Marrese-Vanous study focuses on the Soviet Union. Thus, other intra- CMEA trade is ignored. 37. In fact, because of the difficulty of deriving unit values for each time period, Marrese and Vanous (1984) use this quicker method study to derive estimates for some time periods. 38. See World Bank (1984 p. 176-7) for an example of such data. 39. Even then one must remember that the methods outlined in V(c) make a number of assumptions about the comparability of CMEA countries. 40. See the discussion in Vanous (1984)r 41. Average CMEA TR/$ exchange rates are used rather than those specific to Romania, which are probably less reliable. If the Romanian rates were used, the features of the example would be enhanced. 307/CMEA/P.Murrell/MS/mcc/03.04.86 -80- 42. The structure of non-Soviet intra-CMEA trade is very similar to that of Soviet imports from the rest of the CMEA. Hence, the present assumption seems appropriate when revaluing Romanian exports. 43. The recalculations contain an error due to the treatment of $ CMEA exports as if they were originally valued in TR's. Given the small share of CMEA trade in $'s, that error cannot be significant in the context of the present example. 307/CMEA/P.Murrell/MS/mcc/03.04.86 -81- REFERENCES Bacha, Edmar and Lance Taylor, "Foreign Exchange Shadow Prices: A Critical Review of Current Theories," The Quarterly Journal of Economics, vol. LXXXV, No. 2, May 1971, 197-224. Birman, Igor, "From the Achieved Level", Soviet Studies, Vol. XXX,No. 2, April 1978, 153-72. Brada, Josef C., "Soviet Subsidization of Eastern Europe: The Primacy of Economics over Politics?", Journal of Comparative Economics, Vol. 9, No. 1, March 1985, 80-92. Brainard, Lawrence J., "Soviet Foreign Trade Planning", in The Soviet Economy in a New Perspective, John P. Hardt, Ed. (Washington, D.C.: Government Printing Office, 1976). . Brainard, Lawrence, J., "CMEA Financial System and Integration", in East European Integration and East-West Trade, Paul Marer and John Michael Montias, Eds. (Bloomington: Indiana University Press, 1980), 121-38. Hewett, Edward A., Foreign Trade Prices in the Council for Mutual Economic Assi6tance (New York: Cambridge University Press, 1974). Jones, R.W. and P.B. Kenen, Eds., Handbook of International Economics, (Amsterdam: North Holland, 1985). Kornai, Janos, Mathematical Planning of Structural Decisions, (Amsterdam: North-Holland Publishing Co., 1967). Magenheim, Ellen and Peter Murrell, "How to Haggle and to Stay Firm.. Barter as hidden Price Discrimination" (Mimeo, University of Maryland, 1985). >iarer, Paul, "Economic Performance, Strategy, and Prospects in Eastern Europe", East European Economies Post-Helsinki Joint Economic Committee, 95th Congress, 1st Session, August 25, 1977, 523-566. , "Exchange Rates and Convertibility in Hungary's New E.conomic Mechanism", East European Economic Assessment, Part 1 - Country Studies, Joint Economic Committee, 97th Congress, 1st Session, February 27, 1981, 525-548. , "Economic Performance and Prospects in Eastern Europe: Analytical Summary and Interpretation of Findings", East European Economic Assessment Part 2 - Regional Assessments, Joint Economkic Committee, 97th Congress, 1st Sessiotr (July 10, 1981), 19-95. 307/CMEA/P.Murrell/MS/mcc/03.18.86. -82- Marrese, Michael and Jan Vanous, Soviet Subsidization of Trade with Eastern Europe: A Soviet Perspective, Institute of International Studies, Research Series No. 52 (University of California, Berkeley, 1983). Soviet East-European Trade Relations: Recent Patterns and Likely Future Developments, (Washington, D.C.: Wharton eccnometric Forecasting Association Associates, October 1984). Matusek, Ivan V., "Eastern Europe: Political Context", East European Economic Assessment, Part 2 - Regional Assessments, Joint Economic Committee, 97th Congress, 1st Session (July 10, 1981), 96-106. Michalski, Ryszard, "Intralinks between the Economic Reform in Poland and her System of Cooperation wizh the CMEA Countries", (Mimeo, Warsaw, March 1985). Riley, John and Richard Zeckhauser, "Optimal Selling Strategies: When to Haggle, When to Hold Firm", Quarterly Journal of Economics, 98 (2) May 1983, 267<-289. Van Brabant, Jozef M.P., Bilateralism and Structural Bilateralism in Intra- CMEA Trade, (Rotterdam University Press, 1973). East European Cooperation. The Role of Money and Finance, (New York: PraGger Publishers, 1977). ___________ _ Socialist Economic Integration CCambridge: Cambridge Univeisity Press, 1980). , "Eastern European Exchange Rates and Exchange Policies", Jarbuch der Wirtschaft Osteuropas, (1985), 123-172. _ _, "Exchange Rates in Eastern Europe -- Types, Derivation, Application", World Bank Working Papers, 1985. Vanous, Jan, "Soviet and East European Foreign Trade in the 1970's: A Quantitative Assessment", East European Economic Assessment Part 2 - Regional Assessments, Joint Economic Committee, 97th Congress, 1st Session (July 10, 1981), 685-715. , "Commercial Exchange Rates of East European Countries and Realistic Dollar/Ruble Exchange Rates for Soviet Trade with Eastern Europe", Wharton Econometri.c Forecasting Associates, September 20, 1984. Wolf, Thomas A., "Exchange Rate Systems and Adjustment in Planned Economies", International Monetary Fund Staff Papers, Vol. 32, No, 2, June 1985, 211-47. 307/CMEA/P.MIurrell/MS/mcc/03.18.86 -83- World Bank, Hungary: Economic Developments and Reforms (Washington, D.C.: The World Bank, February 1984). World Bank, Romania Economic Memorandum, August 21, 1985. Zoeter, Joan Parpert, "Eastern Europe: The Hard Currency Debt", East European Economic Assessment, Part 2 - Regional Assessments, Joint Economic Committee, 97th Congress, 1st Session (July 10, 1981), 716-731. 307/CMEA/P.Murrell/MS/mcc/03.18.86