PART FOUR

Appendices

ON SUBJECTS OF SPECIAL INTEREST

APPENDIX A

Three Commodity Trade Classifications

CEMA Trade Nomenclature (CTN)

History

A brief history of CTN, the “Uniform CEMA Foreign Trade Commodity Nomenclature,” can be pieced together from various East European sources. In March 1959 a group of foreign trade statistical experts from CEMA countries recommended adoption of a uniform nomenclature as of 1960 (U-3,* 1968, p. 48), and the CEMA Secretariat prepared a draft CTN based on this recommendation. Apparently after some delay, the draft was examined by another CEMA expert committee in September 1960 and was superseded by a revised and expanded CTN (H-8, June 1964, p. 652 fn.). This last was formally adopted by CEMA members as of 1962 (U-3, January 1971, p. 49).

The 1962 CTN is a slightly modified version of the U.S.S.R.'s “Uniform Commodity Nomenclature of Foreign Trade,” first published in 1954 (U-4, as quoted in U-1, 1960, p. 6). There are indications that the U.S.S.R. nomenclature was introduced about 1950.** Even before the modified U.S.S.R. code was officially adopted, it had been used by some CEMA countries as the “CEMA code.” For example, Bulgarian foreign trade yearbooks published during the late 1950s show that the commodity classification is based on “1953 CEMA” nomenclature. (B-1, 1956, p. 156, and B-7, p. 13.) The 1956 Hungarian yearbook also indicates that commodity breakdowns are based on a “classification employed by countries participating in CEMA.” (H-1, 1956.)

*See Key to the Citations, p. 277 above.

**The January 1971 issue of an English language Soviet journal states: “For more than 20 years now, the Soviet Union has had a Uniform Foreign Trade Commodity Nomenclature, which has been widely used not only at home but also abroad, in other socialist countries.” (U-3, p. 48.)

Both the original U.S.S.R. nomenclature of 1954 and the subsequent 1962 CTN have undergone revisions. These have affected neither the basic principles of classification nor the number and definition of onedigit categories and broad divisions, until the most recent revision in 1971, as noted below. Hence, revisions should not affect the comparability over time of series included in this Compendium. Revisions have added new multidigit commodity designations, reclassified others, and clarified definitions. A partial reconciliation list* for the original 1954 U.S.S.R. nomenclature, the 1962 CTN code (since 1962 the U.S.S.R. nomenclature and the CEMA Trade Nomenclature have become officially identical), and the 1967 revised second edition of the CTN, shown in a U.S. document (W-3, vol. I, pp. 166-215), indicated that the revisions at two- and three-digit levels were minor. A discussion of differences between the 1954 U.S.S.R. nomenclature and the 1962 CTN, including examples and the reason for the revision, can be found in W-6, pp. 27778.

*The reconciliation list is partial because only those codes appear on the list for which trade was reported in U.S.S.R. yearbooks.

All CEMA countries record the commodity composition of trade in CTN even though not every set of CTN statistics is published. The CTN is also used in trade agreements involving CEMA countries.

Classification System

The CTN is a seven-digit code (compared to the six-digit U.S.S.R. code of 1954), with each commodity designation preceded by a code number. The first digit of the code indicates categories, also called sections (of which there are 9); the second digit, commodity groups (58); the third, commodity subgroups (309); the fourth and fifth, commodity positions; and the sixth and seventh, commodity subpositions (over 4000 positions and subpositions, combined) (U-3, September 1968, p. 48). Judging from the date of publication of the source cited, the numbers in parentheses refer to the 1967 second edition.

In constructing the CTN, it was intended that the classification allow division of commodities according to (1) investment vs. consumption goods, (2) fixed vs. variable capital, (3) industrial vs. agricultural commodities, and (4) degree of processing. Within commodity categories, two-digit groups usually indicate either the designation of industrial equipment (category 1 except 16), that of consumer goods (9), or nature of material (2 to 8). The third digit usually identifies the degree of processing from raw materials through intermediate products (however, in Categories 1 and 9, subgroups detail the type of machinery and equipment and type of industrial consumer good, respectively). A discussion of how well the CTN classification system fulfills its objectives can be found in W-6, pp. 274-77. The nine one-digit categories and four CTN broad divisions (in which data are shown in the statistical tables included here) are identified in Table 1 of the Introduction. Codes and designations to three digits (maximum detail in which data are shown in the Data Bank and used in conversion to BEC and SITC) are shown in Table A-1.

Table A-1

CEMA Trade Nomenclature (CTN), Three Digits
(Effective 1962-1970)

Source: Translated by the author (with notes in parentheses added to elucidate meaning) from H-8, June 1964, pp. 652-59, except for codes 13, 130-34 (which appear to be inadvertent omissions) which were obtained from U.S.S.R. sources as shown in U-4, 1962, p.388; CTN 233, obtained from Polish sources, shown in P-2, 1968, p.253; and CTN 178,179, and 907 (which appear to have been added after 1962), from U.S.S.R. sources as shown in W-3, vol. I, pp. 19-60.

CTN Category 6, shown in the original as a separate broad division, was included in Broad Division III, as is the more recent practice in CEMA countries.

Subgroup designated as 593 above is shown in U-4, 1962, p. 394, as 599. Poland makes use of CTN 160 as a separate subgroup. Equipment and Machinery for Mining, instead of grouping these commodities with CTN 161.

The CTN codes and designations shown in Table A-1 are largely based on a listing of the 1962 CTN found in a 1964 Hungarian journal. This source listed all 58 CTN groups but only 297 subgroups, i.e., 12 fewer than allegedly contained in the 1967 CTN (see above). Some of the missing subgroups represent obvious omissions, such as subgroups 130 through 134, but the discrepancy may also be due in part to subgroups added in the 1967 second edition. The CTN shown in Table A-1 was supplemented with additional subgroups shown in Soviet and Polish publications, bringing the number of subgroups to 306. Thus, three of the 309 subgroups are not shown because they cannot be identified.

Revised and Enlarged 1971 Edition

The 1971 revision does not affect the consistency of the statistical series presented here, which extend only to 1969, but in updating the series beyond 1970, the revision would have to be taken into account.

This new edition was prepared by CEMA experts during 1968 and 1969, was officially adopted at the December 1969 (14th) meeting of the CEMA Standing Commission for Statistics, and was introduced as of January 1971 (U-3, January 1971, p. 49). The publication which lists the revised nomenclature to its full seven digits covers more than 600 printed pages and appears to be a full listing (see CEMA-1). The possibility that some codes are not shown cannot of course be excluded. Explicit statements in the CEMA literature as well as the comparison of the 1971 CTN with the earlier versions down to the three-digit level disclose that this revision was more substantial than the ones in 1962 and 1967. A number of two-digit groups and three-digit subgroups have been added or deleted, while the content of still others has been changed. In terms of numbers, groups have been reduced from 58 to 57, subgroups increased from 309 to 317, and positions plus subpositions decreased from the previously reported “over 4,000” to 3,945 (CEMA-1, p. 4).

It is useful to distinguish two types of revisions in the CTN: a reclassification of commodities from one CTN category to another, which should be taken into account when updating the Compendium and Data Bank beyond 1970; and reclassifications, additions, or deletions within CTN categories, which also need to be taken into account when updating the Data Bank beyond 1970.

A comparison of the 1971 CTN with the earlier version shown in Table A-1 reveals three reclassifications across categories:

(1) The transfer of CTN 29 (Cable Products) to CTN 113 (Cable and Wire), i.e., a reclassification of a commodity group from a semifinished metal product in Category 2 to a subgroup belonging to the power and electric equipment group in Category 1. It is interesting to note that a Soviet source states as an explanation that the U.N.'s SITC “also places cable products in the ‘Machinery and Equipment’ section.” (U-3, January 1971, p. 49.)

(2) CTN 54 (Tobacco Raw Materials) has been moved to CTN 726. Thus, raw tobacco is now considered a material for the food industry, a change which improves the internal consistency of CTN because tobacco products have always been included among foodstuffs in Category 8.

(3) CTN 845 (Vitamins) has been reclassified as CTN 967, i.e., moved from foodstuffs (Category 8) to industrial consumer goods (Category 9), a change which appears to be justified to improve the internal consistency of classification.

Changes involving reclassifications within CTN categories are much more numerous: codes have been added, deleted, transferred (i.e., a commodity group or subgroup designated by one code number in earlier CTN was given a new code number) or redesignated (i.e., the same code number assigned a new commodity designation). The most frequent type of change, the “miscellaneous” or “not elsewhere specified” subgroup, has been consistently given the last three-digit numerical position, i.e., the code always ending with 9. Thus, CTN 135 (Other Hoisting and Conveying Equipment) has been changed to CTN 139; CTN 194 (Other Transport Equipment), to 199; CTN 243 (Other Nonmetallic Minerals), to 249; and so on. About two dozen transfers or code additions of this kind can be noted.

A further apparently systematic change has been the moving of spare parts, previously generally included with the appropriate machinery subgroup, to the “not elsewhere specified” subgroup. This change in classification has been deduced from the new three-digit listing by noting that subgroup designations no longer include the phrase, “and spare parts.” But whether the reallocation of spare parts has been consistent, and if so, always into the last subgroup, is not known for certain.

A careful study of the revisions indicates that their primary purpose was to make the CTN internally more consistent without changing any of the basic principles of classification. For example, Group 16 (Equipment and Parts for Complete Industrial Plants) was transferred to other groups in Category 1 in order to classify all machinery according to type. In the old CTN, Group 16 included all types of machinery and equipment if destined for assembling into complete plants. This practice clashed with the general principle of classifying machinery by type and also caused a significant portion of machinery shipments to be effectively lost from detailed commodity statistics.

Instead of listing the complete CTN-1971 to three digits, let us reconstruct the revisions by comparing codes and designations shown in Table A-1 with those published in CTN-1971 (official source: CEMA-1, pp. 623-34). The revisions are grouped into three categories: new codes, transferred codes, and old codes with new designations. These changes are shown in Tables A-2, A-3, and A-4, respectively.

An important innovation of the CTN-1971 is the inclusion of a 70- itern Appendix, “Production Operations Not Included in Sections 1-9 of the Uniform CEMA Foreign Trade Commodity Nomenclature,” which lists invisible commercial transactions. These entries have “O” codes (which can thus be considered a tenth CTN category), and are divided into nine two-digit groups (01 to 09), each comprised of several fivedigit items (CEMA-1, pp. 601-15). The two-digit group classifications are:

Table A-2

Revisions in the 1971 CTN: New Codes

A Soviet source states:

The need to classify such operations arises because the |CTN| is used not only for operational and statistical records but also for accounting and foreign exchange transactions records, where these transactions are relevant. [U-3. January 1971, p. 49.]

Table A-3

Revisions in the 1971 CTN: Transferred Codes

a. Old CTN codes have been deleted except those with an *, which have been given new designations, as shown in Table A-4.

b. The new CTN codes have been added except those identified with **, whose designation has been changed (as shown in Table A-4) or those with ***, whose designation has remained essentially unchanged.

Standard International Trade Classification (SITC)

History

The first edition of the United Nations Standard International Trade Classification (the “original SITC”) was prepared by the organization's Secretariat in 1950. By I960, governments of countries accounting for approximately 80 percent of world trade were compiling their commodity composition according to the SITC system, as had the major international agencies reporting trade. To facilitate regrouping data from the Brussels Tariff Nomenclature, still widely used in West Europe by customs authorities to record commodities crossing frontiers, the SITC was revised by the UN in 1960 (“SITC, Revised”).

Classification System and Limitations

The 1960 edition essentially preserved the original SITC structure down to the three-digit level. The 1,312 five-digit items in the revised code are summarized into 625 subgroups, these into 177 groups, then 56 divisions, which are consolidated into 10 sections (or categories). To facilitate aggregation, the first digit determines section, the second division, the third group, the fourth subgroup, the fifth item, the sixth and seventh (of interest only to particular groups of countries or international organizations) article or subitem. Codes and designations to two digits are shown in Table A-5; codes and designations to five digits can be found in UN- 7.

Table A-4

Revisions in the 1971 CTN: Changes in Commodity Designations^a

a. The exclusion of spare parts from machinery subgroup designations is not shown.

Since SITC was developed from a customs nomenclature, SITC groupings may be according to nature of material, final use, or degree of processing. Accordingly, aggregates obtained by summing SITC codes have certain limitations for economic analysis.* For example, gasoline and various other processed petroleum products (SITC 332) are classified with primary products (SITC 0-4), while unrefined nonferrous metals (SITC 68), and lime and cement (SITC 661) are included with manufactured goods (SITC 5-8). These and similar problems have prompted the U.N. Statistical Commission to design a new “Broad Economic Categories” nomenclature.

*The Statistical Commission of the United Nations observes: “The SITC, Revised … as it stands is not entirely suitable for analysis by end-use .…” (UN-4, p. vii.)

Table A-5

Standard International Trade Classification, Revised (SITC), Two Digits

Source: UN-7, pp. 3-4, 45; except division 92, added by Yugoslavia (Y-2, various issues).

Broad Economic Categories (BEC)

Classification System

A new and improved system, BEC is comprised of only 19 basic categories within 7 large economic classes, as shown in Table 3 of the Introduction. Large economic classes are one-digit entries 1 through 7; basic categories are one-, two-, or three-digit entries not further divided. Each BEC category is defined in terms of one- to five-digit codes of the SITC, as shown in UN-4.

The 19 basic categories (“elements”) of BEC can be combined not only into the 7 large economic classes but also according to two additional and independent aggregation procedures: broad end use (“primary products” and “processed goods”) and basic SNA classes (“capital goods,” “intermediate goods,” and “consumption goods”).

The BEC system classified commodities as “primary” if (1) they are characteristically the products of primary sectors, i.e., farming, forestry, fishing, hunting, and the extractive industries; (2) transformation or processing of primary products accounts for only a small amount of value added; and (3) they are waste and scrap materials. Any commodity not defined as “primary” is classified as “processed.” To illustrate, ginned cotton is classified as a primary commodity since almost all of its total value derives from agriculture. In contrast, canned and prepared foods, which receive much of their value in food processing, are counted as “processed” (UN-6, p. 34).

As far as matching BEC elements and SNA basic classes is concerned, it is recognized that a fully accurate alignment is not possible. The main problem is that SNA classifies commodities according to end use but a given good entering into international trade may often be put to alternative end uses. Accordingly, commodities are generally allocated in the BEC according to their customary end use. (UN-4, p. vii.)

The aggregation of BEC basic categories according to broad end use and SNA basic classes is shown in Table A-6.

Table A-6

BEC Aggregates by Broad End Use and Basic SNA Classes

Source: UN-1.

APPENDIX B

Reconciliation of Classifications

CTN to BEC

CTN and BEC Classifications Compared

The transformation of CTN to BEC is facilitated by the relatively small number of BEC categories as well as by similarities in the principles of classification in the two nomenclatures. The 7 large economic classes of the BEC correspond rather closely to combinations of CTN categories and groups, so any CEMA country's trade specified in two-digit CTN can be regrouped into the 7 large economic classes of the BEC, as is shown in Table B-1.

Table B-1

Correspondence of BEC Large Economic Classes with CTN Commodity Categories and Groups

A highly accurate transformation of CTN to BEC can be made also from only one-digit CTN categories or the even more aggregated CTN broad divisions, if BEC's 7 large economic classes are combined into 4 “larger” economic classes. Since the commodity composition of CEMA countries' trade is available most readily according to broad divisions, it is of interest to examine how accurately these divisions can be aligned with BEC's 4 “larger” economic classes. A comparison of the most aggregate groupings of commodities in the two nomenclatures is shown in Table B-2.

The approximate correspondence of BEC large economic classes and CTN broad divisions and the required adjustments are discussed next.

Table B-2

A Comparison of BEC Large Economic Classes and CTN Broad Divisions

An item-by-item examination of the two nomenclatures revealed five commodities of some importance classified differently in the two systems: Unmilled Corn, Unmilled Oats, Live Animals Not for Slaughter, Tobacco Manufactures, and Vitamins are all included in CTN III but not in BEC 1. The BEC classifies Unmilled Corn and Oats with Industrial Supplies (BEC 2); Live Animals Not for Slaughter with goods Not Elsewhere Specified (BEC 7), and Tobacco Manufactures and Vitamins with Consumer Goods (BEC 6). Since as far as CEMA countries are concerned the only commodity of any importance included in their published trade statistics among those listed under BEC 7 is Live Animals Not for Slaughter, BEC 1 and 7 can be combined and aligned with CTN III except for the four commodities listed.

The main reason for showing Fuels and Lubricants (BEC 3) separately from Industrial Supplies (BEC 2) and Transport Equipment (BEC 5) separately from Machinery and Equipment (BEC 4) is that fuels and means of transport can be used for both industrial and household consumption. Thus, the complete BEC (see Table 3 in the Introduction) presents a further breakdown of fuels (which includes Motor Spirits) and means of transport (which includes Passenger Motor Vehicles) to permit the allocation of these categories between industrial and household uses as desired. (In fact, such an allocation is necessary when aggregating according to SNA basic classes.) However, according to CTN, all fuels are included with industrial supplies and all means of transport with industrial machinery and equipment. Thus,by combining BEC large economic classes 2 plus 3 and 4 plus 5 into “larger” economic classes, no inaccuracy whatever is introduced as far as alignment with CTN Broad Divisions II and I, respectively, is concerned.

To be sure, to the extent that not all CEMA exports and imports of transport equipment and fuels are destined for industrial use, the descriptive accuracy of CTN Broad Divisions I and II is impaired. One may add, however, that since the production and consumption of privately owned motor vehicles in East Europe was (and still is) limited, the assumption that transport equipment and fuels are destined for industry is much less inaccurate than a similar assumption would be for most Western countries.

These two aggregates can be aligned quite accurately except for Tobacco Manufactures and Vitamins, which are included in BEC 6 and excluded from CTN IV.

Summarizing the above, we find the correspondence between BEC “larger” economic classes and CTN broad divisions as follows:

An Original Conversion Key

To convert CTN data to BEC aggregates by broad end use or SNA basic classes was much more difficult than regrouping by large economic classes because a satisfactory transformation involves disaggregation of trade flows in CTN and reaggregation on the basis of the BEC system. To facilitate conversion of East Europe's not fully detailed CTN data to BEC, the original 19 basic categories of BEC were combined here into 11, as shown in Table B-3.

The abridgement involved three changes in the original BEC classification, as follows:

(1) BEC categories 41, 42, 51, 522, and 523 were combined to new BEC 4, Machinery, Transport Equipment, Accessories, and Parts. For most CEMA countries total trade in machinery is available but further details, particularly on spare parts and accessories, tend to be sketchy. Since in the original BEC all five categories here combined were summed as Processed (see Table A-6), no significant information is lost when these categories are combined for greater comparability with CTN. However, in so doing, Spare Parts and Accessories (BEC 42 and 53), which would have been included in Intermediate Goods according to SNA basic classes (see Table A-6) are now included in Capital Goods, which is therefore biased upward. Similarly, all of BEC 51 (to be allocated by the user) is now included in Capital Goods, although a certain portion of trade in Passenger Motor Vehicles would need to be allocated to Consumer Goods.

Table B-3

Combined Broad Economic Categories (BEC) and Corresponding CTN Headings

*Special allocation between BEC 21 and 22 is required for Natural and Synthetic Raw Rubber, respectively.

(2) BEC 32, Processed Fuels and Lubricants, is not broken down into 321 (Motor Spirit) and 322 (Other Processed Fuels) because this CTN detail is generally not available. Once again, this elision does not affect the Primary and Processed subtotals. In the SNA subtotals all of BEC 32 is classified as Intermediate rather than Consumption goods, following the reasoning set forth on p. 328 above.

(3) The final such combination, to new BEC 6, Consumer Goods, of Durable (61), Semidurable (62), and Nondurable (63) articles, raises no problems of aggregation by degree of processing or SNA classes.

Next, a CTN/BEC conversion key was prepared, which made use of that most highly aggregated CTN code which could be assigned accurately to a BEC combined category. The key, classified by BEC headings, is shown in Table B-3; cross-classified in CTN sequence, in Table B-4.

Table B-4

Conversion of CEMA Trade Nomenclature (CTN) to Combined Broad Economic Categories (BEC)

*Total trade less trade specified in CTN.

The following simplifying conventions were found to be useful or necessary:

(1) Food and Beverages, Primary, for Industry (BEC 111) now includes Live Animals Not for Slaughter (CTN 6) to avoid retaining a separate BEC category just for this small item, and, also, to be able to allocate all of CTN Broad Division III to BEC Large Economic Class 1, as was discussed above. BEC 111 also includes (instead of BEC 21, as in the original) the feed cereals oats and corn (CTN 70004 and 70005) because CTN to three digits does not show them separately from wheat, rye, and barley (CTN 70001, -2, and -3, respectively). Since both BEC 111 and 21 are Primary as well as Intermediate goods, this raises no problem of aggregation by degree of processing or SNA classes.

(2) Food and Beverages, Processed, for Industry (BEC 121) omits unrefined sugar (CTN 80002) and some additional smaller items because they are not specified at three-digit CTN. They are included in BEC 122, which thus may be slightly upward biased. This will affect aggregation only by SNA classes, in favor of Consumer and at the expense of Intermediate goods.

(3) Industrial Supplies, Primary (BEC 21) is downward biased because it excludes oats and corn (see (1) above) as well as waste and scrap materials (except those of metals), the latter causing the Primary subtotal to be understated.

Special Problems and Evaluation

A special problem encountered during the CTN/BEC conversion was that for certain CEMA countries or years some CTN entries were not available in sufficient detail to use the conversion key without further adjustment. In these few cases, special allocations were made, and are documented in notes to the appropriate tables.

No attempt has been made as yet to quantify the bias introduced by combining original BEC categories and by simplifying assumptions. However, one may conclude that whenever CTN data were available in detail required by the conversion key, the resulting BEC series provide a reasonably reliable empirical foundation for analysis and inter national comparisons. It is believed that the classification of spare parts with Capital instead of Intermediate goods may result in the most significant bias.

SITC to BEC

The original BEC categories were designed by the U.N. to facilitate conversion from SITC data; in fact, BEC categories are defined in terms of SITC headings. This SITC/BEC conversion key has been prepared and published (UN-4), and it individually assigns more than 600 two- to five-digit SITC headings to appropriate BEC categories (SITC contains a total of 1,312 five-digit items).

The trade of individual East European countries with West Europe as a group (Series IV) is the longest and most important SITC series in this collection which was to be converted to BEC. After preliminary processing of the data (including assigning the SITC code when not given), it was revealed that usually no more than two- or three-digit SITC details would be available, so the U.N.'s SITC/BEC key, requiring up to five-digit SITC detail, could not be applied. However, an item- by-item examination of the U.N. key revealed that in most cases two- or three-digit SITC can be matched accurately to one of the 11 combined (rather than the original 19) basic categories of BEC. Such an “abbreviated” key was constructed and is shown in Table B-5.

The new SITC/BEC key in turn was used as a guide to process East-West trade data, by computer, as summarized next.

In the U.N. publications on East-West trade, the number of component commodity groups has varied by subperiods, increasing from about 20 in 1950 to nearly 200 by 1964. Since 1964 the data have been published with three-digit SITC codes; before 1964 only commodity or commodity group designations were given, to which two- or three-digit SITC codes (or code spans) could be assigned. Even though only onedigit SITC category subtotals are shown in the Compendium, greater commodity detail was included in the Data Bank. The number of commodity series processed was determined, subject to data availability, by two objectives: to obtain a consistent series by SITC one-digit categories and two-digit divisions, and to convert SITC data to BEC. The decision to disregard SITC detail beyond two digits except when needed for more accurate transformation to BEC was made because SITC data even to two digits are voluminous, and also because the more aggregated the commodity groups the better the chances that data in comparable classification for earlier subperiods could be found. Approximately 50 export and import series were processed for each CEMA country.

The following example illustrates how the two criteria were jointly considered. Textile Fibers (SITC 26) is shown in U.N. publications for most years by commodities (wool, cotton, jute, etc.), each a three-digit component of SITC 26. Thus, only SITC 26 would need to be included to attain the first objective. However, conversion to BEC requires that SITC group 266 (Synthetic Fibers) be placed in one BEC category, while all other SITC groups (natural textile fibers and textile waste materials) tally with another category of BEC. Thus, SITC 26 would be too highly aggregated, but including all subgroups of SITC 26 would be unnecessarily detailed. Accordingly, two commodity groups were included: Natural Textile Fibers and Textile Wastes (SITC 261-65, 267) and Synthetic Fibers (SITC 266).

Table B-5

Conversion of Standard International Trade Classification (SITC) to Combined Broad Economic Categories (BEC)¹

1. This key has been constructed primarily for application to trade between individual East European countries and West Europe as a group.

2. Trade assumed to consist of refined sugar and sugar products only. For U.S.S.R. trade with West Europe this assumption is valid on the basis of U.S.S.R. trade statistics.

3. East-West trade statistics usually include components of SITC 09 with 07 and other divisions in section “O.”

4. Trade assumed to consist of unmanufactured tobacco.

5. Includes unprocessed lignite, peat, etc.

6. Includes coal, lignite, and peat briquettes.

To construct East-West trade series that conform with the two objectives of consistency of SITC as well as BEC series over time, it was necessary to adjust the original U.N. data by making special allocations among SITC categories as well as SITC groups.

Allocations among SITC Categories. If in a subperiod a commodity group was found to correspond to an SITC code span which contained different SITC first digits, trade was allocated to two or more commodity designations by a special procedure. To illustrate, trade in Oil Seeds (SITC 22) and Animal and Vegetable Oils and Fats (SITC 4) were each reported for 1952-57 and again since 1964, but given only as a combined commodity group for 1958-63. Since the two groups have different SITC first digits, the value was distributed between the components by interpolation, on the basis of proportions calculated for subsequent and (if available) preceding years. Entries so obtained are identified by a special footnote code in the Data Bank.

The value of commodities so allocated is considerably less than 10 percent for each East European country for most years. Between 1952 and 1963 there were in fact only four commodity groups of importance for which special allocations were made: Oil Seeds, Fats, and Oils; Fur skins and Products (if Undressed, SITC 21; if Tanned and Dressed, SITC 61; if made into clothing, SITC 84); Fertilizers (if Crude, SITC 27; if Manufactured, SITC 56); and Textiles (Yarn and Fabrics, SITC 65; Clothing, SITC 84). Since 1964 adequate commodity detail has been published so that no special allocation was necessary; for 1950-52 additional allocations were made because the statistics then published were highly aggregated.

The accuracy of SITC one-digit categories shown in the Compendium would be affected adversely by these special allocations only if the composition of trade with respect to the allocated commodity groups was atypical for the years requiring adjustments.

Allocation among SITC Groups. For certain commodity groups, the SITC detail required for conversion was unavailable for some years. For example, within SITC 02, Dairy Products and Eggs, subgroups 022-024 (various dairy products) converts to one BEC category while SITC 025 (Eggs) converts to another. For 1957-63 only the combined commodity group was shown, so a special allocation by a method similar to that detailed above was made for each East European country's exports and imports. The additional allocations were also mostly for food items (which are all in SITC category “O”) so these do not affect the accuracy of SITC categories.

For a few commodities, accurate BEC conversion required greater than three-digit SITC detail, which was not available in U.N. publications. The most important was SITC 321, Coal, Coke, and Briquettes (a principal Polish and U.S.S.R. export to West Europe) because BEC classifies coal as Primary, and coke and briquettes as Processed, industrial supplies. For Poland, all SITC 321 exports were classified as Coal (SITC 321.4), based on information obtained from the Polish Commercial Consulate that during the entire postwar period, Poland exported to West Europe almost exclusively coal. For the U.S.S.R., the ratio of coal to coke-briquette exports to the main West European imports combined was calculated for each year from U.S.S.R. sources and the ratios applied to SITC data obtained from West European sources. A few further special allocations were made on a similar basis but the relative importance of the additional commodity groups involved was very small.

CTN to SITC

The United Nations' Conversion Key

In 1965, a subcommittee of the United Nations Statistical Commission prepared a preliminary key bridging SITC and CTN nomenclatures (UN-5). It consists largely of a reconciliation of six- and seven-digit SITC items with five-digit CTN positions. The key is cross-classified in order of SITC and CTN headings. No conversion is shown for less than five-digit entries in either nomenclature. Apparently the U.N. itself has been unable to make use of the key, because of inadequate knowledge of the details of CEMA foreign trade. The only official statement that can be found with regard to the SITC/CTN conversion project states:

In considering the key, the Commission bore in mind the difficulties of reconciling the CMEA classification, which postulated the knowledge available in countries whose economies were centrally planned, of the end use of many of the articles classified, with the SITC. As the use of the key would undoubtedly reveal areas in which it could be improved, the Commission recommended that the Statistical Office of the United Nations should co-operate with the secretariat of the CMEA in preparing a revised version. [UN-6, p. 32.]

That some CEMA countries and possibly the CEMA Secretariat make use of the U.N. key is suggested by the fact that two CEMA countries, Hungary and Czechoslovakia, submitted to the U.N. the SITC commodity composition of their trade not long after the key became available. Another key, linking CTN, SITC, and the Brussels Tariff Nomenclature (BTN), has reportedly been published in the U.S.S.R. (U-5, as shown in W-6, p.277, fn. 10). In the West, however, such conversions have not been made because in most cases only too highly aggregated CTN statistics have been available.

New Conversion Key for U.S.S.R. Total Trade

To obtain adequate correspondence of CTN and SITC, a minimum of three-digit CTN data would seem needed, a disclosure requirement generally not met at the present time.

For U.S.S.R. total trade alone, which is the longest and most detailed (though still incomplete) published CTN series, a transformation to SITC has been carried out by sequential application of a three-part conversion key, shown in Table B-6.

Table B-6

Conversion of CEMA Trade Nomenclature (CTN) to Standard International Trade Classification (SITC)

*Allocation proportions must be determined separately for imports and exports, uniquely for each country.

The new CTN/SITC conversion key was especially designed for CTN data which details all or a substantial part of trade at no less than two- or three-digit level codes, such as are published by the U.S.S.R. for total trade. A primary key was devised to convert two- or threedigit CTN codes into two-digit SITC codes, generally using three-digit CTN except when a two-digit level was sufficient; this would have been the only key needed if all reported trade were specified at the appropriate level. When this was not the case, a less accurate secondary key was employed to transform all remaining two-digit CTN codes (those not included in the primary key because conversion from three-digit CTN was needed but not available). Finally, when only one-digit CTN was available, a tertiary key was formed from one-digit CTN into onedigit SITC.

It is important to note that, in contradistinction to the CTN/SITC key prepared by the U.N. and other keys reconciling two nomenclatures, the purpose here is to obtain accurate SITC aggregates (categories or divisions) rather than a fully detailed CTN-SITC matching of homogeneous articles. Hence, the maximum SITC detail used in the conversion is two digits.

The primary key was prepared by matching commodity descriptions in the two systems. In some cases, exact correspondence between the two codes was not possible since CTN classifies commodities largely according to origin and end use, and SITC classifies them partly by stage of fabrication, partly by industrial origin, and partly by the nature of material. In some cases even three-digit CTN data permit only approximate transformation to two-digit SITC. Since the purpose here is to obtain broad SITC aggregates, differences between the codes was serious only when a given CTN item had to be allocated among two or more SITC categories. To illustrate, Fertilizers (classified by end use) belong to CTN subgroups 340 to 342 (Phosphorous, Potassium, and Nitrogen Fertilizers, respectively), which in the SITC system (classified by stage of fabrication) could be Crude (SITC 271) or Manufactured (SITC 561). Having determined from U.S.S.R. sources that the bulk of fertilizer exports are Crude, the conversion key allocates CTN subgroups 340, 341, and 342 to SITC 26, 27, and 28, respectively.*

The secondary key was designed on the basis of the primary key. To illustrate (see Table B-6), within CTN group 15 the primary key specifies eight subgroups to be converted to SITC 71 and two subgroups to other SITC divisions. If the subgroups of CTN 15 were not at all (or not fully) specified, all of CTN 15 (or its unspecified component) was converted by the secondary key into SITC 71. The tertiary key was devised by a procedure similar to that used for the secondary key.

*U.S.S.R. statistics specify most (but not all) five-digit components of CTN 340, 341, and 342. From the U.N.'s detailed CTN/SITC key it could be determined that most of the exports specified are Crude. In U.S.S.R. imports, the maximum CTN detail is two digits (CTN group 34).

As a general rule, the larger the proportion converted by the primary key the greater the accuracy of the SITC series. Notes to tables give the proportions.

The conversion involved the following iterative steps:

(1) All CTN entries that could be converted directly by the primary key were converted into two-digit SITC codes. For example (see Table B-6), CTN 10, 111, and 12 were converted to SITC 71, 72, and 71, respectively. Note that CTN 110 and 112 were skipped because (assuming that these commodities were traded) published statistics in group 11 do not specify trade in the full three digits. Whether there was in fact trade in CTN 110 or 112 could be determined by subtracting CTN 111 from CTN 11 (see step 3).

(2) Two-digit SITC entries obtained by primary key conversions were numerically ordered and summed so that the proportions of total trade converted by the primary key could be calculated.

(3) Converted CTN entries were subtracted from each of their respective higher-level subtotals and from the grand total in order to obtain so-called remainders. The purpose of this subtraction was to document in greatest possible detail all information which had not yet been converted, in preparation for subsequent iterations. To continue with the above example, CTN 111 was subtracted from CTN 11 while CTN 10, 11, and 12, etc., were subtracted from Category 1, from Broad Division I, and all of them from the grand total.

(4) The secondary key was applied to all the remaining two-digit CTN entries (i.e., two-digit codes where it would have been preferable to convert three-digit components but where these were not available). In the example above, the original data contained an entry for CTN group 11. If there had been no information at a subgroup level, the entire group 11 would have had to be converted by the secondary key. But one three-digit component of 11 (i.e., Ill) was given, which, as the first step, was already converted by the primary key. If this subgroup (or the sum of subgroups, if applicable) exhausted group 11 — a question that is decided when it (or their sum) is subtracted from group 11 — this ends the matter. If the sum of subgroups is less than the given group, there will be a remainder under CTN 11, to be converted to SITC during the second iteration by the secondary key.

(5) Two-digit SITC entries obtained by secondary key conversion were numerically ordered and summed so that the proportion of total trade converted by the secondary key could be calculated.

(6) CTN entries converted by the secondary key were subtracted from their respective subtotals and from the grand total in order to obtain remainders for the third iteration. Thus, to continue with the example, CTN 11 (itself a remainder after CTN 111 was subtracted) and the other two-digit components of CTN 1 converted by the secondary key were subtracted from Category 1, Broad Division I, and from the grand total.

(7) The tertiary key converted remaining one-digit categories (remainders or original data not specified in greater detail) into SITC one-digit categories. Since in the U.S.S.R. only the Machinery and Equipment (CTN 1) subtotal was originally given, there was a need to apply the tertiary key only to this Category (where the unspecified residual in both imports and exports fluctuated between 5 and 15 percent during 1946-1969).

(8) One digit SITC entries obtained by the tertiary key conversion were numerically ordered and summed.

(9) Converted CTN entries were subtracted from the grand total.

(10) All SITC data were merged. Proportions of trade converted by the three keys were combined and tabulated with the “final remainder,” which could not be converted. This final remainder was checked against the “unspecified” column in Tables II. CTN. M.U and II. CTN. X.U; final remainders and unspecified had to be numerically identical.

APPENDIX C

Socialist Trade Prices and the Valuation Problem

Concept of Transaction Value

National statistics on external trade are compiled by most countries according to transaction values. For imports, these values are based on prices at which goods are purchased by the importer, plus transportation and insurance costs to the frontier of the importing country (c.i.f. valuation). For exports, they are based on the prices of sale by the exporter, including costs of transportation and insurance to the frontier of the exporting country (f.o.b. valuation). Since price is a market phenomenon, the concept of transaction value presumes that price is substantially the outcome of the interaction between international supply and demand, even though it is recognized that the recorded value of a commodity may differ on regional markets (UN-6, p. 39).

To compare foreign trade statistics of countries, commodities traded should have a similar base of valuation. Thus, regardless of how the goods traded may be valued or how much they sell for on the domestic market of the exporter or importer, for international comparisons they should be valued approximately at (actual or hypothetical) “world market” prices. This is the only interpretation one can give to the U.N.'s concept of comparable transaction values, even though “world market price” is not an unambiguous concept.

A country whose currency is free from exchange controls or major restrictions can be assumed to trade approximately at world market prices, in response to automatic market pressures. Such a country's transaction values serve to compare its external trade with both domestic transactions and corresponding external transactions by the rest of the world. In contrast, centrally planned economies, which rely on strict exchange controls and other institutional barriers between foreign and domestic markets, have no automatic mechanism maintaining foreign trade prices at or near world market levels. Thus, these countries' trade statistics may not be comparable with either the domestic transactions of the country itself or with the external transactions of the rest of the world, as can be deduced from certain institutional features of CPEs, to which we turn next.

Transaction Values in “National Devisa Units”

What are “National Devisa Units”? Domestic prices in East European countries are generally arbitrary or at least strongly manipulated by central authorities. Official exchange rates are set on the basis of a stated gold content* that is essentially fictitious because currencies are not backed by gold and cannot be exchanged for convertible Western currencies, and also because the stated gold content frequently yields official exchange rates that overvalue East European currencies (according to purchasing power comparisons with Western currencies), often by two- to three-hundred percent or more. To distinguish external trade transactions which have been converted into national currencies at official exchange rates from purely domestic money transactions, the prefix “devisa” is attached to these values by CPEs. Thus, external trade transactions are reported in “national devisa units” (NDUs), that is, in “devisa-rubles” by the U.S.S.R., in “devisa zlotys” by Poland, and so on. The prefix “devisa” is assigned explicitly to trade statistics by East Germany, Hungary, and Poland, whose published foreign trade transactions are in devisa-marks, devisa-forints, and devisa-zlotys, respectively; in other East European countries the “devisa” prefix is implicit and reference to it is seen occasionally in their literature.

Significance of “National Devisa Unit” Values. To elucidate the meaning of values expressed in NDUs, it is necessary to know the prices used in foreign trade transactions. East European countries state that they trade with both other CPEs and Western countries approximately at world market prices. Let us accept this as if it were a fact, for the moment, and trace the recording of transactions accordingly.

In trade with Western countries, prices are nearly always quoted in dollars. The resulting dollar transaction values are converted to NDUs by the official $/NDU rate. In trade with other CPEs, prices are nearly always expressed in devisa-rubles (also called “accounting” or “transferable” rubles), which, given the above price assumption, are simply dollar (or other convertible-currency) world prices, converted into rubles at the official exchange rate of .9 rubles to the dollar. Each East European country (except of course the U.S.S.R.) converts its transactions stated in devisa-rubles to NDUs at the official ruble/NDU exchange rate. It follows that values recorded in NDUs can be reconverted to dollars at the official $/NDU rates because East European countries' exchange rates with the dollar and other Western currencies are consistent with the rate of .9 rubles to the dollar. Thus, East Europe's external trade statistics in NDUs can be converted into dollars at the official exchange rates so that the resulting values would compare with the external trade statistics of the rest of the world if, and only if, transaction prices were comparable to those on world markets.

*Except the East German valuta-mark, which has been defined in terms of the U.S. dollar (4.2 valuta-marks/$). Since the devaluation of the dollar in December 1971, CEMA countries have maintained the nominal gold contents of their currencies but revalued them approximately 8 percent higher in relation to the dollar.

Foreign Trade Prices of East European Countries

Evidence suggests that the assumption that East European countries trade at current world market prices is not correct for trade among CPEs or (fully) for trade between CPEs and non-CPEs.

Prices in Trade With CPEs. CEMA agreements specify that prices are to be based on average world market prices of some previous period as a point of departure. They are then adjusted to take into account various “considerations” stated in such general terms that wide latitude remains for negotiators to set prices. The result, called “adjusted historical world market prices,” usually remain fixed for many years, depending on commodity and trade partner. Considering, further, that trade among CPEs is based substantially on long-term trade agreements specifying quantities of commodities to be exported and imported, and that exports and imports continue to be balanced bilaterally, transaction values in CPEs cannot be said to correspond to current world prices, nor to have been determined largely by regional market forces. The Statistical Office of the U.N. states the problem as follows:

The trade of a centrally planned economy with another centrally planned economy may be carried out on a basis not comparable to that governing its trade with a country whose economy is not centrally planned.… [I]n the absence of data on the unit values of specific commodities entering the trade of the centrally planned economies with one another, it is difficult to assess the possible effects on the trade statistics of this kind of incomparability. [UN-8, 1966, p. 9.]

Recent statistical evidence, comparing unit values or actual foreign trade prices in intra-East European trade with unit values or price quotations of similar commodities on the world market, suggests that the former are substantially higher than the latter. One such estimate, made by this author on the basis of a CEMA statistical study (summarized in H-13, March 1967, pp. 283-300), would place the average markup in the mid-1960s close to 20 percent. The markup appears to have been somewhat higher for manufactures than for primary products, somewhat larger during the early 1960s than during more recent years, and greater in some bilateral relationships than in others (W-11). Even though precise estimates of foreign trade price levels cannot be made at the present time, the conclusion that prices are generally higher in trade among CPEs than on the world market appears substantially correct. These findings, and their significance for interpreting East European statistics, have been confirmed by the following statement in a 1970 U.N. publication:

Intra-east European trade is conducted in almost complete isolation from the rest of the world market, and investigations — whether conducted in eastern or western Europe — indicate that price levels ruling in that trade sector are significantly above those in east-west and intra-west trade. This suggests that trade data … for east European countries have an upward bias, but by how much cannot be ascertained. [UN-2, vol. 21, no. 1, p. 48.]

Prices in Trade with Non-CPEs. In East-West trade, prices in general could be expected to correspond to current world prices. Yet, studies comparing transaction prices in trade between CPEs and non- CPEs with world prices have shown that East European countries offer large discounts in their exports to the West — discounts that appear to be substantially greater than those which could be justified by lower quality. These may be due, inter alia, to Western discrimination in the form of higher tariffs and other obstacles to trade, CEMA countries exporting on the basis of plan directives rather than market considerations, and the need to dispose of unwanted surpluses acquired from other CPEs or other inconvertible currency areas as a result of bilateral agreements. A rough estimate by this author of such discounts would be about 20 percent. Discounts appear to be larger in manufactures than in primary products, and there is some evidence that they have been increasing over time (W-11).

Less evidence is available on prices in East Europe's imports from the West; what there is indicates that import prices are probably close to world prices, perhaps somewhat higher.

Conclusions

Given that prices at which CPEs trade on Eastern and Western markets apparently diverge from world market prices in opposite directions and by substantial margins (depending on trade partner, commodity, and year), official statistics in NDUs or their unadjusted dollar equivalents are not exact and can be misleading. For example, the trade dependence of an East European country on CEMA (or on the combined CPE group) appears to be higher on the basis of official statistics than it would be if calculated at comparable transaction (i.e., world market) prices. It follows, therefore, that the reorientation of trade from one trading area to another may indicate shifts of resources in real terms different than that which would be calculated on the basis of unadjusted NDU or dollar values. Furthermore, since relative prices of machinery and equipment—the fastest growing CTN broad division in intra-CEMA trade — appear to be high on the CEMA market as compared to price ratios on the world market, the rate of growth of intra- CEMA trade calculated from official statistics may be upward biased.

The inevitable conclusion is, therefore, that intraregional and international comparisons should take into account differences in transaction prices which are not reflected in official statistics as expressed in NDUs or their dollar equivalents obtained through official exchange rate conversions. If one could measure the deviation of an East European country's foreign trade prices from world market prices according to trade partners, commodities, and years, trade flows could be revealed. But since information on foreign trade prices (unit values) is too fragmentary to permit such an adjustment for every East European country, the value series in current NDUs have been recorded in the Data Bank without adjustments.* For presentation here, the NDU

*For an experimental adjustment of Hungary's foreign trade statistics at “Adjusted Dollar Prices,” see W-11. series have been converted to U.S. dollars at the official exchange rates shown in Table C-1.

Table C-1

Official Exchange Rates of East European Countries, 1946-1969

Note. For all countries except Yugoslavia, external trade statistics have been recorded in the Data Bank, for all years shown, in NDUs consistent with official currency values in effect in 1969. The official exchange rates of several East European countries were revised at times during the 1946-69 period. Series originally reported in “old” NDUs have been either revised subsequently by the countries themselves in “new” NDUs, or the “old” NDUs have been converted into “new” NDUs by the author before recording the series in the Data Bank. For Yugoslavia, statistics up to 1964 were recorded in the Data Bank at the former rate of 300 dinars per U.S. dollar.

a. New ruble introduced on January 1, 1961. (Prior to 1961, .25 old rubles per dollar.) Pre-1961 series have been revised in official publications to correspond with the new ruble rate.

b. New leva introduced on January 1, 1962. (Prior to 1962, 7.56 old levas per dollar.) Some pre-1962 series have been reported in recent publications in new Ievas, other pre-1962 series have been converted into new Ievas by the author before Data Bank processing (1 old leva = 0.171954 new Ievas).

c. New koruna introduced in 1948. (Prior to 1948, 50 old korunas per dollar.)

d. Prior to 1960, East German official publications reported trade in old rubles. These series have been converted into valuta-marks by the author (1 old ruble = 1.05 valuta marks).

e. Data for 1946-49, 1951-54, and 1956-57 obtained from secondary sources reporting either in lei or in dollars, with exchange rates consistent with the lei per dollar rate shown in the Table.

f. New dinar introduced on January 1, 1966. The new rate as of January 23, 1971 is 15 dinars per dollar.

g. 1946-51 exchange rate was 50 dinars per dollar, but pre-1952 series have been obtained from official publications already using the exchange rate in effect between 1952 and 1965.

h. New lek introduced on August 15, 1965; 10 old leks = 1 new lek.

APPENDIX D

Some Main Problems of Statistical Comparability

Introduction

Strict international comparability of foreign trade statistics continues to be handicapped by numerous differences in definitions and conventions used by individual countries. Statistical standardization has long been promoted by various international organizations such as the U.N. and CEMA, but national statistical practices often fail to conform to proposed international standards because of divergent historical developments and the political, economic, and administrative requirements of national policies.

To improve comparability of international trade statistics, in 1970 the U.N. Statistical Office compiled the explicit and implicit recommendations of the U.N.'s Statistical Commission on the methodology of reporting foreign trade statistics (UN-6). Standardization within the CEMA framework is being promoted by the organization's Standing Commission for Statistics, which is concerned with broader issues, and by the Statistical Department of the CEMA Secretariat, which elaborates specific recommendations. (U-3, September 1968, p. 46.)

The purpose of this section is to summarize statistical concepts and definitions known to be used by individual East European countries in reporting trade statistics, and to compare their practices, as well as recommendations of the CEMA Standing Commission for Statistics, with the most recent recommendations of the U.N. Statistical Commission. Such comparison is useful in allowing for the limitations of the national series presented here for making intraregional and international comparisons.

Four comparative aspects are examined: (1) general and special systems of trade statistics, (2) coverage of merchandise trade, (3) treatment of transport and related expenses (f.o.b. and c.i.f.), and (4) identification of partner country. Two additional aspects, commodity classification systems and valuation, have already been discussed in Appendices A and C, respectively. In each subsection, a general discussion of concepts and the recommendations of the U.N. Statistical Commission will be followed by a summary description of statistical practices among East European countries.

General and Special Reporting Systems

Concepts and United Nations Recommendations

Systems of recording trade differ chiefly in their treatment of reexported goods. Special imports are those directly for domestic consumption or use, including transformation and repair. Special exports comprise national exports — wholly or partly produced or manufactured in a country — plus those which, having originally been included in special imports, are re-exported without the intended transformation, i.e., exports of “nationalized” goods. (The above definition needs to be expanded where a distinction is to be made between national and customs frontiers. If so, special imports would also include imports into customs-bonded manufacturing plants plus imports withdrawn, inward from customs-bonded warehouses and free-trade zones; while special exports would include exports from customs-bonded manufacturing plants. UN-6, pp. 15-25.)

General imports are the combined total of imports for directly domestic consumption plus imports for re-exports. General exports are the combined total of national exports plus re-exports.

The meaning of re-exports is not unambiguous and is of considerable significance. CEMA recommends that all re-exports be included:

Imports comprise foreign goods intended for home consumption or processing and also goods for re-export. The latter imply foreign goods imported and then exported without processing and also goods purchased abroad and then exported direct to a third country. [U-3, September 1968, p. 46.]

According to the U.N., a cargo must cross the national (or customs) frontier of a country to be included in the statistics. Thus, if a good is purchased by country A from country B for resale and is routed through country C (without crossing the frontier of A), “it is customary, expedient and correct to exclude these goods from recorded external trade [of both A and C].” (UN-6, p. 25.)*

According to the U.N., countries using the general trade system may also report separately the part of imports destined for domestic use. This is called “retained imports” and is usually calculated by subtracting re-exports (including “nationalized” goods) from general imports. Retained imports are likely to be smaller than special imports since the latter include goods purchased originally for domestic use but sebsequently re-exported. A recent U.N. source in fact refers to three systems of recording trade: special, general, and semispecial, the latter defined as general trade less all re-exports (UN-1, P. 16).

*For country C the transaction is transit trade, to be excluded from general trade by explicit recommendation (UN-6, pp. 24-25). For country A, the recommendation to exclude the transaction is implicit: “In the general trade system, the … record is made when goods move across the national boundary .…” (Ibid., p. 23.) And elsewhere, “in the case of imports, (the recording country] is the country from which the goods were first shipped to the reporting country.” (Ibid., p. 61.)

The U.N. recommends that countries should, if practicable, report general trade:

In order to maximize the usefulness of international trade statistics, the coverage of the statistics should be sufficiently wide to take account of all merchandise entering or leaving a country from or to another country with the exception of goods simply being transported through the territory. [UN-6, p. 3.]

CEMA's Standing Commission for Statistics also recommends that member countries record general trade (U-3, September 1968, p. 46), but the discrepancy noted above in defining re-exports makes the two recommendations unlike in detail.

East European Procedures

Data in this Compendium are believed to be consistent with the general trade concept as defined by CEMA, with the following known or possible exceptions, by country:

East Germany: The system of recording total trade and trade with the U.S.S.R., total CPEs, MDCs, and LDCs, 1946-48, is unknown, but is probably special trade; trade with individual East European countries other than the U.S.S.R., 1946-52, is unknown, but is probably special trade.

Rumania: For 1948-49 and 1951-54, the system of recording trade is unknown.

Yugoslavia: Not a CEMA member, Yugoslavia has consistently reported special trade.

Estimated Re-Export Trade of East Germany and Hungary

Re-export trade is of considerable interest to foreign trade specialists because, among other reasons, international comparisons of trade relative to national income are more meaningful if based on special trade, with re-exports excluded. Of special interest for East Europe, degrees of actual bilateralism can best be estimated by making allowance for financing imports with re-exports, a practice to which East European countries have been known to resort.* Primarily for these reasons, Haberler has suggested that the re-export trade of CPEs “should be made the object of systematic-factual-statistical studies.”**

CEMA countries do not, as a general rule, publish re-export figures. But for two CEMA members, East Germany and Hungary, an attempt can be made to estimate re-export trade indirectly.

*Wiles distinguishes natural re-exports (“a big entrepot like London buys and sells raw materials and semifabricates”) and financing re-exports (“A wants convertible currency more than anything B can sell for [A'sj domestic use So A buys one of B's principle exports … and resells it. Thus A/B trade is both bilateral and reciprocal; yet A acquires the convertible currency she wants.”) p. 261, with illustrative cases.) One may add that re-exports can arise also from the need to dispose of unwanted commodities acquired as a result of bilateral agreements with inconvertible-currency countries.

**“Theoretical Reflections on the Trade of Socialist Economies,” in W-2, p. 45.

East German Re-Exports: The country's foreign trade statistics by geopolitical regions until 1966 (and by countries until 1960) have corresponded approximately to special trade; more recently, published statistics refer to general trade (CEMA definition). In 1970, revised statistical series of trade by regions were published retroactively to 1949, corresponding to general trade. By assuming, as a first approximation, that the revision was due simply to a change in the definition, differences between the two series (available for 1950-66 by regions and for 1953-60 in trade with the U.S.S.R.) would permit estimating reexport trade, which is shown in Table D-1 on the following page. However, when the resulting estimates are compared with those obtained for Hungary (see last paragraph of this page), we find that, in addition to switching from the special trade to the general trade system, some unexplained statistical revision may also have occurred.

Special imports (Wareneinfuhr) consist of commodities for domestic use, the standard definition. However, both special imports and special exports exclude value added by “improvement and repair trade” (Lohnveredlungen) as well as provisions for ships and aircraft, while general trade includes both of these items. Thus, estimated re-exports also include trade in these categories and could also reflect revisions for reasons not explained.

The estimated dollar value of re-export trade, its total, distribution by regions, and relative importance are shown in Table D-1.

Hungarian Re-Exports: Published foreign trade series cover general trade, but from secondary Hungarian sources it is possible to construct a special trade series for 1951-68. For an eleven-year subperiod, it is also possible to show actual foreign exchange costs of transporting Hungarian imports up to the Hungarian border and Hungarian exports to a “contract” point (outside Hungary's boundaries). Since Hungary is the only CEMA country reporting imports on a c.i.f. basis, transport and related costs need to be subtracted to make Hungarian statistics comparable with those of other CEMA countries. The respective series are presented in Table D-2.

Re-Export Estimates Compared: Comparison of the calculated relative shares of re-export trade, Column (2) of Table D-1 and Column (6) of Table D-2, would indicate that re-exports are more important for Hungary than for East Germany. Considering that East Germany's calculated re-export trade would even appear to be upward biased due to the inclusion of additional commercial items, its reexport figures are surprisingly low. While it cannot be excluded that observed differences between the estimated re-export trade of the two countries reflect actual differences, the gap strongly suggests the possibility that East Germany's calculated re-exports are incomplete because the estimates derived here are affected by unexplained adjustment in the revised series. This possibility is suggested also by the fact that while in Hungary's total trade the value of imports for reexport and re-exports parallel closely, the same series in East Germany's total trade diverge, estimated imports for re-export being substantially larger each year since 1958.

Table D-1

East German Imports for Re-Export, and Re-Exports,
by Region or Country of Provenance or Destination, 1950-1966
(Millions of Current Dollars)

Source: G-1, various issues. Data shown in valuta-marks converted to dollars at the official exchange rate.

Table D-2

Hungarian Special Trade, General Trade, Imports for Re-Export, and Re-Exports, 1951-1968 (Millions of Current Dollars)

a. Special trade and general trade concepts correspond to U.N. definitions, except for the definition of re-exports, which follows CEMA practice (H-11, 1963, p. 411).

b. The difference between “contract” and “border” parities are transportation costs incurred in foreign exchange; in imports transport costs are added to arrive at “imports c.i.f. Hungarian border” and in exports transport costs are subtracted to obtain “exports f.o.b. Hungarian border.”

c. Includes only transport costs incurred in foreign exchange. See footnote above.

Sources: Columns (1) and (2): H-1, 1956, p. 183; H-1, 1957, p. 219; H-8, December 1964, p. 1239. Column (3): H-1. various issues as quoted in H-9, p. 84, and H-10, no. 105, p. 204, or derived by summing columns (1) and (2). Column (4); calculated as the difference between columns (5) and (3). Column (5): H-1, 1967, p.219; H-1, 1968, pp. 258-59

Coverage of Merchandise Trade

United Nations Recommendations

The coverage of categories singled out by U.N. recommendations relate to goods to be included in statistics on merchandise trade, excluded from the statistics, and separately recorded. Where such recommendations are contrary to the general practice of most Western countries, this fact is noted here. (Recommendations are given in UN- 6, pp. 3-15; reporting practices of most Western countries are shown in UN-8, 1968, pp. 7-8.)

To be included:

(1) Transactions in ships and aircraft. (U.N. concern with the recording of these items focuses on changes in country of registration unaccompanied by actual movements of the vehicle from one country to another, i.e., shifts in flag of convenience.) Present practice of most Western countries is to include new ships and aircraft but to exclude transfer of ownership of second-hand ships and aircraft.

(2) Trade on government account, including civilian and military foreign aid shipments (whether or not under loan or grant); war reparations; and military goods (except if consigned to own armed forces abroad).

(3) Trade on the account of foreign concessions.

(4) Trade in gas, electricity, and water.

(5) Parcel post: No recommendation is made as to coverage, except that important commodities of light weight but high value, such as diamonds and gems, should be recorded in full.

(6) Silver ore and concentrates, bullion, unissued coin, scrap, and partly worked manufactures of silver.

To be excluded:

(7) Monetary gold.

(8) Securities, bank notes, and coins in circulation.

(9) Goods consigned by a government to its armed forces and diplomatic representatives abroad.

(10) Temporary trade, such as tourists' effects; animals for racing and breeding; goods for exhibition only; returnable samples and containers; goods shipped for temporary storage; vehicles carrying goods or passengers between countries.

(11) Transit trade.

To be separately recorded:

(12) Nonmonetary gold.

(13) Bunkers, stores, etc., supplied to foreign vehicles or acquired abroad.

(14) Fish and salvage sold abroad off national vessels or purchased in national ports from foreign vessels.

(15) Repairs to ships and aircraft (value added only).

(16) Improvement and repair trade: The general practice of Western countries is to include inward and outward movement of such trade at full value and as regular merchandise trade. Goods moving abroad for improvement or repair are described as “passive,” and those correspondingly received from other countries as “active.” Returned goods are most frequently recorded by the country engaged in active repair trade as re-exports.

(17) Goods on lease (e.g., computer facilities, films). Present practice is to include such transactions at full value as regular merchandise trade.

East European Procedures

Comprehensive information on the coverage of merchandise trade is difficult to obtain for most East European countries, particularly for 1946-55. Some do not disclose any information and others report selective information only. Notes to trade statistics may not indicate whether the stated coverage is also applicable to earlier years. Revised series may be published without explaining fully the nature of the revision. Since the statistical practices of most East European countries differed from one another considerably during the 1940s and 1950s, information available for one country may not shed light fully on the statistical coverage of others. Since the early 1960s a move toward more uniform statistical coverage within CEMA countries can be discerned, but differences continue as noted.

Table D-3 summarizes available information on the coverage of merchandise trade by East European countries; national practices and CEMA recommendations are compared with U.N. recommendations. Additional details and the documentation can be found in the notes to the Table.

The following main findings emerge:

(1) Trade in new ships, aircraft, etc., is included in the data for all East European countries. How used ships, aircraft, etc., are treated statistically is not known, but trade in them is believed to be small to insignificant.

(2) The CEMA recommendation on government trade is to exclude foreign aid, i.e., “goods exported as free assistance or as gifts,” from merchandise trade (U-3, September 1968, p. 47). But for CPEs especially, it is difficult to separate strictly commercial from strictly governmental transactions. For instance, goods may be shipped on credit without obtaining full commercial reciprocity, a possible form of foreign aid.

Table D-3

Coverage of East Europe's Foreign Trade Statistics
Compared with U.N. 1970 Recommendations

✓ = Agrees with U.N. recommendation

incl. = Included in merchandise trade

n.a. = Not available

excl. = Excluded from merchandise trade

1. Made by the CEMA Standing Commission for Statistics (U-3, September 1968, pp. 4648), and probably during the early 1960s.

2. Included in CTN (commodity subgroup 192).

3. Foreign aid is excluded, while trade in military goods is probably excluded.

4. Included in CTN (commodity group 23).

5. Private parcel post excluded.

6. Included in CTN (commodity group 24).

7. If a contractor's raw material is processed and the finished product is re-exported to the country supplying the raw material, then both parties take into account processing costs only. Value of raw materials and finished goods moving across boundaries is recorded separately but is not included in commodity trade. (U-3, September 1968, p. 47.) Apparently, only value added is included in merchandise trade.

8. Movie films for rent are excluded, while films purchased and sold, including license payments, are included.

9. For 1951-54, see General Note to Tables I.M.B and I.X.B in Part Three.

10. Foreign aid is excluded since 1955 (B-6, p. 3, and B-2, 1968, p. 169). The United Nations document stating that foreign aid has been included since 1953 appears to be erroneous. (UN-8, 1968, p. 121, and previous issues.)

11. For 1946-52 coverage is not known.

12. Value added (C-3, 1969, p. 408).

13. For 1946-48 coverage is not known; it is probably special trade, which is known to exclude re-exports, value added by improvement and repair trade, and provisions for ships and aircraft. (G-1, 1967, p. 383.) The years 1949-59 are also exceptions, but only for trade with individual East European countries other than U.S.S.R. For 1949-52, probably special trade; for 1953-59, special trade.

14. Private (“gift”) parcel post is excluded from both general and special trade. (G-1, 1967, p.383.) Earlier source states that all postal packages were excluded from special trade. (W-10, p. 59.)

15. Apparently includes value added only.

16. Includes “government aids granted in form of goods, as far as such aids are based on international treaties” (H-ll, 1963, p. 411). United Nations sources state that such aid is included in imports only. (UN-8, 1968, p. 368.) Also includes deliveries under war reparations (H-11, 1963, p. 411), but only since 1953 (UN-8, 1968, p. 368). Prior to 1958, goods destined for Hungarian diplomatic agencies stationed in Hungary are included. (UN-8, 1968, p. 368.)

17. Excludes parcel post since 1953. (UN-8, 1968, p. 368.)

18. Included in merchandise trade until 1958 at the full value of goods. Since 1959 only value added has been included, calculated (if applicable) as “the value of the raw material charged as fee for job-work.” (H-11, 1963, p. 411.) U.N. sources state that inclusion at value added commenced in 1958 (UN-8, 1968, p. 368), a possible error.

19. Excludes “goods exported or imported free of charge” (P-8, 1967, p. 7). Also excludes “war reparations, purchases of surplus army property, purchase and sale of other military material.” (UN-8, 1968, p. 677.)

20. Excludes “private postal matters” (P-8, 1967, p.7); U.N. sources state that “beginning 1954, parcel post” is excluded (UN-8, 1968, p. 677).

21. Included in merchandise trade (P-8, 1967, p. 7); according to U.N. sources, since 1954 only (UN-8, 1968, p. 677).

22. Value added (P-8, 1967, p. 7).

23. For 1948-49 and 1951-54, coverage is not known. See General Note to Tables I.M.R and I.X.R in Part Three.

24. Includes imports “deriving from the Tripartite [U.S., French, U.K.] Economic Aid; since 1957 imports on account of long-term credit on the agricultural surpluses of the USA, as well as import on account of reparations and restitutions, of subsequent claims and grants from various international organizations.” (Y-2, 1960, p. 8.) According to U.N. sources, the above items have been included only since 1951 (UN-8, 1968, p. 926). Also includes “imports of foreign army surplus property.” (Ibid.)

25. Includes parcel post (Y-2, 1966, p. 9); “Prior to 1954, parcel post of non-commercial character” was excluded (UN-8, 1968, p. 926).

26. Included since 1958 (ibid).

27. Excluded from merchandise trade since 1960; shown separately. “Active improvement trade was included in the data on regular exports and imports” (Y-2, 1966, p. 9) “at full value” (Y-2, 1960, p. 8). Excluded prior to 1951 (UN-8, 1968, p. 926).

Hungary is apparently an exception to the CEMA foreign-aid rule, since it includes (though possibly in imports only) “government aid granted in the form of goods… [if] based on international treaties” (H-11, 1963, p.411). Until about 1955 Bulgaria reportedly also included foreign aid in its merchandise trade. Yugoslavia (not a CEMA member) has included goods received (supplied?) as aid since 1951.

Aid received by CEMA countries after the war from UNRRA was apparently excluded from merchandise trade, as stated explicitly by the U.N. for Czechoslovakia, Hungary, and Poland (UN-8, 1968). Reparations deliveries are also believed to have been excluded, except in the case of Hungary, which reports that reparations have been included in merchandise trade since 1953 (H-11, 1963, p.411). As for the purchase or sale of military hardware, there is reason to believe that they are included in the trade statistics because CEMA recommends that “all goods and material values exported and imported on a commercial basis are the object of foreign trade registration” (U-3, September 1968, p. 46). To be sure, military items are not listed in the published CEMA Trade Nomenclature, nor are they shown in national publications. But since detailed commodity breakdowns, when published, are not exhaustive, the possibility that military items are included in reported trade totals cannot be excluded.

(1) CEMA recommendations (believed to have been made during the early 1960s) are more or less being followed by member countries, although statistical practices are still far from uniform. Except for government trade, as noted above, transactions that CEMA suggests should be included or excluded from merchandise trade correspond quite closely to the 1970 U.N. recommendations. The major difference between U.N. and CEMA standards concerns transactions that the U.N. suggests should be reported separately (e.g., improvement and repair trade; bunkers, stores, etc., supplied to foreign vehicles; and repairs to ships and aircraft) but that CEMA recommends should be included with merchandise trade. However, CEMA's Standing Commission for Statistics recently recommended that, as of 1971, certain invisible commercial transactions — including repairs as well as leasing, items which according to the U.N. should be reported separately — are to be shown separately also by CEMA countries, in a new appendix to the CTN (see Appendix A above for further details).

In sum, it appears that the trend is toward greater statistical standardization, not only in the West and in CEMA but also between CEMA and non-CEMA countries. Important differences do remain; moreover, statistical coverage during earlier years may not always correspond to the known coverage of the same country for more recent years.

Transport and Related Expenses

The U.N. recommends that valuation be c.i.f. imports and f.o.b. exports, border of reporting country. C.i.f. valuation is preferred over f.o.b. in imports because the former often or typically represents the total foreign exchange cost of imports. A further advantage of using different systems of valuation for imports and exports is that the difference between them provides an independent estimate of distribution costs (UN-6, pp. 46-47).

All East European countries except Hungary and Yugoslavia report both imports and exports f.o.b. border of supplying country. Hungary and Yugoslavia record imports c.i.f. and exports f.o.b. (Hungary's estimated foreign exchange cost of transporting imports to the Hungarian border can be found in Table D-2 above.) An advantage of having both imports and exports f.o.b. border of the supplying country is that it is possible to estimate imports of a country from the export data of its trade partner.

Identification of Partner Country

United Nations Recommendations

Differences among countries as to identifying trade partners is believed to be the most important cause of incomparability of mirror statistics. (UN-6, p. 58.) Therefore, the U.N. recommends recording trade on the basis of country of consignment, which is defined for imports as the country from which goods are directly received and for exports as the country to which goods are directly sent. (UN-6, p. 63.) This system offers the greatest possibility of attaining accuracy and comparability of trade flows reported independently by trading partners.* It is recognized that mirror statistics could still remain inconsistent, inter alia, because (1) goods are exported to countries compiling statistics solely according to the special trade system; (2) the coverage of re-exports differs; (3) goods are dispatched in transit through a second country to optional destinations; and (4) cargo shipped to one country is diverted on the high seas to another destination.

The U.N. appreciates that some countries might find it desirable to record statistics on alternative bases. For imports, this could be according to country of production (last country in which goods received appreciable physical transformation) or country of purchase (residence of seller); for exports, according to country of consumption (last known destination) or country of sale (residence of buyer). If the exact method used by the country reporting trade statistics is not known, for imports, the term country of provenance, for exports, the term country of destination is used (UN-8, 1968, p. 8).

Present practices of Western countries vary considerably with respect to the use of these definitions.

East European Procedures

Different sources frequently give conflicting definitions of trade partner as used by individual countries. The problem seems to be, at least in part, that the definition of trade partner given in East European documents, when translated literally, may not correspond in intended meaning to the U.N. terminology. The following information is available by country:

U.S.S.R.: Trade partners are defined as countries of production and consumption. This is apparently the correct U.N. interpretation of U.S.S.R. methodology (UN-8, various issues). Literal translation (by the U.S.S.R.) reads “origin-destination” or “origin-designation” (17-3, September 1968, p. 47, and U-2, January 1971, p. 49), but the intended meaning is clarified by the statement:

*Identical definitions but a slightly different terminology are employed by another U.N. publication, where, for imports, the country of first consignment is defined as the country from which goods were first directed to the importing country, and for exports the country of last consignment is the last known destination to which goods are sent (UN-8, 1966, p. 8).

In most countries, a country of exports is defined as a country of destination, … of imports … as a country of origin. This definition is used by the USSR and some [!] other CEMA countries .… Along with this most widespread definition … there is also the registration … according to countries having a trade agreement or countries where the importer (exporter) lives or countries from or to which goods are delivered. [U-3, September 1968, p. 47.]

Bulgaria: Country of import is “the country from which the commodities are bought,”and country of export is “the country to which the commodities are sold” (B-2, 1968, p. 169, and other issues of the same publication). This appears to correspond to the U.N. definition of countries of purchase and sale. (See also UN-8, 1967, p. 128, and other issues for a similar interpretation.)

Czechoslovakia: Statistics are recorded by countries of purchase and sale. Explanatory notes in the yearbooks state that the criterion is the residence of the contracting party (C-3, 1969, p. 408; also previous issues). The U.N., however, defines Czechoslovakia's trade partners as “countries of consignment” (UN-8, 1967, p. 230, and other issues), an apparent error.

East Germany: Trade partners are defined as countries of purchase and sale (G-2, pp. 93-94). The U.N. indicates provenance and destination (UN-8, 1967, p. 304, and other issues) because in the East German statistical yearbooks (G-l)no definition can be found.

Hungary: Based on official statements, trade partners are defined as countries of purchase and sale: “A partner country is the country where the party concluding a civil law contract with the Hungarian party (first buyer and last purchaser) is domiciled' (H-11, 1963, p.411). The U.N. identifies Hungary's trade partners as” countries of production and last consignment (UN-8, 1968, p. 374, and other issues), an apparent error; since 1971 by countries of production and consumption (H-2, 1972, no. 2-3, p. 51).

Poland: Partners are defined as countries of purchase and sale.

Rumania: Trade partners are given as countries of purchase and sale (UN-8, 1968, p. 700).

Yugoslavia: Imports are by country of production, exports by country of consumption:

The country in which the commodities have been manufactured in the condition as imported has been considered as country of origin. Basic raw materials … have been shown by the producing countries even if the commodities have been shipped from the warehouses of some other country .… The country in which the commodities are consumed or have to undergo further processing is considered as the country of destination. [Y-2, 1969, p. 9, and other issues of the same publication.]

Albania: Definition of trade partner is not known, so provenance and destination are the terms used in international publications (UN-8, 1968, p. 45).

To summarize, all European CEMA countries except the U.S.S.R. apparently record statistics according to country of purchase or sale, while the U.S.S.R. keeps books according to country of production and consumption, a practice which is followed by Yugoslavia also.

APPENDIX E

The Case of the Unspecified U.S.S.R. Residuals

The available distributions of U.S.S.R. trade by known partner countries or by commodity categories leaves unspecified a significant proportion of the total. The problem of these unspecified residuals needs to be taken into consideration in any economic analysis of the geographic or commodity composition of U.S.S.R. trade, since the residuals are not randomly distributed.

Exports to Less Developed Countries

U.S.S.R. foreign trade yearbooks (U-1) report trade by both individual countries and groups of countries, the latter classified as “socialist” (“Total CPEs” in our terminology), “industrial capitalist” (MDCs), and “developing” (LDCs) countries. The composition of the LDC group follows the standard U.N. description for Economic Group II, except for the inclusion of Turkey and the exclusion of Cuba (U-3, September 1968. p.48). Each year the sum of reported trade with these three groups adds precisely to the independently given total of exports or imports; moreover, the sum of reported trade with individual CPE and MDC countries adds to the independently given group totals. The one exception is in the case of U.S.S.R. exports by country to LDCs, for which the sum of individual national exports is substantially less each year than the independently given total exports to the group as a whole. (There is no corresponding discrepancy in imports.) In 1970 this unspecified residual amounted to $792 million, or close to two- fifths of total reported U.S.S.R. exports to LDCs and nearly 7 percent of global exports.

Calculated unspecified residuals are shown in Table E-1 for 194670. It is only since 1955, however, that the residuals can be meaningfully derived because only since then has trade with practically all countries been reported in the foreign trade yearbooks.*

Commodity Composition of Total Trade

U.S.S.R. tabulations of commodity composition for total trade (Tables II.CTN.M.U and II.CTN.X.U) and trade by country (Tables III. CTN.M.U-B through CEMA and III.CTN.X.U-B through CEMA) reveal that in each year a certain proportion of trade (usually larger for exports than for imports) is not specified according to any CTN category. Commodity composition cannot be accurately analyzed without knowledge of which CTN categories are fully specified or which are incomplete, or without information on the composition of the unspecified residual.

*For each year, trade with individual countries is reported in several issues of U-1, among which there may be slight differences in the list of countries included. All issues of U-1 were consulted to obtain maximum country coverage.

Table E-1

U.S.S.R. Trade with Less Developed Countries:
Comparison of LDC Group Total with Sum of Trade Reported by Component Countries, 1946-1970
(Millions of Dollars)

*Total of Sums as percent of Given LDC Total.

**Given LDC Total less Total of Sums.

Source: 1946-54: 17-2,1955-70: U-2 and U-1.various issues. Devisa-rubles were converted here to dollars at the official ruble/ dollar exchange rate.

Table E-2

U.S.S.R. Imports by CTN Categories: Comparison of Percentage
Composition Given (G) and Summed (S), 1946, 1950, 1955, 1958-1969

*Trade in CTN Category 6 is insignificant.

**Total trade in CTN 24, Metallic Ores and Concentrates, is no longer reported. Only one component, CTN 24216, Alumina, is given.

Source: G columns, 1946, 1950, 1955, 1960, 1966: U-2; 1958, 1963-64: U-1, 1964; 1959, 1961-66: U-1, 1959-1963, U-1, 1965, and U-1, 1966; 1967-68: U-1, 1968; 1969: U-1, 1969. S columns, 1946-69: based on Statistical Table II.CTN.M.U.

Table E-3

U.S.S.R. Exports by CTN Categories: Comparison of Percentage
Composition Given (G) and Summed (S), 1946, 1950, 1955, 1958-1969

*Trade in CTN Category 6 is insignificant.

**CTN 25, Nonmetallic Minerals, Clays, and Earth, are not included. In 1966 and 1967, this commodity group represented 1.1 percent of U.S.S.R. exports.

Source: G columns: same as in Table E-2; S columns: Statistical Table II.CTN.X.U.

Recent issues of the U.S.S.R. foreign trade yearbook (U-1) indicate, for selected years, commodity structure —in the sense of aggregated commodity groupings — as a percent of total imports or exports. The sum of the percentages each year is less than 100. Although no CTN numbers are given with the commodity groupings, approximate (one- and two-digit) CTN numbers can be assigned. To obtain additional information on the unspecified residual, the percentages given in the yearbooks (here coded G for “given”) and those obtained from the Data Bank after summation of two- and three-digit entries (here coded S for “summed”) have been compared by CTN categories in Tables E-2 and E-3 for imports and exports, respectively.

Table E-4

U.S.S.R. Trade: Comparison of Alternative Estimates of Percentage Composition by CTN Categories, 1946. 1950, 1955, 1958-1969

The independent pairs of percentages obtained in this manner reveal a consistent pattern for 1946, 1950, 1955, and 1958-69, one that is much the same in imports and exports. For some categories the percentages from the two sources are always identical (except for rounding); for others, the percentages from one source are either always larger or always smaller. Differences between the G and S percentages can be explained satisfactorily in each case by examining the underlying commodity coverage in the two sources, as summarized in Table E-4.

From Table E-4 it appears that accepting for each CTN category the larger of the two percentages, G or S, permits reduction of the unspecified component, compared to taking either the G or the S percentages alone. For imports between 1958 and 1969, the average proportion of trade specified by CTN categories can be increased in this way to 97.6 percent (compared to 93.5 and 96.4 percent on the basis of G or S estimates alone); in exports, the average rises to 88.3 percent (compared with 86.9 and 86.8 percent). The new figures are shown in column 17 of Tables E-2 and E-3, respectively.

Two recent publications from East Europe enable us to go a step further. They provide an exhaustive percentage breakdown of commodity composition for each CEMA country, including the U.S.S.R., for selected years according to five new aggregative commodity groupings. These groupings are identified by descriptive titles only, but it appears that they correspond to certain combinations of CTN categories, as shown in Table E-5. The five new groupings may indicate that CEMA or some CEMA countries have introduced alternative or revised definitions of CTN broad divisions.

Table E-5

CTN Classification: Titles of “Alternative” Broad Divisions and Corresponding One-Digit CTN Categories

Source: CEMA -2, pp. 356-59, 446-47; P-9, p. 114.

Table E-6

U.S.S.R. Trade by CTN Categories: Comparison of Percentage Composition Shown in Various Sources

*Discrepancy with percent shown in column (3) is explained by omission of CTN 25, as shown in footnote to Table E-3.

Source: CEMA columns, 1950, 1955, 1960, 1965, and 1969: CEMA-2, pp. 357, 359; 1967: P-9, p. 114.

Table E-6 compares the larger of either the given G or summed S percentages (from Tables E-2 and E-3) with the newly published “CEMA” percentages. Comparisons in Table E-6 confirm the previous findings of Tables E-2 and E-3 and provide additional information summarized in Table E-7.

A reconstructed distribution of U.S.S.R. imports and exports by CTN categories is shown for 1950, 1955, 1960, 1965, 1967, and 1969 in Table E-8. There still remains a significant component unspecified, which amounts to about $250 million (2.4 percent) of imports and $1.4 billion (11.9 percent) of exports in 1969. The fact that the proportions are considerably larger in exports than in imports suggests that all or a substantial part of the unspecified residual is comprised of commercially traded strategic items.

Table E-7

U.S.S.R. Trade: Comparison of Alternative Estimates of Percentage Composition by CTN Categories, 1950, 1955, 1960, 1965, 1967, 1969

Table E-8

U.S.S.R.: Imports and Exports by CTN Categories Plus Unspecified Trade, 1950, 1955, 1960, 1965, 1967, 1969 (as Percent of Total)

Source: Imports, CTN 1, 3, 4, 6-8, and 9: Table E-2; CTN 2: Table E-6; CTN 5: Table E-2, col. (10) plus (col. 7 less col. 8 of Table E-6). Exports, CTN 1, 3, 4, 6-8, and 9: Table E-3; CTN 2: Table E-6; CTN 5: Table E-3, col (10) plus (col. 7 less col. 8 of Table E-6).

APPENDIX F

Trade with United States, Canada, Japan, and Australia

The statistics below supplement Series IV in Part Two on trade between individual East European countries and West Europe as a group. This essentially rounds out the MDC group, which, besides West Europe, includes the United States, Canada, Japan, Australia, New Zealand, and the Union of South Africa. East Europe's trade with the last two countries has been insignificant.

The present separate appendix has been prepared because in general commodity composition by SITC categories has been available for more recent years for these four countries than for West Europe (since 1961 for the United States and Canada; since 1964 for Japan; in the case of Australia only total trade can be shown).

Trade with the MDC group as reported by individual East European countries in Statistical Series I and “mirror” statistics compiled from trade-partner sources (i.e., the sum of trade with East European countries reported by West Europe, the United States, Canada, and Japan) will not be in exact agreement. There are differences in the systems of reporting trade (“general” vs. “special” systems), in the treatment of transportation and related charges (c.i.f. vs. f.o.b.), in identification of trade-partner countries (origin of imports and destination of exports), and in the dates at which trade is recorded as goods move from origin to destination.

Table F-1

United States Trade with U.S.S.R.: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current Dollars)

* Annual average.

Source: 1946-60: GP-4, pp. 21-22, 32-33; 1961-69: W-20, annual issues.

Note: General trade; country of production (imports) and consumption (exports). Figures may not add to total because of rounding.

Table F-2

United States Trade with Bulgaria: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current Dollars)

* Annual average.

Source: 1946-60: GP-4, pp. 21-22, 32-33; 1961-69: W-20, annual issues.

Note: General trade; country of production (imports) and consumption (exports). Figures may not add to total because of rounding.

Table F-3

United States Trade with Czechoslovakia: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current Dollars)

* Annual average.

Source: 1946-60: GP-4, pp. 21-22, 32-33; 1961-69: W-20, annual issues.

Note: General trade; country of production (imports) and consumption (exports). Figures may not add to total because of rounding.

Table F-4

United States Trade with East Germany: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current Dollars)

* Annual average.

Source: 1956-60: GP-4, pp. 21-22, 32-33; 1961-69: W-20, annual issues.

Note: General trade; country of production (imports) and consumption (exports). Figures may not add to total because of rounding.

Table F-5

United States Trade with Hungary: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current Dollars)

* Annual average.

Source: 1946-60: GP-4, pp. 21-22, 32-33; 1961-69: W-20, annual issues.

Note: General trade; country of production (imports) and consumption (exports). Figures may not add to total because of rounding.

Table F-6

United States Trade with Poland: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current Dollars)

* Annual average.

Source: 1946-60: GP-4, pp. 21-22, 32-33; 1961-69: W-20, annual issues.

Note: General trade; country of production (imports) and consumption (exports). Figures may not add to total because of rounding.

Table F-7

United States Trade with Rumania: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current Dollars)

* Annual average.

Source: 1946-60: GP-4, pp. 21-22, 32-33; 1961-69: W-20, annual issues.

Note: General trade; country of production (imports) and consumption (exports). Figures may not add to total because of rounding.

Table F-8

United States Trade with Yugoslavia: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current Dollars)

* Annual average.

Source: 1946-60: GP-4, pp. 21-22; 1961-69: W-20, annual issues.

Note: General trade; country of production (imports) and consumption (exports). Figures may not add to total because of rounding.

Table F-9

United States Trade with European CEMA: Total, 1946-1969, by SITC Categories, 1961-1969 (Millions of Current Dollars)

* Annual average.

Source: Tables F-1 through F-7.

Note: General trade; country of production (imports) and consumption (exports). Figures may not add to total because of rounding.

Table F-10

Canada’s Trade with U.S.S.R.: Total, 1946-1969, by SITC Categories, 1961-1969 (Millions of Current U.S. Dollars)

* Annual average.

Source: 1946-60: GP-2, 1954,1957-58, and 1962-63 (Canadian $ converted to U.S. $ at the average conversion coefficient shown for each year in UN-8, 1966, p. 138); 1961-69: W-20, annual issues.

Note: Special trade; country of consignment (imports and exports). Figures may not add to total because of rounding.

Table F-11

Canada's Trade with Bulgaria: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current U.S. Dollars)

* Annual average.

Source: 1946-60: GP-2, 1954,1957-58, and 1962-63 (Canadian $ converted to U.S. $ at the average conversion coefficient shown for each year in UN-8, 1966, p. 138); 1961-69: W-20, annual issues.

Note: Special trade; country of consignment (imports and exports). Figures may not add to total because of rounding.

Table F-12

Canada's Trade with Czechoslovakia: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current U.S. Dollars)

* Annual average.

Note: Special trade; country of consignment (imports and exports). Figures may not add to total because of rounding.

Table F-13

Canada's Trade with East Germany: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current U.S. Dollars)

* Annual average.

** 1952-55 average.

Source: 1946-60: GP-2, 1954, 1957-58, and 1962-63 (Canadian $ converted to U.S. $ at the average conversion coefficient shown for each year in UN-8, 1966, p. 138); 1961-69: W-20, annual issues.

Note: Special trade; country of consignment (imports and exports). Figures may not add to total because of rounding.

Table F-14

Canada's Trade with Hungary: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current U.S. Dollars)

* Annual average.

Source: 1946-60: GP-2, 1954, 1957-58,and 1962-63 (Canadian $ converted to U.S. $ at the average conversion coefficient shown for each year in UN-8, 1966, p. 138); 1961-69: W-20, annual issues.

Note: Special trade; country of consignment (imports and exports). Figures may not add to total because of rounding.

Table F-15

Canada's Trade with Poland: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current U.S. Dollars)

* Annual average.

Source: 1946-60: GP-2, 1954, 1957-58,and 1962-63 (Canadian $ converted to U.S. $ at the average conversion coefficient shown for each year in UN-8, 1966, p. 138); 1961-69: W-20, annual issues.

Note: Special trade; country of consignment (imports and exports). Figures may not add to total because of rounding.

Table F-16

Canada's Trade with Rumania: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current U.S. Dollars)

* Annual average.

Source: 1946-60: GP-2, 1954, 1957-58,and 1962-63 (Canadian $ converted to U.S. $ at the average conversion coefficient shown for each year in UN-8, 1966, p. 138); 1961-69: W-20, annual issues.

Note: Special trade; country of consignment (imports and exports). Figures may not add to total because of rounding.

Table F-17

Canada's Trade with Yugoslavia: Total, 1946-1969, by SITC Categories, 1961-1969
(Millions of Current U.S. Dollars)

* Annual average.

Source: 1946-60: GP-2, 1954, 1957-58,and 1962-63 (Canadian $ converted to U.S. $ at the average conversion coefficient shown for each year in UN-8, 1966, p. 138); 1961-69: W-20, annual issues.

Note: Special trade; country of consignment (imports and exports). Figures may not add to total because of rounding.

Table F-18

Canada's Trade with European CEMA: Total, 1946-1969, by SlTC Categories, 1961-1969
(Millions of Current U.S. Dollars)

* Annual average.

Source: Tables F-10 through F-16.

Note: Special trade; country of consignment (imports and exports). Figures may not add to total because of rounding.

Table F-19

Japan's Trade with U.S.S.R.: Total, 1950-1969, by SITC Categories, 1964-1969
(Millions of Current Dollars)

* Annual average.

Source: 1950-59: GP-3, 1962, p. 255; 1960-63: GP-3, 1970, p. 290; 1964-69: W-20, annual issues.

Note: General trade; country of production (imports) and last consignment (exports). Figures may not add to total because of rounding.

Table F-20

Japan's Trade with Bulgaria: Total, 1960, 1964-1969, by SITC Categories, 1964-1969
(Millions of Current Dollars)

Source: 1960: B-1, 1970; 1964-1969: W-20, annual issues.

Note: General trade; country of production (imports) and last consignment (exports). Figures may not add to total because of rounding.

Table F-21

Japan's Trade with Czechoslovakia: Total, 1958-1969, by SITC Categories, 1964-1969
(Millions of Current Dollars)

Source: 1958-63: C-2, 1964, pp. 104-105; 1964-69: W-20, annual issues.

Note: General trade; country of production (imports) and last consignment (exports). Figures may not add to total because of rounding.

Table F-22

Japan's Trade with East Germany: Total, 1951-1969, by SITC Categories, 1964-1969
(Millions of Current Dollars)

* Annual average.

Source: 1951-59: GP-3, 1962, p. 255; 1960-63: GP-3, 1970, p. 290; 1964-69: W-20, annual issues.

Note: General trade; country of production (imports) and last consignment (exports). Figures may not add to total because of rounding.

Table F-23

Japan's Trade with Hungary: Total, 1958-1969, by SITC Categories, 1964-1969
(Millions of Current Dollars)

Source: 1958-63: H-1, annual issues; 1964-69: W-20, annual issues.

Note: General trade; country of production (imports) and last consignment (exports). Figures may not add to total because of rounding.

Table F-24

Japan's Trade with Poland: Total, 1951-1969, by SITC Categories, 1964-1969
(Millions of Current Dollars)

* Annual average.

Source: 1946-1959: GP-3, 1962, p. 255; 1960-63: GP-3, 1970, p. 290; 1964-1969: W-20, annual issues.

Note: General trade; country of production (imports) and last consignment (exports). Figures may not add to total because of rounding.

Table F-25

Japan's Trade with Rumania: Total, 1960-1969, by SITC Categories, 1964-1969
(Millions of Current Dollars)

Source: 1960-63: GP-3, 1970, p. 291; 1964-69: W-20, annual issues.

Note: General trade; country of production (imports) and last consignment (exports). Figures may not add to total because of rounding.

Table F-26

Japan's Trade with Yugoslavia: Total, 1951-1969, by SITC Categories, 1964-1969
(Millions of Current Dollars)

* Annual average.

Source: 1951-60: Y-2, annual issues; 1964-69: W-20, annual issues.

Note: General trade; country of production (imports) and last consignment (exports). Figures may not add to total because of rounding.

Table F-27

Japan's Trade with European CEMA: Total, 1961-1969, by SITC Categories, 1964-1969
(Millions of Current Dollars)

* Excludes trade with Bulgaria.

Source: Tables F-19 through F-25.

Note: General trade; country of production (imports) and last consignment (exports). Figures may not add to total because of rounding.

Table F-28

Australia's Trade with Individual East European Countries and European CEMA, 1949/50-1969/70 (Millions of Current U.S. Dollars)

* July - June.

** Excludes Yugoslavia.

Source: GP-1, annual issues. Original data shown in Australian pounds or dollars (introduced in February 1966 with a value of 0.5 Australian pounds) converted to U.S. dollars (Australian $ = 1.12 U.S. $).

Note: General trade; country of production (imports) and consumption (exports). Figures may not add to total because of rounding.

APPENDIX G

Available East European Foreign Trade Series in Constant Prices, 1938-1970

Note: The series shown have been calculated on the basis of published or derived indices of physical volume or indices of price, based on East European country sources. The statistical series are presented in W-12.

a. CEMA plus (OCPEs and Yugoslavia).

b. Four CTN broad divisions for CPEs plus non-CPEs.

c. By SITC categories 0 through 8.

APPENDIX H

IDRC Soviet and East European Foreign Trade Data Bank, 1946-1969

The IDRC Soviet and East European Foreign Trade Data Bank consists of a computer system and accompanying programs which permit us to handle the data presented in Part Two and Appendix F above. In most cases it contains considerably more commodity detail than could be included in this volume.

The Data Bank also goes substantially beyond the Compendium in providing convenient, speedy, and versatile access to approximately a half million data entries. A major feature of the system for most purposes is its capability of retrieving information for immediate use without further programming, as described below. The system has been further designed so that updating and extensions can be done by routine procedures.

The Data Bank differs from a simple data base. Much more than a mere listing of numbers, it contains and identifies on request all titles, nomenclature definitions, and guides to coding, automatically providing the user a full set of word descriptions to match any data set that may be requested.

The computer programs and subroutines available with the data tape, a highly compact binary file, fall into four broad categories: Access Package, Auxiliary Routines, Special Programs, and Interactive Modes. A generalized diagram of the system, whose main components are described below, is shown in Figure 1.

1. The data tape is structured according to a tree hierarchy, as shown in Figure 2. In each year for which data are included, total imports and exports have been disaggregated by trade partners, by one of three commodity classification systems, or through a crossclassification involving both trade partners and commodity designations. Successively more detailed breakdowns are arranged on the data tape in branch-like fashion. Internally aggregated subtotals are always shown within each branch. Thus, even if values in an original source have been reported only at the subgroup (three-digit) and category (one-digit) levels, synthetic group (two-digit) values are obtained automatically. Briefly, the data tree is designed to be as complete as possible.

The user can identify special situations in the Data Bank via footnote-type codes attached to each entry. For instance, footnote code 8 signals that the value is an internally aggregated “synthetic” number, hence may not exhaust trade at the indicated level. Some twenty different footnote codes appear in the Data Bank, an especially significant auxiliary identification since often the data provided by East European countries have gaps or special characteristics that may differ among classification systems, countries, and periods. For example, in one period the U.S.S.R. may show trade in CTN 24, Metallic Ores and Concentrates, and CTN 26, Ferrous Metals, plus some three- and fivedigit components of CTN 27, Nonferrous Metals; in a subsequent period it may publish only a combined figure for CTN 24-27.

Figure 1
Data Bank System Arrangement

Figure 2
Data Structure

The data tape features a bootstrap mechanism: Any request for data automatically initializes a retrieval procedure, which includes inputs into core, a table of contents, written titles, and a set of internal instructions concerning options such as the currency units in which data could be provided.

2. The Access Package consists of a series of short subroutines which serve in “black box” fashion to connect user and data in a way which permits almost effortless random access to materials of interest. The user need only specify nomenclature, desired degree of detail, a code number for country or countries, whether import or export figures are involved, and trade partner. The data so requested are then loaded from the data tape into core storage, where they become available for viewing, together with relevant country names, commodity titles, footnotes, and other explanatory information. If a data set being requested is not in a Data Bank, this information too is immediately communicated.

3. The Auxiliary Subroutines consist of special processing programs available upon request for performing operations on data, once these have been supplied. Examples are presentations of data in tabular form (programs TABLE 1 or Table 2 in Figure 1), conversion from national devisa units to U.S. dollars (program USDOL), or standard time series analysis on any specified set of factors (program GROWTH). Generally, the Auxiliary Subroutines programs perform single functions on the full Data Bank or subsets and then return for the next command.

4. The Special Programs component consists of an expandable set of algorithms for such specific tasks as editing, presentation of data in special tables, or conversions between classification systems. It also includes analysis procedures, such as methods for calculating mirror ratios from trade-partner sources. These algorithms generally utilize the Access Package and building blocks from the Auxiliary Subroutines.

5. The most versatile service in the software is the Interactive Mode program. It permits the user to select individual items or vectors from the Data Bank in isolation from the remainder of the system, and then to print, compare, and manipulate them via teletype or typewriter terminal.

In general, therefore, the researcher could expect to have at his disposal the entire Data Bank file, the capabilities of Special Programs and Auxiliary Subroutines, and, together with these, a number of manipulating algorithms applicable to interactive analysis.

The data storage system and subroutines have been designed to accommodate future expansion of the Data Bank through updating or inclusion of additional data, such as trade by partner countries, or further commodity detail for years through 1969. It also appears possible to catalogue efficiently entirely new types of data, such as industrial production and input-output transactions, for purposes of being analyzed in conjunction with foreign trade proper.

Copies of the Data Bank or portions thereof can be made available to individual scholars and institutions at reasonable charges. For further information, write to:

DATA BANK

International Development Research Center

Indiana University

1005 East Tenth Street

Bloomington, Indiana 47401