1. In the past two decades there has been a pronounced change in the world economy. This phenomenon, known as globalization, has become an intense preoccupation of countries. Many are now conducting surveys on some aspects of globalization. International organizations, such as the OECD, have begun to study some manifestations of the phenomenon.2 However, no systematic description of the implications of globalization for the system of economic statistics has emerged. This paper introduces a few ideas on this subject.
2. One of the distinctive features of globalization is an increasing cross-national spread of products, markets, firms and factors of production, resulting in globally integrated production networks controlled by transnational corporations (TNC's). The speed of the changes has been accommodated by the reduction in trade and investment barriers, changing transportation and communication technologies and the rapid industrialization of some developing countries. Unfortunately, there is but a dim understanding of the depth of the changes triggered by these movements towards economic integration. In particular, the interaction of technology, industrial organization and corporate strategy, and legal frameworks is not properly understood.
3. One result of these movements is that the location of production is no longer fixed nor of particular long term significance. It is being driven by the comparative advantage of countries. Firms seek the places where the various functions performed along their production chains can be developed most efficiently. These places are not fixed in any one particular country; nor is the existence of national boundaries a factor in location decisions. A firm may conduct extraction activities in country A, fabricating activities in B, manufacturing of final goods in C, Research and Development in D and head office service functions such as accounting or marketing in E. Foreign direct investment and trade are used as complements to each other to achieve optimum operating configurations.
4. The existence of production networks involving more than one country results in a much larger number of complex international transactions than was the case hitherto. Further, these networks are not only widespread, involving many countries in a single production process at one point in time, they are also constantly changing. As firms search for the most efficient production network, the countries in which their production networks are involved at one point are seldom the same as those involved at a later point.
5. The current situation contrasts with the internationalization of economic activity which existed in earlier decades. Before the 1970's, a firm might operate in another country in order to be closer to natural resources or to access markets protected behind tariff or quota barriers. In terms of the latter motive, firms viewed foreign direct investment and trade as substitutes, providing alternate means to achieve a single goal. However, once a subsidiary firm was established in a foreign market, it operated largely independently of its parent with the possible exception of flows of capital and technology3. The international production network with its changing membership and complex interdependent relationships was more of an exception.
6. Policy-makers are now endeavouring to turn the impacts of globalization on their respective countries to their advantage. To inform their decisions, a comprehensive statistical description of the globalization process is required. Such a system must transcend national borders since the firms and activities being measured do so. Further, this system must record the evolution of firms over time - longitudinal tracking - as they form new alliances and start to operate in different countries.
7. This paper starts out by identifying a few of the relevant questions for which the current system does not provide answers. It examines, albeit in stylized form, the requirements of an extended statistical system capable of answering such questions. Finally, it speculates on the kinds of changes it is deemed desirable to introduce in the short and medium terms.
8. The current situation has led to a number of new policy- related questions. The following are some examples:
9. To shed light on these issues requires a statistical framework capable of providing information on the way in which individual firms organize their production, investment and trade within and across national borders.
10. The current statistical system takes national borders as one fundamental point of reference. It is the crossing of a national border which defines a "foreign" activity such as production or investment in a foreign country or trade with a foreign firm, even if that firm is domestically owned and controlled. Measures of GDP4 and GNP5, together with aggregate measures of trade and investment, represent national economies and their relationships with each other.
11. An extended statistical system designed to provide consistent descriptions of the way in which firms organize production across borders should not stop at measuring activity with respect to a firm's operations within national borders. Rather, the concepts of "domestic" and "foreign" should be based on the ownership of the factors of production.
12. The relevant questions to which the statistical system should respond require no distinction if a domestically-owned firm decides to access a foreign market through direct investment and production or through domestic production and trade. All activity carried out by domestically-owned firms must be measured and related to the home country.
13. The aggregate measurement meeting these criteria is a more broadly-defined GNP measure. This "extended" GNP includes all of the production-related activity carried out anywhere by domestically owned firms, and not simply the income earned from those activities as is currently recorded.
14. More specifically, the home country of a firm with subsidiaries in several other countries would record, in its own accounts, the output, employment, labour income, profit, net exports and sales of all of these subsidiary firms. The sales of goods and services between domestically-controlled firms would not be recorded as exports and imports -- regardless of the geographic locations of the firms -- but sales between domestic and non- domestically-controlled firms would be considered exports and imports, even if both firms were located within the geographic boundaries of one country.
15. In order to estimate ownership-based GNP at the aggregate level, GNP at the micro, or firm, level must be defined and recorded in a new data system. However, the relationships between the transacting firms must be systematically registered as illustrated in the examples in Appendix 1.
16. There is no suggestion that an extended GNP should be the only macro-economic aggregate but rather that it be an additional aggregate. The difference in level and composition between it and the traditional GDP/GNP is key information on the degree to which a particular country is subject to globalizing pressures.
17. The two versions of a statistical system are described in the following summary framework:
18. From these considerations arise three main requirements which must be met by an extended statistical system for use in a globalized economic environment:
19. The role of these requirements is the object of the following section. The appendix contains some illustrative examples to highlight the differences between the traditional and the extended statistical systems.
The boundary of economic activity should be determined by ownership of productive resources.
20. In the extended statistical system, each micro measurement requires two components. The first component is information on the transactors and their relationship; the second consists of information on the transaction they are engaged in (eg. information on the commodity being traded and its value).
21. Production is attributed to the country of ownership of the firm. That is, all activity including ore mining, fabricating, manufacturing of the final good, and all centralized service functions are deemed to be economic activity of the country of ownership. Thus, statistical agencies must record he linkage or relationship between firms for each economic activity engaged in by firms, even within local markets. Production activity, employment and income must be reported taking into account ownership in order to be attributed to the correct country.
22. Transfers between parent and subsidiaries in two countries should be subtracted from conventional exports and imports. Conversely, transfers of goods and services between a domestic firm and a foreign firm within one country should be added to conventional exports and imports. Again, this means extending the current system so that all transactions (within or cross border) are identified by the two transactors and their relationship with each other.
23. Since the relationship between transacting firms determines the domestic or foreign characteristic of each transaction, the statistical recording of each commodity bought and sold must carry the link with the seller or purchaser involved. The information required for each sale or transfer of product is essentially information on "origin or destination of deliveries", where the ownership characteristic is included with the origin or destination information.
24. The need to measure transactions at the micro level and to ignore national boundaries requires data sharing and confidentiality issues to be resolved.
25. In the extended system, country A must provide micro-data to country B for all of B's subsidiaries operating in A. Country A must be able to collect all of the required data from B's subsidiaries. It must establish a mechanism for sharing (transmitting) the data and it must work within a framework of confidentiality acceptable to individual firms.
26. In some cases country A may not have access to the data of B's subsidiary firm. Such a situation may arise in cases where the parent firm records the data for the larger enterprise but does not share them with its individual subsidiaries. In this case, A must communicate to B that B should collect these data directly.
27. For an extended statistical system to function, agreements regarding data sharing and confidentiality must be reached between A and B bilaterally. Alternatively, multilateral agreements must be struck through some international statistical forum.
28. Measuring economic activity with respect to ownership also solves a growing problem in the traditional statistical system. It is recognized that parent firms provide many services to subsidiary firms in foreign countries. For the traditional statistical system to be complete, these should be identified and considered as imports and exports of services. Further, since these services may not be priced, transfer prices should be estimated to value them. Since these problems are not being addressed comprehensively and systematically, the current statistical system under-counts economic activity, probably at an increasing rate.
29. In the extended statistical system all the activity accounted for by the complete parent-subsidiary firm is considered and allocated to the country of ownership. Thus, there are no implicit imports or exports between the parent and subsidiary.
30. The properties of the extended system vis-a-vis the traditional system are such that the sum of GDP across all countries is the same in both systems. The extended system is a different decomposition of world GDP.
31. By redefining the boundaries of production and trade and integrating data in this ownership-based framework, policy makers can gain a new perspective on their own countries in the globalized world. For example, they will wish to know which parts of a domestically-owned production chain are located abroad and what caused them to be located abroad. What factors do the other economies possess that offset benefits offered by the local economy? Are they economic factors such as unit labour costs, proximity to raw materials or availabiltiy of a modern capital stock? Are they socio-economic factors such as the education, health and dependability of the labour force? An extended system may not answer these questions directly but in combination with longitudinal enterprise data can provide a basis from which to infer the answers.
32. In the traditional statistical system, the conceptual framework is designed and implemented at the national level. By this it is meant that the statistical agencies collect data using the concepts, definitions and methodologies appropriate to the examination of the economic systems and relationships existing in their country.
33. In an extended statistical system, national surveys conducted in country A, for example, must be designed to measure variables of firms controlled by country B. The measurement of these variables must be identical in every aspect to their counterparts measured in country B since they will be used in the extended statistical system of country B.
34. Using employment as an example, and assuming that the problem is to calculate relative labour productivity between A's firms operating in A and those operating in B, A and B must agree on:
35. Of course the registers of firms must be compatible and consistent. That is, if country A knows it has a subsidiary in country B, the firm register of country B must record the existence of the subsidiary and its relationship with its parent in an identical manner.
36. In order to ensure consistency, countries A and B must enter into an agreement on the statistical exchange of information. Alternatively, such an accord can come from a supra-national authority.
37. This requirement is the dynamic version of the previous one. It recognizes that the search for the optimum global integrated production network is ongoing as firms move from country to country and comparative advantages are constantly shifting. For example, as more firms move labour intensive stages of a production process to a country with a low initial price of labour, that price may be driven upwards, diminishing the advantage. This in turn may trigger a shift out of that country in search of lower labour costs.
38. As the economic situation evolves, new and different types of information may be required to address the changing circumstances. Countries must acknowledge the new information needs, agree on the parameters of the new data needed, and implement changes simultaneously to ensure the continuing consistency and comprehensiveness of the statistical system.
39. Policy makers will wish to track the speed of this process but for the information required to be useful, agreements must be reached to ensure that the recording of events is synchronous in the countries involved. These agreements are more complex than those reached to support the traditional system which are based largely on a static situation. Under the extended system countries must agree to the nature of changing information needs and to the timing of their implementation.
40. The extended statistical system will not meet all needs for statistical information in national economies. Its strength will be in its micro-economic applications. Consider the following issue: a government wishes to institute a package of incentives designed to induce firms to invest more in machinery and equipment. How should it forecast the outcome? Assume, with some realism, investment decisions are typically head office decisions. If so, resident head offices, on average, will react in a certain way and this will affect the behaviour of their subsidiaries abroad. However, resident subsidiaries of foreign companies may not react in the same fashion, largely because the calculations of their head offices would be affected by other factors.
41. It follows that a better forecast would be likely if it took into consideration two aggregates - one relating to all domestically-owned companies, regardless of the location of their subsidiaries, and the other relating to all resident subsidiaries owned by non-residents. Thus, data collected on both the traditional basis and the extended (ownership) basis are important for such policy assessment.
42. Other illustrations of the need for the co-existence of the traditional and extended systems come from the measurement of economic relationships and conditions which continue to be of interest to policy-makers in a globalized world. This interest arises from the needs of governments to focus on their own citizens and their own geographically-defined economy for many purposes.
43. For example, the relationship between income and consumption remains of central macro-economic interest to policy makers. However, income of those employed by domestically-owned firms abroad, as measured by the extended system, is not meaningful in relation to consumption in the domestic economy. Income as measured by the traditional system continues to be required.
44. Similarly, labour market policies concerned with social welfare require information on the domestic labour force and its employment, whether the source of that employment be domestically or foreign owned firms.
45. Lastly, government taxation policy-making would also continue to require information from the traditional statistical system. Since taxation jurisdictions are currently defined by national borders, governments wish to know how much surplus is accruing to businesses and how much employment income is being earned by workers within domestic borders, data currently measured by the traditional system. This will remain true as long as taxation systems retain their current legal basis. If a "globalized" or extended world taxation system were ever implemented, the extended statistical system would provide its supporting information.
46. The traditional and extended statistical models are illustrated in the Annex through the production of a good in a two country, two firm world.
Consider a two country model of a globalized world. The accounts used to describe the traditional and extended statistical models include:6
Production--Goods
Income------Wages
------Surplus (including return on capital)
Employment
Final Domestic Sales
Traded Goods
In this simplified world,
Production = net value of goods produced
Income = wages + surplus
Firm A1 is a domestically-owned firm located in country A while firm A2 is its subsidiary located in country B. Firm B1 is a firm located and owned in country B while B2 is its subsidiary located in country A.
Contry A Country B +-----------------+ +-----------------+ | A1 | | B1 | | | | | | | | | | | | | | +_____+ | | +_____+ | | | A2 | | | | B2 | | | +_____+ | | +_____+ | +-----------------+ +-----------------+Consider 3 stages of production:(1)mining ore,(2)fabricating steel and (3)manufacturing steel widgets for final consumption. The firm which manufactures the widgets sells them in the home market.
CASE 1: Production occurs involving firms A1 and B1. The firms are unrelated and located in different countries.
Firm B1 imports the ore ($24), adds value ($8) by fabricating steel, employs labour (5 people), generates labour income ($5) and exports steel to Firm A1 ($32).
Firm A1 imports the fabricated steel ($32), adds value ($12) by producing widgets, employs labour (6 people), generates labour income ($6) and sells the widgets domestically ($44).
COUNTRY A COUNTRY B
OUTPUT OUTPUT
(Value Added) (Value Added)
Mining 24
Manufacturing 12 Fabricating 8
36
8
INCOME INCOME
Labour Labour
Mining 10 Fabricating 5
Manufacturing 6
Surplus Surplus
Mining 14 Fabricating 3
Manufacturing 6
36
8
NET EXPORTS NET EXPORTS
Goods Exports-
Goods Exports- steel 32
ore 24 Goods Imports-
Goods Imports- ore (24)
steel (32) NET 8
NET (8)
FINAL DOMESTIC FINAL DOMESTIC
SALES SALES
Widgets 44 0
44 0
EMPLOYMENT EMPLOYMENT
Mining 10 Fabricating 5
Manufacturing 6
16 5
CASE 2: Production process involves firm A1 (in country A)and firm A2, its subsidiary (located in country B). Residency of the two firms is different, but ownership is the same.
In country B, the subsidiary firm A2 receives the ore (does not import) ($24), adds value ($8) by fabricating steel, employs labour which is located in country B (5 people), generates labour income in the geographic boundaries of country 2 ($5)and then transfers (does not export) the fabricated steel to its parent firm A1 in country A ($32).
In country A, the parent firm A1 receives the steel (does not import)($32), adds value ($12)by manufacturing widgets, employs labour in country A (6 people), generates labour income in country A ($6), and sells the widgets domestically ($44).
The accounts for case 2 traditional model are identical to the Case 1 accounts.
COUNTRY A COUNTRY B
OUTPUT OUTPUT
(Value Added) (Value Added)
Widgets 44
44
INCOME INCOME
Labour Labour
Mining 10
Fabricating 5
Manufacturing 6
Surplus Surplus
23
44
NET EXPORTS NET EXPORTS
Goods Exports
Goods Exports Goods Imports
Goods Imports NET
NET
FINAL DOMESTIC FINAL DOMESTIC
SALES SALES
Widgets 44
44
EMPLOYMENT EMPLOYMENT
Mining 10
Fabricating 5
Manufacturing 6
21
Points to Note:
CASE 3: Production process involves firm A1 located in country A and foreign subsidiary firm B2 also located in country A. Residency of the two firms is the same but ownership is different.
Firm B2 "imports" the ore ($24), fabricates steel and since country B owns the productive resources, the production ($8)is attributed to country B, as is the employment (5 people)and labour income ($5). Firm B2 then "exports" the steel to Firm A1 ($32).
Firm A1 "imports" ($32) the steel, produces widgets, and sells them domestically ($44). The production ($12), employment (6 people) and labour income ($6) is attributed to country A, the country of ownership of Firm A1.
COUNTRY A COUNTRY B
OUTPUT OUTPUT
(Value Added) (Value Added)
Mining 24
Fabricating 8
Manufacturing 12
44
INCOME INCOME
Labour Labour
Mining 10
Fabricating 5
Manufacturing 6
Surplus Surplus
Mining 14
Fabricating 3
Manufacturing 6
44
NET EXPORTS NET EXPORTS
Goods Exports
Goods Exports Goods Imports
Goods Imports NET
NET
FINAL DOMESTIC FINAL DOMESTIC
SALES SALES
Widgets 44
44
EMPLOYMENT EMPLOYMENT
Mining 10
Fabricating 5
Manufacturing 6
21
CASE 3: ACCOUNTS FOR EXTENDED MODEL:
COUNTRY A COUNTRY B
OUTPUT OUTPUT
(Value Added) (Value Added)
Mining 24 Fabricating 8
Manufacturing 12
36 8
INCOME INCOME
Labour Labour
Mining 10 Fabricating 5
Manufacturing 6
Surplus Surplus
Mining 14 Fabricating 3
Manufacturing 6
36 8
NET EXPORTS NET EXPORTS
Goods Exports Goods Exports
ore 24 steel 32
Goods Imports Goods Imports
steel(32) ore (24)
NET ( 8) NET 8
FINAL DOMESTIC FINAL DOMESTIC
SALES SALES
Widgets 44
EMPLOYMENT EMPLOYMENT
Mining 10 Fabricating 5
Manufacturing 6
16 5
Points to Note:
This case has also been used to illustrate that the world total of each activity is identical in the two situations. The extended system is a different decomposition of world GDP.
CASE 4: The production process occurs between subsidiary firm B2 owned by country B but located in country A and subsidiary firm A2 owned by country A but located in country B. Residency and ownership are different AND opposite.
Firm A2 imports the ore ($24), adds value ($8) by fabricating steel, employs labour (5 people), generates labour income ($5) and exports steel to Firm B2.
Firm A2 imports the steel ($32), adds value ($12) by producing widgets, employs labour (6 people), generates labour income ($6) and sells the widget domestically ($44).
COUNTRY A COUNTRY B
OUTPUT OUTPUT
(Value Added) (Value Added)
Mining 24 Fabricating 8
Manufacturing 12
36 8
INCOME INCOME
Labour Labour
Mining 10 Fabricating 5
Manufacturing 6
Surplus Surplus
Mining 14 Fabricating 3
Manufacturing 6
36 8
NET EXPORTS NET EXPORTS
Goods Exports - Goods Exports -
ore 24 steel 32
Goods Imports - Goods Imports -
steel (32) ore (24)
NET ( 8) NET 8
FINAL DOMESTIC FINAL DOMESTIC
SALES SALES
Widgets 44
44
EMPLOYMENT EMPLOYMENT
Mining 10 Fabricating 5
Manufacturing 6
16 5
COUNTRY A COUNTRY B
OUTPUT OUTPUT
(Value Added) (Value Added)
Fabricating 8 Mining 24
Manufacturing 12
8 36
INCOME INCOME
Labour Labour
Fabricating 5 Mining 10
Manufacturing 6
Surplus Surplus
Fabricating 3 Mining 14
Manufacturing 6
8 36
NET EXPORTS NET EXPORTS
Goods Exports - Goods Exports -
steel 32 ore 24 Goods
Goods Imports - Imports -
ore (24) steel (32)
NET
NET 8
FINAL DOMESTIC FINAL DOMESTIC
SALES SALES
Widgets 44
44
EMPLOYMENT EMPLOYMENT
Fabricating 5 Mining 10
Manufacturing 6
5 16
Points to Note:
_______________________________
1 Prepared by Janice McMechan and Jacob Ryten.
2 For example: The Performance of Foreign Affiliates in
OECD Countries, OECD, Paris, 1994
3 Hence the debate over whether profits of enterprises
owned abroad should be ploughed back as part of the owning
country's current receipts. Even though the IMF's Balance of
Payments manual recommended that unremitted earnings of
branches and other unincorporated direct investment
enterprises and the direct investor's portion of earnings of
incorporated direct investment enterprises that are not
formally distributed should be conceived as providing
additional capital to the enterprises, thus increasing the
value of an economy's stock of foreign assets and liabilities,
not all countries accepted the recommendations. Moreover, it
was recommended that the share of portfolio investors in the
earnings of an incorporated direct investment enterprise that
are not formally distributed should not be entered in the
Balance of Payments. (International Monetary Fund, Balance of
payments manual, Fourth Edition, 1977, pages 104-105)
4 Gross domestic product (GDP) is defined as total goods
and services produced within the geographic boundaries of a
country over a particular time period, regardless of the
residence status of the owners of the productive assets.
5 Gross national product (GNP) is defined as total income
earned by residents of a country, regardless of where the
productive assets which earn the income are located. It can be
calculated as GDP plus income accruing to residents from
investments abroad less income earned in the domestic market
accruing to foreigners.
6 Inventories and the capital account are not considered.