Back Matter
Author:
International Monetary Fund
Search for other papers by International Monetary Fund in
Current site
Google Scholar
Close

Abstract

A companion document to the fifth edition of the Balance of Payments Manual, the Balance of Payments Compilation Guide shows how the conceptual framework described in the Manual may be implemented in practice. The primary purpose of the Guide is to provide practical guidance for using sources and methods to compile statistics on the balance of payments and the international investment position. the Guide is designed to assist balance of payments compilers and statisticians in understanding the relative strengths and weaknesses of various approaches. The material reflects the emergence of new data sources and adaptations in the application of statistical methodologies to changing circumstances. Discussed in the Guide are all of the tasks that a BOP compiler normally performs. Appendices contain a set of model BOP questionnaires and a set of model BOP publication tables. Relationships between the balance of payments statistics and relevant aspects of national accounts are covered as well.

Appendix I. Balance of Payments Compiler Survey on Data Sources and Methods

Introduction

1195. This appendix presents the results of a BOP compiler survey, which was conducted in 1991, on data sources and methods.1 Compilers in 58 countries responded to the survey and provided a considerable amount of information, without which the Guide could not have been produced. Table 1 (on page 242) shows the countries for which a response was received. In some cases, responses were not completed fully or contained conflicting information; therefore, it was necessary to use some judgement in compiling the results of the survey.

Table 1.

Countries Reporting in the BOP Compiler Survey

article image

1196. Of the compilers reporting, 71 percent were from central banks or monetary authorities; 22 percent were from national statistical offices; and 7 percent were from other agencies. A breakdown of agency types by region is shown in table 2 on page 242.

Table 2.

Types of Agencies Reporting in Survey

article image

1197. Table 3 (on page 243) provides a description of the data sources referred to in the survey and in this appendix. These sources include international trade statistics (ITS), the international transactions reporting system (ITRS), enterprise surveys (ES), official sources n.i.e., household sector collections, and partner country sources (including international organizations).

Table 3.

Classification of Data Sources

article image

1198. Table 4 (on page 243) shows the incidence of data sources used by compilers. Compilers sometimes reported sources not actually used in BOP compilation; these sources were excluded from the analysis. Table 4 shows that compilers in all of the countries responding to the survey reported using official sources; compilers in 88 percent of the countries responding use ITS and/or ES; compilers in 84 percent of the countries use an ITRS; and compilers in 62 percent of the countries use household collections and/or data from partner countries to compile their BOP accounts. A more detailed analysis of the use of each of these sources is given in subsequent tables.

Table 4.

Compiler Survey—Incidence of Sources Used

article image

1199. Table 5 (on page 243) shows a cross-classification of sources by the primary BOP classifications in which the sources are used. The table shows that:

  • (1) ITS are used to compile goods items in 88 percent of cases, to compile service items (usually associated with transportation) in 43 percent of cases, and to compile transfer items (usually associated with the identification of goods under foreign aid programs) in 17 percent of cases.

  • (2) An ITRS is used to compile goods items in 26 percent of cases; service items in 79 percent of cases, income items in 66 percent of cases, transfer items in 76 percent of cases, financial account items in 62 percent of cases, and IIP items in 31 percent of cases.

  • (3) ES are used (usually as a supplementary source) to compile goods items in 40 percent of cases, service items in 72 percent of cases, income items in 57 percent of cases, transfer items in 21 percent of cases, financial account items in 66 percent of cases, and IIP items in 48 percent of cases.

  • (4) Official sources are used (usually as a supplementary source) to compile goods items in 21 percent of cases, service and income items in about 70 percent of cases, and transfer and financial account items in 81 percent of cases. For countries that compile stock positions, official sources are used in all but one case.

  • (5) Household collections (mostly migration statistics and surveys of travelers) are used to compile service items (mostly travel) in 62 percent of cases and to compile transfer items in 12 percent of cases.

  • (6) Data from partner countries, etc. are used to compile service items (particularly those associated with foreign government expenditure) in 34 percent of cases, and transfer items (usually those associated with foreign development assistance) in 36 percent of cases.

Table 5.

Compiler Survey—Sources Used to Compile Data Items

article image
Note: Compilers in only 36 of the countries responding to the survey compile either full or partial IIP statistics.

Individual Data Sources

International Trade Statistics

1200. A number of questions were asked about ITS, including whether a country’s statistics are compiled on a general or a special trade basis. As table 6 (on page 244) shows, 61 percent of all countries responding to this part of the survey use a general trade basis. Germany, which uses a special trade basis, reported that it adjusts goods for warehouse transactions and thus makes the coverage of its trade statistics consistent with a general trade basis.

Table 6.

Compiler Survey—ITS Statistical Basis

article image

1201. Compilers were also asked to report the point of valuation used in ITS. Of particular interest was whether imports are recorded on a c.i.f. or an f.o.b. basis. When more than one basis was reported, compilers were asked which basis is used to compile goods items in the BOP. Table 7 (on page 244) shows that 38 countries (75 percent) measure imports on a c.i.f. basis and the remainder use an f.o.b. basis. Some qualifications are needed. The United States uses a free-along-side ship basis, and Canada uses an ex-plant basis. (Both are coded as f.o.b in table 7.) Indonesia, which is coded as c.i.f. in table 7, uses a cost and freight basis.

Table 7.

Compiler Survey—Import Valuation in ITS

article image

1202. Compilers were asked whether goods were recorded on a shipment, documents cleared, documents processed, or some other timing basis. Compilers were also asked about the methods used to undertake currency conversion. The answers to these questions were varied, and there was no easy way to tabulate them.

International Transactions Reporting Systems

1203. Compilers who use international transactions reporting systems were asked about the nature of their systems; questions included (a) whether the system covers foreign currency only or foreign and domestic currency transactions, and (b) whether the system is closed, open, or partial. Table 8 (on page 244) shows that, in the 49 ountries in which an ITRS is used, 55 percent of systems measure foreign currency transactions only. Twenty-nine percent of the systems used are closed, 49 percent are open, and 16 percent are partial.2 In the case of industrial countries that use an ITRS, 92 percent of systems cover both domestic and foreign currencies and 50 percent of the systems are closed. While answers to other questions on the use of an ITRS provided valuable information, it was difficult to compile meaningful tables.

Table 8.

Compiler Survey—Nature of International Transactions Reporting Systems

article image

Enterprise Surveys

1204. Table 9 shows types of ES used by compilers. The most prevalent types reported by compilers were designed to measure: bank transactions (compilers in 67 percent of countries use ES for this purpose); transportation transactions (66 percent); external assets and liabilities (59 percent); service transactions, other than transportation and travel (48 percent); travel (36 percent); transactions in goods (34 percent); the activities of financial intermediaries (29 percent); compensation of employees (26 percent); and the transactions of private aid organizations (26 percent).

Table 9.

Compiler Survey—Use of Enterprise Surveys

article image

Official Sources

1205. Table 10 (on page 245) shows the types of official sources that compilers reported using. In order of frequency, the most prevalent sources reported were: central bank information on reserves (used by compilers in 91 percent of countries); debt management offices (71 percent); government accounts (67 percent); foreign aid accounts (66 percent); approvals of foreign investment (22 percent); education statistics (21 percent); applications to obtain foreign exchange and taxation records (16 percent in both cases); immigration records, other than migration statistics, (12 percent); and health statistics (5 percent). A compiler in one country reported using information from applications to export goods. Several other types of official sources were reported, but mese are not shown in the table.

Table 10.

Compiler Survey—Use of Official Sources n.i.e.

article image

Household Collections

1206. Table 11 (on page 246) shows the types of household collections that compilers reported using. The table shows that compilers in 50 percent of the countries responding use surveys of travelers and compilers in 43 percent of the countries use migration statistics. These two sources are usually used in combination to compile all or part of the travel item. The use of other types of household collections was reported by 14 percent of compilers.

Table 11.

Compiler Survey—Use of Household Collections

article image

Partner Countries and International Institutions as Data Sources

1207. Table 12 (on page 246) shows the use by BOP compilers of partner countries and international institutions as sources of data. Compilers in 21 percent of the countries responding reported using surveys of foreign embassies; in 40 percent of cases, compilers reported using information on foreign aid from the accounts of donors; and in 29 percent of cases, other partner country sources were reported as being used. Compilers in 12 percent of countries reported using data from international institutions.

Table 12.

Compiler Survey—Use of Information from Partner Countries and International Institutions

article image

Frequency and Timeliness of Data Sources

1208. This section discusses the observations of compilers on the frequency and timeliness of data sources used to compile BOP statistics. Compilers were asked to record, for each data source used, how frequently data were available and the elapsed time, from the reference period, in which data became available. Results are presented in table 13 on page 247. Some compilers reported the frequency but not the timeliness of data sources; these cases were excluded from the table. For the frequency of data collections, only three time periods are shown in the table. Sources reported as being available each month or more frequently are shown as monthly; sources reported as being available every six months or less frequently are shown as annual. When compilers reported several frequencies for one source, a judgement was made about the predominant frequency. Data on average time lags are rounded to the nearest week.

Table 13.

Compiler Survey—Frequency and Timeliness of Data Sources

article image

1209. Predominantly monthly sources include ITS, international transactions reporting systems, information on reserves from central banks, foreign investment approvals, applications to acquire foreign exchange, ES of banks, and migration statistics. ES measuring transactions in goods, ES of financial intermediaries, government accounts, foreign aid accounts, debt office records, ES of transportation activities, surveys of travelers, and partner country data on foreign aid tend not to have a predominant frequency. Sources that tend to have an annual frequency are ES of travel, ES of other services, ES of external assets and liabilities, and education statistics. However, a significant incidence of quarterly ES of external assets and liabilities was reported by compilers from countries that rely predominantly on ES to compile quarterly BOP statistics.

1210. As table 13 shows, compilers reported that most data were available within a reasonable time frame. However, compilers frequently reported estimating some items when data were unavailable to meet timetables. Table 13 also shows that the more frequent data sources were usually the most timely. Information from central banks on reserves, government accounts, and applications to acquire foreign exchange tended to be the most timely of all data sources. Generally, the timeliness of ITS, international transactions reporting systems, and ES was good.

1211. The average time lags for data sources shown in table 13 mask some significant variations. If the more extreme time lags are removed, the results show what could be reasonably expected. Adjusted results are shown in table 14 on page 248.

Table 14.

Compiler Survey(Adjusted Results)—Frequency and Timeliness of Data Sources

article image

Compiler Assessments of the Quality of Data Sources

1212. Compilers were asked to assess, by using the following ratings, the quality of their data sources:

  • (A) There is good coverage of transactors, and data are valued and classified correctly.

  • (B) There is less than complete coverage of transactors, but data are valued and classified correctly.

  • (C) There is good coverage of transactors, but data are not always valued and classified correctly.

  • (D) There are coverage, valuation and classification deficiencies, but data are considered to be useful indicators for BOP purposes.

  • (E) The data are of poor quality.

1213. Compiler ratings were converted into numeric values (“A” equaled 1, “B” equaled 2, and so on), which were then aggregated and averaged; results are presented in table 15. Sources with lowest values may be considered, in the opinion of compilers, to be the most accurate. The results should be viewed with care, as the ratings provided are based on subjective perceptions that may not be directly comparable.

Table 15.

Compiler Survey—Compiler Perceptions of Data Source Quality

article image

1214. As table 15 (on page 250) shows, the sources rated most highly, in terms of quality, were: central bank reserves information (average rating 1.2); government accounts and ES of banks (average rating of 1.3 for both); ITS and debt management offices (average rating of 1.5 for both); foreign aid donor accounts (average rating 1.6); and partner country accounts of foreign aid donors (average rating 1.8). Sources receiving the lowest ratings were: tax records and ES of employees compensation (average rating 2.9 for both); surveys of foreign embassies (average rating 2.8); and partial international transactions reporting systems, ES of services other than transportation and travel, and surveys of travelers (average rating of 2.5 for each).

1215. Table 16 (on page 251) shows compiler assessments, which are broken down by region, of the quality of the main sources. Compilers in industrial countries rate ITS quite highly, whereas compilers in African countries rate the quality of ITS at a much lower level. On the other hand, all regions rate official sources about the same. Regional differences in the ratings for ES and international transactions reporting systems are not particularly significant. The consistent 2.0 rating given to partner country sources by non-industrial countries contrasts markedly with the average rating of 39 given by industrial countries.

Table 16.

Compiler Survey—Compiler Perceptions, by Region, of Data Source Quality

article image

Compilation of BOP and IIP Statistics

1216. This section examines the data sources used to compile items in the BOP and IIP.

Goods

1217. Table 17 (on page 251) shows sources used to compile goods items in BOP accounts. Compilers in 45 countries (78 percent) use ITS as the main data source; compilers in 23 of these countries use only ITS; and compilers in the remaining 22 countries supplement ITS with other sources. These supplementary sources include international transactions reporting systems (used in 5 countries), ES (used in 17 countries), official sources (used in 5 countries), and partner country sources (also used in 5 countries). Compilers in 6 countries (10 percent) use an ITRS as their main data source; compilers in 3 of these countries use only an ITRS; compilers in 2 countries supplement the ITRS with official sources; and, in 1 country, the ITRS is supplemented with ITS.

Table 17.

Compiler Survey—Sources Used to Measure Goods Items

article image

1218. Compilers in 7 countries reported that they use neither ITS nor an ITRS as the main source for measuring goods items. The Congo uses a mixture of sources including ITS, an ITRS, official sources (for goods provided under development assistance), and ES. Gabon uses ES for exports and ITS for imports. Swaziland uses ITS but, as it also uses ES to check the ITS results, Swaziland has been counted in the other sources row. Indonesia uses a combination of ITS and ES. Sri Lanka measures imports by using both ITS and an ITRS and measures exports by using an ITRS. The Bahamas uses ES for imports and relies on government accounts and surveys of travelers for exports. Trinidad and Tobago essentially relies on ES of goods.

Transportation Services

1219. Table 18 (on page 251) shows that compilers use a variety of sources to measure transactions in transportation services. Compilers in 40 countries (69 percent) reported that they use ES (usually surveys of transportation enterprises) to measure transportation transactions. In 10 cases, ES are not supplemented with either ITS or an ITRS. Thirteen compilers reported using an ES supplemented with ITS.

Table 18.

Compiler Survey—Sources Used to Measure Transportation Services

article image

1220. Compilers in 33 countries (57 percent) rely on international transactions reporting systems to measure transportation items. In 9 of these cases, the ITRS is not supplemented with either ITS or ES. Seven compilers reported using an ITRS supplemented with ITS; 12 reported using an ITRS supplemented with ES; and 5 reported that they supplement the ITRS with both ITS and ES.

1221. Compilers in 24 countries (41 percent) use ITS to measure certain transportation items, such as freight. Several of these compilers reported that ITS provided them with information on imports of goods on both an f.o.b. and a c.i.f. basis or other data necessary to derive freight and insurance on imports.

1222. Official sources are used by compilers in five countries to measure transportation items; in each case, other sources are used to supplement the official sources.

1223. Respondents from the United Kingdom and the United States reported using surveys of travelers to estimate passenger fares.

Travel Services

1224. Table 19 (on page 252) shows that compilers in 37 countries (64 percent) reported using an ITRS to measure travel services. In 33 countries (57 percent), compilers use household sector collections (typically immigration statistics and surveys of travelers). Compilers in 12 countries (21 percent) use ES; compilers in 8 countries (14 percent) use official sources; and compilers in 5 countries (9 percent) use data from partner countries.

Table 19.

Compiler Survey—Sources Used to Measure Travel Services

article image

1225. It is interesting to analyze the combinations of data sources used to measure travel transactions in the BOP. Single sources are used in 25 cases (ITRS, 16; ES, 2; household collections, 6; and foreign exchange approvals, 1). An ITRS is combined with other sources in 21 cases; in 14 of these, the ITRS is combined with household collections, in 4 cases—with both household collections and ES, and in 3 cases—with other sources. ES are combined with sources other than an ITRS in 7 cases; in 4 of these cases, ES are combined with household collections, in 1 case—with official sources, and in 2 cases—with other sources. Household collections are combined with sources other than international transactions reporting systems and ES in 4 cases.

Education and Health Services

1226. Table 20 (on page 253) provides information on sources used by countries to measure education and health services for the BOP.3 Compilers in 39 countries reported measuring education services, and compilers in 33 countries reported measuring health services. In the countries that compile these items, the main sources are international transactions reporting systems (used in 25 countries) and official sources (used in 18 countries).

Table 20.

Compiler Survey—Sources Used to Measure Education and Health Services

article image

1227. Compilers in 26 countries reported using a single source (an ITRS in 17 cases, an official source in 7 cases, and ES and household collections in 1 case each). Compilers in 7 countries combine an ITRS with other sources—9 with official sources, 1 with a household sector collection (Bahrain), and 1 with partner country data (El Salvador). Compilers in 3 other countries reported using more than 1 source: Australia (ES and official sources), Canada (official and partner country sources), and the United Kingdom (household collections and partner country sources).

Other services n.i.e.

1228. Table 21 (on page 253) shows the data sources used to compile other service items. The main sources are: international transactions reporting systems, which are used by compilers in 41 countries (71 percent); ES, which are used in 30 countries (52 percent); official sources, which are used in 37 countries (64 percent); and partner country sources, which are used—particularly to measure services provided to foreign governments—in 11 countries (19 percent).

Table 21.

Compiler Survey—Sources Used to Measure Other Services n.i.e.

article image

1229. Compilers in 13 countries reported that they use only one source to measure transactions in other services. Eight of these compilers reported using an ITRS; three compilers reported using official sources; one compiler reported using ES; and one compiler reported using partner country data. Compilers in 33 countries reported using two sources; the most common combinations were an ITRS and official sources (13 cases), ES and official sources (8 cases), and an ITRS and ES (7 cases). Compilers in 10 countries reported using three sources; the most common combination was an ITRS, ES, and official sources. One compiler reported using four sources and another reported using five.

Compensation of employees

1230. Compilers in 32 countries reported that they collect data on compensation of employees. Compilers in 22 of these countries use an ITRS for this purpose; in 16 countries, ES are used; and in 3 countries, official sources—government accounts and tax data—are used. Australia reported using a survey of travelers to measures wages and salaries paid to nonresident travelers while they are traveling in Australia.

1231. Compilers in 25 countries reported using a single source—either an ITRS (17 cases) or ES (8 cases)—to measure the compensation of employees item. Compilers in 4 countries reported using both an ITRS and ES. One compiler reported using a combination of ES, official, and household sector collections, while another compiler reported using an ITRS, ES, and official sources.

Table 22.

Compiler Survey—Sources Used to Measure Compensation of Employees

article image

Reinvested Earnings

1232. As table 23 (on page 254) shows, compilers in 36 countries reported collecting data on reinvested earnings of direct investment enterprises. Most (24 compilers) reported that they use ES. Compilers in 8 countries reported that they use official sources (including foreign exchange approvals, approvals of foreign investment, and debt office records) and compilers in 5 countries reported that they use an ITRS. Only one compiler reported using two sources—an ITRS and information from approvals of foreign investment.

Table 23.

Compiler Survey—Sources Used to Measure Reinvested Earnings

article image

Other Investment Income

1233. Compilers in all but 3 countries reported that they collect data on investment income other than reinvested earnings. Table 24 (on page 254) shows that compilers in 37 countries (64 percent) use an ITRS; compilers in 35 countries (60 percent) use official sources; and compilers in 29 countries (50 percent) use ES. Compilers in Belgium and the United Kingdom reported that they use data from international institutions to measure part of income debits.

Table 24.

Compiler Survey—Sources Used to Measure Investment Income, Other than Reinvested Earnings

article image

1234. Compilers in 21 countries reported that they use a single source—either an ITRS (14 cases), ES (4 cases), or official sources (3 cases). Compilers in 20 countries reported using two sources; 8 combine an ITRS with official sources and 10 combine ES with official sources. Compilers in 14 countries reported that they use at least three sources; 13 of the compilers reported combining an ITRS, ES, and official sources.

Transfers (current and capital)

1235. Compilers in all but two countries reported that they measure transfers for the BOP. Ten compilers reported that they use ITS; 44 reported using an ITRS; 12 reported using ES; 47 reported using official sources; 7 reported using household collections; and 21 reported using partner countries or international institutions as sources.

1236. Only seven compilers reported using a single source—either official sources (5 cases) or an ITRS (2 cases)—to measure transfers for the BOP. Twenty-nine compilers reported using two sources; the most common combination was an ITRS and official sources. Most of the other compilers reported using between three and five sources.

Direct Investment, Other Than Reinvested Earnings, Financial Flows

1237. Table 25 shows that compilers in all but 6 countries reported that they collect data on direct investment, other than reinvested earnings, financial flows. Compilers use a variety of sources including international transactions reporting systems (26 countries); ES (26 countries); and official sources (16 countries) consisting mainly of government accounts, debt office records, and information from foreign exchange and foreign investment approvals.

Table 25.

Compiler Survey—Sources Used to Measure Direct Investment, Financial Flows*

article image

Other than reinvested earnings

1238. Compilers in 37 countries reported that they rely on a single source. In 16 cases, this source is an ITRS; in 15 cases—ES, and in 6 cases—official sources. Compilers in 13 countries reported using two sources—in 7 cases, an ITRS and ES; in 3 cases, an ITRS and official sources; and in 3 cases, ES and official sources. Two compilers reported using a combination of an ITRS, ES, and official sources.

Portfolio Investment, Financial Flows

1239. Compilers in 41 countries reported that they collect separate information on portfolio investment, financial flows. Compilers in 15 other countries reported that they measure portfolio investment indistinguishably from other investment.

1240. Table 26 shows that the more common sources for measuring portfolio investment items are: an ITRS (20 countries); ES of external assets and liabilities (17 countries); ES of financial intermediaries (12 countries); ES of banks (18 countries), and official sources (20 countries), including government accounts, debt office records, and information from foreign investment and foreign exchange approvals.

Table 26.

Compiler Survey—Sources Used to Measure Portfolio Investment, Financial Flows

article image

1241. Compilers in 21 countries reported that they use a single source—in 10 cases, an ITRS; in 6 cases, ES; and in 5 cases, official sources—to measure portfolio investment, financial flows. Eighteen compilers reported that they use two sources—12 use ES and official sources; 3 use an ITRS and ES; and 3 use an ITRS and official sources. Compilers from 2 countries reported that they use a combination of ITRS, ES, and official sources.

Other Investment, Financial Flows

1242. Compilers in 40 countries reported that they collect separate information on other investment. Compilers in 9 other countries reported that they collect data on other investment that are indistinguishably included with portfolio investment. Compilers in 9 countries reported that they do not collect data for either portfolio investment or other investment; six of these compilers also do not collect information on direct investment.

1243. As table 27 (on page 256) shows, the most common sources of information on other investment, financial flows are: an ITRS (23 countries); ES of external assets and liabilities (20 countries); ES of banks (23 countries); and official sources (27 countries) including government accounts, debt office records, and approvals for foreign investment and foreign exchange. Belgium, Germany, and the United Kingdom reported using data from international institutions to measure part of other investment, and Venezuela reported using data from partner country sources.

Table 27.

Compiler Survey—Sources Used to Measure Other Investment, Financial Flows

article image

1244. Compilers in 12 countries reported that they rely on a single source—either an ITRS (5 cases), ES (5 cases), or official sources (2 cases)—to measure other investment financial flows. Compilers in 22 countries reported using two sources; the most common combinations were an ITRS and official sources (9 cases) and ES and official sources (8 cases).

International Investment Position Statistics

1245. A majority (36 out of 58) of countries responding to the survey compile some form of IIP statistics. Thirty-four countries produce statistics on stocks of direct investment; 29 measure stocks of portfolio investment; and 30 measure stocks of other investment. However, in a number of cases, the coverage of the statistics is only partial. Compilers in 22 countries reported that they do not compile any IIP statistics.

1246. Table 28 (on page 256) shows that the main sources used to measure stocks of direct investment are ES (used in 24 countries), international transactions reporting systems (10 countries), and official sources (9 countries). The official sources used include debt management offices and approvals of foreign investment.

Table 28.

Compiler Survey—Sources Used to Measure Direct Investment Stocks

article image

1247. Compilers in 24 countries reported that they use a single source—either ES (18 cases), an ITRS (3 cases), or official sources (3 cases)—to compile statistics on stocks of direct investment. Compilers in 6 countries reported that they use both ES and an ITRS, and compilers in 4 countries reported using both ES and official sources.

1248. Table 29 shows the sources used to measure stocks of portfolio investment. Compilers in 11 countries reported that they use an ITRS. Compilers in 18 countries reported that they use ES. The United States relies solely on ES of intermediaries and, in El Salvador and Portugal, the ES are restricted to banks. Compilers in 20 countries reported using official sources.

Table 29.

Compiler Survey—Sources Used to Measure Portfolio Investment Stocks

article image

1249. Compilers in 12 countries reported that they use a single source—in 6 cases, an ITRS; in 2 cases, ES; and in 4 cases, official sources. Fourteen compilers reported that they use two sources; 11 of these compilers combine ES and official sources. Compilers in 3 countries combine an ITRS, ES, and official sources.

1250. Table 30 shows the sources used to measure stocks of other investment. Thirteen countries use an ITRS, 22 use an ES (the compiler from New Zealand reported that their ES did not include banks), and 16 use official sources. The compiler from Germany reported using data from international institutions to measure part of stocks of other investment.

Table 30.

Compiler Survey—Sources Used to Measure Stocks of Other Investment

article image

1251. Compilers in 9 countries reported that they use one source (in 6 cases—ES and in 3 cases—an ITRS) to measure stocks of other investment. Compilers in 20 countries reported that two sources are used. The combinations consisted of ES and official sources (11 countries), an ITRS and ES (5 countries), and an ITRS and official sources (4 countries). One compiler reported using an ITRS, ES, and official sources.

Compiler Assessments of the Quality of Statistics

1252. Compilers were asked to rate the quality of their BOP and IIP statistics by using the same rating system that they used to rate the quality of their data sources (see paragraph 1212). The ratings for statistical quality were also converted to numeric values and then averaged. The results are presented in table 31. The table shows the average assessment of quality for particular BOP and IIP components. The number of compilers who rated a particular component is shown in parentheses.

Table 31.

Compiler Survey—Compiler Assessments of the Quality of BOP and IIP Components

article image
Note: Numbers of compilers providing ratings are shown in parentheses.

1253. The ratings provided in table 31 should be interpreted with caution. In a number of cases, the rating for a component was derived from ratings provided for sub-components.

1254. According to compiler perceptions, the most accurate components are reserves, that is, flows (with an average rating of 1.3) and stock positions (with an average rating of 1.0). The next most accurate items are recorded trade and flows and stocks of portfolio investment and other investment. On average, each of these components received a rating of less than 2. Recorded trade statistics were rated highly (average rating 1.2) by compilers in industrial countries, but compilers in other countries—in particular African countries—rated the quality of these statistics somewhat lower. The overall average rating for these statistics was 1.8. Components that received average ratings of between 2.0 and 2.3 included: direct investment (other than reinvested earnings) financial flows, direct investment stocks, other investment income (all with an average rating of 2.0); education services and adjustments to recorded trade (both with an average rating of 2.1); services other than travel, transportation, and education (average rating 2.2); and travel, reinvested earnings on direct investment, and transfers (all of which received an average rating of 2.3).

1255. The components perceived by compilers to be the least accurate were transportation and compensation of employees; each received an average rating of 2.6. For some components there was often a wide range of ratings given to subcomponents. For example, within the other services and transfers components, the subcomponents measured by using official sources were perceived as being quite accurate, while other subcomponents, such as migrants’ transfers, were perceived to be significantly less accurate. Within the financial account and income components, compilers, on average, assigned better ratings to items associated with the official sector and somewhat lower ratings to non-official-sector transactions.

1256. The assessments by compilers in industrial countries of the quality of their statistics are, on average, higher or equal to the assessments of compilers in other countries for all components except transportation, travel, compensation of employees, and transfers.

Appendix II. Model Balance of Payments Forms

Summary of the Model Collection Forms

article image
article image

Form 1—Exploratory Survey

Form 1—Exploratory Form

General Notes

  1. This form should be completed for the enterprise named on page one and any of its subsidiaries in Gondwanaland. If there are any errors in the address label, please make corrections before returning the form.

  2. A nonresident is an individual, enterprise, or other organization ordinarily domiciled in a country other than Gondwanaland. Gondwanaland branches and subsidiaries of nonresident companies are regarded as residents of Gondwanaland. Similarly, foreign branches and subsidiaries of Gondwanaland companies are regarded as nonresidents.

  3. All values are expressed in Gondwanaland dollars.

Part A. Introductory Questions

article image

Part B. Nonresident Owners

article image

Part C. Ownership of Nonresident Branches and Companies

article image

Part D. International Trade in Goods in 1993

article image

Part E. International Trade in Services in 1993

article image

Part F. Employment of Foreign Workers in 1993

article image

Part G. External Financial Assets and Liabilities as of December 31, 1993

article image