Introduction and Summary

Establishment of the Working Party arose against the background of dramatically acceler ated changes in the environment for international capi The Interim Committee of the Fund's Board ofGov tal flows, as outlined subsequent! y .3 ernors, in its communiqué of September 25, 1989, "encouraged the lMF's Executive Board to continue improving the analytical and empirical framework International Capital Markets in the 1980s underlying multilateral surveillance, including the mea surement, determinants, and systemic consequences of During the 1980s, international capital markets international capital flows."At a meeting in December underwent unprecedented changes that gave rise to 1989, the Executive Board decided that the Fund significantdifficulties for the measurement and record should undertake a study on the measurement of inter ing of capital flows. Most industrialized countries national capital flows. To this end, it was agreed that a removed existing restrictions on cross-border capital working party of national and international balance of flows; many innovative financial instruments were payments experts, drawn from industrial and develop introduced; and financial marketsbecame increasingly ing member countries, be constituted under the chair integrated in a global framework. Consequent! y, cross manship of Baron Jean Godeaux of Belgium. To sup border financial flows expanded rapidly and now port the Working Party, a small group of technical greatly exceed international trade in goods or services. experts was organized and attached to the Statistics The main features of this evolution were financial liber Department of the Fu nd. An interim report was submit alization, innovation, and the changing composition of ted to the Managing Director in December 1990. cross-border financial tlows. ln this process, private This final report is presented by the Working Party in capital flows outpaced official capital flows, and the accordance with its mandate to investigate the principal private flows increasingly were arranged outside the sources of discrepancies in the main components of traditional domestic financial intermediaries. Many global capital flows and to consider courses of action new market participants became active in international to be taken by the Fund and national statistical authori finance. ties for improviog the measurement ofthese flows over time. This initiative was prompted in large part by the recognition that major discrepancies in reported capital flowsintroduce a significantelement of uncertainty into Yfhe reported balance of payments data used in this report are the conduct of economie policy. These discrepancies, based on the Fund' sBalcmce ofPayments Statistics Yearbook( 1990), which became apparent on a global scale in the 1980s referred to in this report as the Yearbook. ln addition, references are (Table 3) and were accompanied by large errors and made to the fourth edition of the Fund's Balance ofPayments Man omissions in national balance of payments statements, ua/. referred to in this report as the Manual.

Table 3. Global Capital Account Discrepancies, 1983-89 (ln billions of U.S. dollars; outtlows (- )) 1983 1984 1985 1986 1987 1988 1989 Capital accouni. total 67.4 44.7 53.2 18.0 32.6 45.1 65.8 Direct investment 9.3 4.8 -9.2 -13.6 -24.5 -19.8 -8.2 Portfolio investment 24.0 32.6 43.4 3.6 -0.7 -11.9 32.3 Other capital 36.2 25.8 27.0 32.7 110.5 76.6 45.6 Reserves/LCFAR 1 -2.0 -18.7 -7.9 -4.7 -52.7 0.2 -3.9

Memoranda: 8 2 Current account - 6 . -72.4 -72.9 -58.9 -45.5 -52.5 -75.6 Errors and omissions 0.9 27.8 19.7 41.0 13.0 7.3 9.8 Source: Bala11ce of Payme11ts Statistics Yearbook, 1990. 'Discrepancy between changes in official reserves and liabilities constituting foreign authorities' reserves (LCFAR).

©International Monetary Fund. Not for Redistribution International Capital Markets in the 1980s

Financinl Liberalization opportunüies avai1able to both domestic and fo reign firms. ln other EC countries, entry of foreign institu Withering of Exchange Contrais tions was eased in the context of the Second Banking Directive, which allows banks (including subsidiaries The end of the Bretton Woods system in the early of non-EC firms) established in an EC member state 1970s and the generalization of floating exchange rates to open branches freely within the Community. More were accompanied by the removal of exchange con recently, J apan eased restrictions on the entry of for trois. The Liberalization of capital account transactions eign financial institutions, and similar liberalization also took place in compliance with the Code of Liberal measures were undertaken in most other OECD coun ization of Capital Movements, which was introduced tries. by the OECD, and with the Council Directives of the European Communities (EC). Other Measures of Deregulation By the mid-1960s, a number of industrial countries had relatively liberal policies on capital ftows; else [n the 1980s, most industrial countries also under where, restrictions remained important. The funda went a process of extensive domestic financial liberal mental change in policy attitudes that began in the ization that focused on deregulation of fees and charges 1970s became widespread in the 1980s. This change and on dismantling restrictions on the permissible was heralded by the al most "overnight" dismantling range of activities by financiaJinstitutions. Interest ceil of exchange controis in the United Kingdom in October ings and other restrictions on remuneration for bank 1979 and the culmination of a longer-term process of deposits, which were widespread in the United States, ljberalization in Japan in December 1980. The German Japan, and certain European countries in the 1970s, authorities aJso significanlly reduced indirect methods were lifted or relaxed to the point of insignificance in of restraining capital movements in the 1970s. Other ali industrial countries. There has been a general blur industrial countries-including Belgium, France, and ring of distinctions between banking and the securities ltaly-progressively liberaJized cross-border financial businesses and a growth of linkages between banking transactions during the 1980s. Sorne developing coun and insurance. From the viewpoint of statistical offices, tries also took important steps during the 1980s toward one consequence has been that traditional sources of the liberaJization of direct and portfolio investment reg data, especially banks, cover Jess of the ongoing activ ulations. ity, and, even for banks, an increasing share of transac The liberalization or abolition of exchange controls tions take the form of operations in securities. had an important effect on many statistical systems and, in sorne cases(for example, the United Kingdom), Financinllnnovation there was a clearly substantial and adverse impact on the coverage and quality of the data. New sources, Against this background of increased liberalization including surveys and estimation procedures, were and competition in the 1980s, deposit money banks and introduced, and the process of shifting data sources is securities firms strove to offer their customers new stiJl in progress. These changes required rapid adapta financial instruments with broad market access and tion by national statistical agencies trying to keep pace minimal transaction costs.4 A variety of new debt witb growing difficuJties. instruments emerged in the domestic U.S. market and the Eurodollar markets and spread to domestic finan Deregulation of Market Entry cial markets in other countries. The introduction of tloating rate notes (FRNs) in the early 1970s was an Until the 1970s, national financial systems typically early innovation thal rapidly overtook the traditional were protected from the entry of fo reign institutions syndicated Joan market. The introduction in 1981 of into domestic markets as part of the effort to isolate note issuance facilities (NIFs) was another important domestic markets from external developments. Com expansion of choices in instruments. 5 More recent! y, petition also was stifled by regulations that imposed equity-related debt instruments have become signifi numerous limitations on financial institutions' activi cant. The market for derivative products also devel ties and portfolio choices. However, as these limjta oped considerably. New contracts for financial futures tions were increasingJy perceived to hinder efficiency and distort credit allocation, many of them were dis carded du ring the 1970s. Simultaneously, restrictions on the entry of foreign institutions were relaxed, and 4A more detailed description offinancialinnovations is included in important strides were made toward greater openness Max Watson and others, "Innovations and lnstitutional Changes in Financial in and equality of competitive opportunity in banking and Major Markets-A Ten-Year Perspective," lntema tiona/ Capital Markets (Washington: International Monetary Fond, securities activities worldwide. 1988) and in Bank for International Settlements, Recent Innovations The United Kingdom-long open to foreign banks in InternationalBanking (Basle: Bank for InternationalSettlements, freed the entry of fo reign financial firms into securities 1986). activities in tate 1986, in the context of London 's "Big $Note issuance facilities are medium-term arrangements tbat allow Bang." Further market and regulatory changes since borrowers to issue short-lerm notes, backed by underwriting com 1986 have significantly widened the range of business mitments of commercial banks, in the Euromarkets.

©International Monetary Fund. Not for Redistribution l · INTRODUCTION AND SUMMARY and options were introduced on major exchanges, lions; in 1989, turnoverreached $120 billion per day in wh ile the turnovervolume of existing types of contracts New York alone. Bank lending to nonbanks in indus increased considerably. Finally, a variety of new trial countries expanded only moderately; borrowers instruments-complex swaps, hedges, and such preferred to tap the bond markets. Following the onset arrangements as international mutual funds, unit trusts, of the debt crisis in 1982, bank lending to developing and country funds for cross-border investments-were countries was sharply reduced, as a reflection of the made available to investors. serious debt-servicing difficulties of certain countries, Te cJznologicaladvances in communications and data and creditworthy sovereign borrowers made greater processiog supported the trend toward financial inno recourse to the bond markets. vation. New developments in computer technology, Capital ftowsin recent years have been characterized software, and telecommunications permitted more by increasing reliance on securities, that is, a greater rapid processing and transmission of information, use of bonds and other debt instruments in comparison improved settlement of transactions, and Jess costly with financing provided by deposit money banks. This confirmation of payments. By linking exchanges in dif change alone weakened the statistical systems that ferent time zones, these advances ushered in 24-hour depended large!y on reports from financialintermediar global trading for most markets. In consequence, the ies. Syndicated loans were increasingly replaced by domestic financial markets of Europe, North America, issuance of international bonds or other types of Euro and Japan became increasingly integrated and linked market paper. The international bond market by large capital flows. expanded continuously during the 1980s, but was mainly limited to borrowers from industrial countries. The estimated outstanding stock of international bonds Trends in International Financial Flows in the 1980s increased from about $260 billion at the end of 1982 As a result of these processes of liberalization and (equivalent to 3 percent of industrial countries' annual innovation, international capital flows expanded rap GDP) to about $1 ,400 billion at the end of 1990 (more idly in the course of the 1980s.6 After 1982, divergent than 8 percent of their annual GDP). saving and investment patterns among industrial coun Cross-border equity trading also was a dynamic seg tries resulted in large current account imbalances and ment of the international financial markets in the 1980s; were financed by major shifts in capital flows. The the volume of international equity transactions sum of the current account deficits of the industrial exceeded $1 ,500 billion by 1989, in comparison with countries increased from an average of $60 billion in only $73 billion a decade earlier. Foreign investors 1982-83 to more than $200 billion by the end of the increased their portfolio allocation in foreign securities decade. The scale of gross capitaljlows, stimulated by in arder to take advantage of possible gains abroad increased cross-border banking and securities activity, and to hedge against domestic risks. Traditional direct the development of the offshore (Eurocurrency) mar fo reign investment was augmented by large cross kets, and the entry of foreign institutions into domestic border mergers and acquisitions and international joint markets, expanded even more ra pidly. Priva te financial ventures, especially in tbe United States and the Euro flows were the most dynamic source of capital flows pean Commuoity. and played an increasingly important role in the financ ing of current account imbaJances among industrial countries. Scaled against the huge volume of interna Economie Policy Concerns tional financial transactions taking place, the discrep ancies in recorded international capital flows are rela Progressive deterioration of the quality of informa tive!y small. tion on international capital ftows can undermine the Cross-border bank lending expanded significantly, conduct of national economie policy and international but irregularly, in the 1980s. The total of outstanding policy coordination in a number of ways: international bank credit (net of interbank daims) rose from $175 billion at the end of 1973 (equivalent to a. lnconsistencies in current and capital account 5 percent of industrial countries' annual gross domestic recording at both the national and global levet product) to $3,430 billion at the end of !990 (about may indicate errors in national information on 20 percent of industrial countries' annual GDP). The saving and investment. Such errors may mislead volume of daily transactions on the exchange markets policymakers in basic choices about fiscal and was boosted by the liberalization of interbank transac- monetary strategies. b. With greater freedom for capital to react to shifts in policy, it is important to anticipate how mone tary policy actions may be affected by cross 6A detailed discussion of the economie factors thal triggered the border capital flows. For example, an intended expansion of international capital tlows is included in Determinants and Systemic Consequences of lntemational Capital Flo ws, Occa tightening of credit markets may be frustrated by sional Paper No. 77 (Washington: lnternalional Monetary Fund, capital inflows, or an intended increase in taxes 1991). Stalistical informalion is drawn primarily from various issues may be aborted by a capital outflow. Similarly, of the Fund's International Capital Markets. better information about capital flows may help

©International Monetary Fund. Not for Redistribution Working Party Program

to guide exchange rate polie y, especiaUy when the same transaction may be recorded under different capital ftows have asignificant immediate impact rubrics in national balance of payments accounts. For on the exchange markets. example, white the balances for direct and portfolio c. The enhanced speed and volume ofinternational investment should approach zero at the global level, a financiaJ transactions also create greater expo capital outftow recorded as a direct investment by the sure to possible crises in the clearing systems for creditor country may be recorded under sorne other international banking accounts, should a major category by the host country if different classification participant be unable to meet scheduled obliga criteria are used. Similarly, an increa se in a country' s tions. The availability of accu rate and timely data reserves in the form of securities will be recorded as on capital ftows can help national authorities portfolio investment in the country issuing the securi assess such risks. ties. Consequent!y, when the researcher finds a dis d. There are now many occasions when groups of crepancy in any of the capital account categories (see countries seek to coordinate their responses to Table 3), there is uncertainty as to whether the problem problems in the international economie system, is one of classification differences or whether one or or wbeo the IMF exercises its surveillance func the other country simply has erroneous or missing data. tion. At such times, timely and credible data on The categories combined in "other" sectors are sub relevant developments in capital ftows are vital. divided by type of transactor, so in this ru bric the e. lt is important for the authorities and the public problems of matching inftows and outftows are even to have accurate information on the structure of more severe. the capital ftows affecting a country, thal is, on the distribution between relatively stable private ftows (such as direct investment), ftowsof liquid private capital (such as short-term deposits in Working Party Program banks), or interofficial financing. Changes in this structure and in the terms of financing may have The Working Party's firstmeeting was held in March policy implications. 1990 in Washington; subsequently, it met at approxi f. For direct investment, capital ftows data giving mately three-month intervals. From the beginning, its information on the types of enterprises being aim was to uncover and correct problems in global financedare useful for analysis of the functioning capital accounts, whether or not such corrections of the host and creditor economies. raised or lowered the global discrepancies. The Work g. Countries with large foreign debts and the ir cred ing Party decided to focus on a fo ur-year interval, 1986 itors need to have accurate statistics about such to 1989, and il organized its study of international capi debts and the ways in which capital ftows, both tal ftowson several levels. Each main category of capi inward and outward, are affecting debt manage tal ftows was studied intensively on a global scale. ment. Data reported by many individual couotries and the procedures used to collect such data were reviewed Additional instances can be listed, but the Working and discussed with national compilers. Certain studies Party's main concern about the connection between were carried out to test the applicability of alternative capital flow statistics and the policy process can be data sources, and a large number of specialized reports stated succinctly: there are strong indications that this and analyses were provided by members of the Work body of information on which good economie manage ing Party. the technical staff, and sorne outside experts. ment depends is undergoing a serious and progressive The research drew extensively on the information and deterioration. The size of recorded discrepancies in the experience of the Fund's Statistics Department and global capital account is not the only sign of deteriora also was greatly assisted by the professional staffs of tion. Many errors in the compilation of such data may the World Bank, OECD, BIS, and Statistical Office of be offsetting, or transactions may be missed entirely. the European Communities (Eurostat). Thus, problems with the figures may be worse than is The Working Party's principal lines of inquiry were immediate! y apparent. Concernssuch as these contrib through compilers in individual countries, especially uted to the decision to set up the present Working those nations with large capital ftows. lt established Party. close contact with many countries to consult with them Il is crucial to bear in mind that research on interna about the details of their published accounts and tional capital ftows is far more complicated than work explore their concepts and procedures. At an early on current account transactions. For the latter, classi stage of the research, it became apparent thal much fications are clearer and transaction balances for a more than the published capital accounts would be larger number of component accounts (such as trade, needed. The Working Party therefore designed and travet, investment income, and transfers) should by circulated to about 70 countries a "Special Question definitionbe zero for the world as a whole. Thus, there naire on International Capital Flows." The question is an indication of how rouch correction is needed for naire requested expanded details of countries' capital each principal category. With capital ftows, however, ftows, investment positions, income accounts, and sta there is much greater likelihood thal the two sides of tistical procedures. The responses led to many insights

©International Monetary Fund. Not for Redistribution I · INTRODUCTION AND SUMMARY into country practices and statistical weaknesses and of earnings. The suggested adjustments correct for mis also served as a basis for fo llow-up inquiries about classifications and for missing or inaccurate data in important points. At a later stage, the Workjng Party national sources. (Statistical procedures to deal with conducted special surveys of smaller groups of coun these problems also are suggested in Chapter 3.) The tries to obtain additional information on direct and Working Party's adjustments have reduced consider portfolio investment, on the composition of foreign ably the original pattern of annual net global outtlows exchange reserves, on debt statistics, and on details of for direct investments. banking data submitted to the BIS as compared with The Working Party believes that a crucial factor in banking figures in the balance of payments accounts. improving direct investment statistics is closer adher Of particular value to the Workjng Party were the ence to the requirements of the Fund's Manual. ln Fund's internationalbankjng statistjcs (lBS), which are particular. surveys of direct investors and affiliates are based primarily on the BIS system, data on interna needed in order to compile information on "noncash" tional securities collected by the BIS. data on cross transactions such as reinvestment ofearnings. National border debt collected by the World Bank, and special compilers also can benefit (as is demonstrated in this ized compilations developed by financial market parti report) from using data collected in partner countries cipants and statistical services. These external sources when domestic sources are inadequate. provided, to sorne degree, a check on certain aspects of countries' national data, especially for bank-related Portfo üo Jnvestment transactions and transactions in portfolio securities. Wherever adequate geographie details were available Most of the recent innovations in international in balance of payments accounts, consistency among financial practices have borne direct! y on the measure the accounts of partner countries was also checked. ment of transactions in securities. Consequently, the However, geographie detruis are limited, so bilateral difficulties of capturing data on these transactions in comparisons yielded significant results primarily for national statistical systems have grown to the point direct investment. where the validity of sorne of the available data is increasingly in doubt. Under current practices the pos sibilities for verifying the accuracy of data on cross border purchases and sales of securities are quite lim Findings for Major Capital Flow Categories ited. The main avenue for verification of national data is comparison with independent statistics on new issues The global capital flow discrepancy indicates thal and redemptions of securities, together with sorne recorded capital outflows have been relatively under detailed data compiled in a few countries. The Working stated. This bias is consistent with the earlier findings Party has employed figures collected by the BIS on of the Working Party on the Statistical Discrepancy new bond issues and redemptions and on outstanding in World Current Account Balances as to the major amounts of bonded debt as major reference points. sources of the discrepancy in the measurement of Reference to the BIS bond data and consultations investment income.7 Jn the sections thal fo llow. many with individual countries have turned up a number of specifie corrections have been suggested for transac cases in which corrections to portfolio flowswere indi tions as reported by individual countries, and results cated (see Chapter 4). On the basis of this research, a have been grouped by type of capital flow. Suggested variety of adjustments to recorded data can be recom adjustments have been summarized at the global level mended to national compilers. The net result of these in Table 4.8 adjustments, shown in Table 4, has been to add net inflows to the Yearbook-reported balances in each of Direct lnveshnent the four years under study. The se adjustments mainly reftect transfers to the portfolio sector of international Most of the problems with direct investment statis bond sales that were recorded by a number of borrow tics have arisen from failure to collect adequate data or ing cou nt ries as loans under ''other capital." Almost from deviations from international standards as estab ali such adjustments were made to the accounts of lished in the Balance of Payments Manua/. On the the issuers of securities, sinee the treatment by, and basis of research detailed in Chapter 3, the Working residence of. the purchasers was usually unknown. Party fo und that extensive modifications are indicated for reported capital tlows, especially for reinvestment Other Capital

7 Capital flows grouped under the broad heading of See Report on the World Current Account Discrepancy (Washing "other capital'' consist of tlows not included in direct ton: International Monetary Fund. 1987). investment, portfolio investment, reserves, or lia 8In sorne cases. cstimated amounts of the unrecorded capital out flow can be traced to individual countries, but these attributions are bilities constituting foreign authorities' reserves not without challenge. ln general, the Working Party has opted to (LCFAR). In the Ye01·hook accounts, this "other" aggregate ils findings into three regional categories: industrial coun group is subdivided by major category of resident tries, developing countries. and offshore financial centers. transactor: {1) official, (2) deposit money banks, and

©International Monetary Fund. Not for Redistribution Findings for Major Capital Flow Categories

Table 4. Global Capital Account Discrepancies and Adjustments, by Major Components, 1986-89 (ln billions of U.S. dollars: outflows ( - )) 1986-89 1986 1987 1988 1989 average Capital account, total1 18.0 32.6 45.1 65.8 40.4 Adjustments2 -15.0 -24.2 - 13.4 -40.3 -23.4 Adjusted capital account, to La l 3.0 8.4 31.7 25.5 17.0

Direct investment -13.6 -24.5 -19.8 -8.2 -16.5 Adjustments 16.9 14.6 6.2 9.5 11.8 Adjusted direct investment 3.3 -9.9 -13.6 1.3 -4.7

Portfolio investment 3.6 -0.7 -11.9 32.3 5.8 Adjustments 3.5 11.9 6.3 7.7 7.4 Adjusted portfolio investment 7.1 11.2 -5.6 40.0 13.2

Other capital 32.7 110.5 76.6 45.6 66.4 Adjustments -41.8 - 102.9 -25.6 -63.9 -58.6 Adj usted other capital -9.1 7.6 51.0 -18.4 7.8

Reserves/LCFAR -4.7 -52.7 0.2 -3.9 -15.3 Adjustments 6.4 52.2 -0.2 6.3 16.2 Adjusted reserves/LCFAR 1.7 -0.5 2.4 0.9 'Balance of Paymenrs Sllltistics Yearbook. Part 2, 1990. 2Adjustments to each discrepancy include relevant corrections to LCFAR as described in Chapter 8.

(3) other private non banks (including sornepublic agen rowers reside, although few countries have closely des). "Other capital" ftows include a wide variety of matching data categories. If the derived data clearly financial activity, ranging from !rade credit and depos show a larger volume of ftows (or stocks) vis-à-vis its with foreign banks to marketable short-lerm instru fo reign banks than does the national balance of pay ments and loans from foreign banks, governments, and ments accounts. however, it is reasonable to interpret nonbank tenders. The private nonbank group includes the diffe rence as an indication of the extent to which many types of financial institutions (such as insurance the balance of payments accounts are deficient. companies, finance companies, and security dealers) The statistical adjustment to "other capital'' aver that have large cross-border ftows. Many countries aged close to -$59 billion over the 1986-89 interval have tended, moreover, to categorize as "other" any under review and accounted for the bulk of the net capital ftows that cannot be specifically identified. As reduction in the global imbalance acbieved by the mentioned earlier, the sector accounts in "other capi Working Party. As shown in Table 5, about half this tal'' are not individually "symmetric" across the adjuslment, close to -$30 billion, was derived from world, nor do they co ver ali transactions of the sectors. the internationaJ banking data. The table also shows When consolidated, however, they should balance at thal adjustments to recorded capital ftows of private the global level, and it is only at thal level thal balance nonbanks dominate these results. This was by far the should be expected. largest single block of missing capital outftows that The main findingsof the Working Party in the "other the Working Party was able to identify. However, as capital" category have resulted from comparisons of lBS data-on a flow basis as calculated by the IMF with national balance of payments figures on changes in the claims and liabilities of private nonbanks with Thble S. Adjustments to "Other Capital" lndicated by banks resident in other countries. While many coun International Banking Statistics, 1986-89 1 tries have captured the externat operations of their (ln billions of U.S. dollars; outflows (-)) banking sector fairly accurately in national balance of 1986-89 payments statistics, most have experienced great dif 1986 1987 1988 1989 average ficulty in obtaining reliable data from private nonbanks Total adj ustments -21.3 -33.7 -5.6 -57.7 -29.6 on their claims on or borrowings from foreign banks. By sector Consequently, the coverage of this activity by many Private nonbanks -16.1 -24.7 3.8 -48.7 -21.4 countries is deficient. Deposit money banks -5.2 -9.0 -9.4 -9.0 -8.2 Data reported to the BIS and the Fund by banks in the principal financial centers can be used to derive the By area amounts deposited with them and borrowed from them Offshore financial centers -11.3 -9.3 -10.8 -3.1 -8.7 by residents of individual fo reign countries (see Chap Olher countries -10.0 -24.4 5.2 -54.6 -20.9 ter 6). This information can be compared, in principle, 1Adjustments for nonbanks are described in Chapter 6. Adjust with corresponding data in the baJance of payments ments for deposit money banks are for offs hore financial centers and accounts of countries where the depositors and bor- are described in Chapter 9.

©International Monetary Fund. Not for Redistribution 1 · INTRODUCTION AND SUMMARY discussed in Chapter 6, comparisons between the bal reported by countries where the funds are invested. ance of payments and lBS data sources must be used should balance at the global levet but, in practice, they with caution. On a global scale the results indicate the do not. As shown in Table 3, net increases in reserves general size of omissions, but at the individual country (outftows) reported by asset holders generaUy exceed levet, national compilers should make a specifieevalua the LCFAR inftows reported by obligors. The imbal tion of the quality oftheir own data and decide whether ance between these ftows has implications for discrep their accounts could be improved by substituting the ancies in other categories of capital ftows. The Working IBS/BIS figures. A number of countries have already Party investigated the reserves/LCF AR discrepancy begun to use these data. primarily from the reserves-or asset-side of the pic Adjustments for deposit money banks in Table 5 ture {see Chapter 8). For the purposes of this study, a bring into the world totals the banking activity in off large number of countries provided detailed informa shore financial centers (OFCs) that is omitted from tion on the types of instruments (such as securities and national balance of payments statements. These deposits) in their reserve transactions for the 1986-89 amounts are reported directly by the OFCs in the con period. This information enabled direct comparisons text of the lBS statistics and are therefore net additions of reserve movements-by type of instrument-to be to capital Row totals rather than adjustments to existing made with LCF AR transactions as reported by debtor estimates as is the case for nonbanks. The banking countries. The same data made it possible to isolate adjustments wcre part of a broader picture for OFCs certain sources of error in the LCFAR figures, to cor thal is explaîned in Chapter 9. rect the rcscrvcs/LC FAR imbaiance, and to make addi Adjustments to the "other capital" category, apart tional adjustments to the portfolio and other capital from those derived from banking data, consist mainly accounts. of reclassifications thal shi fttransactions into the port As part of its reserves-rclated work, the Working fo lio investment and LCFAR categories. An average Party a1so found problems with the recording, by sorne inftow ofabout $10 billion per year has been removed members of the European Monetary System (EMS), of from the resident official sector of "other capital" in gold swaps with the European Monetary Cooperation the 1986-89 period and shifted into portfolio invest Fund. These problems required adjustments to the ment ftows. Similar amounts have been shifted from reported foreign exchange and gold transactions ofsev deposit moncy banks into portfolio investment. Data adjustments for short-lerm securities also affect this era! EMS-member countries. category. (These changes are described in Chapters 4 and 5.) ln establishing the terms ofrefe rence for the Working OveraU Statistical Results Party, the Executive Board called attention to the prob lem of concealed international capital ftows associated with drug money and other illegal activities. The Work Adjusted Global and Regional Accounts ing Party investigated various aspects of su ch ftows,as The Working Party's statistical results affecting the weil as the efTects on reported capital ftows of so-called world capital account can be integrated with adjust capital ftight. lts conclusion was thal information avait ments to the current account prepared each year by the able to statistical agencies on such activities was far Fund on the basis of the 1987 Report on the World from sufficient to allow for a separation of such finan Current Account Discrepancy. The initial and revised cial dealings from the general run of international views of world patterns of international transactions financial traffic (see Chapter II). are shown in Table 6. Comparisons of the reported Almost by definition, capitalRight has not been regis and adj usted balance of payments accounts of major tered in the capital-exporting country but often may country groups reveal the fo llowing major differences: have been recorded in recipient countries. Thus, this phenomenon may weil be a significant contributor to a. The large excess of debits in the current account the observed global excess of recorded inftows over has becn nearly eliminated. There was a substan outflows. However, estimates of the extent of capital tial addition of net credits to developing countries flight have been wide ranging, and they cannot identify and a smaller addition to industrial countries. countries where the funds were invested. The amount The added credits were primarily income on of capital Row associated with drug trafficking is also portfolio investment and unrequited transfer highly uncertain, but surely is much smaller than corn receipts. monty cited estimates of the retail value of drug sales b. Although a sizable discrepancy remains, the in national markets. Moreover, these Rows merge excess of credits in the global capital account indistinguishably with other kinds of poorly recorded (nearly as large as the opposite discrepancy in outftows. the current account in 1988 and 1989) basbeen greatly reduced by the adjustments suggested by Reserves and Liabilities Constituting Foreign the Working Party. The effect of the adjustments Authorities' Reserves has varied on an annual basis, but for the entire ln principle, transactions in foreign exchange four-year period there was only a rninorchange reserves, as reported by holders, and in LCFAR, as for the industrial countries. For the developing

Thble 6. Global Balance of Payments Accounts As Reported countries (other than OFCs) there was a drop and Adjusted, 1986-891 from an average annual net inftow of $l2 billion, (ln billions of U .S. dollars; outftows (-)) as reported, to a much smaUer $1 biJHon inflow. 1986-89 The adjustments made by the Working Party to 1986 1987 1988 1989 average create international capital accounts for the As reported in 1990 Yearbook OFCs resulted in increased net outflows-from World2 an annual average of about $5 billion to nearly Current account -59.0 -45.6 -52.5 -75.6 -58.2 Capital account 18.0 32.6 45.1 65.8 40.4 $14 bilJion. c. Finally, the genera)( y downward shift in the lndustrial countries regional errors and omissions reftects the fact Current account -28.2 -57.8 -52.7 -82.5 -55.3 Capital account -2.4 41.1 41.4 73.6 38. 4 that positive adjustments made to the current Errors and omissions 30.7 16.8 11.3 8.9 16.9 account were larger than negative adjustments made to the capital account. Nevertheless they Offshore financial centers) Current account 3.0 2.0 3.8 5.6 3.6 provide an alternative view of the broad eco Capital account -12.6 -1.7 -2.3 -2.6 -4.8 nomie relationships among sorne major country Errors and omissions -9.6 -0.3 -1.5 -3.0 1.2 groupings and provide a point of departure for Othcr developing countries further explanations of the measurement prob Current account -37.9 3.7 -9.1 -5.4 -12.2 lems that re main. Capital account 35.9 1.2 12.2 -0.4 12.2 Errors and omissions 2.0 -5.0 -3.1 5.8 -0.1 Actions Needed International organizations Current account 4.1 6.6 5.5 6.7 5.7 On the basis of its studies, the Working Party has Capital account -2.8 -8.0 -6.2 -4.8 -5.4 made a series of recommendations for actions to be Errors and omissions -1.3 1.5 0.7 -1.9 -0.3 taken at the national and international level to improve the data on international capital flows. It believes that As adjusted World2 implementation of these actions would produce sig Current account -27.1 -10.2 0.8 -9.5 -11.5 nificant gains in the quality of data available to policy Capital account 3.0 8.4 31.8 25.5 17.2 makers and the public. The role of the Fund in carrying lndustrial countries forward these recommendations will be crucial. Since Current account -22.2 -50.3 -37.0 -63.7 -43.3 capital flow adjustments are very heterogeneous and Capital account 7.6 32.3 55.6 49.2 36.2 specialized, the Fund will need to coordinate its work Errors and omissions 14.6 17.9 -18.6 14.5 7.1 with the activities of national compilers. Many of the Offshore finaocial centers3 specifie findings of the Working Party (including much Current account 5.3 5.2 8.8 12.6 8.0 country detail not shown in this report) can be used by Capital account -29.2 2.1 -15.7 -11.7 -13.6 the Fund in reviewing capital accounts reported by Errors and omissions 24.0 -7.3 6.9 -1.0 5.7 countries in the future. Other developing countries AJl countries can benefit from greater accuracy in Current account -21.0 22.2 14.8 22.0 9.5 their international capital accounts, but those whose Capital account 27.0 -18.6 -1.2 -3.6 0.9 Errors and omissions -6.0 -3.6 -13.6 -18.4 -10.4 transactions weigh heavily in the world totals have a special responsibility to improve the quality of their International organizations statistics. As Table 7 shows-apart from sorne OFCs Current account 4.1 6.6 5.5 6.7 5.7 Capital account -2.4 -7.4 -6.9 -8.4 -6.3 for which comparable data are not available-rela Errors and omissions -1.7 0.8 1.4 1.7 0.6 tively few countries account for the major part of recorded world capital ftows. The Working Party rec Unallocated Current account 6.8 6.1 8.6 12.9 8.6 ommends that these /eading countries make a particu Capital accouot lar effo rt to upgrade and harmonize their capital Errors and omissions -6.8 -6.1 -8.6 -12.9 -8.6 account statistics. 1Current account as adjusted by the Statistics Department. plus adjustments by the Working Party for reinvestment of earnings and for special financial institutions; capital account as adjusted by the Prospects and Costs Working Party. 2World totals as shown indicate discrepancies. because in principle Finally, the Working Party is concerned that ali the these global aggregates should total to zero. evidence points to a progressive degradation of the -'The Bahamas. Bermuda, Cayman Islands. Gibraltar. Hong Kong, quality of data in international capital flows. For exam the Netherlands Antilles, Panama, and Singapore. ple, the recent! y compiled 1991 Yearbook shows a con tinuation of large global capital account imbalances. In the long run, basic improvements in the data on capital flows can be achieved only when national compilers have the active support of officiais responsible for pol icy decisions. ln order to accomplish such improve ments, many countries may have to restructure or

Table 7. Countries Ranked by Reported 1986-89 Capital ongoing. Of course, national compilers will continue to Flows1 depend to a considerable extent on the willingness and (Jn billions of U.S. dollars) ability of market participants to provide the rather spe Cumulative cialized data needed for the preparation of the capital Capital Percent percent accounts. Certainly, a primary requirement for secur Country Rank ftows of total of total ing this cooperation is a convincing explanation to data Japan 1 254.7 26.93 26.93 suppliers that these data are essential and that responsi United States 2 155 .2 16.41 43.34 ble national authorities give a high priority to support Unhed Kingdom 3 127.7 13.50 56.84 German y 4 63.7 6.73 63.57 ing this statistical work. Belgium-Luxembourg 5 57.3 6.06 69.63 France 6 53.9 5.70 75.33 Italy 7 25.3 2.67 78.00 Netherlands 8 25.1 2.65 80.66 Switzerland 9 24.3 2.56 83.22 Addendum: Revised Data from the 1991 Ye arbook Canada 10 17.8 1.88 85.10 Australia Il 13.4 1.42 86.52 As the Working Party was completing its assignment, Sweden 12 11.7 1.24 87.76 Spain 13 10.5 1.11 88.87 the Statistics Department at the Fund made available Panama 14 6.2 0.65 89.52 the statistics to be published in the 1991 edition of the Denmark 15 5.8 0.62 90.14 Yearbook. The new global capital account data are Singapore 16 5.4 0.57 90.71 shown in Table 8 (compare with Table 3). Finland 17 4.9 0.52 91.22 The new edition contains revised data for the period China, People's Rep. 18 4.8 0.51 91.73 Saudi Arabia 19 4.5 0.47 92.20 to 1989 and initial figures for 1990. At the global levet Korea 20 4.3 0.45 92.66 the revised figures show somewhat larger net positive Norway 21 4.0 0.43 93.09 discrepancies up through the 1989 peak. The 1990 dis A ustria 22 3.7 0.40 93.48 crepancy is about equal to the 1986-89 average of the Ku wail 93.82 23 3.3 0.34 data in Table 8. Among the principal components, there Mexico 24 3.0 0.32 94.14 lreland 25 2.9 0.31 94.45 are a few notable shifts: Ali countries 185 945 .8 100.00 100.00 a. For direct investment a much larger negative dis Source: Balance of Paymems Statistics Yearbook. Part 2. 1990, crepancy emerges in 1990; about half of the rise Table C-13. reftects an increase of $13 billion in net out ward 11986-89 average of absolute values ofcapital inftows plus capital reinvestment of earnings. The largest net revi outftows. sion to national data apparent at this writing was for the United States, which raised net inward direct investment ftows by $11 billion in 1987. b. For portfolio investment, there is sorne increase in the net positive discrepancy reported for 1989, adapt their compilation procedures. ln that connection, but the size of the imbalance declines in 1990. the new Balance of Payments Compilation Guide pre c. The heterogenous "other capitaJ" category is not pared by the Fund will be useful. significantlyaltered for the 1986-89 period, but in The Working Party recognizes the need to contain 1990 an exceptionally large net positive balance costs and reporting burdens but believes it is possible, emerges. The main changes in 1990 are large net especially with modern computing facilities, to inftows reported for both deposit rnoney banks enhance the quality of balance of payments data at and the nonbank private sector. Banking inflows moderate expense. To a large extent, the costs associ may reftect the investment of the large increase ated with adapting compilation systems to new scales in global reserve assets reported for the year that and forms of capitaJ ftows are transitional rather than is not matched by recorded LCFAR.

Table 8. Global Capital Account Discrepancies 1984-90 (ln billions of U.S. dollars: outftows (-)) 1984 1985 1986 1987 1988 1989 1990 Capital account, total 43.3 62.2 30.8 61.5 57.3 75.4 57.4 Direct investment 4.6 -8.8 -16.9 -17.1 -19.0 -22.9 -49.6 Portfolio investment 30.6 52.2 10.0 17.5 -0.4 44.9 12.2 Other capital 27.2 26.0 41.7 115.5 76. 1 58.8 157.5 Reserves!LCFAR -19.1 -7.2 -4.0 -54.4 0.6 -5.4 -62.7

Memoranda: Current account -73.1 -74.0 -58.8 -44.5 -52.0 -78.5 -79.3 Errors and omissions 29.8 11.9 28.2 -17.1 -5.0 3.1 21.9 Source: Balance of Payments Statistics Yearbook. 1991.

©International Monetary Fund. Not for Redistribution Addcndum: Revised Data from the 1991 Yearbook d. Finally, the increase in reserve assets reported of opposite signs continue to be reported for the for 1990 is far in excess of the reported increase global current and capital accounts. As noted in in LCFAR. This could reflect a sruft in the hold this report, although many factors are at work, a ings of foreign exchange reserves to banks in strong element linking these discrepancies is the countries that fajl to report such liabilities as difficulty of recording certain types of capital obligations to reserve authorities. Thus, the outflows and of measuring the income receipts imbalances in the "other capital" and "reserve/ corresponding to the rising stock of external LCFAR" categories may be closely related. assets not recorded in the relevant capital Sorne further anaJysis would be in order. More exporting countries. generally, it can be seen that large discrepancies