Abstract

Despite the serious disruptions in early 1995 as a result of the Mexican crisis, total net capital flows to developing countries and countries in transition reached a record $228 billion in 1995. Notwithstanding early fears of widespread spillovers of Liquidity problems in Mexico, markets appeared relatively quickly to distinguish between those countries with sound fundamentals and those that seemed to share some features of the Mexican economy, which then saw at least temporary declines in market access and capital inflows. This regional differentiation is revealed by data on capital flows and securities issues. But even those countries that experienced the most serious contagion effects had almost fully regained access to international financial markets at precrisis spreads by the end of the year. Moreover, some countries that at first glance seem to have been relatively untouched by the crisis, particularly those with pegged or fixed exchange rates, were tested by international investors. In addition to the change in the regional nature of capital flows, there was a change in the composition of capital flows, with a decrease in portfolio investment and an increase in bank lending to developing countries and countries in transition.

Despite the serious disruptions in early 1995 as a result of the Mexican crisis, total net capital flows to developing countries and countries in transition reached a record $228 billion in 1995. Notwithstanding early fears of widespread spillovers of Liquidity problems in Mexico, markets appeared relatively quickly to distinguish between those countries with sound fundamentals and those that seemed to share some features of the Mexican economy, which then saw at least temporary declines in market access and capital inflows. This regional differentiation is revealed by data on capital flows and securities issues. But even those countries that experienced the most serious contagion effects had almost fully regained access to international financial markets at precrisis spreads by the end of the year. Moreover, some countries that at first glance seem to have been relatively untouched by the crisis, particularly those with pegged or fixed exchange rates, were tested by international investors. In addition to the change in the regional nature of capital flows, there was a change in the composition of capital flows, with a decrease in portfolio investment and an increase in bank lending to developing countries and countries in transition.

This annex discusses developments in 1995-96 in the capital markets of developing countries and contries in transition and examines both the available balance of payments capital flow data and data on transactions in the primary and secondary markets for debt and equity instruments. It then describes recent developments in investment flows from the United States and Japan. Finally, the annex highlights recent developments in the banking systems of selected countries, an element of the “fundamentals” that has increased in prominence since the Mexican crisis.

Capital Flows in the Balance of International Payments

Total private capital flows to developing countries and countries in transition increased by 29 percent in 1995 to $211 billion, higher than in any previous year (Table 23).1 Capital flows to Asian developing countries and to the countries in transition accounted for almost all of the increase in flows. The share of total flows to Asian developing countries remained unchanged, at 50 percent, but the share of flows directed to the countries in transition doubled, to 14 percent. Flows to Latin American countries remained essentially unchanged over 1994 levels at $49 billion, resulting in a decline in that region’s share of total flows to 23 percent, from 30 percent in 1994, while flows to Africa actually declined. These changes in the regional distribution of total capital flows generally reflect the differences in severity of spillovers from the Mexican liquidity crisis across countries. The “Tequila effect” was strongest in Latin American countries, such as Argentina and Brazil—also important participants in the international capital markets—and less so in Asian developing countries, which saw only short-lived contagion effects.2 Similarly, developments in Mexico did not affect the countries in transition to a great extent.

In addition to the geographic reallocation of capital flows to the countries in transition, there was an important change in the composition of capital flows—a doubling of "other" capital flow’s, which is manily bank lending.3 These capital flows increased from $43 billion in 1994 to $85 billion in 1995, increasing their share in total flows from 26 percent in 1994 to 40 percent in 1995. Other capital flows rose in all regions except Africa. The share of capital flows accounted for by foreign direct investment declined slightly to 39 percent, although direct investment flows in 1994 had been at an unusually high level. Portfolio capital flows fell sharply, however, both in levels—from $53 billion in 1994 to $43 billion in 1995—and in proportion to total flows, declining from 33 percent of the total in 1994 to 20 percent in 1995.

Table 23.

Private Capital Flows to Developing Countries and Countries in Transition

(in billions of U.S. dollars)

article image
Source: International Monetary Fund. World Economic Outlook database,

Short- and long-term trade credits, loans (not including use of IMF credit), currency and deposits, and other accounts receivable and payable.

A positive sign indicates a net increase.

There also were noteworthy changes in the composition of flows within regions.4 In Asian developing countries, although there was a slight increase in importance of other flows at the expense of portfolio flows and foreign direct investment, the change in proportions was not large. Indeed, the important changes in composition of flows to Asian developing countries occurred in 1992 and 1994. In 1990-92, Asian countries relied heavily on bank borrowing, with the “other” category accounting for more than 64 percent of total capital inflows and foreign direct investment accounting for 30 percent or more. Portfolio capital flows were negligible. In 1992, other capital flows declined sharply, as portfolio investment increased to 32 percent of total flows and foreign direct investment rose to 47 percent. While the share of foreign direct investment has stabilized at about 50 percent of total capital flows since 1992, portfolio investment dropped sharply in 1994 to just below 20 percent, and remained at that level in 1995.

Among the Latin American countries, other capital flows nearly doubled in 1995 to $23 billion, representing 45 percent of capital flows to the region. This performance matched the level of such flows in 1992 both in dollars and percentage terms. Portfolio capital flows, conversely, declined sharply in 1995 to $10 billion, the lowest amount in six years. Countries in this region have tended to rely more heavily on portfolio investment than countries in other regions except the Middle East, where foreign direct investment has been insignificant. Hence, the withdrawal of foreign portfolio investors from the international markets in 1995 seriously affected capital flows to Latin American countries.

Foreign direct investment flows have recently been relatively more important for the Asian developing countries than for Latin American countries, where such flows accounted for only 35 percent of net inflows in 1995. A better comparison, however, is the ratio of net foreign direct investment to GDP, which was 1.9 percent for the Asian developing countries in 1995 and 1.1 percent for the Latin American countries. The relatively higher ratio in the Asian developing countries, however, goes back only to 1993. During 1990-92, Latin American countries attracted more foreign direct investment in proportion to GDP than did the Asian developing countries.

The different patterns of capital flows to Asian and Latin American developing countries have been attributed to a number of factors, and in particular to the relatively faster growth rates among the Asian countries.5 Private capital inflows to Asian developing countries in 1995 were 3.8 percent of GDP, compared with 3.0 percent for the Latin American countries; in 1994, the proportions were 3.4 percent and 3.1 percent, respectively. However, despite having had a higher average real growth rate than the Latin American countries since 1990, the developing countries in the Middle East and Europe received capital inflows of only 1.5 percent of GDP in 1995 (1.1 percent in 1994). Among these regions, the ratio of direct investment to portfolio investment also does not appear to be strongly correlated with growth. The countries in transition have had a slightly higher ratio than the Asian developing countries during 1990-95 despite a negative real growth rate.

The success in attracting foreign direct investment to the countries in transition suggests strongly that it is not just growth that matters, but also the structural characteristics of the economy. In particular, the Asian developing countries and the countries in transition are seen by some investors as providing more attractive investment climates because their goods and labor markets are perceived as being more liberal than those in Latin America, for example. Also, much foreign direct investment appears to be motivated by the desire on the part of multinational corporations to locale production facilities in low-labor-cost countries, an explanation that applies in particular to Japanese direct investment in the Asian region. Another factor that might explain the relatively higher proportion of portfolio capital flows to Latin America compared with Asian developing countries is that Latin American securities markets are more fully developed than those in the Asian developing countries. With more mature and liquid domestic securities markets, these countries provide a greater range of instruments through which foreign funds can be invested, which may explain why they attract relatively more portfolio capital.

Capital flows to the countries in transition increased sharply in 1995 to $30 billion, up from $11 billion in 1994. These countries accounted for 14 percent of total capital flows in 1995, compared with only 7 percent in 1994, As in the other regions, the main source of growth in capital flows was from bank lending. Thus, despite a doubling of foreign direct and portfolio investment, the share of other capital flows in total capital flows to these countries rose to 42 percent, from 24 percent in 1994. The sources of external financing for these countries has diversified considerably since 1993, when foreign direct investment accounted for 83 percent of inflows. Indeed, during 1991-93, other capital flows were negative.

Large capital inflows can have important macro-economic consequences in the recipient countries, including an appreciation of the currency or an expansion in the money supply.6 In 1995, central bank reserve assets in developing countries and countries in transition increased by $107 billion, just over half of the magnitude of capital inflows (Chart 26). Among the Latin American countries, Argentine and Brazilian reserves declined significantly in the first quarter of 1995, The stabilization of financial markets in the second half of the year enabled reserves to rise rapidly. Brazilian and Mexican non-gold reserves returned to precrisis levels in July 1995, while Argentine reserves recovered most of their losses in December 1995 and edged above the precrisis level in March 1996. Mexican reserves increased by $1! billion in 1995, while Brazilian reserves increased by $13 billion. Capital inflows to Brazil have continued strongly in the first quarter of 1996 as reflected in non-gold reserves, which increased by $4 billion. In the Asian developing countries, the increase in reserves accounted for about 50 percent of the capital inflows in 1995, compared with more than 75 percent in 1994, Only Malaysia and India among the main recipient countries experienced a decline in reserves in 1995—in Malaysia’s case despite an appreciation of the currency.

Chart 26.
Chart 26.
Chart 26.

Total Reserves Minus Gold of Selected Developing Countries and Countries in Transition

(In billions of U.S. dollars)

Source: International Monetary Fund, International Financial Statistics.

Foreign exchange markets experienced considerable turbulence in the aftermath of the Mexican devaluation on December 22, 1994 (Chart 27). After declining 53 percent from its pre-devaluation level of MexN$3.47 to MexN$7.45 on March 9, 1995, the Mexican peso stabilized at around MexN$6 to the dollar in the second and third quarters of 1995. In October and November 1995, the peso came under heavy selling pressure, and it depreciated more than 23 percent, reaching MexNS7.9 per dollar on November 15. Since late 1995, the peso has traded in a narrow range around MexNS7.5 per dollar, reflecting several factors including, a favorable response to the 1996 budget, stepped-up efforts to support troubled commercial banks, and several successful international bond placements.

The Chilean and Argentine currencies were not significantly affected by the Mexican devaluation, in part because of the sound economic fundamentals in Chile and the adherence to the currency board system in Argentina. The trading band for the Brazilian real has been gradually lowered since the end of 1994, and the real experienced only temporary fluctuations in response to the Mexican crisis.

In Asia, the spillover effects of the Mexican crisis proved to be short lived in early 1995 for all but a few countries. The Thai baht, for example, appreciated slightly immediately following the Mexican crisis, but came under heavy selling pressure in January 1995. Intervention by the Bank of Thailand tightened liquidity and drove short-term interest rates to 100 percent, which succeeded in strengthening the currency. Similarly, the Hong Kong and Indonesian currencies came under pressure in mid-January 1995. In the Philippines, the peso depreciated by 5—10 percent following the Mexican devaluation, and again came under pres-sure later in the year. Increases in interest rates have countered much of the more recent pressure on the Philippine peso, which has been fairly stable since the fall of 1995. While a policy of passive intervention to maintain exchange rate stability effectively insulated the Indian rupee from any spillovers in the first half of 1995, a decline in capital inflows and an increase in imports gradually put pressure on the currency later in the year. Speculative pressures added to the fundamental weakness, resulting in a 19 percent depreciation between September 1, 1995 and February 5, 1996. The authorities responded to this pressure by intervening in the foreign exchange market, increasing money market interest rates, and introducing measures to encourage net capital inflows. These measures, and a surge of capital inflows in February 1996 from foreign investors, brought the exchange rate of the rupee vis-à-vis the U.S. dollar back down from its peak of Rs 37.9 to Rs 34-35, where it has since remained.

Chart 27.
Chart 27.
Chart 27.

Exchange Rates of Selected Developing Countries and Countries in Transition

(Local currency-U.S. dollar)

Source: Bloomberg Financial Markets.

The large depreciation of the U.S. dollar against the yen in March 1995 resulted in significant appreciations of the currencies of some Asian developing countries. In particular, the Korean. Malaysian, and Thai currencies appreciated strongly against the dollar during this period. This sudden appreciation resulted in a speculative attack on the Thai baht in March 1995. The appreciation of the baht led to speculation that the Bank of Thailand might change the composition of the basket (consisting of the U.S. dollar, the yen, and the deutsche mark), and there was heavy selling of Thai baht in March 1995. On March 8, the Bank Thailand lightened liquidity conditions in the money market causing the overnight interbank rate to rise 300 basis points, which succeeded in ending the attack.

Transactions in International Capital Markets

Bonds

Bond issuance in the international markets by entities in the developing countries and countries in transition increased slightly in 1995 to nearly $58 billion, but remained well below the $63 billion issued in 1993 (Table 24).7 Asian developing countries issued a lower dollar amount of bonds in 1995 than in 1994. but the volume of issues by Latin American countries increased by 28 percent. Issues from the European developing countries and countries in transition increased by 85 percent, owing to sharp increases in issues by Turkey and Hungary. The market was essentially closed in the first quarter of the year as investors sorted out which countries represented reasonable repayment risks and issuers waited until the volatility in secondary market spreads settled down. Total issues in the first quarter of the year reached only $5 billion in 1995, less than half of the average quarterly rate of issue in 1994. Korea alone accounted for 56 percent of the volume of issues in the quarter. By the end of the second quarter, however, the pace of activity had returned to precrisis levels.

While the decline in industrial countries’ interest rates might have given those countries that were relatively unaffected by the Mexican crisis an incentive to hasten their pace of bond issues in 1995, the Latin American countries placed great importance on returning to the markets quickly. In particular, the Argentine. Brazilian, and Mexican issues in the second and third quarters were highly successful in attracting broad-based support outside the United States and therefore in establishing a benchmark price in the postcrisis environment. The Republic of Argentina issued a $1 billion global bond on April 1, 1995; the Republic of Brazil entered the market in May with a Euroyen bond worth S922 million at a 454 basis-point spread over Japanese government bonds; and the United Mexican States issued a $1 billion bond in July 1995 at a 555 basis-point spread. The bond issue was particularly important for Mexico, which had not issued a sovereign bond in the international markets since 1993. The fact that these three countries were able to return to the markets, even at wide spreads, marked the reopening of the bond market to emerging market countries. Issues in the third quarter of the year rose to a record $21 billion, with essentially all historically important bond issuers from developing countries and countries in transition being represented. The pace of new issues declined slightly in the final quarter, but the total volume remained high at over $17 billion.

The pace of issuance in the international bond markets has accelerated in 1996, with first quarter issues totaling $20 billion.8 Western Hemisphere developing countries have increased their share of total issuance, accounting for 54 percent of total volume, compared with 40 percent in 1995. The Mexican government, for example, raised $4.2 billion in the first five months of the year, including the bond issued to replace Brady bonds described above. One of the most important of these Mexican bonds was a $1 billion issue at the end of January 1996. The offering was initially launched at a spread over U.S. Treasury bonds of 445 basis points, but strong trading soon pushed it to 435 basis points. The success re-established the key U.S. institutional investor base for Latin America, ending the reliance on European and Japanese investors for Latin American bond financing. This bond was quickly followed by Argentina in February with a $1 billion five-year global issue at terms that were more favorable than the Mexican issue.

The turbulence in the secondary markets for developing country debt during the Mexican crisis is indicated in the prices of Brady bonds (Chart 28). The Mexican liquidity crisis led to a steep rise in the stripped yield spreads of Mexican Brady bonds in the secondary markets from a precrisis level of 417 basis points over U.S. Treasury bonds to a peak of 1,779 basis points on March 8. 1995. Over the same period. Brazilian Brady bond spreads increased by 788 basis points and spreads on Argentine Brady bonds rose by 1.604 basis points. After the peak of the crisis in March 1995. spreads generally recovered quickly. Rising concerns in October 1995 about the recovery in Mexico led to a temporary increase in spreads for these countries’ Brady bonds. While Argentine and Brazilian Brady bond spreads have since returned to precrisis levels. Mexican bonds still have not recovered their losses.

The differentiation by international investors between bonds issued by Latin American and other countries is highlighted in Chart 28. The increase in spreads in the first quarter of 1995 was much greater for Latin American Brady bonds than for the non-Latin American bonds.9 Spreads on Philippine Brady bonds, for example, only rose by 545 basis points to March 8, 1995.

The wider spreads in secondary markets were reflected also in wider spreads at issue for international bonds. In particular, average spreads for Mexican U.S. dollar-denominated bonds rose from 204 basis points for 1994 to 549 basis points in 1995. The increase in spreads was less dramatic for Argentina and Brazil, where spreads on their dollar-denominated bonds increased by 162 basis points and 35 basis points, respectively, although in the case of Argentina there was only one issue. Spreads have since come down sharply: average yield spreads declined more than 100 basis points for Argentina and Mexico and 65 basis points for Brazil. Just as the secondary market spreads did not widen as sharply for Asian developing countries, so too did yield spreads at issue not widen significantly. Indeed, spreads for India. Indonesia, and Thailand were lower in 1995 than in 1994. However, owing in large part to political developments in the region in the first quarter of 1996, spreads widened by an average 21 basis points for Asian developing countries.

Table 24.

International Bond Issues by Selected Countries and Regions1

(In millions of U.S. dollars)

article image
Source: IMF staff estimates based on Euromoney Bondware

Including note issues under European medium-term notes (EMTN) programs.

Chart 28.
Chart 28.

Secondary Market Yield Spreads on Selected Brady Bonds 1

(In basis points)

Source: Salomon Brothers.1 Stripped yield spread on Par Brady bond for all countries, except for Bulgaria (Discount Brady bond).

With interest rate spreads widening, borrowers adopted a number of measures to reduce their costs. First, the currency of issuance changed to exploit opportunities to target markets where the base interest rate was relatively low (Table 25). These changes in issuance patterns were at least partly driven by circum-stances. The withdrawal of significant investor support from the United States forced borrowers to look elsewhere. Consequently, the dollar’s share of total bond issues by developing countries and countries in transition declined sharply, from 76 percent in 1994 to 57 percent in 1995. This decline in importance of the dollar sector was not shared by the industrial countries—the dollar’s share of international bond issuance by industrial countries increased from 36 percent to 38 percent in 1995. The share of yen bonds in total issuance by developing countries and countries in transition doubled to 26 percent, while the deutsche mark sector recovered by accounting for 10 percent of total issues after a poor year in 1994. The deutsche mark sector has continued to expand in importance, accounting for 17 percent of issues in the first quarter of 1996, while the yen’s share has declined to 16 percent, slightly higher than in 1994. The expansion of the yen and deutsche mark sectors for developing countries and countries in transition mirrors the developments in industrial country bond issuance, as noted in Annex I.

Table 25.

International Bond Issues by Currency of Denomination

article image
Source: IMF staff estimates based on Euromoney Bondware.

The change in currency of denomination also allowed borrowers to mitigate the general decline in bond maturities in 1995. The average maturity of un-enhanced U.S. dollar-denominated bonds declined from 6.5 years in 1994 to 4.3 years in 1995. In the yen sector, however, the decline was from 6.4 years to 5.5 years, while in the deutsche mark sector, average maturity declined only slightly, from 5.1 years to 4.8 years.

Bond maturities have also risen in 1996, and more issuers have been able to issue fixed-rate bonds. In January, the Malaysian electric utility Tenaga issued a $150 million 100-year bond, which was followed by China’s $100 million, 100-year Yankee bond yielding 299 basis points over 30-year U.S. Treasury bonds at launch. In May 1996, Mexico launched a $1.75 billion 30-year global bond following an exchange offer for Brady bonds. The issue marked the longest-maturity bond issued by a Latin American country since the 1982 debt crisis, and established a new benchmark for Mexican bonds. At the same time, Argentina broke another record by issuing a 15-year, DM 500 million bond, the longest deutsche mark issue by a Latin American country. Overall, maturities for bonds issued by developing countries and countries in transition have been increasing since the last quarter of 1995, averaging 6.5 years for unenhanced U.S. dollar bonds in the first quarter of 1996—the same average maturity as in 1994. Average maturities in the yen and deutsche mark sectors, at over 7 years, are longer than the 1994 averages. However, overall spreads have widened in the dollar and deutsche mark sectors.

A second change in the pattern of bond issuance in 1995 was an increase in the importance of sovereign issues in the total. This was particularly true for the Western Hemisphere developing countries, where sovereign issues rose from 17 percent of total issues in 1994 to 50 percent in 1995 as the governments of Argentina, Brazil, and Mexico issued large amounts of debt to refinance maturing obligations and to try to reestablish benchmarks in the international markets. At the same time, the increase in importance of government issues was accompanied by an increase in importance of issues by commercial banks, which issued $2! billion in bonds in the international markets. This accounted for 36 percent of total issuance, up from 32 percent in 1994. The importance of commercial banks in total issues declined in the fourth quarter of 1995. In the first quarter of 1996, the commercial banks’ share reached only 27 percent, while the share of sovereign issues in the total increased to 37 percent.

The desire on the part of some governments to create a more liquid benchmark in international markets was clearly expressed in the innovative SI,75 billion global bond issued by the Mexican government in May 1996 (Box 5). In this transaction, $2,337 million in Brady bonds were swapped for the new global bond, which, because of its broader distribution and commitments on the part of the underwriters, was expected to provide a more liquid benchmark in the secondary market. Moreover, the Mexican authorities believed that the prevailing stripped spreads in excess of 700 basis points did not represent a reasonable valuation of Mexican risk, and anticipated that the new bonds would be issued at a much lower spread. The bond was issued at a spread of 552 basis points over 30-year U.S. Treasuries. By freeing up the collateral behind the swapped Brady bonds, the Mexican government was also able to pay down some of its short-term debt.

Many companies in developing countries raised funds by securitizing foreign currency receivables. Securitization often allows issuers to obtain ratings for their issues that are higher than the sovereign rating ceiling that generally applies to straight bond issues. Hence, they are able to issue bonds at narrower spreads than even their governments could. In 1995, 35 international bonds were secured, compared with only 22 in 1994. Firms in Western Hemisphere countries issued proportionately more securitized bonds than those in Asia. For example, Telefonos de Mexico securitized $280 million in long-distance service receipts from the United States through Bankers Trust.

Mexican Brady Bond Exchange

On April 17, 1996, the Mexican government announced an offer to holders of its U.S. dollar-denominated Brady bonds to exchange their bonds for a new issue of uncollateralized $1-2.5 billion dollar-denominated 30-year global bonds. At the time, there were $15.5 billion in par bonds and $7.5 billion in discount bonds outstanding measured at market prices. The face value of these bonds was $28 billion. The principal of these bonds was collateralized by zero-coupon U.S. Treasury bonds, while the interest payments were backed by Value Recovery Rights (VRRs), which represented claims on oil revenues. The VRRs were slightly in the-money at the time the exchange was announced, with the first payments due in September 1996. The Brady bonds were to be exchanged for a clearing spread—to be determined by a Dutch auction closing April 30, 1996, but subject to a minimum 125 basis points—over a fixed margin (300 basis points) over the market spread over U.S. Treasuries at close of business on April 26, 1996.

The results were announced on May 1, 1996. Of the approximately $3 billion in Brady bonds that had been offered for exchange, the authorities accepted $1,894 million in par bonds and $443 million in discount bonds measured in face value. The yield to maturity on the global bonds was set at 12.40 percent (an 11.5 percent coupon)—the sum of the 6.88 percent yield on the U.S. Treasury bond at close of business on April 26, plus the fixed 300 basis point spread plus the 2.52 percent clearing spread determined by the auction, for a yield spread at issue of 552 basis points over 30-year U.S. Treasuries. On the day before the swap offer was announced, Mexican par bonds had closed at a stripped yield spread of 745 basis points, and discount bonds had closed at a 780 basis point spread.

As a result of this swap, the Mexican government freed up approximately $650 million in collateral that it had been holding on the Brady bonds, allowing it to pay down some of its short-term debt. The average maturity of Mexican debt was also lengthened because the new bond has a 30-year maturity, while the bonds that were exchanged had a remaining maturity of 23 years. The operation also lowered Mexican interest payments by a present value of $170 million.

Despite a number of well-publicized securitized deals from Latin America in particular, credit enhancements in general continued to decline in importance in 1995 as they had in 1994. Only 25 percent by value of the international bonds issued by developing countries and countries in transition were enhanced, compared with 39 percent in 1993. The main source of the decline was in the Asian region, where the international convertible bond market shrank markedly. Only 58 of 240 Asian international bonds were convertible in 1995, compared with 94 out of 260 in 1994. By value, only 35 percent of Asian bonds carried credit enhancements in 1995, versus 53 percent in 1994. This trend remained the same through the first quarter of 1996. However, in 1996 credit enhancements by Western Hemisphere developing countries have declined sharply, to only 9 percent of total issues by value, compared with 20 percent in 1995. In the first five months of 1996, credit enhancements have declined even further, applying to only 19 percent of total issuance.

The International Yen Hand Markets

A key development in 1995 was the change in currency composition of bonds issued by developing countries and countries in transition. Yen-denominated issues increased sharply from the equivalent of $7.4 billion in 1994 to $15.3 billion in 1995. increasing the yen’s share in total issues from 13 percent to 26 percent. Bonds denominated in deutsche mark rose from the equivalent of $1.6 billion (3 percent of market share) in 1994 to $5.7 billion (10 percent) in 1995.

An important reason for the increase in yen-denominated bonds was the low interest rate prevailing in Japan in 1995. The official discount rate was lowered to 1.0 percent in April 1995 and to 0.5 percent in September 1995. Also. Japanese investors were prepared to purchase bonds issued by developing countries and countries in transition at much lower spreads than were investors in the U.S. market. Hence, interest rates on yen bond issues were substantially lower in these markets than in other segments of the bond market, particularly the U.S. market. The expansion of the yen market was further aided by the deregulation of the international yen-denominated bond market, particularly the elimination on January 1, 1994 of the 90-day lockup period before which sovereign yen-denominated Eurobonds could be sold to Japanese investors after initial placement, facilitating access by sovereign issuers to the domestic investor base. Access has been made even simpler with the elimination, after January 1, 1996, of the minimum credit rating requirement for Samurai bonds and, after August 2, 1995, of the 90-day lockup period on all other nonresident yen-denominated Eurobonds.

While yield spreads at issue rose in the yen sector as they did in the U.S. dollar sector, the increase was much smaller.10 Average sovereign yield spreads increased by only 120 basis points, compared with an increase in average spreads of 181 basis points in the U.S. dollar market.11 The average yield spread in the yen market, for example, was only 280 basis points above Japanese government bonds—resulting in nominal interest rates below the U.S. Treasury bond rates. For example, Mexico issued a $1 billion U.S. dollar-denominated bond in July 1995 at a spread of 556 basis points over the Treasury bond yield, and two weeks later issued a ¥100 billion three-year bond at a swapped yield spread of 336 basis points over U.S. Treasuries.

The possibility of low-cost financing in the Japanese markets was attractive to a large number of developing countries and countries in transition. One of the most active issuers in the yen market has been the National Bank of Hungary. Between 1987 and 1995, the National Bank of Hungary tapped these markets 27 times, with net outstanding volume at the end of 1995 of the equivalent of $7.4 billion—nearly half its outstanding net foreign debt of $17 billion.12 The latest issue in December 1995 had a maturity of 15 years and a 200 basis-point spread over Japanese government bonds—implying a coupon rate of only 5.2 percent.

Most international yen bonds are either Euroyen issues or Samurai bond issues. Samurai bonds are yen-denominated bonds issued by non-Japanese residents and sold to investors in Japan under Japanese regulations. This differentiates Samurai bonds from Euroyen bonds, which are issued in the international offshore market (usually in London). There are also yen-denominated bonds issued by nonresidents through private placements in Japan—Shibosai bonds—but the volumes are not large. Table 26 shows the total issues of Samurai and Shibosai bonds since 1985. Issuance activity increased sharply in 1992. The share of issuers from developing countries and countries in transition also has increased in the past few years, from 29 percent in 1990 to 80 percent in 1994.

From the issuers’ point of view, whether to issue in the Euroyen market or in the Samurai market depends on a number of considerations. First, the investor bases are different: the former is exclusively an institutional investor base, while the Samurai market has a large retail investor base. Second, the underwriting fees and other issuance costs are nearly three times higher in the Samurai market compared with the

Euroyen market.13 Third, until 1996. only investment-grade issuers could sell Samurai bonds, a restriction that did not exist for the Euroyen market. Fourth, prior to 1996, Euroyen issues by nonresidents were subject to a holding period restriction before they could be sold in Japan. Offsetting these apparent advantages of issuing yen-denominated bonds is the currency risk that the issuers undertake if they do not hedge their yen exposure. Many developing countries that have issued yen-denominated bonds are believed not to have swapped out of yen. possibly exposing themselves to exchange rate risk, particularly since markets forecast a significant appreciation of the yen over the next two years.

Table 26.

Samurai Bonds Issuance, 1985-95

(Amount in billions of Japanese yen and share in percent)

article image
Sources: Japanese Ministry of Finance.

The difference between the total amount and the sum of the amounts for developing and industrial countries is issuance by international financial institutions.

The narrow interest rate spreads over Japanese government bonds and the high volume of yen-denominated issues by developing countries have raised concerns that the investors have underpriced credit risk and that issuers have underestimated the potential exchange rate risk under the attraction of the very low nominal interest rates. The former concern centers on the importance of retail investors in the Japanese bond markets, particularly since the market has now been opened up to subinvestment grade issuers. The share of Samurai bonds held by individual investors in Japan increased between 1991 and 1994 from 16 percent to 32 percent. However, retail investors increased their holdings of Samurai bonds significantly in 1995 to 48 percent. The same concern about the relative lack of sophistication of the investor base has been raised with regard to other bond markets, including the German market, in which retail investors are important purchasers of bonds issued by developing countries.

Retail investors in Japan may not have fully appreciated the credit risk of the issues in which they have invested in part because of the ratings assigned to these issues by Japanese rating agencies (Table 27). Sovereign or public sector issuers in many developing countries and countries in transition consistently have received higher ratings from the Japanese agencies than they have from the U.S. agencies, even in cases such as Uruguay where it would be difficult to argue that the Japanese investors or rating agencies would have a better understanding of the issuer. Of particular interest is the case of the National Bank of Hungary, a long-time issuer in the Samurai and Euroyen markets despite having a subinvestment grade rating from both Moody’s and S&P. To illustrate how the more favorable ratings given by Japanese agencies have affected average credit quality in the yen bond markets—as perceived by investors that accept the U.S. agencies’ ratings—Table 28 presents issues in each year since 1993 categorized according to Moody’s ratings of the issues. In 1993, 38 percent of international yen-denominated bond issues had a subinvestment grade rating from Moody’s. This proportion increased to 64 percent in 1995. Countries with relatively low ratings, or with little or no history in international markets, have accessed the yen market at narrow spreads. The June 14, 1996 default by a private Bulgarian bank on a ¥5 billion private placement bond it sold to Japanese banks in 1989 has renewed concerns about the credit risk facing Japanese investors.

Secondary Market Trading

The growing volume of bond issues by developing country issuers is supported by a maturing secondary market for developing country debt instruments and derivatives, which provides information on equilibrium yield spreads and on investor sentiment in general. Despite the near collapse of the primary market in the first quarter of 1995 and its slow recovery in the second quarter, transactions in the secondary markets for developing country debt and related derivatives declined only slightly in 1995 to $2,739 billion (Table 29). As in past years, trading in Latin American debt dominated the market in 1995, accounting for 83 percent of total trading.14 Brazilian debt instruments were the most heavily traded securities (38 percent of the total), followed by Argentine and Mexican debt instruments. With only one sizable Brady bond issue in 1995 (by Ecuador), and a number of countries, including Mexico, buying back Brady bonds, turnover in this instrument declined to 58 percent of total trading, compared with 61 percent in 1994.15 Similarly, a continuing decline in the average stock of loans traded in the market contributed to a 28 percent fall in turnover in that segment of the market. These declines were offset by a 41 percent increase in trading in non-Brady sovereign and corporate bonds, indicating the continuing high level of issuance activity in the primary markets in 1995. Trading in local market instruments continued to increase in 1995, accounting for 21 percent of total turnover. The most commonly traded local instruments were those of Argentina, Brazil, Mexico, Poland, and South Africa. The high price volatility in the first part of the year contributed to a 25 percent increase in trading in derivative instruments, although this segment of the market remains relatively small.

Table 27.

Bond Rating Comparisons of Selected Japanese Yen Issues, 1994-951

article image
Sources: Adapted and updated from Andrews and Ishu (1995); Euromoney Bondware: Japan Bond Research Institute (JBRI); Japan Credit Rating Institute (JCR); Moody’s; and Nippon Investor Services (NIS)

Each rating is at time of issuance. In the case of multiple issuance, the most recent rating is shown.

Table 28.

Average Credit Quality of Japanese Yen-Denominated Bonds

(Amount in millions of U.S. dollars and share in percent)

article image
Source: IMF staff estimates based on huromoney Bond ware.
Table 29.

Secondary Market Transactions in Debt Instruments of Developing Countries and Countries in Transition

(In billions of U.S. dollars)

article image
Sources: Emerging Markets Traders Association and IMF staff estimates,

Data for 1993 do not include trading in short-term local market instruments.

International Equity Markets

International placements of equity by developing countries and countries in transition declined by 38 percent in 1995 to $11.2 billion, slightly lower than the volume placed in 1993 (Table 30). Total issues by Latin American countries fell by 80 percent to only $962 million, and both Argentina and Mexico had no issues at all during the year. Issues by Asian developing countries declined by 27 percent. In Europe, a 53 percent increase in equity issues by the countries in transition was more than offset by an 86 percent de-cline in Turkish issues. As with the bond market, first quarter activity in 1995 was almost nonexistent—only six countries sold any equity—with only $622 million sold, a fraction of the average quarterly volume of $4,5 billion in 1994. The volume of placements recovered in the second quarter and quickened through the end of 1995, but remained below 1994 levels.

Table 30.

International Equity Issues by Selected Countries and Regions

(In millions of U.S. dollars unless otherwise indicated)

article image
Source: IMF staff estimates based on Euromoney Bondware.
Table 31.

International Emerging Market Equity Funds

(As of December 31)

article image
Source: Micropal Inc.

Table 31 presents indirect evidence on equity investment flows to developing countries by mutual funds. Equity funds surveyed in the table include global equity funds, regional funds, and funds designated to specific countries. The total net asset value of these funds in 1995 was $108.8 billion, slightly below the 1994 figure, although the total number of funds increased from 954 to 1.254.16 The decline in the total net asset value mainly resulted from the decline of funds dedicated to the Western Hemisphere. In particular, the net asset value of Latin American regional funds fell from $10.9 billion in 1994 to $8,5 billion in 1995. and almost all Latin American country funds experienced declines. The largest declines occurred in funds for Mexico (down by $723 million) and Brazil (down by $504 million).17 In contrast, the net asset value of Asian regional funds increased by $2.1 billion to $34.8 billion. The main beneficiaries of this growth were funds investing in China and “China-concept” funds, Korea funds, Thailand funds, and Vietnam funds. Among the funds that lost capital were those for India. Indonesia. Malaysia, Pakistan, the Philippines, and Taiwan Province of China. The total net asset value of East European regional and country funds jumped from $1.48 billion in 1994 to $2 billion in 1995, and the number of related funds increased from 38 to 62 over the same period.

These developments in the international markets reflect the difficulties in domestic equity markets in these countries (see Chart 2). In local currency terms, the Mexican market lost 35 percent of its value between December 19. 1994 and February 27, 1995. The equity markets in almost all developing countries and countries in transition suffered losses in the beginning of 1995, but the Latin American markets suffered the greatest declines. In all of the countries represented in Chan 2 except India, however, these initial losses were soon reversed. Even the Mexican stock price index returned to precrisis levels on July 6, 1995.

Activity during 1995 for global and international equity mutual funds is presented also inChart 29. which shows net sales of fund units during 1995 and into 1996. The chart also shows the monthly return on a composite index compiled by the International Finance Corporation. With equity returns turning negative in January 1995, and remaining very low for the rest of the quarter, net inflows into these mutual funds were very low. Clearly, however, as returns in the local equity markets recovered in the second quarter of 1995, inflows into mutual funds picked up. Indeed, the How of new money into these funds follows reasonably closely the returns on equity. In October 1995, for example, the decline in the stock index was matched by a drop in inflows into mutual funds, while the January 1996 increase in stock prices was matched by a surge of inflows.

The sharp price adjustments scared off many domestic investors, resulting in declines in liquidity in some countries. Average monthly turnover in Mexico, for example, fell by 14 percent from 1994 to 1995, while that in Argentina and Brazil dropped by 48 percent and 26 percent, respectively (Table 32). The steepest declines in liquidity, however, were observed in countries or regions that were relatively unaffected by the Mexican crisis. Turnover in the Chinese, Indian, Korean, Malaysian, Taiwanese, and Thai stock markets declined by between 35 percent and 51 percent in 1995. The Turkish stock market was strongly bullish in 1995, with turnover increasing 74 percent, and the stock price index rising 60 percent in 1995.

The weakness in prices and decline in turnover in many countries and the increased uncertainty during the serious economic downturn in some of the Latin American countries provided strong disincentives for new issues on the domestic and international markets. New equity issues in Argentina declined by 88 percent, in Mexico by 53 percent, and in Brazil by 30 percent (Table 33). Total issues by Asian emerging markets were essentially unchanged compared with 1994, as decreased issuance in India and Indonesia offset increases elsewhere in the region.

Countries with open but small financial markets may be subject to greater volatility related to foreign investors’ purchases and sales of local stocks as they become increasingly influenced by developments abroad. Since a small fraction of the assets of large institutional investors in industrial countries can often represent a large supply of capital in an emerging market, a small shift in these investors’ portfolio allocations—possibly in response to developments elsewhere in the world—could potentially result in a significant price movement in the local market. The influence of foreign investors can be exacerbated if a large number of domestic residents take their cue from the actions of the foreign investor, thereby reinforcing any volatility that is transferred to the market from abroad. It is usually difficult to examine the influence of foreign investors in the local stock market, because buying and selling by foreign investors is usually not reported separately. In Thailand, however, such statistics are available (Table 34). In 1993, net foreign invesment in Thai shares was the equivalent ol $2.7 billion; in 1994, foreign investors sold $396 million; and in 1995. they again were net purchasers, acquiring $1.96 billion.18 During those years, the Stock Exchange of Thailand (SET) index increased sharply in 1993, declined in 1994. and then recorded a moderate gain in 1995. Chart 30 shows the time series of foreign investors’ net purchases of Thai stocks and monthly returns of the SET index. The graph suggests a positive correlation between foreign investment in the market and the local currency returns. This correlation is confirmed empirically: in a regression of foreign net investment on the SET index return, a B 1 billion net inflow from foreign investors increases the return on theSET indexby 0.4 percent in the month.

Chart 29.
Chart 29.

Net Sales of Global and International Mutual Funds and Change in IFC Emerging Market Investable Composite Index

(Net sales in millions of U.S. dollars and change in IFC index in percent)

Sources: International Finance Corporation and Investment Company Institute.
Table 32.

Stock Market Turnover Ratio in Selected Countries and Regions

(In percent, monthly averages)

article image
Source: International Finance Corporation, Emerging Markets database.