This paper describes how Brazil has reentered the international financial market by issuing sovereign bonds. It covers the period since the initial arrangements for the first issuing, in 1994, until the europound issuing, in June 1996.

I would like to express my thanks to Silvio Carlos do Amaral e Silva, Head of the Division of External Credits at the Department of Foreign Capital, in the Central Bank of Brazil, and to his team, who were responsible for providing most of the required information for this final paper. Special thanks to the people who worked in the Task Force, and to those who remain working with this matter. Carlos Henrique Rodrigues da Cunha Gomes and Maria Teresa Lisita Bello were very helpful ever.

After being out of the international market since 1979, Brazil faced a new scenario in 1994. The agreement with its private external creditors was signed in April and the Real Plan was implemented in June, foreseeing a period of economic stabilization. These two events have paved the way for Brazil to reenter the international market.

Brazil’s international reserves, which had accumulated to roughly US$ 38 billion, an adequate level to afford the needs of the Economic Stabilization Plan, facilitated this reentry. The investor could realize that Brazil, after being absent from the market for that long time, no longer had financing as its main objective. The country aimed to set a benchmark, attracting resources from abroad. The doors had been kept closed for a long time and it was already time to reopen them.

However, the international capital markets in the beginning of 1994 were not favorable to new issues, even more regarding emerging markets, particularly those located in Latin America.

The Federal Reserve had imposed a very restrictive monetary policy and the Treasury yields, as well as the international interest rates, were high. After the Chiapas incident and the murder of a presidential candidate, an institutional and political crisis occurred in Mexico. Argentina suffered the effects of this crisis and the investors were not interested in Latin-American papers, unless these bonds were to pay higher yields than those usually paid.

From the Brazilian standpoint, it was crucial that its reentry could provide a benchmark, meaning a successful reference point from which the interest rates could be positively influenced to both sectors, public and private.

2. Initial Arrangements

In April 1994, the Ministry of Finance authorized the organization of a Task Force (TF), composed of the Central Bank, the National Treasury and the Legal Division of the Ministry of Finance (Procuradoria Geral da Fazenda Nacional - PGFN) officials, responsible for analyzing and proposing procedures needed to arrange the reentry of the Republic in the international capital market. This group was initially told to define the parameters through which financial institutions would be selected to participate in the process.

The TF has begun its activities choosing the institutions to be invited for presenting proposals, as they were play the role of the leader in the bond issue. The following information was considered:

. the top ten eurobond issue dealers, based upon specialized publications;

. the top ten global bond issue dealers, based upon the Capital Market Yearbook;

. the top ten Latin-American eurobond issue dealers; and

. the top ten Brazilian bond issue dealers in the euromarket.

Subsequently, the data was gathered and arranged in a spreadsheet, in order to give a better view of the most active institutions in those markets. From then on, these dealers would be invited to make presentations to the TF, with the responsibility of showing their market perceptions, reporting the issue strategy and giving general recommendations to the venture.

As the National Treasury agent, the Central Bank has invited, based on the reports already mentioned above, the 13 best institutions to make presentations in Brazil about the global bond market. In making their presentations, the institutions defined how the bonds are sold simultaneously in the United States, Europe and Asia and estimated how the market could react to a new Brazilian bond issue.

It was required that the presentations propose the best strategy for the issue and evaluate the most appropriate timing to guarantee a successful operation.

These institutions advised that going back to the market would have to be preceded by some actions:

. the government economic team would make presentations in the financial centers (roadshow);

. the bonds should be registered with the U.S. Securities and Exchange Commission; and

. the country rating would be obtained from the largest rating agencies (Moody’s and Standard & Poor’s).

The following list provides the names of the 13 institutions invited to make the presentations:

Goldman Sachs

Deutsche Bank

Morgan Stanley

Merrill Lynch

Lehman Brothers

Nomura Securities

Salomon Brothers

JP Morgan

Credit Suisse - First Boston

Chase Investment Bank

Citibank NA

Bear Stearns and

Swiss Bank Corporation

Of these 13 institutions, 6 were banks (4 American and 2 European) and 7 were investment bankers -- or securities houses (6 American and 1 Japanese).

The presentations were held in Brasilia and in Rio de Janeiro, lasting for three hours each, on average. After this, the TF tried to define how to fix the criteria to choose the lead agent. These were based upon the following considerations:

• Regarding the issue:
• Volume and market - a preference for proposals that are designed specifically for the global market, and that have a minimum value of US$ 1 billion;
• Financial conditions - a preference for the smallest cost, calculated using the methodology of "internal rate of return";
• Tenor and payment conditions - a preference for the longest tenor and bullet payment for the principal;
• Offered mode - a preference for the least sophisticated operations because multi-currency transactions could be complicated to the investors; and
• A preference for the minimum time between the offer and the launching, the contract and the pre-marketing, and the pre-marketing and the closing date.
• Regarding the dealer:
• Size, tradition and reliability;
• Primary market experience;
• Secondary market experience;
• Sovereign issue experience; and
• Competence.

As presented in the next topic, these objectives were prioritized when choosing the agents to lead the issuance.

After the presentations were made and the items on which the criteria should be based were defined, the Central Bank sent the formal invitations to the dealers in order to ask them to present the proposals regarding the intended issue. The Central Bank decided that the issue would be set at an amount within the range of $500 million to $1 billion, with a maturity of 5 years, and as a global bond issue.

A list of questions was sent to the institutions in order to determine which was most qualified to manage the issue.

The following information regarding the agent qualification was required from the dealers:

• The percentage distribution grouped by region (U.S., Europe and Asia), for each issue of global bond with SEC registration, in the 18 months before then, in which the dealer acted as a leader/book-runner or a co-leader. The objective of these data was to evaluate the capacity of spreading in global bond markets;
• The effective amount placed by the dealer and the quantity of the related issues, grouped by country where each issue took place, in the 18 months before then, in which the dealer acted as a leader/book-runner or a co-leader. The objective of these data was to evaluate the expertise in bond issues;
• The Brazilian state-owned companies bond issues, in the 18 months before then, grouped by state-owned company, effective placed amount and quantity of issues, in which the dealer acted as a leader/book-runner or a co-leader. These data were required to provide an evaluation of the Brazilian economic scenario knowledge and the commitment the dealer would have had with Brazil;
• The transacted volume in the secondary market in the 12 months before then. These data had the objective of evaluating the dealer’s size and participation in the post-market procedures;
• The obtained price and the proposed price in the emerging markets bond issues, in the 18 months before then, in which the dealer acted as leader/book-runner. These data had the objective of evaluating the dealer capability of establishing the price according to the market expectations;
• IDB, IRDB and Asian Bank references. These data would be necessary to provide a qualitative view of the dealers that had already worked with those multilateral organisms;
• Tenor for obtaining the SEC registration, grouped by each issue, in the 18 months before then, in which the dealer was responsible for the SEC registration procedures. These data had the objective of evaluating the dealer capability of obtaining the SEC registration, regarding the agent agility and the quality of the undertaken works;
• Tenor for obtaining the ratings with Moody’s and Standard & Poor’s, grouped by each issue, in the 18 months before then, in which the dealer was responsible for obtaining the ratings. These data had the objective of evaluating the dealer capability of obtaining the rating, regarding the agent agility; and
• Rating upgrades or new grades obtained in issues, in the 18 months before then, in which the dealer was responsible for obtaining the ratings. These data had the objective of evaluating the dealer capability of obtaining the upgrade or a new grade rating.

These data were ranked, weighed and consolidated, from which the best four would have been chosen to have the proposed all-in costs analyzed.

On May 18, 1994, each institution presented its proposal, in two envelopes, one with the qualifications as mentioned and the other with the indication of the operational costs. The former was analyzed primarily and the latter was set apart, being opened after the definition of the most qualified dealers, and just for the institutions that were chosen in the first stage. Then, the analysis of the cost and the experience of the dealers were set on a two-stage process, and the costs were not known until the first stage was concluded.

After examining the qualification envelopes, 4 dealers were effectively chosen. Then, the envelopes with the costs were opened, regarding these dealers only. Goldman Sachs was chosen to be responsible for the first phase of the sovereign bond issue, as it has presented the lowest costs. The other pre-qualified institutions (Nomura Securities, Merril Lynch and JP Morgan) could be selected to lead future issues, based upon the best presented proposal then.

With regard to the process the Central Bank of Brazil adopted to choose the dealers, the Euromoney magazine observed that the market was surprised by the Brazilian authorities. The reason was that the method to choose the mandate winner was reasonably transparent. Moreover, this kind of process had been not used by other Latin-American countries before.

After defining the dealer that would be responsible for leading the first tranche of the issuance, the Central Bank, as the National Treasury agent for this matter, arranged the legal procedures to the operation. For that, acting with the National Treasury and the General Attorney, it has elaborated an official expedient to the President. The main objective of this document was to ask for approval concerning the process and the related activities and studies, as well as regarding Goldman Sachs hiring, since it was the winner in the choosing process.

In order to prepare the issuance, there were several technical meetings between the Central Bank and Goldman Sachs. The date on which the launching could occur was set to July 1994, before August holidays, or between September and October. Then, counting on the remarkable work made by the Economic Department at the Central Bank, the issue prospectus, or offering circular, was elaborated. This document shows to the potential investors all the information regarding the issuer country, let alone the juridical clauses and other related aspects.

Standard & Poor’s was also hired to provide a rating to Brazil. This phase required a lot of arrangements in order to set a schedule of interviews with Brazilian personalities, in the political, entrepreneurial, governmental and academic areas. At the end of these procedures, Brazil was classified as single B, initially with "stable perspective" outlook, accepted by the financial community.

This rating took into account, on one hand, the Real Plan success regarding the economic stabilization and, on the other hand, as negative factors, the big struggle that the government would have to face in order to approve the proposed constitutional reforms, let alone the economic and social inequalities in Brazil.

July 1994 was considered by the TF a very good schedule for the launching. From August on, there would be a significant reduction in the market activities because of the North Hemisphere holidays. From September on, the elections in Brazil would have been in the final phase, capturing all the investors’ attention. In order to make the launching feasible by July 1994, a lot of tasks would have to be made, and the most important are:

• Permanent follow-up of the market conditions evolution, in order to allow the best decision regarding the effective issue;
• Modification of the budgetary directives, including specific item regarding the destination of the resources to reduce the internal debt; and
• Sending of a project of law to the Congress regarding budgetary credits in order to pay for the accessory costs of the issue.

These tasks have demanded several meetings involving the Central Bank and technical bodies in the Ministries of Finance and Planning. Moreover, the implementation of the Real Plan on the 1^st of July embodied deep changes in the economic infrastructure. Considering these facts, the necessary law alterations were sent to the Congress about the end of July, and the intended issue had to be postponed to the second semester of that year.

C. The Crisis

Besides all the effort made in order to issue the bonds, the international financial scenario had been not favorable along the second half of 1994. The Brazilian sovereign bond should have to wait a little more. The external interest rates remained in high plateaus and, for this reason, accessing the market to set a benchmark would neither have been a recommendable task, nor justifiable.

Under the Fed restrictive policies, the US prime rate reached 8.50 % annually in December 1994, after having been at 6.00 % in February, in the same year. The LIBOR reached 6.87 % annually in December, after 3.74 % in February. At the same time, the internal situation improved. After the election of President Fernando Henrique Cardoso, considered the Real Plan father, the inflation seemed to be under control, falling from an average of 43 %, in the first semester of 1994, to 2 % monthly, the average of the last five months of the same year. The international reserves were growing more and more, reaching $ 43.5 billion in September 1994.

But the issuance had to be postponed once more. This was defined from the end of November on, when the Senate Resolution to authorize the Republic to issue the program first tranche had not been approved yet. Moreover, Goldman Sachs confirmed this procedure, saying that the first week in December would have been the last in which the issue could be successful. After then, the most important investors might get out of new issuances until the beginning of 1995.

Since then, the perspective changed. The issue was rescheduled to happen in January 1995, after the summer holidays in Brazil. However, the Senate Resolution 87 was promulgated in December 1994, authorizing the Union to set an external credit operation program in the equivalent amount not higher than $ 2 billion.

At this moment, as a coincidence, the Mexican authorities decided to devaluate the local currency in more than 14 % to avoid large capital outflows. Among other measures, this decision has strongly contributed to start up the Mexican crisis.

That sort of action was taken considering the Mexican balance of payments: the deficit in the current account reached 6 % of the GDP. Hence, the Mexican government was not capable of paying its short term external debt. After the devaluation, President Zedillo’s government adopted the floating exchange rate for the Peso, without Central Bank intervention. In only seven days, the Mexican currency lost about 54 % of its purchase parity value compared to the US dollar.

After that, and during January 1995, the exchange crisis that had been started in Mexico spread to other markets. As a result, high instability wages took place in the stock exchanges and in the capital markets in the emerging countries, especially on Latin-American countries, as Argentina and Brazil. The generated haywire affected sharply new issuances negotiations. The Mexican crisis could have been even worse for Brazil if the Brazilian international reserves were not as high as they were.

The Emergency Plan adopted by the Mexican government aimed to reestablish trust from the financial market and to catch up with the economic growth. The American government and the International Monetary Fund - IMF were very important for that, providing a huge amount to help Mexico recovers. The agreement with the United States resulted in credits of $ 20 billion, and other $ 17.8 billion came from the IMF into Mexico.

Even with the financial aid to Mexico, the international environment was not too good for the emerging markets until the end of February 1995. The Brazilian sovereign bonds transacted in the international market had their prices compressed, reaching the lowest level on February 15: the IDU, negotiated at 86 % of its par value in December, reached 77.5 % on the middle of February.

Only after March 1995 the environment began to change favorably. However, it is worth commenting that this crisis had some positive results to Brazil’s image. The market perception changed and they realized that Brazil is not Mexico or Argentina. Moreover, Brazil management was conducted in the right way and the constitutional reforms would have come at the right moment.

From April 1995 on, the difficulties related to the recent crisis seemed to be solved in Mexico and Argentina. The investors began their return to these markets and the international capital flows to Brazil were increasing. After the turmoil, the main observed effect was a $ 10 billion reduction in the Brazilian international reserves in the period from November 94 through April 95.

As the newspapers affirmed and as aforementioned, the Republic had been mulling the possibility of a major bond issue in the dollar markets since late 1994, when it mandated Goldman Sachs to lead a $ 500 million to $ 1 billion five-year global bond. However, the plan was scuttled after Mexico’s disastrous peso devaluation in December 1994. Since then the markets have been either too volatile or too expensive to warrant a dollar-denominated issue. That is not for other reason that the global bond issue intended by the Brazilian Republic had to be postponed once more. The setting of a benchmark could not take the risk of being unsuccessful. So, the authorities had to wait for the best moment to do it.

The Central Bank follow-up identified an opportunity with enough probability of being successful. It regarded Japanese investors. A yen-denominated issuance directed to that market should be a good deal. The Japanese domestic interest rates, in opposition to the US rates (the prime rate reached 9 % per annum in May 1995), had not been as low as they were by then. In May 1995, they were 1.40 % p.a.

After the follow-up had indicated favorable conditions for a yen-denominated issue, the four pre-qualified institutions already mentioned were contacted (Goldman Sachs, Salomon Brothers, Nomura Securities and Merril Lynch). They were required to present proposals to an issue of about ¥ 20 billion, equivalent to roughly $ 240 million, and a maturity of two years.

The proposals were then presented and the costs were evaluated. Nomura Securities presented the best conditions for the issue:

Amount : ¥ 20 billion

Currency : Japanese yen

Market : Euromarket

Maturity : 2 years

Coupon : 6 % p.a.

Issue Price : 100 % (par)

Fees : 0.875 % (issuing expenses not included)

The Euroyen issue occurred on June 19, 1995, and it was very successful. At the same presented costs, the operation reached about ¥ 80 billion (equivalent to $ 856 million). It was the first emerging country bond issue in Japan after the Mexican crisis, in December 94. Moreover, this was the largest issue ever placed by any emerging country in the Japanese market. The spread over the U.S. Treasury was 481 basis points, at the launching day, and the calculated costs were 6.505 % p.a., including general expenses. The yield to maturity was 6 % p.a.

Nomura Securities was responsible for placing 70 % of the total issue, basically in Japan (¥ 53 billion) and in Europe (¥ 3 billion). From the total placed in Japan, about 41 % were absorbed by banks and 31 % by private corporations.

The specialized press spread the success of the issue. Several appraisals came from the international media and the market considered the issue with enthusiasm and optimism. There is to say that a consensus was set about the most important from this issue. The direct impact in the government finance was not too relevant as the fact that part of the international financial marked affirmed that Brazil proved to be recovered from the tequila effect, stemmed in the Mexican crisis. The allocation for this issue is shown below:

Issue of the Federative Republic of Brazil

Amount: ¥ 80 billion Coupon: 6 % p.a. Maturity: 1997 Issue Price: 100 %

Issuing Syndicate

From the biggest share, allocated to the leader of the issue, Nomura Securities, this was the distribution among the institutions that held the bonds:

Issue of the Federative Republic of Brazil

Amount: ¥ 80 billion Coupon: 6 % p.a. Maturity: 1997 Issue Price: 100 %

Nomura Securities Sharing

The way the yens would be handled, if converted to another currency or swapped through a rate/currency swap is a matter of international reserves management, and the Department of International Reserves Operations, in the Central Bank, is in charge of it

Regarding the secondary market, after the issue, the market prices changed, and on November 6, this bond was quoted at 102.41, 190 basis points over the U.S. Treasury. It reached 103.375 in mid-September, and the extreme spreads over the comparable U.S. Treasury in the last 5 months were 129 basis points (September 16) and 228 basis points (October 29).

The success of the issue leaded by Nomura stimulated the international capital market participating institutions to look at Brazil. They contacted more and more the Central Bank in order to propose new issues, specially in the German market. These offers and the perception of the Deutschemark bond market being receptive to a Republic issue paved the road to start the analysis targeting those markets to a new sovereign bond.

To choose the institutions to be invited to present proposals was a similar task to that already adopted to the intended global bond issuance aforementioned. Then, gathering information about the German market and the share of each institution on it was the first step.

With this information at hand, 14 institutions were invited, among the largest in the international market, specially those which were active in the German market, for obvious reasons. They were required to present proposals to a 3-year maturity issue, and considering an amount at not less than DM 500 million.

After the proposals have been presented, and the costs have been analyzed, the Dresdner Bank AG was considered the best option, and the conditions were set as follows:

Amount : DM 750 million

Currency : Deutschemark

Market : Euromarket

Maturity : 3 years

Coupon : 9 % p.a.

Issue Price : 100.8 %

Offer Price : 99.8 %

As it occurred with the yen-denominated issue, the deutschemark issue was a success. It was proven by the fact that the effective amount placed was DM 1 billion, largely higher than the initially intended, about half of this value. The effective price was 100.5 % with a re-offer price at 99.5 %. The resources entered the country on June 21, 1995, and the Treasury was immediately credited.

The spread over the U.S. Treasury was 410 basis points, on the launching day, less than in the previous issue. The costs were calculated at 9.627 % p.a., considering general expenses, interest payment and redemption fees, besides underwriting, management and selling fees. The yield to maturity was 8.803 % p.a.

Dresdner Bank was responsible for placing 68.5 % of the total issue, or DM 685 million. The remainder was set by the other participant institutions, including the co-leaders and other co-participants. From the total placed by the Dresdner, about 81 % were absorbed in Germany and 6 % in Switzerland. The largest share went to portfolio (47 %) and fund (25 %) managers.

The specialized press once more commented the success of the issue. It was one of the three best occurred in June. Accordingly to the International Financing Review - IFR, the three issues were good but the appraisals went to Brazil, having returned to the market after being absent for more than one decade. And it was a "spectacularly well done" DM 1 billion issue.

Regarding the secondary market, on November 6, this bond was quoted at 105.25, 204 basis points over the U.S. Treasury. It reached 105.45 in mid-October, and the extreme spreads over the comparable U.S. Treasury in the last 5 months were 204 basis points (November 6) and 275 basis points (August 27).

The allocation follows below.

Issue of the Federative Republic of Brazil

Amount: DM 1 billion Coupon: 9 % p.a. Maturity: 1998 Issue Price: 100.5 %

Issuing Syndicate

From the biggest share, allocated to the leader of the issue, Dresdner Bank, this was the distribution among the regions and among the institutions which held the bonds:

Issue of the Federative Republic of Brazil

Amount: DM 1 billion Coupon: 9 % p.a. Maturity: 1998

Issue Price: 100.5 %

Dresdner Bank Securities Sharing - Regions

Dresdner Bank Securities Sharing - Investors

After another successful event, and continuing the sovereign bond program, the Central Bank and the Treasury decided to issue another yen-denominated bond, targeting the Japanese market. The main reason for accessing the Samurai Market, as it is known, was the low interest rates yielded by the Japanese securities issued by that government. Another important factor was that the Japanese Ministry of Finance opened the domestic market to issuers with risk below investment grade (BBB), rating set by the international quality agencies.

The selection of the leaders and managers was made in a similar way to that mentioned above. The expertise of the dealers and the costs were, one more time, key variables to the selection.

Nomura Securities was chosen to lead this issue, and the bond was launched on March 4, 1996. The large demand for the paper proved how successful the issue was. It reflected also the credibility of the Brazilian economic plan to the Japanese investors. The conditions were set as follows:

Amount : ¥ 30 billion

Currency : yen

Market : Samurai

Maturity : 5 years

Coupon : 5.5 % p.a.

Issue Price : 100 %

As it occurred with the previous issuances, the Samurai market issuance reached the desired objectives. After that, Brazil became the leader in Latin-American issuances in 1996, followed by Argentina and Mexico. Because of the large demand, some members of the syndicate argued that they would rather have a larger share in the allocation. The successful offering marked the longest maturity achieved by the Republic since it returned to the international capital markets after a 15-year exile.

Brazil was the first issuer to take advantage of the recent flexible policy adopted by the Japanese government. With this issue, Standard & Poor’s changed the outlook of Brazil from "stable perspective" to "positive".

Besides the reduction in the costs, the longer maturity effectively contributed to the improvement of the external debt schedule profile.

The resources entered the country on March 22, 1996, and the Treasury was immediately credited. The spread over the U.S. Treasury was 320 basis points, assuring a downwards trend. The costs were calculated at 2.922 % per semester, or 5.845 % p.a., considering interest payment, disclosure and redemption fees, besides general expenses, underwriting and management fees. The yield to maturity was 5.5 % p.a.

Nomura was responsible for placing 81.33 % of the total issue of ¥ 30 billion. The remainder was set by the co-managers. From the total placed by Nomura (¥ 24.4 billion), about 37 % were taken by local financial institutions, 33 % by retail investors, and 30 % by others (including corporations). The allocation for this issue is shown below:

Issue of the Federative Republic of Brazil

Amount: ¥ 30 billion Coupon: 5.5 % p.a. Maturity: 2001 Issue Price: 100 %

Issuing Syndicate

Regarding the secondary market, on November 6, this bond was quoted at 101.75, 360 basis points over the U.S. Treasury. It reached 102.5 in mid-September, and the extreme spreads over the comparable U.S. Treasury in the last 5 months were 309 basis points (October 16) and 370 basis points (October 24).

The successful stream of issues resulted in the Central Bank and the Treasury receiving new proposals for other issues, with more and more favorable conditions in the Euromarket, with emphasis in Portuguese Euroescudos - the "Navigator" market -, and in Europounds.

The Navigator market presented some interesting features, besides low costs:

• Maturity - 39 issues in 1995 and total amount of $ 2.9 billion;
• Convertibility - the Portuguese escudo had become a convertible currency in December 1992, after the European Monetary System - EMS have eliminated restrictions to the capital flows;
• Liquidity - there is a ten-year yield curve for public securities, with liquid benchmarks along it; and
• Simplicity - the registration of the process is simple, and the prospectus is not needed.

The opportunity of issuing a Navigator bond coincided with the visit of the Portuguese Prime-Minister to Brazil, favoring the receptivity environment that is usual in such occasions.

The selection of the leading institutions was made as usually, considering the tradition of the agent and the proposed costs. The leading institutions in the Navigator Market, the previously selected institutions for the global bond and some other dealers which have proposed issuances on this market were invited to present proposals. The Central Bank received three proposals for the Navigator bond and the Banco Finantia presented the best conditions for the issue, as follows:

Amount : Esc 12 billion

Currency : Escudos

Market : Navigator

Maturity : 3 years

Coupon : 10.18 % p.a. (LISBOR+2.4% p.a.)

Issue Price : 100 %

This was the first sovereign issue of a non-European country in this market, and that was very important to set a solid reference point for future Brazilian issues on that. Besides its small value, the issue was relevant to open the market, and to make future larger issues possible.

This operation was a concrete manifestation on how the economic relationship between Brazil and Portugal started to change significantly.

The spread over the U.S. Treasury, at the launching, was 262 basis points, confirming the downwards trend since the return of Brazil to the international markets. Taking the coupon as 9.6 % p.a., the costs were calculated at 5.025 % p.s., or 10.050 % p.a., all the fees already considered. The yield to maturity was 4.8 % p.s. or 9.6 % p.a.

The bonds were issued on May 15, 1996. Banco Finantia and its joint-leader, Banco CISF, were responsible for placing 20.8 % of the total issue of Esc 12 billion. The remainder was set by the senior co-leaders, co-leaders, and co-managers.

The allocation for this issue is shown below.

Issue of the Federative Republic of Brazil

Amount: Esc 12 billion Coupon: LISBOR+2.4 % p.a. Maturity: 1999

Issue Price: 100 %

Issuing Syndicate

Regarding the secondary market, on November 6, this bond was quoted at 100.45, 245 basis points over the U.S. Treasury. It reached 101 in mid-October, and the extreme spreads over the comparable U.S. Treasury in the last 5 months were 232 basis points (October 23) and 277 basis points (September 6).

The Europound market also presented some interesting features, besides its reduced costs:

• Performance - the English public securities (gilts) market had shown good performance, with downward trend for its rates;
• Demand - the largest part of the investors in this market is composed of fund managers and institutional investors, accustomed to sovereign issues, usually taking 50 % of them;
• Pervasiveness - this market is geographically widespread, facilitating the access to investors from several countries, contributing to a more effective benchmark; and
• Opportunity - in the previous 12 months the investors directed to sovereign issues had not many opportunities in this market, and this is a good reason to foresee the success for the Brazilian issue.

Once more, the selection of the leading institutions was made as usually, considering the tradition of the agent and the proposed costs. The leading institutions in the Europound market, the previously selected institutions for the global bond and some other dealers which have proposed issues on this market were invited to present proposals. The Central Bank received six proposals for this bond and the Hongkong and Shangai Bank Corporation Markets Ltd - HSBC, representing the Midland Bank plc, presented the best conditions for the issue, as follows:

Amount : £ 100 million

Currency : Sterling Pounds

Market : Euromarket

Maturity : 3 years

Coupon : 9.75 % p.a.

Issue Price : 100.394 %

Re-offer Price : 99.719 % p.a.

The bonds were rapidly placed at a low spread of 275 basis points over the Gilts (the English Treasury bond), contributing to enforce the benchmark for the Brazilian sovereign bonds. The distribution makes it clear that the investors were eager for these securities, spread not only to the English market, but also to other European and Asian investing institutions.

The bonds were issued on June 12, 1996. The spread over the U.S. Treasury was 250 basis points, confirming once more the declining trend. This time, the calculated costs were 10.162 % p.a., already included fees and general expenses. The yield to maturity was 9.863 % p.a.

The Midland Bank plc was responsible for placing 80 % of the total issue of £ 100 million. The remainder was set by the senior co-lead managers, and co-lead managers.

Regarding the secondary market, on November 6, this bond was quoted at 100.5, 261 basis points over the U.S. Treasury. It reached 100.75 on October 4, and the extreme spreads over the comparable U.S. Treasury in the last 5 months were 230 basis points (October 24) and 318 basis points (July 2).

The allocation for this issue is shown below.

Issue of the Federative Republic of Brazil

Amount: £ 100 million Coupon: 9.75 % p.a. Maturity: 1999 Issue Price: 99.719 %

Issuing Syndicate

The sharing by region and the allocation of the leader HSBC are shown in the next tables:

Total Sharing - Institutions/Regions

HSBC Sharing - Institutions

These bonds were issued very recently. The profile of the buyers and the low volume of trade are not favorable conditions to a strong secondary market. For that reason, there is too little else to talk about it. The Japanese investors, for instance, buy the bonds intending to hold them until the maturity. There is not such a significant information about the secondary market besides what was already mentioned.

As it is easy to sum up, since the first issue in June 1995, marking the return of the Republic of Brazil as a sovereign bond issuer to the international capital markets, there were five issuances until July 1996.

The observed trend in the spreads over the U.S. Treasury along these issues was declining, as it is shown in the following graph:

Besides comparing issuances in different markets, and bonds with different maturities, this curve shows that the credibility of the country is going upwards, accordingly with the good results of its stabilization plan, and this is the key event to lower the spread.

The spreads that Brazil must pay in order to have investors buying its bonds are declining and it leads to the idea that the route back to the international markets has been followed successfully.

Moreover, volatility measures are shown below. The duration of a bond can be considered as a measure of the average life of this instrument on a present value basis and the modified duration is useful to estimate the price change that accompanies a change in yield. At the day of launching, the yield to maturity, the maturity, the duration and the modified duration for the sovereign bonds issued by Brazil are shown in the following table:

Brazil’s Sovereign Bonds

Yield, Maturity and Volatility Measures

As affirmed by VAN HORNE, for the typical investor, risk is judged not by a subjectively formulated probability distribution of possible returns, but in terms of the credit rating assigned to the bond by investment agencies. The main agencies for this matter are Moody’s Investors Service and Standard & Poor’s.

As stated by the same author, based on their evaluations of a bond issue, the agencies give their opinion in the form of letter grades, which are published for use by investors. In their ratings, the agencies attempt to rank issues according to the probability of default. The highest-grade bonds, whose risk of default is felt to be negligible, are rated triple A.

The ratings used by the two agencies as well as brief descriptions are shown in the table in the next page. The first four grades are considered to represent investment quality issues, whereas other rated bonds are considered speculative.

According to the same author, the ratings by the two agencies are widely respected and are recognized by various government regulatory agencies as measures of default risk. Standard & Poor’s assigned a B rating to the Republic of Brazil as issuer of these bonds, meaning speculative. Moody’s rating was B1, after the Samurai issuance.

Moody’s pointed out that the structural and temporary improvements in the central government’s finances that underpinned the economic program at the outset need to be supplemented by comprehensive legal, political, and fiscal reforms. They emphasized that these reforms were being legislated and implemented in piecemeal fashion, but that some of the most urgently needed legislation still lays ahead on the agenda. And this seems to be the way to investment grade: stability keeping and reforms.

They also affirmed that if successful in gaining approval for most of the proposed reforms, the economic authorities should eventually possess a more balanced and powerful set of policy tools and a more streamlined public sector following privatizations and staff rationalization. In turn, this would allow the economy to grow at a more dynamic pace without fear of rekindling inflation. However, the success in the issuances seems to show that the investors’ perceptions are quite different.

Credit Ratings by Investment Agencies

The perspectives after the europound issuance were towards a global bond. This was the kind of bond first considered when the return of Brazil to the international market began to be analyzed, as already mentioned.

After a successful stream of issuances, a downwards trend in the spreads and with the market seeming more and more receptive to Brazilian sovereign issues, the next step would be the global. The studies were held from July 1996 on and the issuance was effective in the end of October. Some newspaper articles said that the success was relative.

The 5-year global bond US$ 750 million was totally absorbed by the market but the spread was 265 basis points over the U.S. Treasury, higher than the europound issue and also higher than the Petrobras issue, occurred three weeks before. The leaders were J.P.Morgan and SBC Warburg. The bonds were issued at a price of 99.857.

The newspapers commented that the moment in which the bonds were issued was not lucky at all. Bad trade balance and public deficit figures were published in that week. Moreover, there were suspicion of corruption in Argentina and the Mexican economy was deteriorated.

But the issuance of a global bond with at least relative success came as the conclusion of the process initiated two years before, and proved the very good work developed by the Central Bank, the National Treasury and other organisms to set a significant benchmark to Brazil in the international market scenario. It is more remarkable since the most of the involved team had not dealt directly with issuing bonds before and all the procedures and meanders had to be unshackled very fast.

Most of the time, when receiving proposals to issue a bond, the definitions and conventions are different from each other. This is a very difficult obstacle to be surpassed. The Appendix that follows deal with a method to compare spreads over different Treasury bonds, issued by different countries, in different currencies, one of those troublesome matters.

One of the problems when comparing different proposals for issuing bonds is that sometimes there are alternatives in different markets and currencies. And so, how to compare diversified spreads, over different references? One proposal says that a bond in the Italian market would be set at 200 basis points over the Italian Treasury Bond. Another proposal recommends the German market and a spread of 150 basis points over the German Treasury bond. Which is the best ?

To solve this problem there is a methodology that considers the swap curves between the currency of issuance and the reference (dollar, for instance), using them to calculate the present values of the cash flows and to find out the equivalent coupon in the reference currency.

The first step is to calculate the zero-coupons equivalent to the swap curve rates of the two currencies considered, in order to use them as the references to find the present value of the flows in these currencies. In this example, we want to know what the spread over U.S. Treasury is when the spread over LISBOR (the Portuguese reference) is known. The maturity of the bond is 3 years. Then, the two currencies will be the U.S. dollar and the Portuguese escudo. The information about the swap rates may be found in the Reuters, for instance, or other large information agency.

The first zero-coupon rate is D1, equivalent to the 1-year-U.S. dollar offer rate in the swap curve. The only thing to do is to standardize the length of time, since the swap rate refers to a 360-day year and the desired length is a 365-day year. Hence, if we consider the swap rate as D, D1=Dx365/360.

The second rate, E, refers to the 2^nd period, and is the 2-year-U.S. dollar ofer rate in the swap curve. Then, we have to find the rate that transforms, in two periods, a present value as shown below, in a future value equal to 100+100xE, that is, principal + interest (at E rate). The original flow is

100xE 100+100xE

|-----------------|-----------------|

100

Based on this flow, it is necessary to discount the intermediary parcel from the initially received value, in order to get a zero coupon. The rate at which this discount will be set is the calculated rate for 1 year, D1. So, the cash flow will be

100+100xE

|-----------------|-----------------|

100-(100xE)

1+D1

The zero-coupon rate, E1, will be calculated as the implicit rate that takes a present value like the shown above into the future value of 100+100xE, two periods after. This rate must be multiplied by 365/360 to adjust the day-count disparities.

The rate for the 3^rd period, F, is the 3-year-U.S. dollar offer rate in the swap curve. The original flow is

100xF 100xF 100+100xF

|-----------------|-----------------|-----------------|

100

The first parcel is discounted in the same way it was for calculating E1. So, the new cash flow is

100xF 100+100xF

|-----------------------------------|-----------------|

100-(100xF)

1+D1

The parcel relative to the 2^nd period must be discounted in two stages. The first, as the previous reasoning indicates, brings the parcel to the first period, using E1 as the discount rate, since E1 was the calculated rate for two periods. So,

100xF 100+100xF

1+E1

|-----------------|-----------------------------------|

100-(100xF)

1+D1

As before, the 1^st period new parcel must be brought into present value, at D1 discount rate (for one period). So,

100+100xF

|-----------------------------------------------------|

100-(100xF)- 100xF .

1+D1 (1+E1)(1+D1)

The rate F1, that takes this present value into a future value of 100+100xF in three periods must be adjusted by the factor 365/360 as before.

The same reasoning follows for the other currency swap rate. Then, if we consider the 1-year-escudo bid swap rate as G, the 2-year-escudo bid swap rate as H, and the 3-year-escudo bid swap rate as I, it follows that

G1=G,

H1=zero-coupon rate that transforms a present value of 100-100xH/(1+G1) into a future value of 100xH+100 in two periods, and

I1= zero-coupon rate that transforms a present value of 100-100xI/(1+G1)-100xI/((1+G1)(1+H1)) into a future value of 100xI+100 in three periods.

The adjustments for 365 days are not necessary since the swap curve for escudos is based on a 365-day year.

By this point, we have the zero-coupon rates to the swap curves of dollars and escudos. The annualized bond coupon (R1) equals to the informed spread over LISBOR plus the escudo swap curve rate average, considering bid and offer, 1yr, 2yr and 3yr, per annum. This average is used as a proxy of the future LISBOR. The market bets on it. So, we already have the rates to work the escudos and dollars flows out, looking for the spread over U.S. Treasury. From the escudos side:

100xR1 100xR1 100+100xR1

|-----------------|-----------------|-----------------|

Re-offer

If Re-offer is different of 100, we must adjust it to a rate S, as to reflect this distortion. So, we can generalize by considering S as the rate that takes a present value equal to Re-offer into a future value equal to 100+100xR1, in three periods, and with installments of 100xR1. That is to say

100xS 100xS 100+100xS

|-----------------|-----------------|-----------------|

100

Finding the present value of the payment part of this flow, using the calculated zero-coupon rates, we have

PVEsc=100xS + 100xS + 100xS + 100 .

1+G1 (1+H1)² (1+I1)³ (1+I1)³

PVEsc=100xSx( 1 + 1 + 1 )+ 100 .

1+G1 (1+H1)² (1+I1)³ (1+I1)³

If we call the expression 100xSx( 1 + 1 + 1 ) as T, PVEsc=T+ 100 .

1+G1 (1+H1)² (1+I1)³ (1+I1)³

From the dollars side, the present value is PVUS= C + C + C + 100 .

1+D1 (1+E1)² (1+F1)³ (1+F1)³

The present values must be equal, C is the desired dollar coupon, and hence,

T+ 100 .= C + C + C + 100 .

(1+I1)³ 1+D1 (1+E1)² (1+F1)³ (1+F1)³

Making the necessary operations,

T + 100 - 100 .

C= (1+I1)³ (1+I1)³

C + C + C .

1+D1 (1+E1)² (1+F1)³

After calculated C, including the swap cost (around 3 bps) and subtracting the annualized U.S. Treasury results the spread over U.S. Treasury equivalent to the initial spread over Portuguese LISBOR.

The same methodology may be used for comparing other market references. But it is fundamental that the swap curves be known for both currencies. Otherwise, there must be another proxy to the discount rates that are necessary to calculate the present value of the considered cash flows.

Banco Central do Brasil - several documents to the President and several Exposicoes de Motivos, 1994, 1995 and 1996

CAMILLOTTO, Luis Fernando Goncalves - A Atuacao Mercadologica do Banco do Brasil no Mercado de Eurobonus - mimeo - FGV/EAESP, 1993

CAPELA, John J. and HARTMAN, Stephen W. - Dictionary of International Business Terms - 1^st edition - Barron’s, 1996

DIARIO ECONOMICO - Brasil emite em escudos - 4.17.96

DOWNES, John and GOODMAN, Jordan Elliot - Dictionary of Finance and Investment Terms - 4^th edition - Barron’s, 1995

DOWNES, John and GOODMAN, Jordan Elliot - Finance & Investment Handbook - 4^th edition - Barron’s, 1995

ESTADO DE SAO PAULO, O - BC inicia captacao de bonus ‘Samurai’ - 3.5.96

ESTADO DE SAO PAULO, O - Emissao de bonus encontra baixa receptividade - 10.29.96

FABOZZI, Frank J. - Bond Markets, Analysis and Strategies - 3^rd edition - Prentice Hall, 1996

FABOZZI, Frank J. - Fixed Income Mathematics - Analytical & Statistical Techniques - 3^rd edition - Prentice Hall, 1996

FITCH, Thomas - Dictionary of Banking Terms - 2^nd edition - Barron’s, 1993

FOLHA DE SAO PAULO - Japao classifica risco de bonus brasileiro - 3.6.96

FRIEDMAN, John P. - Dictionary of Business Terms - 2^nd edition - Barron’s, 1994

GAZETA MERCANTIL - Brasil lanca samurai e lidera captacao and Brasil no mercado samurai - 3.5.96

JORNAL DO BRASIL - Os bonus sairam mais caros - 10.29.96

JORNAL DO BRASIL - Um sucesso relativo - 10.29.96

KAUFMAN, George G. - The U.S. Financial Systems - Money, Markets and Institutions - 6^th edition - Prentice Hall, 1995

MARQUES, Andre B.C. - Definicoes no Calculo do Spread over U.S.Treasury - mimeo - Banco Central do Brasil, 1996

REUTERS - GMIE - 3.5.96

RICHELSON, Hildy and RICHELSON, Stan - Straight Talk About Bonds & Bond Funds - McGraw-Hill, 1996

SILVA, Silvio Carlos do Amaral e - Programa de Emissao do Bonus da Republica - IV CAEX - Banco Central do Brasil, 1995

TELEBRAS - minutas de carta-convite, 1995

VALOR - Finantia e Cisf lideram emprestimo ao Brasil - 4.18.96

VAN HORNE, James C. - Financial Market Rates & Flows - 4^th edition - Prentice Hall, 1994

VISAO - Sem dores de dentes - 4.18.96