Comparing Costs of Note Issuance Facilities and Eurocredits

International Finance Discussion Papers

Number 264

September 9, 1985

COMPARING COSTS OF NOTE ISSUANCE FACILITIES AND EUROCREDITS

by

Rodney H. Mills

NOTE: International Finance Discussion Papers are preliminary materials circulated to stimulate discussion and critical comment. References in publications to International Finance Discussion Papers (other than an acknowledgment by a writer that he has had access to unpublished materials) should be cleared with the author or authors.

Abstract

Since early 1984, note issuance facilities (NIFs) have in considerable degree replaced syndicated Eurocurrency bank credits in international credit markets, especially for borrowers in industrial countries, because borrowers have found it cheaper to bypass bank syndications and obtain funds more directly by issuing Euronotes. Factors behind this include an increasing awareness of borrowing possibilities, more freedom for Japanese investors to buy foreign securities, a drop in the relative popularity of bank obligations with investors, and banks' desire to slow asset growth to improve capital ratios.

Measurement of the cost savings to borrowers on NIFs as compared with Eurocredits is difficult because of limited comparison possibilities. Most NIFs arranged so far have not been drawn on through issuance of notes, and few of the actual issuers have arranged LIBOR-priced Eurocredits recently. In the 12 cases, as of mid-1985, where meaningful comparisons could be made, it appears that the cost savings to the borrowers ranged from about 10 to around 50 basis points. All of the savings were in the interest spread, where the NIFs appear to have been about 15 to 55 basis points cheaper. Fee costs were slightly higher on NIFs than on Eurocredits; although front-end fees are lower, the total fee costs for NIFs are higher

because of the annual facility fee paid to banks that underwrite the NIF.

Comparing Costs of Note Issuance Facilities and Eurocredits

by

Rodney H. Mills*

I. Int:roduction

Since early 1984 note issuance facilities (NIFs) have in considerable degree replaced syndicated Eurocurrency bank credits as a vehicle for accessing international credit markets; the substitution has been almost entirely concentrated on borrowers in developed countries. This switch in type of facility has occurred because many borrowers have found it cheaper to bypass intermediation offered by bank syndications and to obtain funds more directly from lenders through the issuance (under NIFs) of Euronotes. These are negotiable instruments with short (up to one-year) maturities that can be issued, retired and rolled over in accordance with borrowing needs, thereby making the NIF a very flexible instrument.

The substitution of NIFs for Eurocurrency bank credits may to a large extent reflect nothing more than greater awareness on the part of borrowers of their opportunities to raise funds at lowest cost and with greatest flexibility. However, at least three other factors also appear to have helped bring about this development. One is the increasing demand for foreign securities from banks and nonbank investors in Japan, in consequence of liberalization measures by the Japanese authorities and regulations that exempt Japanese banks' holdings of securities from the participation limits and funding requirements imposed on their foreign currency loans. In this sense, the boom in NIFs is part of the broader development in 1984-85 referred to as the “securitization” of international capital markets that has seen issues of

* Senior Economist, Division of International Finance, Board of Governors of the Federal Reserve System, Washington, D. C. The opinions expressed in

this paper are solely those of the author, and do not necessarily reflect the opinions of the Board of Governors or other members of its staff.

-2-

bonds, floating rate notes (FRNs), and Euronotes assume much greater importance relative to syndicated loans.

A second additional factor appears to be a decline in the relative popularity with investors of bank deposits and CDs compared with alternative instruments, of which Euronotes are one, together with a change in the relative levels at which banks and some nonbank borrowers can raise funds. Morgan Guaranty Trust Co. has put it as follows:

"The spate of loan problems and operating losses experienced by

a number of banking organizations has undoubtedly changed percep-

tions of the relative quality of bank-issued liabilities. This

has caused some nonbank suppliers of funds to shift part of their

assets from bank deposits toward obligations issued by highly rated

sovereign and corporate borrowers, and made it possible for such borrowers to raise funds more cheaply by going directly to money and capital markets than by borrowing from banks. Indeed, some of these borrowers are able to raise funds more cheaply than banks. The resultant pressure on bank lending spreads, together with higher capital adequacy requirements, has reduced the profitability of direct bank intermediation for certain kinds of business."1

Finally, the replacement of syndicated loans by NIFs has also been affected by the desire of banks themselves to slow the growth of balance sheets, improve capital/asset ratios, and boost income from fees derived from off-balance~-sheet activities.

The shift from Eurocredits to NIFs for borrowers in developed countries has not meant a corresponding removal of large international commercial banks from the lending scene, but has changed their role from one of direct lender to one that primarily involves underwriting loans and only rarely involves direct lending. To date, nearly all NIFS have been underwritten by commercial banks, i.e., the banks stand ready (for an annual fee) over a period of years, usually from three to ten, to buy the borrower's notes or make the

borrower a loan, to the extent that the borrower cannot place the notes in the

market at an interest cost below the "cap" rate. This is the interest rate at

1/ Morgan Guaranty Trust Co., World Financial Markets, July 1985, p. ll.

- 3-

which the banks will finance the borrower .2/ Underwriting banks, which

tend to be the larger institutions, have actually been lenders under NIFs only in the very few instances where they have been required to (i.e., have extended credit at "cap" rates), or where they have voluntarily held Euronotes acquired at the time of issue. Euronotes acquired voluntarily by underwriting banks at time of issue are typically resold by them at a higher price in the market, which comprises nonbank and smaller bank investors. As noted below, most Euronotes are issued through a group of bidders called a tender panel, of which the underwriting banks are almost invariably members.

The large differentials in borrowing costs between NIFs and Eurocredits raise the question why some borrowers who have arranged both types of facilities continue to obtain Eurocredits when NIFs have been cheaper for them. According to bankers, the market for Euronotes of a particular issuer, or group of issuers from the same country, is still small enough so that a modest increase in the supply of notes may drive the yield on outstanding notes upward very sharply, perhaps to levels above Eurocredit costs. Care by borrowers, guided by their bank advisors, to avoid this possibility tends to perpetuate the NIF/Eurocredit cost differentials. Borrowers naturally want to keep down Euronote issue costs, and, from a longer perspective, avoid the

stigma that would be attached to a sharp rise in the yields on their paper.

II. Summary of Findings

The purpose of this note is to quantify the degree to which note issuance facilities are a cheaper form of borrowing than Eurocurrency bank credits. There are serious obstacles in the way of making this calculation. Although the number of NIFs outstanding may exceed 200, Euronotes have so far

2/ Not all NIFs have been underwritten (in the sense used here). Those not

underwritten have been few, but their share of the total is growing.

-4-

been issued in only a limited number of cases. Moreover, many of the Euronote issuers have not borrowed in the Eurocredit market either at all, or at interest spreads over LIBOR (the base rate used for pricing Euronotes), or recently enough so that the terms of the credit can be taken as indicative of what that borrower would have to pay today. These considerations greatly reduce the possibilities of comparing borrowing costs .2/

Notwithstanding these data problems, the comparative terms on NIFs and Eurocredits do suggest that, as of mid-1985, borrowers effectively using the NIF market could raise funds at interest spreads over LIBOR that ranged between 15 and 55 basis points per annum less than the spreads that those borrowers would have had to pay on syndicated Eurocurrency bank credits. Fee costs of NIFs do exceed those of Eurocredits, but on an interest equivalent basis the excess in fee costs appears to average only about 5 basis points, leaving an all-in net advantage of around 10-50 basis points for NIFs over Eurocredits. The wide range of the cost saving of the NIF appears principally to reflect smaller cross-country differences in risk premiums on Euronotes than

on Eurocredits.

III. Comparison of Spreads over LIBOR

On the basis of information in International Financing Review, it is estimated that 174 NIFs were arranged in 1984 and the first half of 1985.

These, together with NIFs arranged before 1984, probably mean an outstanding total of around 200 as of mid-1985. As of June 28, 1985 Merrill Lynch

International NV in London was quoting secondary market prices of Euronotes of

only 51 issues 4/ The disparity between these two numbers reflects the

3/ In addition, some spreads on Eurocurrency bank credits from

1984 have been employed in this study which have been updated by adjustments that can only be approximative, thus introducing an element of imprecision.

4/ As published in International Financing Review.

-5-

oft-reported fact that, so far, the volume of outstanding Euronotes has been only a small fraction of the volume of NIFs in existence. Recent market estimates place the dollar value of outstanding NIFs at around $40 billion and the value of outstanding Euronotes at about $7 billion.

Spreads over LIBOR, or in some cases under LIBOR, on the bid side of the secondary market, as supplied by Merrill Lynch, are shown in Table 1, column 1, for selected Euronote issues. Trading prices in the secondary market for Euronotes are quoted in terms of the spread, usually relative to LIBOR but sometimes to LIBID, the London interbank bid rate. In the table, where the published spread was quoted relative to LIBID it has been converted to a LIBOR equivalent by assuming the usual 1/8 percent spread between LIBOR and LIBID. This has been done merely to facilitate comparisons of the spreads with each other or with spreads in other markets.

We will use the secondary market bid rates for Euronotes as proxies for rates paid by borrowers. In most cases, Euronotes are issued by tender, through a panel usually comprising the underwriting banks and the arranger of the NIF (often an investment bank) and possibly other institutions interested in bidding. At tenders the borrower pays the rates on the accepted bids. Notes are sometimes issued by selling them to a placing agent, in which case the rate paid by the borrower to the placer is set by the borrower himself, who is careful to choose a realistic rate in order to avoid failure to have the whole issue placed with investors, or having to call on the underwriters to take up unplaced notes and pay the penalty "cap" rate. Unfortunately, very little information is reported about the rates actually paid by borrowers. However, in recent months the volume of secondary market quotations has increased greatly.

Market sources assert that, to date, secondary market bid rates at a

given moment have typically been somewhat above, if not equal to, the rates

~6-

paid by borrowers just previously. Discrepancies sometimes emerge because rates bid at tenders are sometimes depressed by over-zealous bidding by panel members anxious to impress the borrower favorably, with an eye to gaining future business from that borrower. But the discrepancies between tender rates and secondary market rates are said to be generally not more than a few basis points. And in any case, tender rates reportedly are never above secondary market bid rates; this is the key point for our comparisons.

The 12 Euronote rates shown in column 1 of Table 1 have been selected because they appear to be meaningfully comparable with spreads over LIBOR on syndicated Eurocredits raised by the same borrower, or by borrowers of similar credit standing in the same country. The list is short compared with the total of 51 quotations provided by Merrill Lynch on the same date, because the other 39 issuers have not arranged Eurocredits recently, on disclosed terms, and at

LIBOR-based rates. It is immediately apparent that the spreads are much lower

ay

than those on syndicated Eurocredits, where 20-25 basis points over LIBOR has been considered "rock bottom" in 1985 and achievable by only a few favored borrowers. The highest Euronote spreads shown in the table are only 15 basis points over LIBOR for Portugal and 14 points for the Korea Exchange

Bank ,2/ and the spreads for both the Statoil and Britoil notes are below LIBOR. (Both of these are actually quoted relative to LIBID.)

Spreads on Eurocurrency bank credits to be compared with the spreads on Euronotes issued under NIFs are shown in column 2, together with the date when the credit was mandated or signed. (See footnote 1 to table.) Seven of the 12 Euronote issuers have arranged LIBOR-based Eurocredits since the beginning of 1984, or have renegotiated the remaining maturities of outstanding loans. For the other five, it is possible to find another borrower of similar

5/ Korea and Indonesia are the only developing countries to have arranged NIFS,

so far as the author is aware.

-7-

Table 1. Spreads Relative to LIBOR (in basis points; no sign = positive spread)

Eurocredits Adjusted Euronotes, Actual (on date (estimated if bid rate 6/28/85 arranged) 1/ made today) 2/ Difference 3/ Borrower (1) (2) (3) (4)=(3)-(1) Australia State Electricity ) ) Commission of Victoria) 3 - - ) Victoria Transport Bor- ) ) rowing Agency 4/ ) - 62.5 (5/84) 50 ) 47 Elders Capital Corp. 5/ ) 6 - - ) Elders IXL Ltd. ) - 62.5 (2/84) 50 ) 44 France 6/ Credit National 3 37.5 (10/84) 25 22 Elf Aquitaine ) 8 - - ) ) VW 8/) Central government ) - 24 (7/85) 24 +») 16 Ireland 1 Central government 2 47 (5/84) 40 38 Italy I.R.I. ) 9 - - ) ) 8/) Korea — 9/ Korea Exchange Bank 14 70 (2/85) 69 55 Norway 10/ Statoil -12.5 37.5 (4/84) 30 42.5 Portugal

Central government 15 62.5 (2/85) 55 40

Eurocredits _ Adjusted Euronotes, Actual (on date (estimated if bid rate 6/28/85 arranged) 1/ made today) 2/ Difference 3/ Borrower (1) (2) (3) (4)=(3)-(1) Spain National Railways 5 50 (5/84) 35 30 7/ Official Credit Institute 10 37.5 (4/85) 35 25 United Kingdom 1o/ Britoil ) - 7.5 - - ) National Electricity ) 8/ ) Council ) - 25° (8/84) 20) 27.5

17 Some spreads are averages of different spread levels over the life of the

loan. The date is the date when the mandate was given, except that the date of signing is used for the Australian loans and the renegotiations.

2/ Author's estimate of spread on Eurocredit that borrower would have had to pay in mid-1985. oo,

3/ Difference between Euronote spread and adjusted spread on Eurocredit.

4/ Guaranteed by State of Victoria.

5/ Guaranteed by Elders IXL Ltd.

6/ Assumes full utilization. The spread is at several different levels

that rise with the degree of utilization. The annual facility fee is not included in the cost shown.

7/ Renegotiation of remaining maturities of earlier loan. 8/ Same spread as in column 2.

9/ Actual average spread on Korean Development Bank loan mandated in May 1985.

10/ Actual quotations are relative to LIBID, and have been converted here to a LIBOR basis by assuming LIBOR at 1/8 percent above LIBID.

~9-

credit standing in the same country that has arranged or renegotiated a LIBOR~based Eurocredit since the beginning of 1984. In this second group, the pairings of borrowers are indicated by brackets.

Comparisons of the spreads on Euronotes and Eurocredits leave no doubt that the NIF has meant big cost savings to borrowers. But the quantification of the savings is made difficult by the fact that the Eurocredits shown in the table were arranged, in many cases, many months before mid-1985, some as long ago as early 1984. Since the first quarter of 1984 Eurocredit spreads generally have been moving downward. Account should be taken of this decline by trying to adjust the spreads used in our comparisons, bringing them down to the levels (as best we can estimate them) that these borrowers would have had to pay in mid-1985.

Because of the drop in the number of Eurocredits arranged in the past two years, it has become more difficult to measure what has been happening to the average level of spreads over a broad range of countries. There are more and more gaps in the data because of the paucity of loans. However, two developed countries, Italy and Spain, have continued to be quite active borrowers in the Eurocredit market. For several years, the Federal Reserve staff has, for internal use, maintained a series on weighted average spreads over LIBOR on Eurocredits arranged for public-sector borrowers in a number of countries. The movements in the spreads for Italy and Spain in the Federal Reserve series, averaged together, during 1984 and the first half of 1985 could be used as a guide to update the Eurocredit spreads data in Table 1. The

Federal Reserve spreads data for Italy and Spain are as follows:

-10-

Spreads over LIBOR (in basis points)

1984 1985 Q @ QQ 4 QL Qe Italy (weighted average) 51 47 41 40 40 32 Spain (weighted average) 53 36 36 43 50 37 Simple average for both 52 41.5 38.5 41.5 45 34.5

The irregular decline in the average Eurocredit spread for Italy and Spain over these 18 months has been used to obtain the adjusted Eurocredit spreads in column 3 of Table 1, except for France, Italy and Korea, where we have very recent spreads that do not need to be adjusted. The adjusted spreads, shown to the nearest five basis points, make no pretense of being precisely those on which borrowers or their bankers would have agreed if Eurocredits had actually been arranged in mid-1985. Their only claim is that they are probably a more accurate indicator of mid-1985 hypothetical borrowing costs than the unadjusted spreads.

Even compared with the adjusted Eurocredit spreads, the Eurcnote spreads indicate much lower interest costs for NIFs than for Eurocredits. The 12 differences in column 4 average 30 basis points. But it is also striking that these differences exhibit such wide variation, ranging from low levels of 13 and 16 basis points for the Italian and French borrowers to a high of 55 basis points for the Korean borrower. This wide range does not reflect imperfect adjustments to the Eurocredit spreads -- the Eurocredit spreads in column 3 for France, Italy and Korea are in fact actual spreads on very recent

credits. Rather, the wide country-to-country variation in the savings from NIFs

-ll-

reflects the substantially smaller cross-country differences in perceived creditworthiness in the Euronote market than in the Eurocredit market. The largest difference between any two Euronote spreads in Table 1 is 27.5 basis points; by contrast, for Eurocredits the range of the unadjusted spreads is 45 basis points while that of the adjusted spreads is 49 basis points. We would, of course, expect lower risk premiums on the Euronotes, because they carry maturit:les of no more than one year in contrast to the multi-year commitments

that Eurocredits entail.

IV. Comparison of Fee Costs

Fee costs are almost always higher on NIFs than on Eurocredits, because NIFs carry a fee that Eurocredits rarely do. That is the annual facility fee, which the banks charge for their function as underwriters of the niF.6/ But the additional cost to the NIF occasioned by this fee is at least partly offset by lower front-end fees on NIFs relative to those on Eurocredits. On an overall basis, differences in fees do little to cut into the savings on the spread when the NIF is used in place of the Eurocredit.

On 108 NIFs arranged in 1984 and the first half of 1985 where the amount of the annual facility fee was disclosed and was set at a constant level, that fee averaged 12 basis points.// In the great majority (91) of these 6/ Traditionally, Eurocredits have not carried annual fees. However, since about mid-1984 there have been several multiple option facilities involving inter alia a revolving “backstop” term loan. An annual facility fee has been applied irrespective of how the facility is used.

j/ The annual facility fee is sometimes called an annual underwriting or commitment fee. These terms should not be confused with the underwriting fee

that is a component of the front-end fee on Eurocredits, or with the commitment fee on undrawn balances.

- 12 -

instances, the fee was in the range of 1/16 to 3/16 percent, i.e., between aproximately 6 and 19 basis points. The facility fee is almost always at a constant level irrespective of how much of the NIF has been utilized through issue of Euronotes or through exercise of other options (e.g., short-term advances), but on occasion it is lower on unutilized amounts than on amounts drawn.

At the average level of 12 basis points in the sample of 108 NIFs, facility fee would wipe out about forty percent of the estimated average cost advantage of NIFs of 30 basis points of spread obtained from the comparisons Table 1. How much is thus offset may vary considerably from NIF to NIF. For example, as shown in Table 2, the high annual facility fee of 25 basis points for Portugal offsets about two-thirds of Portugal's estimated spread savings from the NIF, whereas for Elders Capital Corp. the fee is less than one-third the estimated spread savings.

But the impact of the annual facility fee on the relative costs of NIFs and Eurocredits is tempered by the apparently lower level of front-end f on NIFs as compared with Eurocredits. As is the case with Eurocredits, compl front-end fee information is disclosed in only a minority of instances. However, the author believes to have found such information for 37 NIFs arran in 1984 and the first half of 1985. On an annualized basis, i.e., dividing t fee by the years to maturity of the NIF, the average front-end fee on these = NIFs was 3 basis points. It is extremely hard to find an analogous figure fo front-end fees on Eurocredits because of the limited number of credits. However, a study of front-end fees in the period 1981-83 found that on 74

credits to developed countries the front-end fee, on an annualized basis,

~13-

‘Table 2. Fee Data for NIFs Arranged January 1984 = June 1985 (in basis points)

Annualized Annual Facility Fee Front=End Fee Sample average 12 (108 NIFs) 3 (37 NIFs) NIFs of Euronote issuers shown in Table 1: State Electricity Commission of Victoria 15 Nea. Elders Capital Corp. 10 3 Credit National 10 2.5 Elf Aquitaine None 1/ Nea. Government of Ireland 3/ 12.5 + 25 2/ 3*9 2/ Italy: I.R.I. 3/ Nea. Nea. Korea Exchange Bank 25 Neae Statoil Neae Nea. Government of Portugal 25 6.25 Spanish National Railways 3/ Nea. Nea. Spain: Official Credit Institute Neae Nea. Britoil | 12.5 Nea.

1/ This issue is “Euro-commercial paper” and is not underwritten by banks. 2/ The data refer to three separate issues. 3/ These NIFs were arranged in 1982.

- 14 -

averaged 8 basis points .8/ Currently, the “prevailing” level of Eurocredit front-end fees for developed countries may be lower, since fees tend to move with spreads; considering the decline in spreads since 1983 and the relationship between fees and spreads found in the above-mentioned study, the figure of 8 basis points should perhaps be revised down to 6 basis points. That would still leave the NIF front-end fees about 3 points below the Eurocredit fees. And that differential is wider on an interest equivalent basis, because the front-end fees are paid at the start. NIFs and Eurocredits average about seven years to maturity, and at today's interest rates the excess of the hypothesized Eurocredit fees over the NIF fees is about 5 basis points on an interest equivalent basis. This difference offsets close to one-half of the burden of

the average annual facility fee paid on a NIF.

V. Conclusion

Many borrowers in industrial countries are finding that they can arrange NIFs and issue Euronotes at significantly lower cost than raising a syndicated Eurocurrency bank credit. This results from lower spreads relative | to LIBOR on the Euronotes than on the Eurocredits, an advantage that is only marginally offset by having to pay somewhat higher fees (annual and front-end combined) on the NIFs. The size of the savings obtainable from utilizing NIFs rather than the more traditional financing vehicle suggests that the volume of NIFs arranged will continue at a higher level and that the amount of Euronotes outstanding will grow rapidly. Although most newly arranged NIFs are still being underwritten by banking syndicates, observers believe that the share of non-underwritten facilities will become much more important, especially for top-quality borrowers.

8/ Rodney H. Mills and Henry S. Terrell, “The Determination of Front-End Fees on Syndicated Eurocurrency Credits," The Banker, December 1984.