Key Features of Foreign Bond Markets

In this section we provide information on the scale and scope of select foreign bond markets and identify key features that may offer insights into their development. As noted earlier, a foreign bond is a security issued in a domestic market by a non-resident corporation and usually sold to domestic investors. These securities differ from bonds issued in international markets (e.g., Eurobonds) because they are subject to the regulations of the country of issuance.⁷

We exclude the US market owing to its size and level of sophistication and instead focus on smaller markets to identify characteristics that may provide insights for bond market development in developing economies. Our analysis adds to earlier work on the foreign bond markets present in Australia (Kangaroo bonds –see Batten, Hogan, and Szilagyi 2009) and Korea (Arirang bonds –see Batten and Szilagyi 2007) and the role of supranational lenders (Hoschka 2005).

We use data from the Thompson Reuters RFI database, which provides details on the terms and conditions of 3,132 bond issues. Key features of these bonds (such as market of issue, credit rating, and industry sector of the issuer) are cross tabulated with maturity. An Analysis of Variance (ANOVA) F-test is then applied to provide a statistical measure of the presence of differences between the bond categories. The ANOVA F-tests were conducted on only 3,105 bonds as the dataset included 27 perpetual bonds. Coupons on the bonds analyzed were typically linked to a floating rate benchmark, such as the LIBOR.

Table 3 [ PDF 88.9KB | 1 page ] reports the number of bonds issued and the average maturity of each issue in major non-US foreign bond markets. The foreign bond issues were denominated in the local currency. However, in some markets other denominations are permitted. In this case issues in local and foreign currencies are usually named differently. For example, in the Japanese market issues in the local currency (yen) are termed Samurai bonds, while issues in US dollars are termed Shogun bonds.⁸ The data include all foreign bonds issued from 1 January 1928 to 30 June 2009 as recorded in the RFI Database (N = 3,132). There are 14 foreign bond categories or markets of issue. The largest market of issue was the Samurai/Shogun bond market in Japan (1,161 bonds or 37.1% of the sample), followed by: the Eurobonds market⁹ (767 bonds or 24.1% of the sample); the Swiss Alpine bond market (516 bonds or 16.5% of foreign bonds); and the Australian Kangaroo market (339 bonds or 10.8% of foreign bonds).

The remaining 11.5% of bonds covered a range of smaller markets including issues in the legacy currencies of Europe prior to the introduction of the euro on 1 January 1999 and recent issues in markets such as the PRC (termed Panda bonds) and Taipei,China (Formosa bonds). As shown Table 3, supranational or quality sovereign issuers played an important role in these markets, as they tended to be the first issuers. The first two Panda bonds were issued in October 2005 by the International Finance Corporation and the ADB (CNY1.1 billion of 10-year bonds at a 3.4% yield and CNY1 billion of 10-year bonds at a 3.34% yield).

The F-statistic for differences in the average bond maturity of the different markets of issue (F=65.09, p=0.000, N=3,105) is consistent with variation in the average maturity of each of these markets. The two longest average bond maturities were in the UK (Bulldog issues) with a maturity of 13.1 years and the EU with 12.7 years. The shortest maturities were in Switzerland (1.7 years) and Greece (2.9 years).

Table 4 [ PDF 76.1KB | 1 page ] reports the top 20 of the 445 foreign bond issuers in the sample, along with average and standard deviation of the maturities of the issuers' bonds. The most important issuer was the World Bank-International Bank for Reconstruction and Development (IBRD) (472 issues or 15.1% of the sample total), followed by the United Bank of Switzerland (USB) (305 issues or 9.7% of the sample total) and the European Bank for Reconstruction and Development (EBRD) (102 issues or 3.3% of the sample total). While the UBS issues tended to have shorter maturities (average of 1.5 years) with very little variation (standard deviation of 0.8), the World Bank-IBRD and the EBRD tended to have longer maturities (averages of 6.3 and 9.9 years respectively) with considerable variation (standard deviations of 6.2 and 7.2 respectively). This variation suggests that these issuers looked to maximize pricing opportunities that arose along the entire yield curve in a specific country of issue rather than restricting themselves to a specific maturity bucket. Nonetheless, the F-statistic for differences in the average bond maturity of the various single issuers (F=8.72, p=0.000, N=3105) suggests significant variation among the maturity choice of each issuer. This is consistent with issuers capitalizing upon the unique institutional aspects of each market (such as the presence of investors requiring long term issues) or their own unique aspects.

The obvious differentiator between issuers is their degree of credit quality. This affects investor decisions to buy the securities regardless of the maturity preferences of investors. This appears to be the most important factor driving issuer maturity choice since issuers with slightly lower credit ratings (e.g., in Table 4, Clariden Leu Aktiengesellschaft, Bank Vontobel, and Morgan Stanley all are single A-rated) tended to have the lowest variation in the maturities of their issues (standard deviations of 0.3, 0.1, and 1.7 respectively). This effect is better illustrated for two issuers with similar numbers of bond issues but slightly different credit ratings. For example, Toyota and KIS AB both issued similar amounts of bonds (28 and 27 respectively) with similar average maturities (3.1 and 3.0 years respectively). However, AA rated Toyota had a standard deviation of 0.95, whereas higher AAA rated KIS AB had a much higher standard deviation of 1.4. This is consistent with KIS AB having greater choice, or more opportunities, in foreign bond markets due to its better quality rating.

Table 5 [ PDF 57.1KB | 1 page ] records the number and maturities of foreign bonds in the sample from 1928–June 2009, recording the number of issues at 5-year periods from 1970–2005 and then annually from 2005–June 2009. The RFI database classifies 5 bonds issued by Majzen (Ferrocarriles) from 1928–1952 (maturity 2022–2005) as foreign bonds. In 1971 the Republic of Ireland issued a foreign bond in pounds sterling in the UK. Later in 1971, the World Bank, ADB and other sovereigns issued a series of yen denominated bonds in Japan (Samurai bonds). From 1976 foreign bond issues became more frequent with the number of issues rising steadily. Despite the 2007–2008 financial crisis the number of bonds issued in the period from 2006–2010 (1,193 bonds as of June 2009) still exceeded the number issued in the period from 2001–2005 (827 bonds). Interestingly, in the period from 1985, the number of foreign bonds issued increased while the average bond maturity declined. This is the likely consequence of a developing market, which accommodates issuers with a variety of credit ratings not just supranationals and high-quality sovereign issuers. The F-statistic for differences in the average bond maturities (F= 62.72, p=0.000, N=3,105) demonstrates significant variation in the maturities issued over the last eighty years.

Table 6 [ PDF 61.2KB | 1 page ] presents a summary of the key features of foreign bonds in the sample. The most prominent feature of these bonds is the fact that 88.2% were fixed rate with simple pricing features. Only a very small proportion of bonds carried option features (8.2% were callable; 16.4% were convertible, 5.2% were dual currency, and 0.7% were puttable). More than half of the bonds (58.7%) were bearer securities (24.5% were bearer Eurobond issues), although a significant amount (36.8%) were also listed on the various stock exchanges of their home country, or required registration (14.9% of these were issues in Australia, which prohibited bearer securities). In addition, other than a very small amount of bonds that were either secured (8.5%) or asset backed (1.3%), most were unsecured, placing the significant burden of credit risk assessment on investors, who rely upon rating agencies to perform this task.

The simplicity of the coupon structures places the responsibility of interest rate risk and ultimately currency risk (by definition the denomination of a foreign bond is not the reporting currency of the issuer) on the issuer. This burden can only be accommodated if there are deep and liquid foreign exchange markets to facilitate risk transformation and risk management, including: floating rate interest rate products (specifically forward rate agreements (FRAs), interest rate futures contracts, and interest rate swaps); currency swaps or currency options; and spot or forward foreign exchange markets. These last two markets are critical given that currency swaps require cash exchange at the inception and conclusion of the contract. A key feature of foreign bond markets where there is some scale (Australia, Canada, Japan, and Switzerland) is the presence of both over-the-counter derivatives markets and deep foreign exchange spot and forward markets (see BIS 2007). The absence of risk transformation capabilities appears to be a major deterrent to issuance.

Table 7 [ PDF 66.9KB | 1 page ] reports the number and maturities of bonds by type of issuer for all bonds in the sample. The table divides the market into three main groups: corporations (51.1%), supranationals (28.8%), and sovereign issuers (20.1%). Based on these classifications the supranational issues tended to have the longest maturities (7.0 years), followed by the sovereign issuers (6.6 years) and the corporation category (4.9 years). Maturity is clearly linked to rating quality as the average credit rating of the supernationals was AAA, sovereigns was AA, and corporations was single-A. The F-statistic for differences in the average bond maturity for each type of issuer (F=26.32, p=0.000, N=3,105) confirms these differences are statistically significant.

Within the corporation category banks and non-bank financial corporations were the most important comprising of 26.2% and 14.9% of all issues respectively. Issues by banks had the shortest maturities of all of the classifications, while utilities and railways had the longest (9.6 and 9.3 years, respectively). The supranational group showed the greatest variation with 6.1, which is consistent with issuance designed to maximize comparative advantage given market conditions at the time of issuance rather than a predefined or restricted issuance requirement.

Table 8 [ PDF 55.9KB | 1 page ] reports the number of bonds issued by the credit rating provided by Standard and Poor's and the average maturity of issue for these credit rating classes. The largest number of bonds issued was by AAA issuers (44.4%) followed by the single-A class. The F-statistic for differences in the average bond maturity of issuer ratings classes (F=21.40, p=0.000, N = 3105) demonstrates statistically significant variation in maturity by credit rating. Interestingly, when classified in this way, the average maturity of the AAA class was longer (6.5 years) than the AA class (5.6 years), though both were shorter than the single-A rating class of 8.32 years and the BBB rating class of 7.72 years. The majority (208 or 80.0%) of the bonds rated BBB and below were issued in Japan, which is more accustomed to bond issues with longer maturities. The majority of the issues by AAA rated issuers were in the Japanese and Eurobond markets (1,061 issues or 76.2% of AAAs), followed by Australia (105 issues or 7.5% of AAAs).

The Australian market is particularly reliant on high-quality issuers due to its small government sector and the use of foreign issuers as alternative securities for bank repo transactions and official account management. New Zealand has recently adopted similar rules for official transactions (see Appendix 2 [ PDF 27.1KB | 1 page ]). Importantly, these high-grade issues also substitute for, or enhance, benchmark bond curves, which are critical for assisting corporate bond issuance in a domestic market (Batten and Szilagyi 2007).

Download this Paper [ PDF 215.5KB| 26 pages ].

Comment(s)

There are [0] comment(s) for this entry. Post a comment.