THE COST OF BARRIERS TO ENTRY:
EVIDENCE FROM THE MARKET FOR CORPORATE EURO BOND
UNDERWRITING ∗
João A. C. Santos
Research Department
Federal Reserve Bank of New York
33 Liberty St
New York, NY 10045
Tel: (212) 720-5583
Email: [email protected]
Kostas Tsatsaronis **
Monetary and Economic Department
Bank for International Settlements
Centralbahnplatz 2
4002 Basel
Tel: (+41 61) 280-8082
Email: [email protected]
May 2002
JEL Classification: G15, G24, G32.
Keywords: market segmentation, corporate bond underwriting, euro.
∗
We thank Philipp Hartmann, Philip Lowe, Mark Carey and seminar participants at the BIS, the 2001
FMA meetings and the Launching Workshop by the ECB-CFS Research Network for useful comments and
suggestions. Clifford Dammers has provided unique insight into the workings of the international bond
market. We are also grateful to Craig Furfine for very helpful programming tips, and Denis Pêtre for
helping us with the data. The views stated herein are those of the authors and not necessarily those of the
Bank for International Settlements, the Federal Reserve Bank of New York or the Federal Reserve System.
** Corresponding author.
European market for bond underwriting
THE COST OF BARRIERS TO ENTRY:
EVIDENCE FROM THE MARKET FOR CORPORATE EURO BOND
UNDERWRITING
Abstract
The euro has eroded many of the barriers that segmented the European market for
corporate bond underwriting along currency lines and gave rise to a unified market
comparable in size to the one denominated in US dollars. In doing so, the new currency
made it easier for investment banks to benefit from scale economies in the provision of
underwriting services, lowered the entry barriers to this industry, and made it easier for
European borrowers to benefit from scope economies by combining their purchasing of
commercial and investment banking services. This paper shows that the arrival of the
euro led to a reduction in the underwriting fees of corporate bonds issued in the new
currency and that this reduction was largely due to greater contestability of the
investment banking business in the post-EMU European market. Our paper also shows
that the elimination of market segmentation led to a migration of underwriting business
towards the larger international investment banking houses, particularly those form the
United States, rather than an intensification of the business links between euro area
borrowers and bankers from the same country. Moreover, borrowers that chose American
investment banks appear to have made extra savings in the underwriting fees. Altogether,
these results suggest that borrowers place a greater weight on placing capacity than on
business relationships in the choice of an underwriter.
2
European market for bond underwriting
1. Introduction
The most noticeable feature of the European Monetary Union (EMU) was the
replacement of ten national currencies by the euro.1 The advent of the single currency,
however, has also set in motion processes that, while less visible, are equally important
for the structure of the future European financial system. This paper focuses on the
impact of the euro on the underwriting market for corporate bonds denominated in the
new currency.
Almost overnight the introduction of the euro eliminated many of the economic,
regulatory and psychological factors that had led to the segmentation of the European
market for corporate bonds along currency lines. Exchange rate risk and national
idiosyncrasies in yield curve dynamics, which dominated the pricing of fixed-income
securities, became irrelevant. Regulatory requirements restricting currency exposure of
institutional investors coupled with the generally observed “ home currency bias” , which
had contributed to the segmentation of demand for securities, became obsolete. Market
segmentation, which made it difficult for firms to benefit from the scope economies of
using as an underwriter the bank with which they had a relationship, disappeared. Prior to
the euro, issuers wishing to tap a foreign market would find it advantageous to select an
underwriter with marketing and sales expertise in the currency of issue. This usually
meant selecting an underwriter from the country of the currency of issue, as opposed to a
bank from the borrower’ s home country.
These effects coupled with the erosion of rents that derived from local expertise
in research and marketing reduced the economic barriers to entry for investment houses
1
The eleven founding members were Austria, Belgium, Germany, Finland, France, Ireland, Italy,
Luxembourg, the Netherlands, Portugal and Spain. Greece became the 12th member of the currency union
in January 2001.
3
European market for bond underwriting
from outside the area. As a result, in the post-EMU period, a more homogeneous market
and an expanded investor base emerged, allowing underwriters to compete on a panEuropean basis and making the business of bond underwriting in the euro area more
competitive.
The effects of the euro’ s arrival were significant and materialised in a short
period of time. This presents a good opportunity to study the structure of competition in
the euro-denominated segment of the international bond market. Our paper has two
objectives. The first is to investigate the competitive impact from the arrival of the single
currency on the issuance cost for corporate borrowers. The second is to characterize the
new competitive profile that emerged in this market segment and to draw some general
lessons regarding the structure of the investment banking industry.
Our analysis of the bond underwriting fees shows a rapid convergence between
the euro- and dollar-denominated segments following the introduction of the single
currency. This reduction in the issuance cost for euro-denominated bonds, as compared to
issues in the legacy currencies, has also been accompanied by a decline in conventional
measures of concentration among underwriters. These findings suggest that economic
barriers to entry in the pre-EMU market entailed costs, and the elimination of these
barriers offered borrowers in the new currency a wider range of options.
This part of our analysis relates to the recent literature on the competitiveness of
the bond underwriting business. This literature, however, has focused on the erosion in
the 1980s of the Glass-Steagall barriers, which had separated commercial and investment
banking in the United States since 1933, and on the introduction of the Financial System
Reform Act in Japan in 1993, which allowed banks to enter the underwriting business by
stetting up securities subsidiaries. Gande, Puri and Saunders (1999), for example, link the
4
European market for bond underwriting
overall reduction in underwriting fees for domestic US bonds to the competitive impact
from the entry of commercial banks in the business.2 These results have been questioned
by Roten and Mullineaux (2001), which use a longer sample and conclude that the effect
of commercial bank entry was transitory. Yasuda (1999) examines separately issuers’
choice of underwriter and the price of the issue in the US domestic bond market, and
finds that pre-existing relationships are a significant source of market power for the
entrant commercial banks but only for low reputation firms. Hamao and Hoshi (2000) in
turn examine bank-subsidiaries entry in investment banking in Japan and find that,
despite some evidence of conflicts of interest, these institutions were able to gain an
important market share in a short period of time through an aggressive entry strategy
focusing on issues aimed mainly at institutional investors and by cultivating issuers with
weak ties to their parent banks.
In the second part of our analysis, we examine the specific forms under which
intensified competition in the euro-denominated segment manifested itself in the postEMU environment. In doing so, we draw a parallel between the impact of euro’ s advent
and the effect of tariff reductions on cross-border trade. Two competing views on the
effects of liberalization on international trade flows are of particular interest to us. The
first posits that in the presence of scale economies in production, free trade benefits
countries with larger domestic demand base (see Krugman (1980), Helpman and
Krugman (1985)). An alternative view formulated by Head and Ries (2001) emphasizes
the importance of national content of traded goods in the presence of constant returns to
scale.3 In this latter case, trade liberalization brings an increase in the demand for small2
Since 1988 the Federal Reserve has used the so-called “ Section 20 subsidiary” rule to allow subsidiaries
of commercial bank holding companies to underwrite securities, bypassing the strict separation of
commercial and investment banking activities imposed by the Glass-Steagall Act of 1930.
3
The key idea behind this theory is that because of the national content, for example the climacteric
conditions for the production of certain wines, it is not possible to move production to a different country.
5
European market for bond underwriting
market producers’ products as they gain access to the consumers in a larger market
through exports.
Our paper complements the literature that has studied these effects in connection
with international trade of manufactured goods, by looking at the financial services
sector.4 In the context of the bond underwriting market, the presence of increasing returns
to scale suggests that the market environment after the introduction of the single currency
would favor the larger international investment banks. If alternatively scale effects are
not as important as “ domestic content” (in the form of relationship capital between the
intermediary and the issuer) then greater market openness would tend to increase the
share of home underwriters in the corporate bond market. 5
The analysis of the patterns of intermediation in the European segment of the
market before and after EMU provides evidence in favor of the increasing returns to scale
hypothesis. European corporate borrowers have not exercised their greater freedom in the
selection of an underwriter in order to exploit existing relationships with their “ home”
institutions. Business instead gravitated towards the larger bond underwriting houses,
especially those from the United States, and borrowers that chose the latter underwriters
received some extra savings in underwriting fees. Our results, therefore, suggest that
marketing and placement capability dominated the information synergies in the
relationship between the underwriter and the borrower. The evidence of economies of
scale in the production of investment banking services also suggests a solid economic
rationale for consolidation in the industry both at the national and international levels.
4
Beginning with the deadweight loss calculations of Johnson (1960), there has been a great deal of interest
in the measurement of the costs of trading barriers. See, for example, Feenstra (1992, 1995) and the
references therein.
5
A bank that provides commercial banking services to a borrower would find it easier to underwrite the
firm’ s securities as it could use the information it has accumulated over the course of its relationship with
the firm. See Kanatas and Qi (1998) for a formalization of this argument, and Rajan (1996) and Santos
(1998) for other potential benefits for a firm from relying on its bank for underwriting services.
6
European market for bond underwriting
The remainder of the paper is organized as follows. The next section compares
the market for bond underwriting in Europe before and after the arrival of the euro.
Section 3 presents our methodology, and describes the data and the variables used in our
analysis. Section 4 discusses the empirical results and section 5 concludes.
2. Eliminating entry barriers: One currency, one market
The impact of EMU in the European bond markets was multifaceted, and providing a full
analysis of it is beyond the scope of this paper.6 We therefore focus our discussion on the
impact of the euro on the market for corporate bond underwriting.
Besides eliminating all uncertainty relating to exchange rate risk affecting the
choice between securities denominated in the legacy currencies, a key impact of the euro
was to relax a number of regulatory restrictions in the currency profile of assets and
liabilities of institutional investors in the euro area. These restrictions include currencymatching requirements that put a ceiling on maximum permissible mismatch in the
denomination of assets and liabilities to 20%, and portfolio allocation rules that further
restrict these investors’ ability to allocate funds in foreign-issued securities (Table 1).7
Finally, the euro relaxed the influence of less formal restrictions, such as the so-called
“ prudent man rule” , which complement the well-documented tendency of investor
portfolios to be overweight on domestic assets at the expense of better diversification
opportunities offered by a wider geographic spread of their holdings.8
6
For a more completed overview of the effects of the euro on European financial markets see Danthine et
al (2001), ECB (2001), von Thadden (2001), Detken and Hartmann (2000) and Galati and Tsatsaronis
(2001).
7
These restrictions apply to the technical provisions against liabilities as defined in each regulatory
jurisdiction. Life-insurance liabilities are typically denominated in the local currency as the commercial
practice of these companies has been to use locally chartered subsidiaries rather than cross-border sales for
their international operations (Lannoo (1996)).
8
The so-called “ home bias puzzle” has been documented by several authors (Tesar and Werner (1994),
Baxter and Jermann. (1997) and Glassman and Riddick (2001)).
7
European market for bond underwriting
The combined effect of these factors led to a segmentation of the fixed-income
market along currency lines. By eliminating this market segmentation, the euro
broadened the investor base for any issue denominated in the new currency.
Consequently, issuers in euros address simultaneously a more diverse and larger group of
investors.9 Data covering the extent of cross-border securities holdings within the euro
area are limited, but it suggests asset managers from the euro-area responded positively
to these changes. For example, the share of euro-area bonds in the portfolios of Italian
mutual funds has increased from 8% in 1995 to 23% in 2000. Furthermore, available
information suggests that European investors absorbed the bulk of the new bonds
issued.10
Greater and broader investor demand was mirrored by an improvement in market
liquidity, which further enhanced the attractiveness of euro-denominated bonds.
Comprehensive data on corporate bond market liquidity are not available, but both
anecdotal and sparse evidence suggests that the secondary market turnover for private
debt issues has benefited from the arrival of the single currency.11
Issuers reacted very favorably to these structural changes brought about by the
euro. During 1999, for the first time, the number of euro-denominated issues surpassed
those denominated in the US dollar. The historical gap in issuance volume between the
two currency segments almost disappeared too (Table 2). The key contributors to this
issuance boom were non-financial borrowers from the euro area (Table 3). Lower rating
9
Michael Somers, head of the Irish National Treasury Management Agency, noted colorfully that “ we’ re
all essentially selling the same detergent now” referring to the greater substitutability of government bonds
in the post-EMU period (Financial Times, 2 January 2002).
10
These facts are consistent with the conclusions of Detken and Hartmann (2000) that, during its first year,
the euro’ s role as an international financing vehicle was more important than its role as an international
investment currency. For a more detailed discussion of the trends and nature of international portfolio
flows within the single currency area and vis-à-vis the rest of the world see also Galati and Tsatsaronis
(2001).
11
See Galati and Tsatsaronis (2001) and BIS (2001).
8
European market for bond underwriting
issuers, particularly single A and BBB issuers, also had an important contribution to the
boom in this market, which used to be dominated by sovereigns and high-grade
corporates (Table 4). Finally, euro-area borrowers that were new to the capital markets
also had a key contribution to this surge in bond issuance; 30% of the bonds issued after
the euro were by firms that had been absent from the capital markets at least since 1994
(when our data start).
This boom in the Euro-area bond market may have also been driven by the
changes in the underwriting business that came with the new currency. In choosing an
underwriter, the issuer is guided by two main considerations: the relationship with the
banker and the ability of the latter to successfully sell the issue.12 On one hand, the ability
to draw on the experience and knowledge that comes with an ongoing relationship
between the issuer and the banker would tend to reduce the overall transaction costs to
the two parties.13 Intimate knowledge of the borrower can also be interpreted by potential
buyers as a form of certification of the bond’ s quality.
On the other hand, successful placement of the issue requires expert judgment of
market conditions and accurate gauging of investor demand. An underwriter which is
continuously engaged in marketing bonds with a specific segment is better placed to
gauge market demand and more closely attuned to current investors’ preferences
compared to an occasional player. Moreover, an effective marketing network is key to the
success of an underwriter. The establishment of such a marketing capacity is one of the
major cost lines for underwriters and its maintenance requires the repeated (and
successful) presence in the market. Finally, successful placement is important because it
12
See Marshall and Ellis (1994) for a detailed description of the activities performed by an underwriter and
Smith and Walter (2000) for a review of the investment banking business in the euro-zone.
13
See Santos (1998) for a review of the potential economies of scope associated with the combination of
commercial banking with investment banking activities, and Forestieri (1993) for a survey of the empirical
research on economies of scope.
9
European market for bond underwriting
helps minimize the cost from unsold or under-priced inventory, thereby lowering the cost
of this funding source.
A consequence of the market segmentation along legacy currencies that we noted
above was that borrowers interested in capitalising on a relationship they had with a bank
for the purpose of issuing bonds would find the relationship to be of little value if the
bank did not have, for example, marketing operations in the country of the desired
currency of issue. Underwriters with pan-European ambitions, in turn, were obliged to
establish a series of specialized marketing networks and multiple research teams, thus
increasing the underwriting costs. Segmentation also erected barriers to entry for
outsiders, thus limiting competition and becoming a source of rents to underwriters with
the “ home currency” advantage.
The introduction of the single currency made it easier for firms to select as
underwriter the bank with which they have had, for example, a lending relationship and
thus benefit from the scope economies in the simultaneous provision of both services.
The single monetary policy, and the consequent convergence of yield curves across the
euro area, together with the broadening of the investor base eliminated the need for
multiple research teams covering different countries and allowed securities houses to rely
on a single marketing network to place all issues.14 This simplification eliminated the
need to engage in costly nurturing of relationships with diverse investor communities,
and improved placing ability even if the bank cultivated links with only a subset of a
larger market; something that would be difficult to accomplish in the smaller and
14
The euro-bond market, as a consequence of its larger size, also permits a clearer separation of the
institutional and retail segments of the market. The particular requirements of retail placement, such as
smaller denomination of the bonds, the printing and handling of physical certificates, a more extensive
distribution network and higher transactions costs can add substantially to the cost of issuance. In smaller
markets, borrowers were at times obliged to also cater to the retail clientele but the larger and deeper bond
market since the introduction of the euro has reduced the importance of these considerations.
10
European market for bond underwriting
fragmented markets prior to EMU. Finally, the new currency reduced the competitive
advantage of home currency underwriters and facilitated the establishment of panEuropean operations by aggressive competitors from outside the area.
3. Hypotheses, method and data
3.1 Hypotheses and method
Our investigation of the impact of the euro on the cost of bond underwriting proceeds in
two parts. In the first part, we compare the average underwriting fees for eurodenominated bonds with both the fees of bonds denominated in the legacy currencies and
those of bonds denominated in US dollars. We then relate the changes in fees with the
changes in market concentration in each segment. Finally, we use regression analysis, to
evaluate the impact of the euro on the underwriting fees of euro-denominated bonds.
In the second part of our investigation, we attempt to distinguish between
different channels through which the arrival of the euro might have affected underwriting
fees for euro-denominated bonds. We focus on two hypotheses that emphasize different
facets of the underwriter’ s role: the relationship with the borrower and the capability to
place an issue. If the informational advantage derived from an existing relationship is a
dominant factor in the choice of an underwriter, then in the post-EMU market one would
expect that firms’ greater freedom in the selection of underwriters would lead them to
prefer banks from their own country. In this case, informational economies associated
with the joint provision of underwriting services and other banking services would
explain the reduction in underwriting fees. We label this the national differentiation
hypothesis, borrowing a term used by Head and Ries (2001), who argue that trade
liberalization benefits producers of goods with a distinct national character.
11
European market for bond underwriting
Alternatively, if global selling capability is important, then in the post-EMU one
would expect firms to gravitate towards large international underwriters, regardless of
their nationality. In this case, the presence of scale economies in the underwriting
business would explain the reduction in underwriting fees, for example in the form of
declining marginal cost of market research and placement for underwriting houses with
pan-European operations. We refer to this as the increasing returns to scale hypothesis as
it bears similarities with the theory on trade put forward by Krugman (1980), which
asserts that in the presence of scale economies trade liberalization would favor countries
with larger domestic demand.
In an effort to distinguish between these two hypotheses we examine the
dynamics of the share of “ home currency” and “ home country” underwriters around the
introduction of the euro. We pay particular attention to the choice of underwriter by
issuers that sought capital market funding for the first time in the post-EMU period, as
well as the market share of large underwriters. We also enlarge our fee regressions by
including variables that capture the effect of overall scale of operations for the
underwriter, the size of the market and any relationship between the nationality of the
underwriter the currency of issue and the nationality of the borrower.
3.2 Data
The data used in this study have been extracted from the IFR Platinum bond database
complied by Thompson Financial Securities. This database covers mainly bond issues
characterized as international by meeting one of the following criteria: (a) they are selfdescribed as such in the prospectus stating that the issuer targets investors outside its
home country, (b) the bond is denominated in a currency different than the home
12
European market for bond underwriting
currency of the issuer, or (c) the issue is underwritten (or co-managed) by an
international investment bank.
While the international characterization leaves out a large number of bonds issued
by US borrowers, which are primarily domestically focused, it is fairly inclusive in the
case of bonds issued by European borrowers. Apart from government debt issues, very
few bonds issued by European corporate borrowers are purely domestic. The small size
of national markets obliges issuers to broaden their target audience by issuing bonds that
appeal to international investors.
An advantage of focusing in the international bond market for the purpose of this
study is that we can capture the market segment that is most exposed to international
competition and where issuers are more likely to be influenced by cost factors, including
the underwriting fees that are the main focus of this paper.
In the formal analysis we have selected only fixed coupon bonds issued by the
private sector between January 1994 and June 2001. We have excluded bonds issued by
financial institutions (first digit SIC code 6) because of the possibility that they might not
be competitively underwritten if issued by subsidiaries or affiliates of securities houses.
We have also excluded bonds issued by central or local governments and public
enterprises (first digit SIC code 9) for much the same reason. Many European sovereigns
have issued bonds that were underwritten by the national central bank and it is likely that
large issuers such as emerging market countries might have a special relationship with
the institutions that promote their bonds. The same holds true for the large supra-national
institutions such as the World Bank, which is one of the most active players in the
international bond market. The inclusion of these borrowers in the sample could distort
the results.
13
European market for bond underwriting
The dataset used for the analysis comprises 3110 bonds that satisfy the above
criteria and for which we could obtain all the necessary information. Table 5 provides a
general idea of the composition of the dataset, which appears generally balanced with
roughly comparable number of bonds in the two main currency categories and bonds in
the pre-EMU and post-EMU sub-periods
3.2.1 Variables
We next present the most important variables used in our analysis, namely the fees and
those that relate to the identity of the underwriter. The data appendix includes a fuller
description of all variables used in the regressions.
The measure of fees we use is the Gross Spread, which refers to underwriting fees
expressed as a percentage of the total funds raised. This is the same variable used by
other related studies such as Gande, Puri and Saunders (1999) and Mullineaux and Roten
(2001). The definition is inclusive of several types of fees such as management, selling
commissions, preacipium etc.15 We elected to analyse the gross fee because it is the most
comprehensive of the overall issuing expenses borne by the borrower.
A subtle complication with the chosen measure relates to the possibility that it
might overstate the true fees due to the practice of “ re-allowance” .16 This practice refers
to the selling intermediary offering the bond at a discount to the final investor, hence
passing on a portion of the stated underwriting fee. Since 1989, however, the alternative
arrangement of fixed price offer/re-offer has gradually become the accepted standard in
the international bond market, phasing out the re-allowance of fees. The fixed price re15
In some cases the database provides a breakdown of the gross fees to individual components but in many
instances all the fees are combined into one.
16
Another source of potential misrepresentation of the cost issuance by our selected measure stems from
the fact that a bond issue is often bundled with interest rate and/or currency swap structures. The lack of
information on any swap-related revenue will tend to bias our estimates of the level of fees. However, it is
unclear to what extent its absence should interfere with our analysis of the effect of the euro on these fees.
14
European market for bond underwriting
offer system binds the members of the selling syndicate to the established offer price for
the bond for a period of several weeks following the launch of the issue, negating the
need for an inflated fee figure, at least for those issues with a clear institutional market
focus.17 A discrepancy might still exist in the case of issues that are simultaneously
offered to retail and institutional clients. Higher fees might be stated for those issues in
order to cover the expenses of retail distribution. Large portion of these fees could be
“ re-allowed” to institutional buyers.
The IFR Platinum database does not unambiguously identify issues targeted to the
retail investor and thus makes it difficult to measure this effect directly. Moreover, to the
extent that re-allowance of fees might affect our results it is likely to be in the form of an
upward bias in the level of fees that should not be affected in any direct way by the
introduction of the euro.
An investment house’ s involvement in the preparation and sale of a bond issue in
the primary market can take many forms. This paper focuses exclusively on the role of
the bookrunner or lead underwriter as identified in the database. The bookrunner receives
the original mandate to organise the bond issue and subsequently forms the syndicate of
underwriters and acts as its leader. As a reward for shouldering the main responsibility
for the coordination of the placement as well as bearing the risks of a botched sale, the
bookrunner receives the largest portion of the fees.18 For a number of bonds in our
dataset there is more than one bookrunner. We have dealt with this issue in a manner
most appropriate to the specific context, as described in the data appendix.
17
The first issue offered under this arrangement was a US dollar World Bank bond in 1989. The formula
of fixed-price offer was quickly spread to other currency segments. For an analysis of the offer system and
the details of the international bond market see Fisher (1997).
18
The lead manager is also the leader of the price stabilisation effort during the first few weeks of the new
issue flotation. This operation that aims at keeping the market price close to the offer price can result in
trading losses if the bond is received unfavourably by the investor community.
15
European market for bond underwriting
4 Evidence on the reduction in underwriting costs
We first document the increase in the contestability of the euro-denominated market
segment and the accompanying decline in fees. After that, we discuss the evidence on the
shifting patterns in the selection of underwriters, and the corresponding impact on
underwriting fees, in an attempt to distinguish between the two alternative hypotheses of
increasing returns to scale and national differentiation.
4.1 Competition and fees
The right hand panel of Table 5 shows the average gross fees for the bonds in our sample.
Figures refer to percentage points of the amount raised and are calculated as averages
weighted by the size of the bond expressed in US dollars. There is a clear global
downward trend in fees over the 1994-2001 period as evidenced by the figures in the last
column. Value weighted average fees for 2001 stood at 86 basis points below their 1994
levels, the equivalent of a 37% reduction.19
A closer examination of the time pattern of these figures shows an acceleration in
the global decline in fees around the introduction of the single currency. During the
course of 1999 average fees dropped more than 30 basis points compared to their level of
one year earlier. The bottom rows of Table 5 show the difference between average fees
prevailing before and after the introduction of the euro. The differences are statistically
significant at all conventional confidence levels both for the market as a whole and for
either of the two major currency segments.
19
Value-weighted averages provide a per-dollar-raised measure of bond underwriting cost for borrowers.
Hence they are a more relevant measure of the economic costs of accessing capital market funding. The
corresponding per-issue average fee figures are 1.338% and 0.799% respectively for the pre- and postEMU periods in our sample.
16
European market for bond underwriting
The size of the overall decline in average fees, however, is largely attributed to a
sharp drop in the euro-denominated segment, which for the purposes of this table
encompasses all bonds issued in euro (or the ECU) as well as the ones issued in one of
the ten legacy currencies. Average fees per dollar raised for this category of bonds were
halved between 1998 and 1999, and continued to decline in the following years.
Another interesting aspect of these figures is that average fees in the post-EMU
period are quite similar between the euro and US dollar market segments. The difference
of a quarter of a basis point between the two groups of bonds is not statistically
significant. In the pre-EMU period, this difference stood at 55.8 basis points. Since the
issue size of the average bond denominated in the one of the legacy currencies or the
ECU was about $206 million, the borrower incurred an extra cost of the equivalent of
$1.14 million in underwriting fees by not issuing in US dollars.20
Altogether these results suggest a significant shift in the pricing of underwriting
services in the international bond market around 1999, the impact of which has been
particularly strong in the euro-denominated segment. We next provide evidence in
support of the claim that the high level of fees prior to EMU reflected the segmentation
of the underwriting market for the legacy currency bonds. In addition, we show that the
subsequent convergence to the levels prevailing in the US dollar segment can be largely
attributed to the impact of the introduction of the euro.
The advent of the single currency has had a very significant effect on the
competitiveness of the European market for underwriting services. Table 6 shows the
evolution of three measures of concentration, an Herfindhahl index and percentage shares
20
The pre-EMU euro-denominated issues include 29 bonds denominated either in the euro or the ECU.
The average value-weighted fee for those bonds was 1.37%, falling in between the average fees for legacy
currency (1.685%) and US dollar denominated issues (1.095%). There are no legacy currency denominated
bonds in our sample issued after January 1999.
17
European market for bond underwriting
of the top-5 and top-10 underwriters, for various currency segments for the period 19942001. The Herfindhahl index calculated as the sum over all bookrunners of the squared
percentage share in terms of volume of bonds issued in the specific currency-year pair.
There are three features of Table 6 that are worth highlighting. First, bond
underwriting in the smaller currency segments is generally more concentrated than for
bonds issued in one of the three major currencies. This is especially true if one considers
the legacy currencies of the euro or the Swiss franc. In many of these market segments
the share of the top-5 bookrunners is higher than 70% and there are fewer than ten
bookrunners active in the market (i.e. total share of top-10 equals 100%). Using
commonly applied criteria, all markets in the legacy currencies, with the exception of the
Deutsche Mark segment, could be classified as concentrated and many as highly
concentrated.21
The second point worth highlighting is the stark contrast between these segments
and the one denominated in the euro post-EMU. The latter is the least concentrated
segment of the international bond market.
Finally, it should be noted that concentration measures have been trending
upwards in the US dollar market segment throughout our sample period. This is likely to
be a reflection of the consolidation trends in the US financial sector, and is consistent
with greater concentration in the global market for investment banking services and the
success of larger investment banks in attracting global business. In fact, league tables for
bond underwriting supplied by Thompson Financial indicate that the global share of the
top 5 (top 10) underwriters in the international bond market has increased form 32.7%
(50.7%) to 52.5% (77.9%) over the course of the 90s. An implication for our study is that
21
The US competition authorities use a Herfindahl index value in the 1000-1800 range as an indication of
a concentrated market, and index values in excess of 1800 as evidence of highly concentrated markets that
require regulatory review and potentially action.
18
European market for bond underwriting
while there are fewer competitors globally the contestability of the euro denominated
segment improved dramatically after 1999, as more investment houses were able to
compete for business in the European currency segments.
In the analysis so far we have examined the evolution of average fees without
paying attention to the particular characteristics of the different market segments.
Previous research has established that certain characteristics of the issuer and of the bond
issue are important determinants of underwriting fees.22 Hence we then examine whether
these characteristics alone explain the decline in the fees for euro-denominated bonds.
To this end we run a regression of the gross underwriting fees on a number of
variables that have been found to have some explanatory power, such as: size, rating,
maturity of the issue, the industrial sector of the issuer, and whether the bond is part of an
MTN program. We also include an annual trend variable to capture any global effects and
two variables to control for the familiarity of the market with the particular corporate
name. More specifically, we include a dummy variable to identify borrowers that have
issued more than once in our sample period, and a dummy variable to flag bonds that
have a primarily domestic focus. Moreover, to gauge the shift with the introduction of the
euro we include a set of interaction dummy variables for the pre- and post-EMU periods
with the different currency segments. Finally, we include the Herfindahl index of market
concentration for the particular currency/year combination.23
The results of the regression analysis are presented in Table 7. They are
suggestive of an important shift in the pricing of underwriter services quite specific to the
euro-denominated segment of the international bond market. The left-hand side columns
22
See Gande, Puri, Saunders and Walter (1997), Gande, Puri and Saunders (1999) and Roten and
Mullineux (2001).
23
The index is calculated as in Table 6 but for expositional purposes it is divided by 1000 (hence its range
is 0-10). For full description of the variables the reader is referred to the data appendix.
19
European market for bond underwriting
relate to the entire sample of 3110 bonds. After controlling for bond and issuer
characteristics, there appears to be a significant downward trend in fees of the order of 5
basis points per year. In addition to the overall decline, however, fees for eurodenominated bonds have declined dramatically after the introduction of the new
currency. The difference of roughly 65 basis points between the fees for bonds
denominated in one of the legacy currencies prior to EMU and those denominated in the
common currency after 1999 is highly significant (t-statistic of 9.42). This decline is
much larger than the 15 basis points decline in the US dollar denominated segment of the
market (t-statistic 3.05). In fact, the advent of the euro all but erased the 50 basis points
that separated the two market segments. The difference of barely one basis point is not
statistically significant at any conventional confidence level (t-statistic 0.22).
The other coefficients present a picture that is generally consistent with the
existing literature. Fees increase with market concentration, and are generally higher for
lower grade credits. Bonds with a primarily domestic focus or issued by issuers known to
the market carry lower fees and the same is true for bonds at the two ends of the maturity
spectrum. The relationship between fees and the size of the bond is non-linear. Fees
increase with the amount raised up to about $22 million, but decline for larger issues.
Interestingly this relationship seems to be very robust judging from the high levels of
statistical significance associated with the coefficients.
The right-hand side columns present the results of the same regression by splitting
the sample in the pre-EMU and post-EMU periods. The main thrust of the results is
robust to the sample split but there are a number of interesting nuances. There is an
acceleration in the pace of decline in fees in the more recent period, as evidenced by the
20
European market for bond underwriting
higher value of the linear trend coefficient.24 In the later period the effect of issue size on
fees is somewhat less pronounced but that of market concentration is much more
important. In fact, it appears that the overall significance of the Herfindahl index in the
regression is driven primarily by the post-EMU observations. We interpret this result as a
further indication of the greater openness of the international bond market after the
introduction of the euro has greatly reduced the degree of segmentation. Finally, there is
a decline in the fee discount associated with name familiarity as it can be seen from the
lower value and statistical significance of the dummies for domestic issues and repeat
borrowers.
4.2 Increasing returns to scale vs national differentiation
In the first part of our empirical investigation we provided evidence linking the advent of
the single currency to an increase in the contestability of the euro-denominated segment
of the international bond market accompanied by a marked decline in underwriting fees.
In the second part we will focus in more detail on the specific transformation of the
competitive environment in the post-EMU period. We ask the question of whether
borrowers used the greater flexibility in the choice of underwriter to favour bankers from
their own country, with whom they are more likely to have had an ongoing relationship,
or whether their choice was primarily driven by the global placing capacity of the larger
securities houses. We first examine the evolution of the relationship between the
nationality of the underwriter on one hand and that of the issuer and the currency of
denomination of the bond on the other. We then examine how these relationships
translate to the level of underwriting fees for the different segments of the international
24
While it is true that the coefficient reflects only three years of data (1999-2001) it is more precisely
estimated judging from the high degree of statistical significance (t-ratio of 5.80).
21
European market for bond underwriting
bond market. Finally we enrich our fee regressions with variables that account for these
relationships as well with proxies of the overall size of the market. In our analysis we pay
particular attention to the role of US underwriters in the euro-denominated segment of
the market.
Table 8 provides an overall picture of the trends in the relationship between the
underwriter and the issuer on one hand and the underwriter and the currency of
denomination of the bond on the other. The upper panel shows a slight increase in the
percentage of issues with a “ home currency” underwriter. Local bankers accounted for
58.2% of the total issues after 1998 as compared to 54.5% of the issues in the earlier
period. This overall picture, however, masks the opposing trends taking shape in Europe
and North America. Borrowers from both sides of the Atlantic have shown a stronger
preference for US underwriters for their dollar bonds in the post-EMU period (62.1%
compared to 39.1%) but have relied much less on local underwriters for their eurodenominated bonds compared to those in the legacy currencies (59.5% versus 80.5%). In
fact, the very high incidence of “ home currency” underwriters for bonds denominated in
the legacy currencies is the clearest indication of market segmentation in the pre-EMU
period.
The lower panel of Table 8 looks at the coincidence of banker and borrower
nationalities, a proxy of the existence of a business relationship. There is a pronounced
increase in the share of bonds underwritten by a “ home banker” in the most recent
period (38.7% compared to 20.3%). Once more, however, there are opposing trends
between North American and European issuers. While US investment bankers more than
tripled their share of business originated by US borrowers (46.9% versus 12.9%),
European issuers have moved away from their national bankers. The fact that these
22
European market for bond underwriting
opposing trends are present in both the dollar and euro/legacy segments suggests that the
boom in issuance of euro-denominated bonds did not translate to a profit opportunity for
the area’ s investment banks. It also implies that a business relationship between the
borrower and the underwriter plays a limited role in the choice of the latter.
It is instructive to look more closely to the nationality composition of the eurodenominated bonds issued by euro area borrowers. Table 9 provides the distribution of
underwriter nationality in terms of the number and volume of this category of bonds. The
contrast between the pre- and post-EMU periods is quite stark. Between 1994 and 1998,
US bankers underwrote only three issues out of a total of 169, or 3.6% of the overall
volume. In the post-EMU period they accounted for 41 issues out of 223 and nearly onequarter of the overall volume.
The increase in the share of the US investment houses is even more impressive if
one considers only first-time issuers from the euro area. This group of borrowers should
be the hardest to break in for an outsider as they may attempt to capitalise on existing
relationships with local banks and use them as underwriters. As it can be seen from the
bottom panel of Table 9 the distribution of business across the three groups of
underwriters is broadly similar to that for the overall group and slightly more favourable
to the US houses. This evidence points clearly to the fact that international competition
from the larger US investment houses has been the main new feature of the post-EMU
environment.
An obvious question concerns the precise effect of these shifts on underwriting
fees. Table 10 shows the average fees for different combinations of issuer nationality and
currency of denomination.25 For the issues denominated in one of the legacy currencies,
25
Figures in this table refer to arithmetic averages and are thus slightly different from those reported in
Table 5 which are value weighted.
23
European market for bond underwriting
or the euro, the table distinguishes between those that are issued in the “ home” currency
of the borrower (first column) and those issued in another legacy currency (second
column). The top panel of the table refers to the pre-EMU period. The impact of the
segmentation of the Europe market is evident in the high level of fees of bonds
denominated in the legacy currencies. Importantly, we notice that euro-area borrowers
issuing in their “ home” currency instead of another legacy currency, enjoyed a discount
of about 45 basis points.26 Overall, however, fees for bonds issued in one of the legacy
currencies were significantly higher than those for dollar bonds, independently of the
nationality of the issuer.
In the post-EMU period, fees for euro-denominated bonds were reduced
dramatically to levels comparable to those charged for dollar denominated issues.
Moreover, Euro-area borrowers seem to enjoy a “ home” currency advantage much like
that of their US counterparts. It is cheaper for each group to issue at the home currency
than it is for the other to do the same. This is further evidence of the “ normalization” of
the relationship between the two market segments.
In an effort to measure the effect of the closeness of the relationship between the
borrower and the underwriter on the underwriting fees and the importance of scale
economies, we enrich the list of explanatory variables in the fee regressions we employed
earlier by four variables. We include one dummy variable that flags bonds where the
underwriter’ s nationality matches that of the currency of issue, and one that flags the
coincidence of nationality between the borrower and the underwriter. We further include
a variable that measures the (logarithm of the) overall volume of bonds issued in a
specific currency/year pair and the share of that volume won by the issue’ s underwriter.
26
This result is consistent with the evidence from the regressions in Table 7 of a significant discount in
underwriting fees associated with issues of domestic focus.
24
European market for bond underwriting
The overall results are generally consistent with those presented in Table 7. The
pattern in the differences in fees between the bonds issued in the legacy currencies and
those issued either in US dollars or in euros is similar to that depicted in that table. It is
interesting to note, however, that the spreads are smaller than those reported in Table 7,
an indication that the expanded set of explanatory variables has gone some way into
explaining away the estimated effect of the euro.
A difference between the two sets of regressions is the negative sign of the
coefficient of the concentration index for the overall regression and the pre-EMU
sample.27 This result is counterintuitive, but should arguably be seen in conjunction with
the coefficients we obtain for the variable that measures the overall volume of issuance in
the currency of denomination of the bond. This variable is a very important factor in the
determination of underwriting fees. Fees decline by one quarter of a percentage point for
every ten-fold increase in the size of the currency segment. Moreover, the importance of
the overall activity in a currency segment increases in the later part of our sample. We
showed earlier (Table 6) that larger currency segments tend to be more competitive. We
argue that the change in the sign of the concentration measure is related to the high
degree of collinearity between the two variables.28
The variables that proxy for the closeness of the underwriter with the borrower
and/or the currency of issue have limited overall impact on fees. The coefficients are
rather small and not statistically robust. Only the connection between the banker and the
currency of denomination seems to be significant and this is mainly the case in the preEMU period. In fact, we interpret the fact that a “ home currency” underwriter could
27
The concentration variable is insignificant in the post-EMU sample.
Both variables are defined over currency/year pairs. This means that all bonds issued in euros during
2000 will have the same value for market concentration and the same value for market volume.
28
25
European market for bond underwriting
extract an additional 20 basis points in fees as a reflection of how important was the
effect of segmentation in the pre-EMU bond market in Europe.29
There is a very important discount in fees for bonds underwritten by a US
investment bank. The discount persists throughout the whole sample, but it is particularly
pronounced for the pre-EMU period when it exceeds 40 basis points. It is not clear how
one could explain this persistent fee reduction by US securities houses other than being
the result of their effort to further improve their prominence in the global market. This is
consistent with our evidence of the rapid increase in US houses’ market share among
bonds issued by euro area borrowers in the post-EMU period.
The variables that measure the market position of the underwriter have a rather
weak effect on fees and its sign is not consistent across sample periods. The coefficients
are small and not significant in most cases. The only exception is the coefficient on the
percentage share of the underwriter in the currency segment of denomination for the
bond. The overall sample result suggests that for every ten percentage points climb in the
share of the underwriter for a particular currency segment there is a 6 basis points
surcharge in fess. This result seems, however, to be driven by the pre-EMU sample
because in the more recent period the relationship seems to be economically and
statistically insignificant. We suggest that this is a reflection of the changes in the
competitive environment of the market where underwriters in their effort to boost their
share in a market with fewer barriers pass on to their clients part of the cost savings that
come with larger scale of operations.
As we saw in Table 7, the fee discount associated with borrowers that have
repeatedly accessed the bond market all but disappears in the later part of our sample. In
29
Earlier we documented that legacy currency markets were characterized by a very high incidence of
“ home currency” underwriters compared to the global pattern.
26
European market for bond underwriting
our view, this reflects a maturation of the market and improvements in the ability of end
investors to evaluate previously unfamiliar signatures, thus reducing the effort of the
underwriter to place these bonds.
Overall these results provide evidence that the market for bond underwriting in
the euro area has become more competitive in the post-EMU period. As a result,
underwriting fees have declined. The increase in competition has weakened the
importance of home currency advantage for underwriting houses but has not led to an
increase in the share of the business earned by “ home country” underwriters. Banks that
have benefited from greater market openness are those with global presence and those
that have been able to capitalise on economies of scale by increasing their share in
expanding market segments. The combined result of declining average costs and
intensified competition leads to a negative relationship between overall issuance volume
and fees, but the impact of an increase in the underwriter’ s share of the particular
currency segment is almost negligible.
5. Final remarks
Besides eliminating the exchange rate risk for transactions within the euro area,
the euro has contributed to important structural changes in the European bond markets.
These changes transformed what used to be a highly segmented market for corporate
bond underwriting along currency lines into a single market.
As a result, the single currency made it easier for investment banks to explore the
scale economies in the provision of underwriting services, lowered the entry barriers to
this industry increasing contestability in the industry, and made it easier for European
borrowers to benefit from the scope economies from combining commercial and
investment banking.
27
European market for bond underwriting
This paper shows that the arrival of the euro had an important contribution for the
ongoing reduction in the underwriting fees of corporate bonds issued in the new
currency, and that this reduction brought done the cost of underwriting services in the
euro area to levels similar to those in the US. The importance of this change becomes
apparent when we take into account that the average fee for bonds denominated in euros
in 1994 was about twice the corresponding figure for dollar-denominated bonds. The
paper also shows that the reduction in underwriting fees was largely due to greater
contestability of the investment banking business in the post-EMU European market.
Given that by the end of our sample period, underwriting fees in the euro area had
reached about the same level of these fees in the underwriting market for bonds issued in
US dollar, one is led to believe that the impact of the euro in this regard is sustainable.
The savings from the reduction in the bond underwriting fees are important, but
they are not the only implication of the structural changes introduced by the euro in this
context. Our analysis is mute about the impact the arrival of the euro might have had on
the cost of bond financing in the euro area. However, if the reduction in fees we unveiled
led to a reduction in the relative cost of this source of funding, then the euro has helped
put in train a number of longer-term effects.
Firstly, it will affect the composition of finance for European firms and especially
the blend between bank and capital markets funding.30 This redistribution is likely to
have implications for the transmission of monetary policy and other financial shocks to
the real sector reflecting the differential response of capital markets and banks to these
impulses. Secondly, it will affect the role of European banks in the process of financial
intermediation as a result of the differences between investment and commercial banking.
30
See Mayer (1988) for evidence on the differences in the way that firms finance investment in bank–
based systems and market–based systems.
28
European market for bond underwriting
Finally, it will have broader implications for the financial architecture in the euro-zone
with implications for risk-sharing opportunities and corporate governance that
differentially characterize systems with an emphasis on bank-based and market-based
finance.31 Some of these issues appear to be fruitful areas for future research.
31
Researchers, including Dewatripont and Maskin (1995), Sabani (1993) and Allen and Gale (1997), have
identified important differences between these financial system prototypes. See Thakor (1996) and Allen
and Gale (2000) for an extensive discussion of the implications of the financial system design.
29
European market for bond underwriting
Table 1. Restrictions on institutional investors’ investments (%)
Position limits
Countries
Pension funds
Portfolio allocations to foreign securities
Lifeinsurance
Pension funds
Equities
Life-
Bonds
companies
France
Germany
Italy
Netherlands
a
-<20
PMRa
--
<20
<20
<20
<20
insurance
companies
2
5
0
17
3
2
0
12
0
0
0
10
PMR = Prudent Man Rule.
Source: Davis and Steil (2001) and Lanoo (1998).
30
European market for bond underwriting
Table 2. International bond issues in euros and US dollar
Year
1996
1997
1998
1999
2000
2001a
Number of issues
In euros
887
1054
1001
1799
2094
1470
In US dollar
1390
1682
1343
1536
1375
1250
Volume of issues (billion US$)
In euros
171.0
157.3
219.0
504.3
473.7
438.2
In US dollar
276.7
346.9
438.8
587.5
654.1
615.9
a
Up to the third quarter.
Source: Thomson securities: IFR Platinum database, authors’ calculations.
31
European market for bond underwriting
Table 3. Location and sector of activity of bond issues denominated in euro
Year
Issuer’ s location (billion US$)
Euro-area
issuers
1996
1997
1998
1999
2000
2001a
95.6
92.8
144.7
372.7
346.8
326.2
Euro-area
a
Non-euro
area issuers
58.0
57.7
95.1
169.8
155.2
155.4
Issuer’ s sector of activity (%)
Non-financial
corporations
Banks
19.4
19.6
37.1
154.4
131.1
153.3
95.8
86.5
111.1
252.9
249.9
202.9
Other
financial
corporations
23.4
22.4
48.3
82.3
85.1
76.1
Up to the third quarter.
Source: Thomson securities: IFR Platinum database, authors’ calculations.
19
19
19
19
200
32
European market for bond underwriting
Table 4 Shares of euro-denominated bond issues by rating of the issue (based on volume)
Year
1996
1997
1998
1999
2000
2001 a
a
Bond rating
AAA
AA
A
BBB
BB
B
CCC
64.5
58.5
55.0
35.8
40.3
32.4
25.6
28.5
25.0
26.9
27.3
24.0
8.3
9.2
15.5
25.7
27.0
34.3
0.7
1.2
2.1
9.5
3.9
7.9
0.3
0.8
1.2
0.3
0.2
0.7
0.7
1.5
1.4
1.5
1.0
0.7
0.0
0.0
0.0
0.2
0.3
0.1
Up to the third quarter.
Source: Thomson securities: IFR Platinum database, authors’ calculations.
33
European market for bond underwriting
Table 5. International bond issuance by private sector 1994-2001
Year
Number of issues
USD
1994
1995
1996
1997
1998
1999
2000
2001 c
Total
94-98
99-01 c
Euro b
Total
Total proceeds
(billion US$)
USD
Euro b
Total
82
77
244
19.3
10.7
40.8
116
77
304
25.3
11.6
51.9
116
78
352
30.8
16.9
66.3
148
52
330
47.8
13.6
82.3
175
86
389
67.5
23.5
110.8
239
176
591
93.5
83.0
213.8
192
173
546
105.1
70.9
217.5
131
123
354
96.9
80.0
197.9
1199
842
3110
292.7
180.3
981.2
422
289
1213
114.3
54.0
237.0
634
520
1703
322.1
246.2
682.4
a
Average fees in these columns are calculated on a value weighted basis.
b
This column includes all issues in Euro, ECU and legacy currencies.
c
Up to the second quarter.
Source: Thomson securities: IFR Platinum database, authors’ calculations.
Average gross fees
(in % of amount raised)a
USD
Euro b
Total
1.299
1.516
1.082
1.134
0.860
0.748
0.583
0.586
0.816
1.096
0.636
1.553
1.797
1.543
1.732
1.664
0.803
0.685
0.426
0.888
1.654
0.638
1.377
1.533
1.306
1.222
1.086
0.768
0.579
0.512
0.851
1.259
0.622
34
European market for bond underwriting
Table 6 Concentration in the underwriting market for international bonds
Currency
US dollar
Euro
Deutsche
Mark
French
franc
Italian lira
Dutch
guilder
Yen
British
pound
Swiss
franc
H.I.
Top 5
Top 10
1994
1995
1996
1997
1998
1999
2000
2001
339.7
28.7
46.9
494.2
41.1
61.3
411.5
35.6
53.1
426.7
36.8
55.6
589.2
45.6
67.8
814.4
55.5
79.0
723.6
49.7
76.2
905.7
58.5
80.5
971.2
57.1
80.9
400.6
33.4
53.2
449.9
38.3
60.8
533.8
40.6
61.7
HI
Top 5
Top 10
HI
Top 5
Top 10
1372.2
67.5
91.1
785.5
49.6
71.5
948.3
59.2
74.9
889.4
54.2
76.6
967.9
53.0
71.1
HI
Top 5
Top 10
1511.1
77.5
94.4
2503.9
93.2
100.0
1555.1
79.1
96.0
1323.9
70.8
86.7
1792.7
82.6
92.9
HI
Top 5
Top 10
1758.0
77.5
92.7
1679.1
85.5
100.0
1936.0
87.3
100.0
1344.8
71.8
93.2
2658.7
74.9
93.1
HI
Top 5
Top 10
5344.7
100.0
100.0
2405.6
90.6
100.0
3510.9
100.0
100.0
3012.4
96.2
100.0
1783.6
83.1
100.0
HI
Top 5
Top 10
662.0
45.6
64.6
660.2
48.0
71.2
1238.4
60.7
79.2
714.4
50.9
73.6
957.2
61.3
82.1
675.6
47.0
65.6
633.7
46.5
70.5
790.5
53.9
81.5
HI
Top 5
Top 10
1122.1
65.0
86.2
1054.1
61.8
86.6
1734.3
79.0
94.9
863.7
58.0
79.2
670.8
46.4
70.9
922.5
59.0
80.5
730.5
50.9
78.5
703.7
49.4
73.3
HI
Top 5
Top 10
1488.8
70.3
87.7
1862.7
75.4
89.5
1262.3
68.4
85.6
1350.2
70.6
83.3
1213.8
69.5
90.9
1841.1
81.2
91.1
1498.8
77.7
93.3
2694.9
85.9
99.0
35
European market for bond underwriting
Table 7. Determinants of bond underwriting gross feesa
Variables
All issues
Pre-euro
Coefficient
Constant
Trend
Pre-USD
Pre-Euro
Pre-Legacy
EMU-USD
EMU-Euro
EMU-Other
Concentration
LOG(amt$)
LOG(amt$)_sq
AA
A
BBB
BB
B
CCC
JUNK
MAT(1-5)
MAT(>15)
MTN
Domestic
Repeat
Seasoned
TESTS:
Pre-Legacy = postEuro
Pre-Legacy =
pre_USD
Post-Euro = postUSD
Observ. / R2
a
T-ratiob
-0.347
-0.049
-0.118
0.179
0.388
-0.268
-0.259
-0.319
0.041
0.715
-0.070
0.135
-0.116
-0.242
-0.492
--0.177
0.661
-0.178
-0.139
0.428
-0.189
-0.122
0.035
-1.78
-4.81
-2.74
1.88
8.39
-4.84
-4.23
-5.55
3.99
11.20
-11.03
3.48
-3.71
-6.91
-4.70
--1.83
6.91
-6.83
-4.10
3.48
-3.49
-2.21
0.98
0.646
9.42
0.506
10.76
0.009
0.22
33.84%
3110
Coefficient
-0.909
-0.030
-0.196
0.116
0.345
T-ratiob
-3.66
-2.52
-4.09
1.25
7.30
0.001
1.075
-0.102
0.131
-0.143
-0.401
-0.339
-0.104
-0.425
-0.202
-0.232
--0.294
-0.252
0.047
0.80
11.63
-9.72
2.58
-3.20
-7.16
-4.49
-0.77
-8.35
-5.83
-4.15
--4.22
-3.22
1.02
0.432
10.43
30.40%
1619
Post-euro
Coefficient
T-ratiob
0.343
-0.114
1.44
-5.80
0.049
0.129
1.17
2.80
0.081
0.315
-0.034
0.097
-0.057
-0.065
-0.750
-0.272
-0.140
-0.066
0.446
-0.164
-0.042
0.035
4.25
5.18
-5.71
1.63
-1.27
-1.44
-5.02
-2.91
-3.67
-1.62
3.48
-1.86
-0.53
0.64
0.081
1.97
19.61%
1491
The model estimated the following structure
FEES =CONST + b1 TREND + b2 {CURRENCY×EMU} + b3 (BORROWER=CURRENCY) + b4
(BANKER=CURRENCY) + b4 AMOUNT$ + b6 {RATING} +b7 {MATURITY} + b8 MTN + b9
DOMESTIC + b10 (REPEAT BORROWER) + b11 (SEASONED OFFERING)
b
Inference uses standard errors robust to general forms of heteroskedasticity.
36
European market for bond underwriting
Table 8 Association patterns of underwriters, borrowers and currency of issue
Borrower
EMU
US dollar
Currency of issue
Euro/legacy
Total
Banker nationality matches currency of issue (in %)
N. America
Euro area
Total
pre-EMU
post-EMU
37.0
64.9
71.1
38.7
50.6
58.6
pre-EMU
post-EMU
25.7
41.7
89.3
68.6
63.9
55.6
pre-EMU
39.1
80.5
54.5
post-EMU
62.1
59.5
58.2
Banker nationality matches borrower nationality (in %)
N. America
Euro area
Total
pre-EMU
post-EMU
18.1
61.0
5.9
25.8
12.9
46.9
pre-EMU
post-EMU
12.9
8.3
51.5
40.8
28.6
6.5
pre-EMU
14.8
26.8
20.3
post-EMU
45.9
33.7
38.7
37
European market for bond underwriting
Table 9 European issuers’ choice of underwriter
Underwriter’ s nationality
US
Before EMU
After EMU
After EMU
Euro-area
Other Europe
Euro area issuers in Euro/legacy currencies
Amount
1488
36319
3142
($mil.)
No. issues
3
152
14
Amount
($mil.)
No. issues
Total
40950
169
30501
71912
33380
135793
41
127
55
223
Euro area issuers in euro: First time issuers only
Amount
22301
31572
19879
($mil.)
No. issues
24
52
26
73752
102
38
European market for bond underwriting
Table 10. Average fees by nationality of borrower and currency of denomination
Nationality
of the
borrower
Currency of denomination
Legacy currency
Foreign
“ Home”
currency
currency
Euro
US dollar
Other
All
currencies
1.750
1.168
1.103
1.304
2.000
1.160
1.602
1.139
After the euro
1.350
1.295
1.168
1.199
1.351
1.492
1.227
1.291
0.886
0.779
0.597
0.772
0.644
0.808
0.653
0.694
0.706
0.791
0.652
0.716
Before the euro
US
Euro-area
Other
Total
US
Euro-area
Other
Total
1.715
1.945
1.786
1.780
1.504
1.504
0.669
0.804
0.757
0.686
39
European market for bond underwriting
Table 11. Determinants of bond underwriting gross feesa
Variables
All issues
Pre-euro
Coefficient
Constant
Trend
Pre-USD
Pre-Euro
Pre-Legacy
EMU-USD
EMU-Euro
EMU-Other
Concentration
Volume
Bank=Currency
Bank=Issuer
US banker
Global share
Currency share
LOG(amt$)
LOG(amt$)_sq
AA
A
BBB
BB
B
CCC
JUNK
MAT(1-5)
MAT(>15)
MTN
Domestic
Repeat
Seasoned
TESTS:
Pre-Legacy = postEuro
Pre-Legacy =
pre_USD
Post-Euro = postUSD
Observ. / R2
T-ratiob
0.112
-0.032
0.113
0.240
0.460
0.034
-0.017
-0.387
-0.039
-0.240
0.057
-0.016
-0.254
-0.004
0.006
0.679
-0.066
0.138
-0.092
-0.190
-0.376
0.127
-0.603
-0.189
-0.103
0.429
-0.190
-0.120
0.040
0.56
-3.11
2.31
2.33
9.96
0.55
-0.26
-6.75
-2.49
-6.71
2.03
-0.60
-7.96
-0.98
3.47
10.77
-10.49
3.66
-3.02
-5.51
-5.54
1.21
-11.81
-7.25
-3.02
3.85
-3.39
-2.24
1.16
0.478
6.54
0.347
6.65
-0.052
-1.28
36.55%
3110
Coefficient
-0.463
-0.001
0.122
0.278
0.387
T-ratiob
-1.62
-0.04
1.92
2.72
7.73
-0.063
-0.179
0.196
-0.024
-0.432
0.002
0.007
0.928
-0.091
0.111
-0.123
-0.326
-0.484
-0.295
-0.648
-0.200
-0.192
--0.303
-0.241
0.029
-3.46
-3.56
4.05
-0.51
-6.59
0.28
3.91
9.63
-8.33
2.26
-2.84
-5.77
-5.26
-1.93
-8.52
-5.76
-3.48
--3.92
-3.14
0.66
0.265
3.82
34.28%
1619
Post-euro
Coefficient
T-ratiob
1.225
-0.158
4.38
-7.47
0.599
0.648
7.64
7.80
0.001
-0.448
-0.023
-0.008
-0.179
-0.006
0.000
0.395
-0.038
0.101
-0.047
-0.033
-0.779
-0.316
-0.174
0.009
0.407
-0.164
-0.035
0.053
-0.27
-6.78
-0.64
-0.23
-4.53
-1.06
0.04
5.92
-5.89
1.71
-1.09
-0.75
-5.13
-3.42
-4.56
0.21
3.37
-1.90
-0.47
1.00
0.049
1.14
23.76%
1491
40
European market for bond underwriting
References
Allen, F. and D. Gale. (1997). “ Financial Markets, Intermediaries, and Intertemporal
Smoothing.” Journal of Political Economy 105:3:523–546.
Allen, F. and D. Gale. (2000) Comparative Financial Systems. Cambridge: MIT Press.
Baxter, M and U J Jermann (1997) “ The international diversification puzzle is worse
than you think” , American Economic Review 87(1), pp170-180
BIS (2001) The changing shape of fixed income markets, BIS Working Paper n.104.
Danthine,J-P, F Giavazzi, L von Thadden (2001), "The Effect of EMU on Financial
Markets: A First Assessment" in C. Wyplosz, ed, The Impact of EMU on Europe and the
Developing Countries, Chapter 9, Oxford University Press.
Davis, E P and B Steil (2000) Institutional Investors MIT Press
Detken, C and Philipp Hartmannn (2000) “ The euro and international capital markets”
International Finance, 3:1 pp 53-94.
Dewatripont, M. and E. Maskin. (1995). “ Credit Efficiency in Centralized and
Decentralized Economies.” Review of Economic Studies 62:541– 555.
European Central Bank (2001a) “ The euro bond market” July 2001
European Central Bank (2001b) “ The euro equity markets” August 2001
Feenstra, R. C. (1992) “ How Costly is Protectionism?” Journal of Economic
Perspectives 6(3) 159-178.
Feenstra, R. C. (1995) Estimating the effects of trade policy, Handbook of International
Economics, vol III, Chapter 30, pp. 1553-1595, G. Grossman and K. Rogoff eds, Elsevier
Science.
Fisher, F (1997) “ Eurosecurities and their related derivatives” , Euromoney Books.
Forestieri, G. (1993) “ Economies of Scale and Scope in the Financial Services Industry:
A Review of Recent Literature” . Financial Conglomerates, Paris, OECD, pp. 63-124.
Gande, A., M. Puri, A. Saunders and I Walter (1997) “ Bank underwriting of debt
securities: modern evidence,” Review of Financial Studies 10, 1175-1202.
Gande, A., M. Puri and A. Saunders (1999) “ Bank Entry, Competition, and the market
for corporate Securities Underwriting,” Review of Financial Economics 54, 165-195.
41
European market for bond underwriting
Glassmann, D A and L A Riddick (2001) “ what causes home asset bias and how should
it be measured?” Journal of Empirical Finance 8, pp.35-54
Hamao, Y. and T. Hoshi (2000) “ Bank Underwriting of corporate Bonds: Evidence from
Japan after the Financial System Reform act of 1993” , Mimeo, University of California,
San Diego.
Head, K. and J. Ries (2001) “ Increasing Returns versus National Product Differentiation
as an Explanation for the Pattern of U.S.-Canada Trade,” American Economic Review
91(4), 858-876.
Johnson, H.G. (1960) “ The cost of protection and the scientific tariff,” Journal of
Political Economy 68, 327-345.
Kanatas, G. and J. Qi (1998) “ Underwriting by Commercial Banks: Conflicts of Interest
vs. Scope Economies,” Journal of Money Credit and Banking 30, 119-133.
Lanoo, K (1998) Institutional Investors, Capital Markets and EMU, in Institutional
Investors in the New Financial Landscape, OECD.
Mullineaux, D and I Roten “ Debt Underwriting by Commercial Bank-Affiliated Firms
and Investment Banks: More Evidence,” Journal of Banking and Finance (forthcoming).
Marshall, J.J. and M.E. Ellis (1994) Investment Banking and Brokerage, Kolb Publishing
Company, Co.
Mayer, C. (1988) “ New Issues in Corporate Finance.” European Economic Review
32:1167–88.
Rajan, R. G. (1996): “ The Entry of Commercial Banks into the Securities Business: A
Selective Survey of Theories and Evidence” . Universal Banking: Financial System
Design Reconsidered, I. Walter and A. Saunders (eds.), Chicago, Irwin, pp. 282-302.
Sabani, L. 1993. “ Marked Oriented versus Bank Oriented Financial Systems: Incomplete
Contracts and long– term Commitments.” Journal of International and Comparative
Economics 3:279–307.
Santos, J.A.C. (1998) “ Commercial Banks in the Securities Business: A Review,”
Journal of Financial Services Research 14(1), 35–59.
Smith, R.C. and I. Walter (2000) Investment Banking in the euro-zone. London: Financial
Times/Pitman.
Galati, G. and K. Tsatsaronis (2001) “ The Impact of the Euro on Europe’ s Financial
Markets” , Bank for International Settlements Working Paper no.100, July.
42
European market for bond underwriting
Tesar, L and I M Werner (1994) “ International equity transactions and US portfolio
choice” in J A Frankel (ed), The internationalisation of equity markets, Univ. of chicago
Press, Chicago, pp. 185-227.
Thakor, A.V. (1996) “ The Design of Financial Systems: An Overview.” Journal of
Banking and Finance 20:917–48.
Yasuda, A. (1999) “ Relationship Capital and competition in the Corporate Securities
Underwriting Market” , Mimeo, University of Pennsylvania.
43
European market for bond underwriting
Data appendix
Below we detail the other variables we use in the formal analysis:
GROSS FEE:
AMOUNT:
TREND:
EMU:
CURRENCY:
SECTOR:
RATING:
MATURITY:
VOLUME:
The underwriting fee for an issue expressed as a percentage of the total
funds raised. The definition is inclusive of several types of fees such as
management, selling commissions, preacipium etc. In some cases the
database provides a breakdown of the gross fees to individual
components but in many instances all the fees are combined into one.
We have elected to analyse the gross fee because it is the most
representative of the overall issuing costs.
The natural logarithm of the size of the issue expressed in US dollars
(LOG(Amt$)) at the time of issuance. In order to capture potential nonlinearities in the relationship between fees and the size of the issue we
have also included a squared term (LOG(Amt$)_SQ) among the
regressors.
To capture secular trends in the behavior in underwriting fees, we have
included a time variable defined as the difference between the and
1994 (the first year in our data).
Is a dummy variable that marks the single currency sub-period and is
equal to 1 for all bonds issued after 1 January 1999.
There are three dummy variables that capture effects specific to a
certain currency segment of the market. USD is a flag for US dollar
denominated issues, EURO flags issues in ECU or the euro, LEGACY
flags issues in one of the predecessor currencies of the euro, and
OTHER accounts for all the remaining currencies. Interaction terms of
these variables with the pre- and post-EMU periods are noted as pre<CURRENCY> and EMU-<CURRENCY>.
This is a set of seven dummy variables designed to capture specific
effects due to the nature of the issuer’ s line of business and it
corresponds to the first digit of the borrower’ s SIC classification.
A set of seven dummy variables that denote the rating category of the
bond at issue. We have used the following classification (AAA, Aa1-3,
A1-3, Baa1-3, Ba1-3, B1-3, Caa1-3) for Moody’ s ratings and (AAA,
AA, A, BBB, BB, B, CCC) for Standard & Poors ratings. We have
used the rating by Moody’ s, and in its absence the rating by S&P, if
available. The JUNK dummy variable denotes ratings that are in the
sub-investment grade category .
To capture effects related to the maturity of the bond, we have
classified the issues in three maturity buckets: MATLO for maturities
of less than 5 years and MATHI for issues that have longer than 15
years of maturity at issue.
This is a variable that measures the overall volume of issuance in the
particular currency of denomination of the bond during the specific
year of issue. The variable is expressed as the base 10 logarithm of the
overall volume measured in billions of US dollars.
44
European market for bond underwriting
AFFINITY:
SHARE:
DOMESTIC:
REPEAT:
SEASONED:
This is a group of dummy variables that flag the coincidence of the
nationality of the borrower the currency of issue and the nationality of
the underwriter. The variable Banker = Currency is equal to one when
the issue is denominated in the “ home currency” of the underwriter,
while the variable Banker = Borrower denotes the cases where the
nationalities of the banker and the issuer coincide independent of the
currency of denomination.
The variable measures the share (in percentage points) of the overall
volume of issuance in the specific currency segment during the year of
issuance that is captured by the underwriter. In the case of more than
one underwriter we have selected the syndicate member with the
largest share.
This is a dummy variable that corresponds to a flag in the database that
signals issues of primarily domestic focus.
This dummy variable flags bond issues by borrowers that have
repeatedly accessed the capital markets as evidenced by more than one
offerings recorded in the IFR Platinum database.
This dummy variable flags all “ seasoned” bond offerings defined as
those preceded by at least one previous issue by the same borrower.
45