EUROPEAN FINANCIAL MARKET INTEGRATION:
THE CASE OF PRIVATE SECTOR BONDS AND SYNDICATE LOANS
by
Christian Harm
PROFESSOR OF INTERNATIONAL BUSINESS
UNIVERSITY OF MÜNSTER, GERMANY
First draft: January 1997
First Revision: January 1998
This revision: May 1999
ABSTRACT
This paper uses individual transaction data on bonds and syndicate loans from European issuers to document market segmentation along national lines: banks manage bond issues because of ties to home institutional investors, and syndicate loans due to ties to home issuers. Home currency investment restrictions for many of Europe’s institutional investors are responsible for disproportionate market shares in the home currency markets of bond underwriters. Thus, the introduction of the Euro will greatly increase competition in bond underwriting, which will nonetheless benefit non-EU investment banks most. Many ambitions to create national champions for a European "Bulge Bracket" will be frustrated.
JEL Classification Codes: G15, G18, G21, G24
Key Words: Bond Underwriting, Loan Syndication, Europe, Market Segmentation, Euro, Bank Mergers
OUTLINE
Introduction 1
1. The Data 3
2. Concentration measures in the market for large debt issues 5
3. Why market segmentation? 6
4. The Tests 9
5. Issuers' Views on European Monetary Integration 18
6. Who are the Successful Competitors? 20
7. From Financial to Banking Integration? 22
8. Conclusion 26
Introduction
This paper documents that European markets for debt issues of large private sector firms and banks is still largely segmented along national lines. Moreover, it establishes that the sources of market segmentation are different for syndicated loans, where close ties between issuers and banks are important, and bonds, where close ties between banks and institutional investors are the dominant source of market segmentation.
Market segmentation is theoretically motivated by market structure considerations in markets with problems of informational asymmetries and regulatory constraints on institutional investor portfolios in different European countries. I hypothesize that reputational equilibria are important to overcome rationing phenomena in markets with asymmetric information. Issuers can establish reputations for high quality claims with banks. Banks can establish reputations for their due diligence when marketing financial claims to institutional investors. Both types of reputation serve to reinforce customer-client relationships and represent an impediment to competition, which naturally produces concentrated market structure outcomes, since reputational equilibria gain stability in transaction frequency.
The currently integrating financial marketplace in Europe provides an ideal testing ground for such stickiness in competition, hence indirectly for the existence of reputational equilibria. Restrictions in cross-border capital flows, international expansion of financial institutions, and cross-border advertisements of financial services had left EU financial markets largely segmented along national lines. Today, EU Banking and Investment Services Directives are trying to create a level playing field for financial intermediaries, and many of the previously existing restrictions on international financial transactions are now removed, although notable exceptions are made with respect to prudential legislation.
We are going to make the case below that prudential legislation restricting institutional investors such as pension funds and insurance companies to invest in home currency are the biggest impediment to financial market integration today. Currency restrictions on institutional investor portfolios reinforce previously developed reputational ties between financial intermediaries and institutional investors, and maintain existing business structures despite growing internationalization.
Such reputationally strengthened ties are shown to be particularly important in the market for bonds. In the market for syndicate loans, where all syndicate members share the lending information, due diligence considerations are not as important, and the reputational ties to the issuer are more pronounced as a source of market segmentation.
From here, the paper will proceed as follows. Section one introduces the database used for this study. Section two documents that concentration ratios in underwriting are much lower in Europe when compared to the US, although they are similar in individual countries. This crudely suggests that European markets are still segmented. Section three develops the theoretical argument that reputational equilibria reduce competition and enhance concentration. Section four presents the empirical evidence that European debt issue markets are still segmented along national lines. Section five discusses the results, establishing that portfolio limitations of institutional investors must be due to regulation, not preference. Section six identifies banks from non-EU countries as the principal actors in the non-captive markets, where product know-how has overcome historically grown business relationships. Section seven explores the relevance of an integrating financial marketplace after the introduction of the Euro for the development of transnationally operating banking institutions. It is argued that on balance, the further move to an integrated marketplace is going to benefit the intermediaries that are currently successful. Since these are mostly from non-EU countries, very few cross-border mergers are expected within the EU to create successful competitors in the European league tables. Section eight serves to summarize and conclude.
1. The Data
This study uses transaction data published by the International Financing Review. The database records a wealth of information for an individual bond or syndicated loan. This study uses: name and nationality of the issuer and nationality of the issuer's parent company; name and nationality of the manager and nationality of the manager's parent country; date and amount of the issue, including US$ amount; the currency of the issue; and a dummy variable whether the issuer is a real sector or financial corporation.
Of central issue are the nationalities of the issuer's parent country, the manager's parent country, and the currency. Naturally, ambiguities arose with respect to the parent country of the manager of an issue if the manager had recently been taken over by a foreign intermediary. This represented a problem with many of the UK financial institutions that had been bought by foreign banks, as well as with Credit Suisse First Boston. For lack of an alternative, the nationality of the purchasing bank was coded for the parent company. In a more detailed analysis, it was verified that the purchasing banks had indeed purchased market share in the respective countries.
Likewise, it represented a problem at times to decide on who is the arranger in a syndicated loan. Recently, more and more deals record more than one arranger or lead manager. However, in a different entry the IFR database records one definitive manager or arranger, which was used for the coding of the nationality of the manager variable.
The database was selected to include only issuers with parent companies from one of the EU countries and Switzerland. These together issued 7379 bonds and syndicated loans amounting to US$ 1545 bn. between 1/1/93 and 10/31/96. 862 of these issues, or US$ 174 bn. were managed by the issuer himself or an affiliate company, and were not included in the subsequent analysis, leaving 6517 issues worth US$ 1371 bn. for consideration. These are fairly evenly split between bonds and syndicated loans, with financial issuers dominating the bond market, and corporate issuers domiating the syndicate loan market. German financial companies, especially mortgage banks, account for the largest subset of bond issues at around 40% of all bond issues, while British corporates account for a similar magnitude of syndicate loans. The latter feature of the database may be due to the IFR operating out of London, and recording more British deals. It may, however, also reflect more loan syndication in Britain.
2. Concentration measures in the market for large debt issues
According to Smith (1998), 10 financial institutions in the US managed 83% of the syndicated loans, bonds, and equity issue volume in 1997 and 1996. Table 1 shows that European concentration measures in the debt markets are less extreme. Naturally, the smaller countries have much higher concentration in financial intermediation, while the larger countries feature concentration measures similar to, or less concentrated than, the US market. The US$ market of European issuers is less concentrated than the US market in general, and the European markets overall show significantly less concentration than the US market.
Thus, a cursory glance at the data suggests that individual European markets appear less concentrated than the US market. Also, there seems to be more competition in the issuer markets when compared to the currency markets: an issue we will return to below. Finally, the overall European market is less concentrated than any of the individual markets, suggesting that national preferences still determine market structure. I posit, that this is due to reputational effects between issuers, intermediaries, and institutional investors.
Market concentration measures |
# of banks to account for 83% of issuer market |
Market Size ($bn.) of issuer market |
# of banks to account for 83% of currency market |
Market Size ($bn.) of currency market |
UK |
21 |
402 |
19 |
270 |
Switzerland |
6 |
46 |
4 |
34 |
Netherlands |
16 |
137 |
3 |
58 |
France |
23 |
124 |
11 |
77 |
Italy |
28 |
59 |
12 |
34 |
Germany |
24 |
365 |
15 |
137 |
USA |
n/a |
n/a |
19 |
617 |
Japan |
n/a |
n/a |
12 |
49 |
All Currencies |
37 |
1371 |
37 |
1371 |
Table 1: Concentration measures in various issuer and currency markets. European issuers only. In the USA, 10 banks manage 83% of the annual issue total.
3. Why market segmentation?
Financial intermediation is notoriously plagued by an information problem between borrowers and investors. Banks' business is to raise and process information, else they would be in competition with the post office. Credit rationing due to asymmetrically distributed information has been long recognized in the Economics literature. Next to contracts and vertical integration, reputational effects can stabilize a market between autonomous parties despite the information problem. Such an equilibrium is hypothesized between the issuer of a financial claim, the financial intermediary that markets it, and the institutional investor that ultimately purchases it. This is depicted in Figure 1.
Figure 1: A reputational solution to an information problem
In a reputational equilibrium, an agent misrepresents his "type" and behaves honestly if the future rent-stream from such honest behavior outweighs the gains from one-time cheating. For the issuer, the reputational equilibrium is sustainable as long as the firm has positive net worth, since misrepresentation of the value of the issued claims can lead to bankruptcy or future rationing. The reputational equilibrium between financial intermediary and institutional investor can be supported by high management fees for bringing an issue to the market. Unfortunately, the IFR database could ony establish management fees for very few issues so that this hypothesis cannot be tested.
The anti-competitive implications of a reputational equilibrium are straightforward. The rent stream necessary to "bribe" the agent to be honest always represents a social loss. As long as the gain from one-time cheating does not grow proportionately to the scale of total operations, the more transactions an investment bank manages, the lower each management fee necessary to induce the desired behavior. Thus, as long as there is still competition between investment banks, the social loss inherent in the rent-stream necessary to stabilize the reputational equilibrium between investment banks and investor community is reduced with increasing market concentration. This provides an explanation for why a market for underwriting debt and equity of one trillion dollars per year in the USA can largely be managed by only 10 investment banks.
Reputational equilibria represent a hindrance to competition. Established business relationships in the financial industry may prove more resistant to change. The empirical tests below will attempt to support the existence of such a reputational equilibrium. With legislative and regulatory barrier to competition on a European scale largely removed, I intend to demonstrate that historically grown business relationships have so far resisted change.
4. The Tests
I run logit regressions to determine the probability that the manager of a debt issue is domiciled in a certain country. Thus the independent variables in the regressions to be discussed below are dummy variables coded "one" if the parent country of the managing bank is from the country to be examined in the respective regression, and "zero" otherwise. Such regressions I run for banks from Belgium, France, Italy, Luxembourg, the Netherlands, Spain, Sweden, Switzerland, the United Kingdom, Germany, the USA, and Japan.
Independent variables in the regressions are: a constant; ten dummy variables coded "one" if the parent company of the issuer is domiciled in one of the ten European countries examined in the regressions, and "zero" otherwise; twelve dummy variables coded "one" if the currency of the issue comes from one of the ten European countries examined or the USA or Japan; dummy variables for all years except 1993; a dummy variable coded "one" if the issuer is a financial institution, and "zero" otherwise; and the US$ amount of the issue. The currency dummies serve as proxies for ties between banks and institutional investors, since in Europe the latter are typically constrained to invest mainly in local currency. Then, foreign issuers may choose banks from the country of the desired currency, because banks with established relations to institutional investors can easier place the issue.
Thus, I run twelve logit regressions to determine the probability that the manager of an issue is domiciled in any one of the examined countries including the same 28 independent variables in all regressions. These twelve regressions are run on all data, only on bond issues, and only on syndicated loan issues. Results are shown in figures 2 through 4. On the X-axes, the figures show the probability - given a base case - that a bank from the country examined is chosen if the issuer is domiciled in the same country. On the Y-axes, the figures show the probability - given a base case - that a bank from the country examined is chosen if the issuer demands the currency of the country. The base case is an issuer from a country other than the ten European countries examined trying in 1993 to place a US$ 100 million debt issue denominated in a currency not from any of the twelve currency countries examined.
Figure 2 shows that both issuer and currency effects are statistically and economically significant for almost all countries. Issuer effects are statistically insignificant for Belgium and Luxembourg, while the currency effect is statistically insignificant for Sweden. Moreover, figure 2 also suggests that the currency effect is clearly stronger than the issuer effect. Reputational ties seem more important between intermediary and institutional investor than issuer and intermediary.
Figure 2
The latter point can be more closely examined by distinguishing between bonds and syndicated loans. Since syndicate members are sharing all relevant lending information, a reputation for due diligence is not as pivotal for the arranger of a syndicate loan when compared to the bookrunner of a bond issue. Results for bonds are shown in figure 3, for syndicate loans in figure 4.
Figure
3
The results for bonds in figure 3 show a clear dominance of the currency effect both economically and statistically. Statistically, the issuer effect is only significant for France, Germany, Switzerland, and the UK, while it is marginally significant for the Netherlands. The currency effect is statistically insignificant only for Sweden and Italy, while it is economically insignificant for Spain. In six out of ten countries, the probability of being chosen as a bookrunner of a bond issue is greater than 50% if the bank is domiciled in the country of the currency desired.
Figure 4
The results for syndicate loans are very different according to expectation. While the regressions for Belgium and Luxembourg are altogether insignificant, all other countries show significant issuer effects. France, Sweden, and Spain show insignificant currency effects. The issuer effect economically dominates the currency effect in five countries.
The results on syndicate loans support the traditionally imputed information gathering and monitoring role for banks. Only issuers of a certain size and recognition can afford to tap the bond markets. Here, the information gathering and monitoring role is - in part - contracted out to rating agencies. For lesser known issuers, such contracting out of the monitoring function is uneconomical, and has to be internalized by banks. Thus, reputational equilibria between issuer and banks strengthen the long-run lending relationship. Figure 4 provides ample evidence that this type of equilibrium of historically grown bank/client relationship is still visible despite increasing competition in the European marketplace.
Needless to say, the null hypothesis for the tests was that neither currency nor issuer effect would be observed under a scenario of financial integration in European markets. In practice, the results here establish that even the presumably fiercely competitive markets for underwriting debt of Europe's prime issuers is still dominated by historically grown ties between issuers and banks, and especially between banks and institutional investors.
One further observation is of interest from the results shown above. Typically, Germany is viewed as a bank-based financial system, while the United Kingdom would be described as a market-based financial system. While either financial system has to have strong financial institutions, banks in the German system are at times presumed to gain special advantages with their clients through board seats and equity stakes. Figures three and four suggest rather the contrary: that issuer ties are more important in the UK rather than Germany, especialy for syndicated loans, where industrial clients dominate. At least on an aggregate national level, the anticompetitive nature presumed for the German universal banking system does not show in the data.
Figure 5: Home
issuer vs. other significant issuer effects in the syndicate loan markets
Another question is, whether banks may not only have a strong position in their home markets, but also possibly in other markets. In that case, the home bias due to issuer or currency nationality may be significant, but not unique. Figure 5 shows the probabilities associated with the logit regressions on syndicate loans on the home issuer effect and issuer effects from all other statistically significant issuer countries. For the UK, only two other issuer countries (Italy and Luxembourg) show up significant in the regressions, and far less so than the probability associated with issuers from the UK. Dutch banks are also somewhat successful in the market for issuers from the UK, while German banks are also recognized in the market for Belgian issuers. Other than that, only home issuers significantly increase the probability of a bank to be chosen to arrange a syndicate loan. Not shown for the case of syndicate loans is that European banks are not likely to be chosen as arranger in any of the currency markets outside the home market.
Figure 6 shows that bond markets are slightly more competitive, but the home currency effect still dominates for banks from all countries. Banks from the USA and Switzerland are successful in a number of currency markets, while banks from only three EU countries are successful in one other currency: French banks in the Luxembourg Franc, German banks in the Italian Lira, and Dutch banks in the Swedish Krona. Thus, with the exception of US and Swiss banks, bond markets are dominated by the home currency effect. Not shown for the case of bonds is that French banks have a slightly significant edge in the market for German issuers. All other bond markets feature - if any - only a home issuer effect.
Figure 6: Home
currency vs. other significant currency effects in the bond markets
In summary, we can conclude from the analysis that national effects still strongly dominate the choice of financial intermediary to bring an issue to market. Historically grown bank/client relationships still dominate the market for syndicate loans, while traditional ties between financial intermediaries and institutional investors still segment the bond markets. The latter point is indirectly inferred given an established bias of European institutional investors to invest the bulk of their assets in home currency. This investment bias can account for the home currency effects that are so strongly documented in the regressions. A remaining question, however, is, whether institutional investors invest in local currency due to preference or regulatory restrictions. Here, the borrowing patterns of the issuers recorded in the database can be of some guidance.
5. Issuers' Views on European Monetary Integration
Under the assumption that corporations with long run commitments in foreign countries typically have more revenue than cost exposure in the respective currency markets, they have an incentive to hedge their foreign operating exposure by borrowing in the respective currency. Since the exposure of financial companies is much less transparent, I use the borrowing patterns of non-financial corporations to infer implied hedging behavior.
The database lists 2756 issues amounting to a US$ equivalent of US$ 627 bn. in the sample period. Of this amount, US$ 272 bn. were borrowed in home currency, leaving US$ 355 bn. as debt issued with potential hedging implications. Of these US$ 355 bn., US$ 289 bn., or 81%, were issued denominated in the US$. Even though the US$ is the major invoice currency in world trade, it can safely be argued that the US$ is used as a hedging tool overproportionately to its role in these corporations' foreign currency exposures. All European currencies were only chosen for 15% of the issue volume of non-financial corporations, while these companies' operating exposures in Europe must arguably be higher. It is also extremely unlikely that there is a swap interpretation for the observed borrowing pattern: swaps are based on rating arbitrage, and why should the same European issuer have access to the US$ market, while it has to swap into another European market?
Thus, the corporate borrowing patterns suggest that market participants seem to see only a reduced need for hedging European currency exposure. Issuers in the debt markets perceive the European Monetary System as stable even after the band was significantly widened following the crisis of the summer of 1993. If this, however, is conventional wisdom among issuers, it is unlikely that institutional investors restrict their investments to home currency for reasons of risk preference. If - as documented by European corporate borrowing behavior - the EMS can be regarded as largely stable, then the currency restrictions in institutional investor portfolios can only be explained by regulatory restrictions placed upon pension funds. This is the allegation made by Mantel (1997) for the case of pension funds. Prudential supervision requires European pension funds, but also insurance companies, to invest their assets in the currencies of their liabilities.
With the advent of the Euro in 1999, institutional investors will have large potential to reshuffle their asset portfolios, which will create a small revolution in asset management, but in the medium term also in financial intermediation in general, since a truly integrated European market will apply the same pressures towards concentration as now observed in the US market. Since the currency effect was particularly pronounced in the bond regressions discussed above, it can be inferred that European bond markets will experience a competitive revolution after the formal introduction of the Euro.
6. Who are the Successful Competitors?
Having identified reputational equilibria between issuers, intermediaries, and institutional investors as sources for strong business relationships, we can define banks' home issuer and home currency markets as captive markets. It is then interesting to identify market shares in the non-captive markets, i.e. in the market for issues managed by banks from countries other than those of the issuers or the currencies involved. The leaders in these markets arguably have the best potential to assume a principal position after the introduction of the Euro, since they have already managed to successfully use their product know-how to break into historically established business relationships. A total of US$ 194 bn. in bonds was managed by banks not from the issuer's or the currency's country. The non-captive market for syndicate loans amounted to US$ 259 bn. during the sample period. In total, one third of the issue volume in the database is non-captive. Table 2 identifies market shares of individual banks in these non-captive markets.
Ten banks have been able to capture more than two thirds of the non-captive bond market, while in the syndicate loan market the ten most successful competitors in the non-captive market account for slightly less market share with 62%. 25 Banks account for 83% of the total non-captive market. This number can be compared with the concentration statistics of table 1. The non-captive market is more concentrated than European debt markets in general, but not yet as concentrated as the US market. This is a further suggestion that absent historical ties maintained through reputational equilibria, financial markets exhibit strong trends towards concentration.
Bonds |
Syndicate Loans |
||
Financial Institution |
% Share |
Financial Institution |
% share |
Credit Suisse First Boston |
12,6 |
Credit Suisse First Boston |
12,4 |
Swiss Bank Corp. |
10,9 |
Union Bank of Switzerland |
10,8 |
Union Bank of Switzerland |
8,0 |
Swiss Bank Corp |
6,3 |
Merril Lynch |
7,1 |
Chemical Bank |
6,0 |
JP Morgan |
6,1 |
Banque National de Paris |
5,5 |
Hong Kong & Shanghai Bank |
5,7 |
ABN Amro |
5,4 |
Morgan Stanley |
5,1 |
Deutsche Bank |
5,1 |
Goldman Sachs |
5,1 |
JP Morgan |
4,6 |
ABN Amro |
4,2 |
Citibank |
3,1 |
Banque Paribas |
3,9 |
Barclays |
3,1 |
Table 2: Market shares of individual banks in non-captive markets for bonds and syndicate loans.
Moreover, table 2 indicates that most of the successful competitors in the non-captive European debt issue markets are not EU-based, and are rarely from one of the initial member countries of the Euro. With the evidence presented in section 4 above, the introduction of the Euro will arguably leave its bigger impact on the bond markets. There, however, only two banks - ABN Amro and Banque Paribas - are among the top ten, and they are at the bottom of this list. Among the banks from ranks eleven to twenty, there are only two more banks from the likely Euro member countries, Kredietbank from Belgium, and Deutsche Bank from Germany; and there are two more banks from the UK, Barclays and Hambros. The other banks among the ranks from 11 to twenty are from the USA or Japan.
A similar picture emerges in the non-captive market for syndicate loans: among the top ten, three - Banque National de Paris, ABN Amro, and Deutsche Bank - are domiciled in the Euro area. On position ten is Barclays from the UK. Among the next ten banks, there is only WestLB from Germany from the Euro area, and Nat West and Warburg from the UK.
In all, among the 25 most successful banks in the non-captive European debt issue markets, only 6 banks accounting for 11,5% of market share are domiciled in the countries comprising the first round of the Euro. The successful banks tend to be precisely not the ones from Europe, but rather banks with a long tradition for operating in more dynamic financial markets, such as the Swiss and US financial institutions. Why could that be the case?
7. From Financial to Banking Integration?
Due to their high leverage, banks in all countries are subordinated to a regulatory structure developed in the interest of depositors. Figure 7 depicts the complicated corporate governance structure of banks, where regulators as agents of depositors compete with a board of directors as the agent of shareholders to govern the banking institution.
Figure 7: The governance of banks
Since regulators are largely interested in avoiding failure, they are predominantly survival oriented. Radner (1996) has shown that survival interests are generally inconsistent with efficiency goals such as the ones favoured by shareholders, creating conflict among the two. Typically, due to its legislative foundation, regulation emerges as the stronger force. Prowse (1998) has shown that control changes in US Bank Holding Companies were often initiated by regulators rather than shareholders.
In regulated environments, financial sectors shape first and foremost around regulatory interests, with efficient structures emerging only in areas where regulation allows significant degrees of freedom for the managements of financial institutions. Such forces driven by idiosyncratic regulations can create the observed heterogeneous structures of financial sectors in different European countries. The efforts of the European communities to create a single banking market in Europe by adopting the principle of a single banking license have certainly increased competition in financial services, and have indirectly led to deregulation and increased focus on efficiency in several countries. Dermine (1995) documents how interest margins in European banks converged over the last decade.
In the sheltered banking environments of the past, it was not always necessary for banks to more actively develop bond markets that could be more efficient to cater to the financial needs for larger clients, possibly explaining the relative underdevelopment of financial markets in Europe when compared to the USA. In the deregulating European financial markets opening for competition, issuers increasingly discover the bond markets, and banks increasingly discover the risk-spreading benefits of loan syndication. Financial institutions from countries with a longer tradition in these markets could enter the European markets due to a comparative advantage in managing such products.
These foreign banks, mostly the three large Swiss banks and several of the large US investment and commercial banking houses, were able to use their comparative advantage in managing the relevant products to break into historically developed relationships between issuers, intermediaries, and institutional investors in the individual European markets.
With the natural tendency towards concentration in underwriting, the Euro will further remove barriers to entry into this market, and the now successful financial institutions will likely emerge as front runners in ever more concentrated markets. The merger between Swiss Bank Corp and Union Bank of Switzerland represents a deal most certainly geared at a future lead position in European league tables.
Yet, most other banking mergers observed in Europe today are not. If the hypothesis sketched out above - that European financial regulators have at times been overly protective of their banks, and have thereby inhibited efficiency-oriented management - is correct, then many of the (mostly domestic) bank mergers must be judged as rationalization mergers to eliminate inefficient structures built up in the past. Thus, banks are preoccupied to eliminate excess capacity, particularly in retail distribution, and mergers are primarily engaged in to consolidate overlapping branch networks, and unite information technology investments. Even "bankassurance" at times carries the (not fully convincing) taste of trying to buy more products to fill banks' excess retail capacity. Few mergers seem to be a preparation for a more integrated European financial marketplace.
The lack of cross-border mergers in Europe is no doubt explained by the lack of sufficiently integrated markets for financial services. The primary purpose of this paper was to show that even those financial markets that are most likely to be integrated on a European level are still heavily segmented today. The evidence presented in section four suggests that the most important source of market segmentation in debt issue markets of large private sector borrowers is the regulatory constraint on institutional investor portfolios to invest largely in home currency. The introduction of the Euro will eliminate this source of market segmentation in the coutries that join the Euro in the first round.
With that, it would be tempting to argue that the introduction of the Euro would provide for the first time a reason for cross-border mergers in Europe. This would be a proper prediction if European debt markets were totally segmented now, and all banks would start the era of competition only equipped with their domestic contacts to institutional investors and issuers. However, section six has shown that foreign banks with superior product know-how have already captured significant market share, and are the dominating force in the non-captive markets, where product know-how has proven more important than historically grown business relationships.
With continuing trends towards market concentration in underwriting and syndication markets based on sound reputations, the future leaders in the markets examined in this paper can be expected to come from the pool of already successful financial institutions. Since there are only few continental European banks among them, one can also expect only a few of them to succeed in capturing more market share through mergers. If US market structure is any guide, then the run for the top of European league tables has largely been decided.
8. Conclusion
This paper has used individual financial transaction data to document that the markets to manage bond issues or arrange syndicated loans for leading European issuers are still rather segmented along national lines. This paper argues that this evidence of market segmentation finds its explanation in the fact that financial intermediaries are embedded in reputational equilibria between issuers and institutional investors, and that these equilibria function as barriers to entry. This is supported by the US experience, where an underwriting market of an annual US$ 1 trillion or more is largely managed by only ten financial institutions. Due to historically grown ties to issuers and investors, European markets are far more segmented.
According to predicition, reputational ties between issuers and intermediaries are more important in the market for syndicate loans, while reputational ties between intermediary and institutional investor are more important in the bond market. Since institutional investors are largely constrained to invest in local currency, barriers to entry to the market financial claims to institutional investors are proxied by barriers to entry into individual currency markets.
I establish in logit regressions that the probability for a bank from a certain country to manage a debt issue increases dramatically, when the issuer or the currency are located in the same country as the financial institution. Also, with the exception of Swiss and US banks, who are successful in a number of foreign currency markets, no banking groups have made significant inroads into any foreign markets.
I explore the consequences for banking integration - i.e. the creation of transnationally operating banks in Europe - from these findings. The most successful banks in the non-captive European financial markets are not domiciled in Europe. Presumably superior product know-how has swept Swiss and US banks to the top of European league tables. Further concentration can be expected with the removal of a significant barrier to entry after the introduction of the Euro. Yet, the nature of the reputational equilibrium suggests that the successful competitors in a further liberalized market are the same institutions that already dominate the process at this time.
With that, only few European banks can realistically expect a significant boost in market share from a cross-border merger. The merger wave sweeping through Europe during the last few years largely addresses problems associated with too extensive retail branch networks and efficient utilization of information technology investments. If the analysis presented in this paper is correct, many domestic mergers created to have a national champion in the European "bulge bracket" will likely find their ambitions frustrated.
BIBLIOGRAPHY
Dermine, Jean (1995): "European Banking Integration Ten Years After", European Financial Management.
Dutch Euro Project Group of Pension Funds (1997): "Manual for the Implementation of the Euro by Pension Funds".
Edwards, Jeremy and Klaus Fischer (1994): Banks, finance and investment in Germany, Cambridge University Press, Cambridge, UK.
Jaffee, Dwight M. and Thomas Russell (1976): "Imperfect information, uncertainty, and credit rationing", Quarterly Journal of Economics.
Klein, Benjamin and Keith B. Leffler (1981): "The Role of Market Forces in Assuring Contractual Performance", Journal of Political Economy (August), pp. 615-641.
Kreps, David M. and Robert Wilson (1982): "Reputation and Imperfect Information", Journal of Economic Theory, pp. 253-279.
Mantel, Jan (1997): "Don't worry. It's only a Revolution", Merril Lynch research paper.
Prowse, Stephen (1998): "Corporate Control in Commercial Banks", Journal of Financial Research, forthcoming.
Radner, Roy (1996): "Survival", mimeo, STERN School of Business, New York University, New York, USA.
Ryck, Koen de (1996): "European Pension Funds. Their Impact on European Capital Markets and Competitiveness", European Federation for Retirement Provision, London, UK.
Smith, Roy (1997): "The European Securities Industry Under a Single Currency", mimeo, STERN School of Business, New York University, New York, USA.
Smith, Roy (1998): "1997 Global Capital Market Activity and Market Shares of Leading Competitors", mimeo, STERN School of Business, New York University, New York, USA.
Steinherr, Alfred and Huveneers, Christian (1990): "Universal banks: The prototype of successful banks in the integrated European Market?", Research Report No.2, CEPS Financial Markets Unit, Centre for European Policy Studies, Brussels.
Stiglitz, Joseph and Andrew Weiss (1981): "Credit rationing in markets with imperfect information", American Economic Review, June.
Vander Vennet, Rudi (1997): "Determinants of EU bank takeovers: a logit analysis", Universiteit Gent Working Paper no. 97/29, April.
Williamson, Oliver E. (1979): "Transaction-Cost Economics: The Governance of Contractual Relations", Journal of Law and Economics, April.