Brookings-Wharton Papers on Financial Services

Comments and Discussion

[Globalization of Financial Institutions: Evidence from Cross-Border Banking Performance]

Comment by Raghuram G. Rajan: This is an exhaustive and exhausting book masquerading as a paper. It presents everything you ever wanted to know about banking, but did not dare to ask. It is a very nice read despite its size. I am going to address some issues that the authors did not spend much time on and then discuss their results.

Why do banks go cross-border? To some managers, it seems self-evident: there are cost savings in mergers. You can close branches that face each other across the street. You can share back offices. There also are benefits of cross-selling, which means selling socks and stocks in the same place--a strategy once tried by Sears. Merging also offers the benefits of diversification and risk reduction. Finally, and this is where cross-border may matter, there are advantages of "globalization," a vague word that means all things to all people. It could mean, for example, that your clients are going abroad. It certainly is offered as a rationale for Japanese banks going cross-border: to follow their traditional clients. Some people argue that another reason for going cross-border is to keep abreast of the latest trends. People argue, for example, that you have to have a foot in the United States where a lot of innovation is occurring. Having a physical presence is essential to obtain information about what is going on and to transmit that information to the rest of your operations.

This is a partial list, but it includes some important reasons that practitioners offer for merging. Academics, and to some extent analysts, are a little more skeptical about some of these reasons. We have learned to be skeptical because a large amount of our research shows that mergers typically are not done for the best motives. [End Page 126]

For example, consider scale economies. The evidence is that scale economies tend to be exhausted pretty soon. By the time you reach a bank of $10 billion, there are very few economies left to exploit. Since many of these cross-border mergers involve giants that are already much bigger than $10 billion in assets, it is not clear what scale economies are being achieved.

Information technology could also reduce, rather than enhance, minimum optimal scale. For example, think about phone banking. People say that you need hardware and software to implement a phone banking system and that it will cost a packet. But in practice small banks estimate that these costs are on the order of $50,000 to $100,000 for an off-the-shelf system. You certainly do not need a $100 billion bank to implement a phone banking system. Also the alliances and networks that banks can belong to obviate the need for a branch banking system. For example, I can withdraw money from my Citibank account at an automated teller machine (ATM) in Heathrow airport in London, but I can also withdraw it from my account in tiny Hyde Park Bank using the same ATM. So the alliances tend to equalize the playing field, and again scale economies do not matter then.

Also, there are some diseconomies of merging really big banks. In Japan, for example, Dai-Ichi Kangyo took more than twenty years to merge the operations of the different parts of the bank. And in some big bank mergers the information systems in each bank cannot talk to one another. In fact, the systems sometimes run in parallel rather than being merged because the costs of merging them are enormous. So there might be some diseconomies of scale also.

Perhaps most important are issues of management. Even when you have a merger in the United States across fields, a whole set of new conflicts arises between traders and investment bankers, between investment bankers and commercial bankers, and so on. A cross-border merger introduces cultural differences, creating yet another reason for conflict.

There is this myth that, after a merger, employees will live together happily, will cooperate, and will sell services together, when in fact a lot of conflict arises.

In addition to the conflict, you also have the issue of inter-firm equity. The investment bankers in Germany do not like the investment bankers in England being paid much more than they are, even if they come from [End Page 127] different labor markets. So there is a lot of heartbreak. If you try to force the investment bankers in England to work for less pay, you are operating a human capital-intensive firm, and human capital can leave at any time. That is precisely what Deutsche Bank has been facing. It bought a lot of banks at extremely high costs and lost a lot of people very soon after.

There also are operational risks. The Japanese banks discovered this at extremely high cost. Some trader sitting in New York made unsupervised bets using the bank's capital and nearly sank the bank. It is also not clear that cross-border mergers increase diversification nor that more diversification necessarily reduces bank risk.

And finally, in all of these mergers, there is never a well-articulated plan as to where the cost savings are going to come from and what kinds of benefits clients will get. Why do your clients care that you can serve them in Japan as well as the United States? Is it because of frictions in payment systems? How long is that advantage going to last? Synergies are always more apparent before a merger than after.

In this paper, the authors point out that one of the unstated reasons for cross-border acquisitions is, for example, overconfidence. Because you think you can do things better than anybody else, you go out and buy firms. There is certainly some evidence for managerial hubris in the United States.

Banks find new ways to lose money every ten years. It might well be that cross-border mergers are the new way. Bank managers need a seat when the music stops. The one way to ensure that management has a place is for it to do the acquiring before it gets acquired by somebody else. So are mergers a way for banks to get rid of their excess capital? I do not know the answer.

Now consider this paper. The paper presents evidence suggesting that foreign-owned banks are less efficient than domestic banks, which suggests either that the efficiencies have been mismeasured or that banks have been merging for all of the wrong reasons.

The authors do two things. First, they break up the banks in particular countries based on the owner's country of origin. This is a nice way of carving up the data. Second, they look within the United States and try to see these effects across states. If they can see these cross-state effects in the United States, perhaps we can learn more about the cross-border effects.

Let me talk about the results in the United States. The authors do not see any systematic effect of distance in the United States. They see some [End Page 128] effects going this way, and some effects going that way. But the fact that you are on one side of the country and you own a bank on the other side of the country does not have much effect on the operations of the bank on the other side of the country. This does not necessarily mean that you can run a bank efficiently sitting on the opposite side of the country, but it may mean that the way in which you selected the bank to buy compensates for whatever lack of efficiency you may have. Conversely, it might be that the acquired bank has competent management who can work independently, which is why you bought it in the first place.

In short, the systematic underperformance of foreign-owned banks in another country does not have a cross-state counterpart in the United States. Physical distance does not seem to matter a lot. This implies that currency and language differences may well be key. Another possibility is that the underperformance we find internationally is mismeasurement.

Also, unlike previous papers, this study does not simply find that all foreign-owned banks are less efficient than domestic banks. It offers a more nuanced view. Some foreign-owned banks are more efficient. So U.S. banks are generally more efficient in other countries than domestic banks. Japanese banks are particularly inefficient in other countries. It may well be that U.S. banks have been more efficient since they cleaned up their act in the late 1980s and that Japanese banks have been doing the wrong thing at the wrong time. They have been buying at the peak in the United States and getting out in the trough, and this might account for the fact that they have been less efficient.

I would like to see more on this subject. For example, the Spanish domestic banks are less efficient than foreign banks in Spain. So perhaps if we had more data, we could look at Spanish banks in other countries and ask if they underperform in other countries also. This approach is consistent with the U.S. evidence; for example, U.S. domestic banks typically are more efficient than foreign banks in the United States, but it would be nice to see if this holds for other countries.

Even with the caveats, this study finds that foreign-owned banks underperform in a country. Before concluding that cross-border mergers are a mistake, we should recognize the limitations of the analysis. For example, the nature of banking business may be different for banks from different countries. Perhaps U.S. banks take more risks than other banks in part because they have learned to manage these risks. When we look at profit efficiency, are we capturing the greater risks that banks might be taking? [End Page 129]

Another problem is that of transfer pricing. Let me give an example. Suppose that a Japanese firm interacts with a Japanese bank in Japan, and as a favor to that Japanese firm it provides a service at low cost in the United States. The service is provided in the United States, and the profits are recognized in Japan. And the U.S. operations look really inefficient, because profits show up in the home country. Neither of these issues is important cross-state, which again highlights the importance of understanding that finding.

One way of substantiating the conclusion that cross-border mergers are generally inefficient is to examine the effect of a change in nationality of ownership on the operational efficiency of a bank. While such a fixed-effects regression would not fully address the issues of greater risk and transfer pricing, it would at least address some issues of selection.

The bottom line is that the virtues of cross-border mergers still are not obvious, a conclusion with which I heartily agree.

Comment by Robert McCormack: Most of my response is going to be taken from the slant of commercial and wholesale banking and investment banking, because that is where I lived for so long. The paper has some important strengths. First of all, it is a great compendium of the literature. Just the bibliography alone would be valuable to anyone who wants to understand the trends affecting efficiency in business. Second, by and large, the paper confirms that efficiencies are important in mergers, but the data in the end do not really confirm the home field advantage. The paper is very interesting to read, but I would like to offer a few comments from a practitioner's point of view.

First, you cannot tell much about a global business by looking at local banks. I was surprised by the example of Spain that was used in the paper. I used to have a manager in Spain who took great pride in pointing out how well his business did against the Spanish banks in Spain. And, in fact, the corporate bank in Spain always was the number one bank in terms of returns, profitability, and everything you could possibly measure with regard to efficiencies. The problem was that we had a horrible business in Spain. We did not have a large enough customer base, and after adding in the cost of serving Spanish customers all around the world, we did not make much money in Spain.

One of the reasons that we did not make much money was essentially that our cost base in Spain was too high relative to the business that we [End Page 130] would get out of it; we could not compete with the Spanish banks. The Spanish banks were telling us that they would give away the business in Spain to keep their customers. They had a tremendous home field advantage in that sense, and we did not have much to offer. It was very difficult to make money in Spain. From a business point of view, it is interesting that the articles reviewed said that Citibank was great, but this fact had no actual impact on the way we ran our business.

In contrast, in Japan, Citibank makes no money at all in local corporate banking, yet we have one of the most successful global businesses with regard to Japanese corporations. Japanese corporations are a big piece of our corporate income. We do business with them all over the world because we are the only bank that is active globally as well as in Japan. This means that corporations can receive service from their home. So we do not make much money in Japan, but we do make money all over the world. So judging that bank through the dimension of the local numbers is quite difficult and, I think, a stretch.

Another example is the cards business in Europe Citibank. All of that business is managed out of South Dakota. When you put your card in one of the machines in London, it is picked up in South Dakota, and it is run there. The advantages of doing that globally are tremendous for the bank, so the geographic dimension does not measure global businesses, and it is very hard to draw conclusions from it.

Regarding efficiencies in mergers, I would be very surprised if the evidence showed that there were not some efficiencies in mergers. I also would be really surprised if some chief executive officer went to his board of directors and said, "We are going to merge with Alpha Bank over there. And, by the way, we are going to spend more money." It just is not going to happen.

In almost any merger, you are going to get some assistance when you combine head offices. At this stage of the game and in this day and age, getting the efficiencies is just part of the game. It is like making your numbers. It is something that you have to do, but it is not why these mergers are happening. And I would disagree with the paper's conclusion that it is a driving force.

I think that the driving force in the banking business is fundamentally--and we use this term in Citibank--that corporate banking, meaning commercial and investment banking together, is a tricky business. The math problem is very simple. In order to attract capital in today's world capital [End Page 131] markets, you have to build 10 to 12 percent growth into your business, and you have to make somewhere between 15 and 20 percent return on equity. If you do not do that over time, you are going to have a problem. This is very difficult to do in a world that is growing on a real basis of about 3 percent. You try to run the numbers and figure out how you are going to do it. You have got to get the business someplace. You have to do something different. And that is the math problem that not only banks but all businesses face, especially on a global scale.

The industry has excess capacity. There is huge excess capacity in the banking business brought on by history of the way that banks are created, by all of the geographic barriers that are put on banks, and by a lot of reasons. There are basically too many bankers, and there is too much capital in the industry. Getting the kind of returns that you need in order to satisfy your owners and shareholders is very, very difficult in an industry with too much capacity.

The real field of study here is determining the cause of this excess capacity in the banking business and how it is going to play out in the future. The first issue is the size of the excess capacity. No bank has pricing power. It is almost impossible for a bank through price actions to generate anything on the top line. Because immediately the competitors come in and kill you. And that is true across almost every product in the business. There is no pricing power in the business. There is a lot of crazy competition, which is a sure indication that the business has excess capacity. When you think about it, charging fees for automated teller machines is nuts. It really is. It is antithetical to the .com world, where essentially what you want to do is attract customers. And you do not attract customers by charging them for something that they obviously do not believe is worth the money.

Mergers essentially are a reaction to this excess capacity of the business, which happens in every industry that has excess capital. People are looking for what I would call some sort of strategic accretion.

The mergers need to create some differentiating factor that will allow them to put up barriers to competition strategically, merging cross-border itself and finding businesses that are complementary. There was almost no overlap of businesses in the Citicorp-Travelers merger. And the major reason for that merger was to create complementarity, not efficiencies. The same was true of bankers in Deutsche Bank, who also were looking for strategic mergers rather than just efficiencies. [End Page 132]

What are some of the causes of this excess capacity of the industry? First and foremost is the technology itself. When I joined the bank, decisionmaking was done by telex and maybe by phone, but the phone was very expensive. Today, it is done on the Internet, and it is done in groups on the Internet; it is done instantaneously, and it is done asynchronously, and it is done asymptotically. You do not have to be in the same place, and you do not have to do it at the same time, but you can get it done almost instantaneously, if you have to.

The speed of the technology in a business that is fundamentally nothing more than an information arbitrage is creating the excess capacity of the industry. Most of the industry basically was involved in dealing with the information arbitrage. In banking the arbitrage concerns what is in the marketplace against what the customer's needs are and what the customer's needs are as a group against what is in the marketplace.

If the information becomes ubiquitous and free, there is no more value for the people who are using it to break the arbitrage. In the foreign exchange business, which is the heart and soul of Citibank's business globally, we used to know what was going on in the market, and the customer maybe knew what was going on in the market, and we could always make a spread between the two.

Now the customer knows absolutely as much as, if not more than, what the trader knows, and he can see the market by just turning on his computer. There is no spread in the effects business. The only way that you can make money in the effects business is to front run your customers or to position and gamble in the business.

What happens to all of those folks who were in the effects business and are no longer there? If you look at foreign exchange as a business alone, the number of competitors and the number of true players in the foreign exchange business have shrunk dramatically over the past fifteen years simply because the arbitrage is no longer there.

Every single business that the banks do is suffering from the same phenomenon, and this is creating the excess capacity. The second thing that has created the excess capacity are the turf issues that have all gone away or are about to go away. Certainly, doing away with Glass-Stegall broke down some more turf barriers. It was probably long after the horse was out of the barn, but certainly the turf issues are all gone. Couple that with the technology, and it does not matter where you are anymore. You can deliver services just about anywhere. [End Page 133]

Finally, efficiencies are going to become even less important in the future, while strategic interests are going to become much more important. For example, NationsBank, before it became Bank of America, paid, I believe, more than four times book value for Barnett Bank in Florida. Even assuming some good efficiency numbers, it is impossible to believe that more than four times book value makes sense in the world of 15 to 20 percent return on equity. It only makes sense if you say to yourself, "There is only one Barnett Bank in Florida, and there are only 1.5 million customers, or whatever the actual number of customers, and if I do not buy it, I do not have Florida. And if I do not have Florida, I have lost my strategic position." So the price does not matter because it is driven much more by the strategy of the business than by efficiencies.

I believe that the banking business is subject to Moore's Law. It is no longer more or less a steady-state operation; rather, we are looking at doubling capacity and cost in a relatively short period of time, and the modes of doing the arbitrage that we take as our bread and butter are changing very, very rapidly.

So speed--being the first to market--and investment up-front in the technology are going to determine the success of banks. What is going to drive the mergers is whether or not they can get those kinds of advantages.

General Discussion: Charles Calomiris suggested that some of the authors' findings were surprising, if not counterintuitive. The study showed that French banks operating in Spain are more efficient than Spanish banks. The study also showed that Dutch banks operating in France and Germany are less efficient than domestic French and German banks, but Dutch banks operating in the United States are not very different from domestic U.S. banks. Calomiris speculated that a reason for these oddities is that the authors were making "apples to oranges" comparisons when contrasting bank performance in different countries. An alternative comparison would match, for example, Fleet Bank (a major U.S. bank) that has no foreign operations with ABN Amro (a Dutch bank) operating in the United States. Edward Ettin added that, when making global comparisons, one should be cautious not to compare banks that have large wholesale banking operations with banks that concentrate on retail operations.

Berger and DeYoung both acknowledged the need for matched sampling in future research, although they noted that assembling the data would require a tremendous effort if it were conducted on a systematic [End Page 134] international basis. DeYoung added that although their study did not evaluate the organizational performance of banks on a worldwide basis due to limitations of the data, the study did address at least one organizational issue by examining the relative performance of U.S. banks operating in different regions.

A difference of views was expressed during the discussion on whether European banks had less or more incentive to engage in consolidations than their U.S. counterparts. Those arguing that they had fewer incentives pointed to the fact that European banks no longer need to be physically located in other European countries to operate and retain profit. Those taking the opposite position asserted that European banks could reduce their overall risks by diversifying into other countries. On this view, it was suggested that the authors in their future work use the consolidated results of banks rather than their profitability only in specific locations.