DEPARTMENT OF JUSTICE
I appreciate the opportunity the ABA's Antitrust Section has given me today to set the record straight about how the Department's Antitrust Division assesses the competitive impact of proposed bank mergers. I am aware that there has been some uncertainty about our antitrust evaluation process. In addition, there have been concerns about our alleged differences with the banking authorities, with whom we share statutory authority for reviewing the antitrust implications of bank mergers.
I hope to demonstrate today that what we do is not that mysterious and that, in fact, banks and their clients can and should be able to predict with reasonable precision the outcome of our assessments.
Putting Bank Mergers In Context
Before outlining how we evaluate bank mergers, it is important to put the subject in its proper context.
Over the past several years, the banking industry probably has seen more mergers than any other industry in the economy. In large part this is because of the fragmented way in which the law has required banks to operate: confining them to their home states, and in some cases, to their home counties or cities. Accordingly, denied the right to expand internally, banking organizations generally have been able to expand into different geographic markets -- in particular, across state lines -- only by acquiring banks in those markets. The result is that the Division has been required over the past several years to review approximately 2000 bank merger or acquisition applications each year.
In many cases, the banks involved in these mergers have not competed with each other and thus have not posed antitrust risks. In the overwhelming majority of the others where the banks have competed against each other, it was clear that the market would continue to be competitive even after the merger. As a result, over the last five years we have undertaken a full investigation with respect to only 43 of the approximately 9000 applications filed with the banking authorities, or about 0.5 percent of all proposed transactions. Of the 43 mergers investigated, we have challenged only four, resulting in divestitures in each case of those branches that created competitive problems.
If, as appears increasingly likely, Congress enacts interstate branching legislation this year, two contradictory trends will be unleashed.
On the one hand, with branching restrictions removed, many regional and potential national banks will be seeking to widen their geographic coverage. With perhaps rare exceptions (where the geographic markets of the two banks may overlap to some degree), these market-extension mergers are not likely to pose antitrust dangers.
On the other hand, it is possible that interstate branching will encourage a larger number of smaller banks competing in the same markets to merge in order to realize sufficient economies of scale to compete against the regional and national giants. Almost by definition these within-market mergers are more likely to pose antitrust risks than market-extension mergers.
I do not know at this point whether the mix of mergers -- between within-market and market-extension -- will change with the arrival of interstate branching and so I cannot predict whether the Division will be more or less active than it has in the past in challenging bank mergers. But I can assure you that we will remain vigilant to ensure that the process of consolidation in the banking industry does not hurt consumers by significantly weakening competition.
Screening Bank Merger Applications
The Division uses the same standards to assess the competitive impacts of all mergers, whether or not they involve banks. Those standards are found in our Merger Guidelines, the most recent edition having been issued in 1992.
In brief, under those Guidelines we first define relevant product and geographic markets, using the construct of a sustained 5 percent price increase to determine whether consumers would turn to alternative suppliers.
We then measure industry concentration in the relevant market based on the Herfindahl-Hirschman Index (which sums the squares of the market shares of the competitors in the market). The higher the initial level of concentration and the change in concentration that the merger will produce, the more likely we are to challenge it.
The competitive inquiry, however, is not simply an arithmetic exercise. We don't simply plug some HHI numbers into the computer and then have a program tell us whether or not to sue. Instead, we use the HHI calculations only as a screen, or starting point, in our analysis. Given the large numbers of merger applications we receive each year and our relatively limited resources, the HHI-based screen is useful because it can help point to those merger proposals we should examine more intensively. Thereafter, at the examination stage, we will take account of a series of less quantifiable factors that can make a challenge more or less likely, including: the ease of entry, the pace at which the size of the market may be expanding or contracting, and whether the parties have demonstrated that the merger will yield efficiencies sufficient to offset any anticompetitive effects.
We use an HHI-based screening procedure for banks as well. In a significant respect, it is easier for us to screen bank mergers than other types of mergers because of the ready availability of data on deposits held by banks (and assets or loans of other financial institutions whose activities may put them in the relevant market).
It is important to understand, however, that the criteria we use for screening the many bank mergers we see each year are simply rough rules of thumb, subject to several caveats I will mention shortly. Nevertheless, these rules of thumb should be useful starting points for any bank or other depository institution contemplating a merger.
In brief, we have used a screen of 1800/200/20 over the past several years. That is, in most cases we will not conduct a full investigation unless in every market:
--The post-merger HHI is at least 1800
--The merger must produce a change in the HHI of at least 200
--In the case of bank mergers, concentration in the relevant product market is computed for screening purposes on the basis of all bank deposits plus 20 percent of all thrift deposits in the relevant geographic market.
We make these calculations in most cases using the geographic markets defined by the individual Federal Reserve Banks in the districts where the merging banks do business. Later I will explain why we do not always accept the Fed's pre-defined markets.
We are frequently asked why, at the screening stage, we give thrift deposits a smaller weight than is accorded by the Federal Reserve Board (50 percent). The answer stems from the fact that we use a different approach to product market definition (although as I discuss later, this difference in approach rarely generates a different assessment of the competitive impact of a merger). Whereas the Fed looks at deposit data as an adequate proxy for the "cluster" of services that banks often provide (loans, deposits, and various fee-based services), we at DOJ have long treated banks as multi-product firms and, accordingly, have assessed the competitive impact of mergers in each relevant product or service (deposits, various types of loans, and so on).
As it turns out, although thrifts compete with banks on an equal basis in many services, that is not true with respect to one very important product line -- commercial loans, especially to small and mid-sized businesses. In significant part, this is because thrifts are restricted by law from investing more than 10 percent of their assets in commercial loans. But it is also because, in practice, thrift involvement in commercial lending has been well below the 10 percent threshold. So although we are comfortable at the screening stage using deposits to act as a proxy for various product markets, we have discounted thrift deposits substantially in order to reflect the realities of thrift lending to commercial borrowers in most geographic markets.
In addition, I should explain why at the screening stage we use a looser HHI test (1800/200) than is suggested by the Guidelines (1800/50) as a threshold indicating a likely challenge. The basic reason is that banks face competition in virtually all of their services from non-banks, as well as from out-of-state banks, that often cannot be captured by computing HHI's based solely on deposits. We have recognized the strength of that competition generally by screening out mergers causing changes in the HHI up to 200 even where the post-merger HHI in the market is 1800 or higher.
Nevertheless, as I cautioned just a few moments ago, the 1800/200/20 screen is just a rule of thumb. There are certain exceptions.
One obvious exception is that we will go beyond the simple screen and conduct some investigation of mergers involving large financial institutions, such as the marriage of Chemical and Manufacturers Hanover or NCNB and C&S Sovran. Given the operational complexity of such institutions and the many specialized service and geographic markets in which they operate, it is essential that we conduct a more intensive examination to ensure that they do not have an anticompetitive effect (and that if they do, appropriate divestitures can be arranged to eliminate such effects in the geographic markets where the two may overlap).
A second caveat is that we will often look past the simple HHI screen where mergers involve the top-ranking institutions in the same geographic market, such as the first and second ranking banks, the first and third, or in some cases, the second and the third. In such cases, a merger may give the merged entity market power vis-a-vis certain customers that would not require collusion, the anticompetitive outcome that the HHI analysis seeks to identify and prevent.
A third caveat is that we are reexamining the 20 percent weight given to thrift deposits for screening purposes in light of recent trends indicating that thrifts are moving away from commercial lending and going back to their roots in residential mortgage lending. Of course, once we reach the investigational stage, we will continue our practice of examining the competitive strength of each thrift in any particular product market. In some areas of the country, such as New England where thrifts have been active commercial lenders, this will often mean that thrift deposits will be given a weight larger than 20 percent. In other areas of the country where thrifts have not been active, thrifts may be excluded entirely from the business loan market. Bank Merger Investigations
As I suggested earlier, full-fledged bank merger investigations are relatively rare -- something we do in less than one percent of the transactions we see. But for those mergers that reach this stage, we conduct the same kind of analysis under the Merger Guidelines that we would do for any transaction.
Thus, we typically will ask the institutions to supply documents relating to the way they see the markets and the impacts of the merger on those markets. We will interview competitors and purchasers. We may engage economic experts. And we may employ compulsory process to obtain documents and to take depositions of key witnesses, including employees and managers of the merging institutions. It is entirely possible that after all this analysis is completed we will challenge a merger with HHI figures different from those used as a rule-of-thumb in our screen.
In particular, we may challenge where the change in the HHI is less than 200 (but more than 50) in markets where the post- merger HHI is over 1800, for several reasons. The parties to the merger may be of sufficient size or market standing that we will examine the merger regardless of the HHI screen. Or the parties involved in the merger may be competing head-to-head in a number of markets, one or two of which may cause the merger to "flunk" the initial screen, and where on further investigation, we determine that the anticompetitive effects in one of the other markets where the change in the HHI may be less than 200 are significant.
We are likely to pursue several lines of inquiry once we investigate particular bank mergers.
First, we will examine and refine our definitions of the appropriate product and geographic markets, taking our cues from the specific services the parties offer and where they offer them. Thus, in the typical bank merger case, we will look separately at the markets for deposits (commercial and retail), various types of loans (mortgages, consumer, and commercial), and any other services the parties may offer (trust, cash management, correspondent banking services, etc.). We tend to look especially hard at the type of commercial lending in which the parties are engaged. If, for example, they concentrate their attention on small business borrowers (for example, with loans no more than $1 million) or on mid-size borrowers (with larger loan limits) then those are the markets we will look at. I should point out that once we undertake a full investigation and define the relevant markets with greater precision the HHI figures we then calculate frequently will differ from those we use at the screening stage.
Similarly, we will closely examine the relevant geographic markets. Although we will endeavor to use the Fed's market definitions, in some cases those definitions may not accurately reflect the nature of competition for a particular service. This is especially likely to be true where the Fed's markets are drawn quite broadly, but the particular services the parties offer actually are bought by purchasers in a smaller region -- for example, the inner city rather than a wider metropolitan area. Moreover, it is important to keep in mind that we may define different geographic markets for each of the different product markets. Our touchstone is where customers for specific services are willing to turn, not some arbitrary geographic area that may be developed for some other purpose.
If in any particular case you have questions about how the Division will define the relevant markets, we urge you to contact the staff. A timely visit before a transaction is consummated can save you and your clients a lot of time and money.2 In particular, where competitive problems may be apparent, you may be able to restructure the transaction ahead of time by arranging for divestitures of certain branches where the problems show up.
Second, we will consider whether there has been a history of de novo entry in a particular market (by creating a new institution and not by merger), or whether the market itself is growing rapidly and thus likely to attract entry if current participants begin to earn supranormal profits (or profits above what one would expect given the risks involved).
Third, and conversely, we will examine whether the current deposit data in any way overstate the competitive strength of particular institutions that have experienced or are likely to experience a significant diminution of their past business success. More generally, we will examine whether the current deposit data for all the institutions in the market accurately indicate their competitive strength.
Fourth, we will assess the strength of all non-bank competition, including individual thrifts and such non-banks as finance companies.
Finally, if we receive customer complaints in a market we will probe more deeply to ensure that the merger won't result in competitive harm. Answers To The Critics
We at Justice do not live in an ivory tower. We are aware that the way we analyze mergers has been criticized in some quarters. I would like to conclude by addressing some of the more frequently-heard concerns.
First, some have criticized us for not taking sufficient account of non-bank competition, especially by finance companies (which have taken market share from banks nationwide in both consumer and business lending) and credit unions (which have done the same in deposits).
Several responses are in order. Where finance companies do effectively compete with banks -- for example, in providing automobile loans -- we will include them in the relevant product market and count their participation, data permitting. The same is true for credit union deposits.
Consumer lending concentration, however, rarely poses a competitive problem. Instead, to the extent that bank mergers raise competitive problems they most often have been manifested in business lending. And it is in this context where some have argued that finance company competition should be factored in.
Nevertheless, it is our experience -- based on interviews and document productions over the past several years -- that finance companies generally do not make the kinds of working capital loans to small and mid-size business borrowers that are the bread and butter of many banks. While finance companies have gained market share from banks at the national level for certain types of loans, primarily in asset-backed loans (equipment financing, for example), we are charged under the law with investigating the competitive impacts of particular mergers in specific markets. And, at the level at which do our analysis, we have not yet found finance companies to offer meaningful competition for certain types of business lending for which asset- backed financing is not a ready substitute.
Second, some concerns have been expressed about the fact that we will define the nature of the business lending market differently for different transactions and that this causes uncertainty for those contemplating mergers. Thus, for example, we may define the commercial lending in one case to be loans up to $1 million, while in another case we may use a $5 million cutoff.
The short response to this concern is that we analyze the markets we have before us. If the parties do not lend to larger business borrowers, then we will not analyze that market. Conversely, if the parties lend only or primarily to borrowers under a given size, then we will look to that particular market. This shouldn't be a mystery to counsel who can simply ask their bank clients how they categorize their loans themselves.
The way we analyze markets should cause no more uncertainty than our procedure in all other types of mergers, which often present unique product and geographic markets. Merger analysis is fact-sensitive in banking as elsewhere. The analytical principles and goals are the same, but the answers they produce depend on the facts of specific cases.
Third, some have questioned why we, in some cases, have not used the Fed's pre-defined markets. I hope I have already explained why: we will depart from the Fed's regions where the market realities suggest we should. Moreover, unlike the Fed, which views banks as providing a cluster of services, we view banks as multi-product firms. As a result, we sometimes have to use different geographic market definitions for each of the products we examine.
Finally, some have suggested that our methodological differences with the Fed have caused needless uncertainty in the banking community. I do not believe this to be true. Although DOJ and the Fed may analyze mergers differently, this difference in methodology only generates a difference in result in very few cases. As I noted at the outset, only about a half a percent of all cases over the past several years have reached the investigational stage, and in only a fraction of those have we differed from the Fed.
Nevertheless, the public -- and the banking community in particular -- has a right to expect as little uncertainty from the government agencies that enforce the law as is reasonably possible. In that spirit, I will close by pointing to several things that will hopefully narrow the uncertainty that may be out there.
For one thing, we have been consulting with the Fed Board and staff about the standards and procedures used to assess bank mergers. I think we have clarified our differences and ironed out any procedural and administrative frictions. We are conducting a similar dialogue with relevant staff at the Comptroller of the Currency.
In addition, I hope I have made clear the difference between the standards we use to screen mergers and what standards we apply to those mergers we investigate.
How can those of you who believe that non-banks are viable competitors to banks in small and medium sized business lending convince us this is the case? One thing you might give us is lost business reports from bank loan officers prepared in the routine course of business (rather than those assembled after the fact solely for purposes of the merger filing). To the extent those reports show banks frequently losing business loan customers to thrifts or other non-bank lenders, we might be persuaded in particular cases to give those competitors greater weight for that type of customer than we may otherwise.
At the same time, I can relay to you the experience of our staff that highway traffic counts generally have not been helpful in documenting the scope of geographic markets, and bankruptcy filings have not been that helpful in proving the scope of bank product markets.
If you have any doubts in a particular matter about what information the Division might or might not consider useful in carrying out its responsibilities -- whether in defining the relevant market or in weighing other factors relating to the strength of competition in the market -- then come in to see the staff. Although some of you may not believe it, such visits make the merger analysis process easier, less time consuming and less expensive than trying to outguess or outmaneuver the Division. If a merger has competitive problems, we'll eventually find them. Equally important, if your merger does not pose a competitive concern -- but you don't know whether we'll come to that conclusion ourselves -- then it is your job to present us with the evidence that leads to that conclusion. You can save yourselves time and money by asking us in advance what type of evidence we would consider most useful.
I hope in this short time I have clarified the way we do business so that you can conduct your own business affairs with greater assurance.
1 This figure includes acquisitions by banking organizations of (permitted) non-banking operations, as well as bank assets or branches, for which filings are required by law.
2 We have recently transferred review of bank mergers from our Communications and Finance Section to our Litigation 1 section, headed by Anthony Nanni. Mr. Nanni and his staff are the appropriate contacts for such visits.