This past February, the U.S. Government Accountability Office (GAO) issued a report entitled “Bank Secrecy Act: Derisking along the Southwest Border Highlights Need for Regulators to Enhance Retrospective Reviews.”1 This report, which incorporates feedback from the Consumer Financial Protection Board (CFPB), Federal Reserve, Federal Deposit Insurance Corporation (FDIC), the Financial Crimes Enforcement Network (FinCEN) and the Office of the Comptroller of the Currency (OCC) presents data to support the finding that Bank Secrecy Act/anti-money laundering (BSA/AML) regulations have not kept up with legitimate risk management and cost of compliance concerns from the banking community. As a consequence, the current environment has led to a notable constriction of access to financial services in the counties along the U.S.-Mexico border.
The report, while it produces a modicum of statistical and anecdotal evidence of the challenges of accessing banking services in the border regions of California, Arizona, New Mexico and Texas, is inherently flawed due to factors presented below. Also, with the possible exception of the results of a web survey performed by the GAO, the data presented does not appear to be a statistically reliable—much less significant—sample. In addition, as pointed out by a number of the regulators, a number of relevant factors that could have reasonably been expected to impact the data were not incorporated into the analysis.
That being said, the report does profile real difficulties in accessing banking services that have reverberated through the economies of border communities.
The Method to the Madness
The GAO gathered information on banking account closures and restricted access to new account openings, as well as bank branch closures. This data was gathered for the 33 counties, which have 25 percent of their area within 50 miles of the U.S.-Mexico border. Statistical information was also gathered for comparable counties outside the region within the same four states, in high-risk areas elsewhere in the country and more broadly nationally.
Bank and Community Information
The GAO conducted a national web survey of 406 banks, including 115 from the defined southwest border region for the three-year period of 2014 through 2016. The survey, which had a 46.5 percent response rate, gathered information on account closures and limitations on new account openings related to anti-money laundering (AML) risk, the types of impacted customers, and the reasons for the closures and limitations.
In addition, three site visits were made where five discussion groups included representatives from local banks, as well as from the broader community. The three communities, Nogales, Arizona (approximate population: 20,000), San Ysidro, California (approximate population: 28,000) and McAllen, Texas (approximate population: 142,000) were chosen to provide a mix of urban and rural communities, different rates of branch closures, and different designations as distressed or underserved communities. Two of the discussion groups focused on retail banking customers, while the rest centered on business banking customers.
Moreover, the GAO interviewed four of the five largest border banks. Fifteen others were interviewed and selected to provide a mix of primary regulator, number of branches and asset size. Interviews were also conducted with economic development specialists, chambers of commerce, and industry and trade groups.
The GAO also reviewed a spectrum of BSA/AML-related information. Bank-related data included regulatory filings of suspicious activity reports (SARs) and currency transaction reports (CTRs), enforcement actions and issues flagged during bank examinations.
In addition, documents and guidance issued by federal regulators focused on de-risking were reviewed, as were retrospective regulatory reviews, and executive orders that either proscribed or recommended such reviews.
Lastly, officials from FinCEN and bank regulators were interviewed about de-risking and retrospective regulatory reviews performed for BSA/AML regulations.
The Hard (And Not so Hard) Data
The most prominent statistical number in the GAO report is stark. Banks in the southwestern border region filed three times more SARs for each billion dollars in deposits than the national average. The overwhelming majority of banks in the region (80 percent) claimed to have terminated accounts due to BSA/AML risk, and limited or did not offer accounts to high-risk customers because of related regulatory oversight.
Background: Mexico/U.S. Border Region
Mexico is the U.S.’ second-largest trading partner for goods, both in exports and imports. Key industries in the border region include fresh produce and manufacturing. This includes “production sharing,” an arrangement where U.S. companies keep production costs down by locating some of their operations in Mexico. Border tourism is also a major economic force for border counties. In 2017 alone, three-quarters of a million pedestrians entered the U.S. at the San Ysidro border crossing, spending much of their time visiting and shopping in border communities.
The U.S. State Department considers Mexico to be a major money laundering country. The border area attracts criminal organizations for cross-border money laundering, narcotics trafficking and human smuggling, according to the U.S.’ 2015 National Money Laundering Risk Assessment (NMLRA). That report also states that bulk cash smuggling is the main way drug proceeds are transported across the border. Not surprisingly, all the counties in the border regions are either High Intensity Drug Trafficking Areas (HIDTA) or High Intensity Financial Crime Areas (HIFCA), with a large majority carrying both designations.
The report notes that in the border region, high volumes of cash transactions, as well as the presence of cross-border transactions and foreign account holders contribute to the assessment of the area as higher risk for money laundering. While bank employees interviewed by the GAO say that these risks are managed by more frequent monitoring and investigation activities, they also note that these efforts require a higher investment of resources.
Adding to the AML compliance environment in the border region are the June 2010 regulations issued by the Mexico Finance Ministry. These changes put a cap on the amount of U.S. dollars that could be held by Mexican banks. Both bank examiners and FinCEN acknowledge that this altered the risk profiles of affected banks, most notably in the southwest U.S. The report explains that while (prior to the change in regulations) banks used to get cash from a limited base of Mexican banks, after the restrictions were put in place, firms had to contend with a much larger universe of individuals and companies. That change raises both the risk and the cost of compliance for the same level of business. In addition, FinCEN notes that businesses were getting more cash payments from Mexican customers, and consequently were depositing more cash (and creating elevated risk) for the banks they used.
This trend is validated through the GAO’s review of CTR data: in 2016, southwest border banks filed 30 percent more CTRs than comparable in-state branches, and 60 percent more than high-risk counties outside the region.
The elevated AML risk in the region is also represented by the number of SARs filed. Banks in border counties filed three times as many SARs per billion dollars in deposits as comparable in-state counties, and two and a half times as many as in other HIDTA or HIFCA counties.
Account Closures and Limitations
Most border banks claimed to have terminated accounts related to BSA/AML risk
Most border banks claimed to have terminated accounts related to BSA/AML risk. Similarly, they also did not offer accounts or they limited the number of new accounts to certain customer types consistent with their BSA/AML programs. However, the information does not provide needed details, such as raw numbers of accounts affected, or comparisons of these restricted or closed accounts compared to those affected for other reasons.
GAO’s analysis suggests that counties with a younger average age, those that are more urban in nature, those with a higher per-capita income and those with higher AML risk were more likely to lose branches. In addition, AML risks were noted as being higher in the southwest border region.
Account terminations exhibited some notable variations. While 80 percent of southwest border banks terminated accounts due to AML risk, only 60 percent of banks who also had business outside the region did. In addition, while 95 percent of large banks and 93 percent of medium banks had such closures, only 26 percent of small banks did. The report does not propose theories as to why this happened.
There were four broad categories of business accounts that were closed due to AML risk by large percentages of banks in the border region. In addition to the 70 percent of banks who closed the accounts of cash-intensive businesses, and the 58 percent who closed money services business (MSB) accounts, almost half of the banks also closed the accounts of foreign businesses, regardless of whether or not they were involved in cross-border trade.
However, the numbers of actual closures are not uniformly distributed. While 15 banks on the southwest border closed 5,396 personal accounts, and 16 banks reported 901 business accounts, one “extra-large” bank was responsible for over 80 percent (4,402) of the closed personal accounts, and slightly over half (457) of the business accounts. However, the report notes that these closures amount to less than one-half of one percent of the large bank’s accounts. Similar ratios between the overall numbers and the closures by another extra-large institution exist outside the border region as well.
There appears to be a strong correlation between SAR filings and account closures. Ninety-three percent of banks on the border claim to terminate accounts due to SAR filings, although the data does not show what percentage of the closures this represents. However, three of the 19 banks interviewed required accounts to be closed based on a specific number of SAR filings, while two others claimed that the number of SAR filings caused an account review. And while yet another bank noted that SARs are one of the factors that can cause an account to be terminated, taken together, this still represents less than one-third of the surveyed institutions.
Other reasons for closing accounts that were noted in the report included customer failure to supply requested customer due diligence (CDD) information (mentioned by 80 percent of respondents), and the reputational risk associated with certain business types such as those associated with the gambling and marijuana industries (mentioned by 68 percent). While not discussed in detail, a table in the report notes that increased regulatory oversight led 63 percent of banks to close accounts, almost half cited an inability to manage the BSA/AML risks, almost 40 percent ascribed the closures to compliance costs and slightly over a quarter quoted concerns about personal liability.
Many of these factors are also noted to contribute to limitations on, or denials of, new account openings. Approximately three-quarters of banks reported such limitations or prohibitions applied to MSBs as well as all types of foreign businesses.
The report notes that failure to provide CDD information and SAR filings are the primary reasons cited for account closures outside the border region as well.
One example given for account closures points to special challenges in the border region. One bank previously offered accounts to used clothing wholesalers who exported the goods to Mexico. Mexican nationals would cross the border to buy pallets of clothes with cash, but the bank was unable to identify the source so it no longer services such accounts.
Branch closures in the border were not uniform; 18 counties had no closures during the report period, while five counties lost 10 percent or more of their branches. This is reflected in the three communities that the report’s authors visited: McAllen lost four of 63 branches, Nogales lost a full third of its nine branches, and San Ysidro fared the worst, losing five of 12 branches.
The GAO’s econometric model showed a number of interesting variations in the data:
- Urban counties were 22 percent more likely to lose branches than the most rural counties
- Counties where 70 percent of the population is younger than 45 were nine percent more likely to have closures than ones in which only one-half the population is as young
- Counties where the per capita income is $50,000 were seven percent more likely to see branch closures than ones where that figure is $20,000
- HIDTA counties are 11 percent more likely to have branch closures
- Counties where 200 SARs per billion are filed are 8 percent more likely to experience branch closures than ones that have no SAR filings
According to the GAO, many of the southwest border region’s demographics are not as stark as the extremes noted above. It is roughly as urban as the national average, has a slightly lower average per capita income and slightly younger residents than the national average (60 percent vs. 55 percent). However, the report says that the border region has relatively more AML risk factors, which is reflected in the number of HIDTA counties, and the high rate of SAR filings. In fact, counties which lost branches had 10 times as many SAR filings as those which did not (600 vs. 60). However, given the relatively weak correlation between elevated AML risk factors and branch closures, there is reason to doubt that this is the cause of the closures in the border region.
Feedback from banks validates this. Six of 10 banks said AML compliance was not part of the decision to close branches, although four did say that costs (including compliance costs) could be, and one claimed that closing a branch was one way to deal with significant compliance challenges. Most tied closures to the financial performance of the branch, while three also mentioned that the volume of customer traffic and the adoption of mobile banking were also considered in the decision-making process.
Enforcement Actions and Regulatory Oversight
From 2009 through June 2016, the southwest border counties were subject to 41 enforcement actions from regulators (in addition to two actions from FinCEN) representing 229 violations (33 percent of which were due to SAR issues, and 31 percent due to issues with controls and training). In contrast, nationally there were 576 total enforcement actions, and approximately 9,000 violations. That implies that one in 14 enforcement actions take place in the border region, and about one in 40 violations occur there.
Comments from the interviews revealed that banks tailor the nature of their businesses to ensure the avoidance of regulatory criticism. They believe that the business benefits of maintaining high-risk accounts are outweighed by the potential impact of negative comments from examiners and/or enforcement actions. Having a good relationship with regulators is important to the banks, and one person noted that they felt pressured to close accounts based on the concerns of bank examiners. Several of the interviewees stated that law enforcement and regulatory actions have made them more conservative in their approach. There was anecdotal evidence of defensive SAR filing, and one respondent claimed that his/her institution was not servicing a specific area because they were afraid that regulatory penalties could be large enough to cause the firm to go out of business. It should be noted that some of these reasons are also noted outside the border region.
This is all in line with a speech given by a Treasury representative in 2015, which noted that banks have raised concerns on the cost of compliance, as well as uncertainty about regulatory expectations and the level of potential penalties. That uncertainty leads to assuming the worst case, as the report claims that accounts are being closed not because firms are unable to manage the risks, but because the perceived costs of taking on those risks is too high. Banks perceive that the lack of supervisory and enforcement transparency increases their estimates of anticipated costs, which causes certain businesses to be considered unprofitable.
Much of the data comes from commentary from local representatives, the discussion groups and interviews. These comments provide valuable insight on unique challenges while operating in the border region:
- A regional trade group stated that border businesses prefer cash because of potential exchange rate changes while a peso-denominated check clears
- Produce industry association representatives noted that U.S. produce distributors import from Mexico and pay with a funds transfer. The exporter then withdraws the funds in cash to pay workers. This raises suspicion of money laundering because of the timing of the withdrawal on the heels of the electronic transfer
- Some banks send staff to Mexico to establish legitimacy of businesses attempting to open accounts, while another reviews three months of bank statements to determine normal volumes of business
- Other contributors to compliance costs include translation services, and development of internal expertise on foreign identity documents
- Representatives from Southwest Border AML Alliance stated they believed some of the closed accounts were due to information from law enforcement or government agencies. The report gives the example of government guidance on funnel accounts, which led to funnel account closures.
Effects of Account and Branch Limits
The report contains a sizable amount of information about the effects of border area banks’ account closures and restrictions, and branch closures:
A 2013 study notes that cross-border produce trade accounted for one quarter of jobs and wages in Nogales
- A 2013 study notes that cross-border produce trade accounted for one quarter of jobs and wages in Nogales
- A business owner claimed that the volume of funds deposited from an affiliated Mexican firm caused too much risk and led their bank to close their account. It then took seven months to get a new account opened, which required coordination with banks on both sides of the border
- Economic development specialists claim that Mexican nationals spent one billion dollars in Pima County each year, and that 70 percent of sales tax is paid by Mexican customers crossing the border to shop. However, they make fewer trips now, because accounts in U.S. banks that they use to withdraw funds for their shopping trips are harder to maintain. This, in turn, affects the economies of border communities. Similarly, branch closures lead to less foot traffic that depresses sales volumes, as evidenced by changes in the fortunes of businesses along San Ysidro Boulevard in the wake of branch closures
- Similarly, lost branches reduce means of borrowing, which limit investments in local communities. For example, Tucson businesses have turned to alternative funding sources, including loans from family members, title loans and accounts receivable lending. This is also validated by academic research that finds that closures lead to reduced small business lending as well as employment growth in the immediate area
- Branch closures cause people to travel further to conduct banking activities, to pay higher fees for using alternatives for their financial services, and increased difficulties completing their transactions. People in Nogales and San Ysidro have to travel between 20 and 40 minutes to the nearest bank branch, although one person noted that they travelled over 70 miles to another branch of their bank because they were afraid they would not be able to open an account at another bank
The regulators that were consulted for the report provided feedback on the initial draft of the report, and their feedback was incorporated in, among other places, a separate section of the report. The feedback largely focused on two areas: the steps being taken to address the issues, and pointing out flaws in the report and its methodology.
For example, the FDIC mandated that examiners document the instances where they recommend account closures. However, as of December 2017, no closures had been recommended.
Multiple statements by the Financial Action Task Force (FATF), the OCC and FinCEN have addressed the need to balance AML compliance with the need for financial inclusion and access. However, the GAO responded by saying that the actions taken by regulators do not consider all factors and, the regulatory reviews that they perform are not broad enough.
FinCEN did not get their response to the GAO in time for it to be included in the final report. However, it responded in a separate memorandum.
FinCEN deems the problem of de-risking to be important, and claims to have been actively addressing it by clarifying what drives it and how it can be addressed.
They also participate in multiple governmental initiatives. These include the Multi-State MSB Examination Taskforce, the Conference of State Banking Supervisors, the Federal Financial Institutions Examination Council, the Financial Stability Board’s Correspondent Banking Coordination Group and the promulgation of the 2014 Money Remittances Improvement Act. De-risking is also a major focus of the BSA Advisory Group.
FinCEN believes that restricted access to financial services is due to a misunderstanding by firms of their compliance responsibilities.
FinCEN also had a number of criticisms of the GAO report and its methodology. Perhaps the most fundamental is that GAO’s definition of de-risking is not the standard one. A consequence of this is that the report does not differentiate restrictions imposed based on individualized assessment of account risk, as is the recommended norm, from restrictions imposed due to de-risking, which is the uniform treatment of entire classes of customers with common characteristics. By doing so, the report implies that the compliance requirements themselves are the cause of all perceived account restrictions.
Similarly, the report does not define when regulatory concerns justify account closures, nor does it explain how the GAO determines when regulatory oversight is proper.
FinCEN also notes that the higher levels of SAR and CTR filings may be partially due to the Mexican currency control restrictions, as well as financial activity performed as part of criminal activity. As the report does not attempt to quantify these effects, it is impossible to gauge what portion of the filings are defensive, and which are justifiable on the merits. Similarly, the memo notes that the report fails to leverage law enforcement data or their perspectives on financial crime in the border region, as well as more broadly.
In fact, the memorandum also notes that there is an ongoing trend of branch closures in less densely populated areas that is consistent with the higher rates of branch closures in Nogales and San Ysidro as opposed to that in McAllen. FinCEN notes that if branches were being closed due to BSA/AML concerns, the number of CTR and SAR filings would have decreased—but they have not.
At a basic level, FinCEN concludes that the GAO analysis does not show that the loss of access to financial services in border regions has occurred at a higher rate than elsewhere.
Is the truth in the middle?
If one considers the identifiable trends in the data, both quantitative and qualitative, some things are clear. Whether it meets the standard definition of de-risking or not, there is an identifiable impression of higher standards for opening or maintaining an account that appears to be linked to anti-money laundering/counter-terrorist financing program requirements, and the fear of receiving adverse BSA/AML-related comments in a regulatory examination. The change in the Mexican currency controls, and the attendant shift from receiving cash from a relatively small number of known financial institutions to dealings with larger sets of individual consumers who are more difficult and more expensive to perform due diligence on, amplifies those concerns in the southwest border region.
One thing should be clear: creating financial services “deserts” is not only not in consumers’ interests, nor that of the communities in which they live, it is counterproductive to the goals of regulators and law enforcement officials who are trying to identify, trace and stop financial crime. When traditional financial services firms are not readily accessible, money launderers do not stop their illicit activities. They either find outlets in other communities, merely shift the risk geographically or they find alternative means of moving money, at least some of which will not provide the same level of traceability and oversight as traditional service providers.
How might the higher level of BSA/AML-related closures and restrictions be more effectively addressed? It might help, as noted in the Treasury speech as well as comments from the public noted earlier, to have greater clarity on regulatory examinations and the standards needed to be applied. In addition, perhaps the standards should not only provide a sense of what is considered too lax, but also what is considered to be overly restrictive, absent extraordinary circumstances.
Similarly, if FinCEN and bank regulators issued clear, yet flexible, enforcement guidelines—with clearly spelled-out gradations of response and factors considered in the decision-making process—banks along the southwest border (as well as elsewhere in the country) would have less uncertainty about whether or not the efforts they are undertaking will be deemed adequate and, if they are considered to be inappropriate, how that is reflected and managed by regulators.
In a more general sense, maybe what is needed is a paradigm shift for regulation and oversight: from mandating and ensuring that firms have an “effective” BSA/AML program to requiring one that is truly “appropriate” for the individual institution. Such a change would likely slow the rate of unnecessary BSA/AML compliance-related denial of services, better rationalize the risk tolerances and costs of compliance for regulated firms, and keep more criminal assets in traditional financial services firms, where they can be identified by banks, and subsequently interdicted by the authorities.
- “Bank Secrecy Act: Derisking along the Southwest Border Highlights Need for Regulators to Enhance Retrospective Reviews,” United States Government Accountability Office, February 2017, https://www.gao.gov/assets/700/690310.pdf