Re: Renewal without Change of the Bank Secrecy Act Reports of Transactions in Currency, Docket Number FINCEN–2020–0003 & OMB Control Numbers 1506–0004, 1506–0005, and 1506–0064
Dear Sir or Madam:
The American Bankers Association (ABA) appreciates the opportunity to comment on the Currency Transaction Report (CTR) proposal that the Financial Crimes Enforcement Network (FinCEN) published in the Federal Register on May 14. FinCEN is reviewing the Currency Transaction Report (CTR) form to comply with the requirements of the Paperwork Reduction Act (PRA) of 1995. While no changes are being proposed to the information collected, FinCEN is updating the burden estimate associated with collecting the information.
ABA appreciates FinCEN’s effort to more accurately estimate the costs and burdens associated with CTR filings. It is well established that year after year Bank Secrecy Act (BSA) and anti-money laundering (AML) compliance is cited as financial institutions’ top regulatory expense, and annual CTR filings contribute significantly to those costs. We believe having an accurate estimate of CTR filing burden is critically important to policy discussions seeking to improve the effectiveness of the BSA/AML regime. For many years, there have been suggestions that the threshold for CTR filings be increased from the $10,000 threshold set 50 years ago.4 FinCEN has resisted these recommendations, but without good statistics on the use of CTRs and CTR data, it is impossible to have a thorough discussion about the CTR threshold or other steps to address the CTR burden.
In addition, ABA believes it is important to provide the public with data on aggregate CTR filings. The information should be consistently available, and regularly updated, on the FinCEN website—not just in connection with a PRA notice. Federal regulators, including FinCEN, regularly emphasize the importance of transparency for improving the effectiveness of BSA compliance. For example, Congress is currently debating measures to improve the transparency of the ownership and control of legal entities.5 However, there is a clear lack of transparency regarding CTR filings. FinCEN does not publish data on CTR filings on its website or make it available to the public. In fact, last fall, ABA was compelled to file a Freedom of Information Act request to obtain limited data on CTR filings. Making the data available will contribute to efforts to improve the efficiency and effectiveness of BSA/AML compliance and supporting national security.
FinCEN’s revised estimate of the annual cost to file CTRs is approximately $54,676,044, a figure ABA believes significantly underestimates the burden imposed for the following reasons:
Moreover, since this is a Paperwork Reduction Act exercise, one of the elements that is critically missing from the current analysis is a counter-balancing estimate of the benefits of the information collection. Law enforcement representatives regularly testify about the value of CTR information when investigating and prosecuting cases, but there has been no attempt to quantify that value in the proposal. ABA believes it is important to attempt to quantify the value of CTR data to assess whether the cost of CTR filings are justified by their benefit to law enforcement and national security.
According to FinCEN, 14,276 filers submitted 16,087,182 CTRs in 2019, the last full year for which data is available.6 FinCEN believes that the first step to more accurately analyze CTR burden is to separate filers by filing volume to analyze burden, a step that ABA supports. To analyze the differences among filers, FinCEN has separated filers into twelve tranches which range from the highest volume filers that submit more than 2000 CTRs weekly to the smallest volume filers that submit 6 or fewer CTRs annually. Looking at the largest volume filers (the top two tranches), FinCEN found that the top 306 filers, which filed more than 100 CTRs each week, submitted over three-quarters of all CTRs filed in 2019. Among all filers, depository institutions represented 2/3 of the total, submitting 88% of all CTRs in 2019.
Previously, FinCEN’s burden estimate assessed time and costs associated with the collection of the information, completion and filing of the CTR form, and retention of records, estimating an average of one hour for each CTR.7 However, that estimate was developed in 2002 when most reports were filed manually; since 2011, all CTRs must be filed electronically. There have been many other changes in the last 20 years, so FinCEN now proposes to recalculate the burden estimate by taking into consideration: (1) elements not previously considered, such as collecting data that a filer would not otherwise collect in the ordinary course of business and (2) use of technology and automated processes, including maintaining the technology infrastructure, to file CTRs.
To help ABA evaluate FinCEN’s revised burden methodology and estimates, we conducted an informal survey of members and received responses from 76 banks of all sizes, geographies, and supervisory agencies. Their responses are noted throughout our comments.
FinCEN believes that banks that file more than 100 CTRs per week file all CTRs using an automated process. Based on that assumption, the agency estimates the time required to complete a single CTR is approximately one minute for large volume filers.8 95% of the respondents to ABA’s survey, however, found that the estimate of one-minute was inaccurate.
For non-batch filers, which are reporters that file less than 100 per week, FinCEN estimates that completion of the form takes 20 minutes. ABA members found the estimate to be more accurate, with 44% of the bankers agreeing, but as noted above, others found that the process could easily consume two to three times that amount. ABA encourages FinCEN to conduct more in-depth analysis or surveys of filers to better understand the process they use to file CTRs.
ABA agrees that there has been a steady growth in the use of technology for BSA reporting. However, while there have been steady increases in automation, it has not reached the levels reflected in the revised burden estimates. In our survey, nearly 85% of the banks reported that the process was not entirely automated and that they still rely on manual steps to collect and report information. This was particularly true for any CTR which requires aggregation of transactions to report, and over 60% of the banks that responded to our survey reported that the process of aggregation of transactions for CTR filing is not fully automated. The majority of respondents believe that time spent on manual processes must be factored into FinCEN’s burden estimate.
Banks’ reliance on automation for CTR processing can range anywhere from 0% to nearly 100%, but all banks report that manual review and quality control play a significant role in the process,9 even those that have highly automated processes. For example, some banks use their automated system to flag a transaction, but the decision to file, the collection of information, and the completion of the form are conducted manually.
Of course, the more routine the filing, the more accurate FinCEN’s estimates become about the role that automation plays. However, when the process is more manual, which is more typical for smaller volume filers, it often requires one employee, usually a teller, to complete the form and then a second employee, possibly the BSA officer, to review the form for accuracy before it can be filed, and this takes more time than the 20 minutes allotted by FinCEN. Moreover, for all bank filers, a significant number of CTRs are not routine and require manual processing, which undermines the accuracy of the estimates.
Therefore, ABA believes the estimates need to be adjusted to reflect the fact that manual processes still play a significant role. While FinCEN suggests that manual processing is more likely with smaller volume filers, ABA finds that banks of all sizes still rely on manual processes. Even large volume filers in the top tranches rely on manual processes for quality control, handling exceptions, and processing aggregated transactions.
In addition to the preceding analysis, FinCEN has separated the CTR process into five distinct steps: (1) determining whether a CTR is required, either as the result of a single transaction or an aggregation of transactions across an institution’s different channels; (2) obtaining the necessary information for completing the form; (3) completing the form, either manually or using a completely automated process; (4) submitting the form; and, (5) retaining records.
Generally, ABA finds that this approach is appropriate; nearly 60% of the bankers in our survey agree that the five steps FinCEN has identified accurately reflect the CTR process. However, the majority of survey respondents found that FinCEN’s estimate of the time required by the first two steps – particularly collection of necessary information – can easily add 10 to 15 minutes per CTR, but this factor can add as much as 60 minutes to the time estimates used by FinCEN.
For example, aggregation of transactions is time-consuming. Indeed, one of the most time consuming steps for aggregation of transactions occurs when a bank has to evaluate transactions conducted through an ATM not owned by the institution. Another factor that consumes time is the collection of information about the conductor of a transaction, if that person is not the account-holder. Bankers report that it is not unusual to have a non-customer, such as an employee of a commercial customer, make a deposit to an existing account. Another situation that requires the collection of information on non-customers arises when a courier comes to pick up coins for a commercial customer. Similarly, a non-customer cashing a check drawn on an existing customer’s account requires the collection of information. Finally, to ensure quality of CTR filings, review of individual CTRs against other information on file can also add time to the average.
In addition, there are other steps not reflected in the analysis that are important and which should be included in the burden estimates. These include time spent on audit, quality control, error resolution, and installation and maintenance of technology. Over 95% of the banks surveyed believed that these and other elements are important to the CTR process. Even if some of these are difficult to quantify, it demonstrates that FinCEN’s proposed estimate of the burden significantly understates the cost of CTR filing.
For larger institutions, greater granularity is needed to accurately capture all the different steps in the CTR process, something FinCEN should consider as it works to better quantify the costs and benefits of AML compliance. Building on our prior comments, there are numerous support processes that go into a CTR filing system that have not been factored into the estimate. For example, technology costs are not limited to installing and updating systems, but should include time spent checking interfaces with other systems, monitoring that the system is working properly and that output has been validated—steps that can be both time consuming and expensive. And, it is not just the individual employee collecting and completing the CTR, but there are dedicated backroom staff charged with validating the information and correcting errors before the form is filed, and this can add approximately 25% to the overall costs. One banker estimated that to comply with the regulatory standards approximately 50% of all CTRs require manual intervention to verify the true beneficiary of funds, correct errors, or collect information from other systems not available to the employee preparing the initial form. Therefore, when manual intervention is needed, the one minute estimate is more likely 6 to 7 minutes on average.
FinCEN has bifurcated the costs for collecting information to complete a CTR into depositories and non-depositories. For depository institutions, FinCEN operated on the assumption that the necessary information is likely to be already on file since banks limit transactions to established customers. Therefore, FinCEN allocated no cost to collecting that information.
However, our survey shows this assumption is incorrect. Over 75% of the bankers in our survey reported that large cash transactions are not limited to customers. In fact, even banks that have a general policy restricting transactions to customers make exceptions. When a non-customer has a large cash transaction that must be reported, it takes time to collect the information and, according to bankers in our survey, it often takes more time than allotted by FinCEN. To increase the accuracy of FinCEN’s estimates, some estimation of the number of banks’ non-customers who are subject to CTR filing is needed. In addition, the estimate of time spent collecting information is also needed. Less than 10% of the bankers surveyed felt that three minutes for a transaction with a non-customer was accurate and over 80% of the bankers agreed it would take more than one minute to process a CTR involving an armored courier.
It is difficult for FinCEN to estimate costs of software and hardware, since CTR systems may be integrated into other operations. In addition, costs evolve over time as technology is upgraded. To derive an estimated cost for technology, FinCEN based its estimates on information from an assessment of cross-border electronic funds transfers (CBETF) done in 2008.11 They then used proportional costs to extrapolate an estimate of technology costs. Using that methodology, FinCEN has concluded that the costs of technology represent 15% of the overall costs of CTR processing. ABA survey respondents were evenly split on whether 15% of the costs reflected the role played by technology in the CTR process, with 50% agreeing and 50% disagreeing. FinCEN is planning to conduct a survey of filers to obtain more precise figures, a step that ABA welcomes and would be happy to help support.
ABA appreciates and support FinCEN’s work to better analyze the burdens associated with processing and filing CTRs. We agree that technology plays an increasingly larger role in the CTR filing process, but disagree that it is anywhere near the level that FinCEN believes it to be. Manual processes in CTR processing are far more significant than FinCEN believes and must be factored in the burden estimates. We also believe that there are many steps, not only those identified by FinCEN but those identified by bankers, that need to be reflected in the process to better estimate burden. In addition, we believe that the value of benefits must be reflected in the final analysis.
As noted at the outset, we strongly encourage FinCEN to share information about CTR filing data. That lack of transparency is one of the elements that handicaps the analysis. In the long run, greater transparency would help the industry and FinCEN focus better on the costs and benefits of the CTR process.
ABA looks forward to working with FinCEN to collect the necessary information and take the necessary steps to improve the effectiveness and efficiency of CTRs and the BSA. If you have any questions or need additional information, please contact the undersigned at [email protected].
Sincerely,
Robert G. Rowe, III
Vice President & Senior Counsel