Hoosier Banker 41 could miss parameters that are set too low and generate a large volume of false positives. Conversely, when parameters are set too high, the situation can result in missed suspicious activity. Ideally, model validation includes “above the line” and “below the line” testing to challenge the assumptions that went into the model’s creation. Such an approach reduces the risk of criticism for using default parameters. Don’t set thresholds too low. One bank was criticized in an exam because its model generated a large volume of alerts that were determined not to be suspicious. For example, to identify potential elder abuse, the bank’s model had a parameter that alerted every time a customer over 60 years old had a debit between $2,500 and $10,000. This parameter duplicated another parameter that generated an alert for elder cash withdrawals over $3,000. This and other parameters that were not set with reasonable risk-based thresholds led the examiner to express concern that staff could not adequately review all of the alerts, due to the large volume. Additionally, the fact that staff members were overwhelmed by alerts increased the risk of suspicious activity being missed. In another example, one bank used its model to monitor dormant accounts. It had set parameters to identify any deposit account containing over $1,000 and that had no activity within the previous 180 days. That parameter generated the fourth-highest number of alerts compared to all activity parameters, a sure sign that the setting was inappropriate. Don’t set thresholds too high. Thresholds being set too high can result in suspicious activity being missed. For example, one bank’s parameter that was designed to identify spikes in PIN-based debit card purchases ignored activity below $40,000 within a month. This bank had a $1,500 daily debit card limit, so the maximum that a customer could possibly purchase in a month was $46,500. As a result, the parameter never generated alerts. In another bank, a parameter designed to identify large loan payments and paydowns ignored activity below $75,000. A review of the bank’s product offerings noted that its unsecured loan product had a maximum loan amount of $25,000, meaning that such transactions would never generate alerts. Some institutions outsource validations to eliminate busywork, but this cannot be the intent of performing a validation. The purpose of an independent validation is to objectively consider parameter thresholds and recommend adjustments. Such adjustments may increase or decrease the volume of alerts. Institutions that do not periodically evaluate their parameters could be subject to costly lookbacks, civil money penalties, memoranda of understanding, matters requiring attention and cease-and-desist orders. HB A communications toolkit created by the Indiana Bankers Association helps member banks share good-news stories with media outlets. Additionally, the toolkit provides an optional quick-link webform for submitting updates to Hoosier Banker, such as bank hires/promotions, branch openings, community service and more. • View the toolkit at: indiana.bank/communications-toolkit • Access the webform at: indiana.bank/submit-news For more information, contact Laura Wilson at lwilson@indiana.bank, 317-333-7146. HB Toolkit Helps ‘Tell the Story’ of Banking Tell the Story
RkJQdWJsaXNoZXIy MTg3NDExNQ==