What’s next for PPP and data management?
The first phase of the Paycheck Protection Program (PPP) has wrapped up and the numbers are impressive: more than 5 million loans totaling $525 billion from nearly 5,500 lenders. Any bank that participated will attest that the data processing headaches were just as notable.
In the PPP loan forgiveness phase, the data and operational needs to process and adjudicate the applications are shaping up to be even more complex – when the details are final, that is. Congress continues to debate legislation to streamline adjudication and automatically forgive loans up to $150,000. While a simplified process would be a welcome reduction in banks’ data burden, strategic data management remains a critical part of a successful PPP.
Banks primarily collected demographic and basic business performance data in the PPP application. The forgiveness phase, however, is more detailed. It requires borrowers to provide detailed payroll and non-payroll breakup of the spending, supported by documentation.
The Small Business Association’s EZ Forgiveness Application helps record the spending, but establishing proof is a more freewheeling process. Recordkeeping at small- and medium-sized businesses takes many forms. Some rely on standard accounting software, while others prefer homegrown accounting tools. And a significant percentage outsource payroll.
Banks have the task of converting the patchwork of information into a standardized format. The challenge involves intake, normalization and processing of details into a format that can be fed to middle- and back-office systems and contact centers. Unstructured proof of spending is even more complex, consisting of everything from rent payment invoices to emails documenting job offers to prior employees. Accepting the unstructured documentation is likely to be a manually intensive process.
Managing PPP data is a critical endeavor. With many Americans under financial stress in the wake of COVID-19, banks have a responsibility to safeguard government assistance and ensure it reaches those who need it. Mitigating widespread PPP fraud is part of that responsibility. The Justice Department estimates 12 percent of PPP assistance was misappropriated by fraudsters. Better data collection, validation and management during the application phase can help prevent this misuse and lack of compliance.
Strategies for PPP processing and data management
How can banks manage and make sense of the volumes of non-standard data? A unified PPP database that’s agile, secure and available is essential. Many banks are adept at setting up databases and supporting them with mature data management practices. That’s a good start.
But to effectively gather, process and adjudicate the forgiveness applications, PPP requires a more expanded data management strategy that includes a simple, digital front end supported by document processing and analytics tools, and a strong workflow system to automate business processes and communications.
Many banks established a PPP-specific digital front end to collect borrowers’ raw data. The intake tools range from Excel templates or web forms to direct interfaces with accounting data providers.
Fewer banks took the step of implementing advanced document-processing tools, which are invaluable for automating data parsing, validation and populating PPP database tables. They typically include image-processing tools that perform the same automation for unstructured images. Some document-processing engines use AI instead of rules-based engines to automate data parsing and are capable of learning over time.
Most banks haven’t implemented sophisticated data analysis and rely instead on back-office systems and human intervention for processing. As a result, the PPP processes can become operationally intense and costly, leading to compliance and reputational risks. While simple validations like demographic and business validity checks can be performed with readily available external sources, banks gain a more consistent, repeatable approach with data analysis tools that provide evidence-based, referential data points.
Analytical engines go a step further by crawling through the data to validate information and flag potential fraud. For example, analytical engines can compare data from borrowers with cohort information to identify inaccurate payroll information. They can also spot multiple submissions and forgiveness requests by correlating borrower information across processed loans.
Finally, workflow tools that automate business processes enable banks to streamline and make faster decisions on loan applications. They apply the essential business rules, tag the validated data to a work step and queue it for approval. The tools can also initiate communication with borrowers to gather additional data and track inbound calls. Leading tools also curate the interaction to support future compliance and regulatory reporting.
PPP has highlighted the challenges in data gathering and processing for banks. Modern data management, analytics and tools provide the answers. With them, banks can be responsible stewards of public money and responsive to their customers and shareholders.