Our Work: Permanent Fund Dividend Payment System
Each year, the Permanent Fund Dividend Division (PFDD) oversees the single largest ACH transaction in the nation, distributing more than $500 million—sometimes $1+ billion—in oil revenue dividends to more than 600,000 Alaskans.
However, the PFDD’s payment system, written in COBOL and running on mainframes, was at the end of its life. The PFDD wanted to replace their system by adding new modules to their Dividend Application System (DAIS).
“Distributing nearly a billion dollars demands flawless software. Once again, Resource Data delivers.”
—Robert Wood, Alaska Department of Revenue
After analyzing existing systems and documentation, Resource Data designed and built the new payment modules. These modules include functionality such as
- Calculating the PFD amount
- Processing, in prioritized order, deductions to each individual’s PFD, including garnishments for outstanding debts and voluntary contributions to charities or a college savings plan
- Processing dividend payments through a Disbursement Dashboard
- Managing fraud and collection cases
- Interfacing with numerous external systems, including State and federal agencies and banks
In addition to building the new modules, we also
- Migrated more than 25 years of mainframe flat files (several terabytes of data) into the new database
- Enhanced the online application to support new program rules, such as charitable contributions
With the new payment modules, distribution processing time is now measured in hours instead of days, processing is more accurate, operating costs are lower, and the PFDD has greater flexibility and control when executing distributions.
The new system had to be highly accurate, especially with the PFD program’s high visibility. Therefore, our testing process was very rigorous and aggressive. We assigned a dedicated test lead, made the development team hyper-accessible to the testing group, cataloged and tracked all issues, and met frequently with the testing group.
With each distribution producing several million records and querying more than 30 years of history, we also architected the system to be highly efficient. We tested this efficiency by running the distribution process for all available applicants—even ineligible ones—to ensure it would complete within the 3-hour goal.