B-260165 Compliance 2000 IRS viewed its Compliance 2000 strategy as a way to advance voluntary compliance. The strategy differed from the traditional enforcement approach by recognizing that nonenforcement approaches, such as education and assistance, can boost compliance. To determine when it was best to use each approach, IRS sought to uncover root causes for noncompliance and distinguish between compliant and noncompliant taxpayers across market segments. Compliance 2000 did not work for various reasons. In 1992 testimony, we reported that IRS lacked the necessary compliance data and infrastructure to do research by market segments. We found that IRS had not tracked whether its districts started research projects on the basis of objective compliance data or researched the most noncompliant market segments. We concluded that Compliance 2000 was a worthy idea that needed careful implementation. We stated that IRS needed to use objective data to select research projects and develop an infrastructure for planning, managing, and monitoring the projects. An IRS Internal Audit report in December 1993 had similar findings. The report disclosed that 38 of 50 Compliance 2000 projects were traditional enforcement projects that the districts had renamed as Compliance 2000 projects. And IRS had no database to capture results or provide an inventory of the compliance issues covered. The report concluded that the projects did not represent rigorous research, the managerial controls were weak, and a management structure was needed to provide effective oversight. NORA officials acknowledged such problems and indicated that very few Compliance 2000 projects could be viewed as viable research. Even in the few projects that NORA officials viewed as viable, IRS had not created a database to show whether compliance increased and, if so, what actions prompted those increases. We sought to further confirm these problems by collecting data on Compliance 2000 projects as we visited the 31 DORAS. We confirmed that IRS did not track the methods and results of the projects. As shown in appendix I, we found that many projects were duplicated. Available records were insufficient for us to compare costs and benefits across the projects. 'GAO/T-GGD-02-48. B-260165 We found that IRS learned many lessons about research from Compliance IRS' New Approach Support for Compliance IRS holds high hopes for its new compliance research approach in integrating the Compliance 2000 philosophy with efforts to boost compliance. To act on this potential, IRS has taken steps built on lessons learned from past efforts. While directives are important to set the vision, building support relies on collaboration. In this vein, NORA has developed a cooperative strategy to communicate the research vision, needs, and results as well as generate feedback on the needs of IRS districts and functions. Given such feedback, NORA plans to create a special unit to meet the needs for research on ways to better select and handle workload. NORA is also encouraging DORAS to provide short-term research assistance to districts and functions (e.g., electronic filing and earned income credit). Ultimately, NORA knows that the new approach will have to prove its worth to build the necessary staff support. IRS Is Developing a IRS' new approach depends heavily on CRIS. As envisioned, CRIS is to be IRS' network of databases for identifying the nationwide and district compliance of market segments. IRS is implementing CRIS in the following three stages. • Working File CRIS was used in fiscal years 1994 and 1995 for training DORA staff. It had 75 data elements limited to one market segment. B-260165 DORA Positions and • Interim CRIS was delivered to all DORAS by fall 1995. It expanded to 800 data elements and samples of individual and business filers for all market segments. • Final CRIS is slated to implement its first database, having over 2,500 data elements on a sample of 7 to 10 million individual filers, in fiscal year 1997. It is to interface with other systems being created to aid in storing data and assigning workload. It is expected to contain 3 years of data. On completion, CRIS is to have 10 databases, each with thousands of data Also, realizing that IRS data contain taxpayer and IRS errors, NORA has developed data validation standards. NORA officials believed that these standards will better ensure that the research stems from adequate data. Past research efforts highlighted the need for staff who had research skills. Toward that end, NORA devised a staffing plan that requires certain positions at each DORA, such as a chief, program analysts, and other generalists. NORA also created specialist positions that require skills in statistics, operations research, economics, and computers. Recognizing limitations in having such staff in the field and restrictions on external hiring given the redeployment agreement,* NORA encouraged DORA chiefs to fill positions with the most qualified staff available. NORA expected the number of staff to initially total about six to eight per DORA and grow as workload dictated. NORA also devised a plan to train all DORA staff in research methods. Phase I training, which began in early 1995, described NORA, DORA, CRIS, profiling, statistics, and research methods. Phase II training includes advanced methods in statistics, research, and market segmentation. NORA is also offering customized training to meet the needs of DORA staff. *Redeployment agreement refers to IRS' pledge to protect jobs. Under agreement with the National Treasury Employees Union, IRS guaranteed that employees would not lose their jobs due to IRS' modernization and reorganization and would have an opportunity to maintain their grade through training. Years of IRS service plays a major role in qualifying for redeployment. B-260165 The Infrastructure Has IRS has laid the framework for the infrastructure it believes is needed to manage the new research approach. This framework includes NORA, DORAS, a research plan, and research methods. IRS has plans for other mechanisms to manage the research. NORA and DORA officials said research in the field has often suffered Furthermore, IRS districts are forming Compliance Planning Councils (CPC) at the DORA level to build district support for compliance research, oversee district compliance programs, and provide a conduit to the three district functions. In summary, CPCS are to provide a multifunctional perspective in reviewing district compliance workload. At a minimum, each CPC is to consist of the DORA Chief and chiefs of the three functions. IRS is also forming nine Cooperative Strategy Working Groups (CSWG) to help with oversight, coordination, and implementation of the new approach. CSWGS are to make many of the decisions about compliance research, with the concurrence of the national director for compliance research. NORA is developing an annual research plan and a compliance research |