TREASURY INSPECTOR GENERAL
FOR TAX ADMINISTRATION
A Prototype to Help New Small Business Employers Appears Effective, But More Information Is Needed Before Expanding It Nationwide
Reference No. 2001-30-118
The Collection Partnership Redesign Team implemented the Mentor and Monitor (M&M) Prototype to help small business owners understand their tax responsibilities. The prototype was designed to inform new employers of education and assistance opportunities, assist them in complying with employment tax requirements, monitor their tax compliance, and identify and help resolve delinquencies in paying taxes and filing returns. The M&M Prototype project manager has estimated that, if fully implemented, the M&M Prototype would help 250,000 new employers annually at a cost of $51 million, or about $200 per employer.
The overall objective of this audit was to determine whether the M&M Prototype would help the Internal Revenue Service (IRS) improve service to each taxpayer by providing prompt, professional, and helpful treatment to new businesses.
The M&M Prototype appears to be helping certain new business taxpayers comply with their employment tax responsibilities. Small Business/Self-Employed (SB/SE) Division management used an innovative approach to determine the best method to help new employers comply with their tax responsibilities. The prototype should help reduce taxpayer burden, uncertainty and errors by clarifying tax law requirements, resolving issues early in the process, and offering effective communication and education assistance. However, we have concerns that there may not be sufficient information to evaluate the effectiveness of a nationwide rollout of this prototype.
The Mentor and Monitor Prototype Could Provide Valuable Information About the Behavior of Small Business Taxpayers
SB/SE Division management used an innovative approach to determine the best method to help new employers comply with their employment tax responsibilities. They used a special mail-out with materials designed to assist employers in meeting their employment tax responsibilities, profiled employers to identify those most likely to need assistance, and surveyed employers to obtain feedback concerning the value of the prototype. They also partnered with the Small Business Administration (SBA) to secure 10,000 of the SBAís Programs and Services Brochures. In addition, a database was established for tracking case progression, verifying data input, and monitoring telephone contacts. Conclusive results for the prototype are expected to be known in December 2001.
The Sampling Methodology May Not Permit Valid Statistical Inferences About the Effectiveness of the Prototype
Management did not apply random sampling principles in selecting the samples for the M&M Prototype. Therefore, statistical inferences may not be made about the effectiveness of the prototype and the feasibility of a national rollout. However, the planned analysis of 20,000 new employers could provide valuable information about the behavior of small business taxpayers that would aid taxpayer education efforts.
Incorporating the Prototype With Other Taxpayer Education Programs May Be More Cost Effective by Providing Greater Benefits to More Small Business Taxpayers
The M&M Prototype is currently being developed as a program to help small business taxpayers. However, the mentoring aspect of the prototype may not be practical because it is very labor intensive, would require extensive hiring or reallocation of resources, and would help only a limited number of taxpayers. The M&M Prototype project manager estimated that if the prototype were implemented nationwide it would require approximately 1,100 staff years.
Greater Management Oversight Would Help Ensure That the Prototype Will Achieve Its Goals and Objectives
The IRS Headquartersí National Public Liaison and Office of Tax Administration Coordination provided oversight during the early development of the M&M Prototype. However, they ceased their oversight relatively early, and there has been no oversight during most of the process. Greater management involvement may have identified the sampling problems in the initial planning stages, evaluated the feasibility of a national rollout earlier, and considered incorporating the prototype with other taxpayer education programs.
Summary of Recommendations
We recommended that SB/SE Division management ensure that statistically valid sampling techniques be employed where projections will be used and samples will be compared, consider incorporating the M&M Prototype with other taxpayer education programs, and provide oversight for the duration of the prototype.
Managementís Response: The Commissioner, SB/SE Division, agreed with two of the three recommendations presented in the report. The SB/SE Division will consider providing one-on-one taxpayer education service but only in those cases or markets involving high risks of non-compliance. The Commissioner stated that they are exploring the various channels and will include lessons learned and some of the fundamental principles from the prototype as they continue to define the role of the Taxpayer Education and Communication (TEC) organization. In addition, the Division will follow up with a business case analysis of the compliance results that includes the appropriate level of management oversight. The SB/SE Divisionís Director, TEC, will be the recipient of the business case analysis.
The Commissioner, SB/SE Division, had a different opinion about the sampling methodology used in the prototype. He believed that the judgment sample used would provide sufficient information to evaluate the effectiveness of the prototype. The IRS Office of Research agreed that the judgment sample was an acceptable method that provided a number of benefits; however, they also stated that a valid statistical sampling technique would have been the preferred method.
Auditorís Comment: While we believe that the SB/SE Divisionís judgment sample will provide valuable information to evaluate the prototype, we concur with the Office of Research that a valid statistical sample would have been preferred. We continue to believe that the SB/SE Division cannot make statistical projections from its sample.