Our review of the Franchise Tax Board's (board) audit and collection activities revealed the following:
A primary revenue-generating agency for the State, the Franchise Tax Board (board) processes individual and corporation tax returns, audits certain tax returns for errors, and collects delinquent taxes. Between fiscal years 1990-91 and 2001-02, the board provided an average of $31 billion in annual tax revenues to the State, over 60 percent of the State's General Fund that pays for education, health, welfare, and other public services. Although many taxes are self-assessed by individuals and companies, the board's audit program reviews the accuracy of tax returns, assessing additional taxes when appropriate. In turn, the collection program pursues delinquent taxpayers identified by the board's various assessment activities.
The variety of performance measurements the board uses for the audit and collection programs can confuse decision makers such as the Department of Finance (Finance) and the Legislature about the programs' projected and actual results. A complete performance measure compares all the benefits of a program with all the costs of producing them. However, in budget documents describing the projected benefits that will result from new staff, the board excludes some departmental overhead costs without disclosing this exclusion. However, the board's subsequently published historical reports of program results include all costs. Further, the board's budget documents do not disclose a substantial overlap in benefits that are stated once by the audit program as assessments and again by the collection program as those same tax assessments are collected as revenue. The board believes Finance and the Legislative Analyst's Office (LAO), which advises the Legislature on budgetary matters, are aware of the overlap. Finance confirmed that it is aware of the overlap, but indicates it would support a refinement of these measurements to better capture the benefits. The LAO is also aware of this overlap and considers it when interpreting the board's data.
To increase revenues, the board received authorization for an additional 340 net audit positions between fiscal years 1992-93 and 2001-02. The board justified many of the new positions with cost-benefit ratios (CBR) that projected returns of at least $5 in audit assessments for every $1 of cost. In contrast, our review found that for every $1 of cost, the 340 audit positions returned only 79 cents in assessments over the period although the return on the additional positions improved to $2.71 for fiscal years 1998-99 through 2001-02. Changes in the economy probably affected the return on these audit positions, but a major cause of the low return is that despite having additional staff, the board did not increase the number of hours staff spent performing audits. These hours differed little in fiscal year 2001-02 from those in fiscal year 1992-93. The collection program added 175 collection program positions between fiscal years 1998-99 and 2001-02, projecting increased revenue of $179 million over that period. However, because of limitations in board data, we could not determine the return on the 175 collection program positions.
Although sufficiently demonstrating the overall cost-effectiveness of its audit and collection programs, the board's process for assessing the incremental benefit of recently acquired audit and collection program positions is flawed. The board lacks sufficient data and uses an inadequate methodology to determine whether increases in audit assessments or collection program revenues resulted from additional positions. Rather than using an incremental approach to isolate assessment or revenue pools likely to have been affected by additional audit or collection program positions, the board compares its total projected audit assessments against its total actual audit assessments and its total projected collection program revenue against its total actual collection program revenues. At the highest level of analysis, the board can demonstrate that the audit division returns about $10 in assessments, or potential revenue, for every $1 of cost, and the collection program returns around $19 in revenue for every $1 of cost. However, the board lacks a persuasive analysis to show that additional tax assessments and cash receipts to the State came from the workloads that incremental staff would likely have been assigned to work.
Recently, the board has justified collection program staffing requests based on a process that prioritizes workload according to a cost-benefit ratio. However, the board actually assigns staff based on risk and yield factors calculated by its new Accounts Receivable Collection System (ARCS). This leaves the board unable to adequately demonstrate the reliability of its cost-benefit approach. Now that the collection program has nearly two years of experience using the new system, the board is developing an alternative methodology for justifying collection program staffing needs that uses data from ARCS to better reflect the manner in which the board actually assigns collection program staff.
Finally, the board is not using all of its funding for collection program salaries to actually fill authorized positions, but is instead using some of the funding for other costs. In fiscal year 1999-2000, separate merit salary adjustment (MSA) funding for the board was ended, leaving the board to find another way to pay for its MSAs. In fiscal years 2000-01 and 2001-02, the board's savings on salaries increased by roughly 5 percent. To achieve these savings the board has left unfilled some collection program positions even though the board's budget control language requires it to fill them expeditiously.
To more completely and clearly reveal its programs' costs and benefits, the board should consider using the complete measurement of the audit program's performance that we have described in Table 3 on page 17. This measurement compares all the benefits—the total revenues that result over time from the auditors' assessments of additional taxes—with the total costs to produce them, including the costs of collection. Thus, the board would treat the collection program as another service center for audits. If it determines that its current information system cannot produce the data necessary for such a measurement, the board should consider the needs of a complete measurement when it upgrades or changes its current information system.
If the board decides not to use the complete measurement and continues to use separate performance measurements for the audit and collection programs, it should do the following:
To demonstrate the effectiveness of new collection program positions, the board should develop a methodology for measuring the benefit of these positions by isolating the return resulting from the additional positions and comparing it against a base year.
To more accurately represent the process that assigns work to collection program staff, the board should continue to develop a methodology based on data from ARCS to justify new collection program staffing requests.
For the board to be consistent with the intent of budget control language and Finance, it should not as a long-term strategy leave collection program positions unfilled beyond the normal time it takes to fill a position.
The board indicates it will implement most of the recommendations made in the report. It agrees that improvements can be made to increase the usability of information it provides to Finance and the Legislature, asserting that it has already begun a project to better capture revenue and cost data. The board disagrees with the methodology we used to analyze the additional assessments generated by the new audit staff. The board contends that because our analysis fails to take into consideration tax law changes, tax regulations, case law and precedent, economic conditions, and self-compliance it does not recognize the full value of the new audit positions. We provide comments to clarify and add perspective to the board's response to the audit.