![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
||||||||
![]() |
![]() |
|
|||||||||||
![]() |
Home| Services|Working with ACF|Policy/Planning|About ACF|ACF News | Search | |||||||||||
![]() |
![]() |
![]() |
![]()
In this chapter The major elements affecting the cost of implementing a uniform statewide interest policy include: Based on other states' experiences, we believe that automation and notification are essential to a uniform statewide interest policy. We believe the former is especially critical and know of no state assessing interest statewide that has not automated the calculation and tracking of interest. Periodic notification, as practiced in Virginia and Massachusetts, also seems essential because of its potential for increasing collections. This chapter discusses each of the major cost elements and estimates the costs of implementing a statewide interest policy. New business rules are typically developed by a committee composed of state and county staff. State Policy and Evaluation administrators estimate that it will take a total of 2,080 hours to develop and document business rules relating to interest. They anticipate that the task would be divided equally between state and county staff. State staff would most likely include administrators, programmers and general professionals (Levels III and IV). County staff would mostly consist of legal technicians(Levels III or IV). The estimated average pay rate for these positions is $18.62 per hour, or $26.31 per hour when fringe and indirect costs are included. Thus, the total labor costs of developing business rules is estimated to be $54,725 (2,080 hours x $26.31/hour). This does not include the costs of travel to Denver by county staff involved in the process. This section discusses the system requirements for automating interest (using layperson's terminology), the estimated costs of these requirements, and the time needed to implement the system changes. The estimates are based on conversations with staff and administrators responsible for policy and ACSES. Specifically, we include estimates for programming time, testing, training, development, and documentation and maintenance. The estimates may vary somewhat once business rules pertaining to interest are finalized. We have attempted wherever possible, however, to note where variations to business rules could affect automation and costs. Costs are based on the following time estimates: The time estimates are based on the ACSES team's recent experiences modifying ACSES to accommodate PRWORA requirements. There are eight family classifications of dollars on ACSES that will be affected by interest calculation, including: All eight of these classifications will be affected because we assume that the State will want to: (1) separate interest according to federal distribution rules; (2) not assess interest on costs and fees; and (3) make other distinctions with regard to classification. Costs and fees (and any interest charged on them) are not eligible for federal incentive payments because they are program income rather than child support collections. Payment of interest is subject to federal incentive payments as long as interest is attached to child support arrears. The screens and programs for these classifications are numerous. We discussed with State staff whether it would be possible to calculate interest on fewer classifications and thus lessen the number of screens and programs that would require reprogramming for interest automation. For example, could automation be simplified if interest would always be owed to the family rather than the state (i.e., eliminate classification 1)? While such a simplification may ease the administration of interest and the business rules, from a technical perspective the screens and programs associated with all eight families will still need to be addressed. In other words, variations in policy options will not require less programming effort. The rest of this section identifies which programs, batch processing and screens would be affected by automating interest. These are separated by existing and new programs. At a minimum, ACSES staff estimate that 80 existing programs would be affected by interest automation. Their estimate is based on recent experiences modifying ACSES to accommodate PRWORA, which resulted in changes to 80-120 programs. The program, screens and batch jobs that would be affected by interest automation include:
These 80 programs would require 1,200 hours of programming time (80 x 15 hours per existing program). An additional 600 hours are needed to reprogram the Financial Court Case Summary which is used for credit bureau reporting, license suspension, wage assignments resulting from new hire reporting, financial institution matching and other enforcement remedies. This does not include any modifications to the monthly billing coupons. This estimate also does not include the development of business rules on how interest calculations would interact with these enforcement remedies. As discussed earlier in this chapter, this task would be completed by a workgroup of State and County administrators, general professionals, programmers, and legal technicians. They would answer policy questions such as would license suspension be triggered by the amount of arrears principal or the sum of arrears principal and arrears interest? Would a professional license be suspended if the arrears balance comprises interest only? The total estimate of time to reprogram ACSES to accommodate interest automation is 1,800 hours. ACSES staff estimate they will need to write about 20 new programs for interest automation. This includes programs for:
In all, an estimated 400 hours of programming time (20 staff hours x 20 new programs) will be necessary to develop the new programs required to automate interest. This amount may vary significantly depending on what business rules are developed. At a minimum, program testing takes 8-10 weeks and involves two to three staff. Another 4-6 weeks are required for documentation. The recent experience with the MADJAD for documentation and training involved 12 weeks and five staff. Training involves another 6-8 weeks. Finally, one programmer would respond to user problems for 4-6 months after implementation. These amounts may also vary significantly with business rules. The total estimated costs of automated interest are summarized in Exhibit 14. The major assumptions associated with these estimates are also listed below.
Table 4.1. Exhibit 14: Estimated Costs of Automating Interest for ACSES
Finally, ACSES staff did not believe that the work requirements would differ if simple or compounded interest is used, nor did they believe the programming effort would be much different for statewide automation of interest or automating an interest amnesty program similar to that of Massachusetts. In other words, many policy decisions have a negligible effect on programming time. The programming would most likely be contractual at $75 per hour. The testing, training and documenting would be performed by a general professional III, whose midrange rate is $19.58 or $27.66 per hour when fringe and overhead costs are included. Using these rates, the total estimated costs of automating interest on ACSES are $268,120. One of the perennial issues states face when they first automate interest is whether to enter retroactive interest and if so how. The approaches other states have taken include the following:
These options are discussed in greater detail in Chapter II. Based on conversations with State staff, automatically calculating retroactive interest is not a viable option. So, this leaves the options of (1) starting at $0 interest balances, (2) starting with the interest balance that is already entered in ACSES, or (3) manually calculating retroactive interest. Based on the experiences of other states, however, we do not recommend a manual calculation. Interest was calculated manually by contractors in Alabama and New Mexico. Alabama administrators believed this conversion was generally successful although they acknowledged there were some errors that resulted in appeals. On the other hand, New Mexico encountered many difficulties and errors in their conversion; this included the conversion of interest and other information. Arizona and New York provide examples of the phase-in of retroactive interest. In Arizona, the interest calculator is to be turned on as workers update cases. After three years, not all Arizona cases have been updated to include interest. In New York, counties are responsible for verifying judgments and the interest on the judgment, then entering it onto the state system. From then on, the interest calculation becomes automatic. The extent to which this has been done varies among counties. To automate a retroactive interest calculation using ACSES will be prohibitively expensive, if not impossible, because it will require automating a chronology of payments and order changes for the life of every order. Instead, an early ACSES planning committee, that had drafted a plan to include interest calculations on ACSES that became a lower priority issue when additional federal requirements were imposed, recommended starting the interest calculation with the first delinquent monthly order amount. Another option would be to start the interest calculation for all arrears beginning on a certain date. The early committee dismissed this idea because it has the potential of inflating accounts receivable quickly, particularly when interest is compounded. In our survey of Colorado counties, we asked what approach they favored. Most favored entering $0 interest balances or whatever amount the county gave the state to enter. One county administrator suggested going back through ACSES and zeroing out all interest arrears that could be identified. We estimate the cost of calculating retroactive interest manually to be about $833,756. This estimate is based on the following assumptions: This estimate does not include the costs of supervision or verification. Furthermore, we believe the 25 minute average time requirement is highly variable. In another contract PSI has with the State to update its child support staffing standards, for example, we have asked technicians in several counties how much time it takes to calculate arrears alone, a first step in calculating retroactive interest. The typical response is, “It can take 10 minutes or all day depending on the case.” The issue of how to handle retroactive interest will also need to be addressed in the development of business rules. If interest only accrues on past-due current support as of a specified date, should there be an amnesty program for those cases that have already accrued interest? What about custodial parents? Are they entitled to interest prior to a change in policy? Costs for training county staff on business rules are estimated at $24,817 ($1,442 incurred by the State and $23,375 incurred by the counties). This is based on 13, two-hour training sessions administered over the course of four days to 626 county enforcement technicians at the State office. It assumes an additional 8 hours for trainer preparation and $100 per day for refreshments. It does not consider travel expenses; that is, county staff traveling to Denver and State trainers traveling to the West Slope to hold one training session there. The estimated cost of one notification to all arrears cases is $41,799. This assumes 107,178 notices (which is the number of arrears cases as of January 2000) and $0.39 per mailing. This also includes processing time and printing costs incurred by the Colorado Information Technology System (CITS/CS) in Lakewood, the mailing and processing of envelopes by Mail Services on Sherman Street, and the costs of double-window envelopes (so the county address appears as the return address). Costs are estimated based on prior experiences mailing notifications of tax refund intercepts. They will vary somewhat according to the volume. Undoubtedly, statewide interest assessment will increase customer service needs. Although notification may encourage some obligors to pay, it also may cause some obligors to contact counties to request more information or to dispute the interest charges. We were unable to obtain estimates about the number of telephone calls placed in response to other notifications in Colorado. Yet, our focus groups for the staffing standards project indicate that the volume of telephone calls following tax intercept notices takes 1-5 hours per day more for at least one to two months following the date the notice is sent, and some additional time when tax refunds are mailed out. This translates into about 3-25 minutes per letter sent. (This is regardless of whether the letter is returned due to a bad address or the addressee does not call CSE.) PSI's customer service experience in our privatization offices is about 3.5 minutes per letter sent. Nonetheless, the average number of minutes is likely to vary significantly with the contents of the letter and what actions are required. Assuming (1) a mailing of 107,178 letters, (2) that telephone calls will be handled by County Legal Technician IIIs or IVs, and (3) 3 minutes per notice, the customer services costs are estimated to be $100,050 for the first mailing. This is a conservative estimate. We suspect that the number of calls will decrease with subsequent mailings particularly if the mailings are monthly or quarterly. Exhibit 15 summarizes the estimated start-up costs for automating interest and instituting a statewide policy. The estimated start-up costs are $347,662. This assumes that interest will not be retroactively calculated. If retroactive interest is calculated, it would have to be calculated manually. This is estimated to cost $833,756 at a minimum. Table 4.2. Exhibit 15: Estimated Start-up Costs for Automating Interest and Instituting a Statewide Policies
Exhibit 16 displays the notification costs and increases in customer service costs that result from notification. Notifying obligors once a year of interest charges is estimated to cost $141,849 per year, quarterly notification is estimated to cost $267,246 per year, and monthly notification is estimated to cost $601,638 per year. The customer service response to these notifications is conservatively estimated. In reality, the customer service response could vary widely.
Download FREE Adobe Acrobat® Reader™ to view PDF files located on this site.
OCSE Home
|
Press Room
|
Events Calendar
|
Publications
|
|