In July, 1994, Colorado law was amended to give the State Division of Child Support Enforcement authority to report to credit reporting agencies, all cases in which child support debt or child support arrearages are owed (C.R.S. 26-13-116). The central goal of the intervention is to induce noncustodial parents who are delinquent in their child support payments to change their payment behavior in order to avoid an unfavorable credit rating. It was expected that the impact on collections would be modest in the short-term but that more significant benefits would show up over time as creditors deny credit to delinquent noncustodial parents.

Credit bureau reporting (CBR) had been offered to states as an enforcement tool in the Child Support Enforcement Amendments of 1984 which mandated states respond to credit bureau requests for information on noncustodial parents who were $1000 or more in arrears and resided in state. A subsequent survey of 16 states conducted by the US Government Accounting Office (1994) revealed that eleven went well beyond this legal requirement and were routinely reporting child support payment information to credit bureaus. Limited evaluation data suggested that credit bureau reporting had a positive impact on enforcement with only minimal start-up and operational costs.

Colorado's approach to credit bureau reporting was developed by a committee comprised of state and county child support enforcement personnel. Its central features involve the conduct of the initiative by state-level child support personnel at the State Enforcement Unit (SEU) and the extensive use of automation to select, notify and refer noncustodial parents to credit bureau reporting agencies. The planning process that the Colorado committee pursued is documented in a report by State Enforcement Unit Supervisor, Anita DeVargas Field (1996). Anita Field's report also outlines the measures taken by the state and county personnel to implement this intervention. Finally, Anita Field presents the detailed business rules developed to effectuate the automated reporting of obligors to credit bureau reporting agencies, and the detailed procedures to be followed by child support staff at the county level.

This report describes the results of a third-party, independent evaluation conducted by the Center for Policy Research. The evaluation includes a quantitative assessment of collections due to credit bureau reporting over nine and sixteen month periods of time after notification letters were sent to noncustodial parents. It also presents the reactions of key groups including state and county child support technicians and members of the lending community.

The quantitative evaluation of Colorado's credit bureau reporting initiative involved the comparison of payment patterns in two groups of child support cases: an experimental group comprised of 3,000 cases reported to the credit bureau reporting agencies; and a control group comprised of 3,000 cases that were not reported. Both groups of cases were randomly generated from the total pool of Colorado child support cases that met the criteria for credit bureau reporting. In order to be selected for reporting, an obligor was required to have a child support order and a delinquency greater than $500 on any of his/her court orders. Obligors who had a responding interstate case were not selected for reporting. In addition, orders were not reported for closed IV-D cases, voluntary orders and manual ledgers not on the automated child support system (ACSES).

The experimental and control group samples were selected on August 15, 1995. All cases in the control group were flagged by the ACSES to prevent their subsequent consideration for credit bureau reporting. Cases in the experimental group were subject to the standard credit reporting process.

The first step in the reporting process involved the production of a report that itemized for each obligor all relevant court orders, the balance to be reported to the credit bureau reporting agencies, the total obligation and the monthly amount due. For child support orders that lacked current information on the ledger obligation, the ACSES calculated a monthly amount due (MAD) using a standardized procedure. In the absence of a recent payment amount that had been allocated to the obligation, the arrears on the ledger balance was divided by 24 and the quotient was taken to be the arrears for the monthly amount due.

County child support technicians were instructed to review the reports and make any necessary adjustments to the delinquency balance or monthly amounts due prior to their conveyance to obligors. Technicians had discretionary power to suppress a court order from being reported to the credit reporting agencies (CRAs). They also had power to correct mistaken information sent to the CRAs. Finally, technicians had the ability to establish or modify the monthly amount due that was automatically calculated by the ACSES.

On October 15, 1995, after a 30 days allotment for technician review, the obligor was sent a pre-referral notice informing him of his obligations and the impending report to the credit bureau reporting agencies. Obligors were told that they had 30 days in which to repay their child support obligation or negotiate a payment plan with their child support enforcement technician. Obligors also had the right to request an administrative review hearing or schedule a judicial hearing.

On December 3, 1995, obligors who failed to make necessary payments or negotiate a payment plan were reported to the credit bureau reporting agencies as delinquent. On a monthly basis, child support obligation information was extracted for referred court orders and sent to the CRAs. In situations where the obligor was current with child support payments, the obligor's status was coded as "current." Unlocated obligors were reported to the CRAs as "not located" and were assumed to be both delinquent and not located. For delinquent obligors, credit reports reflected various amounts owed, including terms, balance and past due balance.

To measure the effect of the credit bureau intervention, the evaluators extracted pertinent information on the two groups of cases from the automated child support system. In addition to standard case characteristics, we extracted information on the payment status of each case at several discrete time points: upon generation of the sample, upon notification of obligors in the experimental group that they would be reported to credit bureau reporting agencies; upon report to the agencies; and nine and sixteen months following report. We also extracted payment behavior for obligors in the two groups of cases during time periods that preceded and followed the implementation of credit bureau reporting. Payment patterns prior to credit bureau reporting were captured during the period October 15, 1994 - August 14, 1995. Payment patterns following credit bureau reporting were captured during the period October 15, 1995 - August 14, 1996. Longer term payment patterns were elicited for the period August 15, 1996 - March 15, 1997. All extracted information was analyzed using the latest version of SPSS and patterns for the experimental versus the control group were compared.

The qualitative component of the evaluation consisted of the administration of a written questionnaire to child support technicians and supervisors at the county level in June 1996 and the conduct of face-to-face or telephone interviews with individuals in the child support and lending professions during the autumn of 1996. A total of 88 child support technicians and 28 administrators completed the written, child support questionnaire. Twenty-five individuals participated in in-depth interviews.

The written questionnaire included 20 items dealing with credit bureau reporting and its perceived value. Child support personnel were asked to agree or disagree with statements concerning the perceived utility of the intervention, its impact on child support collections, and its impact on the workload experienced by technicians. They were also asked to assess the quality of the training they had received on the initiative, the support provided by state child support personnel and the accuracy of the obligation information generated by the ACSES.

In face-to-face and telephone formats, child support technicians at the state and county level were also asked to assess the intervention. They were specifically queried on their initial expectations, implementation problems, reactions by obligors, workload impacts, impact on child support payments, and suggestions for improvement. Representatives of major lending institutions, subcontractors that assist lenders and credit bureau reporting agencies were also interviewed. They too were queried about initial expectations, reactions, concerns and impacts on loan activity.

The following presents the results of these investigations. We begin with the quantitative analyses and discuss the impact of credit bureau reporting on collections. This is followed by the qualitative assessments: reactions to the initiative by various professional groups, areas of concern, and suggestions for improvement.

Generation of Experimental and Control Samples

The assessment of the impact of credit bureau reporting on child support payment activity involved the collection of payment information for reported and non-reported cases both prior to and following the initiation of the intervention. Two state-wide random samples were generated consisting of 3,000 obligors, respectively. Cases in the experimental were subjected to the standard credit bureau reporting procedure while control group cases were excluded from reporting.

The SEU and the project evaluators grappled with several threats to the evaluation design in the early months of the study. Most were successfully overcome. For example, reporting did not initially occur for 568 obligors in the experimental group because there was some difficulty or confusion about at least one of the relevant court orders. After some deliberation, it was decided to report these cases for the orders for which there was clarity rather than exclude them from the analysis along with all comparable cases with multiple orders in the control group. Ultimately, cases with financial holds were reported although on a slightly different time frame. This maximized the size of the samples available for both experimental and control group treatment.

A second potential problem that was successfully addressed was the inability to identify administrative activity and administrative review outcomes in cases exposed to credit reporting. Prior to December 8, 1996, there was no way to record administrative hearing activity on the ACSES. All hearing activity was maintained by county technicians using manual techniques. Effective December 8, 1996, a computer screen was available to record administrative hearing activity and child support technicians were instructed on its use.

Still a third threat to the integrity of the sample that was overcome was the simultaneous introduction of a second enforcement remedy: driver's license suspension. Pursuant to legislation enacted by the Colorado legislature in 1995, child support enforcement agencies were given the authority to report delinquent obligors to the Department of Motor Vehicles (DMV) for license suspension. Although the automated process for notifying obligors of an impending report to the DMV was not scheduled to become effective until July 1996, authorization for the procedure went into effect in July 1995. Following several newspaper articles about the initiative and calls from custodial parents requesting activation of license suspension in their cases, some county child support administrators began to undertake notification and reporting on a manual basis.

To preserve the ability to assess the distinct impact of credit bureau reporting on collection activity, all county child support administrators were asked to refrain from DMV reporting in cases in the credit bureau reporting samples. If manual notification and reporting was performed, they were asked to record the date of DMV notification on the ACSES. In this manner, the evaluators were able to capture the incidence of simultaneous enforcement activity involving reporting to both DMV and CRAs in cases in both the experimental and control groups.

There were two ways in which the experimental and control group cases were subject to unavoidable differences in treatment. One was the exposure of the experimental group but not the control group to qualification procedures for credit bureau reporting. Although the experimental and control groups were selected in an identical manner using random techniques, 181 cases in the experimental group were removed from the reporting process following subsequent case review by technicians and the use of additional qualification procedures. The reasons for nonreporting included: the absence of a ledger, a low or zero debt balance, case closure, and absence of an applicable court order. Since the control group was invisible to technicians, it was not possible to subject it to identical qualification procedures. Inevitably, the exposure of experimental, but not control group cases to review by technicians, resulted in some differences in the two samples.

The second way in which the two groups differed was in the amount of information available for each group of cases. Pursuant to the reporting procedure, all experimental group cases experienced the generation of a computer-created calculation of the monthly amount due. All experimental group cases were also subject to monthly calculations of amounts due and past due balances. Since comparable information was not captured for cases in the control group, the analysis was limited to the reduced set of data on the ACSES available to both groups of cases.

Characteristics of the Two Samples

As Table 1 shows, cases in the experimental and control groups were identical in many key respects. An identical proportion (98%) involved a single obligor. Similarly, identical proportions of cases involved a single child (58%), two children (28%) or three or more children (14%). About half (52%) of the cases in each sample were AFDC, 29 percent were non-AFDC, and 19 percent were both AFDC and non-AFDC. Approximately 17 percent of both samples involved cases where the custodial parent and obligor were still married, and 40 percent involved divorced and never-married parents, respectively. Only eight percent of the cases in the control group and nine percent of cases in the experimental group were simultaneously reported to DMV for driver's license suspension.

Some differences in the two samples, however, did emerge as a result of the qualification process and technician intervention. Table 2 compares selected characteristics of cases in the experimental and control groups at five points in time: at sample generation; 30 days after technicians had received reports on obligors in the experimental group and had reviewed their ledgers for accuracy; 30 days after obligors had been notified that they would be reported to the credit bureau reporting agencies as delinquent; and nine and 16 months following report of delinquent obligors in the experimental group to credit bureau reporting agencies, respectively. At each time point, we examined the proportion of orders in different payment, partial payment and non-payment categories, the average monthly support order and the number of orders in the samples.

This comparison revealed that the two groups were comparable at sample selection but became significantly different following technician review. At sample generation, the proportion of cases in various payment categories was roughly the same with about a quarter of each sample in a full payment category (Category 1), a quarter in a partial payment category (Category 2), and half in a non-payment category (Category 3).

By October 14, 1995, however, the two groups were significantly different with respect to the distribution of cases in different categories. The experimental group had a significantly lower proportion of cases in Category 1, which meant regular payment of child support. It also had a higher proportion of cases in Categories 2 and 3, which meant that child support orders had been established but no payments were being made or that an order was in effect but the obligor could not be located. Compared with the control group, a higher proportion of cases in the experimental group also appears to have been closed between initial selection and notification.

These differences persisted when obligors were reported to credit reporting agencies on December 3, 1995, but by August 14, 1996 and March 15, 1997, the two groups were once again virtually identical. Most of these temporary differences were probably due to the delayed action taken on 568 cases in the experimental group with financial holds. It will be recalled that these cases had at least one court order for which there was some difficulty or confusion. Although these cases were ultimately reported to credit bureau reporting agencies, the report did not occur until February 1996. Thus, they are captured in a snapshot of cases taken on August 14, 1996, but not before this date.

Another possible reason for differences soon after technician review was clean-up activity initiated by technicians for cases in the experimental group but not in the control group. In August 1995, child support technicians were sent a report listing cases in the experimental group along with relevant debt information. They were given no information on cases in the control group. Technicians were asked to review these ledger balances for accuracy before the obligor was notified and subsequently reported to the credit reporting agency. In the course of reviewing cases in the experimental group, some were probably closed because inspection revealed that the child was emancipated, the arrears had been paid in full, or the custodial parent had requested case closure. Another probable outcome of the review was the movement of partial payment cases from Category 1 to Category 2. Technicians may also have spotted in Category 1, some arrears-only cases where payment had stopped and moved these into other categories. Finally, some cases with multiple enforcing counties may have been inappropriately classified as located in all counties (Category 2) when an obligor had only been temporarily located in one county. As a result of technician review, some of these cases may have been re-classified as falling into Category 3.

Impact on Payments

Table 3 presents payment patterns associated with credit bureau reporting. Payment behavior is monitored during nine months preceding credit bureau reporting (October 15, 1994 - August 14, 1995). Payment behavior following the credit bureau reporting intervention is monitored during an identical nine month period of time: October 15, 1995 - August 14, 1996. In addition, we monitored longer term payment effects during August 15, 1996 - March 15, 1997. Lump sum payments due to state and federal tax intercepts, attachments of worker's compensation or unemployment benefits and lottery winnings are tracked separately for cases in both treatment categories.

The analysis reveals that child support payments rose in both the experimental and control groups of cases following the credit bureau intervention, but that they rose more steeply in the experimental group. In the first post-intervention period (10/15/95 - 8/14/96), payments were 20 percent higher than they had been in the pre-intervention period for cases in the control group. For the experimental group, payments increased by 26 percent. This six percent difference in increase could not be attributed to other lump sum payments. Since the incidence of reporting to the DMV was low and identical for the two groups, it could not be attributed to the effects of license suspension. In the absence of any alternative explanation for the rise in payment activity in the experimental group, we conclude that payment increased by six percent in the nine months after notification letters were sent to noncustodial parents.

The dollar value of this increase came to $150,000. If this increase is evenly attributed to the 3,000 cases originally selected to be in the experimental sample, it translates into a payment gain of $50 per obligor. By applying this anticipated yield to the 61,139 child support cases in Colorado that meet the criteria for credit bureau reporting, the projected revenue due to credit bureau reporting within a nine month time frame comes to $3,056,950. This figure is very close to the $3.5 million dollar revenue gain projected for credit bureau reporting by the architects of the intervention.

These patterns persisted during a longer term follow-up period, August 15, 1996 - March 15, 1997. When we adjusted the payment data for those seven months to reflect estimated payments over an identical ten month period of time, we found virtually identical patterns of payment for the experimental and control groups. Because the calendar months covered in the assessment did not include the tax intercept season, the level of lump sums for the experimental and control groups were much lower than they had been in the earlier post credit bureau reporting period. Overall, payments for the experimental group were 6 percent higher than for cases in the control group. This differential was identical to the difference in payment detected in the earlier post credit bureau reporting evaluation period (October 15, 1995 - August 14, 1996).

There is little empirical data from other states against which Colorado's experiences may be compared. California officials attributed an annual increase in collections of about eight percentage points to credit bureau reporting. Following an 11-month study of reporting obligors with delinquencies of at least $1000 in Marion County, Indiana, researchers reported that the percentage of current child support payments and arrearages collected were 15 and 16 percent higher, respectively.

Like other states, Colorado's evaluation reveals that the short-term impact of credit bureau reporting is relatively modest, at least during a 17 month study period. It is anticipated that the main benefits of credit bureau reporting will show up over time as creditors deny credit to delinquent noncustodial parents. Naturally, confirmation of this expectation will require the conduct of a longitudinal study over a longer period of time to measure delayed effects.

The reactions of child support and lending professionals toward credit bureau reporting tend to be very favorable although few see it as a technique that has generated a good deal of child support money so far. Nor has implementation of the initiative been as easy as originally anticipated. A GAO report on the experiences of eleven other states with credit bureau reporting systems noted that start-up costs had been modest and that once implemented, ongoing costs were minimal (GAO, 1994). Colorado has found implementation of the credit bureau reporting initiative to be fairly challenging. This section of the evaluation report presents the reactions of key groups of professionals: program architects, computer programmers and state level child support personnel who implement the initiative; child support enforcement staff at the county level; representatives of the lending community and credit reporting agencies.

State-Level Child Support Personnel

In 1993, motivated by the successful initiation of mass case processing techniques in Massachusetts, state child-support personnel decided to attempt to expand the SEU and undertake a variety of aggressive, state-wide, automated enforcement activities. These efforts were contained in S.B. 141 which called for automatic credit reporting, the attachment of worker's compensation benefits and the streamlined attachment of unemployment compensation benefits. The bill passed the very year it was introduced--1994. In addition to authorizing these enforcement remedies, the measure increased staffing of the SEU by 3.5. The measure was required to be revenue neutral. In the fiscal note to the bill, program architects projected that the credit bureau intervention would generate 3.5 million dollars from the State's backlog of 61,000 eligible child support cases.

Although program architects were attracted to universal reporting of obligors, as is the case in California, they adopted a more conservative approach calling for reporting of obligors with a delinquency of $500 or more. The legislative process coincided with the appearance of a number of journal articles on the abuses of credit reporting agencies. While many regard the report of paying obligors as helpful to the accumulation of a positive credit rating, it was decided to restrict the intervention to those with a delinquency rather than extend it to all noncustodial parents.

Colorado's approach, however, does allow for the report of positive payment activity with no adverse affect on credit ratings. Modeled after consumer debt, Colorado's automated child support system creates a monthly amount due comprised of the monthly support order and a portion of past due debt. Like installment payments on credit card debt, obligors who make regular payments have the opportunity to enjoy good credit and be reported as "Current" even though they have big debts. As one state administrator explains:

As long as you pay faithfully every month, you show up as "Current" and your credit rating is not adversely affected. It is kinder and more fair. It can give a person with a huge delinquency a chance to have access to credit.

Implementation of the credit bureau reporting initiative was complicated by several factors. Due to a hiring freeze and a protracted search process, the SEU was not able to retain staff until September 1995--more than a year after the legislative authorization. Although county workers were retained on a temporary basis during the interim to assist with implementation of the initiative, the process was greatly slowed. Detailed procedures were not distilled in a timely fashion and the computer programming process was postponed.

Once the programming began, it proved to be very extensive requiring the full-time efforts of two computer programmers for a full year. Aside from initial programming tasks, computer staff found that they had to change the manner in which reports are generated for CBRs. Originally designed to yield payment reports on individual obligors, the system is now programmed to generate reports for individual court orders. Since some obligors have multiple court orders and different payment histories for each order, this has proved to be the most efficient and accurate way to keep track of payment and obligation information.

The delays in staffing the initiative and clarifying procedures exacerbated the pressure the programmers felt to become operational. They worried about putting the initiative into effect without adequate testing. Newspaper reports on credit bureau reporting, calls from custodial parents requesting the initiation of reporting in their cases and the demands of county child support administrators to bring CBR on line more quickly did not help matters. State personnel and programmers recall that some clients and county technicians expected reporting to begin as soon as the legislation passed and were frustrated with the year-long time lag needed for implementation. They were also frustrated by the gradual reporting process that the state adopted. Because county technicians are asked to check information on arrears amounts that will be reported to the CRAs on each order, the SEU has restricted its sample of cases for credit reporting to 3,000 noncustodial parents per month unless a county requests that the pace of reporting be stepped up.

Ultimately, the first sample of cases was generated in August 1995. In October 1995, 3,000 noncustodial parents were notified that they would be reported to credit bureau reporting agencies. By October 1996, approximately 85 percent of Colorado's noncustodial parents had been reported.

Staff at the SEU are pleased with the intervention time frame since credit bureau reporting was never expected to be implemented until July 1995. In addition, they are pleased that they have been able to attach automatic reporting of delinquent obligors to the DMV (enacted by the legislature in July 1995) to the credit bureau reporting process. The license suspension process only entails consideration of whether an obligor is current or delinquent in his child support payment activity. The system utilizes the sophisticated selection processes and ledger manipulations associated with credit bureau reporting. Thus, it has been possible to "piggy back" the two interventions.

Personnel at the SEU feel that state-level staff experience the workload impact of credit bureau reporting most intensely. One SEU staff person is the central contact for obligors and mortgage companies throughout the nation who call to verify the child support obligations listed on credit reports. As more and more cases are reported to credit bureau reporting agencies, the volume of calls to the SEU rises. Lenders call to verify debt accuracy and to better understand it. Although these calls create extra work, SEU personnel view this as a unique public relations opportunity for the child support community.

Sometimes an obligor's credit history is good except for the child support arrearage. Lenders want to know what is going on with this applicant's child support debt. This is a critical opportunity to explain the initiative and our position that the state gets it money first.

Another goal of the initiative was to cultivate a reliable source of location information. Computer programmers at the state hoped that CRAs would notify the state when obligors were rejected because of inconsistent social security numbers or addresses. Although the error rate is less than one percent, child support personnel believe that these error reports would help them locate hard-to-reach obligors. Program architects also hoped that mortgage companies and other lenders processing loans would call the SEU if they noticed discrepancies between employer and address information supplied by the noncustodial parent on a loan application and information contained on the child support section of a credit report. This type of feedback has also failed to materialize, although SEU staff report that lenders will check the veracity of address and employer information held by the child support agency in the course of a conversation on a debt balance or payment history. While the SEU is pleased that the lending community is willing to assist the child support agency with location information, they are frustrated by the growing industry practice of retaining subcontractors to verify child support debt. Subcontractors refuse to identify the lending institution to the SEU contact person. This precludes the opportunity to contact lenders directly and verify key employer and residence information.

In a related fashion, SEU staff have been only partially successful in convincing lenders to include child support debt into a consumer or house loan. One barrier to more effective negotiations with loan companies is the growing use of subcontractors to verify information contained on loan applications. These firms insulate child support staff from banks and other institutions that make loan and limit the opportunities for negotiation. Where the SEU has dealt directly with the lenders, they have found strong support for child support debt and willingness on the part of some to incorporate it into home improvement or debt consolidation loans. SEU staff report that other lenders encourage clients to pay their child support debt and/or deny loans because of child support debt. As one SEU staffer observed:

Lending institutions are very cooperative. If they can, they will incorporate child support debt into a loan. If they can't, they will deny mortgage loans because of the debt. They recognize child support as a viable debt. They listen to our explanation of what the program is about.

One concern that SEU staff have about the initiative is the high degree of discretion county agencies have in categorizing the payment behavior of noncustodial parents reported to CBRs as either "current" or "delinquent." Although Colorado's initiative is designed to permit obligors with large debts to be rated favorably as long as they are making payment, this view is not uniformly adopted across the counties. State personnel report that some county child support administrators believe that an obligor with outstanding debt is delinquent even if he is making monthly child support payments through a wage assignment. Accordingly, obligors with similar debt and payment patterns may be reported differently to CBRs in different counties with some receiving a favorable rating and others being reported as delinquent. As one state worker put it: "There is a lot of discretion in what cases get reported and how cases are reported as current or delinquent."

A second area of concern is the need to review payment records for noncustodial parents prior to their notification and report to CBRs. Many child support technicians lack the time to routinely update monthly amounts due on automated ledger screens. Thus, the financial information generated on the ACSES and included in the pre-notification report sent to the county child support technician may well contain errors that require remedy. These review requirements were not fully anticipated by state and county workers prior to the inception of the initiative. Nor were they addressed in the limited training provided to county child support workers before the initiative began.

County-Level Child Support Personnel

County child support personnel are divided on the utility and effectiveness of the credit reporting initiative. Supervisors and administrators tend to be somewhat more supportive than technicians in their assessments. In their responses to a mailed survey, nearly half of both groups (41%) characterized credit bureau reporting as a "useful tool". More than half of responding technicians (58%) and three quarters of supervisors/administrators (76%) believe that it is worth the effort. At the same time, relatively few report substantial, immediate financial benefits. Only 10 percent of technicians report that it "encourages lump sum settlements," 15 percent feel that it "brings in clients with repayment plans," and 5 percent say that it "generates a lot of child support money.

These sentiments are echoed in the comments of technicians and supervisors interviewed in September 1996, approximately nine months following notification of the first sample of noncustodial parents about the initiative. While most say that reporting is a "good idea", there is general consensus that the initiative "has not lived up to original expectations." County child support personnel had hoped that the initiative would generate many calls from self-employed or federally employed obligors who wanted to avoid a bad credit rating. Technicians who handle public assistance cases are least optimistic about the value of the initiative. They don't see absent parents in public assistance cases as being concerned with credit since many are unemployed or underemployed and have a "sense of futility" about their child support obligations. Others are just determined to avoid their responsibilities. As one county administrator put it:

This is outlaw territory. Absent parents have figured out how to avoid collection. . . .They stay in the woodwork.

The intervention is perceived to have the most positive impact on obligors who have been paying inconsistently, but have progressed to a better life style and want credit. As one technician explained:

We tell obligors who have an iffy credit rating that making regular monthly payments on arrearages will positively impact their credit rating. They generally accept the idea and are willing to negotiate a payment plan.

Technicians report mixed reactions from noncustodial parents. For example, one technician reported receiving three lump sum payments of $5,000, $8,000 and $11,000, respectively in the first month or two following notification. Others cited examples of obligors calling to initiate a monthly payment plan or a wage assignment. Several see the initiative as useful in generating location information when obligors call to complain about their notification letters. As one technician put it:

I get the location information first from the caller. He wants something from me and I want something from him. I request the location information before we continue to discuss his notification letter.

Overall, one third of the surveyed technicians and supervisors agree that the initiative "generates a lot of calls from obligors and one quarter of technicians and 40 percent of supervisors feel that the initiative helps to locate obligors.

Most technicians, however, report that the response from noncustodial parents has been very modest. Typically, technicians feel that most of the calls they get are from obligors who are making monthly payments and cannot understand why they have been reported. While complaint calls and requests for additional information suggest that the initiative has "got the attention" of some clients, it has not generated new money. Only a third of surveyed technicians believe that the intervention targets "appropriate obligors". They tend to view the intervention as reaching the ones already paying. As one technician explained:

Reporting paying clients generates tons of work, but no additional income. Techs are evaluated on how much child support income they generate.

Technicians are philosophically divided on the pros and cons of reporting absent parents who are making monthly child support payments. Some feel that it can be helpful in creating a good credit history and increase an obligor's chances for increased consumer credit. Others feel that it is unfair to subject those who pay to reporting because of the risk of errors and the failure of technicians to remedy outdated ledger sheet balances.

There is also quite a bit of disagreement about the workload implications of the initiative. Overall, 21 percent of surveyed technicians report that the initiative "creates a lot of extra work" although virtually no one reports requests for administrative hearings. One workload issue is the time required to review pre-notification reports and update ledger and balance amounts. As one technician observed, "Oh, another report. More work that I don't have time for."

While technicians are cognizant of the fact that the computerized ledger information is often inaccurate with only a quarter to one-third expressing confidence in their accuracy, many are ambivalent about spending time "cleaning inaccurate ledger balances" Rather than spend time "cleaning cases of obligors who are paying," they believe that their time is better spent "locating nonpaying obligors and establishing judgements and wage assignments." Still another perspective is that it is useful to refer paying obligors because many have arrearages that were not caught initially. Finally, some feel that it is counterproductive for technicians to be too proactive and that there is some benefit to making errors and inciting obligors to phone the agency:

There is no advantage to me to be proactive. If techs make things too easy for obligors, they will not be motivated to make arrears current or negotiate a payment schedule. Obligors faced with credit denial will be more amenable to a payment schedule.

Virtually all interviewed technicians and supervisors believe that the initiative makes credit bureau reporting much simpler than it used to be using manual techniques. They appreciate the automated approach and favor it over the paper intensive processes they used to follow.

The automated system is less time-consuming for county technicians and increases the time they have to spend on other enforcement work. It has had a positive impact on workload. State involvement has been real helpful with program implementation.

At the same time, there is some confusion about the division of labor between state and county-level workers and the quality of state-level performance. For example, although the state is supposed to be fielding calls from lenders and obligors and negotiating payment agreements, some county level personnel report receiving many calls from obligors and lenders. As she put it:

The state may say that they are fielding calls from obligors who are pursuing loans, but the obligors still contact me, not the state. Mortgage lenders advise obligors that they must contact their technician to make child support delinquencies current.

Part of the confusion may be due to the fact that CBR training was conducted before the computer programming for the initiative was completed. Many view this training episode as ineffective, leaving many technicians unclear about the process and the screens. Although relatively few maintain that they need additional training (17%), only about half of the surveyed technicians report understanding "how to adjust ledgers" or "post a MAD" (64%). More critically, only 22 percent of responding technicians felt that the ACSES screens were "understandable" and 36 percent felt that the procedures for reporting were "clear." With hindsight, state and county technicians both concede that more communication would have been helpful. As one state official put it:

We should have done a public relations blitz to the counties about what they could expect from the initiative. We should have explained the initiative to give them a better understanding of what they should expect and why they should work the pre-notification report. A more adequate initial training program may have changed some negative mindsets. The initial training session was held before the program was in place. Everything was done in the abstract. We didn't have the state capacity to do a public relations effort up front. Certain initiatives need outreach for buy-in from the counties.

Lending Industry Personnel

Lending industry personnel welcome the credit reporting initiative. All interviewed representatives of lending institutions and credit reporting agencies are sympathetic to the payment of child support obligations. As a worker who compiles credit reports for mortgage lending companies observed:

The credit reporting initiative is a good idea. Child support is a legal obligation that the individual has agreed to and has signed a contract. It is the same as student loans. Students who default on loans cannot get any kind of government mortgage. The same rule applies to applicants who are delinquent in their child support obligations.

Lending industry personnel say that they take child support debt seriously and treat it the same as any other type of debt. If an obligor is delinquent, the loan will not be approved until the total amount is paid or a monthly payment plan is established. There is general acknowledgement that child support debt can "mess up a loan, but be good for the person receiving child support."

Industry representatives maintain that they always requested applicants to disclose their child support responsibilities. When disclosed, child support debt was incorporated in their calculation of income-to-debt ratios. Despite this policy, however, lenders doubt that many delinquent obligors were honest about their child support obligations. As one loan officer put it:

Fewer and fewer applicants are listing child support obligations. Some applicants say they know about the credit reporting initiative and that they are making support payments. But disclosure is rare. Most of them deny child support obligations no matter what the circumstances.

The reporting intervention is regarded as a useful way to give lenders information not readily disclosed by applicants. It is perceived to be a helpful way to better assess the credit worthiness of an applicant. Not only does the intervention surface child support debt, it often sensitizes lenders to other credit problems.

I think it is wonderful...Borrowers should have to pay child support before receiving a mortgage loan. Frequently borrowers with child support debts have other collections or public records.

Lenders see the initiative working best with two groups. One consists of clients with a good employment history and responsible credit behavior in areas other than child support. Credit reporting will be effective because these applicants will want to buy a home and will be motivated to keep favorable credit ratings when they realize that they can no longer avoid their child support responsibilities.

Another potentially responsive group of parents are those who pay intermittently or may have late payment problems that require ledger corrections. They are disorganized people with many financial issues and have difficulty dealing with bureaucracies to solve errors and problems. Through credit reporting, they may contact the child support agency and be induced to get back on track.

Another aspect of the intervention that is appreciated by lenders is having a centralized contact person at the state child support agency to verify the accuracy of child support debt. They find this approach more efficient than having to deal with staff in Colorado's many counties.

The chief concern with the initiative expressed by industry representatives is the possibility of erroneous debt information being reported to credit reporting agencies. While lenders call to verify debt balances when processing a loan application and are willing to report inaccuracies they discover to the credit reporting agencies, it takes these firms six weeks to two months to recalculate an individual's credit rating. Inaccurate reporting can unfairly handicap an obligor who seeks a loan. All lending personnel stress the importance of accuracy in reporting payment behaviors and debt balances.

Another concern is lack of communication with the lending industry about the initiative and the procedures and policies entailed in reporting child support debt. Child support agencies need to explain how and why they define an obligor as "current" or "delinquent", particularly in cases where they have big debts but are making monthly payments in a timely manner.

A comparison of equivalent child support cases with and without credit bureau reporting reveals that the intervention increases child support payments by about six percent over a nine month period of time. This differential remained constant at sixteen months following credit bureau reporting. Generalized to the 61,139 child support cases in Colorado expected to meet the criteria for credit bureau reporting, the intervention will produce approximately $3,056,950 during the first sixteen months following implementation. Most program architects and lending industry professionals expect that the financial benefits of credit bureau reporting will be realized over time as delinquent obligors seek loans and discover that they must pay their child support debt before getting approved. Thus, without longer-term research, it is impossible to gauge the full financial benefits of the intervention.

Other states can learn several lessons from Colorado's experience with credit bureau reporting. One is the disagreement occasioned by the report of paying obligors. The child support community in Colorado is divided on the pros and cons of reporting obligors who are making regular child support payments. Although some feel that this gives paying obligors the chance to be reported as "current" and enjoy the benefits of a positive credit rating, others believe that it alarms many paying obligors without generating additional revenue and exposes them to the risk of erroneous reports of payment balances. While some technicians maintain that these reports and the technician review associated with them have revealed arrearages that were not caught earlier on, other technicians insist that reporting paying obligors has generated many phone calls and extra work that has not led to new money and has detracted from other work duties.

Another area of concern is the amount of discretion county child support personnel have in determining the payment status of an obligor. County administrators have a great deal of autonomy in how they deal with payments made via wage assignment, and whether they label obligors with debts who are making payments as current or delinquent. Differential treatment of identical payment and debt scenarios in different county settings may threaten the integrity of the intervention.

Like other child support enforcement remedies, credit bureau reporting appears to have little immediate impact on AFDC clients and those who are intent on avoiding their child support responsibilities. At least during the first nine months of implementation, the intervention generates the most calls from "average Joes," who are making payments and can't understand why they have been reported. It may have motivated intermittent payers and others who have been making small payments to "take more action in order to be approved for a loan." Nevertheless, everyone is optimistic that over "a long haul," the intervention will touch a broader base of obligors and lead to more collections over time.

While the lending industry has been extremely supportive of the initiative, the increased reliance on subcontractors to investigate loan applications creates some problems for child support agencies that would like more direct access to lenders. Subcontractors refuse to identify the lending institution they represent to the child support agency. This minimizes the agency's opportunity to explain the initiative to the lender and explore the feasibility of incorporating child support debt into a mortgage or consumer loan.

One hoped-for outcome of the intervention that has not materialized is the generation of new location information for obligors. Credit bureau reporting agencies do not identify to the child support agencies the reported obligors whose addresses or social security numbers do not match in the course of the reporting process. Although this number is very small, estimated to be less than one percent, these obligors go unreported and are not targeted for more intensive location efforts. The more widespread reliance on subcontractors for investigation of loan applications has also limited the utility of the initiative in generating updated employer and address information.

Finally, reactions to the intervention reflect the need for additional training and communication between state and county level child support workers. Credit bureau reporting places the SEU in a key role in the enforcement process. While many county child support personnel appreciate the automated process and the role played by SEU staff, there is lingering confusion about the workload impacts of the initiative and the division of labor between state and county workers. One unanticipated burden is the need to check the accuracy of automated debt information reported to CRAs. In retrospect, state personnel wish that they had done more of a "public relations blitz" to the counties about the initiative, workload requirements, expectations and time frames. In the absence of state-initiated outreach at the early stages of the initiative, county "buy-in" was less than wholehearted and some misunderstandings occurred.

Field, Anita DeVargas, "A Study to Evaluate the Current Structure and Efficiency of the Credit Reporting Project of the State Department of Human Services, Child Support Enforcement Division," presented to the Credit Reporting Task Force Committee, October 28, 1996.

General Accounting Office, "Child Support Enforcement: Credit Bureau Reporting Shows Promise," Report to the Honorable Sander M. Levin, House of Representatives, GAO/HEHS-94-175, June 1994.