Child Care Best Practices Report
CalWORKs, Stage 2 and 3 and Alternative Payment
Best Practices for Program Integrity Report
Required by Chapter 229, Statutes of 2004
(SB #1104, Committee on Budget and Fiscal Review)
Table of Contents
- Department of Social Services Survey
Notification Cover Letter
Child Care Survey
- Summary From Counties and AP Agencies
Providing Definitions of Fraud and Fraud Indicators
To collect information from other states, as required by SB #1104, the Department consulted publications made available by the Administration for Children and Families (ACF), the NCCIC, and the General Accounting Office (GAO).
To assist in gathering input for this report, the Department of Social Services (DSS) mailed a request for written input to the California Welfare Directors Association (see Appendix 1). In response to this request, 42 counties submitted written input to the Department regarding best practices, definitions of child care fraud, and indicators used locally to identify potentially fraudulent cases. Depending on the county, the written input was prepared either by the county welfare department, the local Alternative Payment (AP) agency, or both agencies in concert. Because the input did not identify which local agency was the author, input provided in response to the DSS letter will be referred to by its county of origin or as input from local agencies.
Input was also solicited and received from the following California organizations: The California Alternative Payment Program Association (CAPPA), the California Child Development Administrators Association (CCDAA), the Child Care Law Center (CCLC), and the Child Care Alliance of Los Angeles. Meetings were held to solicit input with CAPPA and the Child Care Alliance of Los Angeles.
Department staff also met with representatives of the state's public assistance fraud investigators on two occasions to solicit input. These investigators work in the state's Special Investigative Units (SIUs) and formed a subgroup for the purpose of presenting input, the Childcare Fraud Investigators Network. Extensive written input on best practices and fraud prevention and detection strategies was received from the Childcare Fraud Investigators Network.
Extensive input was also received from Sacramento County. The input was prepared jointly by Child Action, the Sacramento County Department of Social Services, and representatives of Sacramento County's Special Investigative Unit. In addition, Department staff met with representatives from Child Action for several hours.
Staff from DSS assisted in this report through consultation, by organizing input from county departments of social services and welfare fraud investigators, and by providing copies of previous communications between DSS and local agencies. We have included the letter used by DSS to solicit input from county departments of social services as an attachment (see Appendix 1).
Consistent with input we received from Sacramento County and the Child Care Alliance of Los Angeles, the body of this report is divided into two topic areas: internal program integrity, defined as fraud or errors caused by local agency employees in program administration; and external program integrity, defined as fraud or errors caused by parents and/or providers. Within each of these two topic areas, fraud, i.e., the intent to obtain an unlawful benefit, is distinguished from errors, i.e., mistakes in reporting or administration that result in an overpayment or underpayment. Each topic area contains a summary of the input followed by recommendations received by the Department for future action or statutory or regulatory change.
This report is followed by tables that summarize data provided by county departments of social services (and by AP agencies jointly reporting) regarding definitions of fraud and the "indicators" used to detect fraud (see Appendix 2). Of the 42 counties providing input for this report, 12 provided information regarding their definitions of child care fraud and 21 provided descriptions of fraud indicators. In reviewing these data, it is important to note that information regarding definitions of fraud and fraud indicators were not specifically requested. Therefore, not all counties reported this data and those counties that did report may not have reported complete information. The absence of information in Appendix 2 should not be used to infer that counties and AP agencies are not using definitions of fraud or fraud indicators. The information in Appendix 2 is intended only to summarize information about fraud indicators in instances where these data were reported to the Department.
California's Alternative Payment Programs began as pilot programs in the late 1970. Their mission was to develop less expensive, easily expandable, child care for low-income families. The programs were enacted permanently into law by Chapter 798, Statutes of 1980 (SB #863, Sieroty). The enabling legislation for Alternative Payment (AP) programs contained little in the way of statewide operational standards or procedural requirements. Consistent with their origins, AP agencies were given flexibility to develop policies and procedures that promoted the delivery of services in their local environment. Even after the implementation of CalWORKs child care, AP agencies retained much of their early flexibility to develop unique operational policies and procedures.
Likewise, California's welfare reform program, CalWORKs, was created largely as a block-grant program with outcome measures. Many operational standards and procedures, especially for supportive services such as child care, were left to the discretion of local agencies.
Summarizing input from a variety of sources involves finding common themes or principles around which specific recommendations can be organized. However, focusing on the elements common to many agencies may result in over-emphasizing what is similar. In contrast, what emerges from the data in this report is a description of differences in operating standards and procedures among local agencies.
Differences in local agency operational standards are not necessarily causes for concern. For the most part, they are consistent with the historical flexibility provided local agencies. If they are an effective local response to a local problem, differences in programs' operations may also encourage program integrity.
The best practices reported by agencies to minimize internal error all involved systems for conducting reviews of active cases. All agencies reported a system for supervisor review of new cases or changes in existing cases. Larger agencies, including agencies in Los Angeles, Sacramento, Fresno, San Diego, and San Francisco counties, reported the existence of a quality assurance unit whose duties involved the regular, random sampling and analysis of case files and/or the review of suspicious files to determine if proper procedures were followed in determining eligibility, need, and provider payment. Some local agencies reported that this quality assurance unit reported directly to the executive or the board to ensure their independence.
Smaller agencies reported other systems for case review. Stanislaus County referred to monthly reviews of randomly selected cases. Calaveras, Tuolumne, Humboldt, Santa Barbara, and Ventura counties mentioned periodic caseload quality control audits. Other small counties mentioned engaging in regular "self-reviews" in which workers would exchange and review one another's cases.
Other suggestions for minimizing internal error involved the use of computer programs to limit the potential for operational mistakes. For example, computers can be programmed to deny payment when a child reaches thirteen years of age, or to recognize when a child's age should cause the agency to change ceiling categories (e.g., preschool vs. school-age).
Internal fraud involves the use of position or authority by an agency employee to obtain benefits or funds to which they are not entitled. Of the agencies that administer CalWORKs Stages 2 and 3 and AP programs, approximately 15 are county departments of social services, 16 are school districts, and 58 are private non-profit corporations. Because of the prevalence of private, nonprofit corporations, many CDE contractors recognize the need to establish controls to ensure integrity of internal operations. Extensive input on the subject of internal fraud was received from Child Action in Sacramento and the Child Care Alliance of Los Angeles.
Both emphasized the importance of a formal program of staff training that emphasized the ethical standards of the agency, the right of the agency to inspect work areas, the obligation of employees to report suspicious circumstances, and the consequences for fraudulent activities. Child Action recommended the creation of a training unit whose responsibilities include committing all agency policies and procedures to writing and training that inculcates agency standards and philosophy in all staff.
Other recommendations included:
- Hiring practices involving background screenings on all employees for prior convictions, drug testing, and validation of social security number.
- Prohibiting conflicts of interest. For example, employees cannot have relatives, roommates, or other related parties who are recipients of services or contracting with the agency as providers unless all such relationships are declared to the agency; and all employee-related cases must be processed by a supervisor or unit unconnected to the employee and reporting directly to the executive or the board.
- Requiring employees to declare all business activities or employment outside of the agency.
- Prohibiting relatives of providers, either licensed or exempt, from working in the accounting or payment processing functions of the agency.
Agencies report enforcing these requirements by using procedures, such as computer database matches, to compare all employee names with the emergency names and phone numbers listed in parent and provider files. Employees can also be asked to sign releases to permit the agency to verify outside income reported to the Employment Development Department or the Franchise Tax Board.
All agencies commenting on this subject thought it important to separate employee duties and functions within the agency. The extent to which duties can be segregated depends on the size of the agency. Child Action reported the following separation of duties:
- Scheduling Unit. This unit scheduled intake interviews with eligibility workers. Interviews could not be scheduled by other employees. The eligibility workers and families assigned to each "scheduler" are rotated regularly.
- Eligibility Unit. These employees interviewed the family and verified eligibility, subject to review by the supervisor.
- Provider Unit. These employees set up contracts with providers based on the needs of the family for care and the provider's rates.
- Payment Unit. These employees generate payments based on the provider's sign in/out sheets and invoice for payment.
Agencies report that computer systems were a necessary tool in separating duties and ensuring internal integrity. All staff had passwords that would only allow them to access the portion of agency operations over which they had responsibility. All computer records indicate the source of information by logon password and date of entry. Computer systems can also be programmed to disallow data entry into closed cases or employee-related cases without two passwords, one of which must be the agency executive.
Sacramento County recommended that standards for annual financial audits be strengthened to include an audit of program errors. Annual financial audits are currently required of all CDE contractors.
The Childcare Fraud Investigators Network provided input based on its perception that insufficient information is provided to parents and providers regarding their obligation to report changes affecting eligibility or reimbursement. Consequently, the Network recommended systems to provide clear, comprehensive information to parents and providers regarding their reporting responsibilities and the consequences of failing to report or inaccurate reporting.
In contrast, all agencies responding to the request for input indicated that there are processes and procedures in place for informing parents and providers of program requirements and the responsibilities for reporting. The extent of such informing varies from agency to agency. Some agencies report distribution of written materials, e.g., parent and provider handbooks. Others report more intensive face-to-face discussions. All agencies report asking for an acknowledgement, in writing, from parents and providers that they understand the reporting responsibilities.Three small, rural agencies reported program practices in this area that differ from typical responses. Yuba County reported that fraud investigators conduct an in-person, home visit with all new parents and providers upon enrollment of the parent in Stage 1. During this visit issues such as proper time accounting (sign in/out sheets) and the rights and responsibilities of both parents and providers are discussed. Information contained in the case file for the parent and provider is verified, and questions from parents and providers are answered. Yuba County reported that this practice resulted in many providers becoming colleagues in reporting suspicious circumstances to the local agency. Likewise, Colusa County reported frequent contact with providers, including requiring providers to submit copies of receipts for parent fees with the claim form.
Mariposa described orientation information emphasizing that sign-in and sign-out sheets had to be completed by the parent on the day and at the time that children were dropped off and picked up from care. Deviation from this practice could result in a referral for further investigation. Even though completion of sign-in/out sheets by the parents at the time children are dropped off and picked up is a requirement of Title 5 regulations (see Section 18065), Yuba County and Mariposa County were the only agencies submitting written input that cited this practice as critical to program integrity.
Most local agencies did not report enforcing the daily completion of sign-in/out sheets. (Attendance sheets are completed at the end of the week or month, and may be retained by the parent instead of the provider.) However, many report comparing recorded attendance to certified need prior to issuing a payment. Significant discrepancies between recorded attendance and approved hours of care would result in follow-up contact with the parent and/or provider to resolve discrepancies. Discrepancies that cannot be satisfactorily resolved lead to further investigation or a fraud referral. In this category are, among others, Butte, Colusa, El Dorado, Shasta, Imperial, Kern, Lassen, Orange, San Bernardino, San Diego, Santa Cruz, Siskiyou, Tuolumne, Contra Costa, Fresno, Calaveras, Amador, Sonoma, and Stanislaus counties. Among these counties, some (Calaveras, Amador) also talk about matching signatures on the sign-in/out sheets to parent and provider signatures on other forms.
Other local agencies appear to accept claim forms from the provider - which may or may not indicate daily hours of care - in lieu of attendance sheets. These forms are signed by both parent and provider under penalty of perjury. In this group are San Mateo, San Francisco, and Santa Barbara counties.
Contra Costa County additionally reported including a section on the attendance claim form that allows parents to report changes in their need for child care. Contra Costa then uses this information from the parent to verify the requested change and update the family file.
Tulare and Ventura Counties report that each client is contacted monthly, the approved hours of care are updated, and the updated hours are compared with the attendance reported through the sign-in/out sheets submitted by the provider.Other local agencies report less frequent than monthly, but more often than annual, updates of family files. Calaveras and Contra Costa counties report requiring ongoing submission of employment or education/training information. Amador and Lassen counties report updating all family files every six months. Shasta reports random files selected for periodic updates of information. Several local agencies reported requiring more frequent updates for files in which the parent was engaged in education and training, including evidence that the parent completed the courses in which they were enrolled.
In concluding this section, we would observe that many of the practices described above that are intended to prevent parent and provider errors also reduce the incidence of parent and provider fraud.
Nearly all counties that submitted input mentioned using the Department of Social Services All County Letter No. 00-53, dated August 29, 2000, and the letter to all county welfare directors dated January 19, 2001, to define and detect child care fraud.
Even though most agencies refer to the definitions of fraud and fraud indicators contained in the correspondence from DSS, above, the definitions of fraud and fraud indicators reported to CDE by local agencies differed from one another and from the DSS letters. For example, some agencies appear to report fraud definitions, such as claiming reimbursement for services that were not provided, as fraud indicators.
Definitions of fraud are important because they provide insight into what agencies consider fraudulent behavior on the part of recipients and providers. Fraud indicators are important because they lead to agency procedures intended to deter fraud. For example, if the agency believes that a commute distance of more than 25 miles between the provider and parent is an indicator of fraud, the agency will solicit, in processing the request for child care, information to determine the distance between the parent and provider's home.
Where agencies provided specific examples of fraud definitions or indicators, we have aggregated them on the attached spreadsheets (see Appendix 2). These attachments are intended to provide a visual summary of the similarities and differences in local operations.
In addition to fraud definitions and indicators, agencies also cited best practices for agency operations. Many agencies cited the importance of developing strong, collaborative relationships between local child care agencies and law enforcement. San Diego County reported that all managers and supervisors participate in four hours of fraud prevention and detection training per year. Line staff receives two hours of training. Many agencies, including those in Sacramento, Ventura, Fresno, Kern, Orange, and Solano counties, reported systems for cross-training between local fraud investigators and child care agencies. Sacramento and Fresno counties cited the importance of being able to immediately refer all suspicious cases to a fraud investigator who could access databases (e.g., Employment Development Department, Franchise Tax Board, criminal histories, court databases, etc.) to assist in resolving questionable information.Nearly all local agencies reported that their public assistance fraud investigators would not investigate a Stage 2 or Stage 3 case unless there was a link to cash aid fraud or Stage 1 fraud. Only Colusa, Calaveras, Fresno, Contra Costa, Imperial, and Amador counties reported that local fraud investigators would accept Stage 2 or Stage 3 referrals.
The Childcare Fraud Investigators Network reported that at times Stage 2 and Stage 3 agencies will not provide requested records to them because of mistaken ideas concerning confidentiality. They recommended clarifying this issue to increase collaboration.
Many counties, among them Sacramento, Butte, Los Angeles, San Diego, and Shasta, reported the importance of the local department of social services and the local AP agency sharing databases and information. Several other counties, among them Ventura, stated that access to one another's computer systems was in development. Smaller counties, such as Colusa and Calaveras, reported the importance of regular meetings and case discussions. Some counties, such as Humboldt, Yuba, and Ventura, reported the benefit of having welfare-to-work staff and local AP staff co-located.
Nearly all counties reported difficulty in knowing whether a license-exempt provider in Stages 2 or 3 was also a cash-aid recipient. A few counties reported conducting database matches for this purpose. Many counties, including small counties, saw this as a problem. Lake County reported performing monthly comparisons, by hand, of all license-exempt providers receiving child care payments against all recipients of cash aid.
The input received from urban areas, such as the Los Angeles Child Care Alliance, relied heavily on computer databases and matches. Examples of best practices included computer programs to search and reveal matches on all parent and provider files for all agencies operating in a county. In addition to identifying duplicate services and payments, these matches would indicate when a parent was receiving child care services from one agency while working as a provider for another agency. Computers could also be programmed to immediately identify matches upon entry of new parent or provider information.
Child Action, from Sacramento County, and the Childcare Fraud Investigators Network also suggested the creation of a statewide database to perform such matches. Child Action believes that the data necessary to conduct such matches already exists in a database maintained by the Department and used for federal reporting purposes.
Information from ACF and NCCIC indicates that other states are pursuing computer data matches as a method of preventing and detecting fraud. On this subject, the ACF states in a working paper distributed to the State Partners Meeting on September 29, 2004: "States with a solid infrastructure, integrated and automated systems, and sufficient staffing stand the best chance of preventing fraud, as well as successfully pursuing fraud when it occurs."
However, federal reports on this subject also note the federal prohibition regarding the collection of social security numbers for families receiving child care funded through the federal Child Care Development Fund (CCDF). These same federal reports recognize that the lack of a unique identifier for parents is an impediment to effective database matches. Pursuant to federal law, social security numbers are not available in CDE's 801 reports for all recipients.
Other input offered by local agencies focused on the process for verifying employment of the recipient. The Childcare Fraud Investigators Network and the Child Care Alliance of Los Angeles suggested that the state consider requiring that all verification of employment be accomplished through direct contact between agency representatives and the purported employer. As a condition of eligibility, parents could sign releases allowing local agencies to contact employers directly to verify hours of employment and income. A few local agencies, including Solano and Sacramento counties, indicated that this was already standard procedure.
The Childcare Fraud Investigators Network stated that verification of employment in Stage 1 may "fall through the cracks" because responsibility for this function is not often clear when the county contracts with an AP for administration of Stage 1 services. They believe this is especially true regarding the periodic checking or updating of family files.
The Child Care Alliance of Los Angeles also suggested measures to confirm that the purported employer was in fact "legitimate." These suggestions included computer matches of the employer's address with all legitimate addresses used by the U.S. Postal Service.
Many agencies reported examining wage stubs, and periodically requiring that parents submit new wage stubs, to ensure that the information provided to the agency was consistent with information provided to the employer. Comparisons included areas such as marital status, number of dependents, and hours of work.
Many agencies, including Sacramento County, complained about vague state rules regarding how to verify eligibility and need for self-employed individuals. A few agencies reported requiring income declarations and business plans, updated monthly, as evidence of self-employment.
The Childcare Fraud Investigators Network was concerned about the lack of standards to verify the appropriate use of child care when a parent's work schedule is variable. While no specific suggestions were offered for best practices, the need for statewide uniformity was identified.The Child Care Alliance of Los Angeles, representing urban agencies, recommended several quality control mechanisms to prevent provider fraud that can be programmed into computer systems. For example, computers can prevent more than one payment from being issued for the same child in the same month. Agencies should require that providers indicate the address where care is being provided. These addresses can be matched against the U.S. Postal Service list of legitimate addresses. Computers can also prevent the issuance of a payment that exceeds the appropriate Regional Market Ceiling (with adjustments).
Riverside County suggested computer programs containing fixed "entry" dates for Stage 2 and "exit" dates for Stage 1 to prevent duplicate billings for the same month.
Many agencies reported requiring face-to-face interviews with providers to verify the provider's identity, their social security number (or FEIN number), and the address where care is provided. These practices are standard among CDE's AP programs. However, they may not be as prevalent for Stage 1 agencies.
Sacramento County expressed concern about the lack of statewide standards that ensure compliance with the Fair Labor Standards Act regarding reimbursement for in-home exempt providers. Suggestions involved documenting that the parent had complied with all necessary federal and state requirements to act as an employer prior to the agency issuing a payment.
Several agencies reported various practices intended to prevent providers from reporting "lost" reimbursement checks, only to have the original and the duplicate check subsequently cashed. These practices included only mailing checks to an address (not a P.O. Box), and printing "Do Not Cash After" dates on both the original and duplicate checks. Probably the most effective suggestion in this regard is requiring that all reimbursement checks be transmitted through direct deposit. Agencies using this practice have arrangements with local banking institutions to open accounts for those providers who do not have a bank account.
A few agencies, including Riverside County, reported contacting providers by phone to ensure that care was being provided during contracted hours. Two agencies, San Luis Obispo and Ventura counties, reported conducting site visits to providers to ensure that care was being provided during contracted hours.
All agencies reported having systems in place for incorporating actions by Community Care Licensing to suspend or revoke facility licenses into the agency's process to contract with and reimburse providers.
All agencies reported using DSS or CDE due process requirements for changes to the eligibility status or hours of child care provided to parents. Both CDE and DSS processes call for written notice to parents in advance of any changes and appeal rights if the parents disagree with the proposed change.No agencies commented on grievance or other dispute resolutions processes for providers, even though such processes are required by CDE regulations. CDE on-site reviews regularly examine local agency provider handbooks and other written materials to ensure that local agencies adhere to regulatory requirements. Sacramento County described a process of giving written notice to providers suspected of not complying with program rules and allowing providers to dispute those notices through a local hearing process.
Few agencies described processes for collecting overpayments from parents and providers, including the use of small claims court (for providers) and voluntary repayment (for parents). However, the majority expressed concerns similar to those of Los Angeles County. The Los Angeles County Department of Public Social Services (DPSS) stated that unless a case resulted in prosecution for fraud, the identification of fraud or overpayments carried no deterrent value because:
- There was no legal authority to collect an overpayment from a client, unless the client voluntarily agreed to the repayment.
- There is no authority to impose a consequence for program violations.
Many agencies have complained verbally about parents who were discontinued from Stage 2 or 3 for fraud, and the only consequence to the parent was that they reapplied for and were granted services in Stage 1.
Both Sacramento County and the Childcare Fraud Investigators Network recommended the enactment of statutory definitions of child care fraud, penalties for those who commit fraud, and a system for tracking and collecting overpayments. CAPPA also stated that any fraud definitions and penalties should be established to apply uniformly to all three Stages of CalWORKs child care. Sacramento County and the Childcare Fraud Investigators Network, among others, recommended additional funding for child care fraud investigators. Sacramento recommended that these additional positions be funded from overpayment recoveries. The Fraud Investigators Network suggested a system for tracking child care fraud separate from other forms of public assistance fraud.
Sacramento County, Childcare Fraud Investigators Network, CAPPA, CCDAA, the Child Care Alliance of Los Angeles, and many others expressed concern that current exemptions from trustline clearance allow children to be placed in at-risk situations. They recommended that state law be amended to require that all license-exempt providers receive trustline clearance to serve subsidized children.
San Francisco County, Sacramento County, and the Childcare Fraud Investigators Network recommended that state rules regarding documentation and verification of eligibility and need be clarified and strengthened, especially in the areas of self-employment, incapacity, and education and training.Sacramento County and others thought that the connection between In-Home Supportive Services (IHSS) and child care merits examination. They specifically recommended prohibiting a person from receiving child care if their employment was to provide IHSS services to a member of the same household. They also recommended prohibiting someone receiving IHSS services from also being paid as a child care provider.
The Childcare Fraud Investigators Network suggested that child care providers who live with the parent and child are a high risk area that merits examination. The Fraud Investigators Network suggested that certain relatives or other adults living in the same home (e.g., boyfriend, step-parent) should not receive reimbursement for caring for children they would otherwise care for in the absence of payment. Alternatively, they suggested a prohibition on reimbursing a provider who shares the same address as the parent.
The Childcare Fraud Investigators Network and others identified the parent's employment for the family day care provider who is providing child care for the parent's child as another high-risk area, because it places the provider in the dual role of provider and employer. They recommended that regulations be adopted prohibiting the reimbursement of the parent's employer for providing child care if the parent works in the child care facility.
Measuring Improper Payments in the Child Care Program: A Pilot Project of the ACF Child Care Bureau. (Working papers for discussion). Washington D.C.: U.S. Department of Health and Human Services, Administration for Children and Families, September 2004.
United States General Accounting Office, "TANF and Child Care Programs. HHS Lacks Adequate Information to Assess Risk and Assist States in Managing Improper Payments." Report to Congressional Requesters, June 2004.
Department of Social Services, "All-County Letter 00-53: CalWORKs Stage One Child Care Fraud and Overpayments," August 29, 2000.
Department of Social Services, letter to All County Welfare Directors, subject: "Potential Welfare Fraud and Child Care Fraud Indicators," January 19, 2001.