III. Implementing Work First:

Advice on Program Administration

Once the basic policies of a work first program have been set, it falls to administrators to put the actual program into place. To do that, they need to hire and train staff, determine how staff functions will be assigned, and clearly communicate the program's message to all staff and service providers. In addition, modifications to management information systems will be required to support the work of staff and provide administrators with the information they need to oversee the program. Administrators must also develop linkages with outside service providers who will work with program participants. Finally, administrators need to understand the costs of their program and to identify strategies to reduce costs. Part III of this guide (sections 14-22) discusses the trade-offs inherent in different management approaches, and provides advice for program managers as they tackle these various challenges in the context of work first.

14. Hiring and Training Staff

Hiring Staff

Staffing takes on added importance in work first programs because so much of the job involves encouraging and motivating participants in addition to the work of processing and monitoring participation. Most programs will make use of existing program staff as they shift to a work first approach. Many will also hire new staff. In assembling the program staff, administrators will need to work within local civil service rules, hiring limitations, and union agreements.

The following guidelines have been recommended by administrators in several work first programs. Most apply to staff at every level, from management to support staff.

• Look for staff who are committed to the program's philosophy and goals, and who believe the program can work.

• Look for staff who are outgoing, energetic, and enthusiastic and who can strike a balance between compassion and challenging participants. Also, look for staff who are supportive of participants and believe in their abilities.

• Look for staff who are organized and can manage a caseload without becoming overwhelmed by the administrative part of the job.

• Hire a diverse staff, who bring different backgrounds and expertise to the program. These might include:

• Staff from the private sector, because that is where most participants will be looking for jobs
• Staff with an employment or human resources background, who can bring expertise in job placement
• Staff with experience in marketing, public relations, or sales, especially for positions as job developers
• Staff who are experienced at making presentations or working with groups, especially for positions that involve facilitating job clubs or other group activities
• Staff with experience as eligibility workers, to bring an understanding of what the eligibility office does and to facilitate coordination between the two offices
• Staff from the organizations that will provide key services, because they know the culture and "language" of those programs
• Staff with specialized training or a social work background, who can serve as a resource for other staff members in working with hard-to-serve participants
• Former welfare recipients, who bring personal experience to the program and whose success can motivate both staff and participants
• Staff who reflect the diversity-in terms of gender, race, age, language, and the like-of the participants being served

Training Staff

Training for work first staff needs to focus both on the tasks required to do the job and on the program's overall philosophy and goals. At least one day of training should focus exclusively on the philosophy, to make sure that all staff are willing and able to promote that philosophy in their work. This will be less of an issue when new staff are hired who already support the program's goals, and more of an issue when current staff are asked to adapt to a new program philosophy. Program administrators may wish to provide training in the goals and operations of work first not only to their own staff but also to staff of outside service providers and partner agencies. In addition, administrators may wish to provide staff with training about other programs and benefits that might affect participants as they move to work, such as transitional benefits, the Earned Income Credit, and income-based housing subsidies.

It is often difficult to find the time and resources to devote to extensive staff training. Some programs have made use of free local resources-for example, by borrowing space at a community college or finding individuals with relevant expertise from community social service agencies or employment services to conduct training sessions. Training can also be done in stages, beginning with the program message and the most immediate skill needs, then reinforcing the message and adding other skills as the program develops. Remember, too, that you can use other forums besides formal training-such as daily supervision and encouraging diverse staff to learn from each other-to communicate key program messages and teach staff relevant skills.

Other suggestions for training include:

• Combine classroom training for new staff with on-the-job training.
• Provide regular, ongoing training, identifying training needs by getting suggestions from line staff and supervisors.
• Bring staff together regularly, without a formal training agenda, to discuss issues and to share ideas and best practices.
• Have managers and supervisors reinforce the main messages from training once staff are back at the office.
• Relate training to the program's goals and explain how it will help staff achieve those goals.
• Use training to bring information back to administrators, including trainees' suggestions for ways to improve the program and make the jobs of staff easier.
• Provide training on cultural diversity related to the social identities and characteristics of participants.
• If you do not have in-house expertise on a particular issue, bring in outside trainers (for example, someone from a staffing service to train job developers).
• Publicize the purpose of training sessions beforehand, so that staff and supervisors can make decisions about who should attend.
• When possible, schedule training for a variety of days in order to give staff options on when to attend.

15. Caseload Size and Specialization

The case manager is the most visible staff member in most work first programs, and is responsible for overseeing a participant's journey from welfare to work. How your program operates will therefore be shaped largely by the decisions you make about each case manager's caseload size and responsibilities. Some program administrators may choose not to implement a formal case management structure, instead having participants work with whichever staff member is available at a given time. However, most administrators find that maintaining regular caseloads allows staff to develop ongoing relationships with participants and more effectively motivate them, guide them, and monitor their progress.

Caseload Size

Caseloads in work first programs being studied in three sites as part of the JOBS Evaluation ranged from 95 to 120 cases per case manager. In general, smaller caseloads (100 or less) enable staff to provide more individualized support, to monitor participation more stringently, and to follow up more quickly on attendance and other problems. Smaller caseloads also afford more opportunity for case managers to use persuasion instead of relying on penalties to increase participation. In contrast, large caseloads (as high as 200 or more) make it possible to process a much larger share of the eligible total caseload through the intake points of the program. The trade-off is largely between breadth of participation and intensity of services.

How large caseloads can be without negatively affecting services also depends on other factors. For example, staff can manage larger caseloads in programs with strong computer systems, clerical and other supports, and specialized handling of some tasks, such as child care or attendance monitoring. If case managers do not have those supports, or if they do not have clear guidance on how to prioritize their work, they can be overwhelmed even with smaller caseloads. The American Public Welfare Association recommends that welfare-to-work program administrators consider the following questions in determining caseload size:

• What are the characteristics (that is, the degree of job readiness) of the participants with whom case managers work?
• What are the functions that case managers are expected to perform?
• What are the performance goals that case managers must satisfy?
• How much access do case managers have to community services?
• What levels of clerical support and automation are available to case managers?
• What are the background and training of case managers?

Caseload issues are different in work first than in other welfare-to-work programs because program activities are shorter-term and participants are moving more quickly into employment. Caseload sizes may need to be smaller, as case managers work more intensively with their entire caseload, closely monitoring the participation even of those referred to outside activities. At the same time, the program may be able to serve more participants over time, because of rapid turnover.

Specialization

Different programs define the case manager's job in different ways. In some programs, case managers are responsible for each step along the participant's path to employment, from orientation to job placement. In others, some tasks-such as assessment or job development-are handled by specialized staff. Similarly, programs may assign caseloads randomly across case managers, or may instead assign cases to specialized case managers on the basis of participants' characteristics or activities. The pros and cons of both types of specialization are outlined below. If you choose to specialize staff functions, you may want to consider adopting a team approach, in which staff with different specialties work together to serve program participants.

• Specialization by program function. Designating certain program elements as the responsibility of specialized staff members can allow staff to focus their attention and develop expertise. If case managers not only have to counsel and monitor participants but also must conduct assessments, facilitate job clubs, and call employers to develop jobs, it can be difficult to give each of these aspects the attention it needs. Different functions also require different skills, and specialization can take advantage of staff with those skills. (Some program elements might also be contracted out to specialized service providers.) Finally, specialization can free case managers of administrative responsibilities so that they can focus on employment. For example, specialized staff might handle scheduling orientations, monitoring attendance, or processing child care payments.

Generalist case managers, in contrast, may develop stronger relationships with participants, and can use information they learn in one role to promote employment through another (for example, using knowledge of a participant's interests and needs when developing job leads). A structure in which case managers are responsible for more pieces of the program can also avoid communication problems between staff and establish a clear locus of responsibility for moving participants into employment.

• Specialization by participant characteristics and activities. Another option is to assign certain groups of participants to specialized case managers. This type of specialization can provide non-English-speaking participants with a case manager who speaks their language or can ensure that teen parents get special attention. In Grand Rapids, Michigan, staff caseloads are coordinated with participants' assignments to different service providers. Specialized staff get to know the service providers, thus facilitating communication and coordination.

A drawback of specialized case management is that caseloads may shift whenever participants' characteristics change, as when a participant completes an education program. Every time this occurs, you risk losing participation and delaying progress toward employment. Separating out participants can also send mixed messages about work expectations by suggesting that some people need special services. Finally, specialized caseloads can be difficult for administrators and supervisors to manage, because they may require different degrees of personalized attention, variations in caseload size, different types of supervision, and different performance standards.

16. The Role of the Eligibility Office

The eligibility (or income maintenance) office is a crucial player in welfare-to-work programs. This section describes the role of eligibility staff in a work first program and offers suggestions for improving communication and coordination between the two offices. Section 17 discusses some of the trade-offs involved in co-locating eligibility and work first or integrating responsibility for both functions in a single staff member.

Functions of Eligibility Staff

The roles that eligibility staff often play in work first, and ways in which some programs have tried to facilitate those roles, can be summarized as follows:

• Communicating the program's message. Eligibility staff often give participants their first explanation of what the program is about and what to expect. In order to do this effectively, eligibility staff must understand the program's philosophy and requirements, must be able to effectively communicate these to participants, and must have the time to add this employment focus to their jobs. As the message spreads, and as more participants get jobs, this function should become easier and the dynamics of the welfare office may begin to change. It also helps for eligibility staff to have visited the work first office and seen at first hand what the program is about.

Eligibility staff have a special role in explaining financial incentives, earned income disregards, and other rules about what happens to a participant's grant when she goes to work. Eligibility staff often understand this better than work first staff, and thus can be a more knowledgeable source of information for participants. They can also be more credible when it comes to convincing participants that work can pay.

• Screening and making referrals. Eligibility staff are generally responsible for identifying mandatory participants and for referring new mandatory and voluntary participants to work first. Eligibility staff also need to promptly refer individuals who become mandatory when exemptions end, when sanctions are lifted, or for other reasons. For both mandatory participants and potential volunteers, eligibility staff need to market the program and encourage participation. However, making referrals to work first may not be a high priority for eligibility staff, and delays in this area are a common complaint of welfare-to-work program administrators. It helps to include this function in the formal job description of eligibility workers and to monitor and reward its completion. Another solution is for the work first office to place a staff person at the eligibility office to assist in referrals.

• Offering employment assistance. Many eligibility staff welcome the opportunity to help individuals move from welfare to work. Eligibility staff can be especially useful in this role when working with individuals who are not participating in the work first program, either because they are not in a target group or have not yet been referred. Administrators can support and encourage this role by freeing up staff time and by providing resources and information about job search, job leads, and available support services-especially child care. Recognition or other performance incentives can set the tone and reward eligibility workers who are especially successful at helping individuals find jobs. Changing the atmosphere of the eligibility office-for example, by hanging employment-themed posters and providing information about employment opportunities-can also help.

• Sanctioning noncompliant participants. Processing sanctions that result from noncompliance with work first involves two-way communication. Work first staff must notify eligibility staff to impose a sanction, and eligibility staff must act on the request. The process works the same way when a sanction has been cured. It is often unclear which staff is responsible for following up on people in sanction status and for attempting to bring them into compliance, and each worker may have information about a given case that the other lacks. Delays or miscommunication in implementing sanctions can be costly and send mixed messages to participants, while similar problems in reinstating benefits can cause financial hardship for participants. Some work first programs have taken on the sanctioning function to avoid miscommunication or delays.

• Processing changes in participants' employment and earnings. Work first staff need to notify eligibility workers when a participant gets a job, and eligibility workers need to keep case managers apprised of changes in employment and reported income. Program managers in one site report that their largest financial losses come from delays in adjusting grants when participants start work. Likewise, case managers often report that participants lose income because eligibility workers fail to apply all of the earned income disregards for which participants are eligible.

Participants who combine welfare and work often take much more of an eligibility worker's time than do those without earned income. To accommodate this additional workload, program administrators might consider reducing the caseloads of staff with earned income cases (for example, by counting two employed participants as three without earned income in determining caseload size). Work first staff can help by providing eligibility workers with much of the detailed employment data they need for adjusting grant levels. Timely and accurate reporting of earned income can also be facilitated by conducting individual meetings or group orientations with newly employed participants, to review changes in their grants and to explain reporting requirements.

Suggestions for Improving Coordination

The following suggestions offer additional ways to improve communication and coordination between eligibility and work first offices:

• Include a discussion of common goals, coordination issues, and communication procedures in the training of new and ongoing staff.
• Create liaisons in each office.
• Put in place a team approach by pairing eligibility workers with work first case managers so that their caseloads overlap.
• Have staff visit each other's offices and learn about their operations and procedures.
• Hold joint staff meetings to discuss coordination issues.
• Ensure that eligibility workers know at least one work first case manager whom they can call for information, and vice versa.
• Make it easy for staff to know who a client's eligibility worker or case manager is-for example, by assigning caseloads alphabetically.

17. Co-location and Integrated Case Management

One way to facilitate coordination between eligibility and work first is to co-locate those services. Another way is to combine eligibility and work first functions through integrated case management. This section discusses the trade-offs involved in each of these approaches.

Co-location

The benefits of co-locating eligibility and work first offices can be summarized as follows:

• Communication. Co-location can facilitate communication and coordination between eligibility and work first staffs.
• The "culture" of welfare. Co-location can help to change the overall atmosphere of the welfare office to one focused on work and can increase eligibility workers' "buy-in" to the program.
• Convenience. Co-location can be easier on participants, especially those with transportation problems, and can reduce delays in participation.

Co-location also has potential drawbacks. In deciding whether to co-locate eligibility and work first, consider the following:

• Message. Locating work first in a separate office can emphasize that the program is not "business as usual," thus helping to establish a distinct program message.
• Atmosphere. Many work first programs try to maintain professional offices modeled on the private sector to complement their emphasis on employment. Welfare offices-particularly in urban areas-may be crowded or noisy, have people constantly coming and going, or have guards and other security features.
• Space needs. If the program includes job clubs and other activities on-site, facilities that can accommodate both offices may not be available.
• Participants' attitudes. Because attitude and motivation are such key elements of work first, it might make sense to keep the offices separate if participants have negative associations with the eligibility office.

Integrated Case Management

Another decision involves whether to maintain separate staff for the eligibility functions of the welfare system or to combine the eligibility and work first functions. A "traditional" approach to case management separates the functions of the work first case manager from those of the eligibility worker. An "integrated" approach combines these functions in a single worker.

An ongoing evaluation of the JOBS program in Columbus, Ohio, which includes a direct comparison of these two approaches, suggests that integrated case management may lead to:

• Significantly higher welfare savings and reductions in welfare rolls
• A higher proportion of participants attending the work first orientation sessions
• Higher monthly participation rates
• Lower monthly sanctioning rates
• Better tracking and monitoring of participants

The integrated approach may have been more successful for several reasons. First, an integrated model avoids any lapse of communication between the two workers and ensures a consistent message. Second, participants may take the program requirements more seriously when they know that their case manager also controls their grant. Finally, the integrated approach may help forge a closer relationship between workers and clients, encouraging participation.

Despite these promising results, administrators should be cautious when considering integrating work first case management and eligibility functions. The employment focus so central to work first can easily get overshadowed by the demands of verifying eligibility, processing grants, and minimizing error rates. Potential problems include the following:

• Workload. Eligibility staff often carry heavy caseloads, so their ability to add a focus on employment is limited. An integrated approach necessitates lighter caseloads.
• Multiple priorities. The diverse responsibilities of integrated workers also make it more difficult to make employment a priority. Even with caseloads of under 100, issues of eligibility and benefits tend to dominate integrated workers' time, and the employment focus can get lost.
• Resistance to change. In some places, it may be easier to create a separate employment-focused program than to reorient the existing welfare bureaucracy.
• Relationships with clients. Eligibility staff often function as the "police" of the welfare system, verifying eligibility and protecting against fraud. This may make it difficult for staff and participants to establish the kind of trust required in work first.
• Staff qualifications. Eligibility staff are often successful because of their ability to process and manage vast amounts of paperwork. They may not have the motivational, counseling, and other skills that make for a successful work first case manager. Similarly, case managers may not have the skills that make for a successful eligibility worker.

Because of the challenges for staff of taking on so many roles, integrated case management must be implemented in the context of adequate resources: highly trained staff; clerical and other supports; and high-quality program services.

18. Promoting an Employment Focus

Because work first programs are defined as much by their philosophy as by the services they offer, the clear articulation of the employment message to staff, service providers, and participants is a crucial part of program implementation. Program administrators need to find ways to emphasize the importance of employment and to focus staff on getting people jobs. In addition, unless all those who interact with participants communicate the same message, participants may not understand what is expected of them. This section offers a number of suggestions for promoting an employment focus in your program, and discusses the benefits and risks of establishing performance standards for staff.

Suggestions for Promoting an Employment Focus

• Emphasize employment repeatedly, and from the top down. It should be clear that support for the new focus comes from a high level, whether that is the governor, the welfare commissioner, or the county administrator. Senior administrators should set the tone by promoting the importance of the program's mission. All levels of management and supervisory staff should follow suit, sending a consistent message down the line. Similarly, the employment message should be reinforced in all interactions with participants as well as in any written materials or notices that participants receive.

• Link the focus on employment to the daily work of staff. Let staff members see how their daily work promotes the goal of employment, and help them relate what they do to the message given to participants. Shorter-term or activity-specific objectives (such as bringing participants into orientation or increasing education completion rates) should also be related to the larger goal of employment.

• Use a new language. A change in the program message can be signaled by a change in the language used by program administrators and staff. Including the word "employment" or "work" in the program's name-and repeating it frequently in the context of program activities-can be a powerful way of sending a new message to both staff and participants. Terms like "outcomes," "job placements," and "job contacts" are other examples of a new language for work first.

• Adopt a motto. Slogans can help clarify the employment message and can make it stand out. For example, Los Angeles adopted the motto "A job, a better job, a career" when it shifted from a program of education and job search to one with a strong emphasis on work first. The motto is part of the program logo and can be seen on posters, in handouts, and on pins worn by staff and administrators. It makes clear that the immediate goal for participants is employment, and conveys the philosophy that even a minimum-wage job is a positive start.

• Make sure that the message is heard and understood. In Vermont, all Department of Social Welfare staff showed that they understood the new philosophy by signing a poster proclaiming the program message. For participants, signing an employability plan or other "contract" is a concrete way to show that they understand what is expected of them in the new program.

• Market the program. Promoting the program and its goals through the media and other means can reinforce the employment focus, foster a positive public image of the program, and make staff and participants feel proud to be a part of it. Marketing the program can also help sell the message to prospective employers and the public.

• Promote the employment focus in agreements with service providers. It is important to communicate the program message not just to your own agency's staff but also to any other providers who will work with program participants. (See section 20 for suggestions on how to do this.)

• Monitor outcomes. Some program administrators either may not want to institute formal performance standards (discussed below) or may be limited in their ability to do so by union contracts or other factors. Even without formalized standards, however, employment and other related outcomes can be monitored and included informally in staff evaluations, and benchmarks can be set for desired outcomes.

• Give staff flexibility in achieving desired outcomes. As you emphasize the goal of employment, give local offices and staff members flexibility to try different approaches to achieve that goal. Staff may better respond to the new focus when they feel that they are given the freedom to carry it out.

• Reward success. Publicly acknowledge the success of individual staff members, units, and offices in meeting program goals. In addition, honor the accomplishments of partner agencies and service providers. High performance can be rewarded with public recognition, certificates of achievement, or prizes, such as gift certificates or movie passes.

• Change the culture of the office. Posters, signs, and videos in the waiting area can all "advertise" the message and heighten the emphasis on employment for the program as a whole. Having pots of coffee available in the waiting area can create an atmosphere of professionalism and respect. Staff should follow the same rules relating to professionalism, punctuality, and "dressing for success" that are recommended for participants.

Performance Standards

How staff are evaluated sends a strong message about the program's goals, and instituting employment-focused performance standards can clearly communicate the program's expectations for staff. One program that has made extensive use of employment standards is in Riverside County, California. Each staff member must achieve at least 15 job placements per month (out of a caseload of approximately 120) in order to meet the standard. In addition, each district office sets performance goals, which are higher than the standards (the current goals range from 20 to 30 participants entering employment each month). Staff members who achieve 30 placements in a given month receive an award. Staff achievements are posted daily, by individual and by unit. The performance of different offices is also publicized, so that they compete against each other.

Administrators in Riverside credit their performance standards for much of the program's success in increasing employment among participants. While staff have multiple job responsibilities, the prominence of the standards makes it clear that employment is the main goal of the program and the main (though not the only) criterion for evaluating staff. Staff with lower performance regularly seek out higher performers to learn "how they do it."

There is no fixed rule for how high to set performance standards. Most programs start with a somewhat arbitrary guess and then adjust the standards according to how well staff do in meeting them.

Despite their potential for heightening a program's employment focus, performance standards present some risks, as discussed below:

• Performance standards can lead to "creaming." Because they are evaluated so heavily on the basis of the end goal of employment, staff may concentrate their efforts on those participants who are most likely to succeed. There may be little incentive for staff to expend a lot of effort working with participants who seem a long way from employment (though administrators in Riverside believe that their program's high performance standards force staff to work with their entire caseload). Combining performance standards with some process standards (such as caseload coverage, described in section 13) can discourage creaming.

• An overemphasis on employment can detract from individual client needs and from other program goals. Recognizing incremental steps toward employment or combining employment standards with other performance measures (such as completion of education) can help counteract this.

• If not monitored, performance standards can lead to inflated outcomes. In an effort to achieve high outcomes, staff may report employment that is not confirmed or does not last.

• Performance standards can have a negative effect on staff morale. This is especially likely if the standards are perceived as unrealistic or unfair. Involving staff in setting the standards can help make them more realistic and counter staff resistance.

• Putting staff in competition with one another can discourage cooperation. Measuring the performance of larger units rather than of individual workers or giving credit to all staff who help a participant gain employment can encourage staff to work together.

19. Management Information Systems

A successful management information system (MIS) works as both a management tool and a support for line workers. The design and implementation of an MIS should keep both of these uses in mind. In addition, an MIS should support and promote the program's goals. In the context of work first, this may mean an increased emphasis on monitoring and tracking participants, as well as a greater role in simplifying and reducing paperwork so that staff can focus on promoting employment. If the system cannot provide all the support the program needs, administrators should look both for ways to accomplish functions manually and for opportunities to modify the program design so that those functions are not as critical.

Designing an MIS

The following are three key areas which should be addressed in designing an MIS:

• Who should be served? The MIS must be able to identify individuals appropriate for participation and make that information available to line staff and managers. A system that tracks who should be called in for participation, who has not participated as required, and who is no longer required to participate can assist line staff in everyday tasks and provide tools for management to assess workloads.

• What services are being provided? Line staff should be able to record activities, and managers should be able to see aggregate statistics on participation, both overall and in various components. It is also helpful if staff can identify specific service providers' locations, schedules, and slot openings.

• Where are participants in the process? Being able to see where participants are in the array of activities and when activities are scheduled to end helps workers manage their caseloads. It also helps administrators check whether bottlenecks are developing and whether participants are more likely to drop out at certain points in the program. In a time-limited program, a key MIS function is tracking where people are in relation to the time limit. This can be especially complicated if different participants face different time limits or if the time-limit clock can start and stop as the status of participants changes.

A good MIS should be user-friendly and give staff all the tools they need to manage their caseloads. Staff at several work first programs offered the following specific suggestions about what they would like a computerized system to do:

• Track dates and deadlines, to alert staff when activities or deferrals end, when notices need to be sent, and when meetings need to be scheduled
• Automatically generate scheduled notices for participants around program activities and other deadlines
• Automatically schedule participants for meetings, orientations, job clubs, and other activities
• Coordinate scheduling with other staff-for example, by maintaining uniform class size in scheduling participants for the next available job club
• Facilitate coordination with eligibility staff, by allowing workers in each office to view information from the other's system and to transmit information over the computer
• Automatically update information on employment and earnings, as well as basic information like address changes, in the work first system when changes are entered in the eligibility system, and vice versa
• Automatically insert case information onto computerized forms to avoid double entry and save staff time
• Help staff quickly calculate for participants what will happen to their grants and total income if they go to work at various wage levels and work hours

A good system should also provide program administrators with reliable data that they can use in their daily management as well as in planning and measuring program performance. Administrators have suggested that in addition to basic program and caseload data, they would like a computerized system to give them easy access to the following information: average wages and hours for participants who start work; the program activities from which participants find jobs; information on job retention and recidivism; profiles of caseloads and placement information for each worker; and information on the number of deferrals and reasons for deferrals.

Finally, a good MIS can help both staff and participants by providing linkages with other agencies or systems that have helpful information. These might include labor department job banks, child care resource and referral agencies, or local community colleges and school systems.

MIS Implementation Issues

In deciding how best to implement an MIS, program managers need to consider the following elements:

• Resources. What system resources are already in place? Can the existing systems be easily modified to add new features? If not, is it feasible to develop a separate system for the new features? Are money and personnel available to develop the system within the desired time? If an automated system cannot be developed or cannot perform all needed functions, a manual system will be needed in its place.

• Staff support. Successful implementation of any system requires acceptance from those who must use it. Staff need to be trained both in using the system and in integrating it into their daily work. In addition, equipment must be easily accessible (ideally, there should be a terminal on each worker's desk), and system support must be available to answer questions as they arise.

• Timing. The timing of program implementation often determines what system changes can be implemented. It takes time to redesign systems, put them in place, and train staff to use them. A system will be most useful if it is fully operational before program implementation begins. If this is not possible, it is important to prioritize the order in which systems functions will become operational and to put in place manual processes to fill the gaps.

Designing a system is not the end, however. Once a system is developed, programs need to devote resources to maintaining and updating the system, training and retraining staff, and solving problems. If not-if technical staff are no longer available to deal with problems or implement improvements-the system may quickly become outdated and lose its usefulness.

20. Interagency Linkages

Chances are that your agency will not provide all program services itself. Some services will be contracted out, while others may be performed by outside agencies without formal contracts. A variety of factors-such as political pressure for privatization, restrictions on government hiring, or a desire to take advantage of the specialized experience of other agencies-may lead program administrators to look outside the welfare department for certain tasks.

You may want to think about contracting out those elements that the program has not done before or has not done successfully. In addition, you may want to contract out more services in the early stages of a program, focusing internally on core functions at first and then taking on more roles over time. When considering new or modified interagency linkages, program administrators should be prepared to encounter opposition from unions or others. Such opposition can divert the attention of administrators and can delay the implementation of work first. Administrators need to assess the extent of potential opposition and determine whether the effort is worthwhile.

The late Sar Levitan, the labor economist, once defined coordination in employment and training programs as "an unnatural act between two or more nonconsenting bureaucracies." The key organizational players in welfare-to-work programs-typically the department of social services and providers of employment, education, and training-traditionally have different missions, goals, targeting strategies, and accountability standards, and use different kinds of information systems.

Forming successful interagency linkages is particularly important in a work first program, for several reasons: the need for a consistent and clear message to participants; the need for close monitoring of participants; and the need to facilitate quick movement of participants into and between activities and into employment. This section contains advice for forming linkages with partner agencies and for putting in place successful contracts in the context of work first.

Forming Successful Linkages

Shifting to work first is likely to involve a process of reappraisal leading to the restructuring of existing linkages and the establishment of new ones. Here are a few things to consider in that process:

• Know your partners: successful linkages are more than good operating procedures. A mutual assessment process-before detailed planning begins-can allow organizations to establish common ground about goals and priorities, anticipate conflicts, and set realistic expectations. This assessment process is likely to work best if it involves staff at all levels of the organization. Remember, too, that it takes time for relationships to develop; establish mechanisms for frequent communication and resolution of problems that arise along the way.

• Do not assume that partner organizations cannot change. Because establishing new relationships can be very difficult, it pays to explore how much current relationships can be altered beyond their traditional roles. Both Portland, Oregon, and Los Angeles, for example, have been successful in getting education institutions that previously focused on long-term human capital development to shift relatively quickly to playing a role in a work first model.

• Sometimes it pays to do it yourself. It may make sense for your program to take on responsibility for some functions rather than try to coordinate across agencies. For example, many work first programs prefer to run their own job clubs so that they can maintain control over the activity's content and can more readily monitor attendance and progress.

• Be careful about imposing workload and resource burdens. If you are asking for a service from another organization, you should expect to pay for it. If you cannot, then you should try to minimize the workload or propose alternatives. For example, it may be overwhelming for a small training provider to report attendance information to many different case managers in a large program. Some programs have responded by designating a single staff member as the reporting contact.

Contracting Out

In establishing formal contracts with service providers, program administrators should bear in mind the following advice:

• Address program philosophy head-on. Service providers may have goals, perspectives, and philosophies that are different from those of work first. Requests for proposals, contracts, performance measures, and payment structures should be designed so as to directly address the work first program's philosophy and how it will be incorporated into the specific services being contracted. Regular monitoring should ensure that the philosophy has in fact been put into practice.

• Establish communication linkages. Communication and monitoring procedures and standards should be clearly addressed in agreements with service providers (see also section 33, on maximizing participation). Mechanisms should also be established to promote ongoing teamwork between line staff at the work first office and contracted agencies. Regular visits to each other's sites as well as interagency meetings, retreats, or conferences are useful communication and team-building tools.

• Carefully estimate participation levels. Developing reasonably accurate projections of the number of individuals who will be served in each program component is critical to negotiating workable agreements with service providers. Such projections enable providers to plan for staffing and participant flow, to predict costs, and to design a payment structure. The projections should take into account both the experiences of similar programs and local caseload characteristics.

• Leave room for flexibility and contingency plans. The program will inevitably not go precisely as expected. Agreements should therefore be flexible and include contingency plans to allow adjustments to be made once program implementation has begun.

• Consider your contracting options. Contracts can be structured in a variety of ways. Service providers can be reimbursed for their costs in working with participants or paid for achieving desired outcomes (such as education completion or job placement). Another option is to contract with multiple providers, allowing participants and staff to choose among them. This can provide leeway for matching participants with programs on the basis of their strengths, location, or special features. It can also improve outcomes by fostering competition among providers.

• Promote the outcomes you want to encourage. Contracts for work first activities or services should promote outcomes that further the program's goals. The outcomes you emphasize for job club might be placements and retention, while outcomes for skills training might include both credentials attained and job placements. The outcomes should be relevant to the service provided, and should be easily measurable.

• Protect against creaming. One danger of emphasizing outcomes is that it might lead providers to target only those participants most likely to succeed, especially when funding is at stake. Contracts should give the work first program control over who gets referred to contracted agencies and should specify the reasons why a participant might be denied service or dropped from the activity. Specifying service expectations-in addition to outcome goals-can also guard against creaming.

• Maintain an oversight and coordination role. Once the contract has been signed, the welfare department's role does not end. The department should maintain a role in oversight and coordination of services.

21. Program Costs

Operating a work first program typically costs less per participant than operating a program that emphasizes education and training, mainly because participants remain in program activities for less time. However, work first programs can vary greatly in their costs, for a number of reasons. These include differences in the average length of participation, the extent of monitoring and case management, which services are emphasized, and the extent to which support services are made available.

To illustrate the range of work first program costs and some of the reasons for the variation, this section looks at the costs of six programs that MDRC has evaluated: Florida's Project Independence; San Diego SWIM (Saturation Work Initiative Model); Riverside, California's, GAIN program; and JOBS programs in Atlanta, Grand Rapids (Michigan), and Riverside.

Table 3 breaks down program-related costs into operating costs-staff, administration, and other overhead-and support service costs-mostly for child care and transportation (see section 7 for more on support services). Operating costs are further broken down into costs to the welfare department and costs borne by other agencies, including Job Training Partnership Act (JTPA) agencies and education providers. On the one hand, program administrators will aim to maximize the amount of resources leveraged from other sources. On the other hand, administrators should be aware of the larger picture: maximizing cost-effectiveness from a total government and taxpayer perspective involves taking a certain amount of responsibility for all program-related costs.

Table 3

Estimated Costs for Single Parents Assigned to a Work First Programa

                           Operating Cost         Support Service   Total Cost    
      Program                 Per Person          Cost Per Person    Per Person   

                      Welfare     Other     Total    Welfare Dept.                  
                       Dept.     Agencies              Cost Only                    


Florida's Project                                                                   
Independence             $  312     $  491   $  803             $118        $   921 

San Diego SWIMb             988        858    1,846              101          1,947 

Riverside GAINc           1,671        817    2,488              123          2,611 

Atlanta JOBS              1,154        802    1,956              882          2,838 

Grand Rapids JOBS           648      2,164    2,812              297          3,109 

Riverside JOBS              919        187    1,105              122          1,227 


Average                     949        886    1,835              274          2,109 

SOURCES: Kemple, Friedlander, and Fellerath, 1995; Hamilton and Friedlander, 1989; Riccio, Friedlander, and Freedman, 1994; U.S. Department of Health and Human Services and U.S. Department of Education (prepared by Hamilton, Brock, Farrell, Friedlander, and Harknett), forthcoming.

aCosts are for single-parent AFDC recipients for a two- to three-year follow-up period and are in 1993 dollars. The costs are averages across all program group members, including both those who did and those who did not participate in program activities. (Control group members are not included.) Roughly 60 percent of program group members ever participated in a program activity. Also not included are the costs of services that program group members received after leaving the work first program.

bThese figures have been inflation-adjusted and therefore differ from those in the cited report. SWIM operating costs include the cost of some community college and training services that participants received after leaving the SWIM program.

cThese figures are based on two to three years of follow-up, rather than the five years of follow-up in the cited report.

The following is a discussion of some of the reasons for the variation in costs among the programs cited in Table 3:

• Operating costs. Project Independence had the lowest total operating cost per program group member. As economic and other conditions in Florida changed, caseloads increased dramatically. At the same time, the program experienced a hiring freeze and child care budget cuts. Facing those constraints, the program spent less on each participant. Grand Rapids' higher operating costs can be explained primarily by a high level of participation in education and training programs, in part due to an extensive network of education and training services offered in the Grand Rapids community. Many individuals were already participating in education or training when they entered Grand Rapids' work first program, and they were allowed to continue in those activities.

• Support service costs. Support service costs were lowest in SWIM. Support service costs in most programs are primarily for child care, and SWIM's program group was composed mostly of parents with children six and older, who did not require levels of child care services as high as those needed at the other sites. Also, SWIM's allowable child care rates during the first year of the program were very low ($1.25 per hour per child). Atlanta's support service costs were much higher than the other programs' costs due to high monthly child care costs, a high number of monthly child care payments, and generous payments made for transportation and other ancillary services.

Costs for Program Activities

Table 4 presents the per-person cost of each major work first program activity for those who participated in that activity. The table is based on data from the work first programs in Florida, Atlanta, Grand Rapids, and Riverside (both GAIN and JOBS). The component costs given in the table include the costs of the activity or service as well as costs for case management to monitor and enforce participation requirements and address barriers to participation. These costs can give you a start in estimating what your own program costs might be, depending on how many people you expect to participate in each component, how long you expect them to participate, and other factors. The range of costs reflects differences across the five programs studied; the third column of the table explains those differences by noting some of the factors that affected the cost of each component.

Cost-Effectiveness

Welfare-to-work programs do cost the government money, but they can also bring a return on the investment when program participants leave welfare for work. Virtually every program evaluated by MDRC in which job search has been the first component for most participants has been found to be cost-effective from the standpoint of government budgets. The savings from reduced welfare, food stamps, and Medicaid, as well as increased revenue from taxes paid by participants, more than paid for the program. Moreover, close to half of these programs have proven to be very good investments for taxpayers. Within five years, programs in San Diego, Arkansas, and Riverside returned to government between $2.34 and $5.50 per dollar spent.

Table 4

Cost Per Single-Parent Participant for Work First Program Components:

The Range of Costs for Five Programsa

                                                                                 
Component       Cost Per Participant      Factors That Affected Costs      
                             (Range)                                             



Orientationb          $25-$100              Length of orientation; group size    

Job search            $270-$1,930           Duration of job search; staffing     
                                            and group size                       

Basic education       $1,610-$4,365         Duration of education; class size;   
                                            providers used                       

Vocational training   $4,395-$6,980         Duration of education or training;   
and college                                 class size; providers used           

Work experience       $340-$1,400           Duration of activity; intensity of   
or on-the-job                               supervision                          
trainingc                                                                        

Child cared           $435-$2,250           Length of participation; type of     
                                            care; age of children; local market  
                                            rates                                

Transportatione       $65-$125              Type of transportation; extent of    
                                            coverage                             

Other support         $105-$115             What is covered (e.g., uniforms,     
servicese                                   books, equipment, registration and   
                                            licensing fees)                      

SOURCES: Kemple, Friedlander, and Fellerath, 1995; Hamilton and Friedlander, 1989; Riccio, Friedlander, and Freedman, 1994; U.S. Department of Health and Human Services and U.S. Department of Education (prepared by Hamilton, Brock, Farrell, Friedlander, and Harknett), forthcoming.

aCosts are two- to three-year costs for single-parent AFDC recipients and are in 1993 dollars. Note that this table uses a different base than does Table 3. Table 3 presents the cost per program group member, including both those who did and those who did not participate in program activities. Table 4, in contrast, presents the cost for each person who actually participated in the program activity cited. As noted below, costs of some components for some sites are not included; this is because of small sample sizes or because comparable estimates were not available.

bNot including Riverside GAIN.

cNot including Riverside JOBS. Another study in seven sites found a similar range of per-participant costs for unpaid work experience that generally lasted three to six months. On the basis of that study, MDRC estimated that it would cost between $2,000 and $4,000 annually (excluding child care costs) to keep a work experience position filled for a year. See Brock, Butler, and Long, 1993.

dNot including Florida.

eNot including Florida or Grand Rapids.

22. Strategies to Reduce Costs

The differences in program costs among the work first programs discussed in section 21 reflect differences in program funding, policy, and implementation; the availability of community services; regional wages and prices; labor market conditions; and the demographic characteristics of participants. Some of these variables are outside the control of program planners. However, given these environmental factors, planners can reduce the amount spent on each participant through the decisions they make regarding program policy and implementation. Research suggests some ways to do so, outlined below.

Remember, however, that there are trade-offs to cost-saving strategies, and program administrators need to strike a balance among costs, benefits, goals, and outcomes. Minimal government investment in the short run does not necessarily save government money in the long run. Conversely, well-managed programs with considerable up-front investments can be cost-effective.

Reducing the Length of Time Spent in the Program

• Close monitoring of participation. By keeping close track of attendance and progress, case managers can ensure that program resources pay for active participation and can reduce the number of months that participants spend in the program. Caseload sizes need to be small enough to allow for close monitoring, or else specialized staff need to be assigned responsibility for monitoring participation (see section 15 for more on caseload size and specialization). Close monitoring also involves coordination with service providers if program activities take place off-site (see section 20, on interagency linkages).

• Short-term activities. Use of short-term activities, such as job search, reduces the amount of time participants spend in the program. Priority can also be given to short-term programs in other areas, such as education and training, where programs can vary greatly in length.

• Payment agreements. In order to encourage service providers to reduce the length of time participants stay in activities, targets for completion of activities can be written into contracts and funding agreements. For example, Riverside County makes incentive payments to schools that succeed in getting participants to make progress and complete their education assignments.

Reducing Unit Operating Costs

• Economies of scale. Programs that enroll large numbers of participants benefit from economies of scale. Overhead costs, such as rent, utilities, equipment, and administrative staff salaries, are spread across a larger base of participants.

• Referrals to outside agencies. Outside agencies (such as adult schools, community colleges, vocational training institutes, and trade schools) may provide and pay for some services utilized by work first participants. The welfare department can reduce its own costs (though not necessarily government costs, since other agencies' costs may increase) by referring participants to these agencies.

• Contracts with outside agencies. Outside agencies that specialize in particular areas or that serve large numbers of participants may be able to offer some program services more cheaply than can welfare departments. Program administrators can take advantage of this by contracting out those program elements. However, the offsetting costs of contract management and oversight must also be considered.

Reducing Child Care Costs

• Type of child care. Child care provided by licensed centers tends to be most expensive, followed by family day care and then by child care provided by friends or relatives. Some programs counsel participants to seek low-cost child care, taking the position that welfare recipients who are using low-cost care will be able to afford this service on their own after leaving welfare. Other programs counsel participants to use licensed child care centers, believing that such facilities offer higher quality and more reliable care, and can therefore provide better support for program participation and employment. (See section 7 for suggestions on designing child care support services.)

• Age of children. Infant child care tends to be most expensive, followed by toddler care and care for preschool-aged children. Program planners may want to take this into account when determining who should participate in work first. However, some administrators have found that it is not more expensive to serve parents with young children because they tend to have fewer children and often rely on relatives for child care.

• Flexible scheduling. Limiting the number of required hours of participation, and maintaining flexibility in scheduling those hours, can reduce child care costs. For example, the San Diego SWIM program referred individuals to activities that coincided with their children's school hours. Parents then needed only some preschool, after-school, "backup," and summer care, so costs were noticeably reduced.

Close monitoring of participation. Close monitoring will ensure that child care expenditures support active participation. In addition, reducing the number of months that individuals participate in work first-through close monitoring and use of short-term activities-will also reduce child care costs (though participants may receive transitional child care benefits after leaving the program).