Culture Poverty & Welfare
January 1st, 2000 21 Minute Read Issue Brief by Peter Cove

Making Welfare-to-Work Fly


After 34 years of experience in the welfare-to-work industry, America Works’ Peter Cove knows a way to help end welfare as we know it—and in the process, help break up the “Welfare-Industrial Complex”: switch to performance-based contracting for social-service providers. Cove argues that welfare agencies will not get the results the public wants--and welfare recipients need--until they institute performance-based contracting.


Civic Bulletin

Consider the following hypothetical situation: the U.S. government wants 20 advanced-design jet fighters. It issues a request for proposals (RFP), and the JetSet Defense Contracting Company wins the bid. The government gives JetSet a 25 percent advance, which the company uses to buy materials and hire engineers and production labor. It builds the 20 fuselages and receives another 25 percent. When it has assembled all the planes, JetSet receives another 30 percent, bringing its receipts to 80 percent of the total. But only four of the planes actually work properly; the rest are unfit for flight. In the end, the government gets 16 fewer planes than it ordered, JetSet gets 80 percent of its revenue, and we all know what the taxpayer gets. Obviously, neither government auditors nor a congressional oversight committee would allow such a procurement procedure to remain in place. The project would be acknowledged an abject failure, and JetSet’s prospects for future contracts severely jeopardized.

But let’s say JetSet isn’t a defense contractor; instead, the government hires JetSet to help former welfare recipients get long-term jobs. Following the prevailing practices in the welfare-to-work industry, compensation for the program would likely be disbursed in a way similar to the above—and, unfortunately, it would achieve similar results. Upon signing a contract to remove 2,000 people from the welfare rolls, the company would get a 25 percent advance. It would earn another 25 percent for bringing welfare recipients into classrooms. Twenty percent more would be paid when a certain number of recipients finish prescribed courses. A week after JetSet places people in jobs it would get another increment of revenue. If it should happen that only 20 percent of JetSet’s graduates are still working two months later, and many are back on the welfare rolls, JetSet would still be set. The company would not only keep most of the cash, but other contracts would follow.

Unbelievably, this hypothetical situation accurately reflects welfare-to-work contracting today, as a quick glance at standard contracts show. The pro forma contract of one large southern city gives the service provider 20 percent for enrolling a welfare recipient, 20 percent for 30 days of training, 30 percent for job placement, and 10 percent each for 45 days, 90 days, and 180 days on the job. Under this contract, a provider earns 70 percent of his total payment even if the welfare recipient is employed only briefly and winds up back on welfare. Similarly, one mid-Atlantic state gives providers 75 percent of the contract for one day on the job.

Do government officials know what “performance” means? It would appear not. A Department of Labor contract specialist told me that all their contracts were performance-based, with payments for enrollment, 30-day retention, and 90-day retention. Further questioning revealed that because the department believes the clients who remain are the “hardest to serve,” its payment structure does not focus on long-term retention, and it often pays equal percentages for each milestone: providers receive 66 percent of their total contract if a client works only one day, 100 percent for two months. Other government employees have proudly related their experiences with performance-based contracting; each one eventually—and unself-consciously—admitted that in reality, if a person works for between 60 and 90 days and then returns to welfare, the contracted agency still receives from 87 percent to 100 percent of the total contract fee.

Where are our models for real performance contracting? The pay schedule of a recent Defense Department aerospace contract grants only 40 percent of the fee based on “safety and quality standards”—the contractor’s inputs (which include considerable overhead in this high-tech industry)—with the rest payable only after meeting mission objectives in a timely and cost-effective fashion. This is the kind of performance-based contract we should expect of welfare-to-work service providers, as well.

Real performance-based contracts force contractors to reduce costs and improve service quality. Under the system I will here propose, welfare-to-work contractors would receive a single payment for their work, and they would get it only after a welfare recipient has held a job for a four- or six-month period—which better reflects the success of a program than how many people complete a class on punctuality.

The data show that the performance contracting approach places more welfare recipients in jobs for longer periods than traditional approaches. A study by the Office of the Comptroller for New York City in 1993 found that only 17 percent of the public assistance recipients placed in jobs through the New York City’s Adult Training Program were still working 90 days later.1  Programs that pay little attention to results perform predictably: they don’t work—and neither do their clients. In contrast, data compiled by New York State in 1997 showed that 88 percent of the people my firm, America Works, had placed were not back on the rolls after three years.2

With such clear results, one would think that performance-based contracting would be all the rage, given the national interest in effective welfare reform. Unfortunately, neither contractors nor government agencies seem eager for the change. When I suggested to the contract officer of a large southern city that she pay my firm only if our clients kept their jobs for an extended period, she replied that she couldn’t even consider such a policy, saying, “I’d lose all my other contractors.” Similarly, I have heard fellow providers call such compensation packages “unrealistic,” “crazy,” and “impossible.” With providers and bureaucrats expressing such a mindset, it is no wonder there is so much resistance to making welfare-to-work contracts more accountable. Their nay-saying points up a difficult question: why are the expectations of social-service providers so low? After 34 years in various welfare-to-work programs, I think I have the answer to that question.

I have found that welfare bureaucracies value process rather than product; good intentions mean more than success in achieving the stated goal. So the people who design welfare-to-work programs and draft their contracts rely almost solely on measures of input—the number of social workers, caseworkers, and trainers in a program, or the number of people completing training or leaving with a résumé—to determine success. Largely ignored goes the essential question—in vocabulary borrowed from our jet-fighter hypothetical—“Can she fly?”

Several factors contribute to this odd cultural phenomenon, this “process over product” thinking.

Government loves process. It’s in the nature of government to focus on controls and inputs. Americans’ deep historical distrust of government—the belief that government employees given too much discretion will succumb to favoritism and corruption—partly explains the wide range of rules and procedures that most agencies must follow before agreeing to a contract. “In the soul of the bureaucratic machine there lurks a control freak. Employees are cogs in a highly regulated machine. . . . Managers do the thinking; workers do the tasks they are assigned. Detailed rules and procedures specify behaviors. Inspectors check for compliance.”3  Bureaucrats are more likely to follow established procedures than to think creatively about innovative ways to meet their goals.

Program operators are pessimistic. The “soft bigotry of low expectations” Texas governor George W. Bush recently condemned in our educational system also pervades much of the welfare-to-work industry. Counterintuitive as it may seem, welfare-to-work program directors often don’t believe they will be successful. They know most programs haven’t worked; why blame themselves for failure? Instead, they have variously argued that their programs are underfunded, that there aren’t enough jobs, that employers discriminate against welfare recipients, and that the pathology of candidates is overwhelming.

Program operators now say we’ve reached the bottom of the barrel. A Bronx-based not-for-profit employee recently remarked, “In comparison to a few years ago, we are seeing a lot more people with really long welfare histories, people who are not taking themselves seriously. They are truly the hardest to place.” Government officials accept this assertion at face value: a Department of Labor official recently noted to me, “The people who are left are the hardest to serve, and do not respond well to performance-based contracts.” Having little faith that their clientele can succeed, program operators refuse to be judged on the basis of that success. One might well argue that it’s even more important to base compensation upon success when dealing with this population.

Scattered programs succeed in spite of the presumed obstacles, and unsuccessful programs have an array of reasons to explain away the positive results: the successful program had lots of money; the organization is run by a charismatic leader whose approach can’t be duplicated; the program took only the best candidates; the economy was uncharacteristically good; or candidates who succeeded would have made it without the program’s help, anyway. Government contractors accept the cynical assertions of their service-provider friends, and successful programs are rarely studied to discover the secrets of their success. By maintaining a compensation system that will reward success and hold failure to account, performance-based contracting will challenge this prevailing pessimism by no longer granting the excuses it engenders.

Critics complain that vendors shouldn’t be forced to enter performance-based contracts because too many factors are out of their control: rising unemployment; a lack of sufficient funding per individual placement; and not enough candidates. All are potential problems. But vendors placing welfare recipients in jobs have the same right as any other kind of vendor to bow out of the funding stream if they sense they can no longer operate in changed circumstances. “[P]rivate agencies are often reluctant to leave the cozy shelter of mutual dependence on the state for the cold world of performance assessment.”4  Delivery from market forces can no longer be an option for the welfare-to-work industry.

Service providers don’t like competition. The plain truth is that many people who work in the helping professions are wary of the competitive marketplace. Most welfare-to-work programs are nonprofits, and most of those drawn to professions in the nonprofit sector either do not have an entrepreneurial spirit or are actually opposed to business and the whole idea of capital accumulation. The sense of altruism that draws them to social work makes it hard for them to accept capitalism’s rough-and-tumble efficiency, which can seem ruthless. They prefer the virtues of partnership and coordination.

A current example of this is the recent emergence of collaboration as a criterion in the awarding of welfare contracts. For some inexplicable reason, if two service providers—no matter how incompetent each might be—team up on a welfare-to-work contract, their combined proposal is prioritized in the award process. In some cases, up to 25 percent of proposal scoring is on the basis of collaboration, yet no studies exist demonstrating that collaborative programs increase job placement or retention rates compared with non-collaborative programs—which, as it happens, tend to be independent for-profit providers. It’s a blatant case of irrational government decree stifling competition. But since it’s based on the feel-good “qualification” of collaboration, shouldn’t we feel good about it? Hardly.

A fundamental misunderstanding of human nature and human society is at work in this cultural predisposition against competition. The thinking goes: someone has to lose when you compete; losing makes the loser feel bad; winners therefore victimize losers—and bear the karmic guilt for it; so everyone loses when you compete. Ideological blindness and soft sentimentality prevent purveyors of this theory from seeing that the exact opposite is the case. Losers learn from their losses; losers want to win next time; winners want to keep winning; so competition makes everyone better. We have ample and easily measurable evidence of competition’s role in human progress from the world of sports: even high schoolers run four-minute miles today, something no human had ever done until 1954.

Moving from welfare to a life of gainful employment means joining the market with all its attendant risks and benefits. It is more than ironic that this task has been left in the hands of organizations suspicious of competition. It is tragic.

Patronage rules. Despite the historical American distrust of government and efforts at reform, much of our politics remains “the art of who gets what.” The federal government spends billions of dollars for welfare-to-work programs, and the resulting contracts have become a prime resource for paying off supporters.

A few years ago, a powerful state politician in New York—responsible for the direction of much social welfare spending—visited America Works. After a tour, she observed that it looked like one of the best welfare-to-work programs she had seen. Would we have her support in future? No. “You see, Mr. Cove,” she said, “there is a community-based organization in my district. Their program is smaller than yours and probably not as good, but on election day, they bring out the votes.”

A senior government official in Indiana once made a similar point during a meeting. Two of the primary factors used to evaluate welfare-to-work contract proposals, he explained, were the organization’s location—local programs were prioritized—and whether the organization was friendly or hostile to the politicians then in office.

Nor is my own experience singular. In Smart Contracting for Local Government Services, Kevin Lavery cites numerous examples of political bedfellowship between municipal officials and nonprofit agencies. For instance, when New York City—long used to awarding de facto contracts to nonprofits in the human-service area—changed to a more competitive bidding process, and awarded five out of 75 contracts to for-profit groups, the nonprofits sued the city and won an injunction to prevent the contract awards. Lavery’s conclusion: “In such circumstances there is a risk that contracting to nonprofits in human services could become a new form of patronage politics.”5  Indeed.

I know of only one time when a politician running for office directly said, “If you contribute to my campaign, I will be sure to get you funded when elected.” Similarly, I have never been asked to “turn out the votes” among the welfare recipients I train. However, keeping a contract system that is largely discretionary and focuses on process rather than product permits implicit patronage arrangements to flourish. No matter how poorly your political ally manages his taxpayer-funded social-service program, a performance-blind contract allows him to escape judgment. Since social-service nonprofits have frequent and direct contact with a politician’s constituents, even the worst-performing ones are invaluable politically, and so the temptation to treat them as pools of patronage is strong.

These factors have combined to create a Welfare-Industrial Complex (WIC), similar to the Military-Industrial Complex (MIC) first identified by President Dwight D. Eisenhower. Ike coined the term to describe a growing coziness between the military and arms manufacturers leading to unnecessary government purchases and inflated costs. Similarly, the WIC is made up of program operators whose livelihood depends on government funding, as well as community activists who have an ideological aversion to market principles; on the other hand, these special interests are aided and abetted by government bureaucrats who cherish the former’s vote-grabbing outreach and whose livelihood depends on continuing to award new contracts. The cross-fertilization of complimentary special interests pursued at the cost of the public interest—so similar to the MIC—is sustained by performance-blind contracting. In this waltz of money and power, the tune—long-term jobs for welfare recipients—no longer guides the dancers. That must change.

For those contemplating the necessary changes to welfare-to-work contracting, I offer the following advice.

Establish an RFP process that is results-driven. Too often, government agencies decide to award contracts on factors that have little or nothing to do with the goals they hope to achieve. For example, the Indiana official I referred to earlier told me he also reviewed a bidder’s financial viability, prior criminal record and litigation history (i.e., have they ever sued the state?). The question of whether a bidder could cost-effectively put welfare recipients to work was never mentioned. Municipalities must establish clear goals on eligibility, length of placement and cost-effectiveness while imposing few procedural requirements. Measuring success by attainment of clear goals gives vendors the flexibility to experiment with different approaches, while giving the government a firm basis for cost and quality comparison.

Whom should a program serve? A municipality might choose long-term welfare recipients, or it might decide to focus on those recently signed on to welfare to stem dependency early. Or it could randomly assign different parts of its caseload to competing providers.

Randomly assigning clients to service providers will deflect the common charge of “creaming.” Providers who only get paid for success, critics say, will refuse to accept candidates less likely to do well. The truth is, no one can tell the difference between cream and sour cream. My staff believes you might as well toss a coin as try to predict who will do well and who won’t among our candidates. Some welfare recipients have insurmountable problems, but the vast majority are like the rest of us: blessed with some assets and burdened with some liabilities. If creaming is impossible, random assignment will eliminate it as a temptation for providers and as an excuse for pessimistic nonprofits.

The ultimate goal of welfare-to-work programs nationwide should be to prepare their clients to leave welfare and enter the workforce for life. Therefore, contractors should be paid only when a job placement is retained beyond the “viability” point—where there is a reasonable chance the client will stay employed without continued vocational support. Based on my experience in the welfare-to-work arena, four months is an appropriate threshold. This observation is supported by a study of Project Match, a Chicago nonprofit, which found that 40 percent of its program’s participants lost their jobs within three months of placement.6 Ongoing support for at least four months is essential to get clients through the difficult adjustment to self-sufficiency.

Government officials who oppose such a strict job-retention requirement on cost grounds often ignore the hidden costs of the current policy. While for-profit welfare-to-work contractors like America Works often charge more to prepare and place a client than do nonprofit providers, their service is cheaper in the long run because their clients are likelier to stay off welfare longer. The taxpayer pays twice for a welfare recipient who leaves her job and returns to the rolls: once for job training, and again for the renewed welfare benefits. A true performance-based contracting system needs to take both costs into account when assessing cost-effectiveness.

For example, if New York pays approximately $8,700 per year in welfare benefits for a family of three,7  is it worth spending $5,000 to guarantee that a client will be off welfare for a year? Any investor would jump at the opportunity to generate such a return. Moreover, many candidates will remain off the welfare rolls for longer than a year, meaning that the return on investment is even greater. Competitive bidding will, of course, determine the per-placement cost, but I believe that $5,000 per placement is about what government needs to spend to assure quality job training and job retention support—though it could be a bit less if contracts go to large programs doing 1,000 or more placements per year.

Monitor and audit the contractor to prevent cheating. To prevent fraud, government agencies must communicate frequently with contractors, other agencies, clients and employers, randomly reviewing case files and conducting work-site inspections and interviews. One common abuse is duplicate billing, charging twice for assisting a single welfare recipient; vigilant tracking of names and social security numbers can end this practice. A national database should monitor enrollment in welfare-to-work programs to ensure that only those eligible receive training and that trainees stop receiving benefits they no longer qualify for in their own—and any other—state. Vendors can also cheat by claiming a participant has a permanent job when he really has only temporary work; monitors might survey employers to determine how long and in what capacity participants have been employed. The New York State Department of Social Services provides a national example of an efficient and effective evaluation process.

Models exist for many of the items on this welfare-to-work wish list. One is the procurement system for welfare-to-work services in the United Kingdom. In the U.K., social-service officials develop specific program goals and determine funding levels and geographic target areas. Prospective bidders are then invited to meetings to discuss the program designs—their projected costs and the likelihood of success of their various approaches. Only after receiving this input does the agency craft an RFP, which is then issued to potential vendors for review and comment. The vendors submit bids after the RFP is finalized, and the government drafts a short list of vendors. Agency officials visit their sites to observe program operations and negotiate final details.

The contrast between the U.S. and U.K. approaches couldn’t be more striking. Here, bureaucrats make all decisions about program design without consultation with vendors. They then issue an RFP whose most salient characteristic is a complete lack of flexibility: vendors are forced to take it or leave it. Site visits are rarely conducted, and any negotiations are over costs. In the U.K., the final decision is based on broad information gathering on program costs and philosophies, and a truly competitive bidding process. Such early and ongoing input from service providers greatly improves the chances that programs will be designed and operated efficiently.

Hopeful signs have begun to appear, though. San Diego County recently launched a pilot program giving different organizations a share of its welfare-to-work program. It divided the region into six districts; contracts in three districts went to private companies, one went to a nonprofit, and two were awarded to county employees. The goal is for all operators to engage 100 percent of referred participants in work. To measure progress toward that goal, the contracts specify a variety of outcome measures, including work participation, 30-day employment, 180-day employment, and progress toward self-sufficiency. An independent evaluator is charged with continually reviewing performance goals to ensure that they remain appropriate and realistic. The initial contracts are for 13.5 months, and the county reserves the right to exercise three additional one-year options through June 30, 2002. The successful providers, we can expect, will be given greater responsibilities. This direct competition has stimulated new service-delivery ideas and innovative program designs. We eagerly await the evaluator’s findings.

Nothing spurs innovation like competition. Exposing the welfare-to-work industry to the vagaries of the market is the best thing not only for the taxpayer but for the industry’s clients, as well.

Like it or not, the American economic system is based on market principles, and the market works. Only companies that operate efficiently and provide quality service for a reasonable cost are able to stay in business over time. Eventually, poorly performing operations fall by the wayside. Vendors providing welfare-to-work services, whether they come from the for-profit or nonprofit worlds, must be forced to adhere to these same principles.

If we can do that, we’ll see new energy unleashed that will fuel creative and successful programs operating cost-effectively and achieving genuine reduction in the welfare rolls. For our part, we at America Works are eager for the competition this will foster—it will make us a stronger company. Welfare recipients will be able to improve their own lives and the prospects for their children’s lives, and contribute their part to the country’s economic well-being. Companies will get productive workers. And, just maybe, government social-service agencies, having been dragged into the market kicking and screaming, will emerge as the beacons of social change they’ve always promised they’d be.


  1. Steven Cohen and William Eimicke, Assessing the Cost Effectiveness of Welfare to Work Programs: A Comparison of America Works and Other Job Training and Partnership Act Programs, New York: Columbia University, April 19, 1996.
  2. New York State Department of Labor memorandum from Pete Landsberg to John Haley, June 11, 1997. Another study found that 80 percent of America Works clients were still off welfare 13.5 months after placement.
  3. David Osborne and Peter Plastrik, Banishing Bureaucracy, Reading, Mass.: Addison-Wesley Publishing Company, Inc., 1997, p. 17.
  4. Steven Rathgleb Smith and Michael Lipsky, Nonprofits for Hire: The Welfare State in the Age of Contracting, Cambridge, Mass.: Harvard University Press, 1993, p. 200.
  5. Kevin Lavery, Smart Contracting for Local Government Services, Westport, Conn.: Praeger Publishing, 1999, p. 155.
  6. D. Kennedy, “The Ladder and the Scale: Commitment and Accountability at Project Match,” Kennedy School of Government Case Program, Cambridge, Mass.: President & Fellows of Harvard College, 1992, p. 11.
  7. This severely understates the true cost because it excludes the costs of programs like Medicaid, eligibility for which is based on welfare eligibility. Nor does this figure include the bureaucratic costs incurred administering these entitlement programs.

Are you interested in supporting the Manhattan Institute’s public-interest research and journalism? As a 501(c)(3) nonprofit, donations in support of MI and its scholars’ work are fully tax-deductible as provided by law (EIN #13-2912529).