State employees in Arizona are in head-to-head competition with the Maximus Corporation. Roughly half of Maricopa County's Temporary Aid to Needy Families program is being run by state welfare workers. The other half is being administered by Maximus employees. If Maximus outperforms the state workers, the company will reap potentially significant financial rewards; if the state workers best the Maximus workers, the company will lose the Maricopa County business.

This is the latest experiment in trying to answer a central question around welfare reform: Can the private sector do a better job than the public sector at administering the new world of public assistance?

It is a question that has only gained in prominence since the federal government's high-profile rejection of Texas' attempt to privatize administration of virtually its entire welfare system in the wake of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996. Despite that setback, neither for-profits nor their advocates gave up. They have continued to push for the welfare administration business--specifically, for the right to determine eligibility--by chipping away at the local-government level. And they have accelerated their longstanding push into the market for administration of welfare- to-work efforts, the highly visible centerpiece of the new welfare reform law.

In fact, private companies had been active in the welfare business long before the 1996 reforms (Maximus first got into the business in 1988). But the federal law, by removing barriers to privatization of the administration of TANF, created vast new opportunities. Lockheed Martin IMS officials say welfare reform services now represent one of its two fastest-growing business lines, with the number of government clients going from zero to 25 in the two years after federal welfare reforms passed.

The argument the for-profits make is a familiar one: that the private sector--fleet of foot, and with employees unencumbered by the old- style "just-write-them-a-check" mentality--will be able to handle the whole job more cheaply and effectively than government, especially if it is given the broadest range of responsibilities. "It's an argument for maximum flexibility in how we handle cases," says Steve Minnich, formerly administrator of the Oregon Adult and Family Services Division, now vice president for welfare reform services with Lockheed Martin IMS. "It allows us more options to do total preparation for independence."

Despite the philosophical and practical arguments for allowing for- profits to do eligibility determination across the three programs that underpin public assistance in the United States--TANF, Medicaid and food stamps--it is that frontier of welfare reform that triggers the most vehement opposition. With money as their central motivation, the critics insist, the private sector just won't do as conscientious a job as government or not-for-profits in trying to help people. The for-profits, the argument goes, will "cream" the easiest-to-handle cases, cut corners and push people into work who may not be ready. And if part of their reimbursement is based on cutting caseloads, they may not aggressively recruit people into the system, or they might simply deem them ineligible when they do show up. Even some of those who think there may be value in privatizing welfare-to-work services balk when it comes to turning over the job of deciding eligibility. "I'm not against privatization in some areas," says Vincent Wood, head of the Children and Family Services Division in the Arizona Department of Economic Security, "but I think government has to be the gatekeeper."

There is one other group, of course, that argues strenuously that only public employees ought to run public welfare programs: public- sector organized labor. Credited with being a key force behind the White House's ruling disallowing the waivers that would have made the massive Texas privatization plan possible, unions have argued strenuously--and in many places effectively--that the sensitive job of determining eligibility and benefits ought never to be turned over to for-profits. The American Federation of State, County and Municipal Employees, in particular, is running an aggressive campaign to discredit the work of for-profits in social services, including publishing long lists of canceled contracts and unfulfilled performance promises.

At the moment, there are only two places where for-profits actually are determining TANF eligibility and benefits. In Maricopa County, where the head-to-head experiment was ordered by the state legislature, and in Milwaukee County, Wisconsin, where the county's entire welfare system has been privatized. Meanwhile, a much more audacious experiment in turning eligibility over to for-profits may be coming soon. In West Palm Beach, Florida, members of the regional Work and Gain Economic Self-sufficiency Board (the public-private WAGES boards are implementing welfare reform in 24 regions statewide), along with Lockheed Martin officials, are planning an end-run on prohibitions against non-governmental entities handling eligibility determination for Medicaid and food stamps. Essentially, for-profits would do the determination, then pass the case through a government employee to satisfy federal oversight rules. If successful, it could signal a whole new era in the private sector's role in social services delivery.

The obvious question at this point, of course, is whether organized labor's free-swinging criticism is accurate, or whether for-profits have proved themselves to be measurably more efficient than government in handling welfare. Because no formal study has been made of the performance of government versus for-profits versus not-for-profits, some argue that it's impossible to say right now. The broadest generality that can be made to date, says Amy Brown, a program analyst with the Manpower Demonstration Research Corporation, a not-for-profit organization that has been studying the effectiveness of welfare-to- work programs nationally for the past 25 years, is that no single sector stands out as being the best. "There are examples of great programs and lousy programs run by all three sectors," says Brown.

In Florida, which has one of the most aggressive privatization attitudes of any state, the two top-performing WAGES boards have taken two different routes: One has contracted with private-sector companies exclusively; the other is working with a local community college. Even the same company seems to perform differently from one WAGES region to another within the state. Lockheed Martin IMS was widely criticized for its welfare-to-work effort in Miami-Dade, which in fiscal year 1998-99 cost more than $38 million. With caseload reductions lagging well behind the state average, the welfare-to-work contract was turned over to a local community college. On the other hand, the company has been commended for its work in West Palm Beach, where the WAGES board spent $6.2 million on welfare-to-work services in FY 1998-99 and cut its caseload by almost 45 percent. "We had 5,600 clients when we started, and we're now down to 1,600," says David A. Luhrsen, head of the regional WAGES board. "Lockheed hired more people and invented new systems to help do that."

In Orange County, California, which spends about $20 million a year administering welfare-to-work programs, government employees are responsible for doing job placement for about 40 percent of the locality's TANF caseload; for-profits handle the other 60 percent. Because the county carefully tracks placement and casework costs, says Angelo Doti, director of the Family Self-Sufficiency Division in the county's Social Services Agency, comparing government and for-profit performance is pretty simple: The quality of placements are essentially the same, he says, and when it comes to per-case costs, "I know that we're operating within pennies of each other."

Given that so far there is little evidence that one sector offers any significant advantage over another in cost or performance, why bother privatizing?

The most obvious reason, say welfare administrators who use private- sector contractors, is that they offer maximum staffing flexibility for government: When a government finds itself facing a large increase in casework, it can quickly turn to a contractor to cover the overflow. Should the caseloads suddenly subside, it's simple enough to scale back the contract.

When Orange County's welfare-to-work caseload soared under California's version of welfare reform, CalWORKS, it made much more financial sense, says Doti, to handle the overflow through contracts than through the hard-to-initiate, hard-to-reverse and structurally expensive process of adding new government employees.

That's the reason Los Angeles County is now considering a return to contracting out, says Raul Ramirez, human services administrator with the county's Department of Public Social Services. In L.A. County, a Maximus contract for welfare-to-work services was not renewed after the election of new county supervisors who opposed privatization. Suddenly faced with a crush of 100,000 new TANF cases, however, that same board is now reviewing responses to a request for proposal to contract out its huge overflow of welfare-to-work cases. Maximus argues that it's a time-proven approach. "When we were invited into Fairfax County, Virginia, it was to deal with a peak caseload," says Holly Payne, president of Maximus' Welfare Reform Division, which is currently doing about $50 million a year in welfare administration and welfare-to-work business. "We worked that off, packed our bags and went home."

But as simple as that might sound, Doti cautions, contracting out is in fact a very tricky and complicated business. The work a government gets out of a contractor is essentially only as good as the contract it negotiates, the quality of contract oversight and the extent to which government plays hardball in holding contractors to performance goals. "It's not easy or inexpensive," he says. "There is a lot of cost associated with designing contracts and then policing performance. You have to nail contracts down to the nickel and dime and nth degree, and then you'd better be ready to staff up to monitor the contractor's performance, both financially and programmatically. They'll test you a bit." Furthermore, Doti points out, it takes a while to develop a good working relationship with a contractor. "Ultimately, what you want is a partnership." In fact, Doti questions whether a competition such as Maricopa County's is a good thing at all. "You want to foster cooperation between the private sector and government," he says, "not conflict."

One kind of competition that seems unavoidable if a jurisdiction wants to contract out, though, is the competition for contracts, Doti acknowledges. "We didn't properly estimate the divisiveness and the appeals and the complaining among competitors about the whole contracting process, which made it all that much more energy- and time-consuming."

In fact, one of the clearest impacts of the whole social services outsourcing phenomenon has been the increasing sophistication and complexity of the contracts involved. Specifically, a whole new generation of "performance-based" contracts is emerging, in which payments are made only after contractors achieve certain performance milestones: typically, a percentage of the total when a client signs up; a bit more upon assessment and training; more upon a job placement; and then final payment only if the client stays employed for a set period of time. But while such a contract might sound foolproof, it can work to the private sector's advantage and government's disadvantage.

In Florida, for example, the state legislature has outlined contract parameters that call for paying service providers 50 percent upon accepting a client, 25 percent upon job placement and the final 25 percent if the client stays on the job at least six months. Phyllis Busansky, executive director of the state WAGES Board, says that such incentive contracts are a good first step, but that they don't differentiate between those who've been through the process before and those who are new to it. In other words, if a client isn't placed in a job or doesn't stay on a job six months, he or she can simply cycle back into the program, and the contractor is reimbursed its 50 percent for enrollment all over again. "I'd like to see contracts where there's more of a distinction between new and repeat clients," says Busansky.

It is that type of detail that will continue to be debated as the massive national welfare-to-work experiment evolves, because one thing is certain: Contracting out, regardless of complications, critics and failures, is only going to increase. "Organized labor may fight it," says Doti, "but the train has already left the station." And that increase is very likely to include the controversial job of determining eligibility--certainly when it comes to TANF, and probably eventually when it comes to Medicaid and food stamps, as well. Some hint of how for-profits would handle that particular job can be found in Milwaukee County, which has become a high-visibility national experiment in an aggressive emphasis on welfare-to-work. Under Wisconsin Works--known nationally as "W-2"--the county has been divided into six regions, all run by for-profit or not-for-profit contractors. The contractor that has been receiving the most attention is Maximus.

Given that Maximus makes money only if it saves some of the $58 million it has been given by the state to run TANF for two and a half years, the Milwaukee County setup would arguably be a textbook case of a for-profit having every incentive in the world to find as few clients and provide as little service as it could get away with. "Actually, they're running an aggressive program in reaching out and converting people to W-2," says Tom Corbett, associate director of the University of Wisconsin-based Institute for Research on Poverty. Researchers report that Maximus' on-the-job performance has been at least as good as that of the not-for-profits working their sectors in Milwaukee County.

While some conclusions about the performance of for-profits might be gleaned from Milwaukee, it is Maricopa County that may end up being the case study to watch. Under the terms of the Maricopa County contract, Maximus will be rewarded in varying proportions for reducing caseloads; for job placements; for job placements where salaries are $3 above minimum wage; for job placements with health benefits; and for outperforming the workers in the half of the county where TANF is still government-run. That comparative performance is being evaluated by an outside contractor.

George Leutermann, who heads up Maximus' Milwaukee County program and who helped negotiate the Maricopa County contract, states flatly that Maximus can do the job better and cheaper. "With government, you pay for process," says Leutermann. "When you contract out, you pay for results."

In a world where the view of welfare has undergone a 180-degree shift--from focusing on process to focusing on results--that is an argument that gets people's attention. It has certainly been the attitude behind the groundswell of support for contracting out in welfare-to-work. If experiments in running the whole show--such as those in Milwaukee and Maricopa counties--play out favorably, it could well lead to a 180-degree shift in the view of whether the private sector ought to indeed be handling more of the job of eligibility and benefits determination. And if Lockheed Martin IMS and West Palm Beach succeed in their plan for privatizing eligibility determination for Medicaid and food stamps, then that shift could pick up speed very quickly, indeed.