Those of us following human services have heard the terms "evidence-based practice" and "differential response" bandied about a lot lately. It seems that after a lifetime of avoiding it, the human services world is beginning to experiment with performance-based contracting. It's an approach that doesn't come naturally to those used to dealing with the most human side of public policy and administration.

While contract negotiation and management have never been human services' strong suit, there's the parallel problem that in many jurisdictions a handful of politically connected and powerful human services providers have often hijacked the contracting process, with public money going to historically influential providers rather than those with a proven track record. At the same time, doing performance-based contracting well requires an information technology system that's up to the job -- something the human services profession has always struggled with.

So it's interesting to read the Rockefeller Institute of Government's report on New York City's groundbreaking welfare-to-work performance-based contracting system, which pays providers to get welfare recipients off of public assistance and into jobs. To say that NYC represented a challenging testing ground for performance-based contracting is a mild understatement. When the Personal Responsibility and Work Opportunities Reconciliation Act passed in 1996, one out of eighteen Temporary Assistance for Needy Families (TANF) recipients nationwide lived in New York City. At its height in 1995, almost 1.2 million New Yorkers were on public assistance, or one out of every eight city residents.

This article appears in our Human Services newsletter. Click to subscribe.

When he was mayor, Rudolph Giuliani made reducing the city's welfare population a priority, and his campaign rhetoric was chock full of references to the need for jobs and the nobleness of work. He appointed Jason Turner to head the city's Human Resources Administration, who immediately turned to outside for-profit and not-for-profit agencies to ramp up job readiness and placement programs. He set up a two-track system based 100 percent on performance-based contracts. One track focused on readiness and placement; the other on placement alone. Why not try to get those who are work ready into jobs as quickly as possible?

Over the 15-year course of performance-based contracting, the city learned four basic lessons, according to the Rockefeller report:

First, and not surprising, they learned that how contracts are constructed is key. For example, the city's initial emphasis on placement seemed to work really well -- except as time passed, it became clear that a lot of those placements weren't sticking. The city had to figure out how to refocus contracts on job retention. As they evolved, the city started tracking how much jobs paid in an effort to reward vendors who found clients better paying jobs.

The second takeaway was that IT systems are critical to doing performance-based contracting well. The city invested in the systems necessary to managing the large amount of data that would soon be pouring in, and that would be necessary to accurately tracking results and reimbursements.

Third, different vendors behave differently when given the flexibility of a bottom-line approach to payment. That presented the city with both a challenge and an opportunity. The challenge was to tease out different strengths among providers -- since each was coming at their jobs differently -- and then use that information to manage contracts in a way that played to various vendors' strengths. Some honed in on more challenging clients (and higher payments); others had success in placements with different employment sectors.

Fourth, and related to one and three, contracts have to be flexible enough to evolve with lessons learned. As New York City officials figured out early on, just paying for placements wasn't good enough; contracts had to also focus on length of placements. As officials learned more about vendor strengths and client challenges, it was able to tweak incentives in a way that played to those strengths. For example, once easy-to-place clients were washed out of the system, the city had to focus more tightly on which vendors seemed the most adept at readying and placing various clients with varying obstacles to work.

For students of David Osborne's Reinventing Government, a quote from Jason Turner's first deputy Mark Hoover summing up the New York City performance contracting experiment will ring familiar (think steering and not rowing): "Government is best at setting outcomes, designing policy and overseeing and supervising performance. It is not great at operational activities and service delivery."

To some, that might sound a bit more ideological than practical. And contracting out isn't the answer to all that ails government. But in human services -- with its long established history of relying on outside providers -- it is clearly the direction we're headed.