Can Social Impact Bonds Help Reduce Homelessness? (Governing)

BY J.B. WOGAN

On any given night in Santa Clara County, Calif., more than 6,000 people are homeless. Annually, that's costing the county more than $500 million. To Dave Cortese, president of the Santa Clara County Board of Supervisors, such a high cost should come with better results. The solution, he says, is obvious: "Devise a program that rapidly treats those folks and turns some from persistently homeless to consistently housed, and you cut down on the safety net they're using."

But that kind of prompt and comprehensive response is difficult for local government, in part because of the high upfront costs. Three years ago, Cortese heard about a new financing tool that tapped into the private and philanthropic sectors for early investors for otherwise cost-prohibitive public programs. If the program worked, the government would use future years' revenue to pay back its investors. The tool, known as a social impact bond or "pay for success" program, was new to Cortese. Even though it was called a "bond," it was more of a public-private partnership for experimental and expensive interventions in human services.

Last month, Santa Clara County announced Project Welcome Home, the latest local government initiative that leverages the social impact bond model. In the next six years, a nonprofit called Abode Services will provide housing and support services to between 150 and 200 long-term homeless people. The nonprofit will assign small caseloads to a multidisciplinary team with training in psychiatry, substance abuse, social work, nursing and vocational rehabilitation. The approach represents a combination of evidence-based practices, and is backed by academic research and recommended by the U.S. Department of Housing and Urban Development.

A group of funders is providing $6.9 million -- mostly in loans -- to make the project happen. Project Welcome Home's goal is to house at least 80 percent of participants for a year or more. If the program is successful, the county will reimburse its lenders as each person hits certain tenancy milestones. For example, lenders will initially be paid $1,242 for every individual who stays housed for three months. The largest reimbursement comes after a formerly homeless person remains in housing for a year:

In the next six years, the county has agreed to set aside about $8 million from its general fund to pay back its lenders if Abode is successful in keeping people housed.

Part of the appeal of social impact bonds is that they force local governments to account for current public spending on a problem and estimate cost savings if it reduced the problem using a relatively new and expensive intervention. That logic is driving the Santa Clara project as well. A study published in May by the Economic Roundtable, a policy research nonprofit in California, found that more than 2,800 people are chronically homeless in Santa Clara County, and each of them costs about $83,000 a year in public spending. Cortese and the rest of the Santa Clara County Board of Supervisors are betting that as the homeless population drops, some of that $83,000 in spending at local jails, emergency rooms and shelters will drop as well.

While the Santa Clara County social impact bond is certainly an effort to control spending on homeless services, local officials have been careful not to make reduced costs the sole objective.  "We are not predicating success based on the amount saved," says Greta Hansen, a county attorney overseeing Project Welcome Home. The primary goal, Hansen says, is to house people and keep them housed for a long time. The project also may bring other benefits that county officials consider desirable even if they don't translate into direct cash savings. For example, wait times at emergency rooms may go down as fewer homeless patients make frequent and repeat visits. That's a "noncashable" improvement in service delivery, Hansen says.

Another noncashable benefit is the increased focus on data collection and performance measurement in a human services context. Independent researchers from the University of California, San Francisco, have designed an evaluation tool to track treatment groups and control group to determine if Project Welcome Home can be credited for improvements in participants' health or decreases in their use of social services. "For so long, it's been a bit of mystery how impactful the [support] services we deliver are," Hansen says. "We were paying the same amount for a service provider who was extremely effective as one that was less effective. We want to tie our expenditure of public dollars to the outcomes we want to see."

Santa Clara County's Project Welcome Home comes at a difficult time for social impact bonds: It launched less than two months after the first and most famous social impact bond in the United States came to an early end. Group therapy for juvenile inmates, the intervention being tried in Rikers Island, N.Y., proved to be ineffective, so the primary funder, Goldman Sachs, pulled the plug. The contract in Santa Clara County includes a similar clause that allows funders to discontinue the project if Abode doesn't meet its housing targets. "There is risk," says Cortese. "No one is saying there isn't risk." But, he counters, anyone questioning the net benefit of the project should consider the counterfactual, "what you would spend on homelessness if you kept doing what you've been doing." That's about $3.1 billion over six years, according to the Economic Roundtable study.

While failure is a possibility, so is expansion. The initial project only deals with a small slice of the county's overall homeless population, but that could change. If enough participants reach their first set of tenancy milestones, Cortese says he would want to explore ways to scale up the program as early as June of next year.

Santa Clara County is one of eight U.S. jurisdictions with a social impact bond project that is up and running. The other seven are in Massachusetts, New York, Ohio and Utah. While they are alike in using nongovernmental funding to cover high upfront costs for an intervention, they seek to address different issues, such as disparities in early childhood education, high prisoner recidivism and chronic homelessness. Other than the Rikers Island project, none have reported final results.