To its many critics, President Lyndon Johnson's War on Poverty is often described as the classic 1960s social program: a well-intentioned but naive effort that spends billions of dollars year in and year out without making much of a dent in poverty.

But imagine if government only paid social-service providers if their efforts actually yielded the desired results. That's the idea behind "pay for success" programs, also known as social impact bonds. Under this concept, businesses and/or philanthropies provide up-front money for programs to address difficult social problems. If they achieve a set of measurable outcomes within an agreed-upon time period, the funders get their investments back, plus interest. If not, government pays nothing.

Pay for success programs seem to be attracting a growing number of bipartisan fans since I wrote about them in 2014, and two new programs illustrate how implementation of the concept is becoming more sophisticated.

Last month, Connecticut Gov. Dannel Malloy, a Democrat, announced a four-year initiative to keep children from 500 families out of foster care. Social workers from the Yale Child Study Center will focus on parents with substance-abuse problems as part of an intensive effort to keep the children in their homes.

No funder has been named yet for the $12 million initiative, but several have expressed interest. If successful, the state will reimburse the up-front money plus a 5 to 6 percent interest payment. It's a pretty good deal considering that Joette Katz, commissioner of the state's Department of Children and Families, told The Washington Post that Connecticut currently pays about $350 million annually for services to children in foster care and institutions.

In South Carolina, Republican Gov. Nikki Haley recently announced a $30 million, four-year program to send registered nurses who specialize in maternal and child health into the homes of low-income pregnant women to teach them parenting skills and ways to keep their children healthy. The effort is funded by foundations and a corporation.

The Connecticut and South Carolina programs highlight the opportunity pay for success presents for prevention programs that can yield long-term savings but might not get funded through the traditional appropriations process. The South Carolina program also addresses the potentially sticky issue of determining whether a program has achieved the agreed-upon outcomes by designating an MIT research group to conduct an evaluation. Such provisions enhance the integrity and perception of pay for success.

As a model that is dependent on what funders are willing to invest in, pay for success isn't the silver bullet for addressing every stubborn and costly human-services problem. But particularly during a time when some once again seeing economic clouds looming, they can be a valuable tool for helping our neediest citizens by focusing state and local government resources on the best kinds of programs -- those that actually work.

Charles Chieppo is a research fellow at the Ash Center at Harvard's Kennedy School. This originated at governing.com.