US policies on poverty have historically been prisoners of ideology. The War on Poverty of 50 years ago, embodied by legislation that created programs such as Head Start, Medicare and Medicaid, and food stamps ceded to a conservative backlash in the 1980s and 1990s. The backlash was premised on the notion that anti-poverty programs created a culture of dependency. That backlash culminated in wholesale welfare reform in 1996 under a Democratic President. More recently—until the victory of Donald Trump and the Republicans in Congress earlier this month—the pendulum appeared to be swinging back.

But our belief in or opposition to robust government anti-poverty programs continues to correlate closely with our position on the ideological spectrum—not on any objective measurement of a program's effectiveness. 

Despite this ideological bias, there is a basis for common ground. Driven by new data and an evidence-based, “what works” approach, we are capable, for the first time, of synthesizing enough information about enough people over enough time to make accurate observations about how—and how much—policies affect them. We can also more clearly see which combination of programs might be the most helpful at which moments in time. The data is already out there. And while there is sometimes a need to supplement that data to correct for inbuilt bias, there is no need to mine for more of it to achieve these goals.

Thus, a quiet, bipartisan revolution has been underway that is privileging this newly actionable data over ideology. Over the past decade, the federal government has moved away from ideology to impact, replacing inputs (such as a program’s organizing principles) with measurable outcomes (such as the Department of Education’s Investing in Innovation Fund, the Department of Housing and Urban Development’s various initiatives, and the White House Office of Social Innovation).

Guarding Against Ideological Monopoly

The prospect of one-party government could undermine the progress toward this new, fact-based approach and send us back to an ideological conflict over the need for—and usefulness of—certain programs. Nevertheless, faced with that threat, one thing Americans of all ideologies ought to be able to agree on is that many government programs need improvement—and it is possible that improvements will make certain changes more palatable to those who are ideologically opposed to them.

Many government programs, for example, measure metrics that correlate with success but are light on data that actually prove it. We monitor government job training programs to assess how many people enroll, how many participants attend training courses, and how many skills they have gained. But these metrics are (imperfect) proxies for the ultimate proof of the program’s effectiveness: How many people who enrolled in the program found and kept a good job? 

The emerging bipartisan consensus that we can use big data to answer fundamental questions about the efficacy of government programs has led to the creation of a commission chaired by Republican Speaker Paul Ryan and Democratic Senator Patty Murray. The Ryan-Murray Commission is tasked with using data to advance “evidence-based policy-making”—or rather, to evaluate which government interventions work best.

It is possible that we will witness the devolution, or even dismantlement, of many federal programs (a longstanding Republican aspiration). If so, an evidence-based approach will be all the more critical for informing the most-effective programs in states and localities.

Recent public-private partnerships in the social sector have proceeded on a similar premise: Government and private (usually nonprofit) companies partner on projects to improve lives in a local community; the private company advances the funds for the project, and is repaid only if the project demonstrates success, as measured by previously determined criteria. These “pay-for-success” initiatives permit government and the philanthropic sector to determine—with a minimum of taxpayer money—which interventions work best. Properly targeted, using large datasets to track outcomes as they emerge (rather than awaiting long-term, longitudinal studies) will allow us to more quickly determine which approaches work and to more rapidly scale up the most successful ones. 

But we must nonetheless handle big data with care. However vast the dataset or varied the data points, often the inbuilt biases of the systems (and their programmers) render results that are inaccurate, offensive, or just plain absurd. Witness, for example, the recidivism evaluation software that was twice as likely to flag black prisoners as high-risk for repeat offenses as white prisoners. Hence, we need to diversify the tech workforce and democratize data collection.

Technology in Service of Underserved Communities

We now have tools that can put data and technology to work in effectively fighting poverty. Here are four examples:

Crowdsourcing. Community organizations have rightly observed that too much big data focuses on poor peoples’ deficits (such as how much they owe, how bad their FICO mortgage score is, or how many services they consume) and hardly at all on their assets (who can find a neighbor a job, which kids aced their standardized tests, or where the best local entrepreneurial opportunities are). These community-level exercises in data sourcing and consequent action have resulted in remarkable rises in employment rates, salary levels, home ownership rates, and even school test scores for participants.

Projects such as the Family Independence Initiative address this shortfall in “assets” data” by allowing underserved families to generate the data and information they need to make informed choices about how to run their finances and make their own investments. 

Online Evaluation and Choice. Most of our lives have improved as innovations such as Yelp and Google started bringing instant information to our screens, and services like Amazon started delivering every manner of consumer good to our doorstep . But the Internet lacks information about and feedback mechanisms for critical services low-income consumers need. Any number of programs and apps can tell you all about local hairdressers, flower shops, and sushi bars, but try finding an online service that identifies and rates childcare facilities within a 10-mile radius.

With the right effort and planning, it ought to be a simple matter to create and incubate such scalable programs. The Internet is the single greatest tool we have for reducing barriers to information and access. We can now get answers to questions that used to take days or weeks to gather, in seconds. And while it may take thousands of dollars to move from a poor neighborhood to a rich one, we can all inhabit the same free space online.

Online Service Delivery. Today, most government safety-net applications are anachronistic and only sporadically starting to make use of online efficiencies,. The private sector, meanwhile, has unsurprisingly been more interested in designing applications to serve wealthier consumers. We need to incubate partnerships between the private and public sectors to connect the substantial proportion of needy Americans with online services designed for them. And we need to use online data to address their needs, much as companies keep pace with consumer needs by tracking their shopping and browsing history.

Some governments and nonprofits have begun to streamline programs and increase accessibility in this way. Idaho, for example, has consolidated its benefits access applications and management online, in one website. This consolidation allows users to apply for programs from a single site, rather than inefficiently accessing several different sites (or physical locations) for each program.

Data and Technology to Expand Opportunity. Consider: More than half of highly qualified high school students from underprivileged backgrounds do not apply to any selective colleges—largely because they lack the information and local support to apply for financial aid, or even to apply to the colleges in the first place.

A groundbreaking partnership between Bloomberg Philanthropies, the College Board, and the College Advising Corps, among others, is focused on using data from SAT scores to identify these high-achieving, underserved students wherever they may be, and then advising them on the college application process through video conferencing, text messaging, and other remote means. And while any such nationwide initiative would obviously benefit those young men and women who might not otherwise have attended college, it would have the salutary collateral effect of providing a generation of new Americans with the skills necessary for the nation to compete in the global knowledge economy.

Seizing the Moment

Even at its most efficient and unbiased, big data and technology will not solve all our country’s socio-economic problems. It will not usher in an age of racial harmony and gender parity, nor will it eliminate poverty and inequality.

Still, there is every reason to put our 21st-century technology to use in service of a 21st-century system of opportunity for all. We no longer have the excuse that we lack the means to target people in need and track the outcomes of interventions. We should use this moment of political change to advance an anti-poverty agenda that is rational, bipartisan, and achievable.