Discussion: (1 comment)
Comments are closed.
Our recommendations are grounded in our firsthand experience with government data systems as former leaders at the city and state levels, where we oversaw programs including cash welfare, food assistance, public health insurance, child welfare, homelessness, probation and corrections, and child support enforcement. In each of these policy areas, we witnessed immense potential for all levels of government—working with nonprofit organizations and academic partners—to harness the power of data to maximize the impact of taxpayer dollars and improve services for the public.
We believe what policymakers need most—especially at the state and local levels—is practical guidance for overcoming the myriad bureaucratic, legal, and cultural hurdles that prevent government leaders from unlocking the full potential of administrative data.
Our key findings culminate in five recommendations for policymakers:
First, tackle data security and privacy concerns by developing a clear and shared understanding of privacy laws, both within government and with the stakeholder community. Implement appropriate technology to ensure personally identifiable information will remain confidential.
Second, create standard definitions for reporting administrative data and require implementation as a condition of local, state, and federal funding. Technological solutions can offer users a “continuum of access” that is aligned with their legal right to see and use the data. Most importantly, data must be used by researchers and policymakers to improve their quality.
Third, take steps to instill a sharing and learning organizational mind-set. Implement a governance framework that is guided by shared values and transparency to facilitate appropriate sharing of administrative data.
Fourth, create ease and comfort with using and sharing data by implementing data sharing in a tiered approach and open greater access over time.
Fifth, at the federal level, standardize the collection of data and aggressively pursue data-sharing agreements with state and local governments. By linking administrative and survey data, US statistical agencies and independent researchers can more accurately report on Americans’ real conditions.
In addition, this paper offers exemplars of where this work has been done well at the state and local level to help address social challenges, from preventing child abuse to addressing homelessness. At the federal level, it provides a road map for the US Census—in partnership with other federal agencies—to take a lead role in driving systemic change in how we share and use data.
Every day we see how data are used to make our lives more convenient. Enterprises such as Google, Walmart, and Amazon are using data to ensure that the products we most want are at our fingertips. We cannot buy a new jacket in a city 300 miles from our home without being alerted almost instantly that our credit card is being used in a place we have never been. We do not choose a restaurant without first checking the online reviews from other customers, and we do not get there without following the quickest route—even taking into account the traffic caused by the fender bender that took place less than an hour before our departure.
At the heart of all these applications are millions of pieces of information about who we are, where we live and work, and how we earn and spend our money. Using data is the prized skill that allows the private sector to respond to our every need or desire. But in the places where effectively using data can mean saving a child from abuse, preventing high school delinquency, or helping a single mother secure a job and child care, the use of data to guide decisions is woefully inadequate.
To be sure, the issue is not that government agencies lack data (indeed they often have more data, more accurately collected than anyone else) but that the repositories of these data are highly protected and bureaucratically controlled. Much of the country’s administrative data—collected by government entities for program administration, regulatory, or law enforcement purposes—is underappreciated, underdeveloped, and underused.
At the same time, much of the information that is being used for government administration purposes suffers from a variety of inherent shortcomings— mainly related to the way in which the data are collected. Survey data, in which respondents self-report information, are often more biased and less complete than less-used administrative data, which a government entity typically collects. For example, the federal government’s Census Bureau relies on self-reported responses to collect and build repositories of information about who Americans are and how they live, but it struggles to overcome problems of nonresponse, lack of accuracy in self-reporting, and other issues that come with the territory of citizens telling the government about their own lives. Administrative data, which do not rely on self-reporting, can shore up the gaps in information left by relying on Census data.
Among the government’s administrative data collections, many are essentially internally maintained commodities. They have developed from the point of service delivery, at the program level. Sometimes they are held in the public agency, sometimes by a nonprofit provider under contract with the government.
Only a subset of state or federal funding authorities standardize data collection and reporting. Even then, the millions of pieces of data are not fully applied to policy evaluation or important statistical reporting. More data remain in agencies’ internal files and not even centralized within an agency. When Linda Gibbs was appointed by New York City Mayor Michael Bloomberg as commissioner of the Department of Homeless Services in New York City in 2002, she counted 32 separate data-tracking systems in just that agency—with little integration or sharing even within the agency’s four walls.
To truly unleash the data’s value, government needs a set of value propositions, tools, and structures through which it can satisfy its governmental obligation to make the resource available to the public in a meaningful way. At the core of our argument is a deep belief that public data sources—de-identified and privacy protected—should be used to benefit the public and advance social progress. Government has a unique ability, and obligation, to ensure equal access to public goods, including the vast knowledge contained in its data systems.
The ultimate goal of this paper is to explain how administrative data could be better used and more widely shared across all levels of government and made available to researchers, nonprofit service partners, policy experts, and decision makers who could leverage these data to improve outcomes. While sharing data has many barriers, none are insurmountable. The ability to link administrative data sets with each other and survey data offers significant potential to answer important questions that neither type of data can do by themselves. In this paper, we will highlight the most significant barriers and offer a set of recommendations and activities to help policymakers knock these barriers down.
Comments are closed.
A newsletter highlighting work on poverty—and efforts to reduce it—from AEI’s Poverty Studies team.
1789 Massachusetts Avenue, NW, Washington, DC 20036
© 2018 American Enterprise Institute