June 28, 2022
Robust federal assistance programs and social services are essential to a thriving society. This is especially the case as people continue to contend with the fallout from the COVID-19 pandemic, which jeopardized livelihoods and put millions out of employment. Government benefits at the federal, state, and local level help people across the country pay for food, housing, health care, and other basic living expenses. But more work is required at the federal level to ensure that these benefits reach everyone in need. For instance, the historic $1.2 trillion Infrastructure Investment and Jobs Act signed into law last year included a $14.2 billion program called the Affordable Connectivity Program (ACP) to help qualifying low-income households pay for internet service. While the program is off to a strong start, improved data sharing between federal agencies, state and local governments, and institutions can leverage existing data from other benefits programs to streamline eligibility processes and ensure those who qualify receive the benefit. Expanding data sharing for benefits eligibility also aligns with one of the goals in the recent executive order to advance racial equity.
We discuss how data sharing could be improved, as well as other steps that the federal government can take to maximize the impact of this benefit on the digital divide. The solutions outlined here can be applied to both current and future programs that help people find housing, prepare children for school, and ensure everyone has enough to eat.
What Does Data Sharing Look Like Now for Broadband Affordability?
The Affordable Connectivity Program (ACP) builds off of the temporary Emergency Broadband Benefit (EBB), which used the National Verifier—a centralized application system established by the Federal Communications Commission (FCC) and operated by the Universal Service Administrative Company (USAC) for the Lifeline program—to confirm applicants' eligibility. Eligibility is often confirmed through one’s participation in other federal assistance programs—including Medicaid, Federal Public Housing Assistance, and the Bureau of Indian Affairs’ General Assistance, among others—which qualify applicants for both Lifeline and EBB benefits. Both the EBB and ACP widened eligibility criteria beyond those qualifying categories for the Lifeline program so a wider number of low-income households could afford internet service during the pandemic. Doing so, however, meant that the FCC and USAC would need to access additional data sources to determine eligibility under new categories. The Lifeline National Verifier (LNV) currently has automated connections at the federal level to the Federal Public Housing Assistance and Medicaid databases, as well as connections with 22 state and territory databases, but the USAC is working on expanding the LNV’s connections to various federal, state and local databases for expedited benefits delivery. Applications that cannot be verified through currently available databases require manual reviews of submitted documentation, which can slow down application approvals and benefits.
The Role of Data Sharing and “Cross Enrollments” In Improving Assistance Programs
Leveraging eligibility data from other government programs and promoting “cross enrollments'' is a baseline necessity to maximizing benefits. Cross-enrollments link existing data and/or eligibility determinations from one public benefit program to determine eligibility for another program, reducing both administrative steps and burdensome enrollment procedures. This also allows applicants to bypass having to manually provide some or all of the necessary documentation. In the case of the ACP’s National Verifier system, USAC already shares data with the U.S. Department of Housing and Urban Development (HUD) to verify participation in the Federal Public Housing Assistance program (FPHA) and with the Centers for Medicare and Medicaid Services (CMS) to verify participation in Medicaid. Cross-enrollment would allow qualified FPHA or Medicaid recipients to automatically be certified for Lifeline and ACP. In order to add new eligibility categories for the ACP, such as participation in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), USAC would need to identify and establish connections with databases that could be used to automatically verify eligibility based on participation in WIC. This means that USAC might have to enter into memoranda of understanding (MOU) or computer/data matching agreements (CMA) with each state and local agency to automate the process of verifying applicants and administering the benefit. USAC would also need to enter into interconnection security agreements with relevant state and local agencies to ensure the data being accessed will be protected in accordance with federal standards for privacy and information security to protect the personal information of WIC participants.
Data linkages can also help with renewal of program benefits. Most programs require periodic re-verification of eligibility, usually annually. This can be an onerous process resulting in eligible participants losing their benefits. Through data sharing, recertification from one assistance program can be used to extend eligibility for another program.
Identifying Barriers to Sharing Data for Better Benefits Delivery
Legal issues, capacity constraints, fragmented data systems, and privacy concerns across federal and state governments can pose significant challenges to sharing data that would streamline cross-enrollments for many benefits programs.
- Perceived legal and privacy barriers: Many government administrators are wary of developing data-sharing agreements with other agencies for fear of legal concerns or issues of privacy in data sharing. These concerns, often unfounded, can delay MOUs and other types of sharing agreements.
- Outdated and incompatible legacy data infrastructure: Just as benefits programs have been developed in fragmented ways, the systems developed to support them were implemented in fragmented ways. Many parts of the government data infrastructure at the federal and state level are outdated and unable to effectively keep up with data demands from new programs like ACP. This is not a new issue; the Government Accountability Office in a 2016 report described the federal government’s heavy reliance on legacy systems.¹ State eligibility systems are in need of modernization as well, as past efforts such as the funding of integrated eligibility systems (IESs) have had mixed success.
- Capacity constraints: Agencies often lack clear procedures for sharing administrative data. Navigating this process, which includes issues of data quality, lack of data documentation, and interoperability and compatibility for linking data lead to capacity and budget constraints that hinder data sharing.
Improving Data Sharing and Infrastructure Is Key for Greater Assistance
The challenges of improving government data sharing and infrastructure are not new. Various attempts have been made over the last three decades to improve and better integrate government data systems, through periodic legislation enacted to address the issue, such as the Paperwork Reduction Act of 1995 and the E-Government Act of 2002, and other administrative programs and pilots, such as the federal State Systems Interoperability and Integration Project. However, there are currently a number of promising legislative and technical initiatives to improve government data sharing, including the following:
- The COVID-19 pandemic highlighted the need to improve and accelerate government data sharing, and sparked efforts to improve the public data landscape, such as CDC’s Data Modernization Initiative (DMI), a large-scale project to modernize the data infrastructure of federal and state agencies. The DMI aims to standardize data use and sharing agreements, transform legacy public data systems, and improve interoperability across federal and local data systems. Though the goal of the DMI is to improve sharing and reporting of health data, the establishment of this framework would positively impact sharing of benefits eligibility data as well.
- The current Federal Data Strategy Action Plan calls for accelerating practices that “increase the sharing and use of data for federal decision-making and operational needs,” and the Chief Data Officers Council recently solicited feedback on how best to achieve that goal.
- Improving data matching algorithms for data sharing, an important aspect of streamlining data linkages, is currently being addressed by 18F, a special technology and design consultancy within the General Services Administration. They have launched an Eligibility APIs Initiative to develop application programming interfaces (APIs) to help federal, state, and local governments update their benefits systems when policy changes. When new benefits programs that require data across various agencies and states—such as ACP—are created, APIs can allow for clearer communication and data linkages for eligibility determinations. Beyond the API work of 18F, capabilities are also needed in the government to employ emerging analytical methods for matching data that may drive more accurate data linkages, including machine learning and neural networks.
Improving data sharing to enable more cross-enrollments will involve solving challenges in several areas:
- A culture and clear framework that enables benefits data sharing: Federal law and most states authorize data sharing for appropriate governmental purposes, including benefits administration. However, legal and privacy concerns are consistently identified as a significant barrier to creating data sharing agreements. While this sometimes results from misinterpretation of laws and policies, it is often due to ambiguous or inconsistent federal and state laws, and absence of clear regulations. This ambiguity gives rise to a culture that is risk averse beyond the actual risks posed by legal rules, leaving a status quo of ad hoc data sharing and lengthy sharing agreement processes. There is a need for clear legal and organizational policies that create a foundation to support the sharing of data while protecting privacy.
Now that federal agencies and most states have chief data officers, CDOs can provide the leadership to establish a culture of data sharing through setting the tone and direction, prioritizing sharing and collaboration, and developing clear data and privacy policies that can accelerate the formation of data sharing agreements.
The Federal CDO Council and the State Chief Data Officers Network are well positioned to build a culture of data sharing in government and to provide guidance and clarification on appropriate data linkages. The CDO Council has a Data Sharing working group, and the State Network is engaging in projects such as generalizing MOU language to streamline the sharing of data. We recommend both groups prioritize facilitating data sharing for purposes of benefits cross-enrollments in their work. This work should include cataloging the regulatory and legal restrictions on data sharing for benefits eligibility and renewal, providing sample data sharing templates that agencies can use in developing their data sharing practices, and identifying opportunities for cross-enrollment. The chief dataofficer groups should draw from the expertise of organizations such as Actionable Intelligence for Social Policy (AISP), which works with state and local governments to develop data sharing capacities.
- Building capacity and an interoperable data infrastructure: Lack of data documentation can make data sharing processes difficult and time consuming. Many agencies do provide useful data dictionaries and documentation, but there is a lack of standardization for data documentation across government, as well as gaps in updating shifting data definitions and other important data changes over time. Sharing data can thus require significant time investment from personnel, who may need to acquire knowledge about key data attributes and variables across datasets. Current government data standardization initiatives, such as the Data Standards Repository (established by the 2020 Federal Data Action Plan) and the National Information Exchange Model, should be leveraged to streamline data sharing for benefits eligibility.
To enable and streamline data linkages, we recommend the development of further cross-government projects (such as the aforementioned API Eligibility Initiative), as well as innovative projects to develop tools and applications that allow for enrollment in more than one benefit at once.
Funding, of course, is an oft-cited barrier to data sharing, and increasing data sharing for benefits eligibility will require significant investments in building a more interoperable data infrastructure, modernizing systems (such as state integrated eligibility systems), and developing administrative capacity for linking data.
What about Privacy?
Privacy must be a priority in efforts to streamline benefits eligibility through expanding data linkages. Vulnerable populations served by benefits programs, who may be wary of data sharing, need to know that strong privacy protections are in place. In expanding cross-enrollment, programs should generally:
- Minimize data collected: People should only need to share the minimum information necessary for determining eligibility. The less information collected from the start, the less personal information is available to be misused and circulated beyond the user’s control.
- Restrict data sharing: Data should only be shared and circulated to fulfill program purposes. Restrictions around how data circulates also improve privacy, as fewer parties having access to people’s information makes this information less likely to fall into the wrong hands.
- Inform and make transparent how data will be used and shared: Users should know how their data will be used and shared ahead of sharing that data. Notice and consent certainly isn’t sufficient by itself, but people should have a baseline understanding of what’s happening to their data, and transparency into these processes.
Currently, under the CMPPA, linking electronic records for administrative purposes related to financial benefits requires agencies to assess the risk of data linkage and develop procedures to protect the data. For privacy protections, this is generally interpreted to mean adhering to data minimization principles, establishing clear data retention guidelines, and ensuring that shared data is linked in a secure environment with access strictly limited to authorized personnel for reasons related to program administration. As government data linkages increase, however, continuing to maintain public trust that data privacy is adequately protected requires adopting emerging privacy technologies as they become available.
A large part of the answer to this challenge may come from work started by the Commission on Evidence-Based Policymaking (CEP), which looks at the country’s data challenges as a whole, studies potential solutions, and provides recommendations. While the commission is exclusively concerned with data that would be used for statistical purposes, its work may ultimately drive improvements in administrative data sharing for benefit eligibility, including improved privacy protections. The Evidence Act signed into law in 2019 made many important changes, such as mandating that federal agencies appoint chief data officers and requiring agencies to make data more open and accessible—with data presumed to be sharable unless prohibited by law or regulation. The CEP also recommended establishing a National Secure Data Service (NSDS) that could conduct temporary data linkages between agencies. Moving forward, benefits programs such as the ACP can further strengthen privacy by taking advantage of the forthcoming privacy designs of an NSDS.
Another one of the CEP committee’s recommendations for a secure data service is to model best practices for secure data linkages by piloting the adoption of privacy-enhancing technologies.² A particular emerging privacy technology cited by the committee is secure multiparty computation (SMC). SMC allows two (or more) parties to perform calculations and functions involving all of their data sources, without any party having to reveal their private data to anyone else. SMC relies on cryptography to compute answers over distributed data sources while keeping data encrypted at all times, without any party able (or needing) to see individual records of the other. Programs like the ACP could potentially use a special case of multiparty computation known as a private set intersection (PSI), running a protocol periodically to determine whether individuals match the criteria for benefit eligibility without having to reveal any underlying information.
For eligibility verification in ACP and other programs, another emerging cryptographic method known as zero-knowledge proofs could potentially be used to enable verifications without the need to transfer any personal information. Zero-knowledge proofs work by passing a number of values from the person needing verification (such as an applicant to the ACP program) to the verifier (such as the National Verifier). The values passed are meaningless if intercepted, but they are values that only someone actually possessing the correct personal information, such as an SSN, could generate (as part of a cryptographic computation). The verifier checks these values to confirm a match with the person—all without the need to transmit any personal information.
Streamlining enrollment processes through sharing data across programs can help individuals and families receive benefits from ACP and other programs without burdensome, redundant applications, eligibility screenings, and verifications. At the same time, reducing steps in eligibility determination and recertification can realize administrative savings through increased efficiencies and the reduction of improper payments. The COVID-19 pandemic has changed the status quo by highlighting obstacles to government data sharing and the need to enable and accelerate better sharing.
At all levels of government, the lack of clear guidelines for data sharing—combined with penalties for improper data use—results in a reticence to share data and inhibits the completion of data sharing agreements in a timely manner. Establishing clear legal and leadership guidance at both the federal and state levels on allowable data sharing and data disclosures, while working to improve data interoperability and infrastructure, can help build more efficient benefit verification processes and increase program enrollments to deliver vital help to those in need.
This work was funded in part by the Bill & Melinda Gates Foundation. The views expressed are those of the authors and should not be attributed to the foundation.
¹ Three years later, the agency found that only three out of the 12 agencies they had examined had implemented their recommendation and made progress in planning to modernize their legacy systems.
² p. 12.