7th Meeting of the Federal Advisory Committee on Data for Evidence Building
Washington, D.C.
Good morning to my fellow members of this committee. As the last of the committee’s members to offer a presentation during our fact-finding phase of work, I want to take this opportunity to pull together some of the threads of our work and offer some high-level observations. For those who know me well, I’m also going to encourage that it’s time for this committee to lean in, offer advice, and begin to make real and rapid progress in our efforts.
Much has happened in the last four years, over the past year, even over the past three months that affects our society and our government’s operations. Across all of this time, the concept of a data service and the need for using data linkages to drive valuable insights for policymakers and society has not lessened. If anything, the need appears even greater today than we realized 4 years ago when writing the Evidence Commission’s recommendations. A global health pandemic, a call to address racial disparities and inequities in policy, and a tremendous economic shock with implications on mobility and success – are all major issues for our government and society to address that require data capabilities.
Our country is at a moment where our data can be used to help solve real problems. But our government is still not effectively organized to do it.
Over the past six months, this committee has heard from experts on issues spanning from privacy and transparency to data governance, to statistical agencies and CDOs, to evaluation and performance, and intergovernmental considerations. We’ve taken a functional lens; we’ve taken a topical lens. We have gained a sense of activities that are underway across government and support structures available outside of government. To me, these presentations have all reinforced key themes from those before us -- the Evidence Commission, the issues the Evidence Act intended to address, and areas the Federal Data Strategy worked to resolve.
Current Context for the Committee’s Work
So where are we now? The Biden Administration has already offered high-level policy statements that promote the use of data and evidence, setting a foundation for our work. But unfortunately, more than two years after the enactment of the Evidence Act, OMB has not yet issued critical regulations for data access under CIPSEA Part D. The OMB staff are fantastic and deserve much of the credit for our convenings to date. Yet these important regulations and the focus on this committee’s work is not yet receiving the level of prioritization at OMB that is needed in this moment. OMB understandably has a lot of competing priorities so I hope this committee can come together and call on OMB to ensure the CIPSEA regulations are among those priorities. I think doing so is also part of our advisory committee’s statutory obligation.
More than two years after enactment of the OPEN Government Data Act, we have a great number of CDOs across government working to better organize and improve data governance and quality. But our new federal CDOs have yet to receive implementation guidance from OMB on many of the responsibilities under their new statutory charge. We understand it is being drafted – but this missing guidance substantially affects the completeness of this committee’s review for how to promote coordination of data for evidence building.
Last month, as the members of this committee know, OMB did step down in its role as the chair of this committee. While many may view that as a negative – I see that it presents us with a real opportunity to truly advise OMB at this critical juncture, as details of the FY 2022 budget are being finalized, as the American Jobs Plan is being negotiated, and as the American Family Plan is being developed.
Our charge as members of the ACDEB is to provide the White House’s OMB advice. I hope we will not shirk this responsibility. Whether you’re a civil servant appointee or from outside government – let’s move forward on making our advice available, known, and public.
As an advisory body, we have an opportunity to make meaningful statements now based on our six months of review. I encourage that we do so even if it's ahead of our statutorily-required annual report for beginning to make headway on the low-hanging fruit and the bigger questions in front of us.
The first thing we can do, in my opinion, is publicly acknowledge as a committee what we have been asked about and what we have been discussing that ties together all three of the paragraphs of this committee’s duties and what Congress intended us to investigate –
The United States government needs the capabilities of a data service to address existing institutional gaps in data sharing, linkage, protection, and use. A data service is a critical piece of our national data infrastructure that is missing today. OMB should take immediate steps to encourage, promote, and resource the establishment of a data service using existing authorities and available resources to the extent possible, then work with Congress to provide any necessary additional statutory authority.
While I am assuming our implicit understanding of this point still stands across all members of the committee – I firmly think we need to say that this committee continues to view a data service as a needed tool for government, and use our collective voice so that OMB, Congress, agencies, and the stakeholder community know that we are indeed moving forward.
There are already intriguing pilots underway at National Science Foundation (NSF) – and America’s DataHub – but at the early stages of those efforts. The Department of Commerce funded pilots over the last several years at the cost of millions to taxpayers, that provide some insights for us. We know we can have a data service, so let’s not merely nudge it forward or apply incrementalism – the need is too great. Let’s work together to create the ability for it to finally come into reality.
With the committee’s agreement on this headline topic, we get to turn to how we do it, what it means, what it does, what are the other areas to address. Which brings me to the topic I was asked to speak about.
Design Considerations for a National Secure Data Service
Last year, former US Chief Statistician Nancy Potok and I teamed up ahead of this committee’s launch to develop a proposal for how to begin designing a data service. The Data Foundation paper was supported through funding provided by the Alfred P. Sloan Foundation, and sought to tackle one of the most difficult questions about a data service – where in government should it be placed. Our hope was that by outlining a coherent, rationale framework the more important questions could be undertaken to make it work well and meet the needs expressed for government data operations.
Our experience suggests we can spend an infinite amount of time debating which agency should operate a data service. In fact, many already have small versions – but not government-wide integrated capabilities. We outlined a framework for considering where a data service should be within government to be successful. We offered a lengthy list of design considerations – and landed on a particular approach that I would like to strongly suggest this committee build upon.
I am going to preview our recommended approach at the outset because there is more detail in the paper than I can cover today, and the full document was shared with this committee’s full membership more than six months ago. Dr. Potok and I recommend creating the data service as a Federally-Funded Research and Development Center attached to the NSF and as a statistical operation under CIPSEA. Doing so assumes that the rules are issued under Part D – that this committee is tasked with providing advice on.
We offer a roadmap for how NSF could do this, beginning the work today. The pilot projects that NSF has already launched suggest not only that NSF agrees but that it is more than well positioned to lead the charge. Of note, this was not the precise recommendation of the Evidence Commission and I will explain how and why it differs, and how it builds upon the commission’s work rather than deviates.
When the Evidence Commission issued its final report to the President and Congress, we did not have the Evidence Act as a federal law. The commission’s design was able to imagine the full possibility and assume all its recommendations were implemented in tandem; in practice, only pieces of the commission’s recommendations are current law. I hope this committee will consider the pieces that are not in law as it proceeds in its work as potential opportunities for further improvement.
The committee is aware of the extensive guidance issued by OMB in the last several years and aware of the Federal Data Strategy from previous sessions, so I will not cover those at length. That is all important context for where we find ourselves in this moment in time.
How should we think about the value proposition as a starting point? The Evidence Commission unfortunately did not give us a roadmap on this point as a way of prioritizing the construction or development of a service. Clearly delineating the value proposition is certainly key, and we should seek to align with emerging topics of interest including those I have already mentioned.
One example of a particular area that a data service would be particularly useful in the current moment is for analyzing disparities and inequities by race and ethnicity. Rather than adding sensitive data elements or fields to administrative data collections, a data service can serve as a resource to support the analytical needs for studying inequities without introducing new unacceptable risks of harm or burden on the American people. Using the data service as a tool to link confidential information and produce summary insights offers tremendous potential for addressing inequities in our society.
We might also consider that since this committee’s recommendations are directed officially to the White House OMB Acting Director – a political appointee – that we align with the initiatives and priorities of this administration in explaining the value.
Attributes of a Data Service
Many of the concepts in our data service framework are derived directly from the Evidence Commission’s work as well as a parallel effort from the National Academies that was in development at the same time. The eight concepts are hopefully not foreign and I also hope will not be controversial to this committee:
Transparency and trust
Legal authority to protect privacy and confidentiality
Independence
Legal authority to collect data from agencies
Scalable functionality
Sustainability
Oversight and accountability
Intergovernmental support
These concepts and attributes are also a starting point for how we think about the data service over time. The Part D CIPSEA authorities help tremendously with aspects of these attributes. The questions about scalability, sustainability add in dimensions of administrative and resource constraints. Oversight and accountability may introduce new constructs and potentially non-existent mechanisms for ensuring good practices are in place. And increasingly we are recognizing the vital role for intergovernmental support, particularly in the context of the major federal government investments to states over the past year.
When we stacked up the attributes and looked at the range of possible options within government about where to create a data service – it was not an infinite list of realistic possibilities. While there are other potential options, we focused on four that align with the work of the Evidence Commission and CNSTAT. As the Evidence Commission suggested we considered Commerce, we also explored a university model, and we took a close look at an option the Commission did not really consider in depth which was at the NSF.
Included in the full paper on design considerations is a matrix that aligns the attributes with the options, and descriptions for each. In developing our understanding, we consulted with experts on each of these agencies, the issues raised in the matrix, and those who knew the landscape – including multiple former Evidence Commission members. One option clearly rose above the rest in this procedure – creating the Federally-Funded Research and Development Center (FFRDC) at NSF. It is an approach that allows for a focused mission that avoids resources being diverted for decennial censuses or other statistical production activities.
For those who may not be familiar with FFRDCs, there are substantial benefits to the approach. FFRDCs can be created administratively and because NSF has an existing federal statistical agency and experience running FFRDCs, the pairing offers real potential. While the term FFRDC is one that may not be familiar, you all know them – Oak Ridge, Los Alamos, etc. – NSF currently operates 5 other FFRDCs.
FFRDCs are public-private partnerships. They create the capability to adopt new technologies, apply personnel practices for ongoing expertise, facilitate close connectivity to government operations, collaborate with non-governmental actors, including being responsive to intergovernmental needs and potential academic or industry users. Importantly, the FFRDC structure provides a clock for review once the contract is issued, meaning it’s a procedure to build – and compete – that supports innovation, protection, and service.
In sum, NSF has experience with FFRDCs, has the statistical chops, and a domestic focus to create a natural starting point for rapid development of the data service that is measured in months and years, not decades.
A Roadmap for an NSF FFRDC Data Service
Dr. Potok and I present a roadmap in the paper that we hope this committee can use as a jumping off point. The roadmap has several key features for consideration, including alignment with the CIPSEA authority already provided to part of NSF, or another CIPSEA designation for the data service. The roadmap stresses the need for a vision of full transparency, where information is provided about what projects are happening and the benefits of those projects.
In implementing any data service, we may need to consider new accountability mechanisms, but should minimally set expectations for GAO and the NSF IG will be key. This advisory committee may even have a role to play. Promoting coordination with other data stewards, including across the federal statistical system, likely requires the White House and interagency councils to offer their support.
Determining criteria for an award will be essential for ensuring a competitive approach is pursued for the FFRDC, focused on capabilities, deployment of technologies, coordination across actors, and staffing. My hope is that the committee members agree at least partially with the vision and approach outlined here, including the need to move forward quickly.
Assuming the committee generally or implicitly agrees, there are at least four key domains of questions for us to consider: value, organization, sustainability, and accountability. There may be others too. These are the topics that must be considered for administering and effectively implementing a data service.
Conclusion
My appeal to this committee is this: let’s avoid the temptation of getting bogged down in the where a data service sits and think about how to make it successful. We need to focus on how to maximize public trust and usefulness. As part of this effort, I hope we will take on the unfinished business of the Evidence Commission, the Evidence Act, and prompt OMB to fulfill our expectations of its important role, including its statutory obligations for issuing regulations and guidance under the Evidence Act, the OPEN Government Data Act, and CIPSEA.
As we work in coming months to advance a data service I also want to stress that we don’t need to aim for perfection, we need to aim for progress. Every option presented by Dr. Potok and myself in our paper was a vast improvement of the status quo and there may yet be other approaches that have benefits, synergies, and relevance.
Based on the current environment and landscape, I strongly encourage we give serious consideration to prompting NSF to take on an even bigger role. And we build from there.
Hopefully I haven’t left anyone in doubt about where I stand on our next steps. I look forward to working with OMB and this committee in moving this important work forward. Thank you for your attention and engagement on these issues.