Data Is Valuable Only If Managers and Policymakers Will Use It
There’s plenty of evidence about how well programs work, people just tend to ignore it.
We’ve known for years that the federal flood insurance program encourages people to live in flood-prone areas—instead of moving to safer ground—all at taxpayer expense. Yet nothing significant has been changed with the program. Similarly, the most common program designed to reduce recidivism by domestic violence offenders has been shown to be ineffective. Yet, not only does the program persist, but dozens of states mandate its use, hindering experimentation with other more promising programs. Additionally, in 2015, undercover inspectors were able to carry 67 simulated bombs and weapons past TSA checkpoints. Two years later, inspectors did it again. The evidence and data were clear in these and many other cases, but government still did not shut down ineffective programs, switch to better models, or improve.
Efforts to make government more data-driven and evidence-based are gaining steam, whether it’s the effort to increase availability of key government data, data sharing by the bipartisan Commission on Evidence-Based Policymaking, governmental use of private sector data, randomized control trials, or analytical capacity within governments.
These are all worth-while efforts. But they are only valuable insofar as they lead to change, and the missing link in current efforts is a systematic focus on government’s ability to follow what the data and evidence suggest. A significant portion of evidence and data does end up resulting in change, but that fraction is far too small. The core problem is government is not built for change, and that includes changes based on evidence and data. To be sure, change is not easy for any organization. Employees may fear change, their incentives may be misaligned, or they may dismiss new evidence out of motivated reasoning.
But governments have additional obstacles to change, and it’s those obstacles that need to be directly and honestly explored and, to the extent possible, addressed.
The first obstacle is political. Concentrated, entrenched interests in preserving the status quo can often overcome the diffuse, future benefits of policy changes. The public might be irrationally invested in the governmental program or effort, such that governments may not want the backlash from following the evidence. Changes that require legislation must also compete against many other demands to get on the time-limited legislative agenda. The above are not amenable to easy solutions, although it is worth trying. For example, the Base Realignment and Closure process was a successful effort to overcome concentrated interests in converting data (the United States had more military installations than it needed) into action (closing unnecessary facilities based on security and military requirements). Congress passed a law establishing an independent commission to investigate and recommend bases for closure. The law required that members of Congress vote for or against the recommendations as a full package, thus changing the calculus for many lawmakers and resulting in far more action than would have been likely with piecemeal votes.
The second obstacle is operational. Well-meaning laws, policies and practices interfere with agencies’ ability to administratively follow evidence. Employee protections prevent or at least slow the ability to reassign or terminate employees based on evidence-based need. Congress often provides budgets that make it hard to move money from one program to another based on evidence. Tweaking a regulation to address an unintended consequence or new piece of evidence generally must follow the same long process as creating a whole new regulation. But we have seen efforts to address these operational obstacles, from Congress providing pilot authorities or more flexibility in allocating money to limited exemptions from certain governmentwide rules.
At root, governments are not set up to easily integrate evidence and data, let alone for the ongoing trial-and-error experimentation and continuous improvement seen in high-performing companies. Making more data available and increasing the amount of evidence on government performance is great, but much of the value of these efforts will be for naught if inertia wins out. In order to ensure these efforts bear fruit, we must revisit some of the core operations of how government works. This will not be easy. Some ways of making it easier for government to adapt to evidence might also create flexibility for ill-intentioned officials to use for ulterior purposes. However, we must thoughtfully reset the balance. With every new piece of data and evidence, the gap between what our government is and what it should be will grow. And government’s functions are simply too important for that to continue.
Adam Neufeld is a senior fellow at Georgetown Law’s Institute for Technology Policy and Law and the Beeck Center for Social Impact & Innovation. He previously served as deputy administrator of the General Services Administration.