"Balancing Data Privacy and Usability in the Federal Statistical System," accepted for PNAS publication

Economic researchers are long-standing users of federal government data in a variety of research settings. Chris Bollinger, Gatton Endowed Professor of Economics and Executive Director of the Kentucky Research Data Center, highlights an important problem in his research: Balancing Privacy and Usability.

Bollinger’s publication, “Balancing Data Privacy and Usability in the Federal Statistical System,” has been accepted for publication in the Proceedings of the National Academy of Sciences (PNAS). “The paper highlights an important problem: we want to protect the privacy of people whose data are collected, either through surveys or administrative data, but these data are crucially important for understanding our society and our economy,” says Bollinger.

The paper is co-authored with eight peers from Johns Hopkins University, Northwestern, Duke, UVA, University of Minnesota and The London School of Economics.

Professor Bollinger, tell us more about this research publication.

“Data collected by the federal government include both surveys and administrative data. Researchers across myriad fields such as economics, public policy, public health, agricultural economics, finance, statistics, and sociology rely on these data to inform decision makers in business and government. What we do in the paper is provide a framework, based on cost benefit analysis, that can help the government balance the tradeoff between data access and privacy. I’ve been concerned because the approaches taken in the federal statistical agencies seem to have largely ignored this. What I hope to do is engage not just researchers but also Congress and others to carefully look at how we make these decisions. I am deeply concerned that we may lose access to important data which will hamper our ability to help inform both business and government leaders about the economy.”

What does this mean for Kentucky?

“It’s easy to dismiss this as an academic issue, but it’s not. I, and many other researchers have long used federal data to address policy at the National, State and Local levels. As the former director of the Center for Business and Economic Research, I’ve worked with state and local government, with Kentucky businesses and industry groups, and with Chamber of commerce to help them understand both policy and the economy in which they operate. The data used almost always derives from federal statistical agencies such as Census. This can affect small or large businesses which rely on these publicly provided data to help them understand their market. It can impact not for profit organizations tasked with job training in understanding the evolving labor market. And it can impact our state government as it wrestles with which policies to implement. Finding the right balance an insuring that these important interests are recognized is crucially important.”

Why is this important to you?

“As both a citizen and survey respondent, I’m committed to securing privacy. As executive director of the KRDC and former director of the CBER, I know how important it is to have detailed data for research. Examining this tension has been interesting and rewarding. I’m also thrilled to finally co-author with my thesis advisor, Chuck Manski, and to work with what I can only describe as an all-star team of people. I’ve known each of them, to varying degrees, for a while and they all had my upmost respect and admiration as researchers. So, to be included with them has been wonderful.”


Authors include: 

V. Joseph Hotz, Duke University
Christopher R. Bollinger, University of Kentucky, Gatton College of Business and Economics

Tatiana Komarova, The London School of Economics and Political Science
Charles F. Manski, Weinberg College of Arts & Sciences, Northwestern University
Robert A. Moffitt, Johns Hopkins University
Denis Nekipelov, University of Virginia
Aaron Sojourner, University of Minnesota
Bruce D. Spencer, Weinberg College of Arts & Sciences, Northwestern University