MNI INTERVIEW: US Data Already Suffering From Underinvestment
MNI speaks with authors of a new report warning U.S. federal statistics are at risk.
The integrity of the data used by Americans and policymakers including the Federal Reserve is in increasing jeopardy and in some instances is already suffering from the impacts of underinvestment, according to the authors of a new report from the American Statistical Association and George Mason University.
Government statistics are reliable right now but the 13 principal U.S. data agencies face increasing challenges to innovate to the extent, depth and breadth is needed, Nancy Potok, Connie Citro and Steve Pierson said in a joint interview, citing factors including shrinking budgets, falling survey response rates, lagging staffing, and the potential for political interference.
"This is an urgent situation. This is not something that can keep being pushed off. The data will deteriorate over time. This needs immediate attention from policymakers," said Nancy Potok, former chief statistician of the United States from 2017 to 2019.
BLENDING DATA
Potok, Citro and Pierson spoke after releasing the new report that includes 15 immediate recommendations ranging from filling gaps in existing legislation, bolstering agencies' autonomy, expanding data sharing authority between agencies that frequently hamper research projects, to providing adequate multi-year budget levels. Funding for most of the principal statistical agencies has declined 14% in real purchasing power over the past 15 years, with the BLS down 18%.
They say, however, that innovation will be key to addressing the declining survey response rates, which is being exacerbated by underfunding and further threatened by Capitol Hill proposals that would constrain agencies like the Census Bureau to a limited number of contacts.
"The ability of the agencies to do the innovation needed for this century is very shaky at the moment and our recommendations are designed to try to turn that around," said Connie Citro, former director of the Committee on National Statistics of the National Academies of Sciences, Engineering, and Medicine.
RESPONSE RATES
BLS Commissioner Erika McEntarfer last month suggested budgetary pressures might require cutting the sample size of the Current Population Survey (CPS) by 5,000 households, a survey that is used to put together the monthly employment report.
"This reduces accuracy," said Citro about the CPS. The CPS response rate has fallen from just above 90% in the late 2000s to just under 70% over the last year. "The trends are all in the wrong direction on that, which is why blended data is needed," she added. The response rate on the Consumer Expenditure Survey which produces market basket weights for the CPI has declined from over 70% in the late 200s to about 40% most recently.
Potok, the former chief statistician, said: "If the response rates keep dropping then the pressure is on the federal statistical agencies to find innovative methods of upping response rates with costs associated or using other methods of filling in the gaps with other sources."
Fed Governor Adriana Kugler recently said she's hopeful non-traditional survey contact measures and shorter, more direct surveys can help improve response rates, "but it may be that more integration of private-sector data with official data provides an avenue for addressing this problem as well."
BUDGET CONSTRAINTS
The U.S. Bureau of Economic Analysis announced earlier in the year that due to budget constraints it was cancelling its tracker of credit card spending and its detailed statistics for the financial business sector.
"Unless this gets resolved, there will be more cuts at places like BEA," said Pierson, director of science policy at American Statistical Association. The U.S. statistical agencies are also handicapped by data sharing agreements that often limit access to data.