Why can’t you see what you’re doing to me, when you don’t believe a word I say? We can’t go on together, with suspicious minds. And we can’t build our dreams on suspicious minds. So sang the King of Rock and Roll, Elvis Presley in the Mark James song Suspicious Minds. While this song gave Elvis his 17th (and last) number one song, its story line, about a distrustful relationship, fits in well with what has been going on with the statistical and data collection agencies in Washington as of late.
Earlier this month, the New York Post reported that as far back as 2010, Census Bureau employees had been fabricating some of the data that went into the calculation of the nation’s employment statistics. It seems that surveyors were fabricating and adding in responses for people and companies that did not exist in order to reach their response quotas. While only one employee has been named, the Post suggests that the fabrication was widespread.
The reason behind the fabrication of data was likely benign. As someone who conducts surveys for a living, I know how difficult it sometimes is to reach sample size minimums. This is particularly true in a case like the survey that Census conducts for the Bureau of Labor Statistics, where a 90 percent response rate is required in each district, and where the timeframe for conducting the survey is short. This does not justify the addition of fabricated surveys into the data, but it does provide an explanation as to why a department staffer might do it.
It is also unlikely that even widespread cheating of this type on the part of census staff would lead to meaningful skewing of the statistics in question – namely employment, unemployment and the unemployment rates. In fact, there are already a large number of built in statistical anomalies in these data that most economists already have to deal with, and this would be just one more. That is a general problem with all data sources, not only those produced by the Census Bureau and the BLS. Good researchers understand that no data source is perfect, and that it is always best to use multiple data sources, to include sensitivity analysis and to document all of the adjustments that are made during an analytic process.
We reported in this blog that we believe that the vast majority of the people who toil away in the underfunded government statistical agencies are conscientious individuals that try to do a good job producing the information that the economy depends on. From the Economic Research Service in the Department of Agriculture, to the Energy Information Administration at the Department of Energy, to the Bureau of Labor Statistics, this alphabet soup of agencies creates tens of thousands of economic statistics and indexes each year. We rely on these agencies to do our job, as do people in companies, trade unions, other government agencies, and not-for-profit institutions all around the world.
The real problem with the story reported in The Post is not that the data were particularly bad because the surveys did not count 90 percent of respondents, but rather that there is now suspicion being leveled against Census and the other statistical agencies. The impartial nature of government statistics cannot be questioned, or else the entire statistical base on which decisions are made comes into question. Think about countries like Argentina, Venezuela and even China that are known to make up their numbers, and decide if these are good places to invest over the long term. How can one purchase Argentine debt if they don’t know what the true inflation rate in the country is? How can one invest in Venezuela if they can’t be sure if the country’s oil production figures are accurate, and honestly, who is willing to hold Chinese equities over the long-term if they really can’t trust the nation’s GDP statistics?
It is important that Congress and the White House examine this issue at the Bureau of the Census immediately, and in a non-political way. This is particularly true when most people in the country believe that the President lied to them about Obamacare, and when Congressional approval is in the single digits. Ignoring this problem is simply too dangerous, but allowing it to continue is even more so. If we are ever going to energize the economy and pull the country out of its funk, we cannot have Suspicious Minds about the basic information underpinning out decisions.