The Bureau of Labor Statistics, the office that publishes the inflation rate, told outside economists this week that a hiring freeze at the agency was forcing the survey to cut back on the number of businesses where it checks prices. In last month’s inflation report, which examined prices in April, government statisticians had to use a less precise method for guessing price changes more extensively than they did in the past…
To calculate the inflation rate, hundreds of government workers called enumerators fan out across cities each month to check how much businesses are charging for products like blue jeans and services like accounting, largely by visiting brick-and-mortar stores. Statisticians in Washington, D.C. roll those figures together into the consumer-price index, a data stream that shows how the cost of living is changing for typical Americans.
If the government’s enumerators can’t track down a specific price in a given city, they try to make an educated guess based on a close substitute: say, cargo pants instead of slacks. But in April, with fewer workers on hand to check prices, statisticians had to base their guesses on less comparable products or other regions of the country—a process called “different-cell imputation”—much more often than usual, according to the BLS…
The inflation rate determines how much social-security benefits go up each year, and where federal tax brackets are set. Private-sector contracts such as wage agreements between companies and unions routinely reference the inflation rate. Payments on $2 trillion of inflation-protected federal bonds hinge on the inflation rate, as do yields on standard Treasury bonds. Businesses, investors and policymakers rely on the reading to guide their decisions. The Federal Reserve is laser focused on inflation data when it sets interest rates for the country.
Surveys require a lot of work to put together. Questions and methods need to be thought through and tested. Data needs to be collected. Analysis requires skill. Sharing results and interpretations is important.
The particular issues outlined above seem to have to do with (1) collecting data, which relies on going out and finding prices, and (2) dealing with missing data, which is related to #1 but is an issue for many surveys. If the survey is utilized by a large number of people, the choices made in the survey process can then affect decisions and policies.
It will be interesting to see what happens here. At what point do academics, policymakers, and others decide that the survey data may not be trustworthy? Which government surveys – and there are many – get priority for funding and having enough employees?
Legally Sociable 