The government collects data on consumer's behavior

When Margaret Humphrey goes shopping, she isn't just shopping for herself; she is shopping for the entire country.

Humphrey is one of the economic assistants for the Chicago regional office of the Bureau of Labor Statistics, the employees that collect the price data that is used to generate the Consumer Price Index, or CPI.

The CPI tracks relative price changes from month to month on a basket of goods purchased by consumers. It is widely considered the standard measure of inflation. The market basket is based on data collected by the Census Bureau and on the economists' conversations with store owners and managers.

Armed with a tablet computer, they travel to outlets ranging from mom and pop stores to auto lots collecting the prices on items that will eventually go into the market basket transformed in Washington, D.C. into the CPI.

The index has been a particular focus lately as inflation as measured by the CPI reached a 17-year high of 5.6 percent at an annual rate in July. The Federal Reserve has also been monitoring inflation with greater scrutiny according to minutes of their August Federal Open Market Committee meeting released on August 26.

Despite all of the attention, there is a great deal of confusion about how the index is calculated.

"You can look at the footnotes and see they are constantly tinkering with it. It is so complicated, they can't easily explain how it is put together," said Chris Martenson, a financial writer and commentator.

Martenson isn't wrong; the process is complicated and data intensive. During July, the 72 economic assistants in the region that includes Chicago went to 5,474 stores, and priced 21,440 items.

To accomplish this, the employees that collect the data undergo extensive training and are under constant watch.

"Our training is awesome," said Greg Howard, a branch chief and 11-year veteran of the bureau. "All of these things are put in place to succeed at getting the most accurate data we can because everybody knows that it makes such a big difference."

The training includes a two-week course in Washington, D.C., between 30 and 40 hours of homework and on-the-job training.

After passing a comprehensive exam, a new employee returns to the branch office, where they labor under constant observation. Areas where the new employee scored lower on the exam are monitored more closely than others. There is a certification process through the branch chief and all of the new employee's data is reviewed.

Even after certification, every employee undergoes constant review to make sure their data is correct.

Exactitude doesn't even begin to describe the way the armies of BLS economic assistants go about collecting consumer prices. Every item on their list is painstakingly detailed.

For example, Howard described how they would go about finding the price of a simple notebook with a specific number of pages, type of binding and even whether the pages can be removed from the book.

"All the way down to wide-ruled or narrow-ruled," Humphrey added, saying that the descriptions are "very specific, so that somebody else could go in there and read the specifications and pick out that exact item."

The descriptions all appear on the tablet, with the most important details, or cost factors, listed first. This is important when an item is out of stock and a replacement item needs to be priced.

Should a replacement be made, the economic assistants will note every detail of the item they have priced. All details of the item are sent to Washington along with the price.

Even on in-stock items, bonuses such as an extra 25 percent in a cereal box are noted along with the price data.

The process of collecting the data has become much easier since the introduction of the tablet about eight years ago. Before that, it was done with pen and paper, requiring reams of pages to be brought along in case the store had run out of an item or there was a substitution, Howard said.

The BLS is also very conscious of changes in consumer behavior.

"We might have XYZ grocery store and we might have 20 items in that grocery store," Howard said. "But every six months, there is a possibility that we might go into that grocery store and might initiate maybe two or three new items in certain categories.

"It is always giving us a fresh look at what the consumer is purchasing."

The changing tastes of consumers also determine what items are chosen for the CPI market basket. For example, while ten years ago standard cathode-ray tube television sets were the most common purchase, today plasma or LCD screens might be the majority of the sales.

When the economic assistants go to an electronic store to price a television, they will talk to the store owner to get an estimate of the percentage of sales that comes from the different types of televisions. Those percentages are fed into the tablet and the computer chooses a type to price based on the probabilities entered.

The computer could still pick the CRT television, but the percentage chance of that happening is much lower than in the past.

Even with the complications of exact descriptions and item substitutions, the CPI always seeks to measure the basket of goods that a typical consumer will purchase.

"The way we do things, I think it is fine really," Howard said. "I have been collecting data for a long, long time and a lot of thought goes into it."

While not a true cost-of-living measurement, the employees of the Bureau of Labor Statistics find that the data they collect does represent a typical consumer's behavior month-to-month.

"It is amazing how real it is," Humphrey said. "Not everything, because you don't always purchase everything that you are pricing, but it is amazing how you can relate it to real life.