The Government’s Inflation Data Caught 'Percentile Dysfunction'

As if we didn’t already know something was wrong with the government’s inflation data, now up pops what looks like a confession. It’s part of a broader issue with government metrics that one economist dubs “percentile dysfunction.”

New research shows there’s a problem in how the prescription drug portion of the consumer price index gets calculated. Another, perhaps more realistic, index of drug prices far outgrew the CPI’s drug metric over a recent five-year sample period, with one part of the alternative metric increasing 59 percent more.

This anomaly should matter to anyone whose income is inflation-linked or who might get sick enough to need medication. As the U.S. ages, that should be a lot of people.

The pharmaceutical price news came out almost perfectly timed for President Trump’s State of the Union address on Tuesday, where he acknowledged the economic pain of rising medication prices. “I have directed my administration to make fixing the injustice of high drug prices one of our top priorities for the year,” he said.

Trump points out what many of us already know: drug costs are spiraling and emptying our wallets. But Trump’s message may be a shock to some technocrats, who swear by the CPI inflation metric.

A new report shows that the CPI data for medications lags far behind those of an alternative data set provided by the health research think tank IQVIA Institute for Human Data Science.

“We find that the IQVIA data indicate a more rapid rate of price increase than reported in the CPI for the aggregate of all prescription drugs,” states the January-dated report from the National Bureau of Economic Research titled “An Evaluation of the CPI Indexes for Prescription Drugs.”

Notably, one of the report’s five authors works for the Bureau of Labor Statistics, compiling the CPI data, which is why the paper has the feel of a back-door ‘fess-up. The man in question wouldn’t comment.

How much discrepancy is there between the two calculation methods used in the prescription inflation data? In the five-year period through October 2013, the difference “cumulates” to an average of 10 percent, the report says. That means that the IQVIA price data had jumped by 10 percent more than the prescription drug CPI did.

But the difference is even larger for certain classes of drugs. Notably, those in the categories of Cardiovascular-renal and Skin/Mucous Membranes saw a discrepancy between the two indexes of 59 percent and 37 percent respectively. In other words, the IQVIA price data had risen by 59 percent and 37 percent more than the CPI data over the same half-decade.

There are differences in the way they crunch and gather the data. The CPI data is calculated using a series of probabilities “that a specific retail outlet will be included in the sample” whereas the IQVIA data is constructed using “a basic national sample” of actual prescription drug transactions, the report says. Or put another way, the former is something of an educated guess, whereas the latter takes a sample of what drugs people are buying.

These findings on medications come on top of a slew of other problems that already make the CPI metric an increasingly suspect number.

For instance, take the cost of a new car. Despite the fact that new car prices are almost always higher than last year’s, the government adjusts those prices down. Why? This is known as the “hedonic” adjustment, and it tries to take into account that the most recent models will tend to have more features than last year’s new models.

But guess who mandates many of the new features? It’s the government, usually under the guise of improved safety. You have one part of the federal bureaucracy dictating evermore gizmos, while the other tells you that the increased price isn’t a price increase.

“Telling people there is little or no inflation is like telling people in West Virginia that we are at record-low unemployment,” says Dr. Philippa Malmgren, founder of London-based DRPM Group and a former economic adviser to president George W Bush. “We see the end result of the inflation, and still people say this is all in your imagination.”

In addition to the problems mentioned in the report above, and the so-called hedonic adjustments, she points out that inflation metrics are often used like a one-size-fits-all caliper.

CPI is a data point, which represents an average, she explains. “But inflation behaves like a wave.”

That means it impacts some people sooner than others and harder than others while there are those who barely notice.

“Economists ignore the human impact of the percentage point and yet are incredibly surprised by populism,” she said. “They’ll drill down on prices but ignore pain. I’d call it ‘percentile dysfunction.'”