The House GOP dials a wrong number on the chained CPI

Your Social Security inflation adjustment is under attack.
(Bradley C. Bower/AP)

We’ve documented how the chained CPI--that new and supposedly “more accurate” inflation index that deficit hawks want to impose on Social Security--is really just an unwarranted benefit cut for the elderly. (You can find some of our columns on the topic here and here.) That’s because it consistently delivers a lower inflation reading than the consumer price index traditionally used to calculate retirees’ annual cost-of-living adjustments, or COLAs.

This week House Republicans termed that a “myth.” The “fact,” according to the GOP majority on the House Ways and Means Committee, is that the chained CPI, had it been in use now, would have given seniors a 1.7% raise for 2014--higher than the 1.5% they’re getting under the traditional measurement.

Unfortunately, they’re not quite correct. The GOP figures work only by torturing the chained CPI beyond its breaking point. Moreover, the committee members’ assertion proves exactly how complicated it would be to use the chained CPI as an inflation index, which they advocate. Finally, their claim is irrelevant.

Let’s take this point by point.


First, the GOP compared the initial third-quarter chained CPI for 2012 to the third quarter figure for 2013. That indeed shows a 1.7% increase year to year.

But the chained CPI, which ostensibly takes into account how consumers faced with price increases in favored goods simply shift to cheaper goods, is typically revised twice, over a period of two years, before it’s final. The goal is to incorporate real data about consumer substitutions.

Comparing the revised figure for the third-quarter 2012 chained CPI to the initial figure for the third quarter of 2013--the latest figures for both years-- produces an inflation figure of 1.46%--lower than the traditional CPI’s figure of 1.5%. (Currently, Social Security bases the COLA on the consumer price index for urban wage earners and clerical workers, or the CPI-W.)

That’s an issue often pointed out by critics of the chained CPI: Which version do you use? If the Social Security inflation raise is to be based on the initial chained CPI, then inaccuracies will be consistently baked into the inflation raise. So much for the “accuracy” of the chained CPI. But if the final revised figures are to be used, the cost-of-living adjustment will always be as much as two years out of date. How is that “accurate”?

In any case, what does one year’s calculation prove? Nothing.

As Ben Veghte, research director of Social Security Works, pointed out this week in a widely circulated email, the reason the chained CPI and traditional CPI figures are even close this year is because overall inflation is very low. In a low-inflation environment, he writes, “In most cases consumers don’t even notice changes in relative prices, and hence don’t substitute. When inflation returns to normal levels, these...indices will differ considerably.”

The fact, not the “myth,” is that the chained CPI consistently measures inflation at a lower rate than the CPI-W. The House GOP knows this. It’s exactly the reason that they favor using the chained CPI--they say it will save money on the inflation adjustment. (They don’t care that the change would cut benefits for seniors, and that the cut accumulates over time.)

So if the whole point of the chained CPI is to produce a lower inflation reading, why are the House Republicans crowing that it would produce a higher reading this year? Could it be because they know their figures are bogus, through and through?