As fiscal-cliff negotiations continue, reports indicate that the two sides are still at an impasse on Sunday afternoon because the GOP insists that "chained CPI" be included in any deal. What is chained CPI? Matt Yglesias explained back in December how using this method to calculate Social Security will result in lower benefits. The article is reprinted below.
On Wednesday I wrote about one of the most widely discussed and high-profile ideas for reducing federal spending on the elderly—raising the eligibility age for Medicare. It’s a simple idea to understand and a pretty terrible one. The other idea that comes up constantly in negotiations and rumors of negotiations and hypothetical grand bargains, by contrast, languishes in obscurity. It involves indexing Social Security benefits to the Chained Consumer Price Index for All Urban Consumers (C-CPI-U) rather than the Consumer Price Index for Urban Wage Earners and Clerical Workers (CPI-W) that’s currently in use.* This change shows up in all kinds of think-tank plans and unofficial sketches, and it has at various times secured the endorsement of everyone from Dick Durbin to Eric Cantor.
Conservatives love this idea because it cuts Social Security benefits. Deficit hawks love it because it might also raise taxes. And negotiators love it because nobody understands what it means. But don’t let people fool you. Even though the details are technical, this is a political question, not a technical one—a benefit cut, perhaps paired with a tax increase, and not an effort to make inflation calculations “more accurate.”
So what’s the difference? It all comes down to substitutions.
The way the index works is that the Bureau of Labor Statistics sends its minions out through the country to find out what things cost. They write this down, and the bureau notes the change over time. Then it weights the change in the price of different things according to how large a share they are in the typical consumer’s overall basket of purchases. An increase in the price of cars is a bigger deal than an increase in the price of violins because the average American spends much more on cars than on violins. The bureau also adds in some fancy math and a bit of hand-waving to try to account for changes in the quality of goods and the arrival of whole new products. The idea is to track the prices of a constant basket of goods over time.
“Chaining” the index means taking a slightly broader view of how the baskets should work in order to account for switching behavior. One reason people buy pork, for example, is that it’s cheaper than beef. But so is chicken. So if pork prices rise, price-sensitive shoppers will probably shift and buy less pork and more chicken. In other words, the price of pork went up, but the overall impact on meat prices is smaller than a naïve look at the movement in pork prices would suggest.
If you want to make use of the chained index sound like a no-brainer, you can do what Bloomberg View has done and use a trivial example like “a shopper might respond to an increase in the cost of Granny Smith apples, for example, by switching to lower-cost Red Delicious.” On average, this kind of behavior-linked chaining makes the chained index about 0.25 to 0.35 percentage points lower than the unchained index. So great news! Unless, that is, you really like Granny Smith apples. If the price of bourbon tripled overnight for some reason, I would, in fact, probably shift to scotch. But I’d be pretty pissed. In the grand scheme of things, whiskeys are fairly substitutable, but they make the different varieties for a reason—people have preferences about this stuff. By shifting chaining assumptions far enough, you could make inflation completely vanish. If people can’t afford medical care, they’ll just buy over-the-counter homeopathic remedies instead! The chained index doesn’t go nearly that far, but the point is that there’s no unique right or wrong answer (the British have their own version of the indexing controversy) for how to treat product shifting, and its impact on individuals’ welfare will vary enormously.
This matters a great deal for Social Security, however, because benefit levels are adjusted upward each year in line with inflation. If Congress decides that chained index is the “right” measure of inflation, benefit levels will be lower than currently predicted and the deficit will go down.
If Congress wants to be more intellectually honest about it, they could also note that tax brackets are pegged to inflation as well. If you switch tax brackets to chained index, then taxes will rise gradually as benefits fall. As far as grand bargains go, that’s fair enough. But in terms of accuracy, it should be noted that the BLS also calculates the Experimental Price Index for the Elderly (CPI-E) for reference purposes weighted to the basket of goods consumed by the elderly and finds that these prices generally rise faster than the regular index. Grandma buys a lot of health care services and isn’t so interested in the falling price of an iPad 2. So if you’re looking to trim benefits in keeping with the spirit of the program, it’s important to note that many Social Security beneficiaries already see quite meager monthly checks. That’s why thoughtful reform plans like the 2005 one from Peter Orszag and Peter Diamond at Brookings or Christian Weller’s 2010 plan for the Center for American Progress specifically include enhanced benefits for the poorest seniors and the very elderly even while switching to the chained index.
The central point is that it’s politics all the way down. Altering the inflation calculation is often discussed as if it’s a magical deficit-reducing alternative to spending cuts and tax hikes. But it reduces the deficit precisely because it’s a spending cut and potentially a tax hike as well. That doesn’t make it a bad idea. Spending cuts and tax hikes are going to be needed sooner or later. But it simply reposes all the same questions about any deficit-trimming program: Can it be done fairly and without unduly punishing the most vulnerable? If done right, it can. It’s certainly possible to make it work, but glib talk about apples and “better” inflation calculations masks a policy that if done sloppily could be quite damaging to the low-income elderly.
Correction, Dec. 10, 2012: This article originally said that Social Security benefits are indexed to the Consumer Price Index for All Urban Consumers (CPI-U) rather than the Consumer Price Index for Urban Wage Earners and Clerical Workers (CPI-W).(Return to the corrected sentence.)