April 3, 2013
When President Obama releases his budget next week, it may include what some tout as an elegant solution to the nation's fiscal problems, a way of measuring price changes known as “chained CPI.”
The thinking goes like this: The debt needs to be reined in. Changes to benefit payments don’t accurately reflect the way Americans spend money. If you fix the benefit adjustments to make it so they do, you would raise money and put the government on a path to addressing the debt.
The proposal enjoys broad support. But the very premise of the change—that chained CPI will make adjustments to benefits and taxes more accurate—may not hold true for everyone, especially the elderly.
The government currently relies on a one-size-fits-nobody measure of inflation. That indicator, known as the Consumer Price Index, is crucial for determining how tax brackets and government benefit payments are updated over time. When prices rise, benefit payments rise in step. But the current inflation measures (components of CPI known as CPI-U and CPI-W) fail to account for how people actually spend money. When the price of certain items rise, consumers turn to cheaper alternatives. If pork gets more expensive, for example, people might start buying beef. If new car prices rise, Americans might turn to used cars. The current measures tracks the price of all items, but it doesn't fully account for the cost of substitution.
Some argue that the government should instead use its chained CPI indicator, an alternative inflation measure that better accounts for that substitution. Because it more accurately reflects how consumers swap out goods to mitigate price hikes, chained CPI rises more slowly than the current inflation measures to determine benefit payments. In fact, chained CPI is projected to grow about 0.25 percentage points more slowly per year over the next decade, according to the nonpartisan Congressional Budget Office.
It may seem small, but the gap widens over time. The Moment of Truth project, a nonpartisan group that advocates taming the ballooning federal debt, illustrates the difference in the chart below. From 2000 to 2011, prices either rose by 29 percent (chained CPI) or 34 percent (current inflation measures). If the government switched to chained CPI next year, it would raise $390 billion in taxes and program savings through 2023, the group says.
Almost exactly one-third of the money raised—$127 billion—would come from reduced Social Security payments, according to the Moment of Truth calculations. Estimates vary, but changes to Social Security Cost of Living Adjustments account for a large share of savings under each. A switch to chained CPI would barely be noticeable for Social Security recipients at first: Benefit payments would be about 0.3 percent less at the start of retirement, the Center for Retirement Research at Boston College wrote in a 2011 paper. But it would compound. By the time the average retiree reaches 85, payments would be about 6.5 percent less per month than they would be otherwise.
Proponents acknowledge that the elderly and other federal benefit recipients will take a hit, but say that that’s a subject for a separate debate. “Using an incorrect inflation index is neither a sensible nor well-targeted way to help these populations,” the authors of the Moment of Truth paper write. “It is not a reason to forgo the deficit reduction that moving to the chained CPI would bring,” the Center for Budget and Policy Priorities wrote in a paper last year. It and others argue for a broad benefits increase and other changes to mitigate the impact of switching to chained CPI.
If it’s accuracy you want, chained CPI won’t get you there, opponents argue.
Seniors spend money differently from the average person. Health care accounted for 13 percent of spending for those 65 or older at the time the Center for Retirement Research wrote its paper. It accounted for 5 percent of spending for the general population. Not only do the elderly spend more on health care, but those costs also grow faster than others.
To better understand the difference in seniors' spending, the government has long tested an elderlyfocused inflation measure (CPI-E). From 1982 to 2010, that experimental measure outpaced standard inflation by 0.27 percentage points annually, according to the nonpartisan Congressional Research Service.
In other words, using the experimental elderly inflation measure (CPI-E), benefit payments would rise faster than under the current inflation measures. Using chained CPI (or C-CPI-U), benefits would rise more slowly, as AARP shows in the graph below. AARP, the huge seniors lobby, has been one of the biggest opponents of chained CPI. The group plans to release in a week or two the results of a survey showing how backing the proposal would affect voter favorability for individual lawmakers.
That’s not to say that CPI-E is perfect, either. Costs may be rising, but so too might the quality of health care services. And it may not account for that. The point, opponents of switching to Chained CPI argue, is not that the CPI-E is the best measure to calculate Cost-of-Living Adjustments (COLAs), but that so far there is no better alternative.
“To ensure that the system is paying proper COLAs, Congress should instruct the Bureau of Labor Statistics to develop a statistically rigorous index of inflation among retirees,” The New York Times editorial board recently wrote. “Until that is done, cutting the COLA on grounds that it is too large would be unjustified and disingenuous.”
More research also needs to be conducted on how—and whether—the low-income elderly substitute goods, according to the Center for Retirement Research. For low-income seniors, most of their spending goes to essential items such as food, housing, transportation and health care. Because they are already spending so little, they may not have as much room to maneuver around price hikes.
“With little ability to respond to price changes, the poor have no mechanism to offset the full brunt of a price increase,” the center wrote in its report.
And Social Security’s top actuary made a similar point in a 2011 letter to Rep. Xavier Becerra: “[T]he degree to which individuals, in different circumstances and with different income levels, are able to change their purchases on a discretionary basis across strata likely varies."
In other words, chained CPI may account for substitutions, but not everybody substitutes in the same way. Like the current inflation measures, chained CPI is a one-size-fits-all or -nobody solution, depending on your point of view.
April 3, 2013