Commentary: The Debate Over the Chained CPI
The news that President Obama's new budget will propose adopting the "chained" Consumer Price Index (CPI) for cost-of-living adjustments in Social Security and other retirement programs, and annual inflation adjustments in the tax code, has intensified the debate on this issue. Some commentators portray this proposal as a test of fiscal rectitude, arguing that the chained CPI more accurately measures inflation -- period -- and that if you're opposed to it, you aren't really serious about addressing deficits. Others, including many progressives, strongly reject the proposal, believing it would impose serious hardship on seniors with modest incomes.
I'm not comfortable with either position.
There are legitimate reasons not to adopt the chained CPI, and many people who aren't affluent would indeed be worse off. At the same time, fears that the chained CPI would impose severe hardship are overblown, especially if policymakers accompany it with a robust package of protections and mitigating measures for those who are very old or have low incomes.
Of particular importance in my view is whether policymakers adopt the chained CPI by itself or as part of a larger, balanced deficit-reduction package that stabilizes the debt as a share of the economy over the coming decade. If the former occurs, I think the case against the chained CPI outweighs the case for it. The chained CPI should be considered only as part of a larger overall deficit-reduction package that is well balanced.
The Case Against the Chained CPI
Most analysts who have studied the issue have concluded that the chained CPI -- which has risen about one-quarter of a percentage point more slowly per year than the regular CPI over the last ten years -- more accurately measures overall inflation than the regular CPI. But that judgment applies to the population as a whole. The chained CPI probably does not more accurately measure inflation for the elderly; in fact, it may well be less accurate.
That's because both indexes measure changes in prices for a "market basket" of goods and services that reflects average consumption by the U.S. urban population as a whole. The elderly population's market basket isn't the same, however, as the overall population's -- chiefly because a larger share of elderly people's spending goes for out-of-pocket health care costs. Since health care costs tend to rise faster than other costs, the chained CPI likely understates inflation as the elderly experience it.
That prompts another key argument against applying the chained CPI to Social Security. As elderly people grow older, their out-of-pocket health care costs tend to rise and they exhaust an increasing share of their assets. Some Social Security analysts believe that for such reasons, Social Security benefits should rise in purchasing power as people move from old age into very old age. But, if the chained CPI slightly understates inflation for the elderly, then people's benefits will erode in purchasing power as they age, because the effects of switching to the chained CPI compound over time.
On a related front, some argue that the chained CPI poses a further problem for many low-income people. The chained CPI rises more slowly than the regular CPI in part because it takes more account of the reality that when the cost of one item rises faster than the cost of another item, consumers will buy less of the former and more of the latter. But low-income people cannot necessarily make such substitutions as easily as other Americans can, especially if they live in inner-city or sparsely populated, rural areas with fewer large stores that carry a wide variety of goods, or if they lack a car or Internet access.
The Case For the Chained CPI -- Under Certain Conditions
The case for the chained CPI rests on several facts: we face unsustainable long-term deficits and need to make changes in both spending and taxes; moving to the chained CPI is that rare animal that simultaneously generates savings on both the spending and the tax sides of the budget ledger; the savings it produces will mount in future decades when we will need them most; and, in most analysts' view, it more accurately measures inflation for the overall population.
(Despite some belief to the contrary, the chained CPI would not have a highly regressive impact on the tax side. Instead, the tax would have a roughly proportional impact over the vast bulk of taxpayers along the income distribution. By 2020, according to the Tax Policy Center, it would reduce after-tax income by roughly 0.2 percent for households with incomes between $20,000 and $1 million. )
Simply saying, however, that because the chained CPI more accurately measures inflation, switching to it doesn't constitute a benefit cut -- as I once wrote nearly a decade ago, before I had studied this issue sufficiently -- goes too far. It oversimplifies the issues and overstates the case.
Assessing the Arguments
Both sides of the chained CPI debate have powerful arguments. We shouldn't lightly cause elderly people with modest incomes to become less well off as they age, through the loss of purchasing power. At the same time, we will need to restore long-term Social Security solvency sooner or later and, as the last major Social Security solvency law (in 1983) illustrates, such legislation almost certainly will include Social Security benefit reductions as well as tax increases. Moving to the chained CPI in Social Security is the sole such benefit reduction that is accompanied by savings in other programs and in the tax code, thereby somewhat lessening pressures on other valuable parts of the budget.
Of note, former Social Security Commissioner Bob Ball (whom I and virtually all Social Security advocates revered) included the chained CPI in several Social Security solvency plans that he developed a decade or so ago. (In the last years of Bob's life, he also developed solvency plans that did not include the chained CPI; those plans contained nobenefit reductions whatsoever and restored solvency by lifting the Social Security payroll tax cap, dedicating all estate tax revenue to the Social Security trust fund, and investing part of the trust fund in private equity markets.)
So, what should policymakers do? The chained CPI is worth considering -- but only if two crucial conditions are met.
First, policymakers must accompany the chained CPI with a strong set of measures both to substantially ease the impact on very old people (through an upward benefit adjustment after people have received Social Security for a specified number of years) and to shield low-income people to the greatest extent possible. These measures to protect very old and low-income people should be regarded as an integral, essential component of the chained CPI proposal. Without these measures, the chained CPI should be a non-starter.
To be clear, no set of protections can fully shield the very old or the poor. That leads me to the second condition: even with a package of protections, policymakers should not adopt the chained CPI unless it is part of a larger package that shrinks deficits enough to stabilize the debt over the coming decade and does so in a balanced and fair manner.
I'm deeply concerned about the effects of ten years of sequestration budget cuts on an array of vital programs and investments, as well as about the House-passed budget's stunning cuts in low-income programs, which would greatly increase poverty and destitution and likely produce levels of suffering that we haven't seen in decades. Instead, we need a balanced deficit-reduction approach that stabilizes the debt without skewering our most vulnerable fellow citizens or compromising the nation's future by shorting critical investments -- one that replaces sequestration and includes substantial revenue from scaling back wasteful, inefficient, or low-priority tax deductions, exclusions, and other "tax expenditures."
If including a chained CPI proposal with the necessary protections for very old and low-income people can enable such a balanced plan to become law, we should be open to it.
Progressives who strongly dislike the chained CPI proposal should consider whether there is any chance that congressional Republicans will agree to raise revenues by curbing tax expenditures without some significant entitlement changes. And if (as I believe) there is no real chance, what's preferable: the chained CPI with protections for the very old and the poor, or measures such as converting Medicare to premium support, raising the Medicare eligibility age to 67 and risking having some 65- and 66-year-olds go uninsured, and cutting Medicaid deeply and making ours more of a two-tier health care system based on income?
Having said this, I am concerned that Republican leaders will adopt the cynical approach of labeling the chained CPI an "Obama proposal" that they are willing to accept but only as part of a package that raises little or no revenue and, thus, does not force them to make any sizable compromises of their own.
The President has shown his willingness to make substantial concessions to reach an agreement. Will Republicans do the same?
 Tax Policy Center Table T13-0104 -- "Tax Parameters Indexed with Chained CPI, Current Law; Distribution of Federal Tax Change by Cash Income Level, 2020," March 8, 2013, http://www.taxpolicycenter.org/numbers/displayatab.cfm?Docid=3862&DocTypeID=1.
 Robert M. Ball, "Just a Little Maintenance: We Can Fix Social Security, If We Can Manage the Politics," Washington Post, July 18, 2004, http://www.washingtonpost.com/wp-dyn/articles/A56847-2004Jul17.html.