President Obama's budget proposes to use the chained CPI as a more accurate way of setting the annual inflation increases for Social Security benefits. It is not a cure-all for the program's long-term funding challenge but would, by itself, eliminate about a fifth of that gap. It does so, however, by cutting benefits to most recipients.
Not surprisingly, the idea is roundly denounced by seniors' groups. It is also just as solidly supported by an impressive group of economists and program experts.
The president's budget is not likely to be enacted in anything like its current form, and this goes for the chained CPI proposal. But making sense of the starkly opposing views of the "pro" and "con" camps provides an illuminating look at how hard it is to tweak even a single feature without affecting many other pieces of the nation's bedrock retirement program.
First off, what is the chained CPI? It's a measure of changes in the cost of living that its developer, the U.S. Bureau of Labor Statistics (BLS), unequivocally says does a better job than the two most widely used current consumer price indexes: the CPI-W, which measures price changes affecting urban wage earners, and the CPI-U, which measures price changes for all urban residents. The CPI-W is the measure used to set annual changes in Social Security's cost of living adjustment (COLA).
BLS Commissioner Erica Groshen, one of five experts who testified late last week at House hearings about the chained CPI proposal, says the purpose of the government's CPIs is to measure not just price changes but the changes in what it costs consumers to maintain the same standard of living. It is incredibly challenging just to measure price changes in an economy as big as ours and one where prices can change with dizzying speed.
It's hard to measure the cost of items that are the same each month—an apple, say—across the 8,000-plus products and numerous retail (brick-and-mortar plus online) places where prices are recorded by BLS agents. But the agency must also try and figure out how our standard of living is affected by qualitative changes in products or even the introduction of new product categories.
Is a $250 smartphone, for example, an inflationary product compared to an older, $100 phone that just made telephone calls? Or is it an improvement in our standard of living? Or some of each and, if so, what's the ratio? The standard-of-living impact of health care changes is especially challenging, and health care price changes are felt to understate living standard gains.
In terms of measuring changes in the standard of living, the CPI-W and CPI-U have real limitations in measuring how consumers adjust their purchase patterns to substitute different products in response to price changes. The chained CPI better reflects those substitutions and thus does a better job of measuring what it costs to maintain our standard of living. It's also much less vulnerable to misstating the rate of price changes than the other indexes. The "chained" label exists because the measure compares purchase patterns from one month to the next and thus chains them together for purposes of looking at consumers' product-substitution decisions.
The chained CPI consistently records lower rates of inflation than the CPI-W. The difference depends on the range of years being measured but, looking ahead, it is projected to average about a quarter of a percentage point a year. That's a tiny number but its impact is cumulative, meaning the total impact in year two is more than half a percentage point, rising to more than a full percentage point after year four (don't forget compounding) and so on.
So, superior or not, using the chained CPI would cut Social Security payments from their projected levels using the CPI-W. It also would trim other federal benefits and would raise federal taxes. The percentage tax brackets are adjusted up every year by the rate of inflation and thus would adjust upward more slowly using the chained CPI, meaning higher taxes.
The Congressional Budget Office says using the chained index would produce a $340 billion swing in deficits over the next decade. That would consist of $127 billion in lower Social Security COLA payments, $38 billion in other federal COLAs, $51 billion in deficit reductions in other parts of the federal budget and $124 billion in higher federal taxes.
Critics say even the current index under-measures inflation for seniors because it far underweights health care expenses, and seniors spend a bigger percentage of their budgets on health care than do younger consumers. They want the feds to refine and use an experimental index that specifically tracks spending of elderly consumers and is called, appropriately, the CPI-E.
There is a lot of informed thought, however, that all the federal price indexes overstate the impact of health care price inflation, including the CPI-E. Perhaps a "best case" solution would be to develop a chained version of the CPI-E, but that's not on the table right now. But this would likely mean using one index for Social Security and at least one more index for other federal programs and for adjusting income-tax brackets. And the way a chained CPI is developed means it is much more likely to be revised in future periods than are the CPI-W and CPI-U measures. Applying such revisions is possible, supporters say, but hardly a snap.
Supporters of the chained CPI also freely acknowledge that using the measure would result in benefit cuts to Social Security recipients, with special hardships for lower-income beneficiaries. Most of them thus propose related benefit increases for the most vulnerable segments of society, and the Obama budget does the same.
Ed Lorenzen was a staff member on the National Commission on Fiscal Responsibility and Reform, better known as the Simpson-Bowles commission after its co-chairman, Alan Simpson and Erskine Bowles. The commission recommended using the chained CPI and Lorenzen noted it raised little controversy and was the change most widely supported by the 18 commission members. That was due largely to the commission's support for enhanced benefits for those most affected by use of a new COLA measure.
As is the case with most, if not all, federal benefits, the biggest dollar winners are the people who make the most money. And this is true of the chained CPI proposal as well. "Overall," Lorenzen testified last week, "there is no reason to maintain an average tax windfall of $450 for those in the top quintile as a result of using an inaccurate measure of inflation in the tax code just to prevent a $25 tax increase for those in the bottom quintile."
"If a technical change with broad support among experts like switching to the chained CPI to more accurately index provisions to inflation cannot win bipartisan support in Congress,' he concluded, "then prospects for putting our debt on a sustainable, downward path as a share of the economy are grim."
If nothing else, however, the complexity of Social Security supports one of the strongly held views of program experts: Changing Social Security provisions should not be done piecemeal but as part of a comprehensive review of the program, last accomplished 30 years ago under the leadership of former Federal Reserve Board Chairman Alan Greenspan.