11/06/2013 12:28 pm ET

House Ways and Means Committee Misleads Hill on Chained CPI

The House Ways and Means Committee added another chapter Monday to the epic misinformation campaign to sell the chained CPI to Members of Congress, the media and the general public as the basis for Social Security cost-of-living adjustments (COLAs). It sent a press release to all Hill offices and the media entitled, "FACT: Chained CPI Would Result in a Larger COLA Increase for 2014." They claimed a COLA based on the chained CPI would be 1.7 percent, a larger COLA than the statutorily scheduled COLA of 1.5 percent. Yet this is by no means a "fact" -- and is highly misleading. The chained CPI rises slower than the current COLA, which is the very reason Republicans support it. And while there are a variety of ways to calculate a chained CPI COLA, the method Ways and Means staff used is the least accurate on a year-to-year basis.

The chained CPI is impractical for use in setting Social Security COLAs because it is based on surveys of consumer spending that are not completed until up to two years after the fact: BLS first issues a preliminary estimate, then a year later revises this to an interim estimate, then two years later corrects the interim to a final value. For any given year, there are often significant swings from the initial to the final estimate (initial estimates are off by 0.31 percentage points on average).

The error-prone method the Ways and Means staff used to estimate a chained CPI COLA for 2014 was to compare the initial estimate of the index value for this year to the initial estimate for last year (even though an interim estimate was available). Using this method, they estimated a 2014 chained CPI COLA would be 1.7 percent. If they had instead used last year's interim estimate, they would have arrived at a less error-prone estimate of 1.5 percent, an amount equal to next year's statutorily scheduled COLA.

Analysts interested in accuracy, rather than in misleading Congress about the chained CPI, choose one of two other methods available for calculating a hypothetical chained CPI COLA. One method would provide distinct COLAs for each cohort of beneficiaries: for each cohort, a given year's benefit increase would be the product of their initial benefit multiplied by the cumulative increase in the chained CPI index since then. This would be very complicated to administer because each year, each cohort of beneficiaries would have a different COLA. And it would be very hard for Social Security beneficiaries to understand. Those receiving lower COLAs would likely perceive this to be unfair.

Another method would, in the first year of retirement, provide a COLA based on a comparison of the initial estimate for this year to the interim estimate for last year. Then in the following year, the COLA would equal the growth in prices over the year, with an adjustment for revisions of errors in estimating the previous year's growth, and so forth in succeeding years. New beneficiaries in any given year would not receive the corrective adjustment, because they would not have been subject to last year's errors; so again, there would be cohort-specific COLAs. This method would also be complicated to administer. Indeed, there is no method of implementing the chained CPI that is not either highly error-prone on a year-to-year basis, or expensive to administer and unlikely for the public to perceive as fair.

What is so misleading about the Ways and Means press release is that no serious analyst would claim that the change in the chained CPI initial estimate from one year to the next constitutes an accurate measure of annual inflation. Ways and Means misled the Hill by taking one pair of initial estimates out of context. Again, since the Bureau of Labor Statistics began tracking the chained CPI a decade ago, such estimates have been off by an average of 0.31 percentage points per year. Hence a COLA calculated as Ways and Means suggested would ultimately end up having to be revised to somewhere between 1.3 and 2.0 percent.

The larger point is this: All non-partisan experts -- from the Social Security Administration to the Congressional Budget Office to the Congressional Research Service -- agree that the chained CPI would cut Social Security benefits significantly over the medium and long term. Indeed, the very reason that the Republican Ways and Means staff support the chained CPI is that it would cut Social Security benefits. For the average worker retiring today, it would cut benefits cumulatively by $28,000 by age 95.

Finally, on a technical note: the reason the difference between the actual COLA and a chained CPI COLA is small for 2014 is that in today's low-inflation environment, there is scant price dispersion. And with scant price dispersion, there is little scope for consumers to substitute cheaper items for desired/needed goods and services. It is such substitution that lowers the chained CPI. Moreover, with inflation under 2 percent, in most cases consumers don't even notice changes in relative prices, and hence don't substitute. When inflation returns to normal levels, the chained CPI will again rise much slower than the current COLA.

Highlighting one year's low COLA differential was disingenuous -- another episode in a persistent misinformation campaign. Congress must not let itself be deceived into believing the chained CPI is either accurate or fair for seniors. What we need is a COLA that actually keeps pace with seniors' inflation, the Consumer Price Index for the Elderly, the CPI-E. Senators Harkin (D-Iowa) and Begich (D-Alaska) have introduced bills that would adopt the CPI-E for Social Security, and they deserve our support.

Workers who have contributed to Social Security through a lifetime of work deserve better than to see their earned benefits cut, especially when this cut is disingenuously sold with smoke and mirrors.