Editor's Note: This piece is cross-posted from the Center for Economic and Policy Research Blog.   According to reliable sources, the Obama administration is seriously contemplating a deal under which the annual cost of living adjustment for Social Security benefits would be indexed to the chained consumer price index rather than the CPI for wage and clerical workers, CPI-W, to which it is now indexed. This will lead to a reduction in benefits of approximately 0.3 percentage points annually. This loss would be cumulative through time so that after 10 years the cut would be roughly 3 percent, after 20 years 6 percent, and after 30 years 9 percent. If a typical senior collects benefits for twenty years, then the average reduction in benefits will be roughly 3 percent.  

There are a few quick points worth addressing:

1. The claim that the chained CPI provides a more accurate measure of the cost of living; 2. Whether Social Security benefits are now and will in the future be sufficient to allow for a decent standard of living for retirees; and 3. Whether this is a reasonable way to be dealing with concerns over the budget.

This are taken in turn below.

Is the Chained CPI More Accurate?

While many policy types and pundits have claimed that the chained CPI would provide a more accurate measure of the cost of living for seniors, they have no basis for this claim. The chained CPI is ostensibly more accurate for the population as whole because it picks up the effect of consumer substitution as people change from consuming goods that increase rapidly in price to goods with less rapid price increases.

While this is a reasonable way to construct a price index, it may not be reasonable to apply the consumption patterns and the substitution patterns among the population as a whole to the elderly. The Bureau of Labor Statistics (BLS) has constructed an experimental elderly index (CPI-E) which reflects the consumption patterns of people over age 62. This index has shown a rate of inflation that averages 0.2-0.3 percentage points higher than the CPI-W.

The main reason for the higher rate of inflation is that the elderly devote a larger share of their income to health care, which has generally risen more rapidly in price than other items. It is also likely that the elderly are less able to substitute between goods, both due to the nature of the items they consume and their limited mobility, so the substitutions assumed in the chained CPI might be especially inappropriate for the elderly population.

While the CPI-E is just an experimental index, if the concern is really accuracy, then the logical route to go would be for the BLS to construct a full elderly CPI. While this would involve some expense, we will be indexing more than $10 trillion in Social Security benefits over the next decade. It makes sense to try to get the indexation formula right.

Are Social Security Benefits Adequate?

While some people have tried to foster a myth of the elderly as a population living large, the facts don’t fit this story. The median income of people over age 65 is less than $20,000 a year. Nearly 70 percent of the elderly rely on Social Security benefits for more than half of their income and nearly 40 percent rely on Social Security for more than 90 percent of their income. These benefits average less than $15,000 a year.

The reason that seniors are so dependent on Social Security is that the other pillars of the retirement stool, employer pensions and individual savings, have largely collapsed. Defined benefit pensions are rapidly disappearing. Defined contribution plans, like 401(k)s have also proved grossly inadequate. Only around half of the work force even has a defined contribution plan available to them at their workplace. In a period of stagnant wages and limited employer contributions, workers have generally been unable to accumulate much wealth in these plans. According to the Retirement Research Center at Boston College, the median value of 401(K) and other defined contribution plans for those near retirement who have a plan is $120,000, enough to get an annuity paying $575 per month.  

For most workers the vast majority of their wealth was in their homes. The collapse of the housing bubble destroyed much of this equity. Counting all forms of wealth, including equity in a home, the median household approaching retirement had just $170,000 in wealth in 2011.

The proposed cut in the annual cost of living adjustment will be a substantial hit to a population that for the most part is ill-prepared to see a cut in its income. The effect of this cut on the income of the typical beneficiary will be larger, measured as a share of income, than the return to Clinton era tax rates on the richest 2 percent will be to the people affected. It is also worth noting that this cut to benefits will affect current retirees, not just people who will be collecting benefits 10 or 15 years in the future, who might have some opportunity to adjust to a cut.

Is the Chained CPI a Reasonable Way to Deal with the Budget?

It is important to remember that under the law Social Security is supposed to be treated as a separate program that is financed by its own stream of designated revenue. This means that it cannot contribute to the budget deficit under the law, because it is only allowed to spend money from the Social Security trust fund.

This is not just a rhetorical point. There is no commitment to finance Social Security out of general revenue. The projections from the Social Security trustees show the program first facing a shortfall in 2033 after which point it will only be able to pay a bit more than 75 percent of scheduled benefits. While this date is still fairly far in the future, at some point it will likely be necessary to address a shortfall.

It is reasonable to expect that the changes needed to keep the program fully funded will involve some mix of revenue increases and benefit cuts. However if the chained CPI is adopted as part of a budget deal unconnected to any larger plan for Social Security then it effectively means that there will have been a substantial cut to Social Security benefits without any quid pro quo in terms of increased revenue. This hardly seems like a good negotiating move from the standpoint of those looking to preserve and strengthen the program.

There is also the question of whether the Social Security trustees will even “score” this cut accurately. In the 1990s there were changes to the CPI that had the effect of reducing the measured rate of inflation by at least 0.5 percentage points annually (Economic Report of the President 1998 Box 2-6). This would have implied a reduction in the annual cost of living adjustment by this amount and a corresponding improvement in the Social Security trust fund’s prospects. However, there is no evidence of this improvement in the program’s finances during this period. In fact the projected rate of real wage growth (the difference between the nominal rate of wage growth and the measured CPI) was 1.0 percent in 1995, before the changes to the CPI. The projected long-run rate of real wage growth had actually been lowered to 0.9 percent in the 1998 Trustees Report (Table II.D.1) which was issued after all the reductions in the CPI had been put in place.

It is important to remember that the trustees projections come from the trustees, not the professional staff of the Social Security Administration. Four of the six trustees are political appointees of the president. It is certainly possible that the cuts associated with the adoption of the CPI will not be factored into the trustees projections just as the even larger cuts associated with the changes in the CPI in the 1990s were not factored into the trustees projections.

Finally, it is worth commenting on the idea of tampering with statistical measures to achieve budgetary goals. The United States has been fortunate in having independent statistical agencies that have fiercely resisted efforts to manipulate data for political ends. In fact, in the 1990s there was considerable pressure placed on the Bureau of Labor Statistics to make adjustments to the CPI which would reduce Social Security and other indexed benefits. Katherine Abraham, the then head of the agency was steadfast in refusing to make any changes to the index that were not justified by BLS research.

The current effort has the spirit of using statistics for political ends, for example by refusing to have BLS produce a full elderly CPI so we would actually know the inflation rate experienced by the elderly. There also has been some discussion of leaving some programs, such as Supplemental Security Income, tied to the current CPI so as not to hurt a seriously disadvantaged population.

Congress can decide the benefit formula for these programs as it chooses. The honest way to cut benefits is for Congress to explicitly vote to cut benefits, not to try to hide a cut behind a statistical manipulation. This is the sort of behavior that encourages public contempt for politicians and the political process.  

For more in the impact of the chained CPI on elderly women, read Bryce Covert's latest.