Well, who can blame her? After all, we have tens of millions of seniors living high on Social Security checks averaging a bit over $1,200 a month at a time when folks like the CEOs in the Campaign to Fix the Debt are supposed to subsist on paychecks that typically come to $10 million to $20 million a year.

Anyhow, her main trick for cutting benefits is to adopt the chained consumer price index as the basis for the annual cost of living adjustment. This would have the effect of reducing benefits by 0.3 percentage points for each year of retirement. This means a beneficiary would see a 3 percent cut in benefits after 10 years, a 6 percent cut after 20 years and a 9 percent cut after 30 years. This is real money. Since Social Security is more than half the income for almost 70 percent of retirees and more than 90 percent of the income for 40 percent of retirees, the hit to the affected population would be considerably larger than the hit to the top 2 percent from ending the Bush era tax cuts.

But Marcus insists this cut must be done first and foremost in the name of accuracy, since the chained CPI is supposed to provide a better measure of the cost of living. She notes but quickly dismisses the evidence from the Bureau of Labor Statistics (BLS) consumer price index for the elderly (CPI-E), which shows that the rate of inflation seen by the elderly is somewhat higher than the overall rate of inflation.

"The problem with that is twofold. That measure is imperfect — the “E” stands for experimental. And, as the liberal Center on Budget and Policy Priorities notes, the burden of higher health costs falls unevenly among the elderly. Average costs are skewed upward by a minority who face very high out-of-pocket expenses, a problem better addressed by fixing Medicare to deal with catastrophic costs."

Actually, the "E" stands for elderly, but let's get to the substance. First, if we are interested in accuracy then the answer would seem to be to have the BLS construct a full elderly index that tracked the actual consumption patterns of the elderly. This would cost some additional money, but we will be indexing $10 trillion in Social Security benefits over the next decade so if we want to ensure accuracy, it would seem reasonable to spend $70-$80 million to put together a full elderly index that actually tracked the consumption patterns of the elderly, looking at the specific outlets where they shopped and the items that they purchased.

It is difficult to know exactly what this would show, but it is possible that even apart from the issue of health care it would show that the elderly experience a higher rate of inflation than the population as a whole. The current index already assumes substantial amounts of substitution in response to price changes at lower levels of aggregation (e.g. different types of cell phones). If the elderly are less flexible in their shopping patterns and a less mobile population then this substitution may have the effect of understating the increase in their cost of living.

The point about health care misunderstands the way the consumer price index is calculated. The index is an average index, which weights expenditures by dollars rather than households. While the health care case may suggest that a small minority is pulling costs up, it is also likely that a small minority is pulling costs down. The vast majority of the population does not buy a new car in a given year. Yet this item, which has barely risen in cost over the last decade according to the BLS, accounts for more than 3 percent of the weight of the index.  This means that new car purchases by a small and relatively wealthy segment of the population have had the effect of reducing the measured rate of inflation over this period.

This issue arises with many other items as well. The BLS makes a point of quickly including new items (e.g. the latest cell phone or tablet computer) in the index to capture their period of most rapid price decline. Since these items will generally be expensive when they are first introduced, they will disproportionately be consumed by wealthier households and likely under-consumed by the elderly.

This provides reason to believe that the CPI understates the cost of living of the typical household and especially the typical elderly household, but Marcus is only interested in finding reasons why there might be an overstatement. There is also the obvious point with health care which is that most elderly households in most years will not have large health care expenses, but that doesn't mean that at some point in their retirement that most elderly households will not see large health care expenses. That is the point of an index that picks up averages.

Marcus's argument that this is:

"a problem better addressed by fixing Medicare to deal with catastrophic costs," is intriguing. When exactly do we anticipate that Congress is going to fix Medicare to deal with catastrophic costs?

The story here is pretty simple. If we want a more accurate index for adjusting Social Security benefits then we would have BLS construct a full elderly index. If the point is to cut benefits then we would do what Marcus advocates and switch to a chained CPI. That will teach those high living seniors.