More on COLAs (and the CPI-E)

Following Friday's announcement by the Social Security Trustees that there would be no automatic increase for Social Security benefits in 2011 and the slew of compelling arguments about why that is appropriate, there has been significant discussion about how prices have changed over the past few years and how Americans aged 62 and older have faired. (To read why politicians should not go the pander-to-seniors-route, see our commentary on the announcement and the political pandering of lawmakers here, the Washington Post editorial here, the Los Angeles Times piece here, and CNN piece here.)

But aren’t costs for the elderly going up faster? Some have asked us. Take for instance, a comment we recently received on our website:

“…the price index used for S.S. [calculation] is based on the "market basket" of wage earners and clerical workers--in other words, the actual spending of retired persons is deliberately excluded from the weights in the index base! Spending by the elderly includes proportionately more out-of-pocket medical expenses than spending by households of working age, and medical costs consistently rise faster than the costs of other components of household budgets…. In view of the typical under-estimation of the inflation actually experienced by seniors and others with high medical expenses in calculating their COLAs, I think a modest one-time extra payment at a time of general economic hardship is not outrageous."

Well, that is a great question and a legitimate concern. Let’s address it…

Since annual cost-of-living adjustments (COLAs) are calculated from third quarter to third quarter changes in inflation (as measured by the CPI-W), here’s a look at price changes over the past few years according to a few distinct measures.

  CPI^ % Change Change in S.S Benefit Levels for Following Year* CPI-E % Change
Chained CPI % Change
2007  203.6 2.3%  2.3  224.8   2.6% 120.4 2.0%
2008  215.5  5.8%  5.8  236.3   5.1% 126.7 5.2%
2009  211.0 -2.1%  0  233.0  -1.4% 125.1 -1.2%
2010  214.1  1.5%  0  235.2   1.0% 126.3 1.0%

Note: CPI estimates shown reflect non-seasonally adjusted data for the third quarter (July, August, September) of the given year, as used by the Social Security Trustees in calculating annual COLAs.
^Reflects third quarter average for the Consumer Price Index for all urban wage earners and clerical workers (CPI-W).
*Based on percentage change in CPI from previous year to current year.

First off, we all agree that while prices (as measured by the CPI) have increased over the past year, they have not yet returned to the levels reached in the third quarter of 2008 - the period when the last COLA of 5.8 percent was set. And remember, COLAs cannot be negative--they face a zero bound.

But what’s also important is how inflation has fallen according to other measures as well, specifically the CPI-E which seeks to measure price changes affecting those aged 62 or older.

The experimental Consumer Price Index (CPI-E), shown in the table above, attempts to measure price changes affecting those aged 62 or older. The CPI-E differs from other CPI measures only in the relative weights of the 211 categories in the measure's basket of goods to reflect purchasing patterns among more elderly Americans. Granted, the CPI-E has not fallen as much as the CPI over the past few years, but even according to this alternate measure, prices have not reached their peak 2008 levels.

But don't energy prices affect older Americans less, meaning that prices for them didn't fall as much after 2008? Some critics contend that since the CPI-E puts less weight on the changes in costs of energy (because older Americans presumably have fewer energy and fuel needs), then prices haven't really dropped for older Americans as much as is being claimed. But extending this logic back to the last COLA calculated in 2008 would mean that older Americans didn't need the large 5.8 percent increase in benefits levels for 2009--since most of the uncharacteristically high COLA for 2009 was based on inflated energy prices during the summer of 2008.

Of course, CPI-E is not without its downsides though (including substitution bias problems, sampling issues arising from seeking to only focus on people aged 62 and older, and not factoring in senior citizens discounts). See CBO's report on inflation measures for more info.

But haven’t Medicare Part B premiums gone up, meaning that seniors should still receive a COLA? Well, not for the large majority of Medicare beneficiaries. With Social Security benefits frozen in nominal terms, Medicare Part B premiums will also remain frozen for about 75 percent of all Medicare beneficiaries due to the "hold-harmless" provision, which prevents Medicare beneficiaries from paying higher Medicare premiums (despite increases in Medicare expenditures) when they don’t receive an annual COLA. So, arguing that Medicare premiums have been increasing as a rationale for a COLA doesn’t apply for most beneficiaries.

Another significant argument for why actual prices have likely not returned to their peak 2008 levels is that many experts and economists believe that CPI measures (including CPI-W and CPI-U) actually overstate inflation as a result of this economic substitution problem (i.e. people changing their spending patterns between categories and not just within them in response to changing prices). To correct for this problem, economists developed another measure of inflation, the chained CPI. CBO concludes that the chained CPI has been on average 0.3 percent points lower each year than normal CPI inflation estimates. So we’re not talking about a “typical underestimation of inflation experienced by seniors”, but more likely an overestimation.

Thus, even according to the chained CPI, prices haven't returned to previous levels, meaning that the actual purchasing power of benefits today are higher than they were last year. And if Social Security relied on the chained CPI, the last COLA would have been 5.2 percent instead of 5.8 percent.

The public and our readers continue to bring up great questions about inflation and Social Security COLAs. But no matter how you look at it, the slight price increases this past year just haven’t been large enough to warrant a new COLA. Yet this hasn’t stopped lawmakers from proposing a CORA—a “Cost of Reelection Adjustment" (or as Andrew Samwick puts it at Capital Gains and Games, a “Cost of Lobbying Adjustment”).