Social Security is widely believed to protect its recipients from inflation because benefits are indexed to the Consumer Price Index for Urban Wage Earners and Clerical Workers (CPI-W). However, the CPI-W may not accurately reflect the experience of retirees for two reasons. First, retirees generally have higher medical expenses than workers, and medical costs, in recent years, have tended to rise faster than the prices of other goods. Second, even if medical costs did not rise faster than the prices of other goods, as retirees aged, their medical spending would still tend to increase as a share of income; that is, each cohort of retirees would still see a decline in the real income left over for non-medical spending. We show that Social Security benefits net of average out-of-pocket medical expenses have declined relative to a price index for non-medical goods by almost 20 percent for men, and almost 27 percent for women, in the 1918 birth cohort.
This is from the abstract of "How Well Are Social Security Recipients Protected from Inflation?" by Gopi Shah Goda, John B. Shoven, and Sita Nataraj Slavov, NBER Working Paper No. 16212.
The title plus a quick reading of the abstract will cause many readers (indeed, I think, most readers) to conclude that Social Security benefits, inflation-adjusted, are declining. But that's probably false and, just as important, the authors don't say that--they just leave that impression.
Consider, first, prices of medical care. Yes, prices are rising, but, as Joseph Newhouse pointed out in a 1993 classic in the Journal of Economic Perspectives, medical prices don't adjust for quality and quality has increased significantly. Think of the drugs, for example, that have replaced long days in the hospital.
Second, the medical care system can do so much more for us now than it did in 1983. (They consider the time period from 1983 to 2007.) Because of that fact, people choose to spend more of their own money on health care. When quality-adjusted price falls and people buy more (law of demand), if the elasticity of demand > 1, then people will spend more. I think it's true that people spend a substantially higher percent of their income on air travel now than they did in 1977, the last year before de facto regulation under Alfred Kahn began. But the fact that they don't spend as high a percent of their budget on "non air travel" doesn't mean that they're worse off.
The authors recognize this. On page two, they write:
Higher medical costs may reflect the consumption of better quality medical care, and retirees may be better off even if they are left with less to spend on other nonmedical goods. Therefore, we emphasize that we cannot draw any conclusions about changes in the utility of Social Security recipients from this analysis. All we show is that Social Security benefits may not be fully inflation-indexed in the sense that recipients with average out-of-pocket medical spending cannot, from one year to the next, purchase the same bundle of non-medical goods with their Social Security benefits. (emphasis added)
It's important that their readers recognize it too.