Again on the Facts of the Chained CPI

It's amazing how sequences in Washington can repeat themselves. 

Back in June, when switching to the chained CPI was talked about as an option in the debt ceiling debate, House Ways and Means Democrats produced a JCT distributional analysis of the revenue-raising part of it. On its face, the analysis seemed to show that the chained CPI would raise taxes significantly more on lower-income earners, but it wasn't what it seemed.

We pointed out that the analysis was flawed in a few ways: it used "percent change in taxes paid" as a metric (so going from $10 to $20 in taxes would count as a 100 percent tax increase); it ignored millions of low-income taxpayers who did not file tax returns; and it assumed that AMT patches were not enacted (which would affect middle-income taxpayers). It should also be noted that the JCT analysis itself showed that the distribution of taxes paid would stay roughly the same across income levels. A week later, the Tax Policy Center came out with a distributional analysis showing that in terms of "percent change in after-tax income," switching to the chained CPI would be roughly distributionally-neutral.

Last week, in response to the inclusion of the chained CPI in Super Committee negotiations, again there was a blanket criticism of the chained CPI pointing to the JCT analysis as a drawback of the revenue impact. And, once again, TPC has produced a distributional analysis showing that the chained CPI is roughly distributionally-neutral. Also, while 30 percent of people in the bottom quintile would see their taxes go up, 97 percent of people in the top quintile would face higher tax bills. And as a share of the total federal tax change, higher-earners pay much more. The top quintile would pay over 40 percent of the tax increases from the chained CPI.


Furthermore, the Ways and Means document noted the impact of switching to the chained CPI on projected Social Security benefits would be greater for older beneficiaries because the impact compounds over time. While this is correct, giving people higher than inflation adjusted benefits makes the COLA mroe than an inflation adjustment - a COLA should be a COLA, not a raise. Keeping an outdated and incorrect measure of inflation in law is a poor substitute for boosting the retirement security of those at risk of outliving their savings. Policymakers could better target assistance to this population, such as through an old-age bump up for those over a certain age, as several bipartisan fiscal plans have proposed.

The chained CPI has come up in deficit talks over and over again because it is a more accurate barometer of the rate of inflation, allowing for programs and many features of the tax code to more correctly account for price changes. We hope lawmakers can agree to finally put it into place.