Chained CPI and Taxes / by Todd Yarbrough

In the most recent versions of the GOP tax plan there is an acknowledgement that it would be beneficial if the U.S. started basing inflation adjustment on what is called Chained CPI-U, instead of the traditional measures of CPI. The thinking goes: taxes and spending are better contained when using a version of inflation adjustment that does not, as many economists believe of the traditional CPI, overestimate inflation.

The traditional measures of inflation used for adjustment are the CPI-U (taxes) and CPI-W (social security). The CPI-U is a measure of the average cost of a "basket of goods" for all urban consumers. The IRS uses CPI-U to annually adjust tax provisions, including tax brackets. The CPI-W is a measure of the average cost of a "basket of goods" for all urban wage earners and clerical workers. The Social Security Administration uses CPI-W to adjust social security benefits over time. In practice, there is little difference between CPI-U and CPI-W, and the reasoning for using one for taxes and the other for social security has actually few corresponding arguments. They are used mostly due to convention, though the CPI-W will typically be slightly larger than CPI-U, as wage earners will be willing to increase their spending that coincides with inflation more on average than all consumers. Further, since the elderly spend a disproportionate amount of their income on high-priced healthcare services, using the CPI-W over the CPI-U ensures steady benefit inflation. To this point, economists have developed CPI-E, an attempt to capture this special type of inflation for the elderly. 

Chained CPI-U (CCPI-U) is a type of CPI which makes different assumptions about how consumers respond to price changes. In CPI-U and W, when prices rise within the basket, we simply assume this means the representative consumers see their cost rise. But consumers have the option of substituting away from these goods with rising prices, which means that many consumers avoid the increase in basket price by altering their basket. In this way, traditional measures of CPI are likely to overestimate the impact of inflation, by ignoring that rational consumers will alter their basket in response to rising prices.

Moving to CCPI-U for Social Security benefits will reduce those benefits over time compared with the current policy position of using CPI-W. This is problematic because 1) old folks are less likely to be able to substitute away from their basket than younger consumers. and 2) payments into the SS trust fund were made at whatever inflation levels were present at the time, so by moving to a lower measure of inflation we'd be reducing benefits to the elderly relative to their own personal contributions. Such certainly has the feeling of being unfair. But what about taxes?

This relates to tax policy a number of ways. First, if the IRS uses CCPI-U instead of CPI-U for tax provisions, then the inflation adjustment will be slower over time. For example, the IRS adjusts the income brackets within our progressive income tax scheme to account for inflation. Because folks experience raises over time that merely account for inflation, tax brackets need to also be moved or we get what is called "bracket creep". Bracket creep refers to earners moving into higher income tax brackets without actually seeing their real incomes rise. By switching to CCPI-U, the tax brackets will be moved more slowly over time, so relative to current policy this will feel like a tax increase for basically everyone, except those at the very bottom of the income scale. Now, it should be pointed out that while this feels like a tax increase, it is really just a decrease in the forgiveness of the tax code with respect to inflation. 

Another issue is that other tax provisions, such as credits and deductions, are also adjusted for inflation using CPI-U. Once again, if we switch to CCPI-U then these provisions will also be adjusted more slowly, relative to current policy over time. This would mean that current credits and deductions will be somewhat less favorably adjusted, and to many would feel like a benefit reduction. 

The Tax Policy Center put out a report that estimated that such a move would result in a regressive change to our tax code, as the burden of the tax increase and benefit reduction would fall most heavily on middle incomes. Those making $30-40 thousand would see their after tax income fall by an estimated 0.3%, while those making $1 million or more would see their after tax income fall by 0.1%.



This peculiarity arises most likely because the middle incomes are more susceptible to tax bracket changes, as middle incomes have more tax brackets than higher incomes. So, while the tax hike is felt by everyone, those in the middle incomes would feel it the most. This is reason enough for many to find moving to CCPI-U as non-starter. 

I think the real issue is that while economists largely agree that CCPI-U is a more accurate measure of inflation, changing policy to reflect this has real consequences beyond measuring inflation. As I've discussed above, it has the potential to feel like a tax increase that specifically burdens middle incomes by a greater amount that higher incomes. The reality is that moving to CCPI-U isn't technically a tax hike or benefit reduction. And if economists feel it better measures inflation in the economy, I'm hard-pressed to argue against its use. Nevertheless, I also feel even stronger that policy changes should not burden middle incomes more than higher incomes, as a matter of progressive principle. So, such a policy change should be accompanied, at least at the beginning of the policy change, by a system which eases the burdens on middle incomes. This could easily be accomplished alongside a move to CCPI-U by creating a provision which would allow middle income tax filers who experience an earlier bump into a higher bracket to deduct this tax increase from their income, in the same ways we allow businesses to handle profit losses when filing income taxes. Simply provide tax filers the option to deduct the difference between higher taxes paid on the higher bracket rate under the new system and the taxes avoided at the lower bracket rate under the old system. For most middle incomes, this amount is small at the beginning of the change (less than $150), so allowing them to deduct will most likely cause no revenue issues.  

There are probably better and easier ways to handle the impact on middle incomes, but the point is that we can always redistribute to a more equitable position with respect to a policy change. So, if a policy change would be regressive, we can overcome such regression by redistributing back to those negatively impacted, in this case middle income earners. But as is the case with SS benefits, moving to CCPI-U for tax purposes will likely produce similar issues of fairness of policy.