Chained CPI Shift Doesn’t Have to Mean a Tax Hike
By: James Valvo
Lost among the frenzy to raise taxes, the only substantive programmatic spending reduction that the two sides agree on is an adjustment to the way Social Security Cost of Living Adjustments (COLAs) are calculated. The current formula uses the Consumer Price Index for Urban Wage Earners (CPI-W); the two sides have agreed to switch to a variation of that metric called chained CPI. Transitioning all Social Security recipients to chained CPI would reduce future expenditures by roughly $217 billion over ten years. President Obama has proposed using chained CPI for only wealthier retirees, reducing the savings by about half. However, the proposed change would not only result in Social Security savings, it would also be a tax hike. But this doesn’t have to be the case.
CPI is a measure of inflation: increasing prices for consumer goods over time. However, it’s a static measure. It doesn’t take into account the basic economic fact that people’s behavior changes according to the price signals they receive from the market. Chained CPI, in contrast, attempts to measure that effect. For example, if the price of a certain good increases because of inflation, some consumers will shift their purchases away from that now-more-expensive good to a more affordable alternative: a basic substitution effect. The effect of this shift is that it makes is easier for people to cope with higher prices. Using static CPI for the Social Security benefits thus provides a higher adjustment than the market really demands. Switching to chained CPI would resolve this issue. It would also result in a large reduction in projected future spending, which is why it’s currently on the negotiating table.
The current proposals to adopt chained CPI would also affect how the marginal income tax brackets are adjusted every year to prevent what’s known as “bracket creep.” Bracket creep is a stealth tax increase because, like prices, wages also tend to rise over time. If the income threshold at which an individual moves from one tax bracket to the next is not routinely adjusted upward, taxpayers find themselves graduating into higher rates when their incomes relative to purchasing power haven’t really changed. To prevent this bracket creep, the brackets are tied to CPI, just like the Social Security COLAs. The tax brackets use CPI-U, while Social Security uses CPI-W.
The current proposals to shift to chained CPI would cover both taxes and benefits. If chained CPI is used for the brackets, it will decrease the speed at which the thresholds are adjusted upward, resulting in a de facto tax increase as taxpayers will run into higher rates sooner than they otherwise would. But this doesn’t need to the case.
Using chained CPI for Social Security benefits makes sense because you’re trying to measure how far an individual’s Social Security check will carry him or her in the market, how much can he or she purchase. That’s a measure of the cost of goods and thus adjusting for inflation (or inflation taking account people’s changing behaviors) is appropriate. However, marginal income tax brackets have nothing to do with the cost of goods; they are related to wage growth. Ignoring the difference between a taxpayer receiving money (income subject to marginal rates) and a taxpayer spending money (using Social Security benefits to buy goods) creates the false equivalency we see in the Obama and Boehner proposals.
A much better solution to solving the bracket creep problem is to tie the yearly bracket adjustments to some measure of wage growth. The Bureau of Labor Statistics already provides various metrics that could be adopted. One possible solution would be to use the Average Hourly Earnings (AHE) data that surveys payrolls but excludes benefits; this would be sensible for bracket adjustments because the largest non-wage employee benefits (employer-sponsored insurance, etc.) are excluded from taxable income. One drawback to the AHE measure is that it only includes production and nonsupervisory employees. Another option is the Compensation per Hour metric, which does include benefits but also covers all employees. Of course switching to a form of “CPI for wages” begs the question: Shouldn’t we use a chained metric in this area too in order to account for individuals shifting toward careers with rising wages? Perhaps, but labor mobility is lower than consumption elasticity, and therefore the shift would not be as large.
Switching to chained CPI for Social Security benefits would reduce future spending but make an already bad retirement deal even worse for beneficiaries. However, that’s no reason to measure marginal bracket adjustments incorrectly. Adjusting benefits according to inflation and brackets according to wage growth makes much more sense.
Mr. Valvo is director of policy at Americans for Prosperity. You can follow him on Twitter @JamesValvo.