Two approaches to Social Security reform

Ezra Klein has recommended the deficit reduction plan put forward by the Bipartisan Policy Center* as "easily the most thoughtful, detailed and credible deficit reduction plan on the table. On social security, the BPC plan overlaps in two major features with the  that of the Bowles-Simpson commission. But one major difference is instructive.

First, the two major points of overlap. Both plans would gradually lift the cap on earnings subject to social security to the range of $180-190k in today's dollars, the level needed to restore a target set in 1977 of taxing 90% of Americans' wage earnings.  Both also propose to slow down the cost-of-living adjustment by moving from the current Consumer Price Index for Urban Wage Earners and Clerical Workers (CPI-W) to a so-called "chain-weighted" CPI (CPI-U) that aims to account for changes in consumers' habits when the prices of particular items goes up. According to the BPC plan, the "chained" CPI-U is estimated to grow .3 percentage points more slowly (I assume annually) than the CPI-W. According to the Bowles-Simpson plan, these two measures would close 61% of the existing shortfall over 75 years.


The most instructive difference is in changes to the benefit formula -- the percentage of their lifetime average indexed monthly earnings (AIME) that workers in various income brackets earn as a permanent social security pension.  Both plans, it should be noted, eschew so-called "means-testing" of benefits -- that is, proportionately reducing benefits for those who have significant other sources of retirement income. Means-testing is a GOP talking point for reasons I can't fathom, since it constitutes a major disincentive to accumulate wealth and would raise only modest additional revenue, since the vast bulk of social security payouts go to non-affluent Americans. 
Both the BPC and Bowles-Simpson plans propose to reduce the percentage of AIME that people in higher income brackets get back in their social security pensions. At present, rounded to the nearest thousand, we all get 90% of AIME for the first $9000 we earn annually, 32% of AIME for all payroll-taxed earnings from $9,000 to $64,000, and just 15% of tax earnings from $64k up to the payroll tax cap of $107,000.  The BPC plan would modestly and gradually shave the upper level from 15% to 10%.  Today, if the cap on wages subject to the payroll tax were raised to $180k, that reduction of benefits from 15% to 10% of earnings over $64k would cost those who had paid the maximum into the system about $5.8k annually, if I'm calculating accurately -- though by the time the cap got to $180, the "bend point" for the higher bracket would also doubtless rise considerably, reducing the hit.



Bowles-Simpson makes far more drastic cuts in the payouts provided under the current AIME formula -- and more importantly, reaches into the middle class. Oddly for those who put so much faith in reducing the number of tax brackets, the Bowles-Simpson authors would increase the number of social security "bend points," i.e. income levels at which the percentage of AIME earned in the SS pension changes.  The plan calls for gradual transition to a formula that at current income levels would pay out 90% of AIME up to $9k (as at present), 30% up to $38k,  just 10% from $38k to 64k, and 5% for income over $64k.  Today, that would reduce the benefits of someone whose AIME translated to $64k annually from $25,700 per year to $20,000. Someone between the two mid-range break points, earning $51K per year, would see a reduction from $21,840 to $18,100 (again, assuming that my sixth-grade math skills, calculator-aided, are up to snuff).



This proposal is really the centerpiece of Bowles-Simpson's social security plan; it's projected to close 45% of the shortfall over 75 years.  The BPC's AIME formula adjustment, in contrast, adds a competitively negligible amount of revenue -- $59 billion by 2040, the last year projected, compared to about $1.4 trillion for both the COLA adjustment and the payroll tax cap raise.  How can the two plans diverge so sharply on this point?


One answer may be that the BPC AIME adjustment would bring in much more after 2040, the last year of projection, compared to (I guess) 2087 for Bowles-Simpson (75 years from now).  Also, Bowles-Simpson proposes major givebacks to two groups: low earners and the very elderly.  The two proposals are to offer a minimum benefit of 125% of the federal poverty line for any individual with 25 years of work, and to boost benefits 5% for individuals who have been eligible for benefits for 20 years: each is projected to reduce the share of existing shortfall closed by 8%. Finally, Bowles-Simpson budgets for a surplus in 75 years, closing 112% of the existing  shortfall.

I wonder if that Bowles-Simpson give-back to retirees who have already had 20 years of social security eligibility wouldn't be better spent on Medicare.  On average, while very old people's medical costs increase significantly, their other expenses tend to drop.

I suppose that the Bowles-Simpson AIME adjustments are "progressive," as the plan claims. AIME is based on individual rather than household income, so I guess those sharply reduced benefits start to bite relatively high on the income scale. AIME is averaged over thirty five years, and so must be considerably lower than the income most people earn at their peak.  $38k is a higher break point than it sounds on first blush.  Still, the cut from 32% of earnings over $38k to 10% of earnings over $38k is a very large bite in what can plausibly be called middle class benefits.

*The BPC was founded by former senate majority leaders Howard Baker, Tom Daschle, Bob Dole and George; its Debt Reduction Task Force is chaired by Pete Domenici and Alice Rivlin.