Tuesday, October 27, 2015

Jeb Bush releases Social Security reform proposal

Former Florida Gov. Jeb Bush has released a Social Security reform proposal designed to balance the program’s finances, establish a new minimum benefit for long-term low-wage workers, and increase incentives to delay retirement.

I analyzed the plan using the Policy Simulation Group’s microsimulation models. You can find the full analysis, including detailed provisions included in the plan, here.



WilliamLarsen said...

"Adjust Social Security benefit formula. Social Security’s current benefit formula replaces 90 percent of Average Indexed Monthly Earnings up to $826, 32 percent of earnings between $827 and $4,980, and 15 percent of earnings between $4,981 and the contribution and benefit base, currently $9,975 per month. Between 2022 and 2040, the proposal would increase the first replacement factor from 90 percent to 93 percent, reduce the second factor from 32 percent to 21 percent, and reduce the top factor from 15 percent to 5 percent."

Andrew when does this take affect. I either missed this in reading it or it was not there.

"Calculate COLAs using Chained CPI. Beginning in 2018, the proposal would use the Chained Consumer Price Index for All Urban Consumers to calculated annual Cost of Living Adjustments for Social Security benefits. It is assumed that the Chained CPI produces COLAs that are 0.3 percentage points lower on an annual basis than would be calculated using the Consumer Price Index for Urban Wage Earners and Clerical Workers (CPI-W), which is currently used to calculate COLAs."

Does this really do anything since the trust fund is projected to be empty by 2034?

By reducing the second and third bracket percents from 32 to 21% and from 15 to 5% would this not reduce the amount of taxable social security benefits and thus the federal income tax that feeds back into the trust fund?

I see this as nothing more than changing the position of deck chairs on the SS Titanic in order to change the list and reduce the flooding rate; negligible. It reduces benefits for 70% of the workers while raising benefits for low wage workers.

For those born after 1985 instead of across the board cuts and get 29 cents in benefits for each combined dollar of payroll OASI tax and credited US Treasury Interest, it will be much lower than this. This proposal could now shift the majority of workers to a zero to negative 1% rate of return. Considering mortgage rates are 4% or higher, it would be stupid to "loan/pay into" Social Security 10.6% of your wages while paying 4%+ to borrow money to buy a home, student loan debt, etc.

I vote no on this proposal.

Andrew G. Biggs said...

The AIME factor changes are phased in between 2022 and 2040.

Yes, reducing COLA s does help the trust fund's near-term solvency, more than any other provision I'd guess, because it reduces the growth of benefits for current, rather than merely future, beneficiaries.

Yes, reducing the replacement factors will reduce proceeds from income taxes levied on benefits.

I haven't completed a full benefits analysis of the Bush plan, but I can tell you that the ratios of PV benefits to PV taxes aren't anywhere near 29 cents on the dollar -- they're far higher than that. IRRs are trickier, since you can get infinite and -100% values, but I'm pretty confident the majority won't receive negative IRRs either.

In any case, poor moneys-worth ratios are baked in the cake for Social Security and there's really not much you can do about it -- if you raise taxes to make things solvent, you make the system a poorer deal, but the same thing happens if you reduce benefits. The only better deal is if the program can pay full benefits without raising taxes, but that's impossible.

WilliamLarsen said...

Andrew, when I calculate the average benefits paid v total SS-OASI taxes paid for an individual, I do not use present value or discount rates. I assume future US Treasury Rates, COLA and Wage growth and index the SSA wage index and bend points. I then calculate what I would call the theoretical value of the benefit. I then adjust this initial benefit by COLA if it exist (trust fund exhausted or less than 20% its zero). I then look at the total benefits paid and divide it by the theoritical value of the combined SS-OASI tax and credited interest.

My program provides me with with the scheduled benefit of each cohort for each year and total cost of OASI for each year. When that cost exceeds the ability of the combined trust fund and OASI revenues, I assume across the board cuts. I use the theoretical "payable" benefit and not the theoretical "scheduled" benefit.

I recently did another stab at this (20+ years later) and got similar results.

When I look at a person born in 1985, they turn 67 in the year 2052. Payable benefits in that year are less than 63%.

So when I say that a person born after 1985 can expect to receive 29 cents in benefits for each combined dollar of SS-OASI tax and credited US Treasury interest, it does not mean their return is negative. The time value of money is about 3-4 times the value of input over a workers lifetime. So for every $1 dollar paid in tax, this one dollar over time will earn about $3.50. However, as the internal rate of return to the beneficiary drops, so does the benefit.

We agree 100% on your last paragraph. Since this is the case for eternity, why do we support throwing good money after bad? In 2001 there was still a chance to salvage over $10 trillion worth of Social Security Taxes by workers yet congress continued the same program. Now with about $2.8 Trillion in Special Treasuries which represents the sum total of 160 million worker contributions and 43 million OASI beneficiaries, the buying power has been greatly reduced. There was no painless way to extract ourselves, but going forward with SS means we just increase our pain and misery.

Andrew G. Biggs said...

I think you're not really calculating money's worth ratios in a consistent way, which is to compare the PV of expected benefits to the PV of taxes, discounted at the trust fund return, usually for a person who survives to retirement age. Taxes are pretty straightforward -- say, for a medium scaled earner you can calculate them directly, the calculate the PV as of retirement. For benefits, you need to project expected benefits, including COLAs, up to the highest reasonable age to which a person can live -- 100 is common. So for age 65, the benefit would be what they'd get at that age; for age 66, it would be the 65 benefit * (1+COLA) * (1 - chance of dying in previous year)/(1+discount rate). Continue that through age 100 or so, then add them all up. If the ratio of benefits to taxes = 1, then you've received an internal rate of return equal to the trust fund yield. Using that approach, which is pretty standard, you'll find typical benefit/tax ratios of less than 1, but not nearly so low as 0.29. Say, 0.75 wouldn't surprise me.