CBO Releases Report on Social Security Reform Options

Last Friday, CBO released a report on 30 options for changing the Social Security system, analyzing both the savings and distributional effects of the options. After CBO released such dismal long-term budget projections last Wednesday (see our analysis here) showing that population aging is set to become the largest driver of entitlement spending through 2035, surely any real reform to our country's finances over the medium- and long-term must make changes to the largest single government program.

The report first focuses in the imbalance between projected revenues and outlays (actually, CBO even expects Social Security deficits to begin this year).

While outlays are expected to increase from about 4.8% of GDP currently to 6.2% in 2040 and 6.3% in 2080, revenues are projected to mostly stay flat at around 5%. This leaves a program imbalance that grows from 0.3% of GDP in 2020 to 1.3% in 2040, staying in the 1%-1.5% range throughout the rest of the projection. The total imbalance over the next 75 years is 0.6% of GDP or 1.6% of taxable payroll.   


In discussing the reform options, CBO groups the 30 Social Security options into five broad categories:

  • Increases in the Social Security payroll tax
  • Reductions in people's initial benefits
  • Increases in benefit for low earners
  • Increases in the full retirement age
  • Reductions in COLA's that are applied to continuing benefit

CBO measures each option by how much of a change in the 75-year outlook an option would make, both as a percentage of GDP and payroll. They also find out when the trust fund would be exhausted under each option, but we generally find it more helpful to look at the cash flows of the program. This way, we can see the effects on the budget, as well as if gradually phased-in changes (such as changes in indexing) would put the program on balance when their effects start to accumulate. 

It is also worth noting, as CBO does, that different changes to Social Security can affect the future economy in different ways and should be taken into consideration. As they note, raising taxes either through the existing FICA tax or by raising the tax cap would give workers incentives to shift more of their wages into tax-free fringe benefits, especially high income workers who have more flexibility in choosing their compensation. Considering that the excise tax enacted in the health care legislation was designed to do the opposite (or at least reduce the amount of fringe benefits taken in the form of health insurance), this incentive would probably be considered counterproductive. Additionally, CBO notes that benefit reductions might incentivize workers to save more for their retirement, and the increased savings rate would help economic growth. This case is especially true of middle- and high-income workers, which have more after-tax income to save in the first place.

The tax increases come in two forms: changes in the current payroll tax rate or changes in some way involving the FICA tax cap (set at $106,800 currently). CBO offers rate increases of one, two, and three percent, phased in at different rates. None of these options would eliminate the cash flow deficit, although the three percent increase would cut the 2080 cash flow deficit significanly, down to 0.2% of GDP. Eliminating the payroll tax cap without providing new benefits would eliminate the program deficits in the near-term but it would fail to completely close the gap in the years beyond 2040, raising a flat 0.9% of GDP through all those years. Raising the cap to $250,000 without providing new benefits would improve cash flow in 2080 by 0.6% of GDP, cutting the program deficit in half. A more politically realistic option (similar to one proposed by Rep. Robert Wexler) that would tax income above the cap at 4% would improve cash flow by 0.1% of GDP in 2080.  

To understand the myriad of benefit reduction options, one must first understand exactly how Social Security benefits are calculated (the CBO report also explains it). Basically, the Social Security Administration (SSA) takes a worker's average monthly earnings over his top 35 earning years and indexes that for wage growth. This number is the AIME (average indexed monthly earnings). The number that SSA comes up with is used to determine the benefit, referred to as the Primary Insurance Amount (PIA). The PIA is determined in 2010 by multiplying the first $761 of the AIME by 0.9, then multiplying the next $3,825 by 0.32, then any amount after that minus the first $4,586 by 0.15 (see here for an example). The points where the multipliers change, $761 and $4,586, are referred to as "bend points" and the multipliers are referred to as "PIA factors." After the initial benefit is determined, SSA then indexes the benefits semi-annually based on price growth; these are "COLAs."

Because there are a lot of moving parts to the calculation of benefits, there are also many ways to reduce benefits, both initially and in the indexing of benefits thereafter. For reductions in initial benefits, CBO offered many options, whether it was increasing the number of earning years included in the calculation, reducing the PIA factors, changing the indexing of lifetime wages in the calculation of benefits, or indexing the PIA factors differently.

The biggest change from the initial benefits options came from the option to reduce the PIA factors by the amount of real wage growth. This would immediately start reducing benefits from current law (by 1.3 percent each year in CBO estimates). The change would improve the cash flow in 2080 by a whopping 2.6% of GDP and the savings would continue to grow over time. Such a drastic option is probably overkill though, and it would be politically unfeasible to cut benefits by 40 percent, which is what this option would do by 2080 relative to current law. 

In fact, many of the big options that would eliminate cash flow deficits in one fell swoop are politically difficult. It is more likely that Congress will enact a combination of smaller options on both the revenue and benefit sides that together would put Social Security in balance. Here are some smaller options on the benefit side:

Option Improvement in 2080 Cash Flow (% GDP) 75 Year Improvement (% GDP) 75 Year Improvement (% Taxable Payroll)
Index AIME to Prices 0.5 0.2 0.5
Reduce Top PIA Factor By One-Third 0.1 0.1 0.2
Index Initial Benefits to Changes in Life Expectancy 0.6 0.2 0.6
Raise Normal Retirement Age to 68 0.2 0.1 0.4
Index Retirement Age to Life Expectancy 0.5 0.2 0.5
Base COLAs on Chained CPI-U 0.2 0.2 0.5
Increase AIME Computational Period to 38 Years 0.1 0.1 0.2

 CBO has done a good job of laying down the options here. Unlike Medicare, which is a more difficult and complex problem to solve, Social Security's options are clear and stir up less disagreement on the technical estimates of reform options. Most importantly, if we enact a plan now, the options can be less painful and they can be phased in more gradually.