Wednesday, April 16, 2008

Social Security Low Cost & the 100/100 Target

100% of scheduled benefits plus sustainable solvency with a 100 Trust Fund Ratio.

The current system of reporting Social Security solvency depends on a projection model. The Trustees ask the Office of the Chief Actuary to come up with a range of possible economic and demographic outcomes and score solvency under models that purport to represent the median (Intermediate Cost), the top (Low Cost), and the bottom (High Cost) of that range. This has led to a certain amount of controversy with one camp arguing that Intermediate Cost by coincidence or design has been too pessimistic, while another camp argues that it is in fact carefully constructed in light of best available information. Well that is not getting us anywhere, quarreling over which 75 year economic model is superior is kind of laughable considering that we don't have consensus over what happened over the last 75 days. (Are we in recession yet?) I suggest a new approach.

Instead of projecting why don't we try targeting? And a good starting point is the current schedule of benefits. To recap the situation. Social Security benefits are adjusted to capture the changes in real wages over a workers lifetime. What this means in practice is that historically Social Security has been a better deal for each generation, offset by some increases in payroll taxes along the way. Under the current schedule benefits in 2041 are set to be about 160% compared to the benefits a similarly situated retiree gets today. However under Intermediate Cost Assumptions the accumulated surplus in the Trust Funds is scheduled to run out in that year resulting in an immediate reset in benefits to 78% of the schedule. Now one way to look at this is to apply Rosser's Equation (after Prof. Rosser of JMU who pointed me to these numbers) and see that 78% of 160% = 125% and call the deal done. Because a benefit 25% better than my Mom gets today guaranteed until I am 84 doesn't exactly rise to the level of personal crisis. Then again the world doesn't revolve around me.

So lets just assume that we want to target 100%. But what about the other 100? What is this Trust Fund Ratio you talk of?

The Trust Fund Ratio is simply a measure of the trust fund balances expressed as a function of time with 1 year = 100. If Social Security income from all sources continues to track total cost you end up with a stable Trust Fund Ratio. More income than cost ratio goes up, less income than cost ratio goes down. By law the Trustees are mandated to keep a minimum Trust Fund Ratio of 100. This allows for a certain degree of short term fluctuation in employment and hence payroll receipts to occur with creating financing gaps, keeping the TF ratio above 100 is a reasonable policy outcome. But there are limits, at some point a high TF ratio becomes actually a threat to the health of Social Security itself. Oddly enough you can really be too rich.

It works like this. Social Security surpluses are currently invested in Special Treasuries. Despite some hysteria these bonds perform just like any other Treasury, they carry specific interest rates and specific terms and are backed by the Full Faith and Credit of the United States. They are just as real as that dollar bill in your wallet. But is exactly because these bonds are real that they can become a problem. To see why you have to examine the past and present projection of the Low Cost Alternative.

From 1997 to 2007 Low Cost always returned the same outcome, a fully funded Social Security system with a flat TF ratio through the 75 year actuarial window meaning that total income tracked total costs putting the whole system in long term equilibrium. Now while this is a perfectly adequate outcome it would not in fact be an ideal outcome, the TF ratio settled out at high enough levels that the cost of paying the interest starts representing an appreciable burden on the General Fund. Not a big burden but enough of one to get noticed, under 2007 Low Cost the General Fund would be required to pay out about 10% of the interest owed in years after mid-century or about 2% of total system cost. Now there is nothing particularly unfair about a 98%/2% split between payroll tax and general fund, on the other hand the obligation never stops even while the original utility of the excess borrowing fades away. Table VI.F8.-Operations of the Combined OASI and DI Trust Funds, in Current Dollars, Calendar Years 2007-85

The 2008 Report tossed us an entirely different complication, under the newest Low Cost projection the system never settles into equilibrium, instead you get outcome I in the figure in this post Shape of Low Cost the tail doesn't flatten and the TF ratio rockets up to 650 in 2085 and rising from there. To see why this is a problem you have to take this process to its logical end. A TF ratio of 2000 at an interest rate of 5% would require General Fund transfers equal to 100% of overall system cost to restore equilibrium. Which is to say that Social Security would lose all semblance of being a worker financed insurance system and simply be transformed to welfare. In fact I would argue that any TF ratio above 1000 is a longterm danger, restoring the system to equilibrium requiring a transfer of more than 50% of total cost and so making SS into at least a partial welfare program. But by a really cruel mathematical irony the only way to drive the TF ratio down is by cutting the revenue from the payroll tax and so in the short term INCREASING the share borne by the General Fund. Instead you need to get this particular tiger's tail under control before the acceleration sets in.


Andrew G. Biggs said...

A couple quick thoughts:

First, while wage indexing means that real benefits will be higher for future retirees, they won't necessarily be a better deal (which relates taxes to benefits, such as with the internal rate of return). People retiring today, for instance, have a lower retirement age and paid lower average taxes over their lifetimes. People in the future will have paid higher taxes (eg, the 12.4% rate would have been in place their entire careers) and the NRA would have risen from 66 to 67. Overall, their deal from the system will be poorer. For instance, see

Second, while 'Rosser's equation' may show that real benefits post trust fund insolvency will be higher than today's benefit, it's not nearly as extreme as you think. The benefit at age 65 for a medium wage earner retiring today is $15,732 while for a similar worker in 2041 it's $20,901, 78% of which equals $16,302. So it's about 4% more, not 24% more.

In any case, though, we shouldn't treat this as if it's not a big deal. The new retiree post-TF exhaustion will receive a little more than todays' level, but still a lot less than he expected to receive. Moreover, these cuts wouldn't just be for new retirees but for everyone. So the 90-year old retiree will also receive the 22% cut; she'll be receiving a lot less than she did before, and with no options to make it up.

The key is to start ahead of time. If you make changes smoothly then there's no need for trust fund exhaustion scenario. If things turn out better than we though we can always adjust along the way, but it makes sense to make small, predictable changes rather than take our chances with large sudden ones.


Bruce Webb said...

"The benefit at age 65 for a medium wage earner retiring today is $15,732 while for a similar worker in 2041 it's $20,901,"

Well I'll differ to you and Prof. Rosser on the numbers. But I would wonder about the definition of 'similar' here. A 65 year old taking retirement in 2041 is taking early retirement and so taking a discounted check to start with. Whereas I believe a 65 year old taking retirement in 2008 is still taking full retirement.

Well no. 10 seconds of research showed that a current 65 year old (and so born in 1942 or 1943) is only taking a partial haircut by retiring at that age instead of taking full retirement at 66. But I still have to wonder about apples and oranges here.

"The new retiree post-TF exhaustion will receive a little more than todays' level, but still a lot less than he expected to receive."

Not if he reads the newspaper. The implied notion that this change would simply come out of the blue is itself a little nutty. The whole argument for 'crisis' depends totally on selling people on the idea that the change in 2041 is important, to simultaneously argue that people won't see that change coming doesn't make a lot of sense.

Either retirees will see a cut in benefits in 2041 or they won't. In either case they have plenty of time to plan. The fundamental problem for privatizers remains what it always has been-convincing people that a guaranteed phasing in of benefit cuts is better than some sudden benefit cut that may or may not happen. I own a calender, the idea that I should simply accept some package of tax increases and benefit cuts simply to avoid having some Gen-X slacker getting sticker shock in 2041 or 2046 or whenever because he never picked up his Kindle to read the news is just a non-starter.