Sunday, January 08, 2006

Invest or Divert

Let's do both.

Social Security financing projected forward is complicated by a confusion of first and second order income streams. Now Cost is pretty fixed, it varies depending on your assumptions about Real Wages, CPI and various demographic figures but is conceptually easy. Each month a certain number of checks in certain amounts have to be mailed out. But income is conceptually more difficult and requires some breakdown.

The simplest component is Payroll Tax. 12.4% of every paycheck up to about $90,000 flows to the Treasury each pay period. No one questions that this is real money really extracted from the real economy. A lesser known component is taxation on benefits. In some cases higher earning beneficiaries pay tax on their benefits. This gets a little murkier. If this tax was applied across the board as a simple reduction in benefits at the top end of recipients this would not show as an economic extraction from the real economy at all. It is only because it is applied to checks actually mailed out that it even appears as an extraction from the real economy. A third component is interest on excess payroll tax invested in Special Treasuries. This money is pretty real as well. To the extent that borrowing from the Trust Fund just replaces other borrowing (a question for another day), these are just dollars the Treasury would have been paying out to some other investor. But then comes the dread Interest on Interest

Now there are Papa Bear scenarios which would have payroll tax and taxation of benefits fully capable of covering costs forever, and then some. See Goldilocks and the Three Social Security Bears. If this happens our conceptual course forward is pretty simple. We invest the interest on the existing bonds and any principal payment in alternative economic vehicles (my choice would be Municipal and School bonds) and then direct the returns from that back into the income stream. But if payroll is more than handling cost what do we do with the overage? One answer is to split it in thirds: one third to be reinvested, one third in payroll tax cuts, and another third diverted to Medicare (others might balance the ratios otherwise). Now given that the General Fund under this scenario is gamefully kicking in its $100,000,000,000 in interest owed, or even making additional payments on the principal, we end up with a portfolio that is not only steadily reducing payroll tax but kicking in substantial amounts to Medicare besides. All while capping the General Funds responsibility to actual interest owed.

There are other Papa Bear scenarios that are not so rosy. But they all have the same end benefit to the General Fund. If we just started paying out interest in full in the form of buying alternative instruments for the Trust Fund and keep economic productivity anywhere above Baby Bear's 2.1% ultimately we just dig ourselves out of the hole. Depending on how close we get to Baby Bear we may end up having to tap some proportion of the Alternative Portfolio. But in each case we evade the Interest on Interest trap.

Sounds like a fairy tale? Put on your green eyeshades and run the numbers.

Goldilocks and the Three Social Security Bears

I just updated a new diary at MyDD with the above title Goldilocks rather than repost it here, I would refer you there where if you wanted you could login (or set up a Scoop account if you are not already a member) and comment.

To summarize: the current model used by the Social Security Trustees, that of presenting Low Cost, Intermediate Cost, and High Cost as a range of economic outcomes has broken down and it has become necessary to reformulate the models. Which process becomes easier by renaming them. I use the Three Bears.

Baby Bear replaces Low Cost. Baby Bear is a model that produces a fully funded Trust Fund with a flat Trust Fund Ratio. This is in practice what Low Cost has produced for the last decade What is the Low Cost Alternative. Rather than argue whether it is an optimistic number or not, or likely to come to pass or not, we can take it for the question it answers: "What set of economic and demographic numbers gives us fully funded Social Security with a flat Trust Fund Ratio?". Which is to say which is the perfect porridge, not too hot and not too cold.

Mama Bear replaces both Intermediate and High Cost. Take Baby Bear and assume the economy performs worse. Once again we need not worry too much about whether this is realistic or not. It is just a model that projects the effects of an economy whose porridge is too cold.

Papa Bear is new on the scene. Papa relies on the economy producing a better result than Low Cost. Now where you set Papa could set off endless debate. I suggest a simple mechanism: just replace the previously projected second year numbers with the numbers just in. For example the 2005 Report projected a set of numbers for 2005 and another for 2006 and then more for the out years. Leaving the numbers for the out years alone, simply substitute real world 2005 for originally projected 2006 and then do the math. Now this porridge is not particularly hot, the assumption that the economy will perform for a single year pretty much as it did in the past year and then dive back down to the numbers of Baby Bear is not optimistic at all. But it does produce a computable surplus above and beyond what is needed to fully fund Social Security. At least it does this year.

The diary goes on to propose a complicated and not totally thought out mechanism for diverting this one year surplus to other uses, either as a rebate or to Medicare. The key is that we avoid the debate about whether we can predict the economy three years, ten years or seventy five years out. Baby Bear is a model which produces a specific desired result. Papa Bear is a testable model: by the end of the very next year you can determine whether you were correct or not. And in between is a mechanism that regulates the flow of income excluding interest into the Trust Fund.

There is another component. Baby Bear assumes that the Trust Fund is real, that the interest on the Trust Fund is real, and most importantly that the interest on the interest of the Trust Fund is real. The latter is the problem. See the post below.

Interest on Interest: a threat

The Treasury Bonds in the Trust Fund are real. At least those purchased by current payroll tax dollars. They are the product of actual payroll dollars extracted from real paychecks. And in turn the interest earned on those Bonds is real. The General Fund would have had to pay those same dollars to a bond investor if they didn't have the Trust Fund to borrow from instead.

But the Interest on the Interest is more troubling. It is second order. The General Fund is simply assuming an obligation for the convenience of not reducing the current payroll tax to make it a real pay-go system. Now the decision to raise payroll taxes in 1983 was perfectly necessary, for the most part it just restored paygo, the amount of excess payroll tax over cost actually collected is vastly overestimated by just about eveyone, the Trust Fund did not break the $100 billion mark until 1988. And investing the suplus in Special Treasuries was equally sensible, why set up any elaborate system when the whole Fund would be depleted by 2023 anyway.

But 2023 has now turned into 2041, and given not very extraordinary numbers may begin to stretch out even farther. Now the effect of reinvesting the interest in the Special Treasuries back into Special Treasures starts to bite. The General Fund starts assuming an unfunded liability, one that compounds over time as interest on interest starts to pile up. Now no one who believes in the Full Faith and Credit of the United States (and I certainly do) doesn't agree that that money is owed, the question is whether this is the best way to do it in the interests of all taxpayers in the future.

And the answer is 'No'. Given that the Trust Fund will likely be in surplus well past 2041 and perhaps forever we need to reexamine how we manage it. And the answer is pretty clear: stop reinvesting interest into Treasuries. It doesn't produce real cash flow into the Treasury, it just masks the cost of borrowing. The solution is obvious, interest on existing Treasuries and any excess of income excluding interest over current cost needs to be invested in an alternative economic vehicle. In the short term this means some borrowing pain, in the long term it shifts the reponsibility for redeeming those assets off of the General Fund. Moreover it offers the opportunity for the General Fund to eliminate any responsibility long term. If in addition to taking all interest on the current Trust Fund into other vehicles, it actually start to pay down the current principal it gradually reduces its interest burdens overall.

But given the actual financing of Social Security this means reducing the flow of other income into the system. Which means cutting the tax on Social Security benefits that applies to more affluent recipients (a relatively small amount of income) or cutting or diverting a portion of the payroll tax.