HOME. . . . . . . . . ESSAYS . . . . .. . . .

I- The General Outline of the Bush Plan

From the 2000 Presidential campaign onward, George W. Bush has signaled his intention to introduce some form of Personal Savings Accounts (PSAs) as part of a reform of the existing Social Security system. In May 2001, the President chartered a bi-partisan commission[1] whose task was to suggest ways to introduce PSAs into Social Security, consistent with six principles the President insisted upon. Those principles were:

1-   Must not change Social Security benefits for current or near-term retirees;
2-   Entire Social Security surplus must be dedicated to Social Security;
3-   Payroll tax must not be increased;
4-   Government cannot be allowed to invest Social Security funds in the stock market;
5-   Will not affect survivors or disability benefits;
6-   Must included individually controlled, voluntary personal retirement accounts.

In principle, there are two general forms of PSAs: Add-On Accounts, which leave the existing Social Security revenue stream intact and use some other revenue source to create the PSAs; and Carve-Out Accounts, which divert a portion of the existing Social Security payroll tax revenue stream out of the Social Security system and into the PSAs.

In the Report of the 1994-1996 Advisory Council on Social Security, both types of PSAs were advocated by various factions on the Council.[2] In the 2000 election campaign the Democratic presidential candidate, Al Gore, advocated the add-on approach to PSAs while the successful Republican candidate, George W. Bush, advocated the carve-out approach. The PCSSS in turn devised three optional plans, one of which (Plan 2) involved carve-out accounts exclusively, one of which was a mix of carve-out and add-on approaches (Plan 3), and one of which was potentially either an add-on or carve-out approach (Plan 1)—although the PCSSS analyzed it exclusively as a carve-out account. Shortly after the issuance of the PCSSS report, President Bush signaled that he was in favor of an approach similar to that of the Commission’s Plan 2, i.e., strictly a carve-out financing scheme.

An essential functional difference between a pure add-on and a pure carve-out approach is that an add-on account incurs no transitional costs while a carve-out approach does. The reason that carve-out accounts incur transition costs is because they divert funding from the payment of present and future Social Security benefits. If principle no. 1 of the PCSSS study is to be honored (no impact on the benefits of current and near-term retirees) then this missing money must be made-up somehow, at least in the short-term. This obligation to make up for the diverted funds is the transition cost of moving to a PSA system using carve-outs from the existing Social Security revenue stream. Since add-on accounts divert no funds they incur no transition costs.

Some would argue that there are also “transition costs” from establishing a PSA scheme even under a pure add-on approach. The argument here is that the funds to establish the PSAs must come from somewhere, and this will presumably be either an increase in the payroll tax or through funding from general revenues. Either way, this is a new tax cost to the taxpayers. We might call this new cost the equivalent of the transition costs to the Social Security system under a carve-out approach. The Congressional Research Service (CRS), in its analysis of the transition costs issue, took this viewpoint.[3]

While I don’t wish to quibble over semantics, equating this type of economic cost to that incurred in transforming the Social Security system into one depending largely on PSAs, is somewhat misleading. While the generic point is true enough: that even add-on accounts have start-up costs, these two types of costs cannot be simply equated with each other for the simple reason that they are not the same thing. The start-up costs of an add-on scheme are whatever the value is of the funds placed in the accounts. Whereas the transition costs in a carve-out scheme are the cost of Social Security benefits that must be made-up in order to hold harmless those Social Security beneficiaries whose benefits are threatened by the diversion of funding out of the Social Security revenue stream. These two figures might be the same in dollar value for a given investment amount (although it is possible that they might not be [4]), but they clearly are not the same thing in any sense of the word.

In any case, it is useful to be quite clear on what the transition costs are (conceptually) under a carve-out scheme like that of the Bush Plan.  The PCSSS report describes these transition costs in this way:

The current Social Security program is financed primarily on a "pay-as-you-go" basis, meaning that most of the payroll taxes paid by today’s workers are used to finance benefits for today’s beneficiaries. .  . . Under a personal account program, workers would be given the option to invest a portion of their payroll taxes in accounts that they would own. . . . During the transition to a personal accounts program, tax revenues invested in the accounts would no longer be available to finance traditional benefit payments . . .

Therefore, funds must temporarily be found to finance personal account investment while simultaneously paying benefits to retirees.[5]

Under a pure add-on scheme a separate source of revenue is obtained (perhaps an additional payroll tax or general revenues) and the existing Social Security system with its flow of income into the Trust Funds and to current beneficiaries is maintained as is. Benefits for those not under the scheme would continue as at present, while benefits for PSA participants would come from a combination of Social Security income and the PSA.

Under the carve-out scheme, the funds diverted from the Social Security system must somehow be replaced. Benefits for PSA participants come from the combination of the PSA and Social Security benefits; whereas benefits for non-participants comes from Social Security payroll taxes plus the reverted transition cost revenues.

 

II- The Value of the Transition Costs

We might think that in general, under a carve-out scheme, the value of the diverted funds (plus forgone interest) is equal to the transition costs under such a scheme. However, since the object of providing for the transition costs is to pay for the full promised Social Security benefits of current and near-term retirees (i.e., those who will not be given an opportunity to participate in the PSA plan) it is their benefit shortfalls that must be made up in the form of paying the transition costs. This may or may not be identical to the diverted funds. So to state this point more precisely we should say: The transition costs are precisely equal to the amount of money required to pay full promised benefits to those beneficiaries who will not have the option of participating in the PSA plan.  This is not in general equivalent to the diverted funds, which will continue to be diverted in perpetuity. It is, rather, the extra funds needed to maintain solvency while non-participants are alive and collecting Social Security benefits.

Under the Bush Plan as currently formulated, anyone who is age 55 or older as of 2005 is precluded from PSA participation. If we assume that the youngest of this excluded cohort can be expected to live to approximately age 95, then the transition costs will begin immediately upon the start of the PSA system (2009 in the current Bush Plan) and continue for another 36 years. Thus the transition period will be 2009 thru 2045, when those workers presently age 55 will attain age 95. Costs during this period will not be uniform, but will rise and fall with the mortality of the various cohorts in the excluded population. Costs will peak somewhere near the mid-point of this period, then gradually decline.

Determining the value of this cost is complicated by the fact that the Social Security system is insolvent over its standard 75-year valuation period. According to the 2005 Report of the Social Security Trustees, the program has a present value unfunded liability of $4 trillion over the next 75 years.[6] Thus, whatever the transition costs are, they will increase the 75-year unfunded liability, at least in the short term.

In theory, these costs should equal only that portion of the diverted funds necessary to pay full benefits to the excluded population. Some part of the diverted income will in fact be connected to plan participants and not required to make up the lost funds to the excluded population. Thus it is not the full value of the revenue diversions that occur during the transition period that defines the transition costs, but only that portion of the total necessary to make up for the shortfall created in the benefits of the excluded population.

The value of benefits paid to the included population should not be part of the transition costs since there is no obligation to maintain their benefits. Indeed, their Social Security benefits under the Bush Plan will be cut by an amount that is equal to the amount of money they have diverted out of the Social Security system plus the interest that amount of diverted money would have earned the Trust Funds. Thus the Trust Fund costs associated with plan participants are fully compensated by the subsequent benefit cuts that plan participants must endure as the price for plan participation.

This distinction—between the excluded and the included populations—implies that the “claw-back” benefit reductions that are part of the Bush Plan cannot be the mechanism for paying the transition costs, unless the claw-backs are deliberately “over-inflated” to produce more revenue than that diverted by Plan participants, so that the excess might be available to pay the transition costs. It is hard to see how that might be accomplished, given the definition of the claw-back. Since the claw-back is defined as the amount of a plan participant’s individual diversion from the Social Security revenue stream, plus 3% real interest, this should exactly equal the value of his/her diverted income. It is difficult to see how, even over the long-run, the Social Security program could be made whole from these claw-backs. It would seem essential that other programmatic cuts be made in the system if the transition costs were ultimately to be repaid by the Social Security system. 

The PCSSS implicitly recognizes this point since both of its carve-out plans include, in addition the claw-back provision and the general benefit cuts due to indexing, a general revenue subsidy to make up the remaining gap in funding. Under Plan 2, there is a general revenue subsidy beginning in 2025 and continuing until 2054. Its cost is ambiguously stated as averaging 1.2% of GDP during the decade 2030-2040.[7] In other words, the PCSSS carve-out plans pay for the transition costs through a combination of benefit reductions (generalized switch to price indexing and claw-backs) and general revenue subsidies. In no sense are the transition costs recouped by “shifting liabilities in time,” or by the claw-back itself. The transition costs appear to be a net new cost to the system. Some dispute this conclusion, which is a topic we shall consider in detail in the next section.

Various estimates for the transition costs have been offered for carve-out plans along the lines contemplated by the PCSSS and the present Bush Administration plan. The estimates for transition costs for PCSSS Plan 2 are complicated by the fact that this plan has other features (such as full wage indexing and a new minimum benefit). Thus there is no estimate for the cost of the transition itself, without the offsetting and/or aggravating factors of the additional policy changes. This is also a problem with the estimates provided by the Social Security actuaries for Plan 2.[8] The current Bush Plan has never been scored by the Social Security actuaries and hence we have no independent check on the transition cost estimates the Administration has provided. Moreover, the Administration has been very unforthcoming about providing such estimates and documenting their basis.[9]

The PCSSS provided two estimates of the present value of the transition costs for its Plan 2 (as of 2001). These costs are said to be either $400 billion or $900 billion. The difference is that the lower figure includes the present surpluses in the Trust Fund as offsets against the transition costs while the larger figure is the complete cost of the transition without crediting the current surpluses against it.

Crediting the near-term surpluses in the Trust Funds as offsets to the transition costs seems to be an illicit procedure. Those near-term surpluses are already more than spoken-for and are part of the $4 trillion unfunded liability of the existing system. If they are not available to the existing system, then its unfunded liability must rise. Counting them as offsets to the transition costs would then appear to be a form double-counting: they are being used once to reduce the unfunded liability of the present program, then again to offset the transition costs. Putting this concern aside, there are two additional problems with these estimates.

Both the high and low estimates assume the other policy provisions of Plan 2 as offsets over the estimation period, which tends to lower the apparent present value cost of the transition, i.e., it is not a pure estimate of the cost of the transition. And of course these other policies face a very uncertain future: they may in fact never be available to offset the transition costs. The second problem concerns the low-cost estimate. This estimate was as of 2001, when Social Security surpluses were still large and growing. If a PSA plan is not put into effect until 2009, there will be a much shorter period of surplus income to utilize as an offset (the annual surplus stops growing in 2012 and is depleted entirely by 2017 according to the 2005 Trustees Report). As we move toward 2017, the values for the two estimates, other things being equal, converge to the larger value.

The liberal Center for Budget and Policy Priorities has authored a study which purports to show the transition costs of the Bush Plan at around $5 trillion for the first 20 years of what I have identified as a roughly 36-year transition period.[10] However, this estimate appears to be flawed by not accounting for the rise and fall in the size of the excluded population, as it appears to use the costs of the first 10 years and extrapolates them through the next 10 at the same cost rate.[11]

Diamond and Orszag have set the 75-year present value cost of Plan 2 at either $2.2 trillion or $2.8 trillion (the latter figure is the amount required to avoid reductions in disability benefits).[12]

The CRS appears to confirm the Diamond and Orszag estimates. In their report on the transition costs, CRS estimates the 75-year present value transition costs for Plan 2 as between $1.5 trillion to $2.3 trillion, depending on the rate of voluntary participation in the PSAs (the lower number assuming 67% participation and the higher figure 100% participation). As CRS concludes: “. . .  with just the individual account component, the 75-year unfunded liability under Model 2 would increase from $3.7 trillion to $5.2 or $6.0 trillion.”[13]

So according to the CRS, the introduction of PSAs through a carve-out approach will increase the unfunded liabilities of the Social Security system to $6.0 trillion, and hence the 75-year transition costs of a switch to a mixed system of PSAs and Social Security benefits will be approximately $2.3 trillion.

 

III- Have We Found the Elusive Free Lunch?-

One rather curious thesis regarding the transition costs is that they are equal to zero. We might call this the Free Lunch Argument. It purports to show that we can transition to a hybrid system of PSAs and Social Security without incurring any net new costs. Thus, something of presumably great value (PSAs) can be obtained without any economic cost. This seems like the promise of the proverbial economic free lunch. So let us examine this thesis more carefully.

Some proponents of PSAs argue that, in the long run, the cost of switching to PSAs is somehow offset by a reduction in Social Security’s existing unfunded liabilities, and so there is no net real long-term costs associated with the transition to a scheme like the Bush Plan.

The PCSSS itself suggested an even more dramatic outcome. They seemed to imply that the introduction of PSAs, by itself, would somehow produce a net cost reduction in the government’s liabilities. In general terms, the Commission argues that we cannot consider the costs of the transition in isolation without considering its savings. They represent the transition cost as being in the manner of an investment, and so the recouping of that investment has to be part of the net computation of its value. This seems a fair enough point, stated in broad principle. They offer this deceptively simple analogy:

Suppose an individual had a $90,000 home mortgage with a monthly payment of $600 over 30 years. By paying an extra $100 monthly, the individual could pay off his mortgage in full within 20 years and thereafter have an "extra" $600 per month to spend on other things. This additional $100 monthly payment is an investment that brings rewards, not a cost. . . . The so-called “transition costs” associated with personal accounts for Social Security are precisely that: saving and investing for the future, to reduce the need to raise taxes, cut benefits, or curtail other necessary government initiatives.[14] (emphasis added)

This has an initial plausibility, but we have to be careful here. It is necessary to keep in mind just what the transition costs are and what the current liabilities of the Social Security system are. If the transition to PSAs, by itself, produces some long-term reduction in the system’s unfunded liabilities that is greater than $2.3 trillion over the next 75 years, then we could accept this analogy. But if the PSAs merely shift costs in time—that is to say, if they replace the same amount of revenue that they divert—then we end up at the end of the process still facing the same $3.7 trillion in unfunded liabilities. In which case this claim that PSAs “reduce the need to raise taxes, cut benefits, or curtail other necessary government initiatives” would clearly be incorrect.

In order for the move to PSAs to actually reduce the government’s unfunded liabilities one of three things must happen: 1) either the benefits to the excluded population must be reduced in some way; or 2) the reduction in future benefits to plan participants must be larger than that required to pay back their own diversions so that the excess can be used to offset the costs of paying benefits to the excluded group during their lifetimes and whatever additional savings we imagine might be realized; or 3) some other program cuts must be made.

Option 1) seems to fly in the face of one of the most basic promises that President Bush has made regarding his plan. His first principle to the PCSSS was indeed that any acceptable plan “Must not change Social Security benefits for current or near-term retirees.” This seems to rule out benefit cuts to the excluded group. There is of course some room for ambiguity here. The President might only have meant that he promises not to cut benefits below their currently funded level, with no promise being made that the full promised future benefits would be paid. Such an interpretation, in addition to implying that the President is being disingenuous, would be incoherent when applied to those not yet receiving benefits. What would it mean to someone age 55 in 2005, who will not be receiving benefits for another 12 years, that his “current” benefits are safe but there are no promises about his “future” benefits. I think it fair to say that the President has implicitly promised the excluded group that they will receive a full benefit throughout their lives at the currently promised level of benefits (regardless of whether or not that level of benefits is presently funded).

(There is potentially a third group in addition to PSA participants and the excluded group, and that is covered workers who are eligible for the PSA plan, but opt not to join. There has been no promise under the Bush Plan that their benefits will be protected, now or in the future.)

Since the claw-back provision of the Bush Plan has been defined in a way that it is equal to the presumed actual cost of the diversion by each plan participant, option 2) is not available. (There is, however, the curious fact that the claw-backs continue in perpetuity while the transition ends at a point in time when the last beneficiary in the excluded group dies. This suggests, once again, that the claw-back has little or nothing to do with refunding the transition costs.)

So there remains the possibility of some future unspecified cuts in the program so there would be a net reduction in the system’s present unfunded liability. The President’s most recent proposal to switch to “progressive indexing” along the lines suggested by Robert Pozen might fit this bill. But notice that this switch to progressive indexing has nothing to do with the transition to PSAs—it could be done with or without a PSA scheme. So to claim that the switch to PSAs somehow has the benefits that the PCSSS implied in their analogy about home mortgages seems without support. 

Now we can return to the more modest claim that the $2.3 trillion in transition costs are somehow already included in the $3.7 trillion in unfunded liabilities facing the system. Then we could argue successfully that shifting to PSAs would be merely shifting costs in time, and not adding costs to the government’s unified budget. And indeed, some appear to be making precisely this claim.

Interestingly, the PCSSS itself seems also to have suggested this interpretation of the transition cost issue. In a separate passage on a subsequent page the Commission Report makes this argument:

All of the plans presented by the Commission provide individuals the option to invest in personal accounts. In all cases, these accounts are at least partially financed by a redirection of payroll tax revenue from the existing system. In return for the opportunity to pursue higher expected returns through personal accounts, individuals who choose the account agree to forgo the benefit that would have been financed by these payroll taxes (plus interest). Therefore, every dollar invested in a personal account reduces the cost of future Social Security payments by one dollar, plus the offset rate of interest that is proposed for each plan . . .[15]

This version of the free lunch argument clearly seems to imply that it is the claw-back provision which is responsible for making the transition to PSAs cost-free. Again, this seems confused since the claw-back provision is defined in such a way that it is strictly limited to the costs of the diversions on behalf of the participating group. It seems therefore almost true by definition that the claw-back cannot touch the costs of maintaining promised benefits to the excluded group.

Perhaps what the claim is here has been habitually mis-labeled by the advocates of PSAs. Perhaps they are not funding the transition costs at all, but rather they are funding the PSAs. So the argument would be that we can create PSAs without cost by simply reducing future Social Security benefits by the amount of the PSAs. This might be true, but it would not have anything to do with funding the transition costs. The transition costs would still have to be borne in some way. Since the Trust Funds are presently insolvent and there is no surplus to use to fund both PSAs and promised benefits to the excluded group, it would seem we cannot do both without incurring some new costs.

To be quite explicit about it: there are two sets of costs involved here: one is the cost to the Trust Funds from the diversions to PSAs for the participating group, and the second is the cost of continuing to pay full benefits to the excluded group. They are not the same costs.

The analysis by the Congressional Research Service seems to support my interpretation. CRS states: “The cost of the transition to individual accounts is separate from—and in addition to—the system’s current unfunded liability over this period.”[16]

The Cato Institute’s Michael Tanner however claims that the Transition Costs under the Bush Plan are not new debt because they are part of the existing unfunded liabilities of the system, and so shifting these existing obligations out of Social Security and into Treasury debt to create personal accounts does not create new debt.[17] Others have made this same argument.

Here is how Lawrence Hunter (an associate of Peter Ferrara at the Institute of Policy Innovation) explains the argument: “Contrary to conventional thinking, allowing workers to invest a portion of their Social Security taxes through personal retirement accounts will incur no net cost to the federal government. The misconception that there will be new costs—i.e., ‘transition costs’—arises from forgetting that Social Security is already, in effect, debt-financed. Consequently: 1) Any debt sold to the public in the short term during the so-called "transition period" to pay all promised Social Security benefits simply refinances an existing liability; and 2) Refinancing Social Security's debt through personal accounts makes possible a reduction, and eventually an elimination, of Social Security's unfunded liability, which otherwise is scheduled to grow without bound.”[18]

The language is a little ambiguous, but what Hunter seems to be suggesting is that the borrowing to fund the Transition is merely the shifting of some portion of Social Security’s existing unfunded liability forward in time, and that is why no new borrowing is involved.

So here is the implicit argument:

1-   The Social Security system presently has some unfunded liabilities;
2-   If workers opt into personal accounts, and we reduce their future Social Security benefits by an        amount equal to whatever they take out of the system, then the future unfunded liability of the        system will be reduced by an equal amount;
3-   Since we have reduced the unfunded liabilities of the system, the money being used to pay for the        transition costs is not new borrowing. It is merely the shifting in time of the unfunded liabilities;
4-   Thus, the transition to personal accounts has no new costs associated with it; it is essentially        “free.”

There are many problems with this argument. One problem is the way that advocates of personal accounts view the unfunded liabilities. Often they argue that we cannot consider the unfunded liabilities when assessing the value of the present Social Security system and so we cannot compare benefits under the Bush Plan to promised Social Security benefits, but only to currently-funded Social Security benefits. This argument implicitly repudiates the unfunded liability. But for the current argument to work, the unfunded liability must be considered real and implicitly funded so that Hunter et. al. can argue that a reduction in this liability constitutes a real savings. But that small matter of intellectual consistency aside, the main problem with Hunter’s argument may be deeper than that. To see what is going on in this transition business, we have to consider matters in careful detail.

 

IV- Following the Money-

Transition costs, remember, are what we have to pay to make good on the President’s promise not to do anything to the benefits of those people currently age 55 or older. In other words, the transition costs are the money we have to come up with in order to keep paying Social Security benefits to everyone now retired and to those who are near retirement. (Let us call these folks the Transition Beneficiaries.)

To make the math easier, let us assume that the current unfunded liabilities of the Social Security program are $4 trillion and that $2 trillion is diverted from the system to create personal accounts (assume interest and so on is accounted for). Under the Bush Plan, when Plan participants retire, they repay the government an amount equal to the amount they diverted into the personal accounts system. This repayment occurs in the form of reduced Social Security benefits. So three things are going to happen: 1) $2 trillion will be diverted out of the Social Security revenue stream; 2) $2 trillion in federal debt will be floated to continue paying benefits to the Transition Beneficiaries; and 3) $2 trillion will be paid back to the government in the form of reduced future Social Security benefits.

To analyze these transactions, we need to look both at the Social Security program and at the impact these transactions have on general government debt (assuming, for simplicity, only debt created or reduced by these transactions).

Here, then, in detail, is the sequence of financial transactions:

Step 1: Current Situation-

At the start of this sequence the Social Security program has $4 trillion in unfunded liabilities, and there is no other relevant government debt (for simplicity). Thus the total federal debt is $4 trillion.

Step 1: Current situation:

S. S. unfunded liability           =          ($4t)
Other Gov. Debt                     =          $0
Total Gov. Debt                      =          ($4t)

 

Step 2: Diverting Funds to PSAs-

In the next phase we divert $2 trillion out of the Social Security system into personal accounts (PSAs). At this point, the unfunded liability of Social Security rises to $6 trillion because the original $4 trillion figure included the full income stream in computing this present value unfunded liability. Removing $2 trillion from the revenue stream thus increases the unfunded liability by this amount.

Also, at this stage, $2 trillion in assets appear in the private markets in the PSA accounts, and hence, this $2 trillion is no longer available to the government in balancing its books. So at this point here is how matters stand:

Step 2: Diversion to PSAs:

S. S. unfunded liability           =          ($6t)
Other Gov. Debt                     =          $0       
PSAs                                       =          +$2t
Total Gov. Debt                      =          ($6t)

Notice that at this stage $2 trillion in new debt has implicitly been created. The government’s indebtedness now stands at $6 trillion, $2 trillion more than in Step 1.

 

Step 3: Replace Diverted Funds-

At the next stage the government must replace the $2 trillion diverted to PSAs in order to continue meeting the payments to the Transition Beneficiaries. Hence, the Treasury floats $2 trillion in bonds to replace the diverted funds. This lowers the Social Security unfunded liability back to its original $4 trillion, but adds $2 trillion in Treasury debt. So the situation looks like this at this stage:

Step 3: Replace Diverted Funds:

S. S. unfunded liability           =          ($4t)
Other Gov. Debt                     =          ($2t)   
PSAs                                       =          +$2t
Total Gov. Debt                      =          ($6t)

Notice that we are still carrying $2 trillion in additional debt, we have merely shifted its location on the government’s books from the Social Security account to the general Treasury account.

 

Step 4: Repayment of Federal Borrowing-

In the final stage, Plan participants must repay the money they diverted from the Social Security system. This is accomplished by cutting their lifetime Social Security benefits by an amount that is the actuarial equivalent of the amount they diverted. Depending on where we allocate this repayment to, it either has the effect of paying back the Treasury for its new borrowing, while leaving the Social Security system exactly where it started—with a $4 trillion unfunded liability—or we can allocate the payback to the Social Security system, in which case its unfunded liability is reduced to $2 trillion, but the Treasury debt is not liquidated. So, at this final step, here is where matters end up:

Step 4: Payback:

S. S. unfunded liability           =          ($4t)
Other Gov. Debt                     =          0         
PSAs                                       =          +$2t
Total Gov. Debt                      =          ($4t)

Or,

Step 4: Payback:

S. S. unfunded liability           =          ($2t)
Other Gov. Debt                     =          ($2t)   
PSAs                                       =          +$2t
Total Gov. Debt                      =          ($4t)

 

So, at the end of this process, we end up right where we started, with $4 trillion in government debt. This is perhaps the basis for the intuition that the Transition Costs are not new debt, because we do not have an increased amount of debt in the government system at the end of the process. But notice that we have likewise not reduced the outstanding government debt either. This, I think, has huge implications for the validity of this argument, which clearly seems to imply some form of reduction in the government’s unfunded liabilities. (I will explain what I think the implications of this fact are in the Conclusions section of this paper.)

So we still have the same amount of unfunded governmental liabilities that we had before we went through this exercise. This entire sequence of financial maneuvers is merely a complicated, round-about, way of ending up precisely where we started, in terms of Social Security’s underlying insolvency. We had $4 trillion in unfunded liabilities before we went through this process, and we still have $4 trillion in unfunded liabilities at the end. But one thing has changed very profoundly: Social Security benefits have been reduced by $2 trillion. This is where the money really comes from to create the PSAs. It does not come from shifting unfunded liabilities, and it does not come from creating new debt, it comes from cutting benefits. This is the source of the new money in the system (the $2 trillion in assets in the PSAs). So there is no economic free lunch after all. The PSAs were not created without any net new costs. There are precisely $2 trillion in new costs, and these new costs will be borne by future Social Security beneficiaries in the form of benefit cuts.

Actually, the most accurate description of the dynamics of the Bush Plan is to say that the federal government is loaning money to workers to establish PSA accounts, and that these workers will pay the government back by accepting reduced Social Security benefits. It actually has little or nothing to do with diverting a portion of the worker’s Social Security payroll tax into a PSA, except in a notional sense. To see why, notice that we could accomplish exactly the same financial result if we just transferred $2 trillion out of the Treasury’s general fund to establish the PSAs, then required workers to pay this loan back through reduced Social Security benefits. In other words, the whole transaction could take place without the Social Security payroll tax ever being mentioned, and the outcome would be identical.

So, have we created new debt to fund the transition to PSAs? Perhaps not in the long run, but what about the short-run? After all, as Keynes used to say, “In the long-run we will all be dead.”

 

V- Timing is Not Quite Everything, But it Surely Matters-

Now let’s assume, for the sake of argument, that in the long-run no net debt is created. Does this mean there is no cost to the Bush Plan? No, because even granting this premise, we still have to factor in the element of time into our analysis—and not the time value of money in the sense of present value calculations, but time in the sense of which cohorts of taxpayers bear what burdens. So let’s look at this sequence of events more precisely, putting the time-dimension into consideration.

The Transition Period has a real and time-based duration. It starts when the personal accounts start and it ends when the last Transition Beneficiary, who is presently age 55, dies. For simplicity sake, let us assume that the Bush Plan were to start this year and that all those persons now age 55 have only 35 more years to live. So the Transition Period would be the years 2005-2040. This is the period during which we will actually have to pay actual dollars to actual Transition Beneficiaries. (If the two periods overlap slightly that does not alter the logic of the example.)

Now, what will really happen under the Bush Plan is that we will issue new federal debt starting in 2005, and we will carry this debt on the books until Plan participants start retiring and start paying this money back in the form of reduced Social Security benefits. This payback cannot start until 2018 (when those workers now 54 retire at age 67).[19] So payback starts in 2018 and occurs gradually up through the lifetime of those persons who opt into the personal accounts system during the Transition Period. This repayment period will extend until the youngest worker who opts into the system during the Transition Period has retired and repayed his diversion.

To make this clear, consider a Plan participant, J. Wilbur Worker, who is age 31 and who opts into the Bush Plan in 2005. From 2005 through 2040 J. Wilbur will be diverting money out of the Social Security system that is needed to pay for the benefits of the Transition Beneficiaries.[20] Wilbur has to pay this money back starting when he retires, which will be in 2041. So we have to carry this federal debt on the books from 2005 until 2041 when J. Wilbur starts to pay it back. (Some repayment starts earlier due to the retirement of older workers, and some starts later due to the retirement of younger workers, but this is a good illustration of the principle involved.)  

Even if we accept that in the long-run the transition costs are offset by a reduction in the unfunded liability or other government debt, this offset does not happen at the same period in time as the accrual of the debt. So the debt will be carried for several decades before being fully repaid. In the long-run, this may balance out; but in the near-term there will be actual real costs to be paid by actual real taxpayers. New Treasury bonds will be issued starting in 2005 and these bonds will pay interest, and the principal and interest on these bonds will be repaid by actual taxpayers with actual dollars.

So what Hunter et. al. are arguing is that, in the long-run, bonds issued by the Treasury in 2005-2040 are the same bonds that would have been issued in 2017-2080, so it is not new debt.[21] But the timing matters here. Debt carried on the books during the years 2005-2040 will have to be serviced (and perhaps repaid) by the taxpayers of 2005-2040, whereas debt that would have been issued gradually during the period 2017-2080 would have been serviced and repaid by future taxpayers.

This almost certainly means that the initial generation of participants will bear an extra burden. They will have to pay both for their own personal account and for the cost of sustaining the Transition Beneficiaries. There is an actual dual-burden during the Transition Period, even if, in the long-run, this burden is cancelled out by a reduced burden decades from now on other taxpayers.

 

VI- Conclusions-

I think I have raised sufficient doubts about the claw-back provision of the Bush Plan to begin to worry about how it could possibly be the payback mechanism for the transition to PSAs. So, if the claw-back provision has no real connection to repayment of the transition costs, how did we manage to pay them off? The answer, it seems clear to me, is that perhaps we did not.

Recall that, at the end of the day, we still have $4 trillion in unfunded liabilities in the Social Security system, and we have $2 trillion in new assets in PSAs. And recall, again, where that $2 trillion in new money came from: it came from reducing future Social Security benefits, i.e., from the claw-back mechanism (via an intermediate loan from the Treasury).

So we took $2 trillion out of Social Security and put it in PSAs; and we reduced Social Security payouts $2 trillion in consequence. This means that, in net terms, nothing happened except that the size and importance of the Social Security program has been diminished and the size and importance of equities has increased. A neat trick perhaps, but we are still in a $4 trillion hole in Social Security. How are we going to discharge this debt? The answer, I think, is by floating substantial amounts of new government debt in the future, or by “clawing back” additional amounts in Social Security benefits.

So, I think perhaps the best way to understand our circumstance at the end of the day under the Bush Plan is to say that we haven’t yet paid for the transition costs, and we won’t really until we discharge the $4 trillion in unfunded liabilities in the system. That $4 trillion in unfunded liabilities was the figure we computed before the Bush Plan was put into effect, and it included paying the Social Security benefits of the Transition Generation. So if we are still in debt by precisely the same amount, then its seems we have not paid the Social Security benefits of the Transition Generation, we have merely once again deferred to some future date the time at which we will finally liquidate our debt obligations. So in a quite straightforward sense, it seems we have not actually repaid the transition costs at all.

Or to put matters another way, the creation of PSAs seems to have had nothing whatever to do with the Social Security system! This is what I was hinting at before: that the idea of a diversion of Social Security payroll taxes into PSAs was at best a notional event. The reality is that the government created PSAs with general revenues via new borrowing, then rationalized ongoing Social Security benefit reductions by a notional allocation to Plan participants of appropriate amounts into their PSAs. Having done so, we are then in a position to demand that Plan participants must accept deep reductions in their Social Security benefits to “pay back” the money they “diverted” out of the Social Security revenue stream.  

But in a real sense nothing happened to the financing of the Social Security system; it is stuck exactly where it was before PSAs were conceived. It is almost as if the diversion of the $2 trillion out of the Social Security revenue stream never actually took place at all. This idea appears to be a pure fiction, designed almost as a ruse to justify deep Social Security benefit cuts. Or to say it another way: the whole purpose of this financial scheme may only have been to shift vast amounts of wealth out of the public into the private sector, thereby dramatically reducing the importance of the Social Security program in the economic life of the nation.

After all is said and done, the government is still as deeply in debt as it ever was and, somehow, future taxpayers still face a day of reckoning.

---

NOTES

[21] The year 2017 is the year that Social Security passes into a negative cash flow position and hence the government must begin redeeming debt in order to continue financing the program.