Share/Bookmark

Discuss: Share in Our Policy Ideas

Tuesday, May 20, 2014

A Response to Larry Kotlikoff

Larry Thompson, Founding Board Member, National Academy of Social Insurance

Larry Kotlikoff advocates replacing the current defined-benefit [Fixing Social Security, May 20, 2014], mostly-pay-as-you-go Social Security system with a system of funded individual accounts financed by mandatory contributions. This general approach was first adopted by Chile almost 35 years ago, and soon became popular among that set of economists most enamored with the universal superiority of markets as a mechanism for allocating resources. In the latter half of the 1990s, the approach was promoted internationally by the World Bank.

Kotlikoff’s particular wrinkle is to urge that all of the assets accumulated in these accounts be invested in a giant, global, indexed fund, managed by the government. That particular feature distinguishes Kotlikoff from other supporters of funded individual accounts; the others generally have not trusted governments to manage the funds and, instead, favored forcing workers to choose among private pension managers. Kotlikoff’s plan now also features an explicit strategy for converting individual account balances into lifetime annuities at the time of retirement, a feature the World Bank neglected to address in its initial enthusiasm for the individual account approach.

Although much can be said about the desirability (or not) of particular details of Kotlikoff’s plan, I believe that the two most important questions one should ask are: (1) would we want to go there? and (2) if we did, is there any chance we could get there from here? My answer to both questions is no.

Our current Social Security program offers workers a benefit defined in terms of wage levels prevailing at the time of their retirement. Individual benefit amounts are based on the worker’s pre-retirement average earnings adjusted by various socially motivated provisions. (Kotlikoff disapproves of some of these social adjustments.) In contrast, a system of funded accounts offers workers a retirement income determined entirely by the pattern of capital market returns over the individual’s working life. (Kotlikoff would also adjust individual retirement incomes upward or downward by taxing everyone to finance those socially motivated additions that he does approve of.) The major difference between the two approaches is that one sets pensions with reference to wage levels prevailing at retirement whereas the other establishes retirement incomes based on capital market returns over the course of the working life.

To cut to the chase: Kotlikoff wants retirement incomes determined exclusively by capital markets while I favor having them determined partially by capital markets and partially by wage developments (labor markets). The U.S. retirement system is currently a mixture of the wage-based Social Security program and capital-based individual retirement accounts financed by workers alone or by workers and employers jointly. By converting the former into a system resembling the latter, Kotlikoff would force everybody to put all of their retirement income eggs into the capital market basket.  One could debate whether, at the margin, the relative weights of the wage-based and capital-based approaches should be altered, but I can’t believe that forcing everybody to rely totally on the capital-based approach to the exclusion of a wage-related approach increases social welfare. Even the Chileans have recently added a defined-benefit element to their retirement system, having concluded that a mixture of defined benefit and defined contribution was superior to exclusive reliance on individual defined-contribution accounts.

Even among those who prefer funded individual accounts to our current Social Security approach, most realize that the immediate public-sector costs of a transition make such a change highly unlikely, even if the adjustment is rather modest and is phased in over a long period of time. Indeed, a significant portion of the international reforms sponsored by the World Bank just before and after the turn of this century have since unraveled, largely because the countries found the transition costs to be larger than they cared to bear. Given our Congress’s current attitudes about fiscal policies, I doubt they would find such costs any more acceptable.

Two of the biggest problems with our current retirement income system are that workers (and/or their employers) are not contributing enough to their defined- contribution accounts and that investment returns in these accounts under-perform broad market averages due to high administrative charges and poor individual investment decisions. Kotlikoff might be of more use to future retirees by concentrating his attention on reforming the current defined-contribution accounts rather than trying to destroy the current wage-based pension program.


Larry Thompson is a founding Board member of the Academy and has been a consultant on Social Security and the administration of pension systems for the World Bank, the Asian Development Bank, and the International Labor Office. He has served as Assistant Comptroller General of the United States and Acting Commissioner of the U.S. Social Security Administration.

Posted on May 20, 2014  |  Add your comment
Keywords:

All Comments

Larry, how right you are! It is hard to believe that this conversation is continuing after 2008 when so many lost so much in their IRAs and 401(k) plans. Both social security and financed plans have risks, and it seems so logical to weight those risks with a system more like what we have today. As a footnote, it seems that too many countries like Argentina and Hungary are starting to step back from their funded pension promises -- which the developers of these systems said would not happen -- in contrast to social insurance systems.
Dear Larry and Emily,

Thanks for your comments. But please reread my postcard-length proposal. There is a government guarantee -- a zero real return -- in the proposal. So your statement, Larry, that I'm leaving retirement income entirely up to capital markets is simply not true. Second, the proposal calls for the government to match contributions on a progressive basis. It also calls for contribution sharing. Wall Street plays absolutely no role in the system and makes not a single dime. The index fund -- of stocks, bonds, and REITS, is entirely market weighted and can be run by a single Apple laptop suspended by virtually invisible wires in the middle of an entire floor of HHS (with everything and everyone else cleared out). Annuitization of each cohort would also be done by this same laptop. The proposal is only about reforming the retirement portion of Social Security. It doesn't propose changes in DI or Survivor benefits. Both of those systems need to be fundamentally fixed, but I'm leaving that for another day.

The fact that Kazakhstan and Argentina and some other countries have made a total hash of pension reform is not a fair criticism of my proposal since what they did is not what I'm proposing. Indeed, I've been among the strongest and most outspoken critics of what they did and what they've done of late.

Other countries, like Singapore, have run successful individual account systems.

But the real issue is not funded versus unfunded. It's not even relying on wages versus capital markets. I'm relying on lifetime financial returns with a very gradual -- 10 year -- period of selling off assets and converting them, on a cohort-by-cohort basis into inflation-index annuities. But the Purple Social Security Plan provides a guarantee that would be backed by taxes on wages if the guarantee were called.

The real issue is generational equity. The real issue is intragenerational equity. The real issue is having a system that people can understand and rely on. The real issue is that our Social Security System, in its design, in its opacity, in its abject insolvency is, in fact, a terrible disgrace to social insurance, where I'm referencing both the insurance and equity aspects of that term.

Finally, let me respond to the proposition that I'm "trying to destroy the current wage-based pension program." The current wage-based pension program has destroyed itself. Table IVB6's infinite horizon fiscal gap in the Trustees Report makes this crystal clear. Over 1,000 economics, including 16 Nobel Laureates, have endorsed infinite horizon fiscal gap accounting at www.theinformact.org for a critical reason, namely any finite horizon fiscal gap accounting is entirely a function of our fiscal nomenclature. Only the infinite horizon fiscal gap measure is independent of how we label government receipts and payments. Stated differently, only the infinite horizon fiscal gap is a well defined economic concept. The 75-year fiscal gap is, in contrast, an entirely linguistic construct. There are an infinite number of internally consistent ways we can label past and future SS receipts and payments and each will produce an entirely different 75-year fiscal gap. That's why the entire economics profession (well close to the entire profession) is saying publicly to do infinite horizon accounting.

The infinite horizon fiscal gap facing Social Security is 32 percent of the present value of the system's taxes. Neither of you is going to make this problem go away by defending the status quo.
Kotlikoff is completely wrong.

Most importantly, having such a large buyer of financial assets will push these into bubbles that will explode so forcefully that 1929 and 2008 will be 'bubblets".
Larry Kotlikoff, in his rejoinder to Larry Thompson and Emily Andrews, closed with an odd jeremiad on the seemingly esoteric issue of so-called “infinite horizon” actuarial estimates. Actually, this was doubly odd in that neither Thompson nor Andrews mentioned the concept in their comments. But Kotlikoff is right to think this issue important, and right to be defensive about it, but for all the wrong reasons, as I will explain in a moment. But first, a little history.

Kotlikoff tells us that 1,000 economists (including a suitably impressive 16 Nobel Laureates) think infinite horizon estimates are the way to go. Well, as Pat Moynihan liked to say, there are some mistakes you need a PhD to make. But all these learned economists are right about one thing: there is nothing sacred about the 75-year estimation period for Social Security. In fact, when Bob Myers produced the first long-range actuarial estimate in 1935 it ranged for only a 40 year period. This choice of estimating period was rationalized by the idea that 40 years would allow time enough for the first generation of covered workers to complete their paying-in years and reach retirement age. It was not chosen arbitrarily. The estimating period then ranged widely for the next three decades, dropping as low as 35 years and reaching as high as 80 years. The 75-year estimation period has been in use since 1965. This estimating period—also developed by Bob Myers—was rationalized by extending the timeframe to include both the wage-earning and benefit-receiving period of a typical participant generation. In other words, the 75-year period also was not chosen arbitrarily.

So the 75-year estimation period is to some degree an arbitrary choice. This is all Kotlikoff is saying when he complains—in the best economist’s jargon—that anything less than an infinite horizon “is entirely a function of our fiscal nomenclature.” In other words, it is a policy choice. But we are not trying to write an economics textbook here, we are trying to make sensible public policy. And even though the 75-year estimating period is not part of some economic canon, it is a precedent, and as a precedent it has enormous value and importance. Our policy choices need to be framed in the traditional rubric of the standard 75-year estimation period, so that the choice between various options is not distorted by the use of different “yardsticks” to measure the size of our problems.

“Infinite horizon” estimates appear for the first time in the Annual Trustees Reports starting only in 2003, during the Bush administration’s push to privatize the retirement program. In that effort administration officials typically used infinite horizon figures rather than the traditional 75-year estimates in making their case, for one simple reason—to make the program’s problems look as big as they possibly could so as to better make the case that the program needed radical reforms. These infinite horizon projections were not added to the Annual Reports to make them more robust or complete. Indeed, the professional actuaries at SSA believe them to be meaningless, but they are obliged to use them so that scholars like Kotlikoff can make Social Security look as bad as he possibly can.

So what does it mean anyway to construct an “infinite horizon” estimate of Social Security’s future liabilities? Consider first what it means to construct a traditional 75-year estimate. The actuaries at SSA must predict, with precision, the likely values of dozens of economic and demographic factors each year for the next 75 years. What audacity! Consider what Bob Myers could not foresee in even his 40-year projections back in 1935: WWII; the Baby Boom; the post-war economic growth; women entering the workforce in unprecedented numbers; the stagflation of the 1970s, etc., etc. Indeed, events even in the near term are certain to make a pig’s breakfast of even the best actuarial estimates. Thankfully, the actuaries revise their estimates every year to keep them tracking as close as humanly possible to our actual experience.

So how do the actuaries construct an “infinite” estimate? Stated abstractly the idea of this mathematical cum economic construct seems mysterious, almost magical, in a way that prevents the non-expert from grasping its real meaning. We tend to assume it is sophisticated in some way beyond our understanding, and being infinite, it must be better than a “mere” 75-year projection. But is it? Actually, there is no such thing as an “infinite” estimate of Social Security’s future income and outgo. The way the “infinite” measure is constructed (at least as of the last time I discussed this with the long-range actuary) is that the actuaries take the traditional 75 year estimate as a starting point, they then carry the estimates out to about year 100, then they extend the values from year 100 out at the rates then prevailing for an additional 899 years. In other words, the “infinite” estimates are actually 999 year estimates! That is to say, in the “infinite horizon” estimates the actuaries at SSA are attempting to predict the future each year for the next 999 years! This is not audacity, it is stupidity (actually, a mix of stupidity and cupidity). Why would anyone be adamant that we should use such dubious projections to frame our Social Security policy choices? Why indeed? The only reason I can think of is to make the problems facing Social Security look as large as possible so as to scare us into embracing one’s preferred “solution.”

So if Larry Kotlikoff wants us to take his “purple” plan any more seriously than the ones he has given us before, he should start with a little truth in advertising. Let’s agree that the long-range (75-year) unfunded liability of the Social Security program requires a 21% increase in the tax rate (rather than the 32% Kotlikoff cites) or a reduction in benefits of 17% (not the 23% he cites), or some combination of these two factors. In dollar terms, the unfunded liabilities of the Social Security program are $9.6 trillion over the next 75 years, not the $23.1 trillion Kotlikoff wants us to use (heaven knows, $9.6 trillion is big enough). And let us further agree that these estimates are ridiculous in their precision. We have no idea what the fiscal condition of the Social Security system will be decades from now. Using these remote projections to justify radical changes in the present (as President Bush clearly tried to do) is little short of policy malpractice. The best we can do is to use the traditional 75-year estimates, update them annually, and make the best policies we can in their light. But puffing up the deficit bogeyman with estimates reaching out 999 years is an absurdity. If we can dispense with this absurdity once and for all, then maybe we can work on some ideas for addressing this deficit, without the propaganda overlay.

Larry DeWitt
former Public Historian, Social Security Administration
Dear Larry (Yet another Larry!),

You are missing the economics here.

Take a look at http://www.kotlikoff.net/content/general-relativity-fiscal-language or read
http://www.kotlikoff.net/content/generational-policy to understand the deep labeling problem you are entirely missing.

There is nothing in economic theory that says whether to call a dollar collected from you by the government a "tax," a "borrowing," or "a 50 cent tax and a 50 cent borrowing," or a "$5 tax, plus a $4 loan back to you," ...

And each choice of internally consistent labels will produce a different and equally vacuous 40-year or 75-year fiscal gap measures. Only the infinite horizon fiscal gap will produce the same measure independent of the labeling convention.

The fact that Bob Myers said we should measure the system's solvency using a finite horizon fiscal gap calculation based on his particular choice of words doesn't make it a sound economic measure.

The system is broke, it's a user's nightmare, it's basic formula for the benefits of a spouse age a has 10 mathematical functions on the right-hand-side. One of these functions is in four dimensions. Some of the functions are mins and maxes and one is a min that's a function of a max. There are also extremely complex functions subscribing the arguments of the 10 functions.

This is the true legacy of Bob Myers, Bob Ball, and other Social Security social engineers -- a saving, disability, and life insurance system that no one can remotely understand, that is driving people nuts, that is horribly unfair along a host of dimensions, and that, to repeat, is 32 percent underfinanced!

Yours, Larry
One further point, Larry. Discounting means making less of. If you discount a dollar received or paid 999 years from now at 3 percent, it has a present value of 1.5 E-13. So economic theory, which says to discount future flows in forming the fiscal gap, understands that a dollar in the future is not the same as a dollar today. It also understands that things that are uncertain need to be discounted with proper risk adjustments. The economists who endorsed The Inform Act at www.theinformact.org include Ken Arrow, who is perhaps the greatest single contributor to the economics of uncertainty and financial valuation. He's at the top of the list of Nobel Laureates who endorsed the bill. This doesn't mean the valuation problem is easy. It's not. But we don't gain anything by driving in NY with a map of LA, especially when we're heading right for the East River.
Larry,
I completely agree with you. We can't put all the savings into the equity market and zero percent of return guaranteed by the government is NOT acceptable given a long saving period of more than 30 years for most of the young generation!

I think the credits required to get retirement benefits need to be raised since people are living longer than generations ago. Alternatively, those workers who barely meet the minimum credit requirement should get a benefit cut. The system should encourage people to work more not to rely on social benefit more.
Larry,
I completely agree with you. We can't put all the savings into the equity market and zero percent of return guaranteed by the government is NOT acceptable given a long saving period of more than 30 years for most of the young generation!

I think the credits required to get retirement benefits need to be raised since people are living longer than generations ago. Alternatively, those workers who barely meet the minimum credit requirement should get a benefit cut. The system should encourage people to work more not to rely on social benefit more.
To all my Larry and other fans who feel I'm being too harsh on the Social Security System as currently constituted,

Let me convey a bit of where my outrage is coming from. I have a small software company (see www.maximizemysocialsecurity.com and www.esplanner.com). Our SS software, which we market for $40, gives people suggestions on how to maximize their lifetime benefits.

I'm not in this for the money. The company barely breaks even. I've run it for 20 years and netted negative $60,000. I run the company to help people and employ great American software engineers.

Because I'm doing this to help people, I'm incredibly anal about getting all of Social Security's details exactly right under the hood. The engineers I've hired are equally anal. We've had four people spend three years -- some full time, some part time -- getting everything we could discover completely straight.

Every day I get emails from people who tell me how they made one huge mistake or another because they didn't understand Social Security's rules. I also get 50 or so SS questions sent to me by people in response to my weekly PBS.org/NEWSHOUR column that answers SS questions.

We have a former Technical Expert on our staff who helps us made sure we understand the system's provisions, and we've received lots of help from top actuaries at Social Security. So when I rail about the system's opacity and inequities, it's from a deep basis of knowledge of the system. It's not because I'm a right-wing nut case who doesn't deeply respect the need for forced government saving and insurance or doesn't understand the clear adverse selection reasons that private insurance markets don't work. Indeed, you won't find a stronger advocate for social insurance than me.

But take a read of this description below of an example of an application of a grace year and see if you think this is something John Q. Public should be able to understand. This is just one of 2,728 provisions in SSA's Handbook! And if you really want to get sick to your stomach, look at https://secure.ssa.gov/apps10/poms.nsf/subchapterlist!openview&restricttocategory=03014. This is a bureaucrat's (fill in the blank), pure and simple.

"Marion was entitled to mother's insurance benefits from 1998 because she had a child in her care. Because she had a non-service month in 1998, 1998 was her initial grace year. Marion's child turned 16 in May 2000, and the child's benefits terminated in April 2000. Marion's entitlement to mother's benefits also terminated in April 2000. Since Marion's entitlement did not terminate by reason of her death and she was not entitled to another type of Social Security benefit in the month after her entitlement to a mother's benefit ended, she is entitled to a termination grace year for 2000, the year in which her entitlement to mother's insurance benefits terminated. She applied for and became entitled to widow's insurance benefits effective February 2001. Because there was a break in entitlement to benefits of at least one month before entitlement to another type of benefit, 2001 will be a subsequent grace year if Marion has a non-service month in 2001."
Before I attempt to address the specifics of Kotlikoff’s postcard-size policy proposal (his “purple plan”), I want to offer a few observations at a higher level.

In my first post, I criticized his description of the problem. The essential points I made there are, in my view, valid (notwithstanding Professor Kotlikoff’s learned papers to the contrary). Just to close that circle, I will say this about his analysis. As best I can follow it, what Kotlikoff does is to construct a theoretical economic model to account for the fiscal dynamics of a system like Social Security across generations. In his model he is able to use as variables factors which can be stated abstractly such that they make no reference to traditional policy constructs like “debt,” “deficit,” “taxes,” etc. He shows (and here I take his word for it) that there are an infinite number of such potential alternative variables (i.e., names of economic factors). He concludes that this means that our choice of variables is arbitrary and hence any particular choice (like “taxes” or “deficits”) is meaningless—in terms of his economic model. He further shows that if one uses an “infinite horizon” type mathematical construct, that this forces the model into a single solution. This, he argues, means that this construct is non-arbitrary, and hence meaningful.

Three rather simple points. One: the idea that because a variable can be represented by different names it is “meaningless,” is highly debatable. This takes us into a realm of philosophical analysis which I for one relish, but which I fear would bore more sensible people even more than trying to follow Kotlikoff’s equations “proving” his claim. Two: whatever economic theory may or may not establish by the use of sophisticated models, it meshes imperfectly (if at all) with actual public policy. To put it bluntly: even if Kotlikoff is right about the arbitrary nature of traditional budget concepts, who cares? Three: what we are after here is the making of sensible public policy, and we do not need to abandon traditional budget or policy concepts to achieve this aim. In fact, making radical changes in traditional measures of policy are generally how one confuses and obscures rather than clarifies the choices before us.

Now, before we get down in the weeds with the purple policy, I want to make a few observations about Kotlikoff’s vision of social insurance. He glibly tells us that he could replace the thousands of bureaucrats in the Department of Health and Human Services (HHS) with a single computer dangling from the ceiling of the (presumably) Humphrey Building in D.C. (Just for the record, the Social Security program has not been part of HHS for nearly 20 years now. Hint: if you are looking for a building to empty out, try Baltimore.) This single metaphor tells us a lot about Kotlikoff’s faulty understanding of social insurance, politics, and policymaking.

What Kotlikoff is on about here is the frustrating complexity of the Social Security program (although here too he misrepresents the magnitude of the problem by listing all the rules and regulations he finds in the Social Security handbook, including those covering the non-retirement programs). But the point is well taken. But Kotlikoff never seems to ask himself the question of why this complexity exists. All of these rules and regulations are expressions of intentional public policies, crafted over decades by policymakers (i.e., our elected representatives). The Social Security “retirement” program is not a simple individual investment account—which is what Kotlikoff wants it to be and into which his purple plan would transform it. As a matter of public policy, America has decided that we should pay benefits on the account of a retiree to: his aged spouse, or to his young spouse if there are minor children in the household; to one or more divorced spouses under certain conditions; to the disabled adult child of the retiree; to the retiree’s aged parents; to certain retirees who were age 72 at a point in time, etc., etc.

Now Kotlikoff dislikes all these complications and wants to wish them away. He prefers a world in which the worker himself/herself is paid a retirement benefit (shared with a spouse) and then let the worker decide who he wants to share his wealth with, and let the devil take the hindmost. This is certainly one point of view. But it is not how this nation has conceived of the role and functions of our social insurance system. America’s social insurance system has evolved over 7-plus decades to meet the unmet needs for economic security that arise in a modern wage-based political economy. Those needs are eternal and universal. Social Security is the best way that our society has ever devised to share the burden of meeting those needs. Social Security is, admittedly, a communitarian approach to economic security. Some prefer more individualistic approaches (clearly Kotlikoff is in this camp). The making of good public policy involves finding the right balance between these two approaches. Perhaps at this point in our history we need to shift the balance more toward the individualistic end of the spectrum. But what we most definitely do not need to do is to tear down Social Security “root and branch,” as Kotlikoff has stated his objective to be.

Also, think for a moment about the potential real-world impact of Kotlikoff’s economic model and what he thinks flows from it. Take just one category of beneficiary whom Kotlikoff would eviscerate. Consider the problem of divorced spouses. One societal dynamic that occurs with depressing frequency is that (typically) a high earning male wage earner divorces his older spouse in favor of a younger model, leaving the divorced spouse bereft of financial means. (Kotlikoff thinks his earnings sharing proposal addresses this, but it won’t, for reasons I will explain in my post on his plan.) Under the Social Security program as currently conceived we require that a long-married divorced spouse be granted a benefit on the wage earner’s account. If this were optional, many wage earners would refuse to make any such provision. Not only are divorced spouses potentially entitled to a benefit, but so too might be the children of the wage earner who are living with the divorced spouse. Indeed, one of the tasks of those thousands of bureaucrats that Kotlikoff wants to replace by a dangling PC is to probe reluctant applicants for other potentially eligible beneficiaries on their account—information which they sometimes try to avoid revealing, especially if it involves a broken family. Filing a claim for Social Security retirement benefits is not just the computation of a benefit amount (a task already handled by computers, thank you very much), it is also importantly about making a knowledgeable assessment of which benefits an individual may be entitled to and who else may be entitled on their earnings record. In other words, a social insurance program is more than a computation of a benefit amount. It is also, on another point, the processing of millions of post-entitlement actions (earnings reports, address changes, etc.). I am sorry, but in the real world of real human beings (messy creatures that we are) a PC is not an acceptable substitute for a well-informed government bureaucrat.

What Kotlikoff is attempting is to use Social Security’s long-range insolvency (as amplified by his inflated account of the problem) as a justification to wreck profound and radical changes in the nation’s social insurance system. There are clearly many in various places (the CATO Institute, the secret dreams of many Republicans, etc.) who share this agenda. But for those of us who do not, we should simply note that we do not have to abandon the Social Security system to address the program’s insolvency. It can be addressed within the rubric of the program’s traditional principles and practices (more about this in a subsequent post). What has prevented this from happening has not been the unwillingness of Social Security’s supporters to acknowledge its problems (both in long-range solvency and intergenerational equity), but rather our lack of progress has been almost entirely the result of the intransigence of those who do not support the program (i.e., those dreamy Republicans and the CATO-ites) to compromise by agreeing to tax increases as part of any comprehensive solution.

In Kotlikoff’s dreamy vision of the future, America’s need for economic security in retirement can be taken care of by a single computer dangling from the ceiling of an abandoned federal building (in Baltimore) and running his company’s simplified software for computing a ROI for each would-be retiree. Speaking for myself, I find this vision more Orwellian than enticing.
Hi Larry (DeWitt),

Thanks for your response to my response. Let me reply below to specific points you make with ** ** indicating what my responses since this editor doesn't allow us to use italics." best, Larry


Before I attempt to address the specifics of Kotlikoff’s postcard-size policy proposal (his “purple plan”), I want to offer a few observations at a higher level.

In my first post, I criticized his description of the problem. The essential points I made there are, in my view, valid (notwithstanding Professor Kotlikoff’s learned papers to the contrary). Just to close that circle, I will say this about his analysis. As best I can follow it, what Kotlikoff does is to construct a theoretical economic model to account for the fiscal dynamics of a system like Social Security across generations.

** The paper with Jerry Green (the senior theorist at Harvard) and me, called "On the General Relativity of Fiscal Language," is not about "my" model or "my and Jerry's" model, it's about all neoclassical models in economics, by which I mean any and every model that assumes agents are rational, by which I mean they aren't fooled by language and they don't construct institutions that are fooled by language. **

In his model he is able to use as variables factors which can be stated abstractly such that they make no reference to traditional policy constructs like “debt,” “deficit,” “taxes,” etc. He shows (and here I take his word for it) that there are an infinite number of such potential alternative variables (i.e., names of economic factors).

** Again, you say "his model," but it's not my model, it's a statement about economic theory in general. **

He concludes that this means that our choice of variables is arbitrary and hence any particular choice (like “taxes” or “deficits”) is meaningless—in terms of his economic model. He further shows that if one uses an “infinite horizon” type mathematical construct, that this forces the model into a single solution. This, he argues, means that this construct is non-arbitrary, and hence meaningful.

** No, Larry. It's not that "the" model is being forced into a single solution. It's that the fiscal policy being run in any and every model we write down in economics can be labeled in an infinitely different number of ways, all of which are internally consistent and all of which will produce different cash flows and all of which will produce different finite-horizon (e.g., 75-year) fiscal gaps, by all of which will produce an IDENTICAL infinite horizon fiscal gap, if one exists. To make this concrete, suppose Congress and the President were to pass a law that swapped everyone's claim to their accrued SS benefits for official government inflation-indexed annuity bonds. This would, overnight, add trillions to official federal debt with no substantive change of any kind in our actual policy. Chile, as you know well, did exactly this back in 1982. But whether or not a law was passed, we can describe our actual policy as entailing these bonds. I.e., you and I can use our own words -- we don't need to rely on the government's choice of words/labels -- to describe how much official debt is outstanding. But your choice of words and mine and the guy down the street can all differ, so we have a Tower of Babel of debt values, each of which, in truth, tells us nothing about our fiscal policy. The fiscal gap over the infinite horizon does tell us about the sustainability of our fiscal policy and is invariant to our choice of labels. You and I and the guy down the street will all use different words and say the cash flows over any finite horizon of the system are different, but we'll all agree on the size of the fiscal gap. **

Three rather simple points. One: the idea that because a variable can be represented by different names it is “meaningless,” is highly debatable. This takes us into a realm of philosophical analysis which I for one relish, but which I fear would bore more sensible people even more than trying to follow Kotlikoff’s equations “proving” his claim.

** Sensible people know that Enron cooked its books. What my paper and other articles show is when it comes to describing its cash flows and reporting its official debts, the government has infinite latitude in cooking its books. But it has zero latitude in reporting the infinite horizon fiscal gap. The reason is simple -- in fiscal gap accounting done over the infinite horizon every dollar the government is project to pay and receive gets put onto the books no matter how it's labeled. This is not hard for people to follow, nor is it boring to say that the cash flows being reported by Uncle Sam for the government as a whole and for SS on its own are fundamentally as arbitrary and fraudulent as Enron's statements of its profits.**


Two: whatever economic theory may or may not establish by the use of sophisticated models, it meshes imperfectly (if at all) with actual public policy. To put it bluntly: even if Kotlikoff is right about the arbitrary nature of traditional budget concepts, who cares? Three: what we are after here is the making of sensible public policy, and we do not need to abandon traditional budget or policy concepts to achieve this aim. In fact, making radical changes in traditional measures of policy are generally how one confuses and obscures rather than clarifies the choices before us.

** This is missing the point to the maximum extent possible. If the labels change the cash flows and the size of the 75-year fiscal gap, then which labels do we use? Economic theory says that the choice of labels is entirely arbitrary and so we need to measure something that is invariant to the choice of the labels -- that's the infinite horizon fiscal gap. Stated differently, traditional measures of policy and the sustainability of Social Security are as reliable as Enron accounting. This is not my opinion. This is what economic theory tells us. And this is why 1000 plus economists have endorsed The Inform Act. **

Now, before we get down in the weeds with the purple policy, I want to make a few observations about Kotlikoff’s vision of social insurance. He glibly tells us that he could replace the thousands of bureaucrats in the Department of Health and Human Services (HHS) with a single computer dangling from the ceiling of the (presumably) Humphrey Building in D.C. (Just for the record, the Social Security program has not been part of HHS for nearly 20 years now. Hint: if you are looking for a building to empty out, try Baltimore.) This single metaphor tells us a lot about Kotlikoff’s faulty understanding of social insurance, politics, and policymaking.

** I know SS has its main building in Baltimore and that the Office of Actuaries is there. I've been calling that office for decades using its area code. It thought you and others, knowing the size of the HHS building, would get the point better if I referenced it. Substitute the Pentagon if you'd prefer. **

What Kotlikoff is on about here is the frustrating complexity of the Social Security program (although here too he misrepresents the magnitude of the problem by listing all the rules and regulations he finds in the Social Security handbook, including those covering the non-retirement programs). But the point is well taken. But Kotlikoff never seems to ask himself the question of why this complexity exists. All of these rules and regulations are expressions of intentional public policies, crafted over decades by policymakers (i.e., our elected representatives). The Social Security “retirement” program is not a simple individual investment account—which is what Kotlikoff wants it to be and into which his purple plan would transform it. As a matter of public policy, America has decided that we should pay benefits on the account of a retiree to: his aged spouse, or to his young spouse if there are minor children in the household; to one or more divorced spouses under certain conditions; to the disabled adult child of the retiree; to the retiree’s aged parents; to certain retirees who were age 72 at a point in time, etc., etc.

Now Kotlikoff dislikes all these complications and wants to wish them away. He prefers a world in which the worker himself/herself is paid a retirement benefit (shared with a spouse) and then let the worker decide who he wants to share his wealth with, and let the devil take the hindmost. This is certainly one point of view. But it is not how this nation has conceived of the role and functions of our social insurance system. America’s social insurance system has evolved over 7-plus decades to meet the unmet needs for economic security that arise in a modern wage-based political economy. Those needs are eternal and universal. Social Security is the best way that our society has ever devised to share the burden of meeting those needs. Social Security is, admittedly, a communitarian approach to economic security. Some prefer more individualistic approaches (clearly Kotlikoff is in this camp). The making of good public policy involves finding the right balance between these two approaches. Perhaps at this point in our history we need to shift the balance more toward the individualistic end of the spectrum. But what we most definitely do not need to do is to tear down Social Security “root and branch,” as Kotlikoff has stated his objective to be.

** Larry. I think if you were to build SS from scratch to meet these social insurance needs, on which we both agree, you would not resurrect the monstrosity that is now in place, with arcane rules that were set up by old white guys decades ago for reasons that may or may not have had anything to do with your or my senses of social justice or social insurance. **

Also, think for a moment about the potential real-world impact of Kotlikoff’s economic model and what he thinks flows from it. Take just one category of beneficiary whom Kotlikoff would eviscerate. Consider the problem of divorced spouses. One societal dynamic that occurs with depressing frequency is that (typically) a high earning male wage earner divorces his older spouse in favor of a younger model, leaving the divorced spouse bereft of financial means. (Kotlikoff thinks his earnings sharing proposal addresses this, but it won’t, for reasons I will explain in my post on his plan.)

** I look forward to this post. **

Under the Social Security program as currently conceived we require that a long-married divorced spouse be granted a benefit on the wage earner’s account. If this were optional, many wage earners would refuse to make any such provision. Not only are divorced spouses potentially entitled to a benefit, but so too might be the children of the wage earner who are living with the divorced spouse. Indeed, one of the tasks of those thousands of bureaucrats that Kotlikoff wants to replace by a dangling PC is to probe reluctant applicants for other potentially eligible beneficiaries on their account—information which they sometimes try to avoid revealing, especially if it involves a broken family. Filing a claim for Social Security retirement benefits is not just the computation of a benefit amount (a task already handled by computers, thank you very much), it is also importantly about making a knowledgeable assessment of which benefits an individual may be entitled to and who else may be entitled on their earnings record. In other words, a social insurance program is more than a computation of a benefit amount. It is also, on another point, the processing of millions of post-entitlement actions (earnings reports, address changes, etc.). I am sorry, but in the real world of real human beings (messy creatures that we are) a PC is not an acceptable substitute for a well-informed government bureaucrat.

** Here we could not differ more, but I'll leave this for another post. **

What Kotlikoff is attempting is to use Social Security’s long-range insolvency (as amplified by his inflated account of the problem) as a justification to wreck profound and radical changes in the nation’s social insurance system. There are clearly many in various places (the CATO Institute, the secret dreams of many Republicans, etc.) who share this agenda. But for those of us who do not, we should simply note that we do not have to abandon the Social Security system to address the program’s insolvency. It can be addressed within the rubric of the program’s traditional principles and practices (more about this in a subsequent post). What has prevented this from happening has not been the unwillingness of Social Security’s supporters to acknowledge its problems (both in long-range solvency and intergenerational equity), but rather our lack of progress has been almost entirely the result of the intransigence of those who do not support the program (i.e., those dreamy Republicans and the CATO-ites) to compromise by agreeing to tax increases as part of any comprehensive solution.

** Now you are trying to label me as a Republican, a member of CATO, and a right-wing nut bag. Not the case. I voted for the President twice. I've voted for Democratic presidents my entire life. I worked as a junior economist for President Reagan, but I didn't vote for him. I thought he and the country needed some help when it came to fiscal matters. I was nearly fired by insisting with Bill Niskanen and others at the CEA on writing the truth about the 1982 tax act, which included, for example, taxing AFDC recipients (in terms of loss of benefits) 1 dollar for every dollar they earned. "

In Kotlikoff’s dreamy vision of the future, America’s need for economic security in retirement can be taken care of by a single computer dangling from the ceiling of an abandoned federal building (in Baltimore) and running his company’s simplified software for computing a ROI for each would-be retiree. Speaking for myself, I find this vision more Orwellian than enticing.

** I suggest you meet with a hundred people a day who are enraged by the operations of the current system and the power bureaucrats have over their lives. The current system is truly Orwellian. **
TYING UP LOOSE ENDS--

Larry Kotlikoff has been very generous with his time and effort in responding at length to my lengthy posts on this discussion, and I appreciate that. Also, I do want to make clear that I have a much higher opinion of him and his work than I do of the ideologically bent “studies” that come out of the CATO Institute and, a fortiori, of the average Republican member of Congress. My “guilt by association” point was only that Kotlikoff may well end up taking us to much the same place that CATO and the like are trying to take us, to wit, a society in which the government has a vanishingly small role to play in the provision of economic security for our citizens. I think this would be a bad place to end up and I fear that, despite being a fine fellow, Larry Kotlikoff’s plans for Social Security may well take us there.

That being said, perhaps one additional admission is in order. Given the general level of my innumeracy, it is a virtual (mathematical?) certainty that I do not understand the econometric studies to which Kotlikoff has directed our attention. I worked my way through them as best I can, and my feeble attempt to state their essential point seems to have missed the mark again. I was trying to state his point in simple lay terms, without the economic gloss. Apparently I failed, (although I must say Kotlikoff’s efforts to do this here in this discussion forum have not been of much help either).


BUDGET GAMES--

It now seems to me that Kotlikoff’s central thesis is much broader than a point about the long-range projections of the Social Security system’s finances. He seems to be broadly skeptical of the general idea of governmental debt and the manner in which the government does its budget accounting. I think I agree with much of his critique here. It is scandalous how diverse federal budgeting practices are used to manipulate our national fiscal picture. Such things as whether future spending is shown from a “current law” baseline or a “realistic” baseline; whether budget projections are five years out or ten years out; whether we use dynamic or static economic assumptions, etc., are well-known ways that policymakers game the federal budget to yield whatever ideologically driven agenda they might have. If that were all Kotlikoff is saying here, then we would have a small divide between us on this score. And I might even be willing to agree that for some purposes an “infinite horizon” projection might be useful (I am still not sure on this point). But I am still not convinced that this is all he is saying here. Let me try it this way. Let me explain in simple lay terms how I see the issue of Social Security’s financing. Then Larry K. can perhaps counter me in simple lay terms, so that all of us can understand what all the fuss is about.

The federal government has a broad taxing power. As long as the Republic stands, the government can raise revenue in the future to fund its expenditures. One of these expenditures is the Social Security program. Because the government cannot fail in the way a private firm might fail, we can use “open group” actuarial estimates to predict what future tax rates (and other policies) might have to be to meet these future expenditures. That is to say, we can assume that the Social Security system will continue to have a future source of funding (payroll taxes) into the indefinite future. We need some sense of our future obligations for purposes of setting tax rates and the like. As far as I can see, this has nothing to do with “labels.” It does not matter whether we call the federal taxing power the ability to raise “taxes” or whether we describe it as the ability to generate the “X factor.” Money is money, taxes are taxes, whatever we label them.

Now, I think I can concede that what the precise value is of our future indebtedness will be estimated to be will depend, in some degree, upon the period of time over which we project those future expenditures and the methodology we use in our estimates. But it does not really matter what the precise value of those future obligations will be at any future point in time, so long as we have a rough enough idea to allow us to make sensible policy decisions as to tax rates, benefit amounts, and the like. To put it in blunt terms, for Social Security planning purposes, it does not matter whether the system’s unfunded liabilities are $9.6 trillion or $23.1 trillion. We can easily recognize that the system is insolvent in the long-term and we can take some policy action now to avert future shortfalls. Then next year, we make another assessment and, if necessary, make adjustments in tax and benefit policies to keep the system on track, and so on. We do not really need to know what the system’s liabilities will be 75 years from now, but we do need to know that we are moving in the right direction at a pace that we can reasonably expect will leave us in good shape 75 years from now. It is sophisticated folly to think that we really know today what the system’s finances will be 75 years from now. It is folly compounded by hubris to imagine that we can perform some econometric magic that will tell us what the system’s finances will be 999 years from now!

I take it that Larry K.’s (quite legitimate) concern here is that the government as a whole is engaged in a mindless over-promising, an over-promising that is being hidden by the short-term horizons typically used in government budgeting. This seems likely to be the case. But I see nothing of value added to the predicament by using projections of such long-term scope that they are themselves meaningless in policy terms. We agree it would be desirable if policymakers could have a vision beyond the next election cycle and could think and plan long-term for the consequences of their policies. But I suggest that trying to convert policymaking to use “infinite horizon” projections will not prove helpful. Ask yourself whether you can explain these projections and the necessity for them to the average member of Congress. I humbly suggest you cannot even explain them to me in a form that makes them seem desirable. What will happen is that politicians who want America’s fiscal problems to look as bad as possible so as to justify their preferred policy changes, might well opt to use them, as long as they appear to justify their ideology. Others will ignore them. And you will not be able to convince ordinary folks that these hyper-long range projections are inherently meaningful and desirable.

In fact, I will make a counter-factual prediction. Let’s imagine that some future demographic or policy earthquake happened such that the “infinite horizon” projections actually showed lower costs than shorter term projections. For example, let’s say we pass a new law today stating that 100 years from now the payroll taxes will be tripled and all Social Security benefits will be cut in half. Under this scenario, the IH estimate would show a lower future cost than the 75-year projections. In such a circumstance, I predict that a political magnetic field reversal would take place, with those who previously supported the use of the IH estimates suddenly coming out against them and those who opposed them seeing a new-found wisdom in their use. Just saying . . .

If we really want to reform federal budgeting practices, the solution is much closer at hand. If we could get the whole federal budgeting system to engage the same rigorous 75-year estimates that the SSA actuaries use to forecast Social Security, many of our problems with budget accounting games would disappear. Social Security’s actuarial projections are the gold standard. They are better than any used elsewhere in government, and better (as far as I know) than those used by any private sector firm with a defined-benefit pension. Rather than trying to convert policymakers to a new (and I think, dubious) metric, why not simply get them to adopt throughout government an accounting practice with which they are already familiar and which they have implicitly accepted for decades.

LABELS--

On this business about arbitrary “labels,” I honestly don’t get much of what Kotlikoff is going on about. But he did give us two examples which he hopes will clarify the situation. He asks what would happen if the government simply abolished Social Security tomorrow and replaced it with a system of government issued inflation-adjusted annuity bonds. Nothing would change in terms of the future costs of the system, but because of federal accounting practices this new debt would instantly show up as debt and its value would be stated in the immediate term. Whereas in the Social Security system these future obligations will show up only gradually, as the 75-year projections stretch out year after year. This proves, he thinks, that our way of accounting for Social Security’s future obligations is meaningless. No, I think it proves that our way of doing federal budget accounting is in some ways arbitrary, which does not come to the same thing.

Let me put it this way: the tax burdens of the future will be whatever they will be, given the demographic and policy realities of the future. We can offer short-range, medium-range, and long-range guesstimates of these future realities (as the Social Security actuaries do now) or we can try to imagine what they might be for all time. In either case, we are wrong. And both types of projections can be gamed, as I showed in my imaginary 100-year policy change.

Now, as I say, I think Larry K. and I agree on some of this, but I’m still not too sure, given his other example. He points to Chile and their privatization of their Social Security system in 1982. He says this is just like his imaginary example where we replace our Social Security system with government annuity bonds. I don’t think so. The point of Kotlikoff’s thought experiment is that the fiscal reality would not change in any way if we made this switch. But the switch in Chile had very real consequences in the here and now for actual people. Many Chileans who switched to the new system ended up with benefits significantly less than they would have received under the old system, and significantly less than their similarly situated neighbors who stayed in the old system. (This, and the “overselling” in Britain, along with other examples, are summarized for the lay reader in a recent book by sociologist James Russell, “Social Insecurity.”)

I think what Kotlikoff “proves” with his thought experiment is that two radically different systems might have the same long-range fiscal obligations, which does not necessarily mean they have the same policy consequences, as the Chilean example proves. And that is what I, and I think other supporters of the traditional system, worry about. That Kotlikoff will lead us to construct a social insurance scheme that is economically “balanced” but that does all sorts of damage to the social fabric.

SIMPLICITY VS. COMPLEXITY--

One last issue before I stop for now.

Kotlikoff can I think fairly be described as outraged by the complexities of our current social insurance system. No doubt complexity can be frustrating, and simplicity can be enticing, but pace William of Occam and Larry of Boston, simplicity per se is not necessarily a virtue. In the present context what ought to matter is whether a given system of social insurance is effective and affordable. And perhaps we can agree that an Orwellian system—in this context—is one that produces perverse results, often the opposite of what it claims.

The problems facing the Social Security system are real, and demand policy intervention. The question of the system’s affordability is a legitimate topic of debate, and the intergenerational equities must be addressed, and the solvency of the system improved. But the question before us is whether the Social Security system delivers the goods and whether the Purple Plan can do so as well, and whether these systems contain the potential for perverse results. But to make that case, we really need to engage his Purple Postcard Plan, which I promise to attempt in my next post!
My name is not Larry, but. . .
Three points in this excellent discussion are not receiving the attention they deserve, in my view .The first is the inequity of a working spouse receiving little more in Social Security retirement benefits than an otherwise comparable non-working spouse. The second is how much modern defined contribution plans and their companion employee benefit programs have grown to accommodate what one writer called the complexities of life. Lastly, I believe creating a basic retirement benefit by combining a modern DC plan with a required life annuity inflation adjusted payout is a pension actuary’s dream assignment-a big challenge with a high probability of success.
For me, it’s the spouse inequity issue that Larry K described in his first post that forces a consideration of a DC approach to Social Security. How else to change this inequity? What is the future of a program that pays retirement benefits to a person that are disproportionate to those received by others who paid the same premiums? I know one of these people well. Her Social Security retirement benefit after 30 yrs. in the classroom is about the same as her entitlement as a spouse. Given her age, my guess is that others in her situation might comprise a single digit percentage of all new retirees in the Social Security system. When that percentage grows to a double digit, I fear our beloved Social Security system will suffer a severe loss of credibility that will empower the people who really do have “an anti-Social Security agenda”.
The four trillion dollar DC market has evolved far beyond the “simplistic individual account programs” derided in this discussion. Investment charges in the largest plans are in the low single basis points for index funds. Principal protected options are available, as are daily valuation, automatic asset allocation and re-allocation and low cost professional management. Very importantly, in Larry K’s suggestion there is a required cohort-based inflation-protected annuity payout, which would capture the mortality gains in retirement for the participants. Thus the ultimate benefit is tied to mortality gains as well as to the capital markets. The inflation risk can be handled by a COLI rider, perhaps one from the private sector design that pays more later than it does at the start. A key provision would be the conversion rate. And perhaps the biggest issue is the transition from the current program. There are obviously so many issues to work out and the Purple Plan is just a start. But outright dismissal of the DC approach seems to assume that there will be an ideal time coming to tackle Social Security reform; but I fear the spousal inequity reality has already put that ideal time in the past.
Incidentally, the complexities of life should be no more a challenge to revising Social Security than they are to any employee benefit program. Survivors are covered by life insurance-Disability is LTD or DI-divorce is the province of the Qualified Domestic Relations Order. There are some very worthwhile programs in employer sponsored programs that deal with life complexities not in our current public Social Insurance System, such as Long Term Care and Dental Insurance (both for retirees and active employees) . My point here is that there is a lot of talent in this country that can design effective, transparent and equitable social insurance programs.
APOLOGIA-

I have already gone on too long in this forum about my views on K.’s scheme to privatize Social Security. But since he challenged his critics to address the specifics of his plan, and since I promised to do so, I fear I must make one more extended post on the subject. But first . . .


THOSE CONFOUNDED ACTUARIAL PROJECTIONS-

I have already written about the absurdity and meaninglessness of the so-called “infinite horizon” cost projections. But perhaps an additional point will prove instructive.

Any projection of future expenditures that has a limit, has a flaw. If we reach out 75 years into the future, then we can always ask, well, what about year 76, or year 100, or year whatever. The way of dealing with this in real-world actuarial practice is to update the projections each year, carrying them another year into the future. This is the sober, sensible, standard actuarial practice by which Social Security policymaking is framed. The way to deal with this flaw theoretically, in terms of pure mathematics, is to deploy a calculus-like procedure to rationalize the projection series to infinity—hence, the “infinite horizon” actuarial estimates favored by K. But are these “infinite” projections of future costs meaningful in a policy sense? Are they anything other than exercises in mathematical theory?

The reason we react with alarm to projections of huge future costs is that presumably someone will be asked to pay them at some point. If our projections are more than theoretical mathematics, then the concern is that real money will have to be spent at real times in the future. So when will these projected future bills actually be paid? In the case of the standard 75-year estimates the answer is concrete and specifiable: within the next 75 years, unless we make some changes, taxpayers will be called upon to pay out the present value equivalent of $9 trillion. When, by contrast, will K.’s $23 trillion in future Social Security costs actually be paid? When will we as a society have to come up with the cash? To this there is no answer. There is no specifiable point in time when we will have paid this total. About the best we can say is: “eventually.” No doubt. If the program continues into the indefinite future, at some point we will have paid out $23 trillion—or any other amount you wish to choose at random. But the fact that we cannot say how far into the future this point will come is because the $23 trillion is an imaginary number. It has no concrete reality. Its use is nothing more than a scare tactic to terrify us into accepting radical policy changes that we would otherwise be loath to adopt. That is how the “infinite horizon” estimates have been used since they were first deployed, and that is how K. is using them now.

If K. wants his Purple Plan to be taken seriously by those of us who actually believe in social insurance, then his first step needs to be the abandonment of the infinite horizon estimates (and all that talk about Ponzi schemes) and start using the standard 75-year estimates that serious policymakers have been using for the last 50 years.


WE HAVE MET THE ENEMY AND HE IS US-

The problem with federal budgeting is not the lack of an appropriate technical procedure for estimating future obligations. Stopping the many accounting and budgeting gimmicks in use in federal finance, and getting a precise fix on our present and future indebtedness, is neither a necessary nor a sufficient condition for addressing the core problem. The core problem has two parts: them and us. Fundamentally, we have to be adult enough to pay for the government goods and services we desire to consume. But like spoiled children, we want to just eat nothing but ice cream and forgo the vegetables. And if all the adults at the dinner table tell us that we can have our childish preferences—that we can have the government we want without having to pay the taxes to fund it—then the “us” part of the equation will behave like spoiled brats, which is what we have been doing for decades now. And the “them” part, the missing adults, are the politicians who for their own ideological/personal agendas have been only too willing to tell us that life is just one big bowl of ice cream. These politicians have been principally, although not exclusively, members of the modern-day Republican Party who, since the lunatic days of the Reagan administration’s supply-side follies, have won political power with the Big Lie that we can have all the government we want without having to bear the taxes to pay for it. The Republicans have for decades been taking the vegetables off the dinner table, with the con-job that, somehow, once the veggies were gone, we would all stop eating so much ice cream.

The idea that the problem with the federal budget is that we spend too much, is nonsense. Manifestly, the government spends exactly what we want it to. The problem is that we don’t want to pay for it. A perfect example, if one is really needed, is our wars in Afghanistan and Iraq. Beginning in 2003 and for many years thereafter, the American public (through its elected leaders) wanted to fight these wars—we wanted the government to spend money to prosecute these wars. But we did not want to pay for what we wanted. We put it on the nation’s credit card, unlike, say, the more honorable way that we fought WWII in which 25% of the cost of the war was funded directly by increased taxes. So we simply borrowed the money from future taxpayers to fund what we wanted to spend today. That is our problem throughout the federal budget. We want the spending but not the taxes (and other revenues) necessary to pay for it.

The problem, to be adult about it, is that we are woefully under-taxed (under-revenued, to be precise). As I say, the culprits in America’s fiscal folly are overwhelmingly the Republicans, with their ideologically-driven idolatry of tax cuts and their refusal to compromise on the issue. But the Democrats are not immune from the pandering impulse. Witness the short-sighted, irresponsible, policy engaged in by the Obama administration in 2011 and 2012 when it cut the Social Security payroll tax rate for two years so as to stimulate economic recovery. The direct effect of this was to worsen the insolvency of the Social Security system. The idea was that this money would be paid back later from general revenues, in the process deepening the general federal debt. So we forwent about $215 billion dollars in actual present tax income and replaced it with an equal amount of future debt. Given that the Social Security program is insolvent over the long-run, and given that the federal budget on the whole has a large overhang of debt as well as a current deficit, the only word to accurately describe such a policy is “idiotic.” Cutting taxes in any part of the federal budget is exactly the wrong thing to do. As to Social Security specifically, income from the payroll tax needs to be increased as part of a comprehensive plan to address the program’s insolvency, not decreased, and anyone who tells you different is a charlatan.

And so, with that backdrop, let’s take a look at the details of the Purple Social Security Plan.


THE COLOR PURPLE-

As far as I can see, there are only two animating principles in the Kotlikoff plan: 1) get rid of government; 2) add on whatever policy plugs are needed to keep the liberals from becoming apoplectic. Well, those and a fetish for simplicity, which I am not sure is an animating principle so much as a convenient argument against the existing system. In any case, let us examine in a little more detail the political color spectrum of the Purple Plan.

K.'s most basic aim is to get the government out of the business of ensuring our retirement security. He wants to abolish the Social Security program (the retirement and survivors part thus far) and replace it with a system of personal accounts based on private equity investments. This is the bright red part of the Purple Plan. Pretty much everything else in his proposal is by way of policy patches to repair the damage done by this first choice. In other words, to block the usual objections from the left, he adds to his core agenda a number of compensating “blue” provisions which he hopes will assuage the concerns of the liberals. Taken together then, one giant red policy and a host of smaller blue patches make for the political color purple.

Now we can certainly concede that K.'s version of privatization is vastly superior to that offered, say, by the Bush Administration in its plan to "strengthen" Social Security by abolishing it. The Bush plan made very few attempts to compromise with political opponents and its plan featured few fixes for the potential policy damages it might have caused (and it featured much dishonesty about its costs and impacts as well as its benefits). So, yes, by all means we should try our best to put in place fixes to repair the potential damage done by abolishing the Social Security retirement program. But before we examine those little blue policy patches in detail, perhaps we need to first ask whether we really need to cause the damage of privatization in the first place.

K. has offered several reasons for his Big Red policy: he says that Social Security is a Ponzi scheme; he claims that the true cost of Social Security is much greater than admitted; that the existing program is insolvent in the long run; that the federal government is a liar when it comes to its finances and the government simply cannot be trusted to keep its promises; that the Social Security program is a nightmare of unnecessary complexity; that the existing program contains outrageous inequities, etc. Thus we must abandon this decades old system that has provided economic security to millions of Americans and replace it with market-based Personal Savings Accounts.

What K. offers to supporters of social insurance is much worse than a Hobbesian choice. Perhaps we should hereafter refer to such deals as a Kotlikoffian Choice. To strain a bit for an analogy, his offer is rather like coming to a homeowner and announcing that he intends to burn down his home, and then he provides a menu of generous, liberal-minded, options for how the homeowner might best be cushioned from this catastrophe. No doubt this approach would be infinitely preferable—and generous and liberal-minded—compared to the fellow who just burns down our house and then moves on to his next victim. But somehow this type of “generosity” in public policy fails to appeal to at least this homeowner.

Secondly, as a would-be radical reformer, perhaps K. is allowed some latitude in rhetoric in making his case. But calling Social Security a "Ponzi scheme" is not innocent hot air. This claim, a favorite of various fringe commentators who think it actually means something, is demonstrably false. As K. well knows, a Ponzi scheme depends for its illusion of high returns on an unsustainable increase in the number of participants in the scheme. Social Security is simply an income transfer through time from the current generation of workers to the current generation of beneficiaries. Such a scheme can be sustained indefinitely, provided only that the balance between the two cohorts can be maintained in some fairly stable ratio. If there were no large fluctuations in demographics from generation to generation, then no honest person would be tempted to equate this kind of temporal transfer as in any way related to a Ponzi scheme. When the demographics do fluctuate significantly (like with the post-war Baby Boom) then temporary adjustments are needed to keep the system on an even keel. This has nothing to do with pyramid or Ponzi schemes, and K. is being blatantly dishonest to say so.

So, then, let’s take a quick tour of the specific features of the Purple Social Security Plan.


THE PURPLE SOCIAL SECURITY PLAN-

K.’s plan has 12 main policy features:

1) Pays existing Social Security beneficiaries their full benefits.
2) Freezes existing Social Security system by filling zeroes in workers’ earnings records for years after reform begins.
3) Requires all workers under 60 to contribute 8 percent of their wages to personal security accounts (PSAs).
4) Each worker's contribution is allocated 50-50 to his/her own PSA and to his/her spouse/legal partner's PSA.
5) Government matches contributions to PSAs by the poor, unemployed, and disabled on progressive basis.
6) All PSA balances invested in a global market-weighted index fund of stocks, government bonds, corporate bonds, mortgages, real estate trusts, and other financial assets.
7) Between ages 61 and 70, PSA balances for each cohort are gradually sold at market and converted to TIPS (Treasury Inflation Protected Securities).
8) All investing is done by a single government computer at zero cost.
9) Government guarantees that PSA balances at conversion equal at least what was contributed adjusted for inflation. i.e., government guarantees participants at least a zero real return.
10) PSA participants who die prior to age 70 bequeath unconverted balances to their heirs.
11) Starting at age 62, cohort TIPS pool makes payout to surviving cohort PSA participants in proportion to their age 60 PSA balances.
12) Distribution from TIPS pool is designed to ensure that real (inflation-adjusted) payout to surviving cohort members does not decline through time.


POLICY 1) The main question here is: how? By diverting funds from the Social Security income stream and directing them to market equities K. is depriving the system of the monies need to keep this promise. This is the well-known (if too little acknowledged) transition problem. When G.W. Bush proposed his version of privatization there was no acknowledgment of the transition costs, which were subsequently pegged at around $2 trillion. That is, in order to keep paying benefits to those already receiving and those near retirement (under the Bush scheme) we would have to impose this amount of new additional taxes above and beyond those then being collected in order to hold-harmless those already dependent on the system. The K. plan has the same obligation. Yet I find no acknowledgment in his writings as to how big his transition costs are and just how he intends to fund them. What he must do in order for his plan to be taken seriously is first to have the SSA actuaries produce an evaluation showing just what the impact of his scheme will be on Social Security revenues. Then he will have to openly acknowledge these new transition costs and specify how he intends to pay for them. If he has done this, I don’t know about it.

By the way, diverting the portion of the Social Security revenue stream due to payroll taxes for retirement and survivors benefits will almost certainly hasten the financing crisis in the disability portion of the program. The DI program is projected to run short of funds by 2016 or 2017 at best. The likely short-term fix for this will be for Congress to authorize a temporary loan from the OASI trust fund to cover disability expenditures until a longer-term fix is in place. But if we cut off the income stream to the OASI Trust Fund, withdrawing money from the OASI will hasten its insolvency. If there are no new monies coming into the OASI Trust Fund it cannot afford to loan assets to the DI Trust Fund. The immediate collapse of the disability program could well be yet another consequence of K.’s privatization scheme.

POLICY 2) Even though all workers under 60 would have zeroes entered into their earnings records going forward, there would still be the problem of the contributions they have already made. If those contributions are enough to render them fully insured, then they will presumably be eligible for a retirement (or survivors) benefits in the future. So K. would have to somehow protect not just current beneficiaries, but also all those now younger workers who are already insured, and their families if the wage earner dies. I see no provision in his plan to accomplish this. And even if those younger workers are not fully insured, they have still made those contributions (as have their employers). What happens to those contributions? How do the younger workers get compensated for their past contributions in the situations where they will not be getting a future benefit?

POLICY 3) This is the Big Red One. Suffice it to note that this proposal dramatically increases taxes on workers while eliminating them on employers. I am not sure this is right on either the equities or the politics, but shifting the entire burden for retirement security onto the workers surely represents a huge sea-change in the way we have traditionally understood the social insurance obligation.

In any case, since K. is proposing the elimination of the retirement and survivors programs (he will apparently be proposing something similar for the disability portion later), we can say that he is proposing to increase the payroll tax on all workers currently under the age of 60 from the current 5.3% (for the OASI portion of the tax) to 8%. This is a 50% increase in taxes on workers. We would have to describe this as (among other things) ironic, given that one of K.’s arguments against continuing Social Security is that it would require a 32% tax increase to sustain the program “to infinity” (21% for 75 years).

POLICY 4) This is the earnings sharing provision, which is designed to meet the objection that his privatized scheme will harm dependent spouses by eliminating their benefit, and to address the supposed inequities inherent in the spousal benefit. This proposal is hardly new; its history is long and tortured. Earnings sharing was considered as a policy option as far back as the 1939 Advisory Council and it has been considered and debated numerous times since. Indeed, in 1979 the Advisory Council of that era considered the question of earnings sharing at length, devoting a section of its final report to the issue. That report gives a fair summary of both the benefits and problems inherent in this idea, and despite the clear preference of some on the Council to move to earnings sharing, nothing ever came of the idea. By the late 1970s and into the early 1980s this was a topic of much political contestation. It was a particular bête noire of feminist scholars in the academy who promoted earnings sharing schemes as a way to empower women. Given this long history of failed efforts to persuade policymakers to adopt earnings sharing, we might wonder whether this suggests certain flaws in the idea. And indeed, there are.

There are lots of variations possible in an earnings sharing scheme, but they all share one common feature. In order to pay the spouse a benefit apart from that of the wage earner the benefit of the wage earner will have to be reduced by (roughly) an equal amount. That is, if we credit half of the wage earners covered earnings to his/her spouse, then this will decrease the wage earners own benefit by (roughly) half. (The only way out is to pay the wage earner an unearned subsidy, which is what the spousal benefit is now, so that will not take K. where he wants to go.) In this regard, Social Security benefits are very much a zero-sum game. This may or may not be a problem if the marriage stays intact, but all hell might break loose when a divorce occurs. Indeed, once primary wage earners grasp the fact that their no-longer-beloved ex is leaving with half their Social Security income, a firestorm of protest will erupt.

The other problem with earnings sharing proposals is that they are rapidly becoming irrelevant. The reason for the spouses benefit in the first place was because most women did not work in the paid labor force and so there was no way for a contributory wage-based social insurance program to reach them. So policymakers came up with the idea of a spouses benefit paid, as an unearned gratuity, on the basis of the entitlement of the primary wage earner (typically a male). But of course even in 1939 some women were gainfully employed. The equity issue arose immediately because some women were getting Social Security benefits without anyone (including the wage earner) having to pay any additional payroll taxes to justify these benefits. So the working woman, who did pay the FICA tax, was seemingly being treated unfairly compared to the homemaker who received this free benefit. (There is a related complaint about the way spouses benefits are paid when the spouse is also a primary wage earner. I won’t go into that here except to note that the same analysis applies.) In 1939 this was a real but small problem, as the vast bulk of women benefited from this provision and only a small percentage of women were “disadvantaged.” But as women surged into the workforce in the post-war era, the portion of women receiving “paid-for” benefits vs. those receiving free subsidies shifted. We long ago passed the tipping point on this issue. The number of pure homemakers who have no income from paid labor is small and continuing to decline. Women are earning their own Social Security benefits as primary wage earners and fewer and fewer are dependent on a spousal benefit for their income security. Eventually this idea of earnings sharing will come to be seen as patronizing and unnecessary—if we haven’t reached that point already. K.’s proposal here is solving a problem that is well on the way to solving itself, and the only reason it is included in his scheme is to buy-off liberal concerns about the elimination of spousal benefits in a privatized system.

All of this is why although it has been considered for decades, earnings sharing has never come close to becoming a reality. Simply put, this proposal is unrealistic and it is unlikely to happen. Thus we should not count on the K. plan being able to protect dependent spouses.

But much worse than the dubious promise of earnings sharing to protect spouses, is the plight of the other dependents whose benefits will vanish entirely under K.’s scheme. Spouses benefits are not the only ones tied to the entitlement of the primary wage earner. Under Social Security, dependents benefits are paid (under certain conditions) to spouses; minor children; disabled adult children; divorced spouses and dependent parents. Under the survivors program benefits are paid to widows and widowers; disabled widows/widowers; minor children; disabled adult children; divorced widows/widowers and dependent parents.

On the retirement and disability side there are about 5 million dependents receiving benefits, roughly half of whom are spouses and half children. On the survivors side, there are well over 6 million dependents of deceased wage earners receiving monthly Social Security benefits. Of these, more than 2 million are children. These 11+ million individuals receive in the neighborhood of $10 billion every month from the Social Security program. It is far from certain that all of them will have anything like the same income security under K.’s scheme. His plan makes a provision for the primary wage earner and his/her spouse, but for no other dependent beneficiary. And the provision for spouses is itself—as I have argued—far from certain.

What K. proposes is that the PSA account be gradually annuitized starting at age 61 and completing at age 70. If the account holder dies before reaching age 70 then any remaining assets in the PSA which have not yet been annuitized, become part of the estate. This could of course be next to nothing in the case of an account holder who is close to age 70 when he/she dies. And if the account holder has already annuitized the account, then there will nothing to inherit. So if the account holder dies after age 70 and leaves behind, say, a widow, there will be no survivors benefit for her and no assets in the PSA to claim. Wiping out the entire survivors benefits system and replacing it with the inheritance of the unconverted balance remaining in the PSA account, is a far cry from ensuring a monthly income for as long as the survivors live, which is what a defined benefit like Social Security offers.

On the other hand, if the account has already been annuitized in the form of TIPS securities, presumably the survivor with inheritance rights will be able to claim the TIPS securities (I am not sure about this, but I believe this is right, if not, then the “survivors” portion of the Kotlikoff scheme is a sheer farce). Even so, the unspent securities in the TIPS account will only go to the person(s) with first-order inheritance rights under state law, or to the person(s) specified in the deceased’s will. But what about those persons whom the deceased did not choose to bequest, such as a divorced former spouse? Or what about alienated children? Or elderly parents whom the deceased no longer wished to support? In other words, K.’s vision for income security is that primary earners get ownership of their retirement funds to do with as they wish, and the devil take the hindmost. That is certainly one way of looking at the issue of income security; but it is not, I daresay, the vision of social insurance. Under social insurance society assumes an obligation to ensure a fair and equitable measure of income security for vulnerable cohorts of the population, and we do so whether the primary wage earner likes it or not. We see this as the “social” part of social insurance.

POLICIES 5-7 & 10-12) Generally, I have no problem with K.’s investment strategy. The kinds of safeguards he is putting in place are sensible and desirable, and they are pretty much the sorts of things one needs—once we have decided to burn the house down. As a card-carrying blue, I applaud him for being more thoughtful than other privatizers. There is only one part of his investment strategy that sounds a bit of a false note.

Consider that K.’s overarching thesis about government finance is that the government in general is bankrupt and that the Social Security system is as well, and that we are so overburdened with debt that Social Security is unsustainable and many other government programs are as well. But of course one core objection to privatizing Social Security is that it exposes retirees and others to unacceptable levels of market risk, in general, and that most investments offer no protection against inflation—one of the most valuable features of Social Security benefits. How does K. allay these liberal concerns? By proposing to invest the annuitized PSA accounts in Treasury Inflation Protected Securities (TIPS)! So when the Treasury is incurring future debt in the form of promised Social Security benefits this is somehow scandalous, yet when it is incurring future debt in the form of TIPS securities this is somehow not a problem? I seem to recall K. getting quite worked up about the use of “labels” that mean nothing in economic terms. He told us that Treasury debt is debt whatever we call it. But somehow Social Security Trust Fund bonds are debt but TIPS securities are not? How is it that we cannot afford the one but can afford the other?

POLICY 8) Although I have commented on K.’s fetish for simplicity before, I cannot resist stating once more that his claims here are hogwash. The task of providing retirement income security does not consist simply of doing an annuity calculation. No doubt such computations could be handled on a single desktop computer for pretty much the whole country. But that is far from all that is involved in establishing and maintaining a system for economic security. We have massive computer systems in the Social Security program for much the same reasons that any large business enterprise does—to handle all the myriad of transactions and factors that are involved in managing a complex business enterprise. Social Security does not just consist of a benefit calculation. It consists of evaluating factors of entitlement; proofs of those factors of entitlement, including the identity of the claimant; changes in circumstance, safeguards against fraud, etc. Indeed, there is a massive reporting system that runs from employers to SSA and the IRS providing annual reports of earnings to form the basis of any future computations as well as the system of payroll tax withholding to actually collect the contributions to the system. Even if we replace Social Security with PSA accounts, there will still need to be the entire bureaucratic apparatus for assessing wage histories, collecting and certifying the PSA contributions, keeping track of changes in employers and in the circumstances of the workers, preventing fraud, etc., etc. The strawman that K. continually provides to the effect that all that is involved is a benefit amount computation, is nonsense. There will need to be a governmental bureaucracy to manage the PSA system, even if it has fewer employees and fewer complications than the existing Social Security system. And again, less complicated is not axiomatically better. Those complications exist for specific policy purposes. If those policy purposes are desirable, then a certain amount of complexity is just the price we must pay for a civilized system of economic security.

POLICY 9) K.’s scheme is at its core little more than the idea that we should replace Social Security’s defined benefit retirement system with yet another defined contribution plan. As Larry Thompson pointed out in his original commentary, this would mean putting all our retirement eggs in one basket, which in and of itself seems like a bad idea.

K. does not guarantee pie-in-the-sky returns from the imagined magic of the market, as most advocates of Social Security privatization typically do. He does not even guarantee that the resultant annuity will be equal to or greater than the promised Social Security benefit. And he implicitly concedes that it is possible to actually lose money when all your investments are subject to market fluctuations. All that he can promise—by way of a government (!) guarantee as a backstop against potential loss—is that every contributor will get back at least as much as they contribute to the scheme. So nobody will actually lose money in an absolute sense. But this is far from an assurance against loss compared to expected returns, or compared to what would have been paid under the defined benefit approach. And need we say, insurance that guarantees you only as much as you put in is not insurance, that’s called a piggybank.

I have to say, I don’t think K. actually believes in social insurance. In fact, he does not even seem to believe in the insurance principle when applied to social provision. The Social Security defined benefits system is a form of insurance. It provides far more to its beneficiaries than they ever contribute in payroll taxes, as all insurance does. It is not a personal savings account. Not everyone who contributes to the system will collect a benefit, as in any form of insurance. (As a point of interest, the system currently enjoys funding to the tune of about $13 billion annually from payroll taxes paid by illegal immigrants, who will never collect a dime in benefits.) But the pooling of monies made possible by the insurance part of the system allows it to pay out much more in benefits to those who do receive them than they contributed. The social part is that this system is operated by government, for the collective and individual good of its participants. As such, it contains elements of internal social policy—such as the weighted benefit formula—that completely disappear from a market-based scheme. If we abandon Social Security for PSAs, we will forfeit our ability as policymakers to tweak internal policy dynamics to produce socially desired outcomes. Some no doubt, see this as a desirable end in and of itself. As a believer in social insurance, I certainly do not. But at least we should all be clear that this is an additional cost of adopting the K. scheme.

K. seems to prefer a “social capitalism,” if I may, in which income security is personal and individual, and the only sense in which social consideration enters in is in that he proposes to backstop the system with government guarantees and investments in government securities. Internal governmental policymaking is jettisoned, in favor of a kind of external policy fencing around the PSA system. So K. wants the government to guarantee at least a zero real return on the PSAs, and he wants the government to issue TIPS securities to annuitize the accounts at retirement. But the government cannot, for example, decide that the plight of the old-old is a special concern and mandate that the PSAs provide a higher proportional benefit to this cohort. The only genuinely social insurance-like provision in his scheme is the idea that the government should match the PSA contributions of the “poor, the unemployed and the disabled.” For this provision, I commend him.


WHAT SHOULD WE DO ABOUT SOCIAL SECURITY’S INSOLVENCY?

K.’s likely rejoinder to all I have written here is that the Social Security program is itself insolvent and so we must have an alternative. His grand assumption is that the program is unsustainable, thus the need to privatize it. But this is manifestly false. The program can be sustained into the indefinite future, if we have the will and wit to undertake the necessary policy changes. So, then, the obligation certainly falls to someone like me to specify just what I would do to fix Social Security.

Actually, solving the long-range (75-year) solvency challenge facing Social Security is easy-peasy, from a policy perspective, although I am sure it will not be so easy from a political perspective. The Committee for a Responsible Federal Budget (CRFB) maintains a helpful online calculator that allows anyone to create their own custom menu of policy changes to restore Social Security’s solvency (see it at: http://crfb.org/socialsecurityreformer/) Using the CRFB Reformer tool, I managed to not only restore the program’s 75-year solvency, but to extend its life by solving 157% of the shortfall. The mix of policies I propose are:
• Eliminate the cap on covered earnings
• Raise the payroll tax 0.5% on employers and employees
• Create a minimum benefit at 125% of poverty
• Cover newly-hired state and local workers
• Allow the Trust Funds to be partially invested in market securities

I realize that not everyone will share my policy views here, and that is what the political debate over the future of Social Security should be about—not the purple fantasies of Larry Kotlikoff.


SUMMING UP-

So as I see it, here are the take-aways from our analysis of K.’s Purple Social Security Plan:

• The plan proposes to abolish the Social Security retirement and survivors programs, replacing them with a mandatory system of Personal Savings Accounts;
• The plan proposes to increase payroll taxes by 50% on employees while abolishing them entirely for employers;
• The plan guarantees a ROI of zero, at least;
• The plan promises a replacement for Social Security spouses benefits that is unlikely to ever be real;
• The plan offers no replacement for the other types of dependents and survivors benefits offered under Social Security, except in the most indirect way, and with no guarantees that the account holder will provide them;
• The plan depends in fundamental ways on government taxing and borrowing and likely increased government debt.

Well, need I state the obvious? In my view, all of this is a very bad deal as a replacement for the Social Security system—a system that has worked (despite all the naysayers, critics, and prophets of doom) for more than 75 years, providing more than $15 trillion in actual, real cash benefits to over 240 million individuals. Social Security works. There is no reason it cannot continue to work, all the way to the “infinite horizon.”
I don't see any reason why dental is not on medicare. a decaying teeth set up affection, that affect the human body. also teeth are apart on the body that has to be cared for, medically.

Post new comment

If you have an account, please login first so we can retrieve your information.

Your e-mail address is kept private and will not be shown publicly.
CAPTCHA
Please note that some comments are moderated, so it might take a few moments for your comment to appear on the site.