INTRODUCTIONS

SUZANNE DEMBO JAFFE:

Good morning. Welcome. On behalf of Mary and I, I would like to welcome you here. I know this is going to be a very productive and lively discussion. Our expressed purpose for the symposium is to review from a technical basis the principle findings of the Brookings and ICF studies.

Before we get into the meat of the symposium, I want to take just a few minutes to give you some history about our role as public trustees, and about how we came to focus on the subject of the investments of trust funds and the unprecedented projected build-up. Some of you know some of the history, but I thought it would be worthwhile for you to understand the context in which this all took place.

Back in the fall of 1984 when Mary and I first became the first public trustees, the papers were filled with headlines which proclaimed the looming bankruptcy of Social Security. Most people, both retirees as well as active employees, did not expect to collect their Social Security checks. There was a crisis of confidence in the system. As the first two public trustees, Mary and I wanted to restore public confidence.

Well, what could we do and what was expected of us? To find out, we turned to the 1983 National Commission on Social Security Reform, and there we learned that they'd recommended that two members of the general public be added to the board. The reasons for that was to inspire more confidence in the investment procedure and help to assure that the demographic and economic assumptions of the program can be developed in an objective manner.

Continuing our search for definition, we turned to the Social Security Act. There, we learned our duties, to develop and issue the Trust Fund Report and to review, if appropriate, and to recommend changes in trust fund investment policies. Mary and I were new at this, so we started asking a lot of questions to try understand current investment policy and practices.

Then we became involved in the preparation of the 1985 Trustees Report. The report projected, back in 1985, that the trust funds would accumulate substantial reserves as we moved toward the next century. The size of the accumulation was unprecedented. In March, we proposed to the Board that we undertake a project to study investment policies. Our project included the following: to evaluate current policy, to review the management of the trust funds from both a statutory and administrative point of view, to consider alternative policies, and to recommend to the Board what future policies should be.

We broke the study down into that portion which could be done by in-house experts, SSA and Treasury, and that which would require outside expertise. Clearly in-house we could deal with the history of the investments, with the policies and procedures and operations as they were taking place. However, we also recognized that we needed outside consultants to examine economic and fiscal implications of the projected build-up on the national capital markets.

We decided to contract two separate studies, allowing us to benefit from the different methodologies and conceptual approaches of two organizations. We asked the outside experts to simulate a variety of economic scenarios and to provide us with analysis of the impact of the changes on the trust funds. We wanted quantitative estimates and the mechanism from which to model using alternative assumptions.

Well, the process which we started back in early 1985 finally resulted in contracts being let to Brookings and ICF in early 1987. Those studies were completed this spring and presented to the Board in April. The Board was in unanimous agreement that the findings should be broadly disseminated and, most importantly, that critical review by other economists was important.

Today, we are going to discuss the technical aspects of those final reports. I know this is going to be an interesting and productive session. Standing here, I feel compelled to reflect on just how far we have come. In 1984, we had a crisis in the confidence in the system. Social Security was going broke. No one under 40 believed that he or she would see one dime's worth of their Social Security benefits.

Now I pick up the papers and I am bombarded with the good news. We currently have surpluses which are expected to continue well into the 21st century. I'd now like to turn the program over to Mary, who will provide you with more specific details about the mandate which was given to Brookings and ICF.

MARY FALVEY FULLER:

Good morning. I am Mary Falvey Fuller, one of the two public trustees representing the Republican side and Suzanne, my counterpart, represents the Democrat side. Before I get into opening the discussion by describing, you know, what we charged our two contractors to do, I'd like to welcome Dorcas Hardy, the Commissioner of Social Security. I am sure she needs no formal introduction to you. She has been a very dynamic leader of the Social Security Administration. She has been very supportive to Suzanne and me as we began our first term.

She has been behind us in terms of putting together the basics that actually the 1983 amendments left undefined. As a member of the Social Security Commission that created these two positions, I had not anticipated that when we -- first of all, that I would be taking office, and then when we did, there were no ground rules in terms of how we would be funded, whether we would have a staff, what it would be like, where the money would come from, any of that, how often the trustees would meet, how our input would be brought into the deliberations of what had been a three-person board.

Dorcas was enormously helpful to us and has been throughout our term. She has been intimately involved in our decision to launch this investment study, to carry it out and to now take it the next steps. So I would like to introduce her to you now and have her give a few remarks. Welcome.

[Applause.]

DORCAS HARDY:

Good morning to you all. It's nice to be here. Susan, Mary. I just came in on the tail end of Susan's comments about good news and confidence. It's true. We have lots of good news and that's why we're here, but confidence I'm not sure we've gotten there 100 percent yet. We all still seem to share our stories and I have mine that I take a poll with either taxi cab drivers or hotel clerks. I figure that's as good as anyone to talk with and to say "what do you think about Social Security and will it be there when you get there" and the answer invariably is no.

So we still have a ways to go on confidence, but I think you're absolutely correct. This is an extremely timely topic and I'm really pleased that the public trustees have decided to go ahead with this and to really pursue it and to have you all here.

First of all, aside from congratulating them for their thoughts and their looking ahead, there are many of you here who are actively engaged in the public debate that I hope is continuing to build with regard to Social Security. These are the kinds of issues that are controversial, but I think they're very important and I think they're not only important for the Federal budget and Federal policy, but clearly for the national economy.

Now I'm not sure, even though I credit the public trustees as being very far-sighted, that they were quite as far-sighted when they started this, to think that we would be having this eruption, if you will, of Congressional and public interest in the issue of the trust fund build-up. I was just reading Irving Crystal's article from the -- I guess it was an op ed piece in the Wall Street Journal last week. He said that the build-up in the Social Security reserves is creating immense perplexity among economists and legislators, who cannot figure out whether it is good news for the nation or bad news, or some undecipherable mixture of the two.

Now I clearly hope that Professor Crystal can at least be proven -- can be proven wrong only in one small area, and that's that I trust that the economists gathered here today will not necessarily concur that the issues are undecipherable. I don't think I need to mention that there are a number of questions that are related to this topic. The questions of the prospect of the reserves eventually absorbing the entire Federal debt, the effect of the build-up on the savings rate, the question of Treasury securities at the present and forms of public securities in the future.

I guess the question that I raise, even though I've now taken Dr. Hamber's class in Economics 101, is will there really be a reserve in the future. Now that's not exactly what today is for, but I keep saying that to myself and hoping that that indeed will be the case when you continue to talk about it.

The studies, I think, that were carried out, both by Brookings and Lewin-ICF, are really a very valuable contribution and I am most appreciative that they have done this work for us. As many of you know, I've really tried to place a lot of emphasis on the need to look at Social Security and to talk about it. That's not particularly a popular thing to do, especially in an even-numbered year, whatever even-numbered year there are, and -- but it's something, that it seems to me needs to be done.

We have a thought or view, I think, in this country, that Social Security is so sacrosanct, that even to mention it is tantamount to discussing or assuming its demise. I just really don't think that's true. I think that a clear and comprehensive debate will be the kind of thing that can ensure the survival of the system.

Now on the issues that are facing all of us are clearly not going to go away. Older Americans have one-third, what is it, one-third of the Federal budget today, and it will be two-thirds of the Federal budget about 40 years from now. So as all of us age, it seems to me that the factors that we look at the economic implications of the Social Security trust fund build-up has become even more important to all of us.

But I think that broader public policy debate, this is certainly an excellent place to start. So I look forward to the forward-looking thinking that I hear from you all and I hope that this will also feed into a bigger discussion of our whole national retirement income system. Social Security is clearly a cornerstone of that system, but there's a lot of other important components, like pension benefits, private savings and the retirement of all of us in the baby boom discussion in a generation will clearly present a lot of strains on all of these systems.

I think that we have an opportunity now to really look ahead and to decide what we want and how we want to structure all of that. So I have also not only been encouraging the public trustees to get into this discussion and debate and to be of assistance, but also the Hudson Institute has decided to initiate a policy review of our future national retirement income needs and those options.

So I would also encourage you at the time that you all will probably be asked for your opinions, thoughts, comments, discussion, that you would be equally enthusiastic about participation in that, as you have with your participation here today.

So I am looking forward to your outcome. I expect all the answers will come forth from this group and you'll tell us exactly all our options, what needs to be done, what happens if -- and I expect that you will have a very good session. Thank you very much.

MARY FALVEY FULLER:

When we launched the study, I will say that it took at least a year to do it, and one of the reasons was that as the two new trustees, there was a lot of concern that we would now add a lot of controversy and perhaps public visibility to the problems of the system and the dynamics of the system that might not be altogether constructive. The Board had pretty much acted in a fairly contained process, where they met once a year and signed the reports that went to the Congress.

Immediately when you think about investment policies, there were all kinds of theoretical alternatives about investing it in the private sector, all kinds of options that were thrown out and debated and we wanted to make sure, as trustees, that we did not get ourselves so embroiled in policy that we were unable to really function effectively as the surveillance group that the trustees are supposed to be. We are not charged as trustees with forming Social Security policy but rather, illuminating the dynamics of how the financial effects of the system work and what some of the problems and opportunities are as we go forward into the future.

So in order to make sure that got this study supported and launched, we defined our objectives and the mandates for the contractors as clearly and precisely as we could. We asked them to take an illustrative range of public policy scenarios. As you all know, it is not possible to discuss Social Security completely in isolation, and what happens to Social Security is very closely linked with what is happening with respect to the non-Social Security budget.

Rather than asking the contractors to postulate or advocate particular policies of how to manage the non-Social Security part of the budget, we asked them to develop a range of scenarios that would be illustrative of the plausible ranges, not the theoretical ranges, but the plausible ranges of what could happen in terms of the policy directions applied to the management of the non-Social Security budget.

We also then asked them to use the II-B assumptions. This was designed to again, not get them into the business of projecting the demographics or the other economic assumptions, but to try to normalize for that so that we would have a common base. We asked them therefore to take these economic scenarios, based on the II-B assumptions, look at the current investment policies, both the statutory policies and the administrative policies, together with this unprecedented configuration of what the trust fund reserves are expected to do, with this great rise in the first part of the 75-year period that we're dealing with, followed by a precipitous drop. So that rather than looking at the beginning and the end, which can happen under a various range of policy options, but rather to look at this configuration as applied to the investment policies that are implemented now, under this range of plausible economic scenarios, fiscal and monetary policy scenarios, and really look at whether the application of current law and administrative practice will in effect be feasible with respect to its effects on the general economy and the capital markets, in light of this expected build-up and then draw down.

So that's what we asked them to do. We had a wide range of people who bid for the opportunity to do the study. We chose two contractors that had the qualifications we thought were necessary, and in addition, we hoped would have a range of points of view so we would have as broad as possible a basis for debate on this subject. Their results, as Suzanne mentioned, were presented to the Board of Trustees in April.

We were very surprised by one fact that came out, which was that their findings were surprisingly similar with respect to the role of Social Security in the economy, and some of the sensitivities of the key variables. Really this similarity, which we did not expect to find, is what prompted us to organize this Technical Symposium.

Either this similarity represents a basis for consensus among you leading economists, in terms of what basis, what framework, we are dealing with as we look forward at the projection of this surplus, or in fact, there may be some factors underlying these analyses which do not show up at the level that the reports were given, but that maybe you might illuminate in the course of your discussions on the assumptions and methodology.

So our purpose today is to try to illuminate whether we will have a consensus, or whether we have some broad directions that look plausible but that underneath it has some issues that will need to be resolved in more detail before we then go forward and use this basis to discuss the policy implications of the results of these studies.

So that's what we're here to do today, and we're committed committed to follow through, either to celebrate whatever consensus that we develop, or to really follow through and ensure that the range of points of view are brought together in following through any issues that come out. So that's what we're here for.

QUESTIONS AND ANSWERS

MR. ENTIN: I'd like to take, nonetheless, seven to ten minutes for general discussion here, which will leave perhaps three or four minutes into the break and three or four minutes into the discussion session for the next presentation. If you could come to the microphone please and identify yourselves -- speak loudly or come to the microphone, whichever you prefer. Sir?

MR. KOTLIKOFF: I'm Larry Kotlikoff from Boston University. In general, I think I agree with the thrust of the findings in the Brookings study, which seem to be suggesting that despite possibly higher Social Security tax rates down the pike, that we are going to have capital deepening and as a result, higher real wages and lower tax rates of other kinds, so that the scenario looks basically fairly rosy. I think that that's true, but I come to that from a slightly different basis.

This is kind of my complaint with the formulation of the Brookings study, which is that with respect to the saving behavior. I think that the saving module here is very crude. It's not taking into account any of the microeconomics of what economists know about microeconomics of savings. We're not showing anything in the study about age-specific saving rates. The saving module does not consider how saving by different age groups is going to change.

It is assuming that the private saving rate stays constant and that national saving changes with kind of pari passu with Government saving. I see no real evidence at all in the literature to support that at all. First of all, the definition of private versus Government saving is a completely economically arbitrary distinction. The definition of taxes and spending and deficits is really an accounting distinction, not an economic distinction.

If you take a more fundamental notion of private saving, for example, if you take private consumption and divide it by NNP minus G, which is the output of the economy leftover after the Government consumes, you look at that notion of the private consumption rate and subtract that from one, so you get private saving rate which is free of accounting definitions. You'll see that that number has changed through time.

So the notion of keeping the private saving rate constant I think is quite arbitrary. I nonetheless think if one was building a more appropriate saving module, that you end up with a very similar answer, because after all, what is going on here is a reduction in the birth rate, and we know from Solow's work, that that's going to lead to capital deepening.

There have been some simulation studies which do, I think, the saving part of the issue better. I think this study does some other aspects of that question better than some other studies. But on the saving side, I think the other studies suggest a very similar kind of answer.

One other quick point is that in the early 1980s, if you properly measure the household consumption, the net national saving rate, if you correct for durables, the net national saving rate has not really fallen relative to the 1970s. So I do not think there is any strong evidence against the standard neo-classical view of savings. Nor is there such need for any need for such ad hoc approach.

MR. ENTIN: Thank you, Larry. Is there any response from the panel?

MR. BOSWORTH: You want us to respond? Okay.

MR. ENTIN: Would you like to? If not, we'll go to the next question.

MR. BOSWORTH: I think that I would agree with Larry, that more could be done with savings. I guess I would disagree on whether or not I think that there's much empirical evidence at present about what theory to adopt. But trying to reflect that, we did run a series of simulations, one in which various measures of the Government savings rate did lead to offsetting behavior in the private sector. So that the private saving rate was made endogenous with respect of fiscal policy.

Basically the effect of that is to dampen anything that happens because of the accumulation of Social Security. Second, there was a section where we did a model looking at age-specific saving rates, using some data the Social Security Administration had plus our own. The problem we found with that was that you get into a big argument over the effect of elderly people on savings. Do they in fact reduce their savings rate?

But equally important, what about children? What we find looking ahead, the dominant effect is not that there are more elderly people. The dominant effect is that there are fewer children. What does that do to the savings rate? Some of the studies are ambiguous on what you do with how you treat children in a model about savings behavior.

Finally, when we took the best available measures we had of age-specific savings rates, and we added them up and constructed a demographically corrected savings rate measure, that implied that that would rise in the future and then fall some, but would remain always above today's level. But then the difficulty was we extended that same calculation back to 1945, and unfortunately there was no correlation between that demographically measured savings rate and the private savings rate measures we had from the national counts.

So although the tendency would exist from demographics, I found the answer very ambiguous about whether we could find anything from historical evidence that was a consistent argument. I think this is the area that a lot more research is being done now, Larry being one of those who is working on it. Obviously one could extend the model to include that stuff when it became available. But at this time, we had to do the project.

I do not think the profession does agree yet on how to treat Government deficits and how to treat demographics and modeling private savings behavior.

MR. MAKIN: Barry and Gary have performed a valuable service in articulating the model that helps one to sort out [inaudible] on Social Security, in terms of these long-run economic assumptions. But two things that occurred to me as I read the paper that probably are worthy of further investigation.

One would come under the heading of open economy issues, and the authors show that they're aware of this. The version of paper that I read assumed that the rate of return on assets accumulated abroad was considerably higher than it was in the United States, to compensate American investors for the risk in investing abroad. I guess in the current environment, I would suggest that perhaps we think about compensating foreign foreign investors for the additional risk for investing in the United States.

I have never been comfortable with one-way risk assumptions. The basic point being there is no reason to suppose that that differential rate of return would persist. I'm not sure [inaudible] part of that. There are other issues, I think, that one needs to consider. Second point. The most interesting finding of the paper, the Brookings paper for me was something that I think Larry Kotlikoff has alluded to and I want to just amplify that.

Essentially, the capital accumulation that is implicit in enhanced saving rates, if you get that from the accumulation of the Social Security trust fund, leads to an anticipated increase in the tax on labor, because the fact is that you get capital deepening, the real wage rises. The way the Social Security benefits formula is articulated, then benefits rise more than total receipts. So if you think about it, you're going to have to increase the tax on the use of labor.

If that becomes well-articulated and everybody knows it, there will be some response. We already see some responses, I think, to the high cost of employing labor both in the form of high real wages and heavy taxes on the active employment in the United States, and is in the form of movement of capital abroad. If it becomes expensive to employ labor to produce things in the United States, if any of the benefits come into this category of payroll taxes, then rational multi-national firms, of which there are many, will simply produce what they want or produce elsewhere to sell it back here, like IBM or Eastman Kodak. There have been examples of that already.

That is just another result that unfortunately is more feedbacks from the rich results you already got, but I think those are my points. Thank you.

EDWARD GRAMLICH:

Well, as my instructions came down to me, I was told to write out comments, which I carefully did and mailed them off, but on the one hand, we'd given the slip in the timing. We don't have time for me to read them all. On the other hand, a lot of what I said when I wrote out the comments has already been said this morning. A lot of new things have been added. So I thought I would throw away my prepared text and just go it alone here.

First off, on the report itself, like Bob was complimentary on the other report, I thought this one was very good. It got about the answer that I would have thought one would get from this kind of policy experiment. It did it carefully. I guess if I have any complaint at all, it was written in a very laborious mode. That is, initially what Joe talked about were three scenarios, but really two. The report describes 17 scenarios. Joe, in his qualitative remarks, said this, that and the other thing could go on with private saving investment and so forth.

When you got down to business and what went on in the models that led to the results there, there were actually pretty rudimentary assumptions and so forth. So it's a little hard to figure out exactly what was going on, and part of my comments, my written comments which I will make available if anybody wants to see them, is a reader's guide to what really seems to be basic about the ICF report.

I'm going to mention just two points about that right now. One is that in contrast to the Brookings model that the ICF did have one model that dealt with the interplay between the short-run and long-run. I mean we all know that if we increase saving in some way by raising the budget surplus or whatever, that there could be a short-run negative effect and then a long-run effect through the capital accumulation and so forth. They do have one model, the MGM model, the Wharton variant, which combines that.

I suppose in that sense it's reassuring to note that you get roughly the same kind of qualitative results when you do it that way as you do in the long-run growth models. Now I suppose we would not be terribly surprised at that, because we are, after all, doing simulations over 75 years. So we do have a lot of time for the [inaudible] to work itself out, but at least in that sense everything is as expected.

The other point that I think is true. Neither speakers made too much of it. In the base path in the ICF model, there is a higher saving base path than in the Brookings model. The Brookings model takes the 1995 unified budget deficit to 1.5 percent of GNP. The ICF model takes it to zero, and then they do their simulations, where we either validate or do not validate the surplus in the OASDI trust fund.

Now in terms of critiques, actually both of these critiques have been made already, but let me repeat them. For all of Joe's words, and he did go through the various kinds things that could happen with private saving and investment, that like in the Brookings model, the ICF model held private savings constant. Bingo, just flat. There are several reasons why we might wonder about that in an exercise like this.

The first one is Social Security itself, and this is a little puzzling to me, but I will just raise the issue. Now in all of the stuff that we've talked about this morning, the deal that the Social Security trust fund makes with people is fixed. That was the deal given in the 1983 amendments, and so from a fully-informed, farsighted, microeconomic point of view, I guess you would wonder -- I guess you would expect that private saving would not be affected by that.

That is, private savers saw what the Congress did in 1983. They've adjusted and they have got to 1995 to get fully adjusted. That's when these simulations really start, and then we go from there. Now the question, then, is okay, all of the sudden we start reading in the papers that Social Security, which we thought might have been on shaky grounds -- the poll results we heard about earlier this morning suggest that -- and we start reading in the papers that the Social Security trust fund is building these enormous surpluses.

They're making the same deal with you, but would your saving be incrementally affected by that? The answer of both models is "no." There may be some other answer that somebody would want to give, but that is something at least is a question here.

The second reason is that interest rates decline, and that may affect private saving. The third, which I think is, to me, probably the most important of these -- again I do not know what the story would be -- is that all of these private savings rates are constant in the face of rather important demographic swings. Now Larry Kotlikoff says that we know about that than I thought we did. I don't know what we know about that.

I personally do not know too much about it, but you would imagine that these massive demographic shifts would alter private saving rate. So at a minimum, you think maybe we would want to put constant private savings rates for people of cohorts of different age and then run this through and see what happens. The fourth is the Barro effect, often discussed. I do not place much faith in it, so I won't spend much time on it.

The second problem is the open economy problem. Again, this has been mentioned and as I understand it, the Brookings model did a simulation -- this is as I understand it. I could be wrong here -- but a simulation that did not generate the results that you saw this morning, that did have a mixed closed-and-open economy model where not all of the saving went into domestic capital accumulation. Some went into claims on foreigners.

I do not mean to criticize Brookings. They at least gave lip service to the problem in their report and they did try to do something about it in their calculations. ICF gave lip service to it in their report but did not do anything about it in the calculations. Both their MGM and their MDM simulations assumed a fully closed economy, which means several things for these purposes here today.

First, as the OASDI trust fund accumulates assets, all of that goes into capital formation, as if this accumulation is validated in the rest of the budget. The second is that there is a reduction in interest rates. The third is that wages follow the capital stock. Now if you have an open economy simulation or if you have a mixed simulation, where you might view the United States as a country that is different from a small open economy, but it is also not a closed economy either; that is, there is some kind of hybrid mixing here, then you might imagine you can get a simulation where you would have some of the national saving not going into -- or some of the national saving going into claims on foreigners, that you could have muted interest rate effect, because you have this world capital market that is setting interest rates. You would also have a muted wage effect on the OASDI trust fund. So some of these calculations done on that would be different as well.

Now there is one puzzle in all of this and that is that as I look at the numbers, suppose we compared two calculations, one of which was fully closed, one of which was hybrid open, and what would you get from the results? I would expect that the one that is hybrid open would get a bigger GNP increase with the saving, because you would not run into saturation of capital. That is, if you run out of domestic profitable investment opportunities, you can lend them abroad and get the same rate of interest. I would expect less of the interest rate change.

As I compare the Brookings and ICF calculations, even the right ones, I don't think that works. I haven't had enough time to do this and I am just picking up numbers from tables here. But there is at least a minor possible discrepancy in regards to that.

In terms of what I think -- if the Social Security Administration hasn't run out of energy on this, what strikes me as the right thing to do about all of this, is that first off, you did the MGM simulation with the short-run and long-run, you found that didn't matter very much and it sure is a lot simpler not having all of that short-run long-run stuff.

So take -- let's take one long-run model. Let's build in at least a hybrid open economy and let's work out the private saving responses by age. I think that what all of this tells me is that those are the complications that seem to be significant, and if there is any taste for further work in this area, I think that would be the kind of thing that one might want to do.

I don't think there is a particular need for a horse race. Apparently, Social Security felt there was a need for a horse race last time. But I think everybody would probably agree that the results are similar enough that maybe we are beyond that, and we can just try to take one combined model and focus in a little more carefully on saving investment, which I think after all are the key things here.

Now I would like to say one final thing on the policy itself. I am supposed to be only on technical side and not comment on policy aspects, but I can't contain myself, so let me take a minute or two and do that. The stuff we're talking about this morning may generated volumes that are so boring that once you put them down you can't pick them up again. That may be.

But I think it's potentially pretty important for the United States, and as one who has been long pushing that we think a little more about the share of output that we're devoting to consumption and investment, and even raise that -- I've also pushed that view -- I think that this kind of thing is something that we ought to think about seriously.

I guess as I read the reports, you ask yourself what would you do about all this from a policy point of view. I guess what I would do about all of this, I'm pretty tentative on this because I haven't really started thinking deeply about this, but it seems to me that the notion that is described by Brookings of close actuarial balance -- I'm sure they didn't get it -- there must be some suggestion about it in present law or something -- but that is a useful notion.

If we can discipline ourselves such to raise tax rates 75 years in advance, which would be quite surprising to those of us with any Washington experience, but if we could, that is certainly a notion that ought to be praised. As I understand it, this is a minor problem for Social Security. It is a significant problem for Medicare. Medicare is the trust fund that is kind of swept in the closet in all of this. Medicare is, as I understand it, dealt with as the rest of the Government. But the close actuarial balance concept would make for a significant change there.

The second thing is that it seems to me that validating the surpluses, such as they are in the trust funds, is probably the right thing to do. We have massive demographic changes coming on and somehow if we are deciding saving policy over the generations, you would think that we ought to adjust for the size of generations in some way. So I think that roughly validating the surplus is probably the right thing to do.

I also think, again partly colored by my own Washington experience -- but I think most people would tend to go along with this -- that as long as the trust fund surpluses are sitting right in there in the budget, it would be real hard to develop the discipline to validate the surplus. I think that -- I haven't really thought through this before reading these papers, but now that I have, I think that I'm coming around to the view that we might want to for these purposes take these trust funds off-budget.

We operate on them in close actuarial balance, and then we do what we do on the rest of the budget. But we don't just keep the money right in there, so that we feel that we could spend it. Now of course you could argue that as long as we know what the trust fund surplus is, etcetera, etcetera, etcetera, that we'll always do the right thing. I guess I'm a little more skeptical than that, and I think that given the proclivity not to save that seems to be coming more common here, I think that just removing the trust funds from the formal budget might help in these regards.

It might not help, but I think I'd rather go into the 21st century with this kind of budget accounting than with the opposite. Thank you.

QUESTIONS AND ANSWERS

MR. ENTIN: Thank you very much. We're 12 minutes into our 15 minute break. What I would suggest is that we take seven minutes now for questions and answers, and push back the break for that length of time and perhaps get the break cut back to about five minutes. It will take us at least that long to rearrange the furniture. Any questions for this panel? Dr. Stein?

DR. STEIN: Well, my comment is really addressed to both, and I follow up on what Mike Gramlich said about what's the significance of all this? I think that's all very wonderful stuff really, and what really bears relevance is the question of national fiscal policy, more than to the question of the trust funds. But when I ask myself what can you make out of that and what the lesson seems to be is if you save more, you're richer.

But then I ask well, is that a good thing? Well, what I'm getting at is what I missed. Is there some explanation of the path that real per capita consumption would follow over these 75 years under these various scenarios, because it seems to me that is something you could look at in order to decide whether you like one or another, that if I assume that under all the big pie savings scenarios, you have more consumption at some time in the 75 year period. But you have, under some assumptions, if you have more of it you also have it later.

So you can have some idea about what you think is an acceptable path of real per capita consumption, and how much of it are you willing to sacrifice now in order to make your grandchildren three times as rich as you are. Well anyway, it seems to me that's the kind of relevant bit of information information and it's easily derivable from what you have got.

I was not really very relieved about the answer, as to where you would invest all this money. At least on some assumptions, I understand that you get up to $12 trillion, which is a lot to put into guaranteed mortgages. Anyway, that does not relate -- even if they are guaranteed mortgages, somebody's going to look at them and guarantee them. That is, it does not relieve us of the problem of whether the Federal Government is going to be in the business of micro-managing the investment of an awful a lot of money.

As I understood it, these models assume that the labor supply is given. I suppose the labor supply is given by the number of workers. But of course the labor supply is really the quality of input, which is affected by the amount of investment in education. Investment of education does not appear in any of these models as far as I can see. I think that could have a significant effect, because it will affect what happens to interest rates, because the interest rate decline, I guess it comes about because of the increase in the capital labor ratio. If labor is measured only by the number of hands that is one thing; if measured by some kind of quality input, that is another thing.

I did not understand in the Brookings model, whether they're assuming -- they seem to be assuming a constant number of Government employees. Is that right? Something is constant about Government employees.

[Laughter.]

Then with the Brookings proposal for the actuarial balance, does that assume that you have some demographic assumptions out for another 75 years beyond the 26 years?

VOICE: Yes we do.

DR. STEIN: Very good.

MR. ENTIN: I'd like to make one comment. Harry Valentine has already promised me that in next year's trustees' report, Appendix G will show those surpluses and build-ups both in nominal and real dollars, and the $12 trillion trust fund build-up is $2.7 trillion in 1988 money, which is still a large number, but it is only $1.7 trillion if you toss in HI, which someone should at some point.

MS. ROBBINS: My name's Aldonna Robbins, and I'm with Fiscal Associates. Both studies, as I read, have what seems to be a somewhat disquieting finding in that when average wages increase, the OASDI system is made worse off. So here we are trying to encourage capital formation to raise productivity and then we end up making OASDI worse off.

The Trustees Reports going back many, many years have a section entitled "sensitivity analysis," where they try to measure the impact of changes in various economic variables such as average wages. Now in this year's Trustees Report, an increase of one percent in average wages, which I believe was mentioned in the Brookings study, ends up making the system better off over a 75 year period by one percent of taxable payroll. In other words, it lowers the long-term deficit. So I wonder if the two -- if the authors of both studies could perhaps reconcile what I see as inconsistencies.

MR. BOSWORTH: I think that's a fairly easy thing to differentiate. You've got to make a sharp distinction between level and growth level and growth effects. If the economy grows one percent faster, real wages -- which is the sort of thing the trustees often do -- is always better off, because the tax revenue runs ahead than the benefit package.

If you increase the level of real wages and hold it constant, the proportionate gain in taxes and benefits is exactly the same in the long run. They are proportionate to one another. In this simulation, because the II-B projection has in the last half of the period, benefits way in excess of tax revenues, an equal percentage increase in benefits and an equal percentage increase in taxes, widen the deficit.

Since for the 75 years as a whole, the II-B projections are in deficit, a proportionate increase worsens it. But if you start it from a system that was balanced, an increase in the level of real wages is neutral. It does not make the system better, and it does not make it worse.

MR. POTERBA: Jim Poterba, MIT. I want to follow up something that Ned mentioned at the very end of his comments, which to me were in the realm of political economies in the technical assumptions about the two studies. It's very clear that the conclusions one draws about national savings depends greatly on the degree of offset that takes place in other parts of the Federal budget.

It seems to me that there, thinking about what assets the trust funds are allowed to hold, may potentially become quite important, not just because the only security backing these Government liabilities are Treasury securities or some other sort of security that the public does not understand very well. They undermine the kind of confidence effect that you mentioned.

I think there's also just a serious question of where it's to bail out other trust funds, or to consolidate things in terms of budget accounts. I would argue that even though they'd be more difficult to manage substantial portfolio, but if one of our concerns is encouraging saving and trying to prevent the use of these funds to just bail out the rest of the Government, there is an argument for forebearing the cost of managing the portfolio and try to hold other sorts of outside securities. I think that those are the ways to try and strengthen the Chinese wall, if you will, between these sets of accounts and the rest of what's happening in government.

DR. MUNNELL: Alicia Munnell from the Federal Reserve Bank of Boston. I want to follow up on Jim Poterba's comments and something Herb Stein said. Both studies assume that the only way we can invest is in the private sector. In fact, not all Government spending is consumption; some of it is investment. An alternative is to do some of that directly through the Federal Government -- bridges and roads, education. I had not thought of this before, but if Ned can toss things off the top of his head, I can too, that might raise questions of whether we want capital budgeting in the Federal budget, as well as taking Social Security out of the trust funds.

The second point I would like to make is that the Brookings exercise, the extension that they performed is really quite useful, because it highlights how silly the path of trust fund assets that we have in current law is. There is no reason in the world to build up the trust fund and take it down, because that assumes that the baby boom [inaudible] is going through system. In fact, we don't have a lid. We have a permanent increase in cost and if you have a permanent increase in cost and you mark your fund, then you want to pre-fund it permanently and keep your trust fund around. Thank you.

MR. ENTIN: Just one more, and then we'll take a five minute break.

MR. VAN DE WATER: I'm Cole Van de Water from the Congressional Budget Office. A couple of comments. First of all, the concern that was raised about the simplicities of the assumptions on savings behavior. Without getting into the substance of the policy question, it seems to me -- I'm emboldened by Herb Stein's suggestion, that the question is indeed whether or not you want more or less Government saving. It seems to me there is where the features of the models that might be relevant are the ones that would potentially be changed through different public policies.

I assume that the response of savings to demography is probably not one of them. Maybe I missed something there, but the effect of interest rates on savings, of course, is placed where there could be policy related feedbacks. Even if the effect of the interest rates on savings may be less important, in one sense, than the demographic issues, it may be more policy relevant. But anyhow, I suggest focusing on the items that policy might influence.

Secondly, pursuing the area of quotas, the question is namely where do the models produce significantly different results. The only place, it seemed to me, there might be such a difference was in the two ICF models and their effects on interest rates. I would be interested in finding out whether people do think those are significant difference or not, that it looked to me that if there were a significant difference somewhere, that might be it.

The third point. I have been perplexed for as long as I've had the paper, the significance of the Brookings' burden measure, the final column on the final page that Gary Burtless pointed out. I think I finally understood it. Maybe I'm the last person who figured it out. But I think there is a problem in that measure because it is adding up two different things.

It includes the Herb Stein questions of what's happening to the level of consumption, but it also takes up a short-run changes in the ratio of Social Security benefits to NNP, which show the ratio of benefits to GNP does go down, because Joe Anderson said at first when you have this stronger saving and higher policy and higher [inaudible].

So there's where this measure is benefitting from that in the short run, even though the benefits in terms of per capita consumption, what I think people usually look at, don't begin to show up for a bit longer. So I think that at the very least those two features should be distinguished, rather than combined in a single measure.

MR. ENTIN: Perhaps I could give the panel just two minutes to respond, and then we'll take the five minute break and please do try to come back promptly after five minutes. Thank you.

MR. ANDERSON: From what I've listened to, my only comment on one [inaudible] all the confusion, which is the impact of changing demographic structure on savings. That is represented, I think, in some detail in the macroeconomic demographic model (MDM). Our assumption was that -- [inaudible] comparing one scenario with another, that a trust fund accumulation or validation scenario would change in any way that demographic influence on savings. So that underlying private savings behavior, as affected by demographic changes, is the same in both scenarios.

The one thing that we forced a change was the amount of real capital accumulation. Of the many issues, those that we did not permit the model to investigate would be negative interest rate feedbacks on private savings and various other negative feedbacks such as the Barro and that would be a useful follow-on study.

For the results that we presented here, we try to present kind of the extreme case. What if all of the trust fund could be invested in the economy versus what is not [inaudible].

MR. ENTIN: I think one consensus that emerges from the panel so far is that a great deal [tape cuts]

That may indicate that we need some sort of behavioral model of the Congress and I would be willing to bet that we get a behavioral model of the public before we get a behavioral model of the Congress.

WILLIAM NORDHAUS:

Well, before I get into the fist fight or the food fight, I would like to say at the beginning that these two studies are very fine studies. I think they are economics at their best, at its best if you like. They are clear. You can understand their assumptions. You can understand their conclusions and they're well-documented. So even though I'm going to play the normal discussant's role of finding faults and nits, I would like to say that these are very useful and interesting studies.

One question I have. One reaction I had to them is

-- reminds me a little bit of the old Pirindello? play, six characters in search of an author. I found in these reports that I knew all the answers, but I wasn't sure what the questions were, and the -- in particular, I wasn't sure what the questions were that related to the Social Security trust funds and trustees and Social Security policy, because most of them -- most of the important issues there were frozen.

Now in my comments, I want to -- I have a number that I won't go over in detail, but I had two comments I'd like to dwell on, one having to do with uncertainty and the other having to do with investment policy. Then I will sprinkle a few other ones in those.

First, with respect to technical comments, it is true that as a number of people have remarked, that the results of the two studies are quite similar, and it is not surprising, as a number of people have remarked, because they both use a kind of canonical model of economic growth, and if you use the same assumptions you get the same results.

Now this is actually true mainly -- I think we are over emphasizing the similarity and canonical nature of the studies, in the sense that they actually rely on the canonical model of production. But I think the other half of the model, which is the supply of savings or wealth-holding behavior, even though they're quite similar, I think there are lots of questions, some of which have been raised here.

The two important ones, having to do with the omission of open economy issues and the reaction of individuals to change in their public pensions. I think as a number of sidebar comments and sort of side comments over this morning indicated, changes in those, particularly the open economy, introducing open economy aspects, will have a significant difference on the results.

Now having said that they are very similar, I would like to express a concern about lack of attention to the uncertainties in the models or in the projections. These take highly disaggregated specifications of the economy and forecast them out more than 70 years. I would think we should treat these as good, best-guess estimates, but we surely should pay more attention to the inherent uncertainties about the projections.

Now in thinking about the middle of the next century, and projecting behavior of that time span, we really do have to remember how much social and economic behavior is likely to change in three quarters of a century. It's impossible to know, looking forward, but just to remind myself whenever I look at a projection, I'd like to think back over the period of time.

So I go back 75 years to 1913, and just ask how much have things changed since then, as a reminder of how much things are likely to change in the next 75, and just I'll remind you of a few things about that period. In 1913, there was no Federal Reserve System; there was no Soviet Union; and Franz Josef ruled the Hapsburg Empire. There was no Federal income tax; women could not vote; Model T's were a rich man's luxury.

Seventy percent of men over 65 were in the labor force as opposed to 10 percent today. The birth rate was double it is today; the infant mortality rate ten times today's rate. The divorce rate was one-fifth of today's rate. Abortions and homosexuality were illegal and AIDS was unknown. Social Security was unborn. Just one other thing. Those of you who are over 40 will remember about the dentist, and I'll just remind you my favorite example of cultural change is that going to the dentist in that era, before novocaine, was about as painful as going to the gallows.

Now I remind you of these things only to suggest that we really must take them seriously in doing our forecasts. As Santayana might have said, "Those who forget history are condemned to believe their forecasts."

[Laughter.]

We all have our favorite, we all have our candidates for the worst forecasts. One of them is, I think it was the Secretary of Treasury in 1933 or former secretary, who said when the U.S. went off the gold standard, this is the end of Western Civilization. But the favorite is, and one that came to be shown very incorrect in only two years was a forecast of a Spanish Royal Commission in 1490, commenting on the likelihood of Christopher Columbus' voyage.

This is quoted from Samuel E. Morison's book on Columbus: "According to the Royal Commission, Columbus plans to sail west to find a shorter route were considered impossible for five reasons" -- I'll only quote four of them -- the second two. "The Western Ocean is infinite and perhaps unnavigable." Three, "if he reached the Antipodes," which is what it was known as at that point, "he could not get back," four, "there are no Antipodes because the greater part of the globe is covered with water and because St. Augustine says so," and fifth, most interesting "So many centuries after creation, it is unlikely that anyone could find hitherto unknown lands of any value."

So this only took two years to be found a faulty forecast. Now to come a bit closer to home, let's go back to assumption II-B of 1986, which is the gospel according to the Social Security Agency and followed by the two consultants' reports. Just to see what happened, I examined population projections of the United States contained in the 1969 and the 1987 statistical abstract, only 18 years apart, and I'll just give you one number. The population now is around 250 million. The projection for 2015 in the last projection, which is the middle one, was for a growth from today to 290 million in 2015. If you go back to the '69 Series B projection, which was commonly used, it was 431 million. So you can see looking forward today, the difference between a growth of 40 million live citizens to 180 million. That's quite a change. I can go through some other numbers, which I will include in my written comments.

I note these, not to point the finger of blame, but simply because we really must take into account in our studies the inherent uncertainty over a period this long. With this in mind, I have one study I would recommend to the public trustees, and that is a carefully designed study of the impact of uncertainties upon the OASDI system, and what I recommend is using a small and manageable model and using the well-established principles of subjective probability and Monte Carlo experimentation or simulation.

You could actually determine over a period of time what the likely impacts of different uncertainties are on the trust fund, such things as economic, demographic, work, work span, work week, earnings, family composition and so on and so forth on the trust funds. It's really important, because even though these projections give us good best guess scenarios for the future, they give absolutely no indication as to whether the future is likely to be anywhere near these projections. But we must look at that.

A second set of comments, which I'm just going to mention the title of, is overall fiscal policy. I think I have nothing to say but the conventional wisdom here, which need not be said again. I would simply add one point, and that is my concern is that having seen the article in the New York Times three or four months ago, which let out one of the best-kept secrets about the buildup of the Social Security trust funds for everybody else outside of the inner circle, it worries me now that that is going to be a -- lead to a national relaxation with respect to the fiscal deficit.

As you look out, it's probably premature. We're going to be spending our nest egg before we even -- the chickens are hatched. But it would really be foolish at this point to relax our efforts to decrease the fiscal deficit and to increase national savings, because of a Social Security surplus that's way into the future.

Now as a final issue, again coming back to Social Security, I'd like to talk about the question of investment of Social Security. It's been touched on before, but I do think there is an important issue here, and that's the question of how should the trust funds be invested.

Now the consultants raise the possibility of investing in non-Treasury securities, but it's more by force majeure almost that you ran out of Treasury securities to invest in. I think actually there's some interesting and important economic issues here which I'd like to raise.

One of my tasks in recent years has been to participate in the management of a large endowment, probably larger than the trust fund endowment now at Social Security, but soon to be outstripped by it. My reaction to the trust funds policy is that these are unconscionably conservative. I'm going to say why this is a dangerous line of thought in a minute, but let me just pursue it.

If you look at the trust fund investment policy, my calculation is it has, using historical data and reasonably decent assumptions -- I think reasonably decent projections from a wide variety of areas, that it looks like those should have yields of about one percent a year. Our calculations are that an aggressive investment strategy could get at least five, maybe six percent a year. Those are both real returns.

Now it is riskier in the simple sense of the word, in the sense of more volatile to invest in these risky assets. But compounding that over time, the higher return does begin almost to dominate the less risky investment. For example, we did a 50-year simulation and found that even though the, in terms of standard deviation, the aggressive portfolio is a little more than twice as risky as the so-called riskless long-term bond portfolio, after 50 years the probability of having a higher terminal value is over 90 percent, because of the compounding of the higher return. A risky or aggressive portfolio would include domestic equities, foreign equities as well as bonds and perhaps -- our aggressive portfolio includes other things as well.

Now having said this, the question is whether the experience of a private portfolio is economically or politically relevant for the Social Security trust funds, and let me deal with each of those. Just briefly with the political issues, I might remind you that the Federal Government currently owns 34 percent of the equity of large corporations through the Federal income tax.

Now the immediate response is that this is socialization, to invest in private equities. But I would call it actually privatization. But on a more significant point, I think we should not underestimate the difficulties, and I don't, of getting -- and the complexities of dealing with other than Treasury issues. But if the returns are high enough, you know, it might be worth thinking about.

There are, it would seem to me, easy mechanisms to get around what is the biggest problem, which is the involvement in corporate management. There's no reason in fact that you couldn't perhaps -- well, there are no reasons in the longer run where you couldn't have non-voting classes of stock, but there's no doubt that you'll get involved in the wrangling about tainted investments in South Africa and Northern Ireland, and I'm not sure that's something I would recommend. But again, if the stakes are high enough one should think about it.

But let's turn to the really important issue, which is the economic issue. We obviously must be wary of the fallacy of composition, and worried about what might be true for a small individual portfolio, such as one that's a very small fraction of the economy. We have to wonder and worry whether that's also true for the nation as a whole or the trustees of a national policy or national institution.

The consultants both opine on that and I think -- I find myself in complete disagreement. I will read from both of them. The ICF report says "it may be that assets acquired by the OASDI trust funds could have greater expected return. Such a policy would simply redistribute the composition of assets and risk in the economy, would not change the overall productivity of the economy or the aggregate rate of return." The Brookings report states "it is important to recognize at the outset that this kind of policy would have little economic significance. It would not in any way change the composition of real assets in the economy, not change the total amount of capital" and so on.

Well, when two eminent sets of consultants come up with the same conclusion, apparently independently, it would surely seem foolhardy to disagree. But I think that they're clearly wrong from a theoretical point of view, and demonstrably wrong from a practical point of view.

Now the error comes -- going back, trying to think of how they could have fallen into this pitfall, it comes, I think, from their finding that most dollar denominated fixed-income securities are perfect substitutes, in the sense that the relative interest rates across these do not appear to change with relative supplies. That's something which I think has been well-known by the financial economists for a while, and it's probably right. I mean it is clearly right.

But it doesn't mean that's true of all assets, and I think it's their extension to all assets which is the problem. We know, just to give you some reasons why we think that's not true for all assets, we know, for example, that the change in outside assets between money and non-money has a big effect on the relative returns of money and non-monetary assets. We are also pretty sure that changes in dollar and non-dollar denominated, U.S. Government assets, will have some effect on the -- okay. Time's up.

So let me just then go on to give one more example, and that is if you think of the U.S. as being part of a world economy, a small part of a big world economy, where the world economy is close to -- treats different assets as close to perfect substitutes, then it's clear that any change in the OASDI portfolio will have a dollar for dollar change on the nation's portfolio.

So to sum up, then it does appear to me that we should think carefully about whether in fact it would be useful to change the OASDI policy of investing only in Treasury securities. The yield on the securities, the yield for the trust fund would certainly be important, and would change the tax rates for the Social Security tax rates. That is certainly not neutral. That is certainly not distributionally neutral, and I believe that this is another area where the trustees might want to consider a careful study of the impact of alternative investment policies. Thank you.

QUESTIONS AND ANSWERS

[Please note in this section most speakers nowhere near a microphone; inaudibles in some places]

MR. PALMER: Thank you. Two observations. One is implicit in this whole symposium, and I think is well understood by people, but I think it's still worth noting, and that is as we move from a pay-as-you-go system to one where these projected surpluses are likely to develop, it's clear that issues related to Social Security are now so intertwined with broader economic policy concerns, and that's really what this has been all about.

Both models show what common sense would show, but show it in a lot more detail, and that is that what really matters about the surpluses, what's important and interesting about them is not what goes on within the Social Security system, but what's going to go on external of the Social Security system in terms of the rest of Government and private sector reactions and adjustment or coordination with those.

I think that's just important for us to know, because it does have implications obviously for things like the role of the trustees, both ways. It means that managing the Social Security system now becomes an integral part of overall management of the economy. People who are charged with the one must be thinking about the set of issues and concerns it raises for the latter. It also means that policymakers who are concerned about broader economic policy are going to find themselves much more concerned and much more involved in playing in the Social Security policy formulation. That has other implications too.

So it's a new ball game in ways that it hasn't been in the past. I think it's admirable that this symposium is laying a technical foundation for that broader dialogue that has to go on.

The second point is -- has to do with this finding that higher real wages, stronger economic growth, capital formation, etcetera, don't change the long-run financing picture for Social Security. They don't help the actuarial balance. It actually seems to hurt it a little bit. It has to start as an artifact, the fact that there's a deficit [inaudible] projection.

I think it's important for us to keep in mind that while it doesn't change that balance, it does represent a higher level of real wage rates, a higher level of real GNP out there in the future. When the time comes to figure out how we are going to adjust to the increased pressure of demographics through some combination of potentially reduced benefits or increased payroll taxes, the larger GNP we have gives more flexibility for however those adjustments are made.

That is, if the size of the pie is bigger, it can be carved up slightly differently as between retirees and non-retirees, and still leave both groups better off than they would have been in the absence of that higher real growth. So its wrong to think in terms of well, this contribution that Social Security may be making to higher savings, real growth, capital formation, that it doesn't really help the system when you look at from within the system.

It will provide a better context overall for thinking about how we are going to make those adjustments that we're going to be facing as we move beyond the retirement of the baby boom, to put the then 75-year projections into some sort of equilibrium. So it's very valuable. I think one extension of these studies might be to begin to look at what the implications for real benefit levels per capita, for real disposable income of taxpayers and so forth might be under these different models, and then how different ways of adjusting benefits and taxes will be for those two groups in the future under these different scenarios, to help drive home the point of the value of having that larger economic growth and higher GNP out there to help adjust to [a COLA rate] consistent to demographic changes.

VOICE: Well, I don't necessarily want to follow Joe Palmer. It was too long-winded. I do ask you to keep it short.

MR. MAKIN: I'm on notice. For my second reactions, I'll be quick. I just want to say that Rudy suggested that we could get a consensus in the room about whether the national savings rate is too high or too low. I don't think we could; I don't think we should. I don't think Bill Nordhaus or Rudy or I or anybody knows what the national saving rate should be.

What we were talking about, what we are talking about today is the result of a policy that is designed to raise the national saving rate, and it becomes a critical issue today because of the way we have designed our systems, which say (1) we have a numerous and generous indexed benefit programs for the elderly, (2) the finger points at the baby boom generation that says that you will be elderly some day, and therefore you will be a numerous elderly generation; and (3) therefore, you must save more.

What that ignores is the possibility that maybe we ought to redesign those programs, because what they are now saying, what the messages coming back from those programs are saying, that we should have a higher national saving rate or else we weren't not going to be able to fund those programs. That may be so. There's a question in my mind is should we really try to satisfy the parameters in those programs, especially when we know from the golden rule, which has been mentioned briefly here this morning, but let me remind you that the golden rule says that changing the saving rate does not have any effect on per capita consumption, real per capita consumption over time. So what's the point?

MR. BURTLESS: [inaudible] doesn't affect the growth rate; [inaudible] affects the level of consumption.

MR. KOTLIKOFF: Larry Kotlikoff. My comment relates to what Rudy seems to be saying and Herb Stein. I agree with Rudy's intellectual basis for changing the deficit. You know, take these figures put them off-budget or whatever. But I think Rudy means to admit that there's no intellectual basis for the current definition of the deficit.

As economists, we should not be trying to help with the Congress with their fiscal issue problems. We should try and put together some numbers which make some sense from an economic perspective about fiscal policy. Now from an economic perspective, what we should be doing is bring together generational accounts which would indicate in present value what each generation is going have to pay to the Government over its lifetime.

From the perspective of generational accounts, we saw in 1983 a change in policy and these generational accounts would indicate that the baby boom generation is going to be able to afford less consumption -- it relates to the issue you're raising about the consumptions, Herb. These generational accounts are not going to change. If policy doesn't change, then through 1990s when we start running "surpluses," these generational accounts will not change.

So we wanted some numbers which are invariant when the policy doesn't change. What we've got is a bunch of numbers which are based on cash flow accounting, which do change even though the policy isn't changing. So I think another study that Social Security could actually foster would be some totally different present value accounts that would show us why our policy does or doesn't change and what it really is.

MR. ANDERSON: I would like to respond to a couple of the comments. Bill Nordhaus' disagreement with both our study and Brookings' findings, because there is no meaningful alternatives to current investment policy that could increase the rate of return to the trust fund. From my view, Bill missed the point, at least the point that we were trying to make in our study.

Our view was that ultimately the income of the Social Security system depends on the taxing power of the Government. To maximize the income of the Social Security system, we want to maximize the overall productivity and efficiency of the economy. That follows for the optimal allocation of the economy's capital, including the public interest to the point that Alicia Munnell brought up, and we also mention that as a possibility in our study.

It's not that we don't believe that some other portfolio could earn over the long run a higher rate of return than one that is entirely Treasury securities. It is that we believe that that's really irrelevant from the social perspective of optimal trust fund investment policy. The trust fund perhaps should be used, along with other instruments, to optimize a country's capital allocation, both between public and private capital and domestic and foreign.

The second point I want to respond is Rudy Penner's, that commenting on the ironic result that increasing the productivity of the economy doesn't help funding the trust fund under current rules, because [inaudible]. We make the point in our study that ultimately the way real capital accumulation and increases in productivity could be used to fund higher benefit outlays that are going to be required in the future, would be to permit a reduction in replacement rates, a reduction in benefit rules without reducing real benefit payments relative to a scenario where no trust fund capital is accumulated.

MR. HENDERSHOT: I'm Pat Hendershot from Ohio State. I guess it's kind of common and nice to take a broad view and talk about the broader big picture. But the current trustees probably have to worry about a narrower view and I'm not sure exactly at all what their [inaudible] is, but it's probably to make sure to the best of their ability that the system stays whole, and that if we have promises to make people large payments at some future date, that we don't renege grossly on those at some point. That would seem to be the public trustee's role.

I think with that in mind, they have to take a little more narrow view, and I think they do have to worry a great deal about how the rest of the Government might respond to these large surpluses. If the rest of the Government doesn't stay in balance or in reasonable balance and just offsets these surpluses in the funds with large deficits, then the system just isn't going to hold up. At some point, 50 years from now or 75 years from now, the economy isn't going to be able to support the payments at that time to retirees.

If that is a problem, or if the public trustees view that as a problem, they really ought to think about some of the things that Rudy was talking about, and other people, about maybe trying to scale down current benefits some, so that everybody takes a little hit on the Social Security system rather than one particular generation.

MR. ENTIN: One of the tables that is not in the Trustees Report might shed some light on what these future promises are and that's the benefits being promised to future retirees as the wages grow. Implicit in the II-B projections are benefits that could approach for college graduates some $20,000 per a year, for a single worker in 1988 dollars. I forget what that amount would look like in current dollars, but clearly it would be useful to have more data available to the research community and to the public, so that they could decide which way to go in making future adjustments.

I wonder if anybody has thought of modeling the alternative benefit formulas that were looked at back in the 70s by Professor William Hsiao, when he looked at price indexing versus wage indexing, or the reports that were done over the last ten years or so by the Social Security Administration discussing methods of changing retirement ages to keep the ratio of retirement years to work years from changing as the population ages.

I think all of these would be very useful studies if someone would will fund them and do them.

MR. BOSWORTH: Just one thing. It goes back to a comment by Paul Van de Water. Maybe we did have the wrong measure. The issue has come up from several people mentioning that we should look at the consumption path [inaudible]. The burden calculation that we made was intended to do that. Under each one of the simulations, we calculated the amount of total consumption available in the future, minus the increased payments of Social Security.

Since we don't have in the model any separate generations, what we were trying to get was how much better off are non-Social Security recipients on the assumption that Social Security benefits are all consumed. It does seem to me that that's the right measure for us to have used. We took half the total consumption difference from the base line, minus the increase of the Social Security benefits from the 1986 level. So that we thought we were capturing the amount of consumption left over, the incremental consumption.

The interest thing we found that came out about that, mainly by accident, is that if you have an actuarially sound system, sort of like a private pension fund, you increase taxes or cut benefits, whatever, it moves out of balance today, but [inaudible]. You eliminate all increased costs in Social Security on future generations.

It turns out that the increase in GNP and increase in consumption generated by saving the Social Security surplus precisely equals the increased cost out for 75 years of Social Security benefits. The point of that is you should keep all this in mind. The demographic curves look dramatic. The increased burden of Social Security is not overwhelming. The real problem is when we started to simulate Medicare. But that's not really a problem of the elderly again; it's that health care costs keep rising, in excess of the rate of inflation, that they've been going up over the last 25 years.

All of us are in trouble having anything left over after Medicare for consumption, except the doctors, medical doctors. But they're such a sickly group.

MR. BURTLESS: I'd like to talk about two points that were raised repeatedly here. One has to do with the criticism of these models, because they do not take into account the possibility that Government [inaudible]. It's surely not true in our model. We all are very unsatisfied about the way you modeled the effects of an open economy. But nonetheless, we do follow them and we think a lot of improvements are needed. We have believable results just like anyone else does.

But the second thing is just to remind people who have repeatedly asserted here that more saving is good and perhaps people have forgotten the conclusion of both of the studies. In both simulations, saving the currently scheduled Social Security surplus is good in the short run for Social Security and more consumption and saving and so on.

It's bad in the long run, because Social Security in the long run is running a deficit. So if we really do believe these II-B projections are long-run, by the end of the period, people's current consumption is lower, their saving is lower, their capital stock is lower. Social Security benefits are perhaps at the same level that they would otherwise be, but that is just it. After the end of the projection period, eventually Social Security benefits are going to be lowered.

If we say that under the current II-B assumptions, it's not good. It's just the opposite of what most of the people have said here. The only time this conclusion changes is if we change the fundamental assumption in both sets of researchers were given, and we assume instead that the Social Security system has an internally consistent rule about the taxes -- a tax rate is like a levy on wage earners -- or the benefits benefits it's going to ultimately pay for. Benefits will be lower or taxes will be higher.

Under those circumstances, Social Security will always contribute to additional national saving, higher real wages and higher benefits. So --

MR. STEIN: So you're saying that over a 75-year period, consumption will be lower and you'll have the higher saving?

MR. BURTLESS: By the end point.

MR. STEIN: By the end, in the last year. I'm saying what's the use of saving it because you can't take it with you. MR. BURTLESS: No, I'm not saying that at all. From the point of view of the Social Security system itself, narrowly looked at, if it adds to national saving and then subtracts from it, from the point of view of the Social Security system it is worse off over the 75 years. That's number one.

Number two, from the point of view of the economy taken as a whole by the end of the entire period, consumption is lower, Social Security benefits are lower, and capital stock is lower.

MR. STEIN: That's based on very suspect assumptions about national saving.

MR. BURTLESS: I agree, I agree.

MR. STEIN: And private and public saving.

MR. BURTLESS: So maybe everyone sees through that there?

MR. MORAN: I guess I'm not quite as sanguine as Gary is about the net contribution of an internally consistent financing rule to public saving, with the premise that we really studying it from a highly negative base case to begin with. I think all of these scenarios make pretty clear that it's going to take a pretty big policy shot to get to the point where Government savings is positive in the first place.

While we all frame the issue differently, you by specifying 1-1/2 percent of GNP, non-OASDI deficit and we doing it another way, I think anything that understates the fact of how far out of that range of balance do we have to be. So I think that --

MR. BURTLESS?: No, but relative to whatever you're baseline case is, saving the Social Security surplus over the 75-year period actually lowers the capital stock at the end, if we assume that the tax rate and the benefit schedule remains unchanged over the 75-year period. On that, it doesn't matter which of the reports you read; you come to the same conclusion. MR. HAMBOR: But Gary, that just reflects -- I'm sorry.

VOICE: No, go ahead, go ahead.

MR. HAMBOR: That just reflects the fact that there's an error and the system's not balanced.

VOICE: That's right.

MR. HAMBOR: But I mean, all we're talking about here in this whole situation is this -- from Social Security's point of view, is the question of how you are going to fund it? Are you going to fund it in a pay-as-you-go context or are you going to fund it by building up the trust fund? That is the whole issue, man. John Palmer made the comment if you build up the fund you will be rich. That's always, of course, going to be true. The question is should we fund the system for the baby boom cohort with an irrational public policy, or should we go to pay-as-you-go system by following some of the proposals that are out now for cutting payroll taxes and so forth.

The important point is recognizing that that's the issue that we're talking about, and then also recognizing there's a real problem with that. We might not validate the surpluses and we might not cut payroll taxes. We might think we have a big fund and we never really get saving.

So I think it's important to focus on the issues from the point of view of how you -- that we're talking about. Are you going to fund the baby boom retirement, recognize that the policy [inaudible] as to whether this current generation of workers should be double-taxed or taxed more, because you've got to do that to get the money. But recognize also that if you don't look at it that way, you're going to miss the point. Build up the fund [inaudible].

Now the Brookings extension just simply says let's maintain the fund. What you're talking about in the scenarios of the study was if you build up the fund and then sell it off and you get the kind of results that you're talking about. What Brookings does in the extension of this is by raising taxes when the system is [inaudible] toward the balance, maintain essentially a permanent fund, at least [inaudible].

MR. HENDERSHOT: I have one question or one suggestion for the Brookings people. You did keep the fund whole by raising taxes. Did you also keep the fund whole by increasing the retirement age?

MR. BURTLESS: Well, we can do that. We thought that from the point of view of Social Security and from the point of view of the economy as a whole, at least the way we were analyzing it, both of those policies are going to give us the same rough profile of a continued build up in Social Security. That's the point. The Social Security would continue to add national saving to every year in the 75-year projection period. Whether you do it through continuously calibrating benefits, so they're lower for each cohort, or whether you do it through raising the tax rates, those cohorts pay for their retirement. We just picked the one.

MR. DAHLMAN: I am intrigued by the Nordhaus comments on how difficult it is to predict and on the problems with the canonical models in economic growth, and nobody really picked up on that in this discussion. We talked about all the uncertainties [inaudible] different scenarios within a simple model. My microfriends tell me [inaudible] range in [inaudible] expectations in macro modeling is that non-stationary models and where sort of the future is around [inaudible] past trends to where they tell you where you're going to be [inaudible] at all.

I'd like to see that kind of a model and just take the II-B assumptions and let the uncertainty go over time [inaudible] propagation of errors. I mean, that kind of a model the only certainty you will get in the end is going to be that a large block of single issue voters will take care of themselves, and I think that's an issue which folks like Carolyn Weaver, who just left, will take up in her study of this at the [inaudible].

CLOSING REMARKS

MARY FALVEY FULLER

Well as the group starts to disperse, I want to make sure that everybody was here, who has been here really understands how much we appreciate the quality of the preparation, the quality of the presentations and the quality of interactions that went on this morning. There are so many issues to cover and so many good points came out, that I think there is a lot of food for thought and a basis for some further work, much of which can be at least assembled and organized in something that will resemble proceedings, but will not be a big tome.

That will be hard to take up, but something that will give you a sort of a roadmap of what kinds of things were raised and what kinds of work could be done and maybe some of the issues that would link the discussion we had today with the upcoming discussion that we'll be having on the policy implications of this study.

A word. Somebody mentioned the role of the public trustees and what we think, you know, we might focus on. We are the first two. We have a four-year term that expires on September. We've struggled in the early parts of our tenure with what our role is, and we found very quickly that the other board members and therefore the board as a whole, because we are two out of five, were very strong on the idea that we really are fiduciaries, not in the sense of having literal fiduciary liabilities. We both checked that out before we accepted the nomination.

[Laughter.]

But our role is to look at what is coming into the fund and what is coming out of under various assumptions over time, and really comment on the financial stability of the trust funds over the short run and the long run up, out to the 75 years. As much as it would be really fascinating and exciting to get into some additional issues with respect to policy options and trade-offs, as trustees, we aren't going to be able to really make proposals in the format of the Board. I think there's the whole question of how the Board would interact with the Congress and the mechanisms that are in place to raise those policy options or debate those policy options.

We do see it that it is within our scope, however, that coming from a study like this, to elucidate some of the implications of the dynamics that we're studying here with respect to the effect and linkage between Social Security and the larger economy on policy options, so that we can more effectively crystalize, from one point of view, what some of the options are as a contribution to the national debate. I think we intend to do that.

There are several of the suggestions that came out today that I think are within our scope, and if we could find a way that's practical to proceed, fund and go forward with some of these things, I think of one of the few examples are looking at the effect of the uncertainty of forecasts and assumptions on the projections and go back in time, actual versus forecasts and trace how those uncertainties have really affected the assumptions and the program as it's gone forward. That is [tape ends]

... special Social Security debt and whether there are other investment instruments that might be considered or appropriate, some of those aspects. I think that would be within our scope.

We will be looking to see, to sort out the issues that were raised today and identify those on which we could go forward, and in any case assemble something that would be useful for the next step, which would be the assessment of the policy implications. So you will be hearing from us. We appreciate greatly your involvement and hope that this is the beginning of a dialogue that will give us the opportunity to really address these problems before they become crises and to really be able to take some reflective time with a range of views to really debate the options and the implications that are before us.

So thank you all very much. Susan, would you like to say a few more words? Okay then, our meeting is adjourned. Thank you very much.

MR. DAHLMAN: Just one final comment. I think we just gathered that none of you made the obvious suggestion that we ought to invest [inaudible] in economic research.

[Whereupon, the conference was adjourned.]