A Conversation with David Salisbury of the Employee Benefit Research Institute

The founding organizations of EBRI at that time were 13 large employee benefit consulting firms, many of whom in the 1960s had been asked by the Kennedy Commission on Retirement Policy to do studies. They, all thinking that they were the only ones that had been asked, diligently did their work and then found out that a whole lot of duplicate work had been done since they were all doing it for free. EBRI was a way for all of those firms to assist the new Commission with data and studies but only pay for it once. So, EBRI came about as an enterprise that was not to advocate, it was not to lobby, our bylaws and incorporation documents actually have a prohibition against doing either of those things, and to do data, to do basic research and education and to build databases over time that would allow anyone that had an interest in retirement programs to be able to track the effectiveness of those programs, what those programs were, what they were doing.

Initially it was principally focused on retirement, and then about 1982, after the commission had finished its work and the decision was made to definitely keep the institute going, that was the point in time when we made the commitment to extend our work to the health area and broaden our work to include employment based health benefits, general health costs and health management issues and research on the Medicare program. We've been doing all those things now since then and maintain two websites, www.ebri.org, where all of our research since 1978 can be found, and www.choosetosave.org, which fulfills the other piece of the original incorporation mission which was public education and worker education on what these programs were, how they could benefit from employer involvement and the ways in which individuals should be considering their own health and financial futures.

Macchia – When you mentioned Jimmy Carter, Dallas, it reminded me of my entry into financial services in 1977, through the insurance door. I can recall back then that to a great extent life insurers were the custodians of a vast amount of pension assets and that defined benefit pension plans were quite popular and typical. It occurs to me that since EBRIs formation in 1978 you've had a view, a consistent view, over a set of phenomenal changes that have taken place in the landscape of pensions and employee benefits, generally. You've seen and studied the transition from Defined Benefit to Defined Contribution, the emergence of 401k programs, all the way to today's level of popularity, and now the effort to institute Defined Benefits into DC plans. I wonder if you see a cycle that's emerging. How do you view this phenomenal transition that's taken place since EBRI's founding?

Salisbury – Well, you asked about my role in all of this. I was the first employee of EBRI and have been the Chief Staff Executive since we started. I came to EBRI from two years with the Pension Benefit Guarantee Corporation, which is principally a Defined Benefit oriented entity, and two years before that with the US Department of Labor in what is now the Employee Benefit Security Administration where I set up their first Office of Policy and Research. I've basically been doing on some level the same type of work since late 1974.

To your point, the original EBRI board was principally made up of people whose background was as pension actuaries. Those firms role in the retirement area was overwhelmingly related to Defined Benefit retirement plans and actuarial work for Defined Benefit retirement plans. There were obviously, in many companies, thrift savings plans and profit sharing plans on the side, but for the vast majority of what would have then been the Fortune 100 and even the Fortune 500 had a Defined Benefit plan as their primary retirement vehicle. And the only thing they did as a retirement plan was a Defined Benefit plan.

You had a handful of exceptions, Procter and Gamble stands out as a company that has always been a profit sharing and Defined Contribution company, and I mention them because if one talks about design of programs that would do what the good old fashioned DB plan did, you'd have to look at somebody like Procter and Gamble where essentially year in and year out the contribution to Defined Contribution accounts has ranged between 15% of pay and 25% of pay, as I stressed year in and year out. Procter and Gamble still does that. Procter and Gamble provides investment options in spite of a fairly heavy emphasis on Procter and Gamble stock.

There's always been the alternative of people taking the lifetime income annuity if they did not want to take a single sum distribution. Because it was a profit sharing plan where the employer was putting money in regardless of what the employee did, that plan always basically assured 100% participation of everyone in the workplace which is one of the other traditional features of a Defined Benefit plan. So, a healthy enough contribution so that people can retire with an adequate retirement income, a set of managed investment options aimed at allowing people to build enough up over time with that substantial contribution and the opportunity for life income security protection through a pooled annuity arrangement if people did in fact want to know that they wouldn't run out of money before they ran out of life.

If one takes that transition to what we're seeing in the work world today as Defined Benefit plans have morphed themselves you really have to go back to the late 70s when Atlantic Richfield Company amended its traditional Defined Benefit pension plan to offer single sum distributions. That led the Internal Revenue Service to do something that in hindsight was a very significant decision. Atlantic Richfield had added single sum distributions for a participant with high income. If your pension is not more than $35,000 per year then you have to take the first $35,000 a year as an annuity. That was the late 70s and so the vast majority of people weren't going to have that big of a pension and could not take a single sum distribution. One might say it was a balanced policy that would assure workers a base income for life on top of social security. The Internal Revenue Service went into Atlantic Richfield and said, "All get a single sum or none can have it."

As I recall, the Chief Actuary at the IRS, at an Actuarial Society meeting said, "And we know that Atlantic Richfield will choose none, and so there will not be lump sum distributions." They didn't understand that CEO very well who wanted a single sum distribution and Atlantic Richfield says fine. The IRS says we have to let all of you have a single sum distribution. Now, some decades later nearly 55% of Defined Benefit plans that still exist offer a single sum distribution at the point of retirement. Essentially out of that Defined Benefits system of those 55%, generally a minimum of 80% of the participants take a single sum distribution. Generally you end up with a maximum of 20% taking a life income annuity. In many cases it's only 4 – 5% that takes a life income annuity. In some Defined Benefit retirement plans of large employers and small employers no one takes the life income annuity.

The notion of Defined Benefit plans that I was introduced to in 1974, frankly, began to be deserted by the largest Plan Sponsors in the United States before we'd even turned the corner into the 1980s. By 1984 when Kwasha Lipton, then a consulting firm, one of the founding firms of EBRI, in a consulting relationship with Bank of America, moved Bank of America to what is popularly known today as a Cash Balanced Defined Benefit plan. A second trend moved forward because part of that transition included a standard reversion of assets to the Plan Sponsor. A process of workers seeing very small balances upon the conversion relative to what they thought they ultimately were supposed to be getting out of that pension plan, and the decisions which Congress and PPA just finally dealt with after the IBM suits.

It took until 2006 for some resolution in 2007, of essentially issues that began to be implemented in the Defined Benefit system in 1984. You then add on top of those combined actions the total restructuring, if you will of the largest enterprises in America. Microsoft didn't exist in 1974, and today it's one of the nation's largest employers. Outside of the state of Arkansas Wal-Mart did not exist in 1974, and today it's the largest single employer in the world. We have basically seen a growth of enterprises dominant in marketplaces, Starbucks didn't exist in 1974 that have grown up as Defined Contribution only companies.

Those Defined Contribution plans have another very different mix than what the profit sharing plan than Proctor and Gamble did. In those plans the employer automatic contribution has always been de minimis. Large percentages of workers have not chosen to participate and even when they did participate the amount that the employer would contribute has generally been quite small relative to what we can describe as the Proctor and Gamble standard. If we then go to the final, if you will, today's transition point as very large American companies, the IBMs, the Verizons, the Lockheed Martins, that were originally tradition Defined Benefit companies and have gone through a morphing process and are now at the final stages of essentially saying that we will no longer have Defined Benefit plans. In the case of IBM, no one will be in those plans post 1/1/08. For companies like Lockheed Martin and Verizon there will be continuation for some that were already there and just transitions for new workers. So, we're seeing different varieties, but what we're seeing when they put in the Defined Contribution plan that goes to your comment of "making them look like Defined Benefit plans", I stress that they don't in any way, shape or form look like traditional Defined Benefit pension plans, if you will, my father's Defined Benefit plan.

Pop died in July months short of turning 94. He retired in 1978. He wasn't given a single sum annuity option; it was a final pay plan, etc, etc. This transition is that we will have automatic enrollment, but you can still opt out. The result found in the data is that between 10 – 30% do opt out, even with the automatic enrollment. We might do automatic contribution escalation. So far that data is that a minority of firms will do that. We might do an automatic employer contribution. A very small percentage of firms are doing that. Those that are doing it are doing it generally with an employer based contribution of 2 – 5%. For most individuals that will not be enough, even if the employee contributes the same amount. If they started saving at 20 the needed rate would be at least10%, and if they didn't start contributing until 35, the required annual amount would be closer to 23% per year.

So you end up having gone from a Defined Benefit and in a Proctor and Gamble case, a profit sharing plan where the employer was putting in enough money to provide a full career worker with true retirement income adequacy, to what has come to be termed a Defined Benefit/Defined Contribution system where most employers automatically put in nothing, where a large portion of employees can fail to participate and where still only about 20% even offer an annuity, life income annuity situation. Essentially none require a mandatory annuitization.

If we look at the experience of the last 30 years where Defined Benefit plans have offered the option, very few people will chose to do what my father had to do, which is take that life income annuity. It's the morphing that changed over that 30, nearly 35 years has been dramatic and I think so dramatic that we really do need to underline that even with all of the changes that are now being talked about and undertaken for Defined Contribution plans, that even if all of those changes take place those programs, in the absence of aggressive individual savings that exceeds anything individuals have done in modern history, will never accomplish or achieve what Defined Benefit plans did, or what profit sharing plans of the Proctor and Gamble variety did. That says, even with that redesign that the principal shift of all of this is that even the best off workers will have to do far, far more for themselves than was the case under those old systems.

Macchia – You know that's a fascinating historical perspective that explains the implications of many changes over the past 30 years. It reminds me of the countless seminars I've presented over the years to audiences of consumers. I can recall about 10 years ago making comments about the transition to the 401k plan that 401k essentially was one of the greatest financial foibles ever foisted on the American public. Back then I might have said that the transition away from DB was motivated by greedy corporations that wanted to improve their balance sheets and free themselves of long-term obligations. I wonder if there's a more sophisticated answer to that, and I wonder if your dad's own experience offers the pristine example. Where having retired in 1978 and now still receiving benefits at age 94, if that singular example crystallizes the larger phenomenon, that it's simply impossible for those types of plans to be financially viable over the long term.

Salisbury – Well, to respond to two pieces of that. Maybe people today would stop and say the Employee Retirement Income Act of 1974 did great things without commenting on whether that's true or not on a net-net basis. The greatest unintended consequence of that law has been the demise of Defined Benefit pension plans and the rise of Defined Contribution plans. Because one of the principal issues, and I've gone back and read all of the prehistory and there's a very fine book that was published some years ago, a couple of years ago, on the history of the enactment of ERISA called "The Employee Retirement Income Security Act of 1974." It was a history, a political history written by a guy named Jim Wooten and published by the University of California Press. It was telling in a cover quote by a guy named Dan Halperin, who was a tax staffer at the time that ERISA was passed and is now a Professor of Law at Harvard, when he said, "This book is a wonderful, detailed, intensive description of the history of an important piece of social legislation."

The social legislation, part of that legislation was that leading up to ERISA all of the focus was on the absence of benefit portability. It was a focus on these programs doing amazingly positive things for people like my dad who did spend 30 years with one company. But there was a lot of analysis in the 60s about so-called portability losses and one of the most intense debates and also reasons for ERISA was to put into the law vesting standards and to say that these programs that only pay benefits to people that have been there for 20 or 25 or 30 years, well, that's wrong. As you know the most recent legal changes, there are cases now that go all the way down to 3 year vesting and immediate vesting. The moment you went with that change you assured a shift in plan design as you fundamentally changed the cost equation. There is a notion and a mythology out there that in the good old days everybody used to work for one company for a full career, yet, if that had been true, you would not have needed faster vesting. You would not have had portability losses.

When my dad retired in 1978 at that point in history 16% of all workers in the private sector and going into retirement, 16% had been with one employer for 25 years of more. At the most stable, the very most stable companies, you will find of the oldest workers today, maybe 25% of the oldest cohort has been with the company for a full career, but when you look at 1952 to present median job tenure the total labor force has always been about 4 years. That's a long way of saying that we've always been a very high turnover society and a long way of saying that very few of us have ever spent one career, had one long career with any employer. Put that in the sense that you're describing.

Pre-ERISA if 12% of the people that I ever hire will retire from me and I'm doing a Defined Benefit plan and I'm promising 60% income replacement if you spend a full career with me and I fund it over time and everybody that leaves, meaning over time 80 plus percent of the people leave, every dime that I in theory contributed for all of them is going to pay the benefits of that small group that stays. ERISA comes along and says all those people that are leaving and not getting anything are losers and that's wrong. The economists entered, bless them, and said this is a denial of deferred compensation. These plans are deferred compensation which isn't how the companies ever thought about them in any individual worker sense, only in the aggregate. The companies thought about them as something to make sure that people that are still with us after a full career can afford to retire or we can retire them and it's an expense and we're paying people over here and over here on the side, if we do well with investment returns we don't have to make any contributions, so how can there be an individualized deferred wage?

But Congress keeps bringing down the vesting period and what do you keep on doing? More and more and more of the money that's going in is being paid out in small lump sum distributions to this huge number of people who leave between 3 years and 15 years or 20 years, the vast majority of the workers. So the majority of the money in a post-ERISA world, particularly a post-GAT amendment 1990s world, and the vast majority of the money going into a Defined Benefit plan gets allocated to short service people. It is not that my dad lives to nearly 94 or that somebody else lives to 100 that makes a Defined Benefit plan a financing, if you will, problem. If the money going in is going in to pay for those long service people the contribution cost can be low, but if I have the kind of leaking problem that fast vesting introduces, then I end up with a real challenge which requires higher contributions or a reduction in future benefits. That ends up when you think of 1984, the movement to cash balance, that the legislative decision to move the system to fast vesting, essentially imposed on an employer a career average type of contribution with a so-called final pay design. By moving to cash balance and defined contribution or lower benefit formulas cost could be kept low, the money was spread to more people, and the long service worker of the future would get a lot less.

Suddenly I say, well what's that interpret into? For an old fashioned pension plan for somebody who'd been with me for 25 years and was approaching retirement age, if I were to actually put enough money in this year to pay for this year's additional accrual, I'd be contributing 25-30% of their salary in their final pay formula. In this new world of equality you go to 1984 we'll just tell everybody you're all going to get 4.2% and you're going to get 4.2% each and every year. Here's the account and we're going to tell you that's what it is. We've got very fast vesting, and we need to put in money for everybody, and there's very little redistribution of money from the leavers to the stayers, so there is not reward for tenure. Why not just do a Defined Contributions plan rather than all of the expenses and legal requirements tied to the Defined Benefits plan.

So, as opposed to the notion that this change has taken place due to greed or anything else, I look at it backward over my 33 years of involvement and say basically what we've done with the law is we've simply designed something legally that is totally and completely in conflict with what the plans were originally intended to do, which was to provide lifetime income security to the people that were still working for you at retirement age. If I was to put that in a contemporary debate, it is the equivalent to the contemporary debate that President Bush has been most vehement on which is changing part of Social Security from Defined Benefit to Defined Contribution which is the equivalent of saying we really aren't interested in lifetime income security being achieved at the lowest possible cost thorough a group pool that redistributes across individuals depending upon life expectancy.

We're interested in capital accumulation and at the end of the train we want to hand the individual that capital accumulation and give them individual choice and ownership and basically say you decide. This is no longer about lifetime income security in the sense that if we're going to make you take an annuity so there is no conceivable way that you can run out of money before you run out of life. We're saying instead we're going to give you money and you can spend 100% of it in the next 12 months and have 30 years where you're living off of supplemental income and Medicaid, your choice. I don't even view it as a corporate ideology issue if I put it on the President, what we're talking about, has been a general movement away of a theory of community and risk pooling and if you will the redistribution that is implicit in an insurance type of arrangement which is what a traditional DB plan was.

Moving to the individual fight, at this point, in that sense if the only true Defined Benefit plan to be pejorative is the traditional old Define Benefit plan that was an annuity only plan then essentially at the moment there are no Defined Contribution plans that are attempting to mimic the true Defined Benefit plan and in fact a growing proportion of the remaining Defined Benefit system is no longer made up of true Defined Benefit plans. It is made up of hybrid programs that in most cases will produce single sum distributions and will put the decision risk of whether money runs out before life or vice versa on the individual as opposed to that being institutionally protected. You then add a component to that which is what I was getting at with the new company phenomenon. If ERISA had been in effect in 1950 traditional Defined Benefit plans never would have come into existence. The huge growth of Defined Benefit plans that took place in the 50s and 60s would not have occurred because if ERISA's funding standards, most particularly if one were to say what if PPA, the PPA amended ERISA had been in effect in 1950 there would be no Defined Benefit system.

It wouldn't have happened because what the genesis of those Defined Benefit plans was the ability to take people who were at retirement age and give them past service credit of 10, 20, 30 years, accept a huge unfunded liability, and to be able to amortize that unfunded liability over an almost infinite time period and to be able to manage exit of a large number of people, so the people coming back from the war could take jobs, and doing all of it with borrowing against the future cash streams of the firm. Today's PPA would say to those companies, you do that and you've got to fund it off within 7 years and within 7 years you've got to be 100% funded. The economic capacity just would not have been there…wouldn't be there today. If you take that environment and say what started happening in post-1974 and you start looking at those enterprises now on the Fortune 100 and 500 that came into existence post 1974, and I don't mean by renaming or merger, I mean companies that actually grew and came into being through new technology, those companies basically to overstate it, it didn't even occur to them to put in a Defined Benefit plan because they didn't have the primary motivation of the 50s which is a whole lot of existing older workers that we want to entice out the door with a pension.

Macchia – This is again a fascinating historical perspective which is very helpful. I'd like to go back to Social Security for a moment if I could, Dallas, because that is a Defined Benefit structure that is fraught with challenges, as you know. We've seen clearly in terms of President Bush's efforts to introduce privatization and other conversations around Social Security in recent years that any talk of changing that system becomes immediately highly political and polarizing. I wonder what you feel will be the implications of politicians not being able to transcend that, continuing to arguably show a lack of political will, what happens if there isn't the capacity to address some of Social Security's inherent weaknesses?

Salisbury – If we limit Social Security as you're using the words to the retirement program, totally separate and apart from Medicare and the health side of it, the reality which is recognized by the administration and is documented by the actuaries is that the social security retirement program has very little problem. The change in payroll tax that would be necessary to have the program be fine for 75 years is de minimis. The change in benefits that would be required of the change in retirement age, the changes that would be necessary to have the current program sustainable into infinity, are minor changes.

Macchia – Can I just stop you there and explore one aspect of that. If the Social Security surplus is being invested in long term treasuries isn't there a sharp implication for the future in terms of redeeming those and potentially having to lower benefits and raise taxes?

Salisbury – You're describing a general fiscal issue as opposed to an issue of Social Security.

Macchia – But aren't they interlocked to some degree?

Salisbury – The degree to which I personally don't believe that they are ultimately interlocked is that ultimately Social Security benefits being paid on a continuing basis is going to be the difference for almost every working American of whether they watch their parents or their grandparents continue to live decently. Or, they welcome their parents and their grandparents into their home and they start supporting them directly. Then adding a second component. In the last Presidential election, approaching 45% of all votes cast were cast by people over the age of 65. By the next Presidential election it will be pushing 50%. By another 2 – 3 it will be nearly 60% of all votes cast in elections will be cast by people 65 plus, unless younger voters start voting at much higher participation rates. Older voters and one looks at the polling, older people even less than their kids don't want to be dependent upon their kids, they don't want to move back in with their kids.

You end up with this dynamic of will there be the political will even with all of the balancing, will Social Security benefits be paid? Call me an optimist or pessimist or fatalist even with the Social Security trust fund being federal debt securities is I believe that those benefit promises will in fact be paid just given the dynamic of the population and the implications of them not being paid. Especially, and I underline the especially since relative to other issue areas, and I'll use medical as the example, is everybody and it's the majority, about 95%, everybody gets a Social Security check.

Listening to a Congressional hearing yesterday 6% of Medicare beneficiaries account for 50% of Medicare spending. If you end up in a dynamic, a political dynamic of the overage 65 population and their children being put up against the wall and the choice is we can continue to promise you health benefits in the event you get sick or we can continue sending you a Social Security check. Which do you want? Income or a promise of health benefits if you get sick? That's like my trying to convince my employees to take no salary and to take health benefits. Do they want health benefits on top of salary? Absolutely, but if the choice is between income or health benefits, base income, they take income. This year 38% of the nation's retirees have on single income source, it's called Social Security. 64% of today's retirees have a primary income source.

More than half of their income is called Social Security. Retirees 85 plus 62% of their income on average comes from Social Security. If one looks the old curves of health expenditures and health benefits is if you've put it in the terms you're putting it in, ultimately the ultimate trade off decisions of the government, let's assume they were going to filch on the dead, but can we filch on the benefits. If given what the options are, now defense is important to me, but is it more important to me than bread on the table. The health insurance promise is important to me, but am I willing to live in the gutter or in a box in order to have health insurance? No, I'm not, thank you very much. Putting it starkly in your push comes to shove type issue, relative to honoring the Social Security benefits promise, could Congress at some point end up saying okay, beginning 44 years from now benefits will be axed. Conceivable. We're going to raise the retirement age, we're going to match it to life expectancy, and we're going to have CPI minus one. It's what I mean by the issue that the adjustments they could make are relatively minor that would secure the program. I'll use President Bush as the example. If President Bush wanted to secure Social Security as part of his legacy, he could get that through Congress before the end of this calendar year if he was willing to accept changes that did not include individual accounts.

Macchia – By doing what?