Retirement 20/20 Update—The Measurement Framework
Cindy Levering, ASA
Actuaries have been thinking about how to improve North American retirement systems in big and small ways for a number of years. Read Bob Berin’s thoughts from the September 1990 Pension Section News, followed by Cindy Levering’s update on the Retirement 20/20 initiative.
Pension Rules—What If?
By Bob Berin
I was asked, as part of a meeting in another country, to "consider a blank sheet of paper and broad-brush a viable private pension system.” Although this was an interesting challenge, nothing came of it because a private pension system requires free investment markets, and this requires a free enterprise system and, ultimately, a democratic form of government.
In the United States, we seem to have lost sight of the original goal of a private pension system: the encouragement of pension programs so that workers and their families can live in dignity in retirement.
One reason we have lost sight of the original goal is because the government is trying to raise tax revenue to meet the budget crisis. However, private pension plans designed to maintain adequate retirement income relieve the pressure on Social Security and other public programs dealing with poverty in retirement. With an aging population, retaining this goal may be more cost-effective than the additional amount of tax revenue presently collected.
A second reason is the complications that have been written into tax law which discourage both new plans and improvements in existing plans. One treasury official recently said “There is little disagreement that pension law is one of the most complex areas of tax law.” This needed to be said. He went on to explain that a “plain English approach” had been adopted “to make the regulations understandable.” This was necessary and it was helpful. This official should be applauded, but the next step (of which he is surely aware) is even more difficult, and may never happen.
Infinite Number of Rules
Reading the proposed regulation for IRC Section 401(a)(4) leaves some room for wonder at the need for such enormous detail. Are simple solutions possible? Or are we dealing with rules to curb perceived or actual abuses that range from the remote, to not likely, to possible, to plausible, etc.? And who can administer such rules? A grid of such rules can never be complete enough and the gaps create uncertainty and the need for more rules. This creates the need for an infinite number of rules which is not possible. Is there any other approach?
Reporting and disclosure are extensive and complex. Even larger companies have difficulty; the cost of administering the programs grows. Benefit design now relates to questions of compliance rather than principally on what may be good practice. Excessive paper work and costs of compliance are seen as burdens, no longer minor in scope.
We seem to be losing sight of the need to encourage retirement benefits. Tax simplification seems to lead to more complications. Could we ever undo the past and start over?
Here are some thoughts on starting over on the actuarial side (not that we could).
- Pension benefit formulas apply uniformly to all employees. At retirement the replacement ratio of plan benefit plus actual primary Social Security benefit divided by final average salary must decrease with increasing salary and be less than 100 percent.
- Test the funding level at each valuation date and make this public information: (a) Unit Credit accrued liability (valuation interest and mortality assumptions) less (b) Market Value of assets, as well as (b) divided by (a). (Final pay plans use current salary, as a proxy, and state so).
- Disclose the relationship between the market value of assets and the amount required to purchase annuities for vested benefits from a life insurance company.
- Employ one of the recognized actuarial funding methods in the regular valuation. Use liberal and conservative actuarial assumptions and fund increases and starting obligations over a 20-year period to determine a range of contributions as a percentage of total payroll. These calculations should use a full set of actuarial assumptions. (The choices of liberal and conservative assumptions are to be discussed in the valuation report and are the actuary's responsibility.)
- Disclose three percentages of total payroll: liberal contribution, conservative contribution, and actual contribution. The last must be between the first two. The maximum contribution is to be 15 percent of total payroll
In establishing a private pension system, the guiding principle should be to keep it simple in every respect and act on behalf of the participants. This is also best for the sponsoring employer and for government.
And everyone involved in the process must realize that someday, if they are fortunate, they will be retirees.
Bob Berin, FSA, AIA , is Managing Director, Chief Actuary for William M. Mercer, Inc. in New York. He is a Vice President of the Society of Actuaries.
As anyone who has ever built anything knows, measurement is critical. As part of the Retirement 20/20 initiative, we realized early on that we needed to benchmark any proposals that we developed for a new retirement system, as well as the proposals of others and existing systems. Only by benchmarking can we determine the extent to which current systems meet the needs of stakeholders, and how various proposals compare to the status quo and to each other. Measurement is an essential part of the debate.
Attendees at the 2006 Retirement 20/20 conference—Building the Foundations for New Retirement Systems—helped us to understand and enumerate the needs, risks and roles of key stakeholders in a retirement system. We realized that these principles could be used as the basis for our benchmarking tool, which we’re calling a Measurement Framework. This framework will, we hope, also serve as a tool to help us communicate our findings to pension actuaries and other retirement professionals, as well as other stakeholders (including the media) in order to build grassroots support for our work.
Development of the measurement framework is well underway. We began by using it to evaluate a traditional final-pay single-employer defined benefit plan as well as a traditional defined contribution plan. This was done primarily as a way of testing the framework, and to help stakeholders understand the framework’s operation by using it to evaluate well-understood retirement plan designs. We’ve identified several other retirement system models that we’d like to evaluate with the framework (for example, the TIAA-CREF program in the United States, the Dutch retirement system, the ERISA Industry Committee proposal, and the YMCA Retirement Fund which was recently highlighted in Plan Sponsor magazine). It also might be interesting to flesh out the details in Bob Berin’s model from 1990 (outlined in the article above) and to use the Measurement Framework to evaluate these early “blank sheet of paper” ideas.
The framework focuses on how well each proposed design addresses the needs, risks and roles for each of the four stakeholder groups: society, individuals, employers, and the markets. The framework focuses in detail on those needs, risks and roles by devoting a section to each stakeholder group. The framework also develops a “summary rating” for each stakeholder group. These four summary rankings are shown in an executive summary, together with ratings related to four of the themes that came out of the 2006 conference and that we continue to explore:
- Does the plan self-adjust to meet changing economic and demographic conditions?
- Does the plan align stakeholders’ roles with their skills?
- Does the plan consider new and emerging norms for work and retirement?
Does the plan align with markets (use market hedging and pooling mechanisms effectively)?
Additional discussion of these questions and the needs, risk, and roles of stakeholders can be found in the 2006 conference report at www.retirement2020.soa.org.
Instead of rating each criteria numerically, we use a five-level color coding system to assess where the plan meets the principle (green), violates the principle (red) or falls somewhere in between (yellow-green, yellow, red-yellow, ranging from better to worse). We’re not trying to be scientific with the ratings, but to show directionally whether these proposals meet or don’t meet the goals we’ve identified in the Retirement 20/20 initiative. One clarification: we’re using the framework to evaluate individual designs, be they for employer-sponsored plans, individual savings mechanisms, or social insurance systems. A retirement system as such is a combination of these things. It would be too complicated to try to use the framework to evaluate the system in totality. But, the framework can help us determine which combination of individual designs might do the best job of ensuring a strong retirement system for the 21st century. One thing we’ve learned in the Retirement 20/20 journey is there’s no single design that can satisfy every stakeholder in every situation. That’s why we also intend to develop a “summary of summaries” as the final element of the measurement framework. This will show how different plans compare across the board in meeting the needs of stakeholders and the four principles/questions outlined above. This should help us to identify which combination of designs will do the best job of meeting the retirement needs of tomorrow.
If you’re curious how our DB and DC plans fared, the traditional final-pay single-employer DB plans do a much better job of meeting the needs of society and individuals than they do of meeting the needs of employers, and the results with respect to meeting the needs of the markets are mixed. The results for traditional DC plans are mixed for all stakeholders.
We recently instituted an advisory group to give input and suggestions on what we’ve done so far with the measurement framework and to give us guidance on the best way of using this tool going forward. Volunteers will be involved in evaluating the designs noted above, and the advisory group will suggest additional models to “run through” the framework. We will post results on the Retirement 20/20 Web site and provide you with updates via the Pension Section News, as we make progress in this area.
In the Chairperson’s Corner of the January 2008 issue of PSN, Martine Sohier stated “the end of the defined benefit era may be what it takes to bring the pendulum back eventually, or to push it to somewhere else—to a system that might be better able to respond to the evolving needs of today’s and tomorrow’s workers.” As noted above, we understand that it’s probably impossible to come up with a single design that will address all or even most of the needs of all stakeholders. Our ultimate goal is to find designs that minimize the amount of “red” and “red-yellow” results in the framework tool, and to offer combinations of elements that form a system that ensures that all needs and risks are satisfactorily addressed for all stakeholders.
In his submission brief to the Ontario Expert Commission on Pensions in October of 2007, Keith Ambachtsheer (founder of KPA Advisory Services, publisher of the Ambachtsheer Letter, and a panelist at the 2007 Retirement 20/20 conference) talked about the concept of “Moral Capitalism” as it applies to retirement systems:
Even over 200 years ago, Adam Smith knew that fostering “the public good” would require more than just applying the rule of law and maintaining unfettered free markets in goods, services, capital, and labor. Sometimes intelligent intervention is required. The time has come to apply this moral sentiment to the design and implementation of systems that can reliably provide an adequate level of retirement income to today’s and tomorrow’s working populations.
As actuaries who understand the concept of risk and are trained at solving problems, we are uniquely qualified to lead this discussion. This is at the core of the Retirement 20/20 mission. We hope the measurement framework will serve as a cornerstone in accomplishing this purpose.
Please contact Cindy Levering at firstname.lastname@example.org, Sandi Kruszenski (Pension Section Chair) at email@example.com, or Emily Kessler at firstname.lastname@example.org, if you’re interested in working on either the Measurement Framework project or any aspect of Retirement 20/20.
Cindy Levering, ASA, EA, MAAA, is a senior vice
president with Aon Consulting in Baltimore, Md. She is also co-vice
chairperson of Pension Section Council. She can be reached at email@example.com.
Retirement 20/20 is the Pension Section’s initiative to rethink retirement systems. The goal of Retirement 20/20
is to consider what’s possible, beyond the limitations of what’s happened historically or what is in today’s tax code. For more information, visit www.retirement2020.soa.org.