Editorial

Retirement Planning

The MIT Human Resources Website allows each employee to study in detail his or her personal retirement benefit profile. The site is well done, an easy-to-use guide to planning retirement options within the framework of MIT’s current retirement system. The MIT Task Force asks us to consider substantial changes to this system: Theme 5 in the Idea Bank entitled “Modern Workforce Policies and Practices,” describes possible changes in retirement options, rules, and programs, all of which are aimed at reducing MIT’s costs. Among them are:

• Revision of the 401k plan [See article in this issue of the FNL]
• Freeze surviving spouse special death benefits
• Eliminate some pension disability benefits
• Change the actuarial method for accrual of defined benefits after 65
• Cap the career-based pension formula in the defined benefits plan
• Change the interest rate used on the cash balance pension plan
• Shift to a 5% cash balance defined benefit pension plan for new hires and current employees with short years of service
• Shift to three-year vesting for new hires
• Eliminate supplemental pension accruals for high-salaried employees
• Adopt an allocation strategy that matches pension assets to liabilities to recover all pension costs on federal grants and contracts.

If we adopt some subset of these recommendations will the machine remain properly geared? It is relatively easy to trace the effects on employees of just one of the changes proposed by the Task Force if all other features of our retirement plans are held fixed and if medium to longer term feedback behavior induced by a change in the structure of the plan are ignored.

However, two or more simultaneous changes can interact in ways that are not immediately identifiable and can lead to unforeseen and unintended consequences – unless someone builds a model of our retirement system that properly accounts for the simultaneous effects of two or more changes in the plan’s structure over a relatively long time horizon and uses it to trace consequences.

There are many, very different stakeholders at play. Each MIT employee wants to know:

• “How does a particular combination of changes in the MIT Retirement Plan affect my current, future, and retirement income stream?”
•  “How do my contributions to retirement change? How does MIT’s contribution to my retirement portfolio change?”
• “How will my benefits at retirement change?”
• “Life is uncertain. Is there any guarantee that restructuring our retirement plans is ‘permanent’? Or can we expect future changes?”

A retirement “program change” can be viewed as a choice of “where to be” in each of the thematic elements of change cited earlier. There are a very large number of choices. Even if we allow only one change in each domain – treatment of the 401k plan or adopting a 5% cash balance defined benefit plan for new hires, for example – our list yields 2¹⁰ = 1024  possible combinations (including the current plan). The number of possibilities is, in reality, at least one to two orders of magnitude larger.

Not only are a large number of options at play here. Our retirement plan is part of a large organizational dynamic system which, in turn, is imbedded in a competitive marketplace for intellectual talent. This has not gone unnoticed in commentary our colleagues have posted about the Task Force recommendations. Here is a sample:

“Reducing retirement and medical benefits is, of course, a reduction in tax-free or tax-deferred compensation. If we believe agency theory, then the best faculty and staff, those that MIT wants to retain, will get market-based compensation. Over time taxable salaries will increase to compensate for the reduction. MIT will book some savings, but the net cost, which is not easily observed, will not be on the books and will be illusory. The slight savings will come as a cost to our culture. It is the people that we will most want to retain who will leave or over whom we will spend time in recruiting wars. We may not be able to attract the people we care about most.

“With retirement benefits reduced, faculty and staff will be less likely to retire…and fewer opportunities for faculty renewal. The bottom line is that, in equilibrium, the proposed savings could likely be, in reality, net additional costs to MIT, but costs that are hidden.”

“Think carefully about the tradeoff between a salary reduction and a reduction in retirement benefits. Reducing salaries by one dollar also reduces Federal and State income taxes by a dollar so that an employee who pays aggregate taxes of 35% is ‘giving up’ only 65 cents. A dollar reduction in after tax benefits costs an employee one dollar. The multiplier effect on investment of tax-free dollars versus taxable dollars over many years magnifies the difference. Of course, we all must have immediate disposable income, but, for most of us, a salary freeze for a substantial period of time sounds decisively preferable to substantial cuts in after tax benefits.”

“I think people, institutions, societies, indeed, countries, are judged by the way they treat the most vulnerable people in their midst. Who are the most vulnerable people in an institution like MIT (apart from the students, whom we all serve)? I think the answer is that in general they are the very young and very old. I hope the administration will consider very, very seriously, the impact on morale across the Institute that reducing benefits, particularly for those approaching retirement, will have.”

“It would be helpful to explain why the Task Force selected this particular example: ‘…a 51-year-old employee with 19 years of service would, if he or she retired at age 65, receive a projected total income (including the basic MIT retirement, 401k withdrawals, Social Security and retiree medical benefits) of 108 percent of his or her annual income just prior to retirement.’”

Back to top

This last “example” may be a case of selection bias: an atypical member of a large population is chosen to illustrate a general feature of the population. The example may or may not be an accurate characterization of most MIT employees. We do not know until a more complete analysis is done. Display of histograms of the ratios of projected total income at retirement to projected income just prior to retirement for (a) non-academic employees for (b) academic employees and for (c) all MIT employees would tell us much more. In turn, account for the effect of length of employment on retirement benefits by partitioning each of (a), (b) and (c) according to length of employment at MIT.

“Our neighbor Harvard seems to offer better retirement benefits: ‘The University provides a number of generous, competitive pension plans. For most staff, once any employee is with the University six months the University contributes to their pension plan retroactive to their hire date. Each month the University contributes 5% of salary below the social security wage base to those under age 40, and 10% of salary below the social security wage base for those age 40 or over. For earnings above the social security wage base, the percentage is higher: 10% and 15% respectively.’

“For employees with five or more years of service Stanford contributes 5% of salary to what is essentially a 401k and matches up to another 5% of salary. In other words, for employees over 40 who contribute 5% of salary, Stanford contributes a total of 10% of salary. This is two times what MIT currently contributes towards retirement.”

Talent has many options. Stanford and Harvard are our principal competitors – and both appear to have more generous retirement plans than we do. There are, as one would expect, dissenting voices:

“MIT’s benefit packages are totally out of control. In most respects, they are even worse than the public sector benefits that are dragging California into bankruptcy. Hire Towers Perrin, or some other benefits consultant and instruct them to do a comparison between other ‘peer’ schools and also between MIT and comparably sized private sector employers. Use this independent study to reduce these ridiculous, crippling benefit packages.”

We recommend that the Task Force engage undergraduates and graduates and give them the task of building a retirement plan model capable of projecting the consequences for any MIT employee – administrator, service employee, assistant, associate, full professor – of a vector of changes in retirement options and plans that span those presented in the Task Force report.

Along with changes in our retirement plans, the Task Force considers major changes in MIT Medical benefits, in employment practices, and in use of space. MIT has been a supportive, stable, and generally equitable working environment for faculty and overall an excellent place to work. We need our current administration to continue to fully engage both faculty and staff in planning and implementing substantive changes, if we are to maintain a collegial, creative, and productive environment.

Editorial Subcommittee
Gordon Kaufman
Jonathan King
Seth Lloyd

	Back to top
	Send your comments