Among the more detailed features of the RIOpenGov.org pension page is the comparison of pensions against final average salaries. A push of the button on the “detailed search” tab shows there to be 3,298 pensioners whose retirements are paying better than their jobs ever did.

These retirees are collecting $20,903,314 more from the state than they did in the final years of their careers. On average, they have been retired for 18 years and still have an estimated 13 years to go — having retired at 56. For all retirees, the corresponding averages are 13 years retired, with 15 to go, having retired at 58.

Interestingly, the number of pensions above pay does not match up with the number of years retired as well as one would expect. Apart from a huge burst of retirements in the early ’90s, the trend line for total retirements isn’t surprising; the farther from retirement pensioners get, the fewer of them there are:

Viewing pensions that exceed pay as a percentage of total pensions, however, yields a more unexpected result:

As a matter of simple arithmetic, it would make more sense for the percentage to continue to grow as the years go up, thanks to cost of living adjustments (COLAs).  That the result does not match expectations merits further investigation, but for the time being, suffice it to say that such peculiarities give some credence to suggestions that pension deals of the past were inordinately generous and should be pulled back.

Among retired public-sector workers whose pensions exceed their working salaries, the average pension is $45,815 per year, which is $6,338 more than the average final salary. For the pension system overall, the average is $31,388, which is $12,021 less than the average final salary.

At this point, it makes sense to look at differences in the results based on the retirees’ former employers, and clearly, not all pension-granting public entities are equal. Of the 184 public employers, 81 have retirees who have surpassed their salaries. For 43 of the employers, more than 10% of all retirees are in this group.

With some notable exceptions, schools in general and teaching staffs in particular dominate the list for retirees who make more in retirement than while working. And teacher pensions tend toward the higher end of the dollar scale.  However, there is no apparent correlation between the percentage of retirees over-earning their working salaries and the average number of years that each entity’s pensioners have been retired:

National Education Association of Rhode Island President Larry Purtill is gaining some traction with his claim that having teachers retire later will hit local communities in the payroll. As described in an October 25 Providence Journal story by Kathrine Gregg (not online):

Sen. Walter Felag, D-Warren, asked the question that Larry Purtill, president of the National Education Association of Rhode Island, has asked on YouTube and elsewhere: how will city and town taxpayers be able to afford all of the top-scale teachers who would now be forced to work until they are 67 years old?

A top-step teacher averages $70,764 a year in Rhode Island; a first-year teacher averages $39,087, according to Purtill.

Purtill has not yet had his turn to testify before the legislators, but in other venues he argues: “If a teacher is forced to go from 62 to 67 to retire and would have been replaced by someone on Step 1, that is an additional $130,479 for one teacher over that five-year period. Has anyone thought about this, and figured it into the actuarial studies? I am assuming that unless state aid increases by a drastic amount, this cost will go directly to the cities and towns.”

There’s a reason Purtill stops his cost clock at five years: because after that point, the older teacher retires and a replacement is hired at step one. Obviously, that teacher will be making less, to start with, than a teacher hired five years earlier, so the scenario of a later retirement actually begins to save the district money. And the district continues to save money for the ten years it takes the new teacher to reach step ten, herself.

To put a number on this dynamic, I averaged all of the salary steps compiled by the Rhode Island Association of School Committees for the 2011/2012 school year. For easier comparison, I then removed the few districts that spread their steps over more than ten years, leaving me with 14 cities and towns. (To check for any distortion, I compared these districts to the broader average available for 2010/2011 and found them to be about 0.1% higher, overall.) Using these averages, I figured out the year-to-year step raises and factored in the 4% increase that the pension actuaries estimate that the overall step system receives each year, on average (inflation plus general increase). By this method, I found step one to be $39,175 and step ten to be $72,393.

It is true that a local district will therefore pay an extra $141,475 for five more years of the older teacher’s salary, compared with a replacement starting at step one. However, seven years after the delayed retirement, the savings in waiting to hire her replacement have erased the extra costs. By the time the later replacement reaches step ten, the district turns out to have saved $44,990, and that number holds for decades, until the year that the earlier replacement would have retired.

Remember that this is savings to the district on salary alone. Most districts offer retirees some sort of health benefit, which will erode the initial salary savings of earlier retirement. What’s more, a good number of districts only pay retiree healthcare until Medicare age, so pushing retirement beyond that point would eliminate the healthcare portion of districts’ other post-employment benefits (OPEB) liability completely.

An important note arises if we expand the inquiry to figure out the cost/savings of later retirement in terms of the older teacher’s pension. Such an analysis is difficult to perform, because changes to the retirement system over time have left a number of categories into which employees can fall; moreover, teachers can retire at various ages, with varying years of service, so an accurate comparison would require teacher-by-teacher analysis.

For illustration purposes, however, I assumed normal retirement at age 62 with 29 years of service on a non-grandfathered Schedule A. (That means the teacher receives the more generous percentage of pay that Schedule A provides but has the retirement benefit calculated as an average of his or her highest five years, not three years, as would be the case for a grandfathered Schedule A pension.) For the second scenario, I added five years to this teacher’s career, but followed the proposed pension reform such that the teacher only accrues an additional 1% of salary for each additional year of work. I did factor in cost of living increases for both scenarios, but I used the reformed rate, which ties increases to pension fund performance.

The upshot is that the later retirement saves the retirement system about $54,000 over the life of an individual teacher. But those savings only materialize under the terms of the proposed reform. Applying a later retirement to the current system, without other reforms, winds up costing money in the scenario I’ve described, because the starting pension rate and higher cost of living adjustments (COLAs) increase the total benefit by so much.

Whatever the case, under the current system, the teacher is working for 29 years and retiring for 25. Under the reform, he or she is working for 34 and retiring for 20. Put differently, under the current system (with all of my assumptions), the older teacher will be collecting retirement for almost the entire career of his or her replacement. Essentially, for just four years will the system be paying for only one teacher for one teaching job. Under the reform plan, that differential increases almost fourfold, leaving fifteen years of the replacement’s career unburdened by pension costs for the older teacher.

One needn’t run the numbers to get a sense of just how quickly Rhode Island and its cities and towns — collectively and individually — will save money.

The only argument that I’ve heard against my suggestion that the proposed hybrid pension system will be more expensive than the current system is that the hybrid offloads market risk onto the employee, alleviating the risk to the employer (i.e., us, the taxpayers). Moderate Party founder and gubernatorial candidate Ken Block made the point on Anchor Rising, and Gary Morse, pension adviser to the Republican who was almost governor, John Robitaille, called in when I was on Friday’s Dan Yorke Show to make a similar point.

My response has been to suggest that, while the objection may be true, the entire pension reform relies on the 7.5% rate of return that the state’s actuaries have assumed. As a reminder, the current system is a defined benefit plan, which promises employees a certain retirement package; the hybrid system would decrease that benefit but put money into a 401(k)-style plan for each employee to fill in the gap. Right now, the major problem is the liability for defined benefit promises already made, and the hybrid doesn’t make those go away.

Unfortunately, I haven’t come across “what if” data from the actuaries that allows a real comparison of the system’s health with different rates of investment return. One can, however, infer the effect of the Retirement Board’s recent lowering of the return rate assumption from 8.25% to 7.5% from the latest actuarial report. Specifically, the Board dropped the rate of return by 9%, and the normal cost of an active employee’s pension (that is, the percentage of payroll that must be put aside to fund an individual employee’s retirement) went up by 22% for state workers and 18% for teachers.

For an accurate “what if” scenario, the actuaries would have to apply the different assumptions to their models, but purely for the sake of illustrating my point about the cost of the hybrid plan, I’ve assumed that same return-rate-to-normal-cost ratio applies consistently to the equation. If that’s accurate, then the hybrid plan doesn’t cost the taxpayer less than the current system unless the market returns less than 5% on investment:

The solid lines show the effects of an under-performing market on the current defined-benefit plan; the dotted lines show its effects on the hybrid plan. And, yes, the lines do cross eventually. However, under those circumstances, the total cost for the pension system to the employer, even under the reform, will be well above the annual cost that currently has everybody in a panic — over 60% of payroll for teachers and nearing 80% of payroll for state workers:

Of course, the solid lines, which track the current system, show just how scary the situation will be if the General Assembly does nothing, and a defined contribution component ought to be part of the solution (if not all of it). That doesn’t mean, though, that reformers should take the system on the table just because it incorporates some worthwhile concepts. If this reform passes, it’ll be years before the panic level rises sufficiently for further action by elected officials, and by that point hundreds of millions of dollars will have been siphoned into untouchable defined-contribution accounts.

The complexity of the pension problem comes in the multiplicity of angles from which the numbers can be presented. When it comes to the cost of current employees’ pensions, General Treasurer Gina Raimondo’s hybrid pension system will cost more than the system that’s already strangling Rhode Island.

The headline grabber — because it’s such a large figure — has been the unfunded liability. Essentially, that’s the difference between the benefits promised over the amortization period and the accumulated resources expected to be available to pay them.

It’s critical to realize that both sides of that equation are predictions. General Treasurer Gina Raimondo induced her recent shock to the pension system by leading the Pension Board in a reduction of the predicted annual market return on the system’s investments of just 0.75%, from an 8.25% expected return to a 7.5% expected return. In tweaking its assumptions, the board effectively increased the unfunded liability of state worker and teacher pensions from $5.40 billion to $6.83 billion, according to the most recent actuarial report. The overall problem, however, requires predictions about everything from a female desk clerk’s life expectancy upon retirement in twenty years to the average raise that RI school committees will give to their teachers over the next several decades.

Although interested parties like to pick and choose from among them, various factors have contributed to the existence of the liability. Given the lag time between a particular year’s pension payments and its full effects on the system, elected officials have had plenty of incentive to promise future benefits that outstripped the budget dollars that they were actually willing to put aside in the present, and union leaders have been happy to play along, expecting those promises to be fulfilled somehow, someway.

Another factor has been the market. Even responsible officials, who put aside every penny that the actuaries called for would have wound up underfunded, because the actuaries were basing their predictions on an 8.25% return, when the system has only realized 2.51% per year, considered over the past five years, 2.28% over the past 10 years, and 6.57% annually since 1995.

And then there are all those predictions. The actuaries currently expect a female teacher who retired last year to live for 24.2 more years. They predict that the same teacher retiring in 2030 will only expect an additional 1.1 years of life,  or 25.3 years of retirement. With the rate of advancement in medical technology, does that really seem plausible?

The incentives and unpredictability of pension planning are what make defined-contribution option so attractive. With a defined benefit, the employer promises employees a certain amount of future income, and it is up to the employer (or, in the public sector, the taxpayer) to make sure that the money is there; with a defined contribution, the employer promises to put a certain amount of money aside, and future income will depend on the plans and investment returns of the specific employee. The dollar amounts may turn out to be exactly the same, but it isn’t a “liability” in that the employer hasn’t pledged to make up the difference for unmet expectations.

Leading up to the release of Treasurer Raimondo’s pension reform proposal, advocates for public workers and retirees have been downplaying the importance of reducing this liability by separating the theoretical cost of pensions for current employees from the amount that municipalities and the state have to put aside to catch up on the current debt. The mechanism for this argument is the “normal cost” of a current employee’s pension — that is, the percentage of payroll that must be invested in order to fully fund pensions as if the past liability did not exist.

For teachers, the “normal cost” is now pegged at 11.82% of payroll, and since the teachers contribute 9.5% from their own paychecks, the city and state combined only have to put in another 2.32%. That’s what NEA-RI Executive Director Robert Walsh meant when he wrote in the Providence Journal that “for the vast majority of existing employees, the facts support no further changes” to the pension system. Get over the liability hump, in other words, and it’ll be clear sailing for public-sector retirees.

The question that hasn’t been asked, yet, is what Raimondo’s hybrid defined-contribution/defined-benefit plan will do to public costs when calculated in these terms. The answer is that it will actually increase them, as the following table shows.

Reformers should be wary of two facts that emerge from these eye-glazing numbers:

1. Five percent of current retirees’ contributions will be entirely insulated from any problems that might arise with the defined-benefit component of the system — whether they be low investment returns, proof of unrealistic predictions, or a failure to meet recommended funding.
2. Combining the defined-benefit and defined-contribution payments, the amount of money that state and local governments will spend on the retirements of current employees every year will increase — more than doubling until such time as grandfathered employees leave the system and continuing on at a higher rate even then.

This means that the savings of the overall pension reform rely entirely on changes to the existing system (e.g., COLAs and retirement age), because the hybrid will cost tens of millions of dollars more every year… even when the “pension crisis” is completely resolved and even if the assumptions prove accurate. And changes to the existing system are precisely the controversial elements that have battle lines being drawn.

With all the lines on the field and the questionable allegiances, taxpayers should be sure that somebody is defending their interests.