Friday, December 2, 2016

Instead of Making Defined Contribution Look More Like Defined Benefit, ...

I don't think I've ever ended the title of a blog post with an ellipsis before. But, surely, there's a first time for everything.

Lately, the benefits press has written an awful lot about what must be the latest trend in employer-provided retirement benefits -- making the defined contribution (DC) plan look more like the defined benefit (DB) plan. Perhaps I am missing something, but it appears that this "major" initiative has two components to it (that's right, just two):

  • Communication of an estimate of the amount of annuity a participant's account balance can buy
  • The option to take a distribution from the plan as either a series of installments or as an annuity
Let's consider what's going on here.

Annuity Estimate

Yes, there is a huge push from the government and from some employers to communicate the annual benefit that can be "bought" with the participant's account balance. Most commonly, this is framed as a single life annuity beginning at age 65 using a dreamworld set of actuarial assumptions. For example, it might assume a discount rate in the range of 5 to 7 percent because that's the rate of return that the recordkeeper or other decision maker thinks or wants the participant to think the participant can get.

I have a challenge for those people. Go to the open annuity market. Find me some annuities from safe providers that have an underlying discount rate of 5 to 7 percent. You did say that you wanted a challenge, didn't you?

I'm taking a wild (perhaps not so wild) guess that in late 2016, you couldn't find those annuities. In fact, an insurer in business to make money (that is why they're in business, isn't it) would be crazy today to offer annuities with an implicit discount rate in that range.

But, annuity estimates often continue to use discount rates like that.

Distribution Options

Many DC plans offer distribution in a series of installments. Participants rarely take them, however, For most participants, the default behaviors are either 1) taking a lump sum distribution and rolling it over, or 2) taking a lump sum distribution and buying a proverbial (or not so proverbial) bass boat.

Why is this? I think it's a behavioral question. But, when retiring participants look at the amount that they can draw down from their account balances, it's just not as much as they had hoped. In fact, there is a tendency to suddenly wonder how they can possibly live on such a small amount. So, they might take a lump sum and spend it as needed and then hope something good will happen eventually.

Similarly, if they have the option of getting an annuity from the plan, they are typically amazed at how small that annuity payout is. And, even with the uptick in the number of DC plans offering annuity options, the take rate remains inconsequentially small.

A Better Way?

Isn't there a better way? 

Part of the switch from DB plans to DC plans was predicated on the concept of employees get it. They understand an account balance, but they can't get their arms around a deferred annuity. So, let's give them an account balance.

Part of the switch from DB to DC plans was to be able to capture the potential investment returns. Of course, with that upside potential comes downside risk. Let's give them most of that upside potential and let's take away the worst of that downside risk. That sounds great, doesn't it.

Once these participants got into their DC plans, they wanted investment options. I recall back in the late 80s and early 90s that a plan with as many as 8 investment options was viewed as having too many. Now, many plans have 25 or more such options. For what? The average participant isn't a knowledgeable investor. And, even the miraculous invention commonly known as robo-advice isn't going to make them one. Suppose we give them that upside potential with professionally managed assets that they don't have to choose.

Oh, that's available in many DC plans. They call them managed accounts. According to a Forbes article, management fees of 15 to 70 basis points on top of the fund fees are common. That can be a lot of expense. Suppose your account was part of a managed account with hundreds of millions or billions of dollars in it, therefore making it eligible for deeply discounted pricing.

There is a Better Way

You can give your participants all of this. It seems hard to believe, but it's been a little more than 10 years since Congress passed and President George W Bush signed the Pension Protection Act (PPA) of 2006. PPA was lauded for various changes made to 401(k) structures. These changes were going to make retirement plans great again. But, for most, they didn't.

Also buried in that bill was a not new, but previously legally uncertain concept now known as a market-return cash balance plan (MRCB). 

Remember all those concepts that I asked for in the last section, the MRCB has them all. Remember the annuity option that participants wanted, but didn't like because insurance company profits made the benefits too low. Well, the MRCB doesn't need to turn a profit. And, for the participants who prefer a lump sum, it would be an exceptionally rare (I am not aware of any) MRCB that doesn't have a lump sum option.

Plan Sponsor Financial Implications

Plan sponsors wanted out of the DB business largely because their costs were unpredictable. But, in an MRCB, properly designed, costs should be easy to budget for and within very tight margins. In fact, I might expect an MRCB to stay closer to budget than a 401(k) with a match (remember that the amount of the match is dependent upon participant behavior). And, in a DB world, if a company happens to be cash rich and in need of a tax deduction, there will almost always be the opportunity to advance fund, thereby accelerating those deductions.

Win-Win

It is a win-win. Why make your DC plan look like a DB when there is already a plan that gives you the best of both worlds.

Monday, November 28, 2016

Saving for Retirement in a President Trump World

I know this sounds like a politically charged topic, but it's not intended to be. I just want to pose the situation to let readers know a few things. While the Trump platform didn't particularly address retirement issues, how will retirement be affected?

In order to understand this, let's consider a few key non-retirement items that both the Republican-controlled Congress and President-Elect Trump have weighed in on to one extent or another:


  • Repeal of the Affordable Care Act (ACA)
  • Replace the ACA with a framework that is expected to feature competition across state lines, high-deductible health plans (HDHPs) with Health Savings Accounts (HSAs), and a la carte shopping for health plans (you insure what you want to insure to the extent that such coverage is available)
  • Simplified (somewhat) Tax Code with lower marginal tax rates for most taxpayers
  • Elimination of the Head of Household status for filing taxes
As I've remarked many times, our current retirement system for American workers is not what it was, for example, 30 years ago. It's no longer the norm to have the solid three-legged stool of
  • the defined benefit (DB) pension plan from the company you spent most of your career with
  • your personal savings in a high-return savings or money market account (you probably even had your choice of a free toaster or alarm clock when you opened the account)
  • Social Security
Recall that 401(k) plans were in their infancy and even employees who had them didn't tend to make heavy use of them. 

Under the new [proposed] regime, personal responsibility will be king. Out of your higher after-tax income (not significantly higher for most Americans unless the economy booms to where employers are inclined to offer higher compensation to their employees), you'll need to save for retirement and make HSA deductions to accumulate a rainy-day fund just in case you have a high-cost medical expense. Is that practical?

Under the ACA, $10,000 annual health care deductibles are not all that uncommon. So, you'll want to build up your HSA account to ensure that you're not bankrupted by a large medical expense or even by one that you choose to not purchase coverage for. Let's say you choose to defer the family limit for 2017 -- $6,750. Further, since you've been reading about it online, you need to defer, say, 10% of your family income of $75,000 per year (or $6,250 per month) (well above the national median of about $52,000) or $7,500 to your 401(k). Let's add to that your $2,000 per month house payments (including escrow) and your $750 per month car payments (including insurance) and see where you are before basics like utilities, food, and clothing.

We start with $75,000 and let's pull out 22% of that for federal, state, and FICA taxes bringing you down to $58,500. Let's take out another $4,000 per year for health care through your employer (we'll assume they are subsidizing it) and you're down to $54,500. Now subtract your HSA, 401(k), house payments, and car payments and you're down to $7,750. That's $650 per month to pay for food, clothing, gasoline, and to build up a rainy day account, and you haven't even had a chance to buy anything because you just wanted it.

As they said in the movie, something's gotta give. What's it going to be? My guess is that the first thing you cut back on is your retirement savings. You'll worry about that in some future year. Or, will you? Recent history suggests you won't. 

People who retired 30 years ago tended to be much more prepared for retirement despite their lack of 401(k) plans. They were intended to be merely supplemental to employer-provided pension benefits. Is it possible that the Trump Administration should consider the benefits of incentives to get us back into a DB-biased world? Just asking.

Thursday, November 3, 2016

A Modest Retirement Benefits Proposal With Apologies to Jonathan Swift

If you were to ask an employee what are their most important employee benefits, they might tell you that they are their health coverage and their retirement benefits. And, for the last several years, with the passage of the ACA (ObamaCare if you prefer), their health care benefits are fairly well mandated.

Yes, employers have an option -- they can play (provide at least minimum essential coverage) or pay (a penalty for not providing such coverage. And employees not covered by their employers have an option -- they can get coverage (play) outside of their employer either privately or through a federal or state exchange or they can pay a penalty as part of their Shared Responsibility.

To a large extent, the Affordable Care Act came about because of what was viewed (perhaps rightfully) as a health care crisis in the US. Cost of care was increasing rapidly. The number of uninsured was growing. Insurers were perceived to be denying treatment to their insured, not because the medical option being pursued was not the best medical option, but because it was a better business decision for the insurer (the cost of care might include the savings related to the decreased cost of future care). So, for better, for worse, or for some combination, the federal government has largely told us what our health coverage should look like. Depending on your personal situation, you might choose to be the judge of how well that works for you.

Suppose that applied to retirement benefits as well. After all, many would tell us that we are in a retirement crisis as well. Yesterday, I saw some data that while I happen to not believe it said that 50% of Americans over the age of 50 have less than $5,000 in retirement savings. Frankly, if that is remotely close to the truth, we do have a retirement crisis.

Suppose we modeled a retirement benefit system after the ACA. And, suppose this was in addition to Social Security. What would be minimum essential coverage under the Affordable Retirement Act? Let's take a shot at it.


  • Everyone has an account balance into which their employer makes a mandatory contribution and the employee also makes a mandatory contribution. 
  • To the extent that the employee contributes more than that up to a point, their employer must match it. That match is to be more generous for the lowest-paid employees and phases out altogether for employees earning at least $250,000.
  • Withdrawals from the account will not be allowed before retirement.
  • When you do retire, your benefit will be distributed in the form of an annuity with death benefit protection much like a Joint and Survivor Annuity and will have annual cost-of-living adjustments.
  • Those annuities will be provided by insurance companies participating in retirement exchanges. The insurers will not be able to collect data on the participants to whom they are providing annuities, so they will likely not be able to understand the risk they are taking on.
  • Insurers will be limited in the cumulative profit they can earn on these annuities.
  • Employers and or employees who do not participate will have to pay into the system on behalf of others.
How does all that sound? Can you make it work for you?

Special note: there may be as little as nothing in this post that anyone agrees with including its author.

Wednesday, November 2, 2016

Interpretive Guidance Issued, But Pay Ratio Determination Still Difficult

Staff at the SEC recently issued interpretive guidance on the pay ratio rules under Section 953(b) of Dodd-Frank. Regular readers will know that I have written on the pay ratio many times. What many regular readers (those who are used to dealing with benefits issues, for example) may be less familiar with is that the interpretive guidance of staff carries full weight. That is, it's not just a suggestion.

Before digging into the guidance, let's recall what the pay ratio disclosure is. Generally, companies that issue definitive proxies in the US must, beginning with fiscal years starting in 2017, disclose the ratio of CEO compensation to that of the median-compensated employee in the company. And, for those purposes, compensation is that used in the Summary Compensation Table of the proxy. So, it includes (oversimplifying a bit), for example, the value of equity, deferred compensation, and qualified retirement plans provided by the employer.

The employee population for this purpose is not limited to full-time workers or to US workers. So, companies with lots of international employees who are compensated in a variety of different ways may have difficulty with this determination.

If you need a refresher on what the final rule said and how you might handle it, there is useful material here.

Back to the interpretive guidance.

As you may recall, you may determine who your median employee is by using simplified definitions of annual total compensation so long as your facts and circumstances support it. For example, if it's clear that compensation for the median employee will resemble W2 pay, then you can use information from tax or payroll records to determine who the median-compensated employee is. But, if you provide pensions or equity compensation broadly, this may be inappropriate to your situation. In fact, for many companies, determination of who their median-compensated employee is will be the most difficult part of the process.

What staff made clear is that rate of pay (hourly rates or salary) is not reasonable to use as a consistently applied compensation measure (CACM). Staff gave examples where those rates could be part of the process, but in order to make the measure reasonable, companies would need to know how many hours each hourly-paid worker actually worked and for what portion of the year salaried workers were employed.

The interpretive guidance makes clear that in determining the median employee, the population may be evaluated using any date within 3 months of the end of its fiscal year. Once the population as of that date is selected, the employer can go through this process:


  • Identify the median employee using either annual total compensation or a CACM ... by
  • Selecting a period over which to determine that CACM (the period need not include the selected date so long as based on the facts and circumstances indicate that there will be no significant changes (undefined term, of course) between compensation used and actual compensation for the fiscal year)
It is up to employers to determine, again based on all the facts and circumstances whether furloughed employees should be considered in the population. Employees who were hired during the year or who worked less than the full year due to a leave of absence may have their pay annualized. But, part-time and seasonal employees may not have their pay annualized.

And, finally, the staff weighed in on how to determine who might be or not be an employee for purposes of the pay ratio. Essentially, the determination of whether a worker who is not a common-law employee of the employer should be considered an employee for these purposes is based on, you guessed it, all the underlying facts and circumstances. Primarily, the employer should look to whether it sets the compensation of, for example, contract workers, or if that pay is set by someone else.

So, for example, if the company advertises that it will pay contract telephone callers $15 per hour, then they would be employees for these purposes even if they are not employees for tax purposes. On the other hand, a worker who is brought on board through a temporary staffing agency or for a specific contract is likely not a worker.

Well, this clears it up for you, doesn't it? Okay, I thought not. Continue to follow me here for more updates as they come rolling in. Or, let me know if you have questions.



Thursday, October 27, 2016

Analytics and Hiring

Can you use analytics to help you in your hiring process? The conventional wisdom says no. The big data proponents say yes.

Let's consider how this might work at a larger (undefined term) company. Suppose you asked each manager to list all of the characteristics of their best employees. Have them do it with no limits. They tend to be of a particular height, have a particular eye color, and speak loudly. Perhaps they tended to graduate from larger primary state colleges. And, they tended to major in one of three disciplines.

Among the key questions then becomes whether those particular data points are predictive or not. If they are, then depending on what they are, then either before an interview or after an interview, you should be able to put the characteristics of an applicant into your model and determine whether they would be a good hire for you.

You don't think it works, do you?

Let me challenge that. Is professional baseball a job? If you read or watched "Moneyball", you know that one of the early adopters of analytics (they call it sabermetrics in baseball) were the Oakland As under the leadership of their GM Billy Beane. Perhaps his leading disciple was Theo Epstein. Epstein is fairly unique in being the General Manager of the Red Sox who broke the curse of the Bambino and may turn out to be the Cubs GM who breaks the curse of the goat. In any event, Epstein has made highly controversial moves and made two struggling franchises into winners.

I know; that's baseball; it doesn't work in real jobs. Or does it?

Frankly, I don't know. I haven't tried it. But I know people with expertise who think it does.

Try it in your company.

Let me know how it goes.

Friday, September 23, 2016

More on the Public Pension Controversy

The media has fallen in love with the ongoing controversy over public pensions. In the last few weeks, article after article has appeared in major print and electronic media mostly discussing how poorly funded the majority of public pension funds are and placing blame everywhere they can. This is not to say that there are not problems and it's also not to say that there's not plenty of blame to go around, but as in most any controversy, there's more than one side to this issue.

Before delving into it, let me provide some background for those who may not be particularly familiar with public pensions.

Virtually all public pension plans are traditional defined benefit (DB) pension plans in which most, if not all, of the benefit is provided through contributions from the sponsoring governmental entity (some do require significant participant contributions in order to get that benefit). The benefit is typically related to final year's (or the average of the last three or five years) compensation. As a result, spikes in compensation in the final year or years of a career produce much larger pensions. And, when overtime pay is included in those amounts, workers are smart enough to know that their pensions can easily be boosted by getting as much overtime as possible.

As I said, most of these pensions are funded largely through contributions from the governmental entities. Where do those contributions come from? Well, for most of those entities, the very large majority of their budgets come from tax dollars. So, if the recommended, or even legally mandated contributions get too large, there are only two choices -- ignore those requirements or raise taxes (of course in most jurisdictions, future benefits can be reduced prospectively, but that's a different story for a different day).

The current debate centers on public plans in California. What we have recently learned is that the California Public Employers Retirement System (CalPers) values pension liabilities in two ways -- one using what the New York Times referred to as the actuarial value and another referring to it as a market value.

The next two paragraphs briefly discuss a rationale for each basis. For the sake of both simplicity and brevity, both are oversimplified and should not be taken as a precise rationale for either.

The actuarial value discounts obligations using a discount rate at the expected long-term rate of return on plan assets. Proponents argue that this is correct because plans are expected to have a degree of permanence and as assets are invested for the long haul, the obligations that they support should be discounted on that same basis.

Market value discounts obligations using a current settlement rate; roughly speaking, that is the rate at which you could go out to the insurance market to settle those obligations. Proponents argue that in an efficient market, there is no risk premium in investments and therefore, this is the only appropriate basis on which to discount.

What we have read about in the news is that governments that thought that the plans that they sponsor were well funded and that wanted to pull out of the system are learning that to do so will cost them money that they will likely never have.

So, which method is correct? I suppose I could argue that it depends upon which media piece you read. You see, what is happening is that most of the articles have interviewed experts (or self-proclaimed experts) on only one side of the debate. So, they present that side.

The reality is that neither side is correct. It's not as simple a question as the strong proponents on either side would have you believe.

So, here's my take (you knew I would get to it eventually).

Public plan trustees need to understand the true costs of the plans they sponsor. On average, when an employee retires, their pension should be fully funded. This is often not happening. In years when investment returns are good, they use them to reduce contribution requirements. In years when investment returns are poor, they say they can't afford to make the appropriate contributions. It doesn't take an experienced actuary to tell you this is a problem.

I would urge that governments embark on prudent funding policies that build up surpluses in strong years in order to pay for shortfalls in lean years. Doing so will have only minimal effect on the tax base. Studies to understand these issues should be par for the course.

I also urge the popular media to realize that this is not a one-sided issue. While it may make for a great read to tell of the sad tale of a small Citrus District pension plan that is woefully underfunded, it's a small part of the story.

Within the Conference of Consulting Actuaries (yes, I am biased, I serve on the Board of Directors), there is a Public Plans Community that regularly discusses issues such as this. For those who are interested in a view of the problem from people who understand the issues, it's an excellent read.

Monday, September 19, 2016

Lessons From California Public Pensions

Over the last few days, the print media (at least we used to call it print media) has hit hard on public pensions in the State of California. The New York Times hit hard on the differences between the "actuarial approach" and the "market approach." The Los Angeles Times took on a pension deal from the late 90s. Both of these are symptomatic of the issues that all of taxpayers, legislators, workers, and actuaries face in the public pension world.

Let's take a step back. Some of the most generous of all public pensions are those available to public safety officers primarily police and fire. I could give you lots of reasons why this approach is correct and why it's not. You might disagree with my analysis on all of them, but that's not what's important here.

Historically, people who have chosen careers in police and fire have thought of their careers differently than people in most other professions. Those careers are very risky and they are physically demanding. Many in those professions would tell you and me that lasting more than 30 years or so is just not practical. And, if we take that as a correct statement (I do), then it is reasonable that such public safety officers be eligible to retire from that career earlier than we would expect in most other careers. After all, if you were trapped in a burning building, would you be happy if the people trying to save you were just trying to hang on until normal retirement age at 65, but weren't really physically capable of handling such a demanding task?

Over time, states, cities, towns, and other governmental organizations found the answer. Provide those public safety officers with a significant incentive to retire early (compared to other careers) and if you do that, you don't have to pay them all that much. So, what that has historically achieved is that costs for current public safety employees have been relatively lower and costs have been deferred into their retirements.

That's a problem. It doesn't need to occur. The correct answer for a governmental employer, or other employer, is to realize that deferred compensation (that's what a pension is) is earned during a employee's working lifetime. Therefore, it should be paid for during that working lifetime. After all, once an individual in any profession retires, they no longer provide a benefit to the organization that they previously worked for. Further, the burden to pay for those pensions should not be passed on to future generations of taxpayers.

In the past, I have commented about various actuarial cost methods. An actuarial cost method is a technique for allocating costs to the past, the present, and the future. Looking at things as any of a taxpayer, legislator, employer, or employee (I happen to not be all of those, but even so, I can put myself in the shoes of those that I am not), the correct answer has the following characteristics:

  • An employee's pension is paid for (funded) over their working lifetime. Once they retire, the cost of providing their benefit is over. (Understand that actuarial gains and losses make this an inexact science, but we should be close.)
  • The cost of providing that employee's pension should be level. That is, it should either be a constant percentage of their pay or a constant dollar amount. As an employer, I can budget for that. 
Let's consider an example of that second bullet. Suppose I pay a public safety officer $60,000 per year (I know -- in some jurisdictions that seems high and in others it seems low) in salary. Further suppose that their deferred compensation costs me 10% of pay annually. Then when I am budgeting for that person, I know how to budget every year. If their pay goes up by 5%, so does the cost of their pension, roughly. This year, I budget $66,000 for current plus deferred compensation. Next year, with that 5% budgeted increase, I budget $69,300 for current plus deferred compensation.
There is an actuarial cost method that does exactly this. It's called Entry Age Normal (EAN). When I entered the actuarial profession back during the days when we used green accounting paper rather than spreadsheets, in my experience, EAN was the actuarial cost method of choice. But, it had its downsides. 
  • Neither the accounting profession nor the federal legislators accepted it as the method of choice.
  • Employers were advised that their current cash cost would be lower using a different actuarial cost method. It's easy to say you will fix that problem later on.
  • Observers understand a method where you pay for benefits as they accrue, but nor one in which you pay for benefits as they are allocated by actuaries.
So, now we have come to a crossroads. Many of the largest public pension plans are horribly underfunded regardless of how you determine funding levels (some have been funded responsibly; others have not). Getting them well funded requires cash which can only come from increasing taxes or from taking money from elsewhere in the budget (dream on). Legislators want to get re-elected which means you don't raise taxes. 

Hmm, I see a problem here.

The problem extends to private pensions as well, but there are  good solutions there. Since EAN is not available as an actuarial cost method anymore (we could choose to have our valuation done using the legally prescribed Unit Credit actuarial cost method, but fund not less than the EAN cost although that is very rarely done), we need to look in other places. 

Plan design is an excellent lever in this regard. Suppose we had a plan design  that even under a Unit Credit cost method allowed us to achieve exactly what we are talking about here. And, suppose that design allowed for all the benefits of defined benefit plans (DB) including market-priced with no built-in profits annuity options, professional investing, no leakage, portability, and virtually no cost volatility. Wouldn't that be an ideal world?