Controlling longevity risk is achievable, but clients have to decide what they're willing to give up to get it.
This article first appeared in the December 2010/January 2011 issue of Morningstar Advisor magazine. Get your free subscription here.
Choosing an approach that ensures clients won't outlive their retirement savings is no easy task. To manage longevity risk, advisors and their clients must weigh an incredible number of factors, including, ultimately, one that is unknowable: how long a person is going to live. The financial industry, however, is on the case, rolling out products and innovations every year that offer new ways to guarantee investors a steady stream of retirement income.
To get a handle on longevity risk and the best ways to control it, we asked three leading experts in the field to discuss the challenges advisors face in finding an optimal solution. Moshe Milevsky is a finance professor at the Schulich School of Business at York University in Toronto and author of Are You a Stock or a Bond? Create Your Own Pension Plan for a Secure Financial Future (FT Press, 2008). John Ameriks is a leading researcher at Vanguard, where he heads the firm's Investment Counseling & Research group. Thomas Idzorek is chief investment officer and director of research at Ibbotson Associates. Our conversation took place Oct. 18 and has been edited for clarity and length.
Christine Benz: I'd like to start by talking about what you think belongs in investors' tool kits when it comes to managing the risk that retirees will outlive their assets.
Moshe Milevsky: When you're accumulating wealth toward retirement, mortality as an asset class doesn't really come into play. Insurance is something you purchase to protect your family. Nobody really thinks of it as an integrated asset-allocation decision. The same thing happens with downside protection. I don't really view put options as something that I'd use on a regular basis to grow my wealth as I'm getting closer to retirement.
But once I start to withdraw money from a portfolio, both of those product classes--downside protection in the form of put options and synthesized puts and longevity insurance, which is essentially a short position on mortality rates--start to play a role in asset allocation. In other words, your health and mortality become an asset class that you can invest, leverage, and perhaps even arbitrage. So, I would start with traditional linear instruments that we're all familiar with, plus an augmented set that essentially consists of derivatives on mortality and other asset classes.
John Ameriks: I think standard asset classes are going to serve people well throughout both accumulation and distribution. Once people get a little older, once they stop thinking about how much life insurance they need and begin to worry about outliving their money, then they need to think about what kind of instruments might be appropriate.
I want to point out that most people start off in retirement with a pretty big annuitized chunk of wealth in the form of Social Security. So, it's not really about needing to do things that are new. It's just thinking about what you have, and whether it's sufficient to meet your needs, and then trying to decide whether additional longevity protection--an immediate annuity or a company pension plan--is something that makes sense in your situation.
Milevsky: The effect of Social Security really depends on what wealth decile you're in. At the lowest decile, 95% of people's total balance-sheet wealth is in Social Security. In some sense, they're overannuitized. In the top decile, only about 10% of people's total balance-sheet wealth is in Social Security. So, you really have to slice it depending on people's level of wealth. If you just look at the average, I'd agree--a lot of people are overannuitized. But if the typical advisor is advising someone who has $500,000 to $1 million in net worth--and they don't have a defined-benefit pension--you're not looking at the average anymore.
Ameriks: That's a really important point, and it gets missed in the policymaking discussion about annuitization. The interesting financial-planning issues, and the place where most guidance is needed, is in that group of folks that Moshe mentions--$500,000 to maybe $1.5 million. They have been wealthy enough to accumulate a substantial amount of money and they've had an affluent lifestyle. Social Security, yes, will provide a backstop and floor for them, but it may be well below what they've become accustomed to. It's in these circumstances where the idea of needing additional protection of the downside comes into play.
Milevsky: As I like to tell people, retirement-income planning can't be done with an elevator speech.
Ameriks: What's difficult about the retirement-income phase is that it's at the point when you try to articulate exactly what you need it for. The answer is going to dictate the approach. I'm not so sure that we need fancier products and technical engineering. What we need is a better appreciation that there are tools that are far better at achieving certain outcomes than others. There's a lot more that needs to be learned about exactly what someone's trying to achieve before a portfolio can be put together.
Milevsky: You have to ask your clients different preference questions in retirement-income planning as well. How much do you love your kids? That's part of the risk tolerance. Is a bequest or legacy a big part of your portfolio? That's going to make a difference on product allocation. How healthy are you? Are you healthier than average or not? That's part of product allocation that goes beyond asset allocation.
Ameriks: Or how much do your kids love you?
Milevsky: Exactly. Are they going to backstop your retirement if something goes wrong? Are you moving in with them? These are very uncomfortable questions to have for a stockbroker who's used to asking about stocks and bond mixes. Traditional insurance-only advisors probably have an advantage with these conversations, but the asset-manager advisors will have to catch up. You can't eat alpha in retirement.
Benz: All of you have done research that points to some form of annuity as being beneficial to retirees' portfolios. Which types of products do you think make the most sense within this framework?
Thomas Idzorek: The problem with talking about annuities is that there are so many different flavors. There are indexed-linked annuities. There are immediate annuities in both the fixed and variable flavor. Then, there's the deferred variable annuity, now typically with a Guaranteed Minimum Withdrawal Benefit for Life on top of it. And there's the pure longevity insurance, which I never know what to call--some people refer to it as a deferred immediate annuity or an advanced life deferred annuity. We think that immediate annuities, deferred variable annuities with the GMWB for Life, and the pure longevity insurance all could play an important role in a retirement-income solution.
Milevsky: I've gone beyond labels and names. I think that the word "annuity" is almost meaningless. You have to do a DNA analysis on an annuity before you even consider it. "GLWB" may sound good, but if they're promising 3% for life and they're charging you 400 basis points a year in all-in fees, I don't think it belongs in an optimal portfolio. And vice versa with immediate annuities. If people are going to start demanding refunds and 30-year payment certains and all these things that eliminate the mortality credits, then I'm not sure immediate annuities make a lot of sense. They may just be a bond substitute.
The term "annuity" is a very broad term and misunderstood. What we need is some sort of nutritional disclosure of what the annuity really contains, so that we can eliminate some of this confusion. I would like to see a DNA analysis done of any annuity product before it's given to the public. Then, we can understand the magnitude of mortality credits, what's the fixed-income component, how much downside protection we're getting. Are we really getting inflation protection if we need the market to step up? A lot of these annuity products do not belong on the shelf. Just because a lawyer or a state insurance regulator classified it as an annuity doesn't necessarily mean an economist will. In fact, I dare say that many financial economists would be shocked at how little annuity there is in today's annuity.
Benz: Let's back up and talk about how an advisor would match an investor with the right type of product. How does the advisor figure which type of annuity is appropriate?
Ameriks: The key question is, Is a retiree fundamentally comfortable with the idea of giving up access and liquidity to a portion of what they've saved in exchange for a guaranteed flow of income, no matter how long they or their spouse lives? If somebody right away says that they are uncomfortable with that, I think advisors have to take a step back and make sure that people really do understand the trade-off. That's really the crux of things. For some people, it's a deal that makes sense, and for some, it doesn't.
A lot of people have missed the boat on annuities because they had a very hard time giving up liquidity and making this "bet" that they'll live a long time. I think some conversations need to be had with the children of these individuals, because in many cases they are going to be on the hook if a parent doesn't manage to successfully finance their retirement. This is what makes it hard. Not only are these questions that you have to ask your client, but also they're questions that may involve other people who need to be involved in the conversation.
Milevsky: That's absolutely important, but at the same time, many 55-year-olds find it tough to confront their aging head-on and involve their children in that kind of an emotional conversation. Maybe when they're 70 they're comfortable, but by then, it may be too late.
To echo what John said, I don't think we need more products as much as more ways to communicate this in a palatable way to individuals, in a way that they understand without turning them off. It's almost as if we need more communication material than product material.
Benz: The behavioral aspect of this is really valuable terrain to mine. It does seem that, despite all the academic literature, advisors and individuals have an aversion to the concept of an annuity of any type.
Idzorek: Part of that goes back to the way so many advisors have been trained over the years. The typical investment curriculum is focused on investment-only traditional asset advice. Insurance products for the most part aren't covered, so we have a void of knowledge around insurance products.
Ameriks: The behavioral aspects are really important, too. We need to understand that there may be a lot of misinformation. Moshe, I love this notion that "annuity" doesn't mean much anymore. It's lost its meaning because there's so many different flavors of it. And people don't react well to that kind of uncertainty, coupled with the notion of irreversible commitments. We're going to spend a lot more time studying some of these behavioral aspects, because there's no question that when you look at the choices that people are making they do not seem to be predisposed to annuitize. But we have to focus on both sides. There are plenty of rational reasons why people avoid annuitization.
Milevsky: Just because I'm in favor of annuities in general doesn't necessarily mean I'm in favor of every single annuity, which is where I think the confusion around annuities has to be clarified. I wish the true annuities were called "personal pensions," or something that's a little bit more favorable. Why would you not want a personal pension? It's like apple pie and mother's milk. But with annuities, you have all this emotional baggage because of the many years of negative press that you almost have to play defense from the beginning. For many people an annuity will always be an expensive, mis-sold tax shelter. Nothing will change the perception. I often talk to people at seminars and presentations, and as soon as I mention the word "annuity" half the audience freezes like a deer in headlights. When I use the phrase "personal pension," you can just feel the nostalgia. That's behavioral finance.
Ameriks: Vanguard did a research paper a little while back looking at the choices people made as they retired from companies and what fraction of them decided to take their pensions in the form of an annuity or a lump sum. A traditional pension benefit is a type of annuity, right? In a cash balance plan, 83% of people elected to take the cash balance. Even in a traditional pension plan, where no one has been shown a balance before--it's first shown to them when they ask what is the lump-sum equivalent--73% of people take the lump sum. Is that at all rational, or is that driven by some type of behavioral issue or ignorance? What is explaining that decision? Part of it's individual. There may be some of the issues around financial advisors and their training. But we really don't have a good understanding of why people make the decision the way they do.
Milevsky: One of the other factors that affects the decision of whether to take the lump sum or the annuity is the performance of the market in the past six to 12 months. Shlomo Benartzi and a student of his at UCLA have produced convincing evidence that people tend to pick the annuity or the pension after a recent bear-market period. If the market's been up strongly in the past six months, people like to take their lump sum because they're thinking they'll get wealthy. It's the recency effect. Which once again means that all of this is leading to a behavioral perspective.
Benz: In a lot of ways, if we're talking about longevity insurance, a deferred fixed annuity seems to be the perfect product for advisors and clients. What's your take on this product?
Milevsky: It's certainly a way to leverage up the mortality credits and discounting. If I could find somebody to sell me one at age 42 that gives me absolutely no refund, no death benefit, no commuted value for the next 30 years, and at age 72, or perhaps 80, if I make it, they'd give me an income for life, I've solved a bit of my retirement-income problems. But the problem is that you might have to have a Ph.D in economics to appreciate the value of something like that. That's the ultimate behavioral problem. Because most people look at that and say, "If I don't live for 30 years, I get nothing. There's no cash value, no commuted value. Why would I do that?" That's exactly where behavioral economics comes up against the rational economic model.
But I'm a fan of these products. The next generation of them will combine protection against longevity with protection against the market at the same time. They are called Return-Contingent Life Annuities. What I really want is an annuity that pays me income for life starting at the age of 75 only if the market returns haven't met my expectations. As I approach retirement, what scares me are two things: I live a very long time, and my portfolio has done horrendously. What I want is protection that is triggered by these two independent events. When you price something like that out, I would say that that is where the value is the best; it's going to be the cheapest price you see. That's where I see design going. I'll be the first in line to buy.
Ameriks: The big difference now is that most longevity policies are aimed at a big up-front lump sum for a payoff that is 20 or 30 years away and only under particular sets of circumstances. I agree with Moshe that the value of it is maximized when there is no residual or cash value to the product. But we run into some problems, both behaviorally and from insurance regulation rules in some places. I am not up to speed on regulation, but for a while, there was a problem with offering something that had no value in between when the premium was paid and when the benefit was paid.
Milevsky: Compliance departments at insurance companies hate these products. I've sat in meetings where they just cringe at the thought of somebody dying a couple of months later and the family having nothing to show for it. In fact, what they've said is that our risk of a lawsuit so far exceeds the utility value that we don't even want to consider selling something like that. Can you blame them?
Ameriks: Of course, traditional immediate annuities that people buy have the same issue if they're not purchased with some kind of refund or death benefit. Again, people seem to value those features. But what I wonder about from a behavioral perspective is if people recognize how much having those refunds eats into the value of the longevity pooling that the insurance product is trying to do.
Milevsky: It worries me. John, I've seen numbers where the trend is toward piling on features, riders, and guarantees that water down the pure longevity insurance. There are more refunds that are being selected, more payment certains, more joint-lifes, more cash refunds, liquidity. If anything, the ideal SPIA (single premium immediate annuity) is becoming more rare, even if it was rare already. The so-called annuity puzzle is getting worse.
Ameriks: It just goes back to what I was saying. The biggest problem is telling people what the deal is. The deal is you're giving up access to these assets at this point in time in exchange for guaranteed promises in all states of the world in which you're alive going forward. If that trade makes sense to you, then you can continue the conversation and talk about whether an annuity makes sense. If that trade doesn't make sense, then I think there's a limit to how much coaching and information we want to give, if fundamentally there's no desire to make this trade-off.
Idzorek: The way the mutual fund world is evolving, I suspect that pure longevity insurance will play a greater role in solutions in the future. Target-date funds are becoming more popular. You have the first generation of managed-payout funds. I think that there'll be a number of retirement-income solutions that take some sort of target-managed-payout fund or target-date depletion fund and then marry that up with pure longevity insurance. That way, investors keep control of the majority of their assets and annuitize a much smaller fraction of their total liquid wealth. Should they make it past whatever that magical depletion date is, then that insurance would kick in for them.
Benz: I have a feeling the answer to this question is going to be a big "it depends." What percentage of a person's portfolio should go into an annuity product?
Milevsky: Yes, and I guess the word "depends" isn't going to cut it. OK, here is a shot. Using reasonable parameters, I can justify numbers such as 30% to 35% in SPIA products at the standard retirement age. If you're older, more mortality credits, so it would be higher. If you're younger, it would be lower. What that means is that a vast majority of people already have a third of their income annuitized, so they don't need more. If you have strong bequest motives, it would be zero. If all you care about is sustainability, it would be higher.
Ameriks: It is very difficult to say, because situations vary so much. The key question for people is, What do I need if the worst happens? The markets are poor, I run through my assets, and I don't get what I expect from my other sources of finance. What is that bottom-line level of spending that I need every year? Then you just have to back into it. If that number is too big for your portfolio, if you start getting uncomfortable that you won't have enough leftover to deal with emergencies and everything else, then something's got to give. You have to have a discussion around what's discretionary and what's not. A lot more stuff may be discretionary than people expect--which goes back to a lot of research that's been done on consumption changes at the point of retirement.
Idzorek: I, too, am in the "it depends" category. To me, the question is, Are financial advisors learning what it should depend on? I think they are. How wealthy is somebody, and how does that compare with their retirement income?
What are their existing sources of guaranteed income? Do they have a pension? What fraction of their total annual income need is that going to cover? Again, how do they make that critical trade-off between whether they want to guarantee their own income or whether they want to leave the maximum bequest to their heirs? These are all critical elements to trying to figure out what is that appropriate retirement-income solution.
Areas of Caution
Benz: Are there any other annuities that you think in general just don't add up for consumers?
Ameriks: It comes down to the numbers. We just put out a research report that looks at different ways for people to generate an income stream. We looked at immediate annuities, managed-payout funds, Guaranteed Minimum Withdrawal Benefits of a particular flavor. As drawdown approaches, it comes down to the price and what is the benefit that they're providing.
The one area of caution that I would give on annuities and GMWBs, in particular, is that these things are advertised as if you get liquidity and you get a guarantee. But when push comes to shove, you don't get both. You have one or the other. If someone is in the money with their guarantee and looking to exercise the option to take whatever remains of their assets, well, they would have to give up the guarantees. So, one scenario I worry about with that structure is that if somebody is advanced in age and there have been bad markets and their guarantees are in effect and they're looking at a care event where they need $10,000 or $20,000, that's going to extinguish 30% or 40% or 50% of what they've got left on that contract. When they do that, their future income's going to drop as well. I am never 100% sure that investors understand this.
Idzorek: That's a great point. When we developed what we refer to as "retirement-income solution questionnaires," we really tried to hone in on what is the likelihood of that investor being able to meet the conditions of that contract and stay in the contract.
Milevsky: The answer to, Should I GMWB or not? and How much to GMWB? very much depends on the fees involved. About a month ago, I purchased my first VA, which sounds crazy for somebody who's 42 and has a pretty decent pension. I bought a VA with a GMWB because I thought that they were being ridiculously priced by the company. I think that they had made a mistake. They did not price it high enough, and they were giving a guarantee that was simply too good. I purchased it not because the optimal allocation for me at 42 was x% of this annuity. I just think this was being mispriced. So, I purchased it almost as an arbitrage. I could see the exact same thing happening in the other direction, where you look and you say, "Theoretically, I really like the idea of downside protection and liquidity and longevity. But I just can't pay 400 basis points for this."
Benz: That seems to be a likely scenario right now. A lot of annuities were too cheaply priced in the past, and now it seems like insurers are compensating for that.
Milevsky: Some of them are withdrawing from the market entirely. They just have so much of this long-dated liability on their books that they don't want to sell anymore. Others are definitely ratcheting up the fees, watering down the benefits, reducing the asset allocation that you can have. The pendulum has now swung. The direction is "de-risking."
When I talk about asset allocation and product allocation for retirement, I like to discuss strategic product allocation versus tactical product allocation. Strategically, I like GMWBs. Strategically, I like annuities. Tactically, it depends on current pricing, interest rates, as well as what your circumstances are and your bequest motives.
Ameriks: It's a very important point. Pricing will change around an awful lot. I hope it's one thing a lot more people will appreciate post-crisis; the costs of these guarantees and the price of providing insurance against the equity markets can change radically. How an insurer is going to have to deal with those circumstances can be very important and something that an advisor would want to have an opinion on, at least, as they think about whether a product is going to be used.
Milevsky: If I could summarize, annuities and anything with the "A" in it are part of the retirement-income discussion. They're part of the retirement-income portfolio. It's not the main component, and it's certainly not the only component. It's like the daily recommended vitamins and minerals. We happen to be experts on zinc. You don't want to have too much of it, but you don't want to have zero. The perception is that annuities are what retirement is really all about. But annuities are really just a small component. Long-term care and nursing homes, what will my family do, will I be able to make financial decisions? These are such big parts of retirement-income planning that it almost swamps the annuity-or-not-annuity discussion.
Ameriks: That's right. But I also think advisors owe it to their clients to discuss these issues with them and at least have some idea of what the opportunities are for the simplest of the products that we've talked about. Say to clients, "Here's what you would get per year for a lump sum in terms of a guaranteed income going forward." It's important to put structure around anything that you're going to do with a portfolio when it comes to retirement. When you start thinking about what a drawdown strategy might look like or what other types of products are going to be relevant, just having that information is really valuable.
Christine Benz is Morningstar's director of personal finance.