The use of artificial intelligence and data analytics in life insurance
The emergence of data analytics and machine learning is providing insurers and reinsurers with new insights into how they drive and monitor their business.
The American pension plan is in flux, but not for the first time in U.S. history. In this episode of Critical Point, Milliman consulting actuaries Kelly Coffing and Becky Sielman discuss why it’s a turbulent time for retirement security in America, and what the next generation of pension plans could – and should – look like given emerging technology and the changing nature of the U.S. workforce.
Announcer: This podcast is intended solely for educational purposes and presents information of a general nature. It is not intended to guide or determine any specific individual situation and persons should consult qualified professionals before taking specific action. The views expressed in this podcast are those of the speakers and not those of Milliman.
Jeremy Engdahl-Johnson: Hello and welcome to Critical Point, Milliman’s podcast. This is Jeremy Engdahl-Johnson in Milliman's media relations team and I'm excited to be joined today by Kelly Coffing, who is a consulting actuary and principal in our Seattle office. Hello, Kelly.
Kelly Coffing: Howdy.
Jeremy Engdahl-Johnson: And Becky Sielman, who is a consulting actuary and principal in our Hartford office.
Becky Sielman: Great to be here.
Jeremy Engdahl-Johnson: All right. We’re here to talk about pensions, and we’re actually at a pension meeting right now and there's an enormous thunderstorm going on outside. So if people hear that-- just they can have that context. But it does seem maybe fitting. Are we in stormy time here in terms of retirement securities?
Becky Sielman: Storm clouds are coming.
Kelly Coffing: I actually think we’re at really pivotal time in retirement in this country. We’ve really had two predominant plan designs for what, the last 40 years, 80 years, perhaps, but for the last 40 we've had two designs and neither of those designs has really lived up to the expectations we had for it. And I think we’re at a time now where we’re ready for some new options.
Becky Sielman: Defined benefit plans were the predominant form of providing retirement benefits starting a century ago, but now we have a generation of people who have only had 401(k) plans to provide them with retirement income, and we’re increasingly hearing about people who feel that they cannot afford to retire because they don't have enough money built up in their 401(k) and other defined contribution plans. Meanwhile, plan sponsors fled from defined benefit plans for a variety of good reasons. One of the most important one was the volatility of the contributions from year to year, and also the volatility of the balance sheet for private employers. Balance sheet liability is volatile. Balance sheet assets are volatile because the market, the investment market, is volatile. So it was a financial pain in the neck to be the plan sponsor of a defined benefit plan.
That's why people went to defined contribution plans to begin with. From the plan sponsor’s point of view, you have a wonderful stability of the contribution. You don't have financial volatility. But if your employees are getting to retirement age and feeling like they can't afford to retire, you, as the employer, have a whole new set of problems.
Kelly Coffing: Right. I think you end with some problems for sponsors in terms of workforce management. If people aren't retiring when you expect them to, it can cause a big logjam. So a lot of employers are experiencing where they lose their sort of mid-level talent, the people that they thought were going to move into their top spots, but the top spot folks aren't retiring because they don't feel like they have enough assets to retire, and so they're losing that mid-level talent. And they think then when the Baby Boomers do finally retire, they’re going to lose a lot of institutional knowledge. There's a lot of data that suggests that somebody staying on an additional year to try to get to the point that their 401(k) balance is good enough to retire on actually costs the company a huge amount of money, not just in terms of lost talent but actual dollars. So if we had plan designs that cued people when the right time to retire was and that really gave them sort of the facility and confidence to retire, I think we could get back to a place where we would do a better job of allowing people to transition into secure retirement.
Becky Sielman: It seems to me that there’s a tail wagging the dog when the retirement program is incenting employees to keep working because surely the retirement program should be enabling employees to retire.
Kelly Coffing: Agreed. If it's a retirement plan, it ought to actually do its job. I want to talk a little bit more about this problem with defined contribution plans because I don't want to vilify defined contribution plans. I think that they are a really, really good tax-advantaged way to encourage people to save for their own retirement. I just think people also need some annuity income, and people who listen to me talk know that I love to talk about my grandparents.
My grandpa was a drummer in a rock 'n' roll band, early rock 'n' roll, and I think my grandma finally convinced him he needed a real job, so he became a mailman. But he didn't do that job for forever. So he was a mailman, and my grandma pretty much stayed at home with the kids. So when they got to retirement, what they had was Social Security and then some years of the civil service retirement and they had some personal savings. And when my grandma died she was down to-- I guess my grandpa died first.
And then when my grandma died, they were down to not very much assets left. She had a monthly income from Social Security, a monthly income from civil service, both inflation-adjusted, and just enough-- barely enough money in the bank for another car repair and maybe reroofing the house. But they actually had a secure retirement. For middle income Americans, they had a secure retirement, and I think sometimes when we look at the decision-makers in these plans, they aren't looking at our folks where the typical household income is $60,000.
They’re thinking about something way bigger than that. So for my grandparents making middle income wages, they were able to have a retirement where they were snowbirds-- they took a couple of trips, they did some fun things, and they didn't run out of assets. And without that lifelong income, that's not how that would've been for them. And so we really-- I want to challenge us as a nation to get back to the point that we’re providing people with secure retirement. Behavioral economics is a big deal in the news right now. And Richard Thaler, who I have a massive crush on, just won the Nobel Prize in economics for his work in behavioral economics. And as it turns out, people are simply terrible at managing money, and I don't think any amount of education is going to make us better at that. We aren't good at managing a pile of money. We invest in our plans too late. We invest too little. We get in and out of the markets at the wrong time. And then sort of to add insult to injury if we transition those assets in retirement into an IRA, we pay retail fees on those. And I just think we should have that as a way to reroof the house and have a car repair but we should also have some lifelong income to handle the basics.
Becky Sielman: And I’d say even if we have people who are really good at managing their money, they can't manage how long they live in a productive way. So my grandpa was a savvy financial kind of guy who amassed a very comfortable amount of assets, but he also had Social Security and he was a longtime government employee, so he had a pension with inflation adjustment from a governmental plan sponsor. His financial plan in part, though, hinged on he was going to drop dead of a heart attack. He and his doctor made this pact between them that they were going to drop dead of heart attacks in their sleep. And he didn't. Bless his heart-- he lived to be 98. I used to call him on his birthday and tell him, you know, "Hey, Grandpa, you made it to 95. Your chances of making it to 96 are really pretty high." He loved having an actuary in the family. I mean, I know everyone loves having an actuary in the family. But at the end of his life, all he had left was Social Security and his state pension because all of the investments he had amassed got used up because he lived a long time.
Kelly Coffing: Yeah. I think those stories are important stories because we have a generation of people now retiring without that safety net, without that lifelong income as a safety net. And so the question is how do we get sponsors? How do we get companies, governments, other entities to get back into the business of providing plans for people that provide lifelong income? And the answer is we do that by not getting in the way of their running their businesses. They need stable contribution requirements regardless of what's happening with investment markets, regardless of what’s happening with interest rates.
Jeremy Engdahl-Johnson: Milliman just put out its most recent multiemployer pension funding study last month, and there's some good news in there. Fifteen percent of the plans are 100% funded. Fifty-four percent of the plans are at least 80% funded, so that's a good portion of them. But there's 10%, 131 plans total, that are less than 50% funded. And I know that that's provoking some complicated decisions for those trustees. Can you talk little bit about that?
Kelly Coffing: Yeah. We're in a really dire situation. I mean, we've seen the portion of our workforce that's unionized decline over the years. Some of the industries that were typically serviced by union workers have declined a little bit, and so we’re seeing a lot of those plans are what we consider to be very mature. So they’re typically traditional defined benefit plans. Those plans that Becky mentioned earlier are really volatile, require volatile contribution requirements. And so those plans typically are in a position where they have very few active workers compared to their inactive or their retired workers. And plans can get to the point, in pretty short order, where a market downturn really creates unsustainable contribution requirements-- that is, there's just not enough active workers going to job sites every day to fund those benefits, and so it's a real strain not just on the participants, not just on the union, but on those employers that are working in that space. You just can't run profitable businesses with the contribution requirements. So we’re seeing some of those plans fail. And particularly a plan once it's, say, 50% funded, it's very, very difficult for a plan to recover. So I think our mission, in part, we need to do what we can for those plans, but we need to put successor plans in place that won't have that same vulnerability in the future.
So what we want to do with retirement risk is minimize it to the extent possible and then take those remaining risks and allocate them in a rational way so that down the road another 50 years we aren't having this same problem. This is not the first time we've seen retirement plans fail. We actually are really on our third generation of retirement plans. There were some railroad plans back in the 1800s that we set up and they wouldn’t be legal today. They had things like 40-year vesting requirements meaning that your benefit was zero until you had 40 years with the railroad and then you had a benefit.
And railroad executives offered these benefits really with the expectation no one would work 40 years for the railroad, so they didn't fund them. And lo and behold 40 years later, some guys showed up and said, “Hey, we’d like our benefits,” and they were a little bit SOL. So that was sort of round one. And then round two was really sort of the failure we see that prompted the establishment of ERISA, the Employment Retirement Income Security Act, which is the big federal law that we are governed under now. And so we saw a group of pension plans fail there, again, with this idea that the plans just weren't adequately funded.
And I would propose that this third generation we’re seeing isn’t because ERISA failed. It's more that traditional defined benefit plans aren’t sustainable for a really long period of time. ERISA is sort of founded on the notion that entities last forever, and that the workforce in a plan will grow. And if those two assumptions aren't met, pension plans fail. And it would be easy to blame all the people who set up those pension plans. But when we look at when we first designed these traditional defined benefit plans, nobody had a laptop. Nobody had a computer at all. We were essentially designing these plans sort of, I don't know, on an envelope, for lack of a better phrase. We were designing these plans with the tools we had available to us at the time and without the ability to project those results 30 or 50 or 100 years into the future. And I think it's a mistake for us to not now, with the computing power we have, design plans that are going to be robust against those major retirement risks. So there’s interest rate risk, and investment market risk but there's also longevity risk and inflation risk. And so plan design should really look at all four of those risks and eliminate them as possible and then allocate the remaining risk in the most rational way.
Becky Sielman: I think there's another factor at work as well. In the decades after ERISA was passed, we saw increasing complexity layered on top of defined benefit plans. ERISA permitted contributions to be deductible. They’re a tax shelter, therefore, in some circumstances. And we had an army of practitioners busily working out ways to maximize the effectiveness of the tax deduction for pension contributions, which was countered by new regulations designed to counter abuses. And then we had another round of complexity designed to get around the regulation which, of course, prompted another round of legislation or regulation to cut down on the abuses. And at the end of that process of back and forth of modifying plans and then modifying the playing rules, we ended up with many plans that were just subject to a lot of requirements to provide notices or information to participants, information to the regulators, more stringent oversight, more twists and turns, more unintended consequences of behavior. And the system in some places just seems to collapse under the weight of all of that which, I think, is a pity because all of the complexity has obscured and made much more difficult the process of simply delivering retirement income to employees.
Kelly Coffing: So you do blame ERISA.
Becky Sielman: No. I could just as easily blame the practitioners who thought up clever ways around the rules. We had, you know-- It was like watching a tennis match. The ball goes over the net into, we have new regulations. The ball goes over the net to the employers, we have new complexity. We have new ways of getting around the onerous aspects of new things and then the ball gets lobbed back over to the regulators. I think we saw many plan sponsors just get fed up with dealing with the complexity, and maybe we've had plans get thrown out-- the baby with the bathwater. We throw out this good mechanism for delivering retirement income because everything that went along with it was just too complicated, too expensive to administer, too hard to understand, and too volatile to live with from a financial point of view. So if we’re going to do a do-over here, let's not succumb to the temptation to make it complicated. Let's not succumb to the temptation to make it hard to understand.
Jeremy Engdahl-Johnson: So speaking of the future and certainly you guys have uncovered some examples of employers who are looking to make that plunge and get back into the DB business, what else is there? What are we supposed to do about retirement security in this country when, say, the gig economy exists and people just don't have traditional employers anymore?
Kelly Coffing: Let’s divvy up the American population a little bit here. Three hundred thirty million Americans, perhaps, and of those about 150 million work, and some are young and some are old and some are out of the workforce. But roughly half of those people have no employer-sponsored plan at work. So whether that's the gig economy or they work for a small employer but whatever it is, they have essentially no employer-sponsored coverage at all. And really the way that the laws of our country currently work, the major way to provide benefits beyond Social Security is through either individuals making contributions to an IRA or through employer-sponsored plans. I would like to see us be able to aggregate groups of individuals into plans that they can contribute to those and then they can have some longevity pooling. So they may essentially be contributing what we could think of as like a defined contribution balance. But instead of having a pile of money at the end, which they can't manage very well or most people can't manage very well, instead we’re able to turn that pile of money into some lifelong income. There’s a bunch of ideas out there, and there's some proposed legislation that, in fact, has passed the house which would really allow us to do this aggregation. In this case, the law would allow small employers to aggregate or employers to aggregate. But I could see us over the next couple of years really reaching down to the individual level and producing some retirement planning. That said, it's tricky under ERISA now, but that doesn't mean we shouldn't change how it works.
Becky Sielman: I can see a future where saving for retirement is an automated AI-driven process. We have apps right now that plus up spending. So you go to Starbucks and you pay the whatever and 91 cents for the whatever you love to drink in the morning. And the extra nine cents of change you might get goes off into a savings account. Maybe that nine cents could be two cents towards your retirement, two cents towards paying off your student debt, two cents towards saving up for that blow-out vacation. How many cents do I have now?
Jeremy Engdahl-Johnson: You have three left.
Becky Sielman: I have three left.
Jeremy Engdahl-Johnson: Only two cents for the retirement piece?
Becky Sielman: Well, let's make retirement three cents. But maybe that breakdown of your nine pennies is something that changes over time based on information or artificial intelligence that's looking at maybe you need a penny going towards your credit card debt this month. Maybe it's January and you have a big credit card bill you ran up for your Christmas spending, and maybe all nine pennies need to go towards paying down your credit card debt. But maybe there are other months where your spending habits enable more of those nine cents to go towards your longer-term future retirement savings, paying down student loans.
Kelly Coffing: We see some states trying to solve this problem typically through defined contribution plans. But Senator Harkin, who’s retired now, really had some ideas about not just doing that through defined contribution arrangements, but really through some sort of grouped annuities. So I think as we see states starting to tackle this problem, I think it's time really for us to help create some paths forward for folks, individuals, people who work for small employers, but sort of the other half of Americans who don't have sponsored plans at work. And the other thing that, I think, we’re seeing is, because we’re also seeing some movement back to defined benefit plans-- that's small still but I think we can also take the generation of people who are retiring on defined contribution only and help them provide a better, less costly way to get lifelong income out of their pool of assets.
Becky Sielman: Let's talk about pooling. I think pooling risk is a critical concept when it comes to talking about defined benefit plans versus defined contribution plans. A defined benefit plan is taking a collection of individuals, historically a collection of individuals who were employed by the same organization. And a defined benefit plan is pooling two critical things. The first is the investment function. So defined benefit plan, the investments of the entire plan are pooled and are invested professionally. And are invested for the lifetime of the plan whether that's a couple of decades or indefinite, but a defined benefit plan represents a pool of assets that can be invested indefinitely in a mix of risky investments like equities and less risky investments like fixed income. Whereas an individual investing for themselves might invest in riskier assets while they're young and accumulating assets. But once they’re older and retired and starting to draw down on their investments, most individuals are, and should be, being more conservative in their investments. So an individual’s investing patterns looks very different from a pool of investments, and that pooling is an important way in which a defined benefit plan is a more efficient way to translate money going into the system into retirement income because that defined benefit plan as a pool of assets can perform better for longer than an individual's assets can.
Kelly Coffing: Both from the standpoint of better fees, professional management, as well as being able to have sort of an evergreen allocation. That is a higher allocation to equities over the long haul so that people really get better exposure to the market without the risk of the market. And so you really see sort of a threefold reason why participants do better in a defined benefit plan.
Becky Sielman: Just from the pooling of the investments.
Kelly Coffing: Just from the pooling of the investments. When we add longevity into that, this idea of pooling longevity, we really see a huge, huge benefit in terms of being able to provide for the dollars invested the maximum dollars of benefits out of the plan.
Becky Sielman: Let's talk about pooling longevity risks. So we have a population of people and some are going to die young and some are going to live a long time. In a defined benefit plan, those people balance one another out. For every person that’s going to live a very long time, there might be a person that's going to live not as long. They can both get lifetime retirement income because a defined benefit plan is a whole collection of individuals with different life expectancies, and we can pool the risk that some will live long times and collect a lot of benefits because there’ll be some people who don't live a long time and won't collect a lot of benefits. The pooling of risk underlies any kind of insurance, life insurance, car insurance, homeowners insurance. We pool risks so that everyone can get protected, both the people that will experience catastrophic events and the people who won't. In a defined contribution plan, each individual has to somehow manage their own longevity risk. So I don't know how long I'm going to live. I could maybe guess based on how long my parents have lived and my grandparents and maybe I can then do some planning around I think I’d like to live a long time, but I don't really know.
Kelly Coffing: But when you run out of money, I don't want you to live in my basement. I mean we could probably work something out.
Becky Sielman: I have a nice bedroom in my basement, Kelly. You can come and visit me. But no, no, not for life. Not for life.
Kelly Coffing: And that actually is one of the things that's a little bit of a problem for defined contribution plans. We talked earlier about this idea that people are not good at managing the pool of money. We talked just now that the defined benefit plan does a better sort of lifelong income thing. But there's some evidence now that when people know that they have to plan to live to be 90 they spend less than they could. And so the problem with that is we actually want people to be enjoying their retirements. So a combination of some annuity income and that defined contribution arrangement allows people to do some spending and actually enjoy a better standard of living while they're alive, which their kids may not like that, but if it's retirement income it may not actually be for their kids.
Becky Sielman: One of the downsides of a defined contribution plan is that people are, with good reason, concerned about overspending their retirement assets and running out of money before they run out of lifetime. And as a result, we see a lot of underspending from defined contribution plans. That means that there’s money left over when they die, which their kids love. But if you think about an employer who wants to fund retirement for their employees, I don't think the employer wants to spend money providing an inheritance for their former employees’ children. They want to provide retirement income for their employee. And having money left over when someone dies means there’s money that was not being used to provide retirement income.
Jeremy Engdahl-Johnson: It's interesting. I think you two are the only pension actuaries I know who have actually been opening new plans. Maybe this is why you're on this podcast. But talk about that. The conventional wisdom is that nobody is opening pension plans anymore and yet we have several examples of it.
Becky Sielman: I work with a lot of municipalities, and in my neck of the woods, there was a trend that started about 25 years ago for municipal employers to close their defined benefit plans and have new employees covered by defined contribution plans. So they now have a generation of employees that have been covered by defined contribution plans. Last summer, one of those employers called me up and said, “You know, it's becoming increasingly problematic to just have a defined contribution plan for my employees because they are not feeling financially able to retire and live off of their defined contribution plan, and it's costing me increasing amounts in disability costs and other costs to have my employees continue to work. I want, I need to have them feel able to retire. Can you help me?”
And what we looked at was avoiding the things that they had done in the past that they didn't like. So they had had a defined benefit plan, and there were certain aspects of their particular defined benefit plan that were pretty icky. And they have right now a defined contribution plan and its ickiness is that people aren't retiring. So we started with the basic premise that retirement programs are not a binary choice-- that is only defined benefit or defined contribution. That there is a spectrum of possible retirement programs and that there are not just those two options, but maybe there are new options that don't look like a traditional defined benefit plan or a traditional defined contribution plan that might have desirable characteristics or at least avoid the undesirable characteristics of the traditional plans this employer had experienced and not liked. What we’ve ended up doing with them is establishing a moderate position of a modest defined benefit plan side-by-side with a modest defined contribution plan. So rather than being at one extreme or the other, they’re in the middle. Each plan is modest in scope so there is a modest level of lifelong retirement income from the defined benefit plan. There’s a modest amount of investment risk that employer is going to bear through the defined benefit plan, and employees are going to bear through the defined contribution plan. We've also focused on keeping it simple. No complicated plan provisions. No plan provisions that might incent unwanted behavior on the part of the employees. No big plays in investment risk. We’re not going to try to hit and hit a homerun through risky investments. We’re going to keep the investments very conservative, very simple, very straightforward. So every step along the way we’re aiming for simplicity, for keeping risks at a minimum, for keeping contribution volatility at a minimum. Just keep it simple, stupid.
Kelly Coffing: And I think that's a really interesting approach because, while I work in a space where we tend to have DC and DB, we don't think of them necessarily as companion plans in as meaningful a way as you're describing here. And I think for a lot of sponsors that sort of traditional defined benefit arrangement invested more conservatively is a good approach. I work a lot with union plans where you sometimes-- there's a desire in some of those plans to have a more aggressive asset allocation. So we've worked with sponsors to transition from their traditional defined benefit plans which add risk to the plan-- that is, every year we put in enough contributions to cover the benefits that accrue that year, that is, we essentially fund the new liability that's accruing each year. But because we’re invested in a stock market, those benefits can get underfunded in the future, which means in the future, we have to find a way to fund benefits earned in the past, and that's a big problem when you have three retirees for every active. And so we see these sponsors really transitioning to, in large part, to sustainable income plans so that those benefits we fund them each year as they accrue and then they stay funded ever after because in the case of a sustainable income plan the benefit, the defined benefit, the lifelong income actually adjusts each year based on the market returns. And then we have some internal savings that we do to smooth out the ride for retirees. So that in up markets, retiree benefits are improving, and in down markets, retiree benefits are staying flat, and they’re doing that in combination with they’re transitioning out of their traditional defined benefit plan into this new sustainable income plan, and then they often have a defined contribution plan on the side. And I actually just got a call this morning from a trustee who said, “Hey, we negotiated to put another dollar-- take a dollar out of the defined contribution or the 401(k)-type plan and put it into the sustainable income plan or the SIP.” And so that's a great move for retirement security for those folks because they, right now, are getting mostly defined contribution benefits. So we’re optimistic that we’re seeing sponsors really take action to sort of change their arrangements to provide lifelong income to their participants.
Jeremy Engdahl-Johnson: All right. Well, we opened this conversation with Kelly talking to us about how it really is kind of a pivotal time in pensions. It sounds like we’re at what is this iteration four for defined benefit plans, does that seem about right?
Kelly Coffing: I think that's right.
Jeremy Engdahl-Johnson: So that is an exciting new age that we’re embarking on, and the two of you are doing a lot to make that happen. So, I'd like to thank you both for joining us. You’ve been listening to Critical Point presented by Milliman. To listen to other episodes of the podcast, visit us at milliman.com or you can find us on iTunes, Google Play, Spotify, or Stitcher. We'll see you next time.
Critical Point Episode 16: How pensions in America have changed (and should change again)
The American pension plan is in flux, but not for the first time in U.S. history; this episode discusses why it’s a turbulent time for retirement security and what the next generation of pension plans could – and should – look like.