Retirement in America: Underfunded 401(k)s and Forgotten IRAs

[AUDIO OUT] KRISTEN BROADY: --Director of the economic mobility project at the Federal Reserve Bank of Chicago. Thank you for joining us for the Economic Mobility Project's third event-- Retirement in America-- Underfunded 401(k)s and Forgotten IRAs.

We have a great lineup for you today. Enrichetta Ravina, Senior Economist here at the Chicago Fed will lead us off with a presentation on retirement savings adequacy in US-defined contribution plans, followed by Senior Economist Shanthi Ramnath, who will present her research on abandoned retirement savings accounts.

Following their presentation, Dr. Annamaria Lusardi will share her insights about their research. We'll then welcome our moderator, Editor in Chief of The Balance, Kristin Myers, who will lead us in a panel discussion where experts will discuss their perspectives on retirement savings and policies that can help Americans save a sufficient amount.

Finally, I will return to close our event. Now, let me share a few quick housekeeping notes. All attendees have been muted. A recording of today's event will be posted on our website, as well as a meeting summary. Now I'd like to turn it over to Enrichetta Ravina, who will present her research on retirement savings adequacy. Enrichetta?

ENRICHETTA RAVINA: Thanks, Kristen. I'm going to share my screen and get started. So this is joint work with Francisco Gomes from the London Business School, Kenton Hoyem at Bank of the West, Wei Hu from Financial Engine Edelman, and my work as well at the Federal Reserve Bank of Chicago.

In this paper, we propose a new methodology to evaluate retirement savings adequacy. And specifically, we then apply this methodology to a large panel of US workers who have 401(k) account. The reason why we believe this question is very relevant today is that we have seen a very large rise in defined contribution schemes.

In these type of schemes, it is the workers and not the employers who can choose the level and the allocation of their retirement savings. And so this is potentially a plus. The drawback is that this setup might also leave some people vulnerable to the effects of low financial literacy, times inconsistency, and other behavioral biases, with the risk of leaving them financially vulnerable when they reach retirement.

So is there a retirement crisis in the US? Can we evaluate retirement adequacy? Well, the literature so far is the conflicting findings. On one side, there are many papers-- I'm going to cite one here, the Boston College National Retirement Risk Index-- that tells us a large fraction of the US population is not saving enough for retirement. A large fraction of the US population will have to lower their standard of living once they reach retirement.

On the contrary, other studies-- and here, I also cite a few more-- find that the vast majority of US workers is actually saving adequately. So we join this debate, and we revisit these questions by using this new methodology. Let me tell you first our main findings, and then I will describe the methodologies, the sensitivity analysis, and I will provide some results and some information about recent patterns in the 401(k) plan space.

So the first thing that we find is that actually based on the current account balances in our sample, which is representative of the US population that has a 401(k)-- account balances, income, saving, investment behavior-- we find that 3/4 of workers are not saving enough for retirement.

We also find that 1 in 3 workers will have to decrease their standard of living once they reach retirement by more than 10% per year. The dispersion across workers in this level of preparedness is related to characteristics of the employer plan, account balances, but also worker savings behavior, which could potentially be changed going forward.

The shortfall that we document becomes worse if we consider the possibility that the worker wants to leave a bequest to their children. If we are open to the possibility that is going to be difficult to fully extract the equity in someone's house to finance consumption, either because there are financial frictions or because people don't about all the products that are available.

And finally, we consider the possibility of lower investment returns going forward. So we have some expectation of returns embedded in our work, but we could be wrong, and the returns might be lower. Last point about the findings is that retirement adequacy is a little bit in the eye of the workers because it really depends on how risk averse they are.

A worker that is more risk averse-- that dislikes uncertainty more-- then we will be more worried than another worker about the likelihood of a shortfall. And so the same worker will require a higher level of retirement wealth just to be the same level of utility and comfortableness. So risk aversion is very important for our results of maintaining standard of living.

So we examined data. Let me tell you a little bit about the methodology. We look at data of actual savings, investment decisions of more than 350,000 US workers in 401(k) plans. So our methodology differs from the previous ones in the following way.

We are not assuming that the people optimize any function. What we are doing is we are just taking our workers, look at where they stand in terms of account balances, the contribution rates, and characteristics, and then we simulate forward-- 10,000 scenarios for each of our workers in the sample.

And based on these two scenarios of income, probability of unemployment, and other features, we estimate, in each of these 10,000, the wealth that they would have once they reach age 65. Like I said, the starting point is their account value, income, and all the characteristics of this worker-- age, salary, tenure in the firm-- that we see in the sample.

We predict the future patterns of investment like savings and investment allocation by looking at the past. And we look at the past not only for the specific worker we are considering, but also for workers that are similar to this worker for age, salary, tenure, industry that they work in so that we can have a more robust prediction equation.

But we don't assume that these workers are optimizing in any way. We just look at about like a similar person, a representative person behaves in the past, and we infer the future. And then we use income shock, a lot of information about the environment, institutional features like Medicare, Medicaid, Social Securities, and then we estimate their wealth.

So what is our measure of retirement adequacy? So we calculate a ratio-- the consumption replacement ratio-- which is simply the ratio of the consumption that we estimate this person will be able to afford in retirement in each of the 10,000 scenarios. And we compare this retirement consumption to consumption in working life-- during the working life.

And the idea is to say, can these people keep their standard of living? Of course, we need an adjustment because at retirement, people don't have work-related expenses anymore, might cook at home more, or might change their bundle. So we apply a standard adjustment, and this is what we find.

75% of Americans don't save enough for retirement or don't save enough to maintain this ratio-- this standard of living-- at 1. And then there is a probability, for a smaller fraction of them, of a sizable shortfall in consumption, actually.

Although demographics, plan, firm features affect these findings. Well, all is equal, if someone contributes more, then they will have more wealth at the end age of 65. If people invest more in equity-- in stocks as opposed to bonds-- they will also be doing better on average.

If they work at companies with more generous employer plans contributions-- this tends to be companies that are older. They tend to be not publicly traded. And they tend to be company with high profits and that they invest a lot. Higher balances, higher employer contributions are good. Financial literacy and the workers education are also extremely important and people living in areas with higher financial literacy, all these people do better.

And finally, at least in our sample, younger workers-- millennials-- are doing OK. They still have a big horizon in front of them, and so we can evaluate their adequacy so far as good and sound, while people that are in their 40s and are older experienced more significant predicted shortfalls.

So I have a minute left. Let me tell you, since we do this analysis with the slightly older data about around 2015, let me tell you about some patterns that have recently emerged in 401(k) balances. And then I can list some policy solutions that our research work points to.

So first, we see that balances for the average household have increased. These are data from Vanguard in the last decade. But the median isn't. So that the lower-income part of the worker population is not investing more in their 401(k)-- or not by much-- now than they were before, although there is some improvement most recently.

People tend to-- there is a bigger fraction of people contributing in a way that maximizing the employer match, which is great. But then-- this is the TIAA Institute-GFLEC Personal Finance Index. We see that financial literacy has not really been improving too much over time, and a lot of people still lack basic financial knowledge.

Policy discussion-- so what can we do to improve retirement savings. And of course, we have to recognize that some large fraction of the worker population in the lowest income deciles simply doesn't have the resources to save. So that also is a most important issue.

But among those that have the resources to save, one proposal that we evaluate is having age-dependent minimum contribution rates so that early in life. People who have expenditures related to education, children, housing, can actually save less in their 401(k). And as they age and these expenditures are taken care of, they can increase their savings.

However, in order to have a large fraction of the population to save adequately, we would need an average savings rate of 10%. And at the moment, in our sample, that savings rate is about 6.3%. So there should be also a significant increase in savings.

Another possibility is to limit the withdrawals that people are allowed to take once they reach age 59 and 1/2. Another way to stop leakages in the 401(k) system is to have rollovers when people change jobs that are automatic. This helps people that have part-time jobs, seasonal jobs-- they change jobs often, and so they might not transfer their balances and continue to save.

And finally-- and we will discuss them later-- that are portable, state-sponsored retirement accounts which could also help. So this is all for me. Now I'd like to introduce the Chicago Fed Senior Economist, Shanthi Ramnath. She will discuss a research on abandoned retirement savings accounts. Thank you.

SHANTHI RAMNATH: Thank you, Enrichetta. I will share my screen now. everybody-- OK, that worked. Well, thank you. I will be presenting joint work with Lucas Goodman from the Treasury and Anita Mukherjee from University of Wisconsin-Madison.

Before I begin, I just want to say the standard disclaimers that the views that I present are not those of our respective institutions and that all confidential data met the disclosure requirements, and no information was disclosed.

So just jumping right into the motivation. So our project was motivated by the fact that a large share of retirement wealth in the US is held in these tax-preferred retirement plans like individual retirement accounts and employer-sponsored plans like 401(k)s and the like.

And these accounts placed a heavy burden, as Enrichetta discussed as well, a heavy burden on the individual to manage and consolidate them and keep track of them over time. So at the same time as we've seen this sort of growth in these tax-preferred plan participation, we've also seen an increase in using policies like default enrollment by employers, where people are opted into participating in their employer plans rather than opting-- and then choose to opt out.

So whereas this has had an increase in participation, you can imagine situations where people who have multiple jobs could end up with lots of different accounts, and it's on them to sort of consolidate and keep track of them throughout their life. So this has sort of led to a concern among policymakers that people may forget or lose track of these accounts. And there's even been, in the past, some policy proposals to sort of address this.

But to date, there hasn't really been that much empirical work to sort of get a sense of the scope of this issue. And so that's where-- I don't if this is-- OK, there we go. Sorry about that. So that's where our project is stepping in. And so we're hoping to fill this gap in knowledge.

So we're looking at two research questions-- the first being a very straightforward one which is, what is the prevalence of these abandoned accounts? Does this exist? How big is this problem? And in order to look at this, we're going to use individual tax return data and information return data that spans 2003 to 2020. And this analysis is going to focus exclusively on individual retirement accounts, rather than sort of the employer plans. And this is basically because the tax data has a wealth of information about these accounts that it just doesn't for employer plans.

So that sort of gives us a sense of what the current state of these abandoned accounts is. The second part of our research is looking at what causes these accounts to be abandoned. And specifically, with the increase of these default policies, do we see that there's sort of a different risk of abandonment when we have policies that exploit people's inattention to increase participation? What happens sort of at the back end when they're kind of need-- they're going to need to remember that these accounts exist?

So in terms of this being a straightforward question, it turns out to be not so straightforward when we think about how to actually answer it. So I think in terms of thinking about using administrative data, we think that it's uniquely positioned to answer this question because when you think about survey data, you can't really ask somebody, did you forget that you have an account?

And so what we're going to do is use this tax data based on what we observe people doing in terms of contributions and distributions. And one challenge is that these accounts are designed to be long-term savings accounts. So it's completely reasonable for someone to not touch that account for an extended period of time. And so we're going to use a combination of this observed behavior and the tax law to identify accounts that we think are likely to have been abandoned.

So the tax law that I'm talking about are required minimum distributions. And basically, at age 70 and 1/2-- and now it's actually been increased to 72-- an individual must start taking distributions from their account in order to avoid a lot. And if they don't do that, they are faced with a fairly large penalty.

So using our data, we sort of come up with different rules in order to try to identify people who are meeting that Required Minimum Distribution or RMD. And the key point of this is that we can use these rules to identify whether or not somebody is aware of this account, which we based on whether or not they appear to have taken that RMD over the course of 10 years after they're required to do so.

So that 10 years is somewhat arbitrary, but we think we've picked a fairly conservative number of years to identify something that's been abandoned. OK, so what do we find? We find that currently, the share of abandoned accounts ranges by cohort from between 0.2% and 0.4%. So not a huge number of accounts that are being abandoned.

When we think of what these accounts look like relative to accounts that have been claimed-- so we do observe distributions being taken from them-- unsurprisingly, they are much smaller. But the point that we want to make is also that even though they're much smaller, they're still potentially a substantial amount of an individual's income.

So looking at the 90th percentile value of an abandoned account, if we turn that sort of wealth into a flow measure and scale it by somebody's income, that amounts to about 11% of their income. So even though these accounts are being abandoned and are small on the whole, we do see that there are still some large accounts with substantial amounts in them.

So just to turn to the next part of the paper, like I said, this is sort of putting a number on the current problem of abandonment. And so we think that there's reason for this to grow in the future as people who have been impacted by these automatic enrollment policies sort of age into retirement.

And so the next part is going to look at a specific policy called automatic IRA rollovers-- not automatic enrollment-- to think about whether or not these sort of default accounts are different than something that was created with a more active choice.

So just a quick overview of what an automatic rollover to an IRA is. When people separate from their jobs-- so if they change jobs and they have an employer account-- essentially, the employers can choose to hold on to that money and maintain that account for them. But if the account is below $5,000 and above $1,000 and they choose not to maintain it, they are required to roll it over into an IRA for that individual.

Anything below $1,000, they can distribute as cash. And so we're going to exploit that policy in order to learn more about the accounts and the people who hold the accounts of these sort of default creations. OK, so the challenge here is that our definition originally was based on taking RMDs, but these automatic rollovers are going to affect mostly working-age population, so we no longer can use that definition.

And instead, what we do is we look at things that we think could causally be related to future abandonments. So the first one is looking at this intermediate outcome, which is failing to update your address when you move. So if you move and you file from a new address on your 1040-- your individual tax return-- but don't tell whoever is maintaining your IRA, then we sort of see that as a potential for you to later lose track of that account.

And then the second is to zoom in that small subset of people who are actually at the RMD age when they're automatically enrolled into an IRA. So what we find is that people whose IRAs are created with this auto-rollover are less likely to update their address. So we think that that's a possibility of them being less likely to find that account later on.

And we also find that they're less likely to take that RMD. So together, we find that this is suggestive evidence, at least, that the IRAs that are created by this automatic rollover-- so sort of taking advantage of people not making an active choice-- do appear to be at more risk for abandonment. And so I just want to stop there, and this is sort just a quick overview of what we find. So that's the end of the paper results.

And just in terms of the policy discussion, things that could help combat accounts being abandoned. There could be increased education that goes along with plans that use automatic enrollment. And just by giving people a greater understanding, they may be less likely to forget that account later.

There's also this idea of single retirement accounts that follow you from job-to-job. And the idea being that the more simple the system is, the easier it is to keep track of. And then finally, this last idea that comes out of Brookings and actually, one of the panelists is a coauthor on, is the idea of a retirement dashboard which, among many other things, would allow people to centralize their account information and help keep track of accounts over time.

So I will stop there. And now I'd like to introduce Dr. Annamaria Lusardi, Professor of Economics and Accountancy at the George Washington School of Business, who will share her insights about our research. Thank you.

ANNAMARIA LUSARDI: Thank you very much, Shanthi. And just let me say how happy I am to be here today to comment on these two very important papers. The paper both shows that many factors go into play when we look at the retirement saving in an age of defined contribution pension. And certainly, how complex the choices are in this system.

And let's be clear-- the shift from DB to DC pension is a monumental change. And it has and is going to have consequences and implications for the well being of workers, and more so now that this DC pension system is maturing, and we will start soon observing people that have lived all of their life with DC pensions.

This is why it's so important to look at more recent data than in the past, when people were living still, perhaps, with half of the pension in their DB type of pension. So the findings, to start with, are quite somber. 3/4 of workers in the sample shown by Enrichetta might not be saving enough for retirement. And so they might be facing a decline in their consumption when they retire.

And it is far from clear that workers are doing the best when it comes to these retirement system. Some workers are simply forgetting or not using their retirement account and the money foregone, as Shanthi has shown, is not always small.

So let me just say what an impressive piece of work. This is, first of all, a very large amount of data has been used and brought to try to answer these questions how people are faring, basically, in these defined contribution. And I first want to praise the author for not just doing this incredible analysis, but also being able to find this data and bringing this data to bear to this question.

In particular, the data set by Edelman Financial Engines is quite unique because it provides information not just on the workers, but also on the employer, you know, the characteristics, which type of firms the worker work at, which is really impressive. And the tax data on the IRA and the state unclaimed property record, which is also confidential data, is quite rich and is quite unique in providing information.

So this is a really impressive work, and the analysis certainly is not easy, and you methodology has been put in place to study adequacy. And also the definition and calculation, Shanthi and co-author, of being very careful and clever but also precise in trying to address these questions.

So what we are able to learn from this work is the many factors that go into place when it comes to the world of the DC pension. And it's not just how much you contribute or save for retirement, but it is also how you save for retirement, how you allocate your wealth for retirement, the type of fees that you are going to be charged in the DC pension, how much the employer is contributing, and what is the match?

And what people do when they change the jobs. So what's the leakage that there is in these accounts, for example, when they change jobs, when they get unemployed. Also, what happens when you get access to the account without a penalty at 59 and 1/2? And also, how do we accumulate wealth in these accounts?

So this paper, I think, are documenting that what people do in these accounts do not seem to be conducive to accumulating enough wealth for retirement. I think this is consistent with other work. I have to say it's consistent with my own work as well where often, when we look at people, we see behavior that's not consistent with having good wealth. We see people in financial distress very close to retirement, with even debt collector contacting them.

So when people should be at the top of their wealth accumulation, debt collector are contacting them. I don't know that we get to this 3/4 of workers might not have enough, but there is, I think, enough signs of financial distress.

But let me start at the end, meaning, let me start with decumulation rather than the accumulation side of the life cycle, which is what Enrichetta has been showing us. It seems that people-- not many, but the fraction which seems to be rising-- are simply forgetting or not using the wealth in their retirement account.

This wealth gets lost at the end. And it really begs the question, what do we about the decumulation side of the life cycle? We are focusing a lot on the accumulation side, but what happened in the decumulation side? And this is really striking and surprising. Some people may not use the money.

And so what I think the big suggestion from the paper of Shanthi is that we really need to turn the attention to that part of the life cycle as well. And not just that people might be forgetting, but what if-- as the paper of Enrichetta shows-- if people get to retirement with too little wealth, might they then gamble with their wealth?

So for example, to try to do activities or so on where they want to take more risk. For example, becoming entrepreneurs later in life and so on, and so blow out kind of the wealth that we have accumulating so far. The assumption in Enrichetta's paper is that people then use it for consumption, but perhaps that wealth is not counted or that wealth is accumulated, taken off, and then used in a different way.

And I think this also reminds us that we have to study a pension taking into consideration the state of the labor markets. And I'm going to come back to the young, but the fact that people change jobs so often so far, it's going to create many, perhaps, of these accounts that Shanthi was talking about-- account that have very little wealth. And therefore, there is no incentive for the employer to keep them, and potentially the people are not using it.

And what if these account start from maybe early in life, then people are not investing adequately. And so they just become very little account that gets abandoned later on. So I think we need to really focus on this. I think it's such a good warning that this seems to be happening.

And going back again to the paper of Enrichetta, we really see there clearly that there is indeed money coming in in these account. But we have to remember that there is money coming out already in the account well before retirement. And there are so many-- there is such a possibility of leakage because people change jobs, because people get unemployed, there is all of the early withdrawal, and then withdrawal potentially at 59 and 1/2.

So why is this so important? Policymakers have adopted policy such as automatic enrollment into account. But I would argue, given these two papers, this is not enough. Even the portability of retirement account is not going to be enough because then if people misuse or misallocate their account-- in particular, if they are small-- they are not helping, potentially, people.

So we need a system that makes it easy for people to save for retirement, but also that know what they are doing. And one potential problem of these automation is that people are not informed-- they don't have information of what is in their account. And I want to report another striking statistic.

According to a survey by TIA, 64% of people that have target date funds-- remember, a target date fund change allocation as you age, so you don't have to do much-- actually, 64% think that the target date fund provide guaranteed income. So they actually think they are going to use at retirement, they are going to have guaranteed income, the majority. So they don't even which type of fund they have and what does their fund do.

And so both paper, in a sense, do mention financial literacy some time. One says, while it is easy to put most of the blame on limited financial literacy, other factors also play a significant role. I completely agree. There are so many factors that play a role in the retirement system, but I actually think that the system of defined contribution is not going to work.

You cannot get away from the importance of financial literacy in this system because you need to make automatic enrollment work and work for everybody. And one size doesn't fit all. And also, you think of the decumulation side, when people will have to make an active choice. And if they have never made an active choice, chances are they are going to make a terrible one, including abandoning small account.

So I think without basic knowledge, people are unlikely to do well, even using policy that there are designed to help them. So these policy are too much one-size-fit-all, and people are too heterogeneous. So that's why I think we need a different system. I want to say very quickly-- talk very quickly about the young because the paper says, perhaps we need like a saving rate that changes over time.

But I think careful here because I think perhaps some of the young people dealing with that, it's not obvious that they should be saving. So the young, in particular, leave jobs a lot, so I think that's another consideration to take into account. And also, the paper by Enrichetta shows that preferences as well play a role. Risk aversion plays an important role.

And I think we need to consider preferences carefully here and also make sure that we measure those preferences well. And my paper shows, unfortunately, that risk is what people the least. So I think it's particularly hard to measure risk aversion well.

So let me conclude by saying that, in a sense, we need a lot more than what is currently done. I think policy, like automatic enrollment, portable IRA portability, are really important, but they are not enough.

And I would recommend financial education in school and in the workplace. Not because they are going to be enough to fix the system, but because to operate in this system, people need to a thing or two about how this system and how defined contribution work. Thank you very much. And now I think we will have room for questions, right?

KRISTEN BROADY: Thank you, Dr. Lusardi. I'd now like to introduce our additional distinguished panelists. We are joined by J. Mark Iwry, who is a visiting scholar at the Wharton School at the University of Pennsylvania and nonresident senior fellow in economic studies at the Brookings Institution, and Dr. Sita Nataraj Slavov, who is Professor of Public Policy at George Mason University and nonresident senior fellow at the American Enterprise Institute.

Enrichetta and Shanthi will also join our panel discussion. And now I am happy to introduce award-winning journalist, producer, and editor-in-chief of The Balance, Kristin Myers, who will serve as the moderator for the panel discussion. Kristin?

KRISTIN MYERS: Thanks so much, Kristen. It feels weird to say that.

KRISTEN BROADY: I know.

KRISTIN MYERS: Because we have the same name. I'm really excited to be here. I really enjoyed all of those presentations. It's not often that I always get to do deep dive in on retirement. But I want to get straight into this panel conversation because we have so much ground that we want to cover here. So I'm just going to throw this first question just straight out to the panelists about the policy changes that really could be implemented to help people-- but especially those with lower incomes with the limited financial literacy, which Dr. Lusardi was just talking about.

What policy changes could be implemented to help them save enough for retirement?

MARK IWRY: Kristin, if I can begin.

KRISTIN MYERS: Of course.

MARK IWRY: Begin my video at the same time. So I'm Mark Iwry. I'm delighted to be here at this excellent conference, especially with my friend, Annamaria, and my former colleague, Shanthi, both of whom are doing great work.

I think probably, Kristin, the three most important policy initiatives or reforms that we have developed and implemented or at least proposed to address these problems are part of a fairly coherent strategy for policy to improve retirement outcomes.

And those are, number one, automatic enrollment and other behavioral automatic features in 401(k)s, in 403(b) nonprofit saving plans for people who are eligible for those plans. Number two, automatic enrollment for people who are not eligible for those plans-- people who don't have an employer 401(k) or other plan available to them.

And number three, a saver's tax credit to try to make saving more remunerative for people who are in the lowest tax brackets. So if I may, just to quickly tell you what those three are, automatic enrollment in 401(k)s and similar plans refers, as we know, to the idea that people can be put into a plan if they're eligible for it and have the right to opt out.

This means not only that they're participating, but typically, there's an automatic increase, that their level of contribution might go up from 5% the first year to 6% the next year, et cetera. And that they're automatically, by default, invested in diversified investments such as a target date fund, as has been mentioned by Anna, or some other asset-allocated fund, as well as automatic rollovers, automatic annuities.

What we want people to do, by way of best professional advice, is made a default to the extent possible in the plan with the ability to opt out of it, to do whatever you want, not participate at all. Number two, for the 55 to 60 million workers who don't have access to an employer plan at work, we need to get them into the system, or at least give them a realistic chance.

Some of them won't want to save, shouldn't save, perhaps. But a majority of them probably could save. And automatic enrollment into IRAs is the solution that several of us have developed and that is now being adopted at the state level, which we'll talk about soon, and is being proposed to Congress for adoption on a nationwide basis.

This would not affect employers that have a plan, only employers that don't sponsor a plan in our voluntary employer plan system would have to simply make their payroll system available for employees to use it to save the employee's own money in the employee's own individual retirement account or IRA, with no employer responsibilities for investing, no fiduciary liability, no compliance with 401(k) rules or with ERISA labor law rules. The employer is able to just be a conduit.

Thirdly, we can make saving not only more convenient and easier for these people-- a large chunk of the 55 or so million who are not eligible for plans and more of the people who are eligible but are not participating-- we can also make it more remunerative by leveling the playing field in favor of the lower-income, lower-wage, and lower-wealth populations.

And that is by giving people a tax credit for saving in one of these plans or in an IRA on their own, if their income is below something like the median income in America. So that is a existing tax credit. But it was proposed to be much more robust.

And now, Congress is considering, after 20-some years, is finally considering the original proposal for this tax credit, which is to make it refundable-- available to people who don't have an income tax liability, as well as those who do. To make it a 50% credit across the board. And to take the credit and put it into the person's saving account.

KRISTIN MYERS: Right. Annamaria, Sita, I don't if you had anything to add on to what Mark was just suggesting as some of those policy changes?

SITA SLAVOV: Sure, I can add a couple of things. I know that we have mostly been talking about private employer-sponsored retirement saving, but what I would add is that, actually, for most people, understanding Social Security is hugely important because for most people, especially people with lower incomes, that's their biggest retirement asset.

So it is very important to understand what you're going to get from Social Security and then how you need to optimize your own retirement saving around that. And that includes not just how much and when to save, it also includes how you draw down on your saving when you hit retirement. For example, for some people, it may make sense to tap into IRAs and 401(k)s first and delay receiving Social Security because you get a pretty generous adjustment if you delay.

So making those two things work together is very important to have education about. I agree with Mark that automatic enrollment is a powerful tool. I do have some reservations about how it's designed. And in particular, I'd like to see more academic and policy discussion about what the right default is.

In particular, whether it makes sense for everyone to be saving a constant percentage of their income in every year of their life, which tends to be the default. It kind of ramps up to this constant percent. Or whether it might make more sense to concentrate saving towards middle age rather than when you're young and are struggling to maybe pay rent and make student loan payments.

KRISTIN MYERS: I very much feel that. There's this struggle that you have in your 20s. So Dr. Lusardi, I want to come to you then for this next question because there is obviously a lot of documentation on how there's gaps that exist in these retirement accounts, based on race and gender particularly. A lot of reasons for why some of those gaps exist. How do you think we need to go about mitigating them?

ANNAMARIA LUSARDI: Yeah, it's a very, very important question. And I think these two papers as well show the heterogeneity that exists among people. So you have to consider that among some of the people you mentioned, these are the lower-income people, in a sense. And so you have to make sure that we design structures that work well for them.

For example, because for them, liquidity might just be very important. And so you have to be careful in taking away important liquidity. We also need to consider, and particularly for women, that they take some time off the labor market to raise kids and so on. And so I think that the pension system has to be structured thinking as well of the structure of the labor market, you know, what is the compensation there? We do see that women, for example, also are paid less as well.

So I really do think that we need more-- get away from this one-size-fits-all because I often feel one size fits one-- and try to take into consideration the differences that exist among this group and try to design programs that are better targeted to them and to their specific need. Otherwise, you might not help them eventually, and it might be really much harder for them to save for retirement.

SHANTHI RAMNATH: Now, Shanthi, I was listening, obviously, to that presentation about those retirement accounts, particularly with lower balances, and how they are most likely to be abandoned. Can you walk us through just a little bit more about why that might be and really, how some of these accounts-- and you did touch on it just a little bit, but I'd love to hear more about how important those accounts could be to those people that are abandoning them.

SHANTHI RAMNATH: Sure. So I think that there's probably at least two reasons why these accounts exist and are more likely to be abandoned. There's the case of somebody who was at a job very briefly and was automatically enrolled, and they might be high income, and the cost of-- the hassle of just getting that money, and finding the account, and getting the paperwork, and it just might be too much. And if it's the small-balance account, then it's just not worth it. And they just kind of say, hands off. I don't really need it.

But then there's the other case, where people may just not realize they have this money. And unfortunately, that's probably going to be more likely to happen for low-income people, based on our results. So we do sort of regressions to see whether or not different characteristics are associated with these lower balances being forgotten. And yeah, we find that income is a predictor of forgetting.

And so for people who are at the lower end of the income distribution, even though it's a small balance, it may be a larger share of their overall income. And so if you end up with lots of different accounts, together it could actually be a substantial amount. And so what we're-- I think, in terms of these two different types of accounts, they're going to have very different welfare implications and have different implications for the policies that we want to use to address them.

KRISTIN MYERS: And Enrichetta, I do want to bring you into this conversation. As we were just hearing, portability is such a huge factor, especially if you don't want to go through the hassle of having to roll over one of your accounts or track down all of your accounts. So talk to me a little bit about that portability, where you're automatically transferring it from one employer to another. How difficult would it be to implement something like that?

ENRICHETTA RAVINA: So well I believe that definitely-- and Mark is an expert here-- it requires some coordination across employers. So a centralized system would definitely be good. So there would be some costs of implementing it.

But I want to argue that the cost would be smaller relatively to the potential benefit, especially for workers that are lower income, part time, seasonal, they change jobs many times, they might have low financial literacy, and so find it difficult to do the rollover by themselves or to open an IRA if they join an employer that doesn't have a retirement plan. And so I think that it would be a very beneficial move with a relatively low cost.

KRISTIN MYERS: And I want to come back to you, Mark, and also Sita, and you both were hinting at it a little bit. But I want to focus in on some of the challenges, honestly, when it comes to automatic enrollment. Obviously, hearing some of the benefits about it boosting retirement savings, but where does that become tricky and challenging, do you think, to implement? But also, what are the challenges that it could present to the people that are automatically enrolled?

MARK IWRY: Well, Kristin, for one thing, as people have mentioned, automatic enrollment has had a great positive effect on participation. And we have had-- especially lower-wage people, people of color, who have been able to save have been induced to say far more than they otherwise would have far more often-- people who weren't saving at all.

And that's a dramatic change in our 401(k) system and a change that we have implemented at the state level and would like to implement nationwide for people not eligible for plans. Now, automatic enrollment isn't good for everyone. It's not meant to be. That's why you can opt out.

People who are too poor or in financial straits shouldn't necessarily save for the long term. Emergency saving might be a better priority for them, if they can save at all. People who are very well off shouldn't be saving at a default, automatic enrollment level that's intended to be for most people, but not for various subsets.

And people who are automatically enrolled are usually automatically stepped up to higher levels. But Sita, to your point, people at different ages might have a different sort of lifecycle needs and should be saving more or less, depending on how old they are and their personal situation.

So automatic enrollment is mainly to get people into the system and to give them some kind of rough endorsement, if you will, of a saving level, but emphasizing, as Annamaria has done so wonderfully throughout your work, the importance of financial education, of financial wellness, of giving people the understanding they need of how to adapt general systems and procedures to their individual circumstances.

KRISTIN MYERS: Sita, do you want to add to that?

SITA SLAVOV: Yeah, so there's nothing there that I really disagree with. Automatic enrollment is a powerful tool and people need to be educated in terms of making decisions that are right for them individually.

Now, a lot of people do tend to stick with the default. And I think the question that I would raise and that I think it is important to talk about, both in terms of policy and academic discussions as well as in terms of educating people, is, if we don't think people are saving optimally for retirement, and if we think they need a nudge in the right direction, It would be good to have some kind of benchmark for what optimal saving does look like.

And I have argued recently that if that benchmark is what we call in economics a lifecycle model-- that's commonly used as a benchmark in academic studies-- then it often doesn't make sense for people to save for retirement when they're young because for many young people, money's pretty tight, and contributing to 401(k) or an IRA would require sacrificing spending at a stage where money is pretty tight. It might make more sense to concentrate your saving later.

Now, what that would look like for policy, I don't really know, but I think it's important to have these conversations.

MARK IWRY: And I would suggest that those are points very well taken and that we are actually, in a rough-justice way, addressing that. When my Treasury colleagues and I approved and defined automatic enrollment originally in 1998 and tried to promote it in the private pension system-- without much success for the first couple of years, I might add, but eventually, it's taken on.

We were advised by Brigitte Madrian, Richard Thaler, and others that looking at a few cases-- and Shanthi, you remember this-- people were sticking. To your point, Sita, people might be automatically enrolled at 4% of pay, and 5 years later, they were still passively saving 4% of pay.

So we fixed that by encouraging, allowing, and defining, and promoting automatic increases. Now, of course, it doesn't mean that everyone should go up 1% a year. But the fact is that with about a 6% to 7% average contribution rate by those who are saving and a recommended rate, perhaps, I think, Shanthi, you're talking about this, maybe 10% or 12%, or Enrichetta, you were talking about this, I guess, we're actually more or less in the ballpark because the typical automatic enrollment starts now around 5%.

Some plans are a little less. Some a little more. And it increases automatically, so people go into the double digits very commonly. And the states are illustrating this adaptation to heterogeneity, that people are starting at 5%, they're going up to 6%, 7%, et cetera. And a third of them are opting out, demonstrating that people really can opt out if saving isn't right for them or if they believe that it isn't right for them, even though it's been made easy.

KRISTIN MYERS: So then, Mark, who's doing it right, would you say? Which states right now are doing it right?

MARK IWRY: Well, I'd say, for example, all the states that are now implementing-- California, Oregon, and Illinois-- a couple of others are just starting so they haven't really got a track record-- Connecticut, Maryland, Colorado. But those three states that have now 600,000-plus workers saving due to the California, Illinois, and Oregon automatic IRA programs, which are part of this nationwide effort, now a dozen states have adopted this legislation.

Even in just those three, and it's still relatively early days, we've had about a million people autoenrolled. About a third of them have opted out. 2/3 are still saving. And these are the people we're concerned about. These are people with an average wage from their employer of $25,000 a year, according to the state estimates.

They've probably got a couple of jobs, a couple of earners, so maybe the family income is $35,000 or $40,000 or $45,000 or $50,000, but we're talking about people below the median household income in the United States-- the highest-hanging fruit in our private pension system. And the gap is there for them because our system is private-sector dominated, apart from Social Security, the private pension system.

And employers don't have to cover people, don't have to have a plan, so the industry tends to cater to the clients who are affluent and leave behind the lower wage as well as, unfortunately, people of color and women. So these policies are designed in a really rough way to directionally address the problem.

KRISTIN MYERS: So I know that we're essentially at time, but I wanted to just very quickly throw it out to the rest of the panel. Quick 30 seconds apiece, if you wanted. I want to know really what we actually think we can expect going forward from the US in terms of policy changes that can or might be implemented.

ANNAMARIA LUSARDI: I just want to say that policy changes are needed and are urgent, mostly because now we are facing inflation, right? And this is going to play an important role in terms of saving, per se. And also because with the pandemic we have made available to people to withdraw from their retirement account. And so those who withdraw will have to find a way to save again more for retirement.

And I just want to hint yet again that when we are creating some of this policy and creating some of the contribution rate, people take it as a suggestion. And starting very low might give a suggestion that you don't need to save a lot for your retirement. So we have to be extremely careful in making these policies. And I think we have to try to stay away, as much as possible, from one-size-fits-all.

KRISTIN MYERS: I'm going to just go around on the screen. Mark, I'm going to come to you next, and then I'd love to hear from Shanthi, and Enrichetta, and then Sita to wrap this up. How many policy changes do you think we can expect going forward?

MARK IWRY: Well, what I think we can and should expect is that in a civilized society, we make retirement security more readily available to the most vulnerable, the people who've been left behind. And that's about a third-- at least a third of our working population.

I think realistically, what we can expect from Congress is a more incremental bill that could be enacted as soon as this December that hopefully will have some of these components in it-- better portability to help get retirement funds from one plan to another plan, former employer, new employer.

That would have, in particular, a refundability for the saver's tax credit and an increase in this credit, which 10 million taxpayers are using now, but which tens of millions could use and get much more out of if Congress goes ahead and makes it depositable or refundable into an account. And a number of other enhancements to automatic enrollment continuing to push forward this relatively coherent, integrated policy strategy that we've developed.

SHANTHI RAMNATH: I think I was next.

KRISTIN MYERS: You are next, you are correct.

SHANTHI RAMNATH: Yeah, well, I think the nice thing about retirement security is that it does appear to be a bipartisan issue that has support from both sides. And so I do agree that I think a lot of the changes will be incremental. But there may also be room for the private sector to sort of step in and help with things like retirement dashboards or trying to help people sort of manage the accounts on their own.

So I mean, I think in terms of policy space, it's likely to be sort of small changes at a time to strengthen the existing system. And so the question is, how do we get the people who aren't really being supported by the current system?

KRISTIN MYERS: Enrichetta, your thoughts?

ENRICHETTA RAVINA: I'll go next. So I think it would be very important to expand coverage to those people that work for employers so they don't have a 401(k) so that at least there is the option to save. And then of course, make it portable. Customize it more because sometimes workers think that the default is also the implied advice, and there is inertia, and it's difficult to move away. So these would be the things on the agenda on my side.

KRISTIN MYERS: And last but not least, Sita, love to hear your thoughts to wrap up today's conversation.

SITA SLAVOV: All right, so what can we expect? That's more of a political question. So I'm not going to answer that. But what can we hope for? In addition to what everybody has talked about and the general theme of improving retirement security for people, I would also come back to Social Security.

I would hope to see Social Security reform because that is something that many people depend on for most of their retirement saving, and it's in trouble. It would be good to fix it in a way that brings people certainty.

KRISTIN MYERS: All right, I want to thank, of course, all of the panelists for joining me in this conversation. And I want to toss it back to my name twin, Kristen Broady, Director of the Economic Mobility Project over the Chicago Fed.

KRISTEN BROADY: Thank you so much, Kristen. And thank you to our presenters, Enrichetta and Shanthi, for sharing your research on such an important topic. Thank you Dr. Lusardi, Mr. Iwry, and Dr. Slavov for such an informative discussion. And thank you to our audience for joining us this afternoon.

We'll be sending a post-event survey, so please be on the lookout for it. We value your feedback. And again, thank you for joining us. Look forward to seeing you at our future programs. Have a good rest of your day.

SITA SLAVOV: Thank you.