Episode 48: Housing Wealth and Retirement with Jaclene Begley

Shane Phillips 0:04
Hello, this is the UCLA housing voice podcast and I'm your host, Shane Phillips. This week's episode is with Dr. Jaclene Begley, and we're talking about how unexpected changes in housing wealth affect people's decisions about when to retire, or even to reverse retirement and return to the labor market. As our listeners are well aware, housing wealth has a massive impact on society, and for most households, it is the largest source of wealth that they have. On top of all that, or maybe because of that, the presence or absence of significant housing wealth is a very strong predictor of other important life outcomes. Retirement decisions are really just one area that housing wealth impacts our lives., and we could just as easily talk about its role in driving the racial wealth gap, facilitating higher education and entrepreneurial opportunities, or influencing voting and other forms of civic engagement, and hopefully, we will talk about those things in later episodes. But this conversation is an important one because there aren't really many bigger decisions in one's life than when to retire. Jaclene and her colleague find that when people experience negative unexpected shocks to their home value, they're less likely to retire or collect social security and more likely to come out of retirement compared to households who don't experience those negative shocks. We talked about the nuances of those findings and what they might tell us about future housing policy reforms. And since this is one of our first episodes talking about housing wealth in detail, we also discussed the consequences of relying on mass homeownership as a substitute for a strong social safety net. Spoiler. It is not great. The Housing Voice Podcast is a production of the UCLA Lewis Center for Regional Policy Studies with production support from Claudia Bustamante and Jason Sutedja and Divine Mutoni on transcripts. You can reach me at shanephillips@ucla.edu, and you can give the show a five star rating at Apple or Spotify. For the planners out there, you can also get AICP credit just for listening to the show. With that, let's get to our conversation with Dr. Jaclene Begley.

Jaclene Begley received her doctoral degree from NYU and is now a Senior Economist and Research Director at Fannie Mae, and she's here to discuss her research on how housing wealth affects people's work and retirement decisions. Housing wealth is really at the center of so many housing policy decisions, and a lot of the politics around housing. So we're definitely going to talk about the role this wealth plays in society more generally, as well. Jackie, thanks for joining us, and welcome to the Housing Voice podcast.

Jaclene Begley 2:45
Thank you, it's great to be here.

Shane Phillips 2:47
And Paavo is my co-host today. Hey, Paavo.

Paavo Monkkonen 2:49
Hey, Shane. Jackie, thanks so much for doing this. I'm excited to talk about your paper.

Jaclene Begley 2:50
Thanks, me too.

Shane Phillips 2:55
We always start off with our guests giving us a quick tour of a city or a town that they know well. And since Jackie is based in DC, she's going to be touring us around there. DC is not new to a lot of urban planning nerds like us but are there any favorite places you'd like to take visitors or tell them to check out?

Jaclene Begley 3:11
Right? So I think DC has a lot to offer urban planners. What I think is most interesting about DC is the fact that it's a city and then also has sort of like a state government component to it. But my favorite things to do in DC are just walking around different neighborhoods and exploring the architecture. I recently discovered the Main Avenue Fish Market which I read some trivia is the oldest fish market in the US. That's definitely interesting to check out. And also the obvious answer is all the Smithsonians are free, and before I lived here, I'd never been to the National Archives or the Library of Congress, and I think they're both underappreciated.

Shane Phillips 3:50
Can I ask you how Fannie's offices compared to HUD? HUD has famously unpleasant offices.

Jaclene Begley 3:56
Fannie Mae is the exact opposite. Fannie Mae is a brand new office at the former Washington Post headquarters site. It's a brand new building, thoroughly modern, you know, high tech so...

Paavo Monkkonen 4:09
As befits its role in the nation's economy. Before we get into the paper, I wanted to ask you about the research team at Fannie Mae. So I understand this is a growing effort on the part of one of the GSEs?

Jaclene Begley 4:23
Yeah, so the team that I'm part of is fairly new in the last seven years or so, and we are part of an emphasis to grow our thought leadership capacity. And so my team specifically focuses on housing affordability research for external consumption, academic quality, long term papers, focusing on lots of different topics right now an emphasis on racial equity. And then we also help internally with questions the business has when they need an understanding of the research.

Paavo Monkkonen 4:51
Cool. Sounds interesting.

Shane Phillips 4:53
And can you actually give us a quick, you know, elevator pitch on what Fannie Mae does. Paavo use the acronym GSE, which is a Government Sponsored Enterprises of which there are several just in the housing finance space. So what is Fannie Mae's role in all of that?

Jaclene Begley 5:10
So I'm here today representing my own views and not those of Fannie Mae nor the Federal Housing Finance Agency. But in general, Fannie Mae's role is to buy mortgages from banks; we help support the secondary mortgage market. So we buy mortgages from banks and then package them into securities and then sell them on the secondary mortgage market. We also, though, have some emphasis on things like helping moderate middle-income households. So like we have an emphasis on supporting the mortgage market in the middle-income space. And then we also help with underserved markets. So we have this focus on duty to serve markets, which are things like rural markets, where typically there's less lending there. So we also have some focus to help these areas.

Shane Phillips 5:55
Got it. So today's article is titled 'The Effect of Housing Wealth Shocks on Work and Retirement Decisions'; it was co authored with Sue and Chan, and it was published in 2018 in the Journal of Regional Science and Urban Economics. You've studied how people's retirement decisions are affected by changes in the value of their homes., and there's a lot of great stuff in this article including what I found to be very large impacts on ddecision-making and surprising differences between people affected by moderate price shocks versus larger price shocks. Before we even talk about housing wealth, though, there's a broader and older academic literature that looks at how wealth more generally influences household decisions not just housing wealth. And I think that's a good place to start. Can you tell us what the literature says about how asset value or wealth fluctuations, again, especially for non housing assets, like stocks, affects people's behavior? What are the impacts on decisions around work and retirement specifically? And what other important ways can positive or negative shocks to wealth affect household behavior?

Jaclene Begley 7:03
Great, thanks. That's a really important question, this is important background context for the paper. So most generally, if we assume that people are making trade-offs between working and other fun activities that economists would call leisure time, we would basically assume that any sort of increase in wealth would decrease your incentive to work and increase your leisure time. So the literature on non-housing assets generally find that declines in asset values through stocks really only have a weak effect on individual labor force participation, which is surprising because like I said, you would assume that if you have an increase in wealth, you would decrease your labor market participation. But the difficulty in the literature is that when stock markets increase or decrease, it's usually correlated with the broader economy. So labor markets are also experiencing these positive or negative fluctuations at that time. So it's really hard to untangle these two effects. So we don't know if decisions around retirement are due to the labor market rather than the stock market. On the other hand, there are studies that look at unexpected effects of things like inheritance or winning the lottery, and in those cases, we actually see some really clear influences on people's behavior. So people are more likely to retire early if they win the lottery. This suggests that unexpected shocks to household wealth really will affect retirement or labor market participation in general but It's hard with a lot of the other, with things like stocks, to untangle these effects.

Paavo Monkkonen 8:32
I wonder, and you might not know the answer to this question, but I wonder whether the correlation between people's earnings and their prevalence of owning stocks is figured into this literature at all, just because you'd imagine the people that have wealth, make more money. And so when the labor markets doing well, they have the opportunity to make even more money. And so they opt not to take the leisure and work more.

Jaclene Begley 8:55
That could be, I don't know the literature as well. It makes sense, though, like stock market, people who own stocks are definitely different than the broader homeownership population. And then the other thing that the literature finds is really, there's a lot of status quo bias. So if you're working and you're planning on retiring, maybe at 65, or 67, a stock market increase may not change your decision

Paavo Monkkonen 9:21
Right

Jaclene Begley 9:22
It's hard to unpack that, like you might just continue to work, whereas a decrease in your savings might incentivize you to work longer because now you're like, I won't have as much wealth. But I mean, for a lot of people, maybe having just some more wealth is just an added bonus, but not enough to make you retire.

Paavo Monkkonen 9:38
That makes sense, and you lay out several reasons that the standard economic theory about how wealth shocks affect decision-making might not apply to housing wealth. And in some cases, changes to housing wealth might actually have the opposite effect of shocks on other kinds of wealth. What are some of those reasons and what makes housing different?

Jaclene Begley 9:58
Right, so this is really important. So the majority of Americans are homeowners and in particular, the majority of older adults are homeowners, the homeownership rate is really high among seniors. Housing wealth is also the largest form of wealth holdings for most Americans. So as I was saying many more Americans own homes than own stocks. But housing is also unique because there's this consumption value to it meaning that you live in your home, it provides shelter, you have an emotional attachment to your home, and to the specific location where you live. So this may make homeowners more detached from the pure asset value of their home. It's also a less liquid asset so it takes longer to sell a home, there's high transaction costs to selling, and it's also harder to know the true market value of your home. So stock markets, we have a really clear understanding of the value, your house, it's more fuzzy. We have some indicators from AVMs (Automated Valuation Models) but you actually don't know what the transaction price will be ahead of time.

Shane Phillips 10:54
Right, you can only find out after selling.

Jaclene Begley 10:56
Right, the true market value is what a transacts at on the market, and it's a lot harder to see.

Paavo Monkkonen 11:02
Although that has changed a lot. I feel like people are much more aware of their home value these days than they were 15/20 years ago...

Jaclene Begley 11:08
If you assume that AVMs are accurate

Shane Phillips 11:11
If you believe Redfin. AVMs are what?

Jaclene Begley 11:13
Automated Valuation Model.

Shane Phillips 11:17
Okay

Jaclene Begley 11:17
So I know from looking at my own value of my own house that Redfin versus Zillow gives me wildly different estimates.

Shane Phillips 11:26
And it changes a lot like month to month in a way that the actual price probably is not.

Jaclene Begley 11:31
Right, they're noisy but yeah, that is a fair point, right now, people may be more aware that they can gauge at least sort of the market value of their home. And the last thing I was going to say is that there's also we know, like strong requests motives in the housing; so people tend to want to save their housing to give it to their children, some of my dissertation work shows this. So all of this means that people may be less likely to react to changes in housing wealth compared to other forms of wealth.

Shane Phillips 12:01
And this correlation that you brought up between labor markets, and housing wealth seems especially important. Most of the previous research on housing wealth shocks and retirement decisions didn't find much impact if any, and the labor market housing market relationships seems like it might go a long way toward explaining why. Could you dig into that relationship a bit further and why it's important?

Jaclene Begley 12:25
Yeah, so it is really important, and I think this is a real challenge in the empirical literature because of the fact that housing and labor market markets are so highly correlated. And so when there's a national recession or local economic downturn, the housing market is also going to decline. And when the labor market is doing really well, for example Amazon is coming to DC, the housing market is also going to respond and become more expensive. So in the case of retirement, people may be forced to retire earlier than expected if the labor market is declining, which is not actually a reaction to housing wealth declines, but a reaction to the fact that they're, you know, getting left out of a weak labor market. And so a challenge for us in our identification strategy in this paper is to find a convincing way to separate out labor market effects from housing market effects, to really understand the true effect of you know, housing value declines on people's labor market decisions.

Shane Phillips 13:20
Yeah. And it seems like it might actually, this might actually be more of a problem for people at or near retirement age, because you hear about, you know, when someone's industry is shrinking, or going away, you know, during the Great Recession, for example, how older people who had, you know, maybe done a different kind of job for the past 30 years, were in some cases struggling more than others to find new employment, maybe in an entirely different industry. And so, I don't know how this might affect the results, but it does seem important, you know, at that household level.

Jaclene Begley 13:52
Yeah, that's really fair, right. So it's harder for older adults, potentially, to find another job or switch industries, and there is evidence of this in the literature that makes it even more challenging for people to recover from something like the great recession.

Shane Phillips 14:10
So we're talking about price shocks, basically, unexpected increases or decreases in the value of people's homes. But in the context of this article, how big are the price shocks that we're talking about? How big or small does a households housing wealth need to depart from their expectations or from just its starting point, in order to be categorized as either a positive or a negative shock, and maybe even a moderate size shock versus a large shock?

Jaclene Begley 14:42
In our paper, we're looking at any deviation from zero, so we're talking about positive shock or negative shocks and we put them into buckets. So we do for positive shocks, zero to 5%, then five to 10%, and then over 10%, which we consider a large shock to housing values. And then the same for negative we look at zero to five, negative shock, and then five to 10, and then greater than 10. And we did experiment with different ways of modeling the shocks and the different sizes of shocks but these buckets were the ones that made the most sense at the end of the day with the data.

Shane Phillips 15:14
Okay, great, kind of stepping outside of the article itself a bit, I think it's helpful to start kind of early here with a little bit of a conversation about just housing wealth, generally and home ownership in the United States, and how it factors into retirement decisions in particular. What I wanted to ask you to do actually, though, was kind of make a case for and a case against our reliance on homeownership as a tool of savings, as a means of saving for retirement, having something to pass on to your children, all these things that people wanted of housing. I think these things have come up in previous episodes but we've never really talked about the wealth side from a research perspective. So I know there are positive and negative arguments so how do you think about those things?

Jaclene Begley 16:04
Right, so there's lots of benefits to homeownership, it sounds like we've talked about these in prior episodes. From a wealth building perspective, for one thing, housing is an accessible form of wealth. So this means that people buy a home, they can live in a home, a lot of people want to own a home, like the government incentivizes us to own homes. So housing seems to be a simpler way of owning than owning a stock, for example, which is more complicated, and housing as a forced savings mechanism. So the whole mortgage market system is set up so that you put a little bit of your equity away with each mortgage payment, you're building up equity in your home. And so over time, it's this natural for savings. We also have lots of tax benefits to owning a home that make it you know, tax advantageous to own a home. And then the fact that there's also this consumption value is nice, because you get to live in your home, and then you also derive wealth from it. So it's this really unique asset in that sense. On the downside, as we see in the paper, well, first of all, like there's a few things in the paper, we see, you know, the timing of when you buy or sell your home can affect how much wealth you have in your home at the end of the day. And so it's not that everyone's going to, you know, retire when their housing wealth is up, some people potentially want to retire and their housing wealth has dropped. And so it's not theirs for savings, but it's not as predictable as you would hope for something about, you know, this is gonna be my retirement wealth. And you know, I'm going to use this in my old age. And then there's also this other aspect of it in the literature that we don't really talk about in the paper, but that we really don't see that people downsize or move as they age. So people really aren't tapping into their housing wealth, the way you would expect, for the amount of time that we push, like housing wealth is that the best thing you should be doing, this is the predominant form of savings for most Americans, we just don't see people using their wealth.

Shane Phillips 17:48
I don't want to digress too much on the topic of housing wealth, but I do think a lot of those points are important to have in mind, and, you know, I think it is better to have a system where housing wealth isn't people's main form of savings, or patrimony. But those systems are hard to develop, if you look around the world, right, a few countries have developed a robust, non-housing wealth method of having retirement savings accounts, but most countries are still using real estate as a way, you know, kind of one of the main ways people build wealth. But so speaking of the transition to retirement, I think this is a super important question, and I wanted you to kind of give us a better understanding of what we know about how people make this decision to retire.

Jaclene Begley 18:32
So this is really important, and something that I was surprised about when I started working on this paper. I think that I expected retirement to be a clearly defined decision for most people, but you actually don't see that. So in the data, you see that people will say, "Oh, I'm retired", but I should back up, and say, we're using the Health and Retirement Study in this paper. And so part of a big focus of that dataset, it's a biannual panel where they interview people over time - a big focus of that is obviously retirement and working decisions. So there are a number of questions in the panel about are you retired, how much are you working, how many hours a week do you work, and one thing that was surprising to me was that people will say, "Oh, I'm retired", but then they'll also say, "Oh but I'm working", or "I'm retired, but I work 10 hours a week", right? So I think people self-identify as retired, but they're also still working, and some of this might have to do with being not in their original career level job, working maybe a bridge job, or you know, retiring officially, but then still doing some consulting or working part-time, or maybe working in a different field now. So this for us was a challenge, trying to figure out how do we define retirement.

Shane Phillips 19:42
I think there's also, you know, something I've heard about is how people at least this is more true of unemployed people, but someone who is say, 64/63 whatever, who is unemployed, but then transitions to retirement - just that title change like functionally there's no real difference - they were unemployed before they're unemployed after, and yet calling themselves retired, changes their own self-construction and actually become happier people because of them. So I think there might be an aversion to describing yourself in certain ways. You know, you'd rather call yourself retired, even if you're still working, rather than call yourself like, a part time employee at the, you know, local grocery store or something like that.

Jaclene Begley 20:29
Right, I think that's fair, and then you see that on the flip side, right? You don't want people, people who say they're retired might not want to admit that now they're working again. And that was one of the concerns we had too with modeling retirement reversals was maybe people are working, but they don't want to say that they're not retired because, you know, there may be some stigma associated with that.

Paavo Monkkonen 20:48
Yeah, it's it's a super interesting question. I wonder whether you're worried about cohort effects also, just because I think that a lot of this kind of cultural associations with how long someone should work, you know, when you should be retired, or when you should not be retired, I think changes a lot over time.

Jaclene Begley 21:05
Yep, I think that's really fair, and I don't think we really have any baby boomers in our sample, or we get just a small amount of baby boomers in our sample, but you can imagine that that might change, and so there definitely is a probably a cohort effect going on.

Shane Phillips 21:18
So when you're measuring these positive or negative shocks to housing values, I think it's important to know that you're not just looking at absolute changes in value, you're also trying to take into account the expectations people have about the future, which I imagine in most cases are expectations of growth. So for example, if a house fell in value from 400,000, to 350,000 over a two-year period, you know say during the Great Recession, or right after, the actual price shock relative to expectations is probably larger than 50,000 because the owner may have expected the value to rise to say 425,000, instead of fall to 350. Without getting too deep into the methodology here, could you explain why you took that approach to measuring price shocks and why that was important to the methodology of the study?

Jaclene Begley 22:11
Sure, so there are a few reasons. One I think I didn't mention earlier was that we really wanted to capture the unanticipated component of housing value changes. And the reason for that is that economic theory states that people should respond to unanticipated expectations, whereas anticipated expectations, they would have already incorporated into their behavior. So for example, if I expect I'm going to get a bonus this year, and I have a good sense of how much it would be, then I'm probably already changing my spending or planning savings around that. But then if I, if I get a surprise bonus, or I don't get a bonus, even though I was expecting to get this mount, I would be more like likely to react to that, and we should see a reaction there. So we were looking for the unanticipated change in housing values because we think that's what really would influence behavior. And we're also trying to follow some of the literature on this so there's other studies that show that these expectations are maybe what matter for measuring shocks. But we also build on these prior studies, because we are looking at sort of like metropolitan area growth trends and applying them to local zip codes. And one of the reasons we're doing that is we're trying to separate out these broader labor market issues, labor market changes from local housing market changes. So as I said before these, the conundrum we have is how do we get the labor market fluctuations and take them away from housing price fluctuation. So we were using this expectations model, and this metropolitan combined with zip code level trends to look at, sorry separate that out to get it another way to help us figure out how do we know the true effect of housing price shocks on people's behavior?

Paavo Monkkonen 23:46
Yeah, I think it's a really clever way to get at that issue. Because if you think about like a metropolitan labor market doesn't fluctuate at a neighborhood level, as much as the housing market fluctuates at the neighborhood level within a metropolitan area. So very clever way to do it.

Jaclene Begley 24:02
Thank you, that's exactly what we were thinking, you know, the metropolitan markets probably really correlated but if we could get at this zip code level effect, then untangle that

Shane Phillips 24:12
The expectations stuff is really interesting to me, and I'm wondering, you know, maybe there's just more literature on this, but do we have any sense for how much expectations really matter? I mean, clearly, you just said, in some ways, it's what is determining decision-making changes but I guess I'm wondering about scale here, the comparison of the absolute change versus the expectations, and whether people maybe feel the gap between expectations and reality, more or less strongly than the just the absolute change in their housing value. I don't really ask this as a question relating to the study here but it does interest me because I think it may have some bearing on the politics of housing affordability. Like if we want to improve affordability which is many of our goals here, one way we could do that is by causing prices to fall over time, one way or another, and you can easily imagine a big backlash to that from homeowners who would in some real sense, be losing money. Alternatively, we could improve housing affordability over time by slowing or stopping the rise in prices, but not necessarily turning the price direction to a negative so long as incomes were rising at a faster pace than housing, it would be getting more affordable in real terms. But the more that expectations matter, the more it seems that even that latter scenario could face serious opposition and backlash. Again, even if people's home values were still going up, and not losing any value.

Jaclene Begley 25:47
Right, so I've been thinking a lot about this too, lately, particularly with another project I've been working on where we're looking at itemizing the full costs of ownership over the lifecycle, including things like transaction costs, which I think people don't think about. So I think there's a lot of expected and unexpected variation in owning but then I think your main point is really about this, the unique challenges of housing and how we understand housing affordability; housing being an asset, but also a consumption good. So there's a lot of literature that does a great job - I'm thinking particularly of this John Quigley and Steven Raphael paper in the Journal of Economic Perspectives that talks about housing affordability, they do a really nice job of laying out the competing forces at play here. But I think the most important part from that paper is they state maybe the obvious that rising housing prices are great for current homeowners, not good for people who want to be first-time homebuyers, and at the end of the day, I think affordability for homeowners over the last 10 years or so hasn't been a challenge; because interest rates have been so low housing has been more affordable, homeownership, not housing, homeownership has been more affordable. What's been a challenge is really for getting first-time homebuyers into homeownership. And then to back up to this question of expectations, I think this is an empirical question that we don't have a good answer for whether or not it's your expected gain versus your actual gain that's really going to make someone more of a NIMBY or less than a NIMBY with caring about how to how to affect housing affordability broadly. But I do think homeowners at the end of the day vote to protect their asset values, and I'm not sure whether people would differentiate between a decrease in value versus a decrease in the expected increase in value. At the end of the day, both of those might just be bad news for a homeowner. So if I lose $50,000, versus I just didn't gain the 10,000 I was expecting to gain, I think homeowners will think about both of those as a problem for someone who's thinking about their home as an investment. But I also think that tying it all back together to like the cost of homeownership, if you try to stop the rise in prices, it may be hard for people to actually break even on their homes, when you factor in the transaction costs on their investments and other costs related to being a homeowner.

Shane Phillips 28:01
So the three outcomes you're really measuring here are decisions to retire, decisions to reverse retirement to come out of retirement, and then the timing of claiming social security benefits. I think those first two are pretty straightforward in what we're talking about but can you talk about the significance of claiming social security benefits or the timing of it, I think some people may not even be aware, especially maybe our younger listeners, that you know, the amount that you can claim depends on when you actually claim it.

Jaclene Begley 28:33
Right, so that was again, started, we looked at people transitioning towards retirement, people transitioning away from retirement, and then people transitioning to claim social security. And social security is a unique question, because like you said, you can claim social security early, and if you do you get less of a benefit, but you get a benefit over a longer period of time. Or you can claim social security at the age that's appropriate based on your birth year, at age 65 - 67, I think right now, or you can claim it later. And if you claim it later than that, you get more of a benefit per month, but you get you have less years to collect it over time. And so these benefits should all be actuarially fair, and there's a long literature like trying to game out when people claim social security and why. And for us, this was more of an empirical question, like there's a lot of literature that shows that social security claims are happening when you retire, but not necessarily. And so we were wondering if people might start claiming social security as a way to improve liquidity, even if they weren't actually retiring. We didn't see a lot of that. Ultimately, empirically, in the data, we saw that social security outcomes were really similar to the retirement outcomes because they are correlated, but we were curious about whether people, if your housing values are going down, maybe you are going to claim social security but not retire just to get the additional payments.

Shane Phillips 29:54
Got it. So now getting into the results. Again, you and Dr. Chan looked at the impact of Housing wealth shocks on people's decisions to retire, or reverse their retirements by returning to the workforce, and also claim Social Security benefits. Tell us what you found comparing people with positive and negative housing wealth shocks and wealth shocks that were either moderate or large in either of those directions. There's a lot of surprising stuff in here about how big the effects were, how they differed by gender and other characteristics, and how they varied over time, especially pre and post Great Recession.

Jaclene Begley 30:32
Right. So for men, we find that relative to our reference group and in the data, basically, negative shocks of moderate level are associated with reductions in the probability of transitioning to retirement, meaning people are working longer. And then when we we look at larger negative shocks, we expected them to have a bigger effect on retirement delays. But we actually don't find that in the data. And we think this is because of this challenge of controlling for labor market effects, which we throw a lot of things in the model, we try a lot of robustness checks, controlling for things like local unemployment, unemployment, wage growth, looking at specific industries of employment in MSA but we still just don't, we don't see anything there for large negative shocks, and we think we just can't fully control that in our models to find an effect there.

Shane Phillips 31:20
And I did want to just say the the sizes of the effects here, it was for negative shocks of zero to negative 5%, we were talking about a 14% reduction in decisions to retire, and for negative five to negative 10% reductions or price shocks in the home values, it was a negative 18%, or an 18% increase in decisions not to retire I guess would be one way of putting that. But that's a really significant amount, I mean, almost 20% of people who might have retired during that period didn't because of you know, 5 to 10% decline in their expected home values.

Jaclene Begley 31:57
That's not 100% the way to interpret it, it's an 18% decline in the probability of retiring, but it's 3.7 percentage reduction. But yes, it's not it's not a trivial number, and it does, it does show that people are reacting to these negative shocks in housing values. On the flip side, though, we actually didn't see anything for positive shocks, which is, you know, consistent with what literature has found, and what I was saying before about status quo bias where people may not incorporate positive shocks into their behavior but a negative shock we did find was associated with a retirement delay.

Shane Phillips 32:35
It does seem to point toward I mean, coming back to that last question about expectations, just it seems like if we can avoid people's home values, experiencing negative shocks relative to expectations, that's good - that has negative impacts on people, it's something that people are going to feel and maybe have responses to in the way they vote and so forth but there's almost like no obvious upside to positive shocks beyond their expectations. It's just, they don't even notice them, or their behavior doesn't change in response to them. So if we could at least limit that from happening, these positive shocks beyond expectations, that might be like a good starting point, and then we can worry about the negative stuff later.

Paavo Monkkonen 33:17
I wonder, so Jacklene you had cited some other papers that didn't take this approach, and they hadn't found any relationship between price fluctuations and people's retirement decisions. Could you just mention how these results differ from those.

Jaclene Begley 33:32
There are some other papers that look at housing price changes and retirement decisions, a lot of them are using metropolitan level housing price data, and not all of them are trying to get at a shock to values. Some of them are just looking at the changes over time. And then some of the papers are also using a pre-Great Recession time period. We were starting to work on this during that tail end of the Great Recession so a lot of the prior literature was just using the broom, and I think it's important to know I hadn't said this before that prior to the housing price boom, and the Great Recession of the bots, we didn't see these crazy fluctuations in housing values, I think this was surprising to people. Now maybe less so given that we've had this a decade ago, a financial crisis but I think to have these fluctuations in housing values, the boom and bust, were were a unique situation. And so now we've got a lot more information, we have Zillow, we have AVMs, we've had experience with the Great Recession; now things might be different but at the time, I think that this was unique, right, this was this was a unique point in time to measure these shocks.

Paavo Monkkonen 34:36
Yeah, although I mean, it is impressive how short people's time horizons can be. I mean, I think, talking to people and looking at housing markets in the US right now, you know, people already forgot about the Great Recession to some extent, right?

Jaclene Begley 34:51
I had to remind you because at one time I was presenting this and people were like, why were people shocked that their housing values.......

Paavo Monkkonen 34:58
Housing goes down? No way

Jaclene Begley 34:59
Yeah, I was like....

Paavo Monkkonen 35:01
I think but you know, and thinking about where they went down and where they did it in a metropolitan area. I don't know if you've thought about this at all but you know, you can imagine it's the lower priced parts of the region that lost much more value than the higher price parts of the region where people don't have to sell during the bust. Does that factor into this work at all?

Jaclene Begley 35:21
Yeah, so we didn't unpack that in the paper as much but that is part of our theory as to why particularly with things like retirement reversals, we don't see as much of this movement. And what we think is, it's because the areas that were hit the worst, that had the most negative shocks, likely are different and have, you know, different labor market access than other parts of the metropolitan area. So even though we are trying to unpack the metropolitan area from the neighborhood, there's still going to be some local variation that we can't capture. But like I said, our contribution really is getting at these more local effects. A lot of the other papers, even if they say, local, at the time, are using metropolitan data. The Zillow data was still fairly new at this time too like the zip code level data that they have available.

Paavo Monkkonen 36:10
Right, and we'll get into the further heterogeneous effects but I wanted to just get your thoughts on this, the big picture takeaway, that you know, when things are bad, you get a negative shock, you go back to work. When things are good, you get a positive shock, you just stay working. What does that say about humans, like where's this leisure time that people are supposed to be accessing with all that wealth?

Jaclene Begley 36:34
Well, it may have to do with the fact that this is housing wealth, and like I said, before, it's trickier to access it, it's lumpy, where but you know, it's still wealth, so I think maybe people are just more reactive to the loss in value whereas the positive, it's not like a stock where you can cash out right away. And I think also with retirement, like I said, before the status quo bias like people have in their head, I'm going to retire at 67, maybe you're not going to retire at 64, then just because your housing costs went up. Whereas on the flip side, if you are going to retire at 67, and your housing wealth is dropping, now you're like, "Oh, well, maybe I'll work another year"

Paavo Monkkonen 37:10
That makes sense.

Shane Phillips 37:12
And maybe you know, you're not working less, I mean, maybe you are working less, but you're still employed. And so your status doesn't change but also maybe you're working the same amount but your consumption just gets a little fancier, because you have more money, or you can pull from a home equity line of credit, and so forth. So your vacations are a little nicer, you get a bigger boat, or whatever. You know, as one does,

Jaclene Begley 37:36
We do look in the paper at like partial transitions to retirement, but we find the biggest effects are people who are like fully retired or not fully retired.

Speaker 1 37:46
And then Pavo mentioned this other heterogeneous effects. So you talked about men, how about for women,

Jaclene Begley 37:53
Right, so it was important that we separate the sample by men and women. A lot of the labor literature finds differential effects of men's reaction to economic downturns versus women's, and we don't really find an analogous effect for women. So we do see some effects for women, when we look at those that have some subgroups, like those that have fully employed partners or spouses. But we don't, in general, see a strong response from women, it's really men. And then we also, we also look at other things like retirement reversals in our sample. And we, for the most part, don't see a lot of movement on retirement reversals. But again, for some subgroups, we see that, like men who are self-described as fully retired, and those who have mortgages are more likely to reverse retirement. And this goes back to what I was saying before, we think that there's something different about the labor markets, the hardest hit, potentially like these local markets, we're just not able to capture people who might want to go back to work, but can't find a job.

Shane Phillips 38:55
And so that is the retirement decisions, the decisions to transition into retirement - pretty big effects there especially for men or really primarily almost only for men. What about decisions to reverse retirement, to come out of retirement, and also to claim social security benefits during this time period?

Jaclene Begley 39:13
Right, so for reversals, for people coming out of retirement, we really don't see a lot of movement in the data. We only find effects for a subsample of the population. So for men, those who are self-described as fully retired, and those who have mortgages are more likely to come out of retirement as a result of the negative housing price shock. Part of it is when we looked at our sample of people who were already retired, which is who those are the people who come out of retirement, they just look different than the full sample. So there are people in worse health, they were older, lots of reasons for that would make it more difficult to actually come out of retirement. So we only found effects on a small sub-sample. Our social security claims which we weren't sure the direction of how that would go. We weren't sure people would think about it as coincident with retirement or If people would claim social security early to get the liquidity benefit of having the income stream from social security. We actually found that the results were pretty similar to retirement decisions - so the negative shocks, people were more likely to claim social security when they experienced negative shock.

Shane Phillips 40:16
I think one other thing in the results that surprised me it was that there were no differences, or you didn't find any differences by race or ethnicity. And I think even, you know, thinking about what Pavo was saying earlier about how you might expect more impacts on communities with lower home values, which would disproportionately be communities of color, you might imagine that there would be some disparate impact there. So how do you think about the fact that you didn't find any differences by race or ethnicity?

Jaclene Begley 40:45
That's a good question. I don't have a great answer. I know we didn't find any differential effects when we looked at the sample across race or ethnicity. It could be the behavior across groups with similar and we didn't have a large enough concentration of people in the zip codes that had really big negative shocks that we can tease out on effect, it may just be a sample size issue in the data. But we didn't wind up seeing it, and we note that in the paper as you rightly point out.

Shane Phillips 41:12
Maybe given that you found that in places that had the largest negative shocks, the labor market outcomes or at least, this is one possible explanation for why you didn't see the same impact on retirement decisions is maybe because the labor market changes a way that offset the changes in decision making, so maybe that's something that was going on here, too, is because there's probably those communities might have experienced much larger negative price shocks. And so, they're all kind of more likely to fall into that category.

Jaclene Begley 41:43
It could be that we just weren't able to get an effect on the coefficients. Yeah, it could be that, I think there's something there in that story.

Paavo Monkkonen 41:50
So you mentioned sample size, and so you're using survey data, I wonder how reliable you think these survey data are, and whether you think you'd find maybe different results iff you had a full dataset of everybody's employment history.

Jaclene Begley 42:04
I think the Health and Retirement Study is a really robust data source that, has a lot of questions on labor market participation, and housing values. That being said, it's all self-reported like you mentioned, and people have individual biases on their retirement status, of what their housing is worth, their expectations, how they answer these questions - I mean, it's a panel dataset so it follows people over time, and they do answer questions differently from wave to wave, which is, you know, interesting, too. So, it's a really robust data set to use but to your point, right, it's administrative data, and people may not know how much, I mean there's evidence in the literature (that) people don't know how much their house is worth. People misreport their housing values and seniors in particular, because they've lived in their homes for so long, (and) just don't have a good sense of the market value of their home. So there's going to be biases in the data that we can't control for or tease out.

Paavo Monkkonen 43:03
I mean, presumably, the claiming social security will be the least biased measure so I think it's useful that you include that as well.

Jaclene Begley 43:10
. Right? Hopefully, people are reporting that accurate.

Shane Phillips 43:13
They might not think of themselves as retired yet or thinking of themselves as retired even if they're working but you know, their claiming social Security or not, it's probably pretty reliable.

Jaclene Begley 43:22
Right, and since it's a panel, you can see like, did they say they started reclaiming it at a reasonable age, too, right?

Paavo Monkkonen 43:28
Right

Shane Phillips 43:29
Jackie, you touched on this a little bit but as I was reading this, I kept coming back to how important timing and really luck is when it comes to the benefits of homeownership. And I feel like this study really highlights that in some ways; there's the luck of whether you happen to be in a place with a strong and growing labor market, which is totally out of the hands of individual homeowners, and the timing of whether you bought in 2003, or 2007, or 2011, which just strikes me as something that regular people mostly can't, and really shouldn't have to game out so carefully when it comes to choosing where they live, such a basic and important fundamental decision. Ideally, people could move when it makes sense for them, their jobs, their families, whatever but when housing is used as a primary source of wealth, then it's inevitably going to be treated as an investment, and that's often in conflict with the consumption benefits of housing. Because something that's a great investment will become more unattainable as its price goes up, I think that's almost definitional. I guess where I'm going with this is just a wonder how you felt about these results, less as a researcher maybe and maybe more at a philosophical or an emotional level. I'll say for my part, they kind of bummed me out. In that I feel like in a better world, the value of people's houses wouldn't play such a big role in these important decisions like when to transition from one phase of your life to another.

Jaclene Begley 45:00
Yes, I agree with you that it's a total bummer. I feel like there are a lot of difficult questions around our societal, like emphasis on homeownership as a wealth building tool, but not a lot of like policy solutions or no perfect menu of solutions for how to handle these real questions about how do we deal with wealth at the end of life? How do we deal with the fact that housing is a consumption vehicle, but also like a wealth building vehicle? Paavo mentioned that we're not unique as a country for our reliance on homeownership, like this is a common thing that many nations have. But I think to your point, you know, there's still a lot to learn and the timing and sustainability of homeownership and who really benefits from appreciation particularly these questions about local nuances and markets and the timing and role of mortgages, racial disparities across groups, I think these are all really important questions that are gaining more traction recently but overall, are surprisingly under studied. There are a few handful of papers from the Great Recession period that look at, you know, is housing sustainable, did people benefit and who benefited from homeownership, but there's only a handful and I really think there's a lot more we could do to better understand these questions. And then we can start to think that, you know, prescriptive policy solutions that actually help targeted policy to help, you know, create solutions to these questions, or these issues. And I do think like our reliance on housing wealth is a problem, and we need more nonhousing related policy tools.

Paavo Monkkonen 46:33
Yeah, I mean I echo the call for more housing and demographics research, I think it's an understudied connection there. And we didn't talk about how housing wealth might affect younger households decisions about, you know, where to live, and how to form families and all these things but there's definitely a role of housing wealth and in many other aspects of people's lives.

Shane Phillips 46:52
And the Lewis Center is always accepting gifts and grants to study exactly these things.

Paavo Monkkonen 46:59
So turning back to more pragmatic concerns, I'm curious to know what you think could or should be done with these findings in terms of public policy. So I think we've still we've talked about some of the negative implications of the research but it certainly seems important that we know, this relationship between housing wealth and retirement decision really exists, which is something that previous research didn't identify. So what, you know, there are opportunities for reform that could come out of this work.

Jaclene Begley 47:27
So my personal opinion is that we need a better social safety net. You know, there's been a lot of reform since the Great Recession on particularly in the mortgage market but I think we need a better solution than housing as the main source of wealth for households. Or maybe if housing is the main source of wealth, we have some other more robust social security system that provides a better income security for households over time, we need to do more to ensure that older households have access to adequate resources as they age, including, you know, age-appropriate housing, housing that can be modified to allow for people who do want to age in place or a better option, that's housing for people who want to age in place, more housing supply would provide more diversity of housing units that would allow people to downsize and actually tap into their housing wealth.

Paavo Monkkonen 48:13
So like different kinds of housing typology in the same neighborhood. Wait wait so like apartment buildings next to single-family homes?

Jaclene Begley 48:23
Exactly!

Paavo Monkkonen 48:25
So one of the things, I mean you alluded to this idea of the reverse mortgage, can you tell our listeners what that is and why it hasn't worked?

Jaclene Begley 48:33
So the reverse mortgage is a tool to allow older households to tap into their housing wealth, so it's basically, it's a mortgage, similar to a HELOC, but you don't have to make payments on the mortgage, and the balance accrues until you sell the home or pass away. They historically have been underutilized, and there's been some changes to the HUD program in particular in the last few years. And there's a lot of growing research into why people aren't using reverse mortgages, or why people may or may not use key locks to tap into their housing wealth. But the take-up rates are low in general so I think, for now at least we can't rely on something like that to solve our problems. We probably just need lots of different options for households.

Paavo Monkkonen 49:19
And what about there was an idea at some point that was like an insurance hedge where you could buy some kind of product that was inversely correlated to your housing markets' fluctuation so that when your housing price goes down, this product's value went up?

Jaclene Begley 49:34
Yeah, I remember this too. This must have been in John Quigley's class, that we talked about that.

Paavo Monkkonen 49:40
I think it was this Robert Shiller dream product, but I realized,

Jaclene Begley 49:45
Yeah, as far as I read, I haven't seen it and that was more than a decade ago.

Shane Phillips 49:50
It sounds like the kind of insurance program that only an economist could love, the thing that just sounds so great in theory, but in practice, no one is interested in for one reason or another. Maybe because they just don't like to think about, they like that upside., I don't know. Maybe no one expects their home values to go down.

Jaclene Begley 50:06
In all fairness, though, I just think that we got to do more than rely on housing.

Paavo Monkkonen 50:10
Right

Jaclene Begley 50:10
We need a broader safety net of things that are not just focused on the home.

Paavo Monkkonen 50:15
Yeah, I wonder, maybe a second extension of this study is a similar study in a country that has a broader safety net because I wonder if in a Germany or something where negative housing shock would produce the same effects, that would be a double bummer.

Jaclene Begley 50:32
Yeah, I mean, that's a good idea, like looking at this in a country that doesn't have the same reliance on housing wealth, and has more of a safety net for retirees, and to see if you see the same kind....

Paavo Monkkonen 50:42
Because to some extent, you know, you're making this calculation about your future savings, the money you're going to have in retirement, you're going to pass to your kids, and then something changes. And so it's like, how much of it is psychological reaction to like, some loss aversion thing, where you're like, "oh, no, I have a little bit less", and how much of it is like actual rational calculation about, "Oh, that changes my plans, so I'm going to work a little longer". And I think in a country where, it's not like you're going to ever be living on the street, right, in a country that would definitely take care of your old age no matter what, you know, you still might have the psychological effect, even if you don't have the, like, rational calculus effect.

Jaclene Begley 51:19
Right, I think that's really fair, because that is something also the literature finds, right? There's loss aversion and housing prices, so people are less likely to sell their homes if they think it's not going to sell for the sticker price they had in their head. And that's one thing we can't unpack or untangle on this study. And using a different country setting where you could really figure out how much of this is loss aversion versus how much of this is we need this wealth or relying on this wealth, would be really helpful for understanding a policy solution.

Shane Phillips 51:47
Yeah, I do think it's important to just highlight what you're both gesturing toward here, which is that all of these decisions are kind of happening in a context where people have to be concerned that if they make the wrong decision, if they take their retirement too early, then they might die poor, like their money might run out before their life does. And because we have such a weak social safety net, you know, they might become homeless, they might, you know, not be able to afford medical expenses, what have you. And so I think, just keeping that in mind, but like there's very real consequences for making the wrong decisions,in part because we rely so much housing wealth, and have so little to back it up. I just want to underscore that again

Jaclene Begley 52:30
Yeah, that's really fair. And I think that that, you know, there's literature that shows that older adults don't downsize as they age and part of it, the explanation is the safety net aspect of it right? You may need your house in 10 years, you don't know how long you're going to live so people don't want to tap into that wealth.

Shane Phillips 52:47
All right, Jaclene Begley, thank you so much for coming on the Housing Voice podcast.

Jaclene Begley 52:51
Thank you for having me.

Shane Phillips 52:57
You can read more about Jaclene's research on our website lewis.ucla.edu. Shownotes and a transcript of the interview are there to the UCLA Lewis Center is on Facebook and Twitter. I'm on Twitter at ShaneDPhillips, and Paavo is at Elpaavo. Thanks for listening. We'll see you next time.