Skip to content

Meta Brown, Student Loans and the Economic Activity of Young Consumers

May 8, 2014

previous  view previous video in this series | view next video in this series  next

View more about the conference »


   James Bullard, Federal Reserve Bank of St. Louis (5:54)

Keynote Address

- Introduction
   Ray Boshara, Federal Reserve Bank of St. Louis (4:05)

- Keynote Address
   Neil Howe, Founding Partner and President, LifeCourse Associates and President, Saeculum Research (36:03)

   Keynote Q&A (11:04)

   Extended Interview with Keynote Speaker Neil Howe (28:43)

Plenary One — A Micro and Macro Look at Younger Americans' Balance Sheets

- The State of the Balance Sheets of Younger Americans
   Lisa Dettling, Board of Governors, Federal Reserve System (14:25)

- Links Between Younger Americans’ Balance Sheets and Economic Growth
   William Emmons, Federal Reserve Bank of St. Louis (21:40)

- Discussant:
   Steve Fazzari, Washington University in St. Louis (18:14)

   Plenary One Q&A (15:06)

Plenary Two — Student Loans

- Student Loans and the Economic Activity of Young Consumers
now playing  Meta Brown, Federal Reserve Bank of New York (21:12)

- Does Parents’ College Savings Reduce College Debt?
   Melinda Lewis, University of Kansas (18:11)

- Discussant:
   Alex Monge-Naranjo, Federal Reserve Bank of St. Louis (19:01)

   Plenary Two Q&A (18:22)


   Michael Sherraden, Washington University in St. Louis (2:52)

   Clint Zweifel, Missouri State Treasurer (6:26)

   Tishaura Jones, City of St. Louis Treasurer (6:14)

Concurrent Session I

- Moderator:
   Julie Birkenmaier, Saint Louis University (4:36)

- Toward Healthy Balance Sheets: The Role of Savings Accounts for Young Adults’ Asset Diversification and Accumulation
   Terri Friedline, University of Kansas (22:23)

- Financial Decisions of Young Households During the Great Recession: An Examination of the SCF 2007-09 Panel
   Wenhua Di, Federal Reserve Bank of Dallas (19:00)

- Discussant:
   John Sabelhaus, Board of Governors, Federal Reserve System (14:38)

   Session One Panel Response (6:57)

   Session One Q&A (5:41)

Concurrent Session II

- Impacts of Child Development Accounts on Change in Parental Educational Expectations: Evidence from a Statewide Social Experiment
   Michael Sherraden, Washington University in St. Louis (13:09)

- Trends and Patterns in the Asset Holdings of Young Households
   Ellen A. Merry, Board of Governors, Federal Reserve System (15:19)

- Discussant:
   Trina Williams Shanks, University of Michigan (7:06)

   Session Two Q&A (28:58)

Plenary Three — Homeownership

- Moderator:
   Todd Swanstrom, University of Missouri–St. Louis (5:54)

- Homeownership and Wealth Among Low-Income Young Adults: Evidence from the Community Advantage Program
   Blair Russell, Washington University in St. Louis (15:57)

- Aggregate and Distributional Dynamics of Consumer Credit in the U.S.
   Don Schlagenhauf, Federal Reserve Bank of St. Louis (21:53)

- Discussant:
   John Duca, Federal Reserve Bank of Dallas (13:59)

   Plenary Three Panel Response (6:40)

   Plenary Three Q&A (8:46)

Plenary Four — Economic Mobility

- Moderator:
   Jason Purnell, Washington University in St. Louis (2:00)

- The Balance Sheets and Economic Mobility of Generation X
   Diana Elliott, Pew Charitable Trusts (17:58)

- Coming of Age in the Early 1970s vs. the Early 1990s: Differences in Wealth Accumulation of Young Households in the United States, and Implications for Economic Mobility
   Daniel Cooper, Federal Reserve Bank of Boston (17:11)

- Discussant:
   Bhashkar Mazumder, Federal Reserve Bank of Chicago (16:15)

   Plenary Four Panel Response (3:14)

   Plenary Four Q&A (19:02)

Closing Reflections: From Research to Policy

   Michael Sherraden, Washington University in St. Louis (15:59)

   Ray Boshara, Federal Reserve Bank of St. Louis (9:30)

Thank You / Adjourn

   Julie Stackhouse, Federal Reserve Bank of St. Louis (5:59)


Below is a full transcript of this video presentation. It has not been edited or reviewed for accuracy or readability.

Jim: Well good afternoon again, everybody. And thank you. We are, you know, we’re making conference history here. We’re on time for everything so far today. So great credit to everybody here and to the event staff as well. What an interesting morning discussion and it was interesting even just hearing some of the conversations during the break and we feel very fortunate that not only were those conversations happening, but it clearly, it seems like a very nice segue to our second discussion, our second plenary session on student loans today, which came up multiple times even this morning. So I’m very happy to introduce our moderator for this session, Rajeev from the University of Missouri. Thanks, Rajeev.

Rajeev Daroli: Thanks, Jim. Well my name’s Rajeev Daroli. I’m from the University of Missouri and I’m an assistant professor there. I thank Ray and Bryan for having me participate in this great event. Somebody who studies loans at, I think it’s really important and I appreciate that what student loan debt is being talked about today is in the broader context of other financial and personal decisions. And I think that’s really important. I don’t have to tell people in this room. I’m sure we’ve all seen the well-publicized trends of increasing student loan debt and increasing student loan defaults. What’s frustrating but at the same time a source of inspiration for me as a researcher is that it’s so not that well understood, both the causes and the implications of student loan debt. And so, you know, for the researchers in the audience you all know that identifying causes and effects is extremely difficult, and especially in the context of, like, student loans because we have changing higher education supply, we have an evolving student bodies, difficult labor conditions, later market conditions in the past couple of years. And this is all interrelated at a time when we’re trying to understand the complex cost-benefit analysis decisions of students who often aren’t acting in ways that we think they may, that they should act.

So even though it’s really difficult it’s really important that we understand these causes and implications of student loan debt so that we can decide whether or not we want to design policy, and also what’s the appropriate way to design that policy? So a person would have to be a part of this, because I think we have two papers here which provide a very nice contribution to our understanding here.

The first paper, which is by Zachary Bleemer, Meta Brown, Donghood Lee, and Wilbert van der Klaauw, the title is actually a little bit different. It’s called “Debt, Jobs or Housing: What Keeps Millennials at Home.” This actually talks about one of the implications of rise in student loan debts, which is an increasingly important thing that we need to know. The second paper, which is going to be presented by Melinda Louis, with co-authors William Elliot, IlSung Nam, and Michal Grinstein-Weiss, is called “Does Parent’s College Savings Reduce College Debt?” And this is another nice contribution, both trying to understand what is one of the causes of increasing student loan debt, but at the same time also points to potential policy solutions to increase in student loan debt as well. So I’d like to welcome Meta up here to present her paper. We’ll follow a similar format at the other session. Each presenter will have about 15 to 20 minutes and then the Alexander Monje Noranjo will also have about 15 to 20 minutes as our discussion. He is from the Federal Reserve Bank of St. Louis. Then we’ll have time for questions.

Meta Brown: Thank you, and thanks so much to the organizers for this terrific conference and for including us in it. We feel very privileged to, I feel very privileged. So I want to talk about “Debt, Jobs, or Housing: What Keeps Millennials at Home?” You can see a little bit of mission create from the program to the title of this presentation. This was sort, the mission was to look at housing and student debt. And this is what really caught our attention and our interest. This is very new, very recent work with these New York Fed colleagues and I’m here from the New York Fed. It’s highly preliminary and complete. This is sort of news to us and we think it’s very exciting so I’d love to hear feedback and questions and also please don’t cite us until you talk to us, although we love to talk about our stuff, as everyone does. And this, these slides represent obviously our argues and perspectives on the question and not Janet Yellins, at least not that I know of.

Okay. So young Americans are increasingly living at home with their parents. Various sources suggest a steep upward trend in co-residence with parents for 20-sometimes. Various researchers work in the CPS, a recent Pew Research Center study. The latest Sallie Mae annual surveys of college students say that college students are increasingly staying home during college to sort of cheapen the experience. Oops, that’s not the way to say that. But you know what I mean. And so the idea here was to take the Federal Reserve Bank of New York’s Consumer Credit Panel, a very large, very recent and administrative data source and document the living with parent’s trend in the state. A source, and investigate relationships among young people’s residence choices and local economic conditions over the ’99 to 2013 panel that we can observe. With the eventual goal, we hope perhaps, of addressing the drivers of millennials homeward trend. Is this about jobs and a lack of available, good-use employment opportunities? Is this about patterns in housing prices and the cost of living independently? Is this about the rising, the escalating reliance on student debt among young people?

And so there’s been a lot of popular discussion of delays in household formation, delays in first-time home-buying. We hear a lot from the National Association of Realtors on the drop and the share of young people buying houses, or the share of the market for houses going to first-time homebuyers from about 40% pre-crisis to 30% these days. And some discussion, particularly from the CFPB on the link between household formation, home purchase and the student debt growth. This is all in response and very interesting current research. John Duca, who I think will be here tomorrow, he has a really neat paper noting the close commune and a 30-year time series on American adult’s co-residence with their parents and the poverty rate, which actually has been very striking. It’s quite insightful. Because great minds think alike, I’ve been working with this with my CCP team at the New York Fed on this stuff, and our wonderful CCP researcher colleagues at the board are also thinking about very similar ideas. And so you saw Lisa and Joanne earlier. They’re also looking at the delinquency of young people’s debt and the decision to move home or move away. And so all of this is very closely related to the issues we’re thinking about.

So what do we do in this context? We take the, again, the Federal Reserve Bank of New York’s Consumer Credit Panel, which is based on a 5% random sample of Equifax, Credit Bureau, Credit Reports, where Equifax is one of the three leading national credit bureaus, as you may well know. And so these data, it’s a huge data set, of course. My co-author Liam van der Klaauw, effectively the PI’s of the data set. You can talk to them if you’re interested in these data at any point. They contain everything you would think would be on a credit report, and also very fine geographic information down to the census block and age measure, a risk score, like a FICA score, and individual and household level sort of information and groupings.

And we put all this together to look at patterns in co-residence and economic conditions that young people face. It turned out that on measuring co-residence with parents was maybe unsurprisingly challenging in the Consumer Credit Panel, because that’s not the point of credit reports, of course. But also surprisingly challenging—oops, sorry. I skipped one. Well, then I’ll just hammer. Sorry. Surprisingly challenging even in survey data. So I can tell you that our measure of co-residence with parents are not on this slide, is, we take a young person, for example at age 25. If she is living with an elder 15 to 20 years older than her and not in a very large household, but in a household with 9 or fewer adults, we’re calling this person co-residing with parents. We step over to the CPS to vet this approach. It seems reasonably credible and I could spend a very long time telling you about why I think it’s credible, but I’ll just tell you we’d be happy to chat more about that later in this brief talk. So this is our means of inferring co-residence with parents in the CCP.

As we go forward we pull in also data on local house prices here from the correlogic at the zip code level. These things are indexed to be 100 in every zip code in January of 2000 and in our sample across zip codes and over the 14 years that we look at, the standard deviation of house prices is 80, which is to say 80% of the January 2000 house price levels. So we’re working with a lot of house price variability over this time period in these regions. We include some state level youth unemployment measures that we calculate from the CPS, which range from 2% to 22% over our period of time, over our 14-year period in the states that we look at, and we also roll in some BLS county-level unemployment data, which has a standard deviation of 3.7 percentage points. So that’s, those are all the data. Let’s go through, like, eight pictures in a couple regressions to tell the story.

Here’s what we’re getting for the trend in co-residence with parents from CCP. We see an increase from 19 among 30-year-olds, and throughout this talk 30-year-olds will be green and 25-year-olds will be blue. The obvious color choices. So we see an increase from 19% of 30-year-olds in the CCP in 1999 to above 30% appearing to co-reside with parents based on the measure I’ve described to you, and a similar upward trend, but obviously a much higher level for 25-year-olds from 29% to upwards of 45% living with parents as we measure it in the CCP over this time period.

So if folks are, if young people are more and more living with parents, where aren’t they living, so to speak? This figure, again, blue for 25-year-olds shows the steep upward trend in living with parents, and at the same time as we inferred in the CCP, a downward trend in living alone, which is to say living at an address that has only 1 adult. And a similar downward trend in living with 2 or more adult roommates. What we don’t see is a downward trend in what is a good approximation of marriage and cohabitation. So the living with an adult of similar age, if you go over the CPS, it’s a very good approximation of being married or cohabiting with one partner. And we see this occur stable over the time period. So it doesn’t appear to be, at least based on the CCP, the case that young people are increasingly moving home with parents and delaying moving in with marital or cohabiting partners. That seems quite stable. What we’re seeing is less and less living alone and living with groups of roommates.

And the story is more or less the same for 30-year-olds, although of course levels are different. Fewer 30-year-olds live with parents. More of 30-year-olds live in married or cohabiting households. But the patterns, the trends seem to be quite similar.

Homeownership, of course, is a very important residence choice for these young people we’ve talked about a lot so far. So here’s our pass at the homeownership question. We take a look at the following. The National Association of Realtors note that historically 30 has been approximately the median age of first home purchase in the U.S. We look at 30-year-old’s homeownership as inferred from home secured debt. Very few 30-year-olds own their homes outright. So it’s a reasonably good measure. And we see an increase from 2003 to 2007 of maybe about 2 percentage points in the homeownership rate among 30-year-olds, but then a very steep decline following the Recession and the housing crisis of about 9 percentage points in our sample from 2007/8-ish to 2012 in homeownership. 25-year-olds obviously, and as discussed earlier, don’t own homes quite as often as 30-year-olds, so we have 15% homeownership rate from 2003 to 2005, fairly steady here. And actually the decline in homeownership for the 25-year-olds begins much earlier so they have a long window of decline from 2005 to 2012. And more recent data we have shows this continuing into 2013 so that they go from about 15% homeownership in 2005 down to below 10% by 2012.

So putting all this together, we see this substantial and ongoing trend of moving in with parents away from purchasing homes, and in general postponing in these more recent cohorts, the postponement of these major, early adulthood milestones.

Okay. And of course we’ve talked a lot about the growth of student debt. The Consumer Credit Panel we believe is a good place to measure student debt and we said a lot publicly about student debt based on the panel, so here’s one picture of what’s going on that might be particularly relevant to this discussion to student debt among young people. Because as we’ve heard, many older cohort members are also borrowing in the student loan market. Here we see from 2003 to 2013 in the lighter blue curve on top, at the age of 25 an increase from 25% of 25-year-olds participating in the student debt market to above 45% of 25-year-olds participating in the student debt market. So the prevalence of student borrowing over this 11-year-period has changed enormously. And the mean debt conditional on borrowing increasing from above 10,000 to above 20,000 over that same period. So we see an escalating reliance on student debt for higher education, an escalating reliance on student debt at the age of 25. So one immediate question is, what role does this accumulating student debt burden play in the housing decisions and the residence decisions of young people. And here’s a picture I pulled from a blog that we posted on the Liberty Street Economics Blog. Theories last year that we’ll be updating I believe on Tuesday. Excuse me. And this depicts homeownership. It’s green at age 30. Homeownership at age 30 over the same period divided into those who have a history of student debt in the dark curve, and those who have no history of student debt, the lighter curve. You can see from 2003 to 2008 these two curves move along approximately parallel with 33% of those with student debt experienced owning homes, and about 29% of those non-student borrowers owning homes. This is a pretty stable relationship and it makes sense. People with student debt have been to college. They’re probably more likely to own homes.

And then suddenly at the [tape blanks]. You hit the Recession and the housing market crisis and everybody’s homeownership rate drops off a cliff, but the homeownership rate of student borrowers drops considerably more steeply so that by 2013 we’ve seen a drop of 11 percentage points in the homeownership rate of 30-year-olds with student debt history on a base of 33%. And this is much steeper than the drop we see for non-student borrowers so that by 2013 student borrowers are even substantially less likely to own homes than non-student borrowers, reversing a long standing trend. Of course none of this analysis is causal, but it does, it begins to suggest an important role for student debt in the residence choices and homeownership choices of young people.

Okay. So we have five minutes to go or less. I’ll just put up a very ugly picture to motivate a couple of regressions that I’d like to describe to you quickly and then wrap it up. So the ugly picture, the purple series shows the house price index over the period with the pronounced boom bust and we can see a strong recent recovery in house prices. The red lines, the red series indicate total unemployment and youth unemployment similarly showing a pronounced boom bust and then recent ongoing recovery. And the yellow series shows the increasing reliance on student debt of young people. And so the question is, how do these green curves, living at home with parents and owning a home, relate to these various changes in economic conditions.

So I’ll just tell you briefly about a couple of regressions. So here we model the decision to move home to parents for someone who was living independently between the ages of 23 and 30 where the outcome is 1 if you move home. The sample is conditioned on everybody living independently in the community. We’ll look over a two-year window. And we model this as dependent on, among other things, changes in the local economic conditions at the independent living site. Right? So over the two years how does total unemployment in that location change? How does youth unemployment change? How do house prices change? And then also we take a measure of the sort of student debt reliance of the latest cohort [tape blanks] in that location.

And then after this we’ll do something analogous with the decision to move away from parents to independent living. Okay. So we’ll start out with the decision to move home to parents from an independent state. We see youth unemployment is a very important determinant of the decision to move home to parents so that these point estimates imply that a one percentage point increase in the state-level youth unemployment, it leaves two, if you by argument, .24 percentage point increase in the probability that this young person is going to move home over the two-year period. So it makes sense. Every place we look, youth employment markets leave people to move, young people move home when they’re living independently, and weak youth employment markets, and they move away when they’re living with their parents in weak markets.

But we can also look at conditioning on youth employment opportunities. We can look at local economic conditions more broadly. We can think about unemployment rates coming down, house prices going up. In other words, a locality that’s doing better over the two years than it was before. What does that do to the decision to move home? It appears that a young person living in an up-and-coming and rising economy is more likely to move home, particularly through the cost of housing. This house price estimate, recall, there was an 80-point standard deviation in the sample. If we think of adding 10 points to your local housing market index, this is associated with a .28 percentage point increase in the probability that you as a young person move home with your parents. All of which sounds a little bit puzzling but we interpret this as evidence that a local economy that’s becoming more expensive drives the young person home.

Also, maybe not unexpectedly, states with higher student debt reliance among grads are much more likely to have young people moving home with parents.

So we do the same thing for the decision to move away from parents. Again, weak youth employment conditions lead to much more moving home to parents. As you can see from these very large point estimates. Local economic conditions in the parent’s location now. So if you think about local economic conditions for the parents improving, for example, unemployment rates going down, house prices going up, this leads the parent, the child to be more likely to move away from home. So economic conditions where the child is located are improving and yet the kid goes away, which we interpret to be evidence that at least in one story parents are bank-rolling the move-out of these young children and when their economic circumstances improve they’re more able to buy the kid out of the house. When their economic circumstances weaken, they might have more trouble moving the child out of the house, and so the child winds up staying home.

Excuse me. And again, oops. Student loans are associated with more living with parents. So states that are much more student loan reliant per graduate are places where kids are much less likely to move away from their parents. So to conclude, from ’99 to 2013 the Consumer Credit Panel youth show a persistent upward trend in living with parents, accompanied by downward trends in homeownership, living alone and living with groups of roommates. Student borrowers in particular retreated from housing markets in the wake of the Great Recession. Trends in co-residence with parents most closely matched the escalation of student debt. In homeownership we see more business cycle sensitivity. And failure of the youth residence choices to respond to the recent labor and housing market recoveries to us remain a puzzle. Based on our regression results, our transition analysis reveals that moves out of home are most responsive to youth labor market conditions. That strengthening parent local economics tend to drive youth out. Strengthening youth local economies, however, tend to drive youth home, which is to say that it’s not aggregate fluctuations and macro-economic conditions that will be explaining the trend toward home, but perhaps the relative prices and success of local economies where parents tend to live and children tend to live that’s going to be driving this trend.

And finally high local student debt cost of college are associated with substantially higher rates of moving home for young people, and a failing to move home for those living at home. I’ll stop there.