May 9, 2014
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)
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
Meta Brown, Federal Reserve Bank of New York (21:12)
- Does Parents’ College Savings Reduce College Debt?
Melinda Lewis, University of Kansas (18:11)
Alex Monge-Naranjo, Federal Reserve Bank of St. Louis (19:01)
Plenary Two Q&A (18:22)
Concurrent Session I
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)
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)
Trina Williams Shanks, University of Michigan (7:06)
Session Two Q&A (28:58)
Plenary Three — Homeownership
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)
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
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)
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.
John Sabelhaus: Great. Well, I am John Sabelhaus from the Federal Reserve Board and as it says at the bottom nothing I say here can be blamed on Janet Yellen. The key questions in these two papers, I think they're two really great questions. So one, did young people behave differently than others during the Great Recession and in ways that led to disproportionate financial distress? That's why we're here. But also the second paper, does acquiring a saving account contribute in a dynamic sense to improvements in young people's balance sheets?
So I have a few overarching comments. Both these are really great papers. They both really capture the goals and the overall spirit of this conference very well. Both clearly exhibit very careful statistical work on complex data sets. Both are working with data sets that have a longitudinal component to it, and that's difficult to work with but it does allow you also to look at groups and how they're changing over time. You're trying to explain how differences in a group at one point in time, how those translate into differences within that group at another point in time. And both really address tough questions that could have immediate policy implications.
My fourth comment, I think both are very suggestive about prescriptive actions, things that we could possibly do. But I also believe that a lot more research and testing is needed before we take the policy implications away from these papers. In short, I think we see strong correlations here, but I'm not sure that we see the causation that could be argued.
So somehow or another Bill Emmons and I ended up with the same literary reference and the same exchange, theoretical exchange. And I know it because I've been working with wealth data and lifecycle models for a long time. And Fitzgerald is said to have written that the rich are very different from you and me. And Hemingway is said to have answered yes, they have more money, meaning they're really not different. They just happen to be in that situation. So they meant things like character and values. That's what the exchange was about. But we boring economists, when we talk about lifecycle models we think about utility function parameters and risk aversion and things like that.
But I think there is a nice analog here. You know, are the young different? And there's a corresponding exchange so some would say the young are very different from you and me. And some would respond, yeah, they're at an earlier point in the lifecycle and that's the key difference between them.
The distinction here is that the second view, what I call the Hemingway view, focuses on how the environment faced by young people and their planning horizons are very different, and how macro shocks interact with those differences to lead to a series of outcomes.
So I want to focus on this as sort of a framework for thinking about these two papers. Rather than think about the details about how the econometric specifications were done I want to step back and think about looking at young people in the context of these overall well-known differences in environment and planning horizons and lots of other things. So I want to show you a picture that shows earnings uncertainty is very different for the young at all points in time. And they shift proportionally over time. The young have nonfinite retirement financial priorities. This is sort of obvious. So participation in retirement plans is predictably low. Wenhua talked about age 40 as being arbitrary. I don't think it's arbitrary at all. When we look at data on lifecycle wealth accumulation 40 is a magic number for a lot of reasons.
The third is young people live in a world where education is much more expensive. It may be more valuable as well but we need to think about the fact that it's just more expensive. And again a theme I'll come back to, the common macro shocks interacted with these understandable differences by age to generate predictable differences in outcomes over time.
So here's a series of a few pictures, a couple from my work and one from another. This is using the Social Security CWHS, Continuous Work History Sample, one percent Master Earnings File. One percent sample of all Social Security earnings records. And we're measuring two types of variances. This is from a standard lifecycle earnings equation that has shocks. It has uncertainty. So there's a predictable component of earnings but you also get shocks. Some of these shocks are transitory and those are on the left axis in the line labeled transitory. And some of these shocks are permanent, meaning when they happen they shift your whole profile for the rest of your life. Why do I show you this picture? It's really important for understanding young people's consumption and saving behavior because there are things about young people that you can't understand unless you see that they face very different uncertainty environments. So things like the buffer stock model, Chris Carroll's buffer stock model, works much better when you acknowledge that variances are very different across these age groups.
The second picture is from a paper that I do. As was mentioned I do a lot more thinking about middle aged people and old people, thinking about retirement accounts. And this is from a paper about early withdrawals from retirement accounts, something I think is a big problem right now. And so here we have three age groups. We have two data sets. The stars are SCF years, so 2004, 2007, 2010. The lines are from IRS tax data. So there we can fill in every year. And what we see is that the patterns are very much the same, minor differences. But what we also see is a big jump that happens and in fact if I drew this, if I redid this graph for age 40 you would see that's where it really jumps, where participation really surges.
But then the other thing we see is that the macro environment is such that participation dropped off for all of these groups and maybe a little bit less for the young than it did for the middle aged and older groups. And the reason is the young who are there who are participating is because they've been auto-enrolled and other sort of nudge type devices. So the third one I could have used SCF data but I will rely on Meta Brown and her colleagues in New York to show you that total student loan balances are increasing for all age groups. This is something they've exploded—we all know this—over the past decade. But every one of these bars is getting wider by roughly the same proportion. And in fact I think there's actually been somewhat of a shift towards more borrowing among middle aged and older families as opposed to younger families.
So confessions of a Hemingway fan. I really believe that the young are just middle aged people waiting to happen. I really enjoy reading Fitzgerald. I do. I really love—The Beautiful and Damned is one of my favorite books. And I enjoyed these papers.
But I really think in order to understand these differences between the young and the old we really need to think about their lifecycle behavior. We need to think about the particular environment in which we're studying these. I also think we need to step back and say sometimes a lot of negatives that we've identified also have associated positives. So the classic example is that the earnings and employment prospects are much better for people who went to college and/or back to college. It's why people do it. There is variance in the outcomes. We know that. But on average it's generally a good thing to do.
There's also some lifecycle differences in improved outcomes during this period of the Great Recession. So my colleagues from the Board, Lisa Dettling and Joanne Hsu, talked yesterday about how young people not being in the housing market between 2002 and 2006 actually in some ways gave them an advantage. They bought into a much cheaper housing market. And again their behavior when they bought a house is sort of associated with things like getting married and going out on your own. These might be being delayed a little bit by the fact that they have higher-education debt, but in general the lifecycle patterns are very clear.
So when thinking about this again we'll have a few things to say specifically about the papers but before I do, the last little bit of framing is about policy tradeoffs. And I think when we think about policy actions to overcome what we perceive as behavioral shortcomings we really need to establish why those behavioral shortcomings exist in the first place. So education loans, they may not be perfect but they're a pretty efficient way to use risk sharing to invest in specific human capital. Education has become more expensive. That's an issue we should be thinking about as a society. Why has it become more expensive? But given that it's become more expensive education loans are actually a pretty efficient way to do it. We may need to tweak the rules. Think about income-based repayment, lots of other things. We're focused on these. But moral hazard principles tell us that the alternatives that sound better sometimes really aren't. And the idea of people making an investment and taking on some of the risk is something that has an appeal from an economics perspective.
I also believe, thinking about the second paper, consumers can and should be protected from bad behavior of financial institutions. What CFPB is doing is absolutely what needs to be done. But I think we need to be a little bit more circumspect about telling individuals what to own. I have three kids between the ages of 18 and 24 and they're all still pretty much guaranteed to do the opposite of what I tell them to do.
So the first paper, Financial Decisions of the Young Versus Old, the abstract says and I think the takeaway message, younger households are significantly different from older households. This is probably what got me on the Hemingway-Fitzgerald kick. I was reading this and one of the numbers Wenhua showed is that the transition into liquid asset poverty between 2007 and 2009 is basically the same for people less than 40 and over 40, which I'm flipping through the paper very quickly trying to understand how these two go together. And what happens is they're running a transition equation into liquid asset poverty and this equation has many demographic and economic controls. So the coefficient on the dummy for being less than 40 is positive. You've got a bunch of other variables and then you've got a coefficient on age that's positive. And this is interpreted as the young being more likely to become asset-poor given their economic and demographic situation. The way we're trained to talk about this is controlling for all these other factors they're more likely to become asset poor.
So I contend there is an alternative interpretation, which is that that coefficient doesn't tell us what we do know. It actually tells us what we don't know. And effectively what we're doing is we're imposing homogeneity on all of these other variables, that the young should react to income and to all of these other variables in the data set exactly the way that the old should. But if you step back and you accept my framework that the young are at a different point in the lifecycle and there may be things in the data that we're not seeing, right? They're different but in ways that are not captured by this particular econometric specification and data set. So to think about this as a takeaway message that they're more likely to transition into asset poverty I think is a little premature. I think we need to think about exactly why those other differences might exist.
So on acquiring a savings account the set data from the 1990s tells us that young people who acquired a saving account were also more likely to subsequently—this was very carefully done—subsequently acquire substantial equities and retirement accounts. And again from my perspective from a lifecycle model, knowing that different types of people get these different types of shocks and these populations are separating in ways that are not observable to the econometrician, that's what a permanent shock is, right? Some unobservable separation of the population into wider earnings distributions as a cohort ages so there's some correlated but unmeasured factor that's basically underlying both the decision and the outcome.
And I don't doubt that savings accounts could have a positive effect, that somehow putting people in a savings account could have a positive effect on their outcomes, but I think that this is one of those things that just really calls out for experimental evidence and a focus on the specific behavior that we're trying to encourage by this. And one of the things I would ask—again maybe it's the three young people I'm still responsible for—are savings accounts really the key for young people today? Is this something that it rings very true in our ears but one of my former colleagues at the Board left the Board. He left the Board and he went out to California to live in Silicon Valley and work on apps that are designed for young people to help them make good decisions with respect to their finances. And the sorts of decisions we want them to make that might be even more first order is to not borrow too much, not borrow money that they can't repay and things like that.
So if I had to choose one sort of policy if I was thinking about using again in the spectrum of nudging versus education, behavioral economics shows us that we can nudge people. We could in principle nudge people into saving accounts. That makes good sense. But I think we should look at all the tradeoffs before we're focused on certain outcomes. And there may be other things that may be more directly linked to what it is we want them to do.
I will admit this is a little bit of a macroeconomist's perspective when I think about the Great Recession and I think about dynamics and what really happened to the subset of the population that was most affected. I focus more on things like being too indebted as opposed to not having a savings account. So if we can nudge people to open savings accounts we can certainly improve their outcomes along other dimensions and I think before doing the nudging we should explain the choices by some people for example not to have certain types of accounts and what it is that we're really trying to get at.
And thank you very much. I very much enjoyed the papers.