Tom Heintjes: Welcome to another episode of the Economy Matters podcast. I'm Tom Heintjes, managing editor of the Atlanta Fed's Economy Matters magazine, and today we're joined once again by one of our regular guests: Melinda Pitts, an economist at the Atlanta Fed and the director of the Atlanta Fed's Center for Human Capital Studies. Melinda, it's always a pleasure to have you back on the podcast.

Melinda Pitts: Glad to be here.

M. Melinda Pitts, a Research Center Director of the Center for Human Capital Studies at the Atlanta Fed, during the recording of a podcast episode.

Photo: David Fine

Heintjes: Melinda has coauthored a brand new research paper titled "Driven from Work: Graduated Driver License Programs and Teen Labor Market Outcomes." That's a mouthful, but basically it looks at driver's license restrictions for young people, and their impact on their labor force participation. Melinda, I recently sat in on a detailed discussion of this paper with your economist colleagues here at the Atlanta Fed, and since then I've really looked forward to getting you into the studio so I could talk to you about it—so thanks for coming back on. There's a lot I want to get into, but first I should mention your coauthors on this paper, Laura Argys of the University of Colorado at Denver and Tom Mroz of Georgia State University, who is also a research fellow here at the Atlanta Fed. I want to give them credit for their research on this paper.

Melinda, I was intrigued by this research because it's something that so many of us can relate to. I remember when I got my license as a teenager and could drive to my first job, and it really opened up a lot of options for me—and for my poor parents, who no longer had to be my chauffeurs. But from what I see in your research, some states have placed restrictions on novice teen drivers, and this has affected their ability to be part of the labor force.

Pitts: Yes. It's always nice when our research can be relatable. I think this is part of the reason I enjoy working on microeconomic issues, because it's about decisions in the household. And I do also remember the day I turned 15, I got my license—I lived in Mississippi. I had the same privileges as everyone else, and I was very overconfident in my abilities—which terrified my mother…rightly so. And so I was very pleased that, when my son turned 16 here in Georgia, he didn't have that same freedom. And yes, it definitely is going to impact the labor force, because anytime you restrict someone's freedom, for whatever reason, you limit what they can do.

Heintjes: Well, this sort of situation sounds like a case maybe of good intentions and unintended consequences—would you say? Novice drivers are statistically the most likely to get into auto accidents, so legislators probably saw placing restrictions on their driving as having some sort of benefit for them—which it does. It might have in some ways, just not their workforce participation; right?

Pitts: Right, teen drivers are more likely to get into accidents. At the beginning of our period of analysis, in 1995—which is before most states had GDLs, or graduated driver's license programs, in place—and just going on accidents that were reported to the police, teens were involved in approximately 3,000 fatal accidents and over 1 million nonfatal accidents.

Heintjes: That's nationwide.

Pitts: That's nationwide. And that's about three to four times larger than prime age adults, middle age adults. So it's a big differential, and that's just in one year. So the driving restrictions have helped. There's been a decline of about one third for fatal accidents and about 60 percent for nonfatal accidents. So there's definitely been a benefit in terms of lives saved, reduced injuries, and less money spent on car repairs.

Heintjes: Yes. I should note that we're going to be referring to "GDLs" a lot throughout this conversation, so people should note that those are graduated driver's licenses. And I want to get you to briefly define what a graduated driver's license, or GDL—as you call them in your paper—is. Are these in many states, and do they differ among states?

Pitts: They are. They're in every state right now—they weren't when we began. But what a graduated driver's license is, is there are three stages to getting a license, so you gradually get your license. First is the permit stage, where you practice driving with a licensed driver. Then there is the restricted stage, which is the focus of our research, which has the characteristics of curfews, which range from 8 p.m. to 1 a.m.—although I'm not quite sure 1 a.m. is a binding curfew for a 16-year-old—and then it has passenger restrictions. So for example, in Georgia, the first six months after you get your driver's license you can't have any passenger under the age of 21 that is not related to you. And then the second six months you can add one passenger, and then the second year you can have three passengers under age 21. They do somewhat differ among states—like I told you, the curfew expanded, and some people do this for two years, like Georgia. Other states only have it in place for six months or short periods of time.

Heintjes: Melinda, before we get too much further into this topic, let me take a step back and ask you, what led you to initially look into this? I know you watch rates of labor participation closely. Were you seeing something among the rates of teenagers in the workforce that maybe caught your attention?

Pitts: Absolutely. We saw a huge acceleration in the decline in teen labor force participation [LFP], beginning in the late 1990s—both in absolute terms, and relative to the rates of participation in other age groups. So there was almost a 20 percentage point decline. In 1995, about half of teens participated in the labor force, and in 2015, that had dropped to below 30 percent.

Heintjes: Wow, that's huge.

Pitts: It was a huge drop. And it's even bigger when you think about it relative to prime age adults. In 1995, teens were about half as likely to work compared to a 21- to 54-year-old. Now it's one-fourth as likely to work—so this is a significant decline in the labor force. And obviously as a labor economist, I care about what's going on, and the Fed—we have the dual mandate—we care about what's going on with the labor force. And so we thought this was worth looking into. And this is where it pays sometimes to use your personal experience: I had seen how it impacted my son and his friends, and so I wondered if it was big enough to explain some of the decline. It's important for us to understand the reasons behind these declines.

Heintjes: Sure. Let me ask you: Does your research indicate that the GDL has a meaningful impact on teenage participation in the workforce? When do we start to see that impact?

Pitts: We do find a meaningful impact. We find that almost half of the decline in LFP since 1995 in the states in our sample can be attributed to the presence of GDLs. As states implemented their plans at different times, the share of the decline associated with the GDL was increasing throughout the time period of our analysis. And then once all states had the plans in place, the decline appears to have leveled off.

Heintjes: Interesting. Well, there's a body of research—and in fact, you reference this in your paper—that links some work done during high school with increased earnings later in life. So not being able to work during high school, or maybe being forced to work less, is not necessarily a trivial matter. It can have implications for down the road, so to speak—right?

Pitts: Yes, but they are definitely being incentivized to work less, because the benefit of work has declined because the cost of travel has increased. You mentioned your parents having to be your chauffeurs: so if you can't drive after midnight, and you have the curfew, mom or dad has to come pick you up—so that's a big cost. But you are correct: moderate hours of work during high school has been found to be associated with both higher grades and higher wages in early adulthood, and that impact on wages has been found to persist up to ten years.

Heintjes: Wow, so it does reverberate in a meaningful way, interesting. Let's get into the nitty-gritty a little bit: tell me about some of the specific groups you looked at, and what you were looking for, or looking at, within these groups.

Pitts: Our main analysis focuses on drivers aged 15 to 19 and examines the impact of being covered by a GDL. And like I mentioned before, we find about half of the decline in LFP for the states in our samples can be attributed to GDL, with the effect being slightly larger for males than females. We also looked at what happened when teens aged out of GDLs: Did they just go right back to the same patterns? But we're finding a little bit of a persistence in the effect, so that means that teens formerly covered by GDLs participate in the labor force at a lower rate than teens that were never covered by GDLs. And there are a couple of reasons that we came up with—the first would be persistence. Behavioral norms were established, and it's slow to change.

Heintjes: That's what they've known.

Pitts: Yes, this is what they know. But there are also some suggestions that teens are delaying licensure. There are some surveys that have found a 20 to 30 percent decline in licensure during this time period for teens, with cost and GDLs being mentioned as contributing factors.

Heintjes: You mentioned looking at 43 states, so your data was not just in the southeastern states. I imagine with states' guidelines for licensure varying a bit, it must have complicated your work, if apples weren't always apples from state to state, in terms of legal driving for novice drivers.

Pitts: Yes, we did use national data. There were eight states that already had GDL plans well in place before our period of analysis, so we could not include them. But the remaining 42 states, plus the District of Columbia, are a part of the analysis. The details of the programs definitely differ across states, as does the timing of the introduction, but from a statistical point of view, that's a good thing. It reduces the probability of another unrelated change driving these results.

Heintjes: Got it. So actually, that's a nice segue into my next question: I wanted to talk about the actual data you used. How did you gather the data, and how available was it to you in the conducting of your research?

Pitts: All the data we use in this analysis is publicly available, which is nice. The information on the GDL programs came from the Insurance Institute for Highway Safety. They have a great website that details all the differentials of the states, and a history of the GDL programs going back to 1995. The information on the teens and their workforce behavior came from the Current Population Survey [CPS], which is a joint product of the Bureau of Labor Statistics and the U.S. Census Bureau. It's the monthly survey of households and the data we use for key labor market statistics like the unemployment rate and labor force participation rate, among others—and as I said, all of this is publicly available.

Heintjes: Yes. I often see your work referring to the CPS, so I know it's a well you drink from frequently.

Pitts: Yes. And when we produce Labor Report First Look, and these other tools, that's where a lot of our data comes from.

Heintjes: How granular could you get with these GDLs? I know some states restrict in half-year increments, and in some states, drivers who are 15 and a half aren't able to drive with the same liberties that 16-year-olds are, for example. How were you able to account for that sort of "partial year" situation?

Pitts: So we were able to get very granular. The CPS is a monthly survey, and we survey households four months on, eight months off, then four months on. So we followed these teens over time, over a 16-month period, and we match the driving restrictions that would be in place down to the age in months. We assigned everyone the GDL plan based on their earliest potential driving age—so we don't actually know if they have a license. That's not a question on this survey, but we know they have the ability to have a license. For example, we could have someone age 16 years and five months with a different GDL restriction than someone age 16 years and six months, in the same state and month in the sample.

Heintjes: Wow, that is impressive granularity. So is this situation a zero-sum situation, in some ways? If novice drivers are losing out on opportunities to participate in the workforce, are other people benefiting?

Pitts: Yes. Other research has found high elasticity of demand for teenage workers. What that means is if the cost of hiring a teen goes up relative to other age groups, employers will be highly likely to gravitate to the other groups. For example, it appears that the impact of the GDLs was exacerbated by the Great Recession. Perhaps as firms had increased availability of older workers that did not face any driving restrictions, they were able to replace the teens with these older workers. So there may be a little of that, but it's likely not completely zero-sum given the higher cost of older workers.

Heintjes: Your research discusses the employment effects of exemptions for different situations and how they differ across genders, for example. Can you briefly discuss what you found about GDL exceptions and how they affect teenage labor force participation?

Pitts: Yes. Some states anticipated that there might be an impact on employment, so they allowed for exemptions for work, among a few other things, and usually these exemptions are just for curfew, not for passenger restrictions. Not all states allow them. Georgia, for example, where we live, does not, so you have to be home by midnight, no excuses. Among those that do have them, some require explicit documentation—such as a signed document from your employer stating that you work for them and are coming home from work at late hours—while others just require that you be able to say you're coming home from work. They don't require explicit documentation on hand. And the impact goes just as one might expect: in states where there are no work exemptions, there are larger negative impacts on labor force participation than in states with some work exemptions—with strict ones, that the negative impact on LFP is a little smaller. And the lowest negative impact of GDLs on LFP is in states with minor work exemptions, where you just have to get a note from your parents or just need to be able to say, "Yes, I'm coming home from work."

Heintjes: What did you find in terms of difference by gender? Your paper touches on that, and I wanted to get you to address that briefly.

Pitts: Sure. We did find a difference by gender: the labor force participation of females was more responsive to the exemptions than the males. The negative impact of GDLs on LFP was relatively much smaller for females than males, when work exemptions were allowed. Conversely, females were more negatively impacted by GDLs with no work exemption, like in Georgia. We did some further exploration and found that females—we tried to look early on, making sure the behavior reflected pre-GDL behavior—females are disproportionately more likely than males to work in industries that had late-night hours, like hospitality or retail, and males are disproportionately working during the daytime hours. So the benefit of the exemptions—or the burden of no exemptions—falls more heavily on the labor force participation of female teens.

Heintjes: I see. Melinda, my next question is somewhat outside the scope of your work, so please bear with me. As I was reading your paper I couldn't help but wonder if rideshares, electric scooters, and things like that are affecting how young people get to work. In the big cities you have transit, but that doesn't do much for you if you're not near transit, obviously.

Pitts: Well, we actually think that is part of our work. We want to understand where this is driving the labor force participation, where the impact falls, and we tried to capture these factors that are more readily available in metropolitan over nonmetropolitan areas, by estimating them separately. So rideshares—which, technically, teens are not allowed to enter into a contract, so they're technically not allowed to do those—but those are going to be more available in cities, in major cities and metropolitan areas. And while the magnitudes differ slightly, the results suggest that this is not just a rural issue or just a metro issue, but it's important across the whole state.

Heintjes: Sure. Well, Melinda, I always like to ask our economists this type of question when I talk about their work: what would you like to see policymakers take away from this work? Given the importance of labor force participation—and the importance of teenagers' work on future earnings—what would you ask a policymaker to consider in reading your work?

Pitts: Well, personally, I'm a bit reluctant to say we need to alter policies that are saving lives and preventing so much injury. Especially as a parent, I found GDLs to be great. However, we also need to think about the fact that an estimated 250,000 fewer teens were in the workforce in December of 2015 due to these policies, and that's not something we should ignore. Some things that we found the literature to suggest are that we're saving money, really, by not having these accidents in people's lives. We maybe need to put a little bit more money into things to enhance their work readiness—apprenticeship programs in the school is an example of a program that's been suggested—more school-based activities where they have the transportation from the bus. And from a monetary policy perspective, this is just important because we need to understand what's going to happen when we change. The GDLs will mute any response we'll see in teen labor force participation, and we just need to understand that and take that into account when evaluating the economy as a whole.

Heintjes: Well, that's great. And, everyone listening: I want to be sure to note that we'll have a link to Melinda's paper on our website, frbatlanta.org, and I really hope you'll check it out. It's a really interesting situation to think about, and it definitely took me back to driving to my job as a stock clerk when I was in high school. Melinda, thanks again for being back on the podcast today—it's always fun to have you on.

Pitts: Great—thanks, Tom. Glad to be here.

Heintjes: And that brings us to the end of another episode of the Economy Matters podcast. I hope you'll join us next month when I'll sit down with the Atlanta Fed's Nick Parker and Brent Meyer to talk about their Survey of Business Uncertainty, which they oversee here at the Bank—and I can say with no uncertainty that you'll want to be here for that conversation. Take care, and we'll get together again next month.