Probable Causation

View Original

Episode 86: Elizabeth Luh

Jennifer Doleac

Elizabeth Luh

Elizabeth Luh is a Postdoctoral Research Fellow at the University of Michigan's Criminal Justice Administrative Record System (CJARS).

Date: January 3, 2023

A transcript of this episode is available here.

See this content in the original post

Episode Details:

In this episode, we discuss Dr. Luh's work on the effects of financial sanctions:

“The Impact of Financial Sanctions: Regression Discontinuity Evidence from Driver Responsibility Fee Programs in Michigan and Texas” by Keith Finlay, Matthew Gross, Elizabeth Luh, and Michael Mueller-Smith.

OTHER RESEARCH WE DISCUSS IN THIS EPISODE:

TRANSCRIPT OF THIS EPISODE: 

Jennifer [00:00:08] Hello and welcome to Probable Causation show about law, economics and crime. I'm your host, Jennifer Doleac of Texas A&M University, where I'm an economics professor and the director of the Justice Tech lab. My guest this week is Elizabeth Luh. Elizabeth is a postdoctoral research fellow at the Criminal Justice Administrative Record System or CJARS at the University of Michigan and she's on the job market this year. Elizabeth, welcome to the show.

 

Elizabeth [00:00:31] Hi. Thank you for having me again.

 

Jennifer [00:00:34] Today, we're going to talk about your research on the effects of financial sanctions in the criminal justice system, but before we get into that, could you tell us about your research expertise and how you became interested in this topic?

 

Elizabeth [00:00:45] So I graduated with my Ph.D. in economics in 2020, a long time ago, it feels like, from the University of Houston. I became interested in this topic, I think, because ProPublica one night released an article about parking tickets in Chicago and how they were leading to like insane bankruptcy rates and so they had the data available online. So I downloaded it and I started poking around on it and that's kind of how I started becoming interested in like the impacts of financial sanctions from like, you know, criminal infractions or criminal offenses on outcomes.

 

Jennifer [00:01:17] So your paper is titled "The Impact of Financial Sanctions Regression Discontinuity Evidence from Driver Responsibility Fee Programs in Michigan and Texas." It's coauthored with Keith Finlay, Matthew Gross, and Mike Mueller-Smith. So what are driver responsibility programs?

 

Elizabeth [00:01:32] So driver responsibility programs have been around for quite a while. They were pioneered by New Jersey in 1983, and they called their version of their driver responsibility program the merit rating plan surcharges and like I think five other states have at some point had a form of the driver responsibility program, so New York, Virginia, Texas and Michigan so far have all had driver responsibility programs, and they're pretty much all very similar. They have two goals to reduce traffic fatality and increase government revenue and the programs always try to achieve this in two ways. One is a point system for traffic infractions. So things like speeding tickets, not stopping at stop signs. So if you see the number of fractions, every additional fraction comes with not just the traffic fine, but also a $25 or $50 fee.

 

Elizabeth [00:02:18] So the second part of the driver responsibility program are just automatic surcharges that are assigned upon conviction for more severe criminal traffic offenses. So think driving without a license, driving without insurance, driving under the influence, or driving with a suspended license. These surcharges tend to be much higher. So in Texas, they range from $300 to $6000, in Michigan, it was $300 to $2000. So the fees from driver responsibility programs aren't classified as like criminal fines. They're classified as administrative fines. So the only punishment available for not paying these surcharges is license suspension, which as you kind of notice when I talked about the second part of the driver responsibility fee program, driving with a suspended license itself is a driver responsibility fee triggering offense. So that kind of tends to be the biggest criticism about this program, which is that once you get one driver responsibility fee, you're at higher risk of getting future driver responsibility fees from not being able to pay them potentially.

 

Jennifer [00:03:12] Right. So you get your license suspended because you didn't pay them.

 

Elizabeth [00:03:16] Yeah.

 

Jennifer [00:03:16] And then you have extended license problem and then it just compounds. Yeah. And so these are big fees. This is like much more than your standard driving your speeding ticket with like, you know, a hundred bucks or something.

 

Elizabeth [00:03:28] Right.

 

Jennifer [00:03:28] It's $6,000 is steep. So what had we previously known about the effects of financial sanctions like this?

 

Elizabeth [00:03:36] So a lot of financial sanctions research has been, I think, in a lot of non-economics literature, which I think is great. This is definitely a interdisciplinary topic. So I think the most I think the name that comes up the most when you talk about criminal financial sanctions is Alexes Harris. So she's a professor at University of Washington in sociology, and she has a great book about drawing blood from stone, about the impact of financial sanctions, where she goes and interviews inmates and people who have had contact with the justice system on how these financial sanctions have impacted their lives. And she, along with other coauthors in a different paper, have linked these financial sanctions to financial instability, poor labor market outcomes and future recidivism behavior. In terms of economics research, I think, you know, there hasn't been that much compared to like other topics in economics, but in recent years, I think there's been quite a few papers that have come out.

 

Elizabeth [00:04:26] So Steve Mello has a paper that looks at how traffic tickets impact financial health or financial stability, and he finds that traffic tickets lead to these fines that you get kind of lead to lower access to credit and increased financial distress. Myself and Ryan Kessler, we each have papers on parking tickets in Chicago, and luckily for us, we find similar results, which is that parking tickets lead to higher rates of bankruptcy and for him, he also links it to credit report data and he also finds lower access to credit.

 

Elizabeth [00:04:59] So I think what also is very famous when you think about financial sanctions is what triggered all these studies and I think renewed interest in these is what's called the Ferguson report, which was released by the Department of Justice, which looked at kind of in the aftermath of like the Ferguson situation, what happened, why this happened and they found that the city of Ferguson tended to use financial sanctions, criminal financial sanctions as a way of taxing citizens in the community. And you know that the community tend to be lower income and individuals of color. So, you know, generally these financial sanctions are kind of associated with disproportionate harm on lower income individuals and individuals of color.

 

Elizabeth [00:05:41] There's been other past research, like Hansen has a paper about increased DUI sanctions leading to deterrent effects. And Dusek and Traxler also find similar effects when they looked at increased speeding fines. There's a lot of research actually. Now they say this and a lot of like qualitative work. So I think by thinktanks like the Hamilton Project and the Fines and Fees Center and the Brennan Center have all found that these fines and fees are associated with a lot of negative outcomes. But recent work there's been a RCT by Pager and coauthors that looks at debt forgiveness in Oklahoma County. So they randomly forgave debt of about $2,000 and they looked at impacts on recidivism 1 to 2 years after, and they also found no impacts on recidivism. So that's kind of one of the outliers and new research.

 

Jennifer [00:06:30] Yeah. And so I think, you know, most people going into this probably do have those like the Ben Hansen paper and the others that you mentioned in mind where a lot of people have really focused on the effect of these kinds of fees on like not these specific fees, but like being just over a blood alcohol content threshold where you get a much higher fee.

 

Elizabeth [00:06:51] Right.

 

Jennifer [00:06:51] And then Ben and and these others find in their various papers that it has a really big deterrent effect, a specific deterrent effect and so that is one possible way,.

 

Elizabeth [00:07:00] Yeah.

 

Jennifer [00:07:00] That that could be affecting behavior, but of course we have all these concerns about them and so that's been driving, I think, this interest in and seeing what else can we find out. And yeah, I mean, while you listed a bunch of papers, I think I definitely think of this literature as being really thin. We just don't know very much. So what makes this so difficult to study? So, you know, there's been all this interest, especially outside of economics, for a long time and the potential harms of financial sanctions, but it's been really difficult to quantify causal effects here. So what are the main hurdles to figuring out whether and how much these financial sanctions matter?

 

Elizabeth [00:07:36] So I think the biggest hurdle, like just from identification perspective, is fines aren't assigned randomly in the U.S. especially, they're usually tied to a conviction. I mean, there's like some debate about cash bail and like whether or not, you know, those are assigned upon like those aren't assigned upon conviction. Right. There are pretrial kind of things. When I'm thinking about financial sanctions, like most of them are assigned upon conviction. So it's hard to disentangle like any sort of negative impacts you find later down the road if they're driven by the negative impacts from the fines like having to pay these increased fines are they're driven by the fact that you were also convicted for an offense. And so I think, you know, the lack of exogenous assignment of fines generally is kind of what makes this topic a little harder to study.

 

Elizabeth [00:08:18] On the other hand, data, I think, is a big issue on studying financial sanctions. So there's a lot of data on like convictions so that's easy to study, but there's not a lot of great data and systematic data about fines. I think this is kind of because for me, it just drills to like two big things. One is that financial sanctions is a huge category of like monetary sanctions that can be assigned, right. You have like restitution, you have fees, you have surcharges, you have fines and they all vary across state, right. So certain states allow judges to waive them certain fines and fees judges can waive them within state but can't across other states or across other offenses. The states also vary by how they collect this data, too. So, you know, CJARS I'm so lucky to work at CJARS and we get all this great access to administrative court data. But then, like some court systems will give you like each of the fines for each of the financial sanctions laid out exactly and how you paid them when they paid them.

 

Elizabeth [00:09:15] And some will just give you like the total amount paid so you don't even know what was assigned upon conviction at all and judges can waive them, too. So even if you think that they were assigned, it could be that, you know, judges utilized their discretion and they weren't assigned in the end. And there's also not any rules that require them to collect this data. Right. So I know a lot of people hate on UCR for, you know, as the way of measuring criminal justice data in any form or criminal activity in any form. But there's nothing similar like that for fines, right. The U.S. doesn't require any collection of that data, so sometimes it just doesn't go collected.

 

Jennifer [00:09:49] Yeah. And I think this also part of the reason here is that the various courts or agencies that are collecting these fines and fees, you know, if they're collecting data, it's for their own administration it's not for like a research. 

 

Elizabeth [00:10:05] Right.

 

Jennifer [00:10:05] The goal is not to document all that stuff for potential researchers later its to keep track of who owes what and whether it's been paid. I have vague memories of talking with some court a while back trying to get information on driver's license suspensions and fees, and they were like, Oh, no, as soon as someone pays the fee, we're just like deleted from our system to save space.

 

Elizabeth [00:10:25] Oh my God.

 

Jennifer [00:10:25]  And it was like it was just heartbreaking. It was like,.

 

Elizabeth [00:10:28] Yeah.

 

Jennifer [00:10:28] So the data is just gone. Like you only know about outstanding fees, which is of course what they need.

 

Elizabeth [00:10:32] Right.

 

Jennifer [00:10:33] It means that studying the impacts in that state would be impossible.

 

Elizabeth [00:10:36] Right.

 

Jennifer [00:10:37] So the data you will have are very cool and we will talk more about them in a bit, but let's first talk about the policy changes you use. So you have two policy changes in two states. You've got Michigan and Texas and you use them as natural experiments to address this question, giving you some traction on the identification issue you mentioned. So tell us about these policies. When do they go into effect and what do they do?

 

Elizabeth [00:10:59] Okay. So these policies actually coincidentally had very similar timing. So Texas passed its version in June 2003, and it went into effect on September 1st, 2003, in Michigan, literally right after they passed their own version, August 2003 and it went into effect October 1st, 2003. So very similar timing. So both these policies were rather similar. I'm going to talk about Texas first. So in Texas, the program essentially didn't deviate much from New Jersey. Again, had the two systems, the points system with like lower traffic infractions and then automatic surcharges for more severe offenses.

 

Elizabeth [00:11:35] So unlike New Jersey, I think the surcharges were a little bit more severe and also compared to Michigan, I think compared to pretty much any of the other states. So in Texas, the surcharges would last over three years. So you would pay like let's say you were convicted of driving under the influence for the first time. You would pay $1,000 for the next three years, so $3,000 total. So these fines range from $300 to $6000. So rather high. I think if you think about like on average, what a traffic fine looks like. In Michigan, they were a lot lower because they were only assessed over two years. So they ranged from $300 to $2000.

 

Jennifer [00:12:12] And then if you didn't pay.

 

Elizabeth [00:12:14] Oh, right. Yeah. So if you didn't pay, your license would be suspended for both those states. So very similar in that sense. And for both of them, driving on a suspended license was also itself a driver responsibility fee triggering offense. So again, like the problems I stated earlier, they were present in both of these programs.

 

Jennifer [00:12:30] Right. And then remind us what the stated goals of these policies were.

 

Elizabeth [00:12:34] Yeah. So the goals of both of these policies were similar, but a little bit different. I think that's kind of important. So in Texas, you know, they had a very high drunk driving rates in Texas. So their focus of the driver responsibility fee was to reduce drunk driving, which is why you also see kind of why I think you also see like the higher fines. So that $6,000 fine is for people who are repeated drunk driving offenses. So which is missing from Michigan's program. So Michigan's program is only for essentially any DUI in Texas. There was like an extra punishment if it was like a repeat, a DUI.

 

Elizabeth [00:13:05] And so in Texas, they had the high drunk driving problem, but their emergency departments and their trauma system was really underfunded and overstretched. I think they estimate that, like back at that time before the driver responsibility fee program, that there were 16 emergency departments and 8 emergency rooms per 1 million people. So that's very like underfunded. So the whole idea of the driver responsibility program in Texas was that that money was going to go towards funding these like trauma systems where, you know, people who were victims of drunk driving or people who were drunk driving would go to if they caused a car accident. Right. So kind of help fund that whole thing and also reduce drunk driving.

 

Elizabeth [00:13:40] In Michigan, they were actually just facing a budget shortfall overall. So the money from the driver responsibility for you would actually go towards this general fund, which is, you know, money could go anywhere, essentially and they had a very high highway fatality rate overall, not just drunk driving. So they passed this bill to kind of tackle their highway safety and to also tackle their budget shortfall.

 

Jennifer [00:13:58] Okay, great. And then so what are the various ways that these policies might have affected behavior?

 

Elizabeth [00:14:06] Yeah. So there are a few ways that people could respond to this. Fine. So as you kind of touched on earlier, there could be a specific deterring response. So people see these higher fines and they become safer drivers in response. So you would see like people just the reduced caseload essentially like people would be committing less severe traffic offenses or less driving infractions. On the other hand, you know, as I also touched on earlier, both these bills were passed in essentially times of like budgetary distress. Right. So police officers could respond by ticketing people more and then other law enforcement agents could also respond by, you know, increasing convictions or pushing more cases to trial.

 

Elizabeth [00:14:42] So different responses that could also lead to the opposite effect of the deterrence effect, which would be to increase the caseload. There could also be, you know, different responses depending on the characteristics of the individual like some characteristics might just be more appear more sympathetic to judges, and so they're less likely to convict or they're more likely to give them a warning instead. Or, you know, these higher fines could make it so that hiring attorney might become more attractive as a concept. So maybe people with higher income might be more likely to hire a lawyer and then escape conviction in that way.

 

Jennifer [00:15:14] Yeah. And then at some point also, as we were talking earlier, you might have broader impacts on behavior where you are if you have this new $3,000 fee to pay off, maybe you work more.

 

Elizabeth [00:15:27] Yeah. So getting to as these fines, you know, having these fines might change your behavior. Like let's say it doesn't change your behavior before getting the fine before this like all the stuff I talked about before was like upon conviction, essentially, but post-conviction, these fines could also matter. So maybe these fines now induce you to work more. Right. You have these higher fines. You want to pay them off, so you increase your labor supply response.

 

Elizabeth [00:15:50] On the other hand, though, it could be the case that, you know, you don't get to keep your license, you can't pay within the first year. Your license is suspended. So then now you actually can't work. You don't have a legal means of going to work. Your job is contingent on you having a license. You lose your job. So you could actually see labor go down. You could also see like a criminal response to, you know, you don't have a car now potentially, or maybe you don't have a car now potentially so you can't commit crime, right. Maybe having a car or having a license was contingent on you committing crime, for example. Or it could also be one way to increase crime. You know, maybe it's the case that you have your car, your license suspended, you continue driving anyways, you're more likely to pick up future driving convictions. Or maybe, you know, you can't work anymore, so you turn to crime to generate income. So these are all very possible responses that we might see because of these fines.

 

Jennifer [00:16:41] And so how do you use the policy changes that you described before to measure the causal effects of financial sanctions on these various types of behavior?

 

Elizabeth [00:16:51] Yeah. So what's great about this policy change in the way the law was written is that there's a very clearly defined treatment definition. So it all revolves around like when you're convicted. So essentially the way the laws are written was that the fines would only kick in if you were convicted after the effective date. So in Texas case, let's say you were convicted on August 30th, then you would avoid these driver responsibility fees because the effective date is September 1st. So, on the other hand, if you were convicted on September 2nd, now you have to pay this $300 to $6000 fine that you didn't have to pay if you were convicted like two days earlier.

 

Jennifer [00:17:28] Great. All right. Well, let's talk about the amazing CJARS data. What data do you have for all of this?

 

Elizabeth [00:17:34] Yeah. So as you touched on, we used data from CJARS. So I think I don't know if you define it yet, but it's the Criminal Justice Administrative Records System. So this is essentially longitudinal individual data on criminal behavior. So we can essentially see when people are arrested, we can follow people through different stages of the criminal justice system.

 

Elizabeth [00:17:52] So like arrest, conviction, incarceration, probation and parole, it's longitudinal, too, which means that we can create like rich criminal histories and future recidivism behavior based on a focal event. We link this to data from the U.S. Census so what's housed on the FSRDC, and I can't remember that I Federal Statistical Research Data Center. So this gives us access to tax filings like W-2 and 1040, along with very rich and accurate demographic information so from the Social Security Administration numerate file and what's called the best race files. So what's nice about these is that in the data from the law enforcement agencies oftentimes doesn't record like demographic information correctly, like notably like Hispanic heritage. Right. So Hispanic is an ethnicity and they oftentimes just record race. So what's nice about this is we can get self-reported race and ethnicity identification along with accurate date of birth information. So all this stuff linked together is the data that we use.

 

Jennifer [00:18:51] And we should plug a little bit more CJARS is as you mentioned is in the RDCs. So if you a listener are a researcher, you can access it too and there's a lot of information on the CJARS website about how to do all of that and what the data look like and what states are included in all this. It is totally amazing and super fun to see projects like this come out of it after so much work has gone in over the.

 

Elizabeth [00:19:14] Yeah.

 

Jennifer [00:19:14] Years to making these data exist. Okay. And so what outcomes are you most interested in in this paper?

 

Elizabeth [00:19:20] So what we are most interested in is recidivism and labor market outcomes in the 1 to 10 years following the focal event. In this case, it's like the first conviction for a DRF related offense.

 

Elizabeth [00:19:32] So and I think the reasons we're interested in these is because of the mechanisms I kind of highlighted earlier, right? The labor response that you might see the increased or decreased response or the increased recidivism or decreased recidivism response.

 

Jennifer [00:19:46] But then you're also able to link to romantic partners.

 

Elizabeth [00:19:47] Right. Yeah.

 

Jennifer [00:19:50] Which is amazing.

 

Elizabeth [00:19:50] Yeah. So we also look at partner's response to so, you know, these individuals, the ones with like these severe criminal traffic offenses like they tend to have lower labor market attachment. Right. So it could be the case that we don't see a labor market response because they were never going to respond on the labor margin anyways because they don't work or.

 

Jennifer [00:20:10] They don't they don't have a job. It's hard for them to find a job for some reason.

 

Elizabeth [00:20:13] Yeah, and we're looking at like W-2 and 1040, so that tends to limit us to formal employment. So if there was an informal market change, we wouldn't be able to pick that up. So it could be the case that their partners, the romantic partners, are the ones who are going to change their labor market response. And so we also link these individuals to their partners and we look at their recidivism and labor market outcomes also, and then whether or not they are more likely to commit commit a DRF related offense also, because, you know, potentially they're driving more because their partners, the individual who got that driver responsibility fee, aren't able to drive anymore because their license was suspended, so.

 

Jennifer [00:20:47] Okay, great. And that data and so in your linking based on who's filed taxes together, is that right?

 

Elizabeth [00:20:54] So this is another CJARS plug. But there's a I think Brittany Street has done some great work linking partners to are linking family members, it's called a relational crosswalk. So anyone that they've lived with in a household together over time. So what she uses in that is any sort of survey data along with 1040 tax filings. So and then also like paternity and maternity like filings. So like if we observe two individuals have a child together, then we say we don't know like exactly what their romantic relations are, but we know maybe at some point there was romantic relation.

 

Jennifer [00:21:26] Right. Okay. Amazing. So before you get into the main analyzes that you're you're really targeting here you first consider whether the case loads in case details look smooth through the implementation dates of these policies. So you need that for your regression discontinuity design because the only thing you want to be changing is whether someone is as assigned one of these fees you don't want the cases to look different on either side. So this yielded some surprises or at least one surprise. So what do you find when you do this?

 

Elizabeth [00:22:02] Yeah, so as you can tell, really there's a few surprises. So let me talk about the unsurprising one, the boring one first. So in Michigan, you know, we look at the case loads to make sure that we don't see any sort of specific deterrent response or that law enforcement agents, you know, change their conviction behavior because they want to raise money. So in this case, you know, we see that it's smooth across the implementation date, which means that we don't see that specific deterrent effect. So no drop in caseload or we also don't see a response from law enforcement agencies with an increase in caseload from increased convictions. So that's great.

 

Elizabeth [00:22:35] In Texas on the other hand, we do see something interesting, which I think kind of intuitively makes sense. So for Texas, we see something very different, which is that we see people an increased caseload just to the left of the implementation date and a drop to the right. So this means that this is kind of evidence of manipulation, which you don't want to see, right. We want to see essentially that the treatment or in this case, the when the DRF goes into effect is as good as random. And this manipulation would kind of indicate that it's not as good as random. We dig in further and we look at like the sort of demographic characteristics of who's participating in this manipulation. We see that there's an increase in the likelihood of being right to the left of the cut off and an increase in pre conviction income measured using 1040 tax filings. So seems to be the case that people who are white and who have higher income are the ones who are manipulating around the cutoff or around the implementation date.

 

Jennifer [00:23:29] So just in other words, here. So what they're what seems to be happening is that people know that this new law is coming and they know that they were charged with or they got this type of ticket or were charged with one of these eligible offenses. And so they find a way to race to finalize their case and get a disposition and get everything paid before the implementation date so they don't have to pay the big extra penalty that they would have gotten if they had waited one more week. So they plead guilty early or something like that. Is that right?

 

Elizabeth [00:23:58] Yeah, exactly. And we can see this also, if we just look at if we just look at the number of days on average from offense to disposition, you see a drop in the number of days like on average. So on average it's like 240 days between the time you offend to the time you're convicted, but just to the left of the cutoff, the number of days, just by about 40 days is 200 days, which seems to be that people are pushing up their conviction dates or maybe pleading to have an earlier trial or to get an earlier conviction so that they can avoid these increased DRFs, these increased fines.

 

Jennifer [00:24:29] Yeah. And I mean, to be clear, this is totally rational, right? Like in some.

 

Elizabeth [00:24:34] Yeah.

 

Jennifer [00:24:34] Ways economists, we would absolutely expect that people might behave in this way. So you see that and that is interesting. And it also complicates your analysis in Texas. So what do you all do? What do you do to to kind of address this in Texas?

 

Elizabeth [00:24:48] So really, the main avoidance is through expediting your processing time. So that's like a time limited response, right? If you offend after the effective date, there's nothing you can do to push your case date to before September 1st. Like, there's nothing you could do. Like time has gone on. So this time little response mindset. What's nice is we can use a donut because it's not something that's going to persist over the long run is just around the cutoff, essentially. So we're just going to do a donut strategy, which is, you know, if you imagine a donut, we're just going to punch a hole right around our implementation date. So we're going to exclude the 60 days just around the implementation date like a donut. That's what we do.

 

Jennifer [00:25:26] Yeah. You're going to drop all the cases for people who were charged?

 

Elizabeth [00:25:31] Convicted.

 

Jennifer [00:25:31] Convicted 30 days before or 30 days after, because those are the people that you're worried there's something weird about those people.

 

Elizabeth [00:25:38] Yeah.

 

Jennifer [00:25:39] The convictions are not when they're supposed to be. And so just drop them and then use all the other data and in RD form the way you planned.

 

Elizabeth [00:25:48] Yep.

 

Jennifer [00:25:48] All right. So what was your first stage effect of these policies on fees levied? What actually were these policies implemented as planned and what was the impact on the fees?

 

Elizabeth [00:25:59] Yeah, so these policies, like I said, since they weren't criminal fines, they weren't subject to like judge waivers or, you know, other sort of discretionary actions that judges or courts can apply upon, like with these financial sanctions only. So these what we see in our first stage is very strong. So essentially, like once the fine goes into effect, you essentially see if you're convicted of a DRF related offense, so not just DRF eligible, your likelihood of getting a DRF rises by 95% Michigan and 74% in Texas. And then if you just look on like the amount of monetary burden they are subject to now on average, this is like an increase of 1400 dollars that they're subject to that they weren't subject to before In Michigan and then Texas, it's like 2500 dollars. So very huge first stage, very huge change in the amount of fines that they have.

 

Jennifer [00:26:50] And then moving on to the second stage, the main results. What was the effect of financial sanctions on labor market outcomes?

 

Elizabeth [00:26:58] So what we find is actually, I think, very surprising because we went into this project thinking that, you know, a lot of the past, you know, after reading all the past work, that there were going to be very huge negative responses to these finds and what we find is actually nothing. In Michigan, we see an actual increase in labor response that's insignificant and very small. So like very null effect, like an increase in income of like $400 per year, but not significant. And going in the opposite direction of what past research would tell you, which is that people that there's a negative labor response to it traps people in a way that further that it traps them in a poverty cycle that they can't get out of.

 

Jennifer [00:27:38] Right. Because they lose, they lose their license and.

 

Elizabeth [00:27:40] Exactly.

 

Jennifer [00:27:40] They can't work anymore. This has been the sort of conventional wisdom story out there. Yeah,.

 

Elizabeth [00:27:44] Yeah. And then in Texas, we see a similar story, even though a very different state, the policy is similar, but like very different amounts. And again, we see very null precise impacts on earnings like $202 over 11 years. So it's like $20 over 11 years and it's insignificant. We also, you know, these maybe looking over ten years wasn't the right thing. We also kind of check over every additional year after the fine goes into effect. So, you know, we also look in the short run to maybe their short run impacts that attenuate or kind of degrade as time goes on and we also see in the short run that there's no significant impact on labor market outcomes.

 

Jennifer [00:28:22] Okay. So it's not affecting whether or how much people are working, at least in the formal labor market.

 

Elizabeth [00:28:27] Yeah.

 

Jennifer [00:28:28] We had that caveat earlier but that agreed it's surprising. So then what was the effect on recidivism.

 

Elizabeth [00:28:34] So yeah, with recidivism, again, we don't see any effects on total recidivism. So just any type of, you know, future criminal convictions for any type of offense. Very even more null effects over the ten years that we follow these individuals after their initial DRF conviction.

 

Elizabeth [00:28:52] The only response that we find in terms of like recidivism response is actually just driven by the policy itself. In Michigan, we find an increase in likelihood of getting a future conviction for driving with a suspended license, which as I mentioned earlier, is kind of just generated by the policy, because if you aren't able to pay your fines, you get your license suspended. And if you continue driving on it, then you know you'll get another license, conviction or driving on a suspended license conviction. So that's kind of what we attribute that increase to just generated by the policy itself.

 

Jennifer [00:29:24] Mm hmm. Okay. So also surprising, especially in the context of those earlier papers that found, like, you know, high fines for DUI is, for instance, if you're just over the threshold, you'll recidivate less. We don't seem to be seeing that action here. And so the third outcome you look at is what happens, the romantic partners. So maybe nothing is happening with this main target person because their romantic partner is bearing the full burden of this extra money they have to pay off. Is that what's happening?

 

Elizabeth [00:29:53] No. Yeah. So the romantic partners also don't respond either to the driver responsibility fees. So we looked at their total earnings response. We don't see any change there. And again, these aren't just like insignificant changes. These are like very small relative to the mean, like a change in like $500 over 11 years on annual income and again, it's like it's insignificant and then like relative to the mean, it's like pretty much nothing. Like on average their income is like $30,000. So $500 increase on like $30,000 isn't going to change your income substantially. And again, we find similar results in Texas, too, where we don't see any substantial change on labor market.

 

Elizabeth [00:30:33] And then if we look at recidivism, too, we don't see any sort of recidivism response either. These individuals aren't committing more crime because their partner has a driver responsibility fee. We also look at, you know, maybe because they're driving more because their partner's licenses are suspended from the driver responsibility fee. Maybe, you know, they are driving more and putting themselves at higher risk of getting their own driver responsibility fee and we also don't see any significant change there.

 

Jennifer [00:30:57] Okay. So these huge fees certainly are, you know, in some way a burden on people. They are expensive. These are big fees. These aren't tiny little fees, but no impact on any of these outcomes that we might expect them to effect. So what are the policy implications here? What should policymakers and practitioners take away from all these results?

 

Elizabeth [00:31:22] So I think one thing is that the it's really easy to motivate wh I think from like a voting perspective, why we should use these financial sanctions, right because it's like you're taxing a subset of the population that might not have the ability might not be able to vote. Right. So I think it's when you think about why these are so popular, I think this is kind of a big reason why.

 

Jennifer [00:31:44] And it's an appealing it's an appealing target because like, they've committed a crime, right?

 

Elizabeth [00:31:48] Yeah.

 

Jennifer [00:31:48] Yeah.

 

Elizabeth [00:31:49] Right. And it's like these fines don't generate any income, like they're not generating the revenue that you need from a local government perspective, and they're not creating the positive change that you were hoping to see. Also with these policies, like there's no specific deterrent effect so people aren't committing less crimes. On average, these policies in Michigan and Texas, like in the earlier so which is essentially the time that we're studying in our paper right. We're studying within like the first one and a half years surrounding the cut off or surrounding the implementation date. The payment rate of these was like 30%. So it's like people weren't paying these. You were suspending licenses for a huge portion of the population, which is potentially creating more crime, which we didn't find, but, you know, which, you know, suspending licenses, though, is I think maybe we didn't see we didn't see a response here, but it could have had responses that were not able to measure. So I think one major takeaway from this is that we shouldn't keep adding financial sanctions upon conviction for people, people don't pay them and they don't generate the deterrent response or the positive response that you're hoping to see. Aand then in terms of raising local revenue, they're a massive failure on that.

 

Jennifer [00:32:56] Yeah. So, I mean, it's interesting because the push against these types of fees has been really driven by this idea that the fees create tremendous hardship. Right. That.

 

Elizabeth [00:33:07] Yeah.

 

Jennifer [00:33:07] It is trapping people in the cycle of poverty, making it more difficult to work. People are losing their jobs, their homes. They're forced to commit more crime. And you're not seeing any of that.

 

Elizabeth [00:33:17] Yeah.

 

Jennifer [00:33:18] So then it's like, well, then maybe these fees aren't so bad, but when you you're right when you go back to kind of what the stated goals of these policies were, it was to raise revenue and to deter really bad behavior like repeat DUIs. And you're not finding it's not deterring that that bad behavior and it obviously raised some money, some people paid the fees, but it wasn't a very efficient way to raise money. And it's certainly a it's what economists would call a really regressive tax. We're taxing generally very poor people. So not a great way to fund government in general, but when payment rates are 30%, you know that it's not a great moneymaker. And I think you mentioned where I've seen you present this, that both Michigan and Texas have repealed these fines, right?

 

Elizabeth [00:34:01] Yeah, because Michigan repealed their version in 2018. Texas repealed their version in 2019. And I think by the time they repealed it, like there was about $3 billion in debt. So 600 million from Michigan and 2.5 billion from Texas. So huge amount of debt. And then they like immediately reenacted a lot of licenses back on because of like essentially they were only suspended because they weren't able to pay their DRFs.

 

Jennifer [00:34:27] Mm hmm. Yeah. Amazing. And so and so they obviously had not seen your paper yet because it wasn't written yet.

 

Elizabeth [00:34:33] Yeah.

 

Jennifer [00:34:33] Why did they repeal the policies?

 

Elizabeth [00:34:36] They were very unpopular, I think.

 

Jennifer [00:34:38] Okay.

 

Elizabeth [00:34:39] What's called nonpartisan. Like everyone decided they were awful because they weren't raising any money like they tried like Texas tried two different types of, like, modifications to the program. So they try an indigency program, which could go on a payment plan and then get your license reinstated again while you're on this payment plan. It never really took off. Michigan tried a similar thing where you could essentially tried their own version of payment plan. It was a little bit different, really never took off, did not increase revenue in any way. People just like weren't responding to these. And I think it was also very obvious from a public policy perspective that the main goal, which was in both states like highway safety in some form or the other, was not happening either. In fact, like I think one of the biggest detriments from Michigan was like in 2008, they already saw that like highway fatalities were actually higher compared to when they first started the policy, which isn't causal evidence, but for policymakers, you know, it was as good as causal evidence.

 

Jennifer [00:35:31] Right it's not a good sign.

 

Elizabeth [00:35:32] Yeah. Yeah. That this policy wasn't very good.

 

Jennifer [00:35:35] Yeah. And then if they're not bringing in much money, then, okay, let's just scrap this whole thing.

 

Elizabeth [00:35:39] Yeah, yeah.

 

Jennifer [00:35:40] Yeah. So have any other papers related to this topic come out since you first started working on the study?

 

Elizabeth [00:35:46] Yeah. So I talked about the Pager et al paper, which looks at debt forgiveness in Oklahoma County. So what's nice about this paper is it didn't come out when we were first writing this paper, so it felt like our paper was the only outlier that was finding null effects, but when this paper came out, it was such good affirmation that, like we weren't doing anything crazy or bad. But they also find null effects of debt forgiveness on future recidivism outcomes in the 1 to 2 years following forgiveness. I'm working on a future paper with Mike and a new coauthor, Carl Lieberman, at Census, where we look at a whole set of fines. They're not just driver responsibility fees, we're just looking at fines, assigned upon conviction for misdemeanors and felonies. And we so I can talk about this because our results were just disclosed, but we also find no impacts on recidivism, labor market outcomes, and we also link the fines to ACS outcomes. And we don't see any change on like household expenditures or on self-reported total income. So it.

 

Jennifer [00:36:41] Oh wow.

 

Elizabeth [00:36:41] Seems to be the case that these fines don't aren't really changing anything. Tyler Giles, who's out Wellesley College, looks at driver's licenses spend. And similar to our results, he finds that, you know, driver suspending driver's licenses doesn't change recidivism behavior. And I think he's working on getting credit reporting data to look at the financial side of his outcomes. So, yeah, a lot of great work that's come out and will come out.

 

Jennifer [00:37:07] Yeah, and I'll say a little bit more about the Pager et al study. So this was the late Devah Pager had started this project before she passed away and I think and then a bunch of colleagues kind of, you know, kept going and put this paper out this year, earlier this year.

 

Elizabeth [00:37:22] Yeah.

 

Jennifer [00:37:23] And that she'd run a randomized controlled trial where she was forgiving debt. And that's interesting partly because we love RCTs, but partly also because I think it's nice to have, you know, different types of policies here because, you know, debt forgiveness is different from not giving you the debt in the first place. Right.

 

Elizabeth [00:37:44] Yeah.

 

Jennifer [00:37:44] And so it could be that forgiving the debt has no impact because the damage has already been done. Right. You've already had to go some period of time without your driver's license and you've already lost your job. And so, okay, great, someone comes along and forgives your debt, but it's too late. And that's not a potential explanation for your results. There's just like giving you giving you this huge additional fine or fee just doesn't seem to have any impact.

 

Elizabeth [00:38:11] Right.

 

Jennifer [00:38:12] At all. Yeah. All right. So lots of ongoing work in the space, which is very cool. Null effect papers are always a little tricky. It's no one likes null effect papers, but, but this is an area where the null effect is super surprising, I think, to a lot of people. So it has been very interesting to see these papers come out.

 

Elizabeth [00:38:30] Yeah.

 

Jennifer [00:38:31] So what's the research frontier? What are the next big questions in this area that you and others are going to be thinking about going forward?

 

Elizabeth [00:38:37] Well, with my new paper that I'm working on about, you know, looking at a broader set of finds, it's like for me, I was like actually really surprised that we didn't see any response, especially looking at those ACS outcomes like at least you think like expenditures maybe would change like some sort of consumption indicator. I think this also is a good indication that we should think more broadly about like what I think the role of conviction on outcomes, because I think what's tied with all these fines and fees is like a conviction. And maybe it's like these fines don't matter because it's conviction that's like truly causing the harm on outcomes. And I think that's kind of like the next big thing to tease out. I think there's also been a lot of work going forward, like in Harris County, for example, where they're actually lowering the amount of fines that are being assigned. And I think that'll be interesting to see how outcomes change and how like local revenue responds to and maybe how voting changes with that whole change in mind assignment there. So I'm interested to see how that goes on. I think what's also there's a lot that I want to do on this. It's hard to size it down.

 

Jennifer [00:39:41] Yeah.

 

Elizabeth [00:39:42] It's just like so crazy to me that like, people keep assigning these fines and fees and there's so much of them that are assigned like, it's like I think the Brennan Center estimates that once you exit incarceration, you, like, have to have to pay about $13,000, which is insane like if the whole point of the justice system is to be like, rehabilitative, like it's hard like for anyone, I think, to do anything well, with $13,000 in debt without any jobs or work employment history or any credentials. So I just I think that raises that question.

 

Jennifer [00:40:11] Yeah, we've gotten ourselves into a situation where this is the way that we fund these local government agencies.

 

Elizabeth [00:40:17] Yeah.

 

Jennifer [00:40:17] Courts in the criminal justice system are really heavily reliant on these these fines and fees and that puts all kinds of really terrible incentives in place as well, even beyond just pushing them to levy these fines in the first place, but as you mentioned earlier, like one thing you might have seen here or when one could see if you had the data is that police might write more tickets because they need to raise more money. And this is that person who thinks a lot about this, whose work I'll plug here is Mike Makowsky and he has some great policy briefs and he wrote something for the Hamilton Project a few years back. We can link to that in the show notes that basically just makes this point like it's just the incentives are all misaligned here.

 

Elizabeth [00:40:56] Yeah.

 

Jennifer [00:40:56] And, and part of it is that these, you know, the police departments and courts all keep a big share of the revenue they bring in. And so they just have an incentive to keep doing it. And so, you know, we're economists, we know, we know, we know our incentives and it's something something we could be fixing if we wanted to.

 

Elizabeth [00:41:15] Right.

 

Jennifer [00:41:16] But yeah, lots lots more work to do in this space. My guest today has been Elizabeth Lu from CJARS at the University of Michigan. I will repeat that Elizabeth is wrapping up her postdoc this year and so is on the econ job market. Elizabeth, thank you so much for talking with me.

 

Elizabeth [00:41:30] Thank you for having me again.

 

Elizabeth [00:41:37] You can find links to all the research we discussed today on our website probablecausation.com You can also subscribe to the show there or wherever you get your podcasts to make sure you don't miss a single episode. Big thanks to Emergent Ventures for supporting the show and thanks also to our Patreon subscribers and other contributors. Probable Causation is produced by Doleac Initiatives a 501(c)3 nonprofit, so all contributions are tax deductible. If you enjoy the podcast, please consider supporting us via Patreon or with a one time donation on our website. Please also consider leaving us a rating and review on Apple Podcasts. This helps others find the show, which we very much appreciate. Our sound engineer is Jon Keur with production assistance from Nefertari Elshiekh. Our music is by Werner and our logo was designed by Carrie Throckmorton. Thanks for listening and I'll talk to you in two weeks.