Tornado watches, warnings, and sirens don't all mean the same thing — and if you live in Oklahoma, you know you don't even run to a shelter every time a siren goes off. You calibrate your response to the actual level of threat. The Supreme Court says due process works the same way.
In this episode, Gwen and Marc break down Mathews v. Eldridge — the due process balancing test that has governed how every federal agency designs its procedures for the past fifty years. The test asks three questions: How serious is what you stand to lose? How likely is the government to get it wrong without more process? And what would it actually cost to do more? The answers determine how much process the Constitution requires before the government acts — and whether you get a hearing at all before your benefits stop, your license is suspended, or you're barred from flying.
Working through Social Security disability terminations, ten-day school suspensions (Goss v. Lopez), civil service firings (Loudermill), and the no-fly list, they show how the same three-factor framework produces dramatically different results depending on context — from a full evidentiary hearing to a conversation in a principal's office.
The sharpest tension: Mathews said that paper review of medical evidence was good enough to terminate disability benefits without a prior hearing. But for conditions like chronic pain, depression, and fibromyalgia — where credibility is everything — a paper review misses exactly what a hearing would catch. The constitutional minimum and the practical reality diverged, and eventually Congress had to step in.
Mathews doesn't just tell courts how to evaluate procedures after the fact. It's the design specification agencies are supposed to use when they build their systems in the first place — and when they don't, courts use it to force a redesign.
00:00:14 --> 00:00:17 Hello, and welcome to Administrative Remedies,
00:00:17 --> 00:00:19 because you can't fix what you don't understand.
00:00:19 --> 00:00:21 Brought to you in part by the University of Tulsa
00:00:21 --> 00:00:24 College of Law. I'm Gwendolyn Savitz, an associate
00:00:24 --> 00:00:26 professor here at TU and the associate dean of
00:00:26 --> 00:00:28 research and intellectual life. And I'm Mark
00:00:28 --> 00:00:31 Rourke. I'm the dean of the College of Law. We'll
00:00:31 --> 00:00:33 be breaking down complex doctrines with real
00:00:33 --> 00:00:35 -life analogies and examples to demystify the
00:00:35 --> 00:00:37 world of administrative law for everyone trying
00:00:37 --> 00:00:39 to understand how government actually works.
00:00:40 --> 00:00:41 Agencies are the main way the federal government
00:00:41 --> 00:00:44 gets things done. It's not through Congress for
00:00:44 --> 00:00:46 reasons we'll be addressing over the course of
00:00:46 --> 00:00:52 this series. Mark, we are in Oklahoma. The famed
00:00:52 --> 00:00:55 Tornado Alley, Oklahoma. And I think a lot of
00:00:55 --> 00:00:57 people outside this part of the country have
00:00:57 --> 00:01:00 a very dramatic picture of what tornado season
00:01:00 --> 00:01:02 looks like. Sirens go off. Everyone runs to the
00:01:02 --> 00:01:05 basement. Yeah, that's not really how it works.
00:01:05 --> 00:01:08 No, not even close. So a tornado watch comes
00:01:08 --> 00:01:10 out. What do you do? Honestly, almost nothing.
00:01:11 --> 00:01:13 Watches are so common, you can't change your
00:01:13 --> 00:01:15 life every time one gets issued. You'd never
00:01:15 --> 00:01:18 leave the house. Yeah, same. A warning's different,
00:01:18 --> 00:01:20 though. A warning at least means that something's
00:01:20 --> 00:01:23 actually been spotted on radar, a storm that
00:01:23 --> 00:01:25 might produce a tornado. So now you're paying
00:01:25 --> 00:01:28 some attention. Yeah, you're checking the forecast,
00:01:28 --> 00:01:30 maybe thinking twice about whether to be out
00:01:30 --> 00:01:32 driving, but you still don't go to a shelter.
00:01:32 --> 00:01:35 No, for sure. And a warning covers a pretty wide
00:01:35 --> 00:01:37 area, and it doesn't even mean there's a tornado
00:01:37 --> 00:01:40 at all. That means the conditions in your county
00:01:40 --> 00:01:42 or nearby counties have produced something that
00:01:42 --> 00:01:45 might spawn tornadoes. So you adjust, but you
00:01:45 --> 00:01:48 don't panic. Right. And then the sirens go off.
00:01:48 --> 00:01:50 Yeah, this is the part people outside of Oklahoma
00:01:50 --> 00:01:52 don't really understand. They really don't. So
00:01:52 --> 00:01:55 the sirens here cover such a wide area that when
00:01:55 --> 00:01:57 they go off, most of the time the actual threat
00:01:57 --> 00:01:59 is miles away from you. They're basically a signal
00:01:59 --> 00:02:03 to turn on the local news. Right. So you go and
00:02:03 --> 00:02:05 check what the storm chasers are saying. You
00:02:05 --> 00:02:07 find out where the tornado actually is and whether
00:02:07 --> 00:02:09 it's moving towards you. The local stations here
00:02:09 --> 00:02:11 are really good at sending storm chasers to add
00:02:11 --> 00:02:14 to the weather radar. Yeah. And only if they're
00:02:14 --> 00:02:16 reporting something heading your direction, that's
00:02:16 --> 00:02:19 when you actually move. An interior room, a shelter,
00:02:19 --> 00:02:21 some people go outside. Ideally not outside.
00:02:22 --> 00:02:24 But yeah, it's easy to be on the coast watching
00:02:24 --> 00:02:26 the news and thinking, oh my gosh, the sirens
00:02:26 --> 00:02:28 were going off and people didn't even seek shelter.
00:02:28 --> 00:02:32 How stupid are they? But that really misunderstands
00:02:32 --> 00:02:35 what the sirens are in practice. Yeah, each stage
00:02:35 --> 00:02:37 really gets a different response. The stakes
00:02:37 --> 00:02:40 haven't changed. Your family's safety is at stake
00:02:40 --> 00:02:43 the entire time. Right. What changes is the probability.
00:02:44 --> 00:02:46 A watch means storms are possible. A warning
00:02:46 --> 00:02:49 means something's been detected. Sirens mean
00:02:49 --> 00:02:51 it's close enough to investigate. Confirmed rotation
00:02:51 --> 00:02:54 heading your way. Now you're taking action. Yeah,
00:02:54 --> 00:02:57 the risk goes up at every stage, so your response
00:02:57 --> 00:02:59 then escalates. And it's worth putting this in
00:02:59 --> 00:03:02 perspective. Yes, tornadoes can be dangerous,
00:03:02 --> 00:03:04 but every time you put your kid in a car and
00:03:04 --> 00:03:06 drive somewhere, you're doing something that's
00:03:06 --> 00:03:09 genuinely dangerous. But you do it anyway because
00:03:09 --> 00:03:11 the alternative would be you never go anywhere.
00:03:11 --> 00:03:12 You don't keep them home from school because
00:03:12 --> 00:03:15 car accidents exist or because they might catch
00:03:15 --> 00:03:17 a cold. You put them in a car seat. You drive
00:03:17 --> 00:03:20 carefully. You accept some residual risk as part
00:03:20 --> 00:03:23 of life. And tornado season is really the same
00:03:23 --> 00:03:25 calculus. And here's something interesting. The
00:03:25 --> 00:03:28 government could actually fix the sirens. They
00:03:28 --> 00:03:31 could design it so that... Each siren only triggers
00:03:31 --> 00:03:33 when that specific siren is in the path of the
00:03:33 --> 00:03:36 tornado. Okay, but they haven't done that. No,
00:03:36 --> 00:03:39 they sure haven't. So instead, the sirens cover
00:03:39 --> 00:03:41 this huge area, which means that people learn
00:03:41 --> 00:03:43 to treat them as information rather than commands.
00:03:44 --> 00:03:46 The system's design pushes the cost of sorting
00:03:46 --> 00:03:49 it out onto you. So the process itself could
00:03:49 --> 00:03:52 be designed better, but because it isn't, people
00:03:52 --> 00:03:54 have compensated that by ignoring certain parts
00:03:54 --> 00:03:57 of the system. Right. But notice what you're
00:03:57 --> 00:03:59 constantly doing. The stakes here never change.
00:03:59 --> 00:04:02 Your family safety is at stake the whole time.
00:04:02 --> 00:04:04 But you're calibrating how likely the danger
00:04:04 --> 00:04:07 is right now and what it would cost to take more
00:04:07 --> 00:04:10 precautions. Probability and cost of response
00:04:10 --> 00:04:12 are different. The Supreme Court said due process
00:04:12 --> 00:04:15 works the same way. You calibrate the process
00:04:15 --> 00:04:18 you need to the situation. But the legal framework
00:04:18 --> 00:04:21 adds a variable that the tornado analogy doesn't.
00:04:21 --> 00:04:24 So in the tornado, the stakes are constant. It's
00:04:24 --> 00:04:28 always your family's safety. But in Matthews,
00:04:28 --> 00:04:31 the stakes change. Losing your welfare benefits
00:04:31 --> 00:04:33 is different from a 10 -day school suspension,
00:04:33 --> 00:04:35 which is different from getting your car towed.
00:04:36 --> 00:04:38 The amount of process the Constitution requires
00:04:38 --> 00:04:41 depends on three factors. The private interest
00:04:41 --> 00:04:43 at stake, the risk of getting it wrong without
00:04:43 --> 00:04:45 more process, and the cost of additional safeguards.
00:04:45 --> 00:04:49 Last episode, we looked at Roth, Sitterman, and
00:04:49 --> 00:04:51 Loudermil to ask, do you have a property or a
00:04:51 --> 00:04:54 liberty interest at stake? A legitimate claim
00:04:54 --> 00:04:56 of entitlement, not just a hope or expectation.
00:04:57 --> 00:04:59 If the answer is no, you're not entitled to any
00:04:59 --> 00:05:01 process. Government can act without a hearing
00:05:01 --> 00:05:04 full stop. Like Professor Roth himself, a one
00:05:04 --> 00:05:07 -year contract did not equal an entitlement to
00:05:07 --> 00:05:09 renewal. So there was no hearing required by
00:05:09 --> 00:05:13 the court. Today we ask, assuming you do have
00:05:13 --> 00:05:15 a property or liberty interest at stake, how
00:05:15 --> 00:05:17 much process does the government owe you before
00:05:17 --> 00:05:20 it acts? Before, that's the key word. It's the
00:05:20 --> 00:05:22 whole thing. Everyone agrees that you eventually
00:05:22 --> 00:05:25 get some kind of hearing. The fight is about
00:05:25 --> 00:05:27 timing. Can they cut off your benefits first
00:05:27 --> 00:05:29 and give you a hearing later? Or do they have
00:05:29 --> 00:05:31 to hear you out before they take anything away?
00:05:31 --> 00:05:34 And the answer is? It depends. Of course. Yeah.
00:05:35 --> 00:05:38 But Matthews tells you what it depends on. Roth
00:05:38 --> 00:05:40 tells agencies whether constitutional procedures
00:05:40 --> 00:05:43 attach at all. Matthews tells them what the minimum
00:05:43 --> 00:05:46 constitutional floor is, what those procedures
00:05:46 --> 00:05:49 have to look like, and specifically how much
00:05:49 --> 00:05:51 has to happen before the deprivation versus after.
00:05:52 --> 00:05:53 To understand what Matthews did, you have to
00:05:53 --> 00:05:56 understand what came before it. We already talked
00:05:56 --> 00:05:58 about Goldberg v. Kelly a couple episodes ago.
00:05:58 --> 00:06:01 That was the welfare termination case. Right.
00:06:01 --> 00:06:04 New York was terminating welfare first and offering
00:06:04 --> 00:06:06 a hearing after that. The court said that wasn't
00:06:06 --> 00:06:09 good enough when survival was at stake. Right.
00:06:09 --> 00:06:11 What we didn't necessarily focus on was how much
00:06:11 --> 00:06:14 process the court actually required. Here, read
00:06:14 --> 00:06:17 what Brennan said about why a paper review wasn't
00:06:17 --> 00:06:19 sufficient. Written submissions do not afford
00:06:19 --> 00:06:22 the flexibility of oral presentations. They do
00:06:22 --> 00:06:24 not permit the recipient to mold his argument
00:06:24 --> 00:06:26 to the issues the decision maker appears to regard
00:06:26 --> 00:06:29 as important. Particularly where credibility
00:06:29 --> 00:06:31 and veracity are at issue, written submissions
00:06:31 --> 00:06:34 are a wholly unsatisfactory basis for decision.
00:06:34 --> 00:06:37 So that's the key insight, that there are situations
00:06:37 --> 00:06:39 where seeing a person, watching them testify,
00:06:39 --> 00:06:42 assessing their credibility, is fundamentally
00:06:42 --> 00:06:45 different from reading their file. And Brendan
00:06:45 --> 00:06:47 is drawing on something really basic and evidence
00:06:47 --> 00:06:49 -well. The whole reason we have live testimony
00:06:49 --> 00:06:52 is that demeanor matters. So from that reasoning,
00:06:52 --> 00:06:54 the court required a full suite of procedures.
00:06:55 --> 00:06:57 Written notice, the right to appear in person,
00:06:57 --> 00:07:00 the right to a lawyer, witnesses, cross -examination,
00:07:00 --> 00:07:03 an impartial decision -maker, a written decision,
00:07:03 --> 00:07:05 essentially the entire architecture of a civil
00:07:05 --> 00:07:09 trial. Right, for welfare termination. And that's
00:07:09 --> 00:07:12 a remarkable amount of process. It is. So after
00:07:12 --> 00:07:14 Goldberg, the question is whether this applies
00:07:14 --> 00:07:17 everywhere. Since new cases kept arriving in
00:07:17 --> 00:07:19 all sorts of different contexts. Give me some
00:07:19 --> 00:07:22 examples. Like school suspensions that need to
00:07:22 --> 00:07:24 be resolved in minutes. Yeah. You can't hold
00:07:24 --> 00:07:25 a trial every time a kid gets into a fight in
00:07:25 --> 00:07:28 the hallway. Right. Or employment terminations
00:07:28 --> 00:07:30 where something serious has happened and the
00:07:30 --> 00:07:32 employer needs to act really quickly. So Goldberg
00:07:32 --> 00:07:35 wasn't wrong. It just wasn't universal. So six
00:07:35 --> 00:07:38 years later, the court built a framework in Matthews
00:07:38 --> 00:07:40 that can calibrate the process to the context
00:07:40 --> 00:07:43 rather than just applying one template to everything.
00:07:43 --> 00:07:46 In Matthews, George Eldridge is receiving Social
00:07:46 --> 00:07:49 Security disability insurance based on chronic
00:07:49 --> 00:07:52 anxiety and back problems. What we famously call
00:07:52 --> 00:07:55 SSDI. So not really a means -tested process,
00:07:55 --> 00:07:58 but based on his work history and his payroll
00:07:58 --> 00:08:01 tax contributions. Right. Still absolutely a
00:08:01 --> 00:08:03 property interest under Roth. He has a legitimate
00:08:03 --> 00:08:05 claim of entitlement to those benefits. And it's
00:08:05 --> 00:08:08 undisputed. The question is what happened when
00:08:08 --> 00:08:11 Social Security decided he was no longer disabled.
00:08:11 --> 00:08:13 Right. They sent him a questionnaire. He responded.
00:08:14 --> 00:08:16 His doctor submitted reports. A state agency
00:08:16 --> 00:08:19 reviewed the medical evidence on paper. And then
00:08:19 --> 00:08:21 Social Security sent a letter explaining the
00:08:21 --> 00:08:24 decision. And he could seek reconsideration and
00:08:24 --> 00:08:26 even eventually get a hearing. Right. It was
00:08:26 --> 00:08:28 just the benefits terminated before the hearing.
00:08:29 --> 00:08:32 That's exactly what Goldberg said was unconstitutional
00:08:32 --> 00:08:35 for welfare, though. But the Supreme Court said
00:08:35 --> 00:08:37 disability benefits are different. And in explaining
00:08:37 --> 00:08:40 why, they gave us the three -factor test. Okay.
00:08:40 --> 00:08:42 Identification of the specific dictates of due
00:08:42 --> 00:08:45 process generally requires consideration of three
00:08:45 --> 00:08:48 distinct factors. First, the private interest
00:08:48 --> 00:08:50 that will be affected by the official action.
00:08:50 --> 00:08:53 Second, the risk of an erroneous deprivation
00:08:53 --> 00:08:55 of such an interest through the procedures used
00:08:55 --> 00:08:58 and the probable value, if any, of additional
00:08:58 --> 00:09:01 or substitute procedural safeguards. And finally,
00:09:01 --> 00:09:03 the government's interest, including the function
00:09:03 --> 00:09:05 involved and the fiscal and administrative burdens
00:09:05 --> 00:09:08 of the additional substitute procedural requirement
00:09:08 --> 00:09:11 would entail. So, these three factors balance
00:09:11 --> 00:09:13 against each other, the private interest, the
00:09:13 --> 00:09:16 risk of error, and government interest. Right.
00:09:16 --> 00:09:19 First factor. The private interest that will
00:09:19 --> 00:09:21 be affected. This is really looking at how much
00:09:21 --> 00:09:23 does this matter to the individual. In Goldberg,
00:09:23 --> 00:09:26 it was about survivor, literally the means to
00:09:26 --> 00:09:29 live. Where does disability insurance fall on
00:09:29 --> 00:09:30 that spectrum? The reason we're not getting the
00:09:30 --> 00:09:34 same result is because the court says SSDI is
00:09:34 --> 00:09:36 serious, but it's not the same level as welfare.
00:09:36 --> 00:09:39 And this is where that really important distinction
00:09:39 --> 00:09:44 matters. So SSDI versus SSI? Yeah. For SSDI,
00:09:44 --> 00:09:46 you qualify based on your work history and payroll
00:09:46 --> 00:09:50 tax contributions. It is not means tested. Unlike
00:09:50 --> 00:09:53 SSI or Supplemental Security Income, which is
00:09:53 --> 00:09:55 means tested. So these are two kinds of disability
00:09:55 --> 00:09:59 benefits. Eldridge was on the first one, SSDI,
00:09:59 --> 00:10:02 and the court reasoned that this type of recipient
00:10:02 --> 00:10:06 was more likely to have other resources. Whether
00:10:06 --> 00:10:08 that assumption was actually true for most SSDI
00:10:08 --> 00:10:12 claimants is... Definitely a question at least
00:10:12 --> 00:10:15 worth flagging. Because, candidly, a lot of people
00:10:15 --> 00:10:18 on SSDI are not exactly flush with backup income.
00:10:19 --> 00:10:21 They definitely are not. But the court was still
00:10:21 --> 00:10:24 drawing this line. And this was also before welfare
00:10:24 --> 00:10:27 was overhauled in the 1990s. So at the time,
00:10:27 --> 00:10:31 welfare itself did function as a more true safety
00:10:31 --> 00:10:34 net, like a backstop for people who had literally
00:10:34 --> 00:10:37 nothing else. SSDI, by contrast, was a benefit
00:10:37 --> 00:10:39 you earned through work, and the court assumed
00:10:39 --> 00:10:41 that meant you had other resources to draw on
00:10:41 --> 00:10:43 while you waited for a hearing. Okay, let's talk
00:10:43 --> 00:10:45 factor two. All right, the risk of erroneous
00:10:45 --> 00:10:47 deprivation through the procedures used and the
00:10:47 --> 00:10:50 probable value of additional safeguards. So,
00:10:50 --> 00:10:53 really, two questions in one. How likely is the
00:10:53 --> 00:10:55 government to get it wrong, and would better
00:10:55 --> 00:10:58 procedures actually fix it? Right, this is where
00:10:58 --> 00:11:01 the two cases really diverge. In Goldberg, they're
00:11:01 --> 00:11:03 saying that welfare eligibility will turn on
00:11:03 --> 00:11:05 these really subjective factors. Is this person
00:11:05 --> 00:11:07 really looking for work? Do they have hidden
00:11:07 --> 00:11:10 resources? Those are credibility determinations,
00:11:10 --> 00:11:13 exactly what Brennan said would require an in
00:11:13 --> 00:11:17 -person, face -to -face evaluation. But in Matthews,
00:11:17 --> 00:11:18 the court says that disability is more objective.
00:11:19 --> 00:11:22 And here's where they say it. The decision whether
00:11:22 --> 00:11:24 to discontinue disability benefits will turn
00:11:24 --> 00:11:28 in most cases upon routine, standard, and unbiased
00:11:28 --> 00:11:31 medical reports by physician specialists concerning
00:11:31 --> 00:11:33 a subject whom they have personally examined.
00:11:34 --> 00:11:36 That's where the criticism comes in. Because
00:11:36 --> 00:11:39 we just spent an entire episode talking about
00:11:39 --> 00:11:43 consultative exams, doctors the claimant didn't
00:11:43 --> 00:11:45 choose, 30 -minute appointments, examinations
00:11:45 --> 00:11:47 ordered by the very agency that was deciding
00:11:47 --> 00:11:50 the case. Routine, standard, and unbiased is
00:11:50 --> 00:11:53 doing a lot of work in that case. It really is.
00:11:54 --> 00:11:56 At least for some conditions, the court's assumptions
00:11:56 --> 00:11:58 probably hold. A broken bone will show up on
00:11:58 --> 00:12:02 an x -ray. But chronic pain, depression, fibromyalgia?
00:12:03 --> 00:12:05 In those types of conditions, the real question
00:12:05 --> 00:12:08 is, how much is this limiting you? Paper review
00:12:08 --> 00:12:10 will miss exactly the kind of thing an oral hearing
00:12:10 --> 00:12:13 would catch. Which is interesting because Brennan
00:12:13 --> 00:12:15 already said this in Goldberg. Written submissions
00:12:15 --> 00:12:18 are wholly unsatisfactory when credibility matters.
00:12:18 --> 00:12:21 And for a lot of disability claims, credibility
00:12:21 --> 00:12:24 is central. The claimant says the pain is debilitating,
00:12:24 --> 00:12:27 the medical records show a diagnosis, but can't
00:12:27 --> 00:12:29 quantify the experience. This is the most criticized
00:12:29 --> 00:12:32 part of Matthews. But it's decided at a time
00:12:32 --> 00:12:37 when more of the claims probably were more objective.
00:12:37 --> 00:12:39 The court treated disability like a documentary
00:12:39 --> 00:12:42 question. And how's that held up? Congress did
00:12:42 --> 00:12:45 eventually strengthen ALJ hearing protections
00:12:45 --> 00:12:49 and allowed benefits to continue for some claimants.
00:12:49 --> 00:12:52 This kind of reflects the reality of the limits
00:12:52 --> 00:12:54 of paper review. Here, the practice outgrew the
00:12:54 --> 00:12:57 constitutional minimum. Okay. What's the third
00:12:57 --> 00:12:59 factor? This is the government's interest. This
00:12:59 --> 00:13:02 includes the fiscal and administrative burden
00:13:02 --> 00:13:05 that pre -termination hearings for every disability
00:13:05 --> 00:13:08 case would be enormously expensive. How expensive?
00:13:09 --> 00:13:12 Well, we have thousands of ALJs. We have support
00:13:12 --> 00:13:14 staff. They're already backlog stretching months.
00:13:14 --> 00:13:16 If you require a hearing before termination,
00:13:16 --> 00:13:18 benefits keep flowing until the hearing happens.
00:13:19 --> 00:13:21 Even if the recipient actually isn't disabled
00:13:21 --> 00:13:24 anymore? So the government is paying benefits.
00:13:24 --> 00:13:28 It may never recover. Yes, exactly. It's worth
00:13:28 --> 00:13:31 noting that the government interest also includes
00:13:31 --> 00:13:33 the interest in making sure people get the benefits
00:13:33 --> 00:13:36 they should actually get. So it's not just trying
00:13:36 --> 00:13:39 to limit the cost as much as possible. But once
00:13:39 --> 00:13:41 we have all of these in mind, we balance them.
00:13:41 --> 00:13:44 A meaningful private interest, but not survival
00:13:44 --> 00:13:48 level in disability cases. The theoretical low
00:13:48 --> 00:13:51 marginal value of additional procedures, given
00:13:51 --> 00:13:53 the documentary nature of the evidence. And a
00:13:53 --> 00:13:55 high government burden if you require additional
00:13:55 --> 00:13:58 procedures. So a post -termination hearing is
00:13:58 --> 00:14:00 constitutionally adequate. Right. That's the
00:14:00 --> 00:14:03 holding. You get a hearing, just not before termination.
00:14:04 --> 00:14:06 But if you're sitting at home and you can't work
00:14:06 --> 00:14:08 and the benefits you've been living on just stopped
00:14:08 --> 00:14:11 and the Constitution says, just wait, we'll get
00:14:11 --> 00:14:14 to you. Right. But remember, there was at least
00:14:14 --> 00:14:16 notice about what was happening and there was
00:14:16 --> 00:14:19 an opportunity to submit additional documentation.
00:14:19 --> 00:14:22 So it's not that you're going to automatically
00:14:22 --> 00:14:26 just get a letter in the mail one day that says
00:14:26 --> 00:14:29 your benefits have stopped. There is something.
00:14:29 --> 00:14:33 It's just not going to be an oral hearing. This
00:14:33 --> 00:14:35 gap between you'll get a hearing eventually and
00:14:35 --> 00:14:38 you have no income right now can be a year or
00:14:38 --> 00:14:40 more. That is a long time to go without benefits
00:14:40 --> 00:14:43 you've been relying on. And one of the things
00:14:43 --> 00:14:46 to make clear at this point, right, is when you
00:14:46 --> 00:14:49 do have that hearing and you do get those benefits,
00:14:49 --> 00:14:52 you get caught up. Right. Right. You do. If they
00:14:52 --> 00:14:54 took it away erroneously, when they determine
00:14:54 --> 00:14:57 they took it away erroneously, you'll get all
00:14:57 --> 00:15:00 of those back benefits. And so the court is weighing
00:15:00 --> 00:15:02 the system's cost against this person's life.
00:15:02 --> 00:15:06 Yeah. And here the system won. You can definitely
00:15:06 --> 00:15:08 disagree with how the court characterized these
00:15:08 --> 00:15:11 factors. A lot of people do. We've already said
00:15:11 --> 00:15:13 that these assumptions don't all necessarily
00:15:13 --> 00:15:15 hold up in practice, but this is what they're
00:15:15 --> 00:15:18 reasoning. Did they basically do this everywhere?
00:15:18 --> 00:15:21 To some extent, yes. We can look at Goss v. Lopez.
00:15:21 --> 00:15:24 This was decided between when Goldberg and Matthews
00:15:24 --> 00:15:27 were decided. So they're obviously not applying
00:15:27 --> 00:15:29 the Matthews test, but they're functionally looking
00:15:29 --> 00:15:32 at the same factors. In Goss, students in Columbus,
00:15:32 --> 00:15:35 Ohio, were suspended for up to 10 days without
00:15:35 --> 00:15:38 any hearing. OK, the first question here is,
00:15:38 --> 00:15:40 do these students have a property or a liberty
00:15:40 --> 00:15:43 interest? The answer is yes. Ohio law required
00:15:43 --> 00:15:45 free education through high school. That created
00:15:45 --> 00:15:48 an entitlement under Roth. Plus, they had a liberty
00:15:48 --> 00:15:51 interest in their reputation as well as the punishment
00:15:51 --> 00:15:54 by the state. So suspension implied wrongdoing.
00:15:54 --> 00:15:57 Yeah, so they definitely cleared the threshold
00:15:57 --> 00:15:59 here. They have some interest that needs to be
00:15:59 --> 00:16:02 protected. So now we talk about how much process
00:16:02 --> 00:16:05 they're due. Let's look at these factors. Factor
00:16:05 --> 00:16:08 one, private interest, education matters, but
00:16:08 --> 00:16:10 we're also talking about a 10 -day suspension,
00:16:10 --> 00:16:13 not expulsion. It's just a temporary interference.
00:16:13 --> 00:16:17 Right. So factor two, the risk of error. School
00:16:17 --> 00:16:20 discipline situations are fast moving. A fight
00:16:20 --> 00:16:22 breaks out. Principal needs to respond, but mistakes
00:16:22 --> 00:16:25 happen. The wrong kid gets blamed or the story
00:16:25 --> 00:16:27 is more complicated than it looks. Some process
00:16:27 --> 00:16:31 would catch those errors. And the last one. The
00:16:31 --> 00:16:33 government interests in maintaining order and
00:16:33 --> 00:16:35 schools can't really hold a full trial every
00:16:35 --> 00:16:38 time a student misbehaves. Okay, let's read what
00:16:38 --> 00:16:41 the court requires. We only hold that in being
00:16:41 --> 00:16:43 given an opportunity to explain his version of
00:16:43 --> 00:16:46 the facts at this discussion, the student first
00:16:46 --> 00:16:49 be told what he's accused of doing and what the
00:16:49 --> 00:16:53 basis of the accusation is. That's it. Tell the
00:16:53 --> 00:16:54 student what they're accused of and give them
00:16:54 --> 00:16:57 the chance to explain. This is really minimal.
00:16:57 --> 00:16:59 This can happen in a conversation in the principal's
00:16:59 --> 00:17:02 office. We don't need any sort of formal hearing,
00:17:03 --> 00:17:06 no lawyers, no cross -examination. This is dramatically
00:17:06 --> 00:17:09 less than Goldberg. But more than nothing. Right.
00:17:09 --> 00:17:12 This is the low end of what the test is going
00:17:12 --> 00:17:15 to say we need. So Goss is really minimal process.
00:17:15 --> 00:17:19 All you need is a conversation. Notice of what
00:17:19 --> 00:17:21 you're accused of and an opportunity to respond.
00:17:22 --> 00:17:24 We also talked about Loudermill last episode
00:17:24 --> 00:17:25 when we mentioned the bitter with the sweet.
00:17:26 --> 00:17:28 Let's work through Matthews with that one. Remember,
00:17:29 --> 00:17:31 James Loudermill was a civil service employee.
00:17:31 --> 00:17:34 He had a property interest in his job, but he
00:17:34 --> 00:17:36 lied on his application to get the job, so they
00:17:36 --> 00:17:39 fired him. He had a full hearing afterward, but
00:17:39 --> 00:17:41 nothing before. All right, let's look at the
00:17:41 --> 00:17:44 factors. First one, the private interest. Okay,
00:17:44 --> 00:17:48 losing your job is economically serious. Not
00:17:48 --> 00:17:50 survival level like welfare, but more significant
00:17:50 --> 00:17:54 than a school suspension. Right. And losing your
00:17:54 --> 00:17:56 job is also different from disability because
00:17:56 --> 00:17:59 there's no back pay guarantee here. So if you're
00:17:59 --> 00:18:00 unemployed for months while you wait for a hearing,
00:18:01 --> 00:18:03 you might lose out on that income even if the
00:18:03 --> 00:18:05 court determines you were wrongly discharged.
00:18:05 --> 00:18:09 Okay. Factor two, risk of error. Higher than
00:18:09 --> 00:18:11 Matthews, employment termination often turns
00:18:11 --> 00:18:15 on disputed facts like, did he really lie? Was
00:18:15 --> 00:18:18 it a mistake? Was the lie material? Right. We
00:18:18 --> 00:18:21 could have a situation here where he didn't realize
00:18:21 --> 00:18:23 he'd been convicted of a felony because he'd
00:18:23 --> 00:18:26 reached some plea agreement that resulted in
00:18:26 --> 00:18:28 probation. He was never in jail and just didn't
00:18:28 --> 00:18:31 understand it was technically a felony. Yeah.
00:18:31 --> 00:18:33 And an oral explanation actually could matter
00:18:33 --> 00:18:35 here. It could shape his credibility and be able
00:18:35 --> 00:18:37 to respond to these things. It could also be
00:18:37 --> 00:18:40 a situation where he generated a resume through
00:18:40 --> 00:18:44 AI and it created jobs that he never was at.
00:18:44 --> 00:18:46 That's for our students out there in the world.
00:18:46 --> 00:18:50 Don't use AI in that way. Right. That gets us
00:18:50 --> 00:18:53 to the third factor, right, which is the government's
00:18:53 --> 00:18:55 legitimate interest in removing problem employees
00:18:55 --> 00:18:58 quickly. So here's the thing. What process was
00:18:58 --> 00:19:02 Loudermill actually asking for? Not a full Goldberg
00:19:02 --> 00:19:04 trial, just notice of the charges and a chance
00:19:04 --> 00:19:07 to respond. And that's pretty cheap. I mean,
00:19:07 --> 00:19:09 it's a meeting with your supervisor. It's a letter
00:19:09 --> 00:19:12 and a response. Right. So the argument that the
00:19:12 --> 00:19:14 government can't handle the burden loses force
00:19:14 --> 00:19:16 when the additional procedure is being requested
00:19:16 --> 00:19:19 or minimal. And that's a really important point
00:19:19 --> 00:19:22 about how we analyze these. When you're making
00:19:22 --> 00:19:24 a Matthews claim, you're saying. Here are the
00:19:24 --> 00:19:26 additional procedures I should have gotten before
00:19:26 --> 00:19:28 deprivation. And here's how they would have dramatically
00:19:28 --> 00:19:31 reduced the likelihood of error. The court needs
00:19:31 --> 00:19:33 to compare how much more you're asking for against
00:19:33 --> 00:19:36 what difference it would actually make. So again,
00:19:36 --> 00:19:38 we've got notice and an opportunity to respond
00:19:38 --> 00:19:41 before termination and a full evidentiary hearing
00:19:41 --> 00:19:45 afterwards. I want to go back to something. In
00:19:45 --> 00:19:47 Matthews itself, the government won. And in Goss,
00:19:47 --> 00:19:50 the students got minimal process, basically a
00:19:50 --> 00:19:53 conversation with their principal. Even in Loudermill,
00:19:53 --> 00:19:56 pre -termination process was just notice and
00:19:56 --> 00:19:59 a chance to respond. But in Goldberg, it's the
00:19:59 --> 00:20:02 outlier where the individual got serious process.
00:20:02 --> 00:20:05 Is there a pattern we're seeing here? Really,
00:20:05 --> 00:20:08 realistically, Goldberg was an outlier. That
00:20:08 --> 00:20:11 was the Supreme Court at the height of sympathy
00:20:11 --> 00:20:14 for people getting government assistance. But
00:20:14 --> 00:20:16 in most cases, it's true that government interests
00:20:16 --> 00:20:19 are easier to quantify than individuals suffering.
00:20:20 --> 00:20:22 These tests aren't always in favor of the government.
00:20:23 --> 00:20:26 Like Goldberg. Right. So in Goldberg, the individual
00:20:26 --> 00:20:29 won and won big. Factor two here cut hard for
00:20:29 --> 00:20:31 the claimant because credibility mattered and
00:20:31 --> 00:20:34 a paper review couldn't capture that. OK, but
00:20:34 --> 00:20:36 what about the others? So in Loudermill, factor
00:20:36 --> 00:20:39 three barely helped the government because the
00:20:39 --> 00:20:41 requested procedure was so cheap. All they're
00:20:41 --> 00:20:43 asking for was a conversation before firing,
00:20:43 --> 00:20:46 something that costs almost nothing. The test
00:20:46 --> 00:20:48 itself doesn't automatically have a built in
00:20:48 --> 00:20:52 thumb on the scale. So where does the tilt then
00:20:52 --> 00:20:55 come from? Certain categories of cases, particularly
00:20:55 --> 00:20:58 mass adjudication systems like disability, when
00:20:58 --> 00:21:00 we're talking about one government employee's
00:21:00 --> 00:21:02 termination, the cost of a pre -termination hearing
00:21:02 --> 00:21:05 is trivial. When you're talking about millions
00:21:05 --> 00:21:07 of disability determinations, the cost of pre
00:21:07 --> 00:21:09 -termination hearings for all of them is enormous.
00:21:10 --> 00:21:13 The test is sensitive to context, which can be
00:21:13 --> 00:21:16 both a strength and a source of criticism. So
00:21:16 --> 00:21:19 really... We need some kind of balancing test.
00:21:19 --> 00:21:23 What would the alternative be? In Goldberg, the
00:21:23 --> 00:21:26 alternative is everything. Yeah. So we could
00:21:26 --> 00:21:28 do full trial type hearings before the government
00:21:28 --> 00:21:31 could act. This would absolutely paralyze the
00:21:31 --> 00:21:34 administrative state and doesn't exist in any
00:21:34 --> 00:21:37 form now. Or no constitutional floor at all is
00:21:37 --> 00:21:39 an alternative. Right. Which would leave individuals
00:21:39 --> 00:21:42 the mercy of whatever Congress decided to provide.
00:21:42 --> 00:21:45 So Matthews is at least a middle ground. It's
00:21:45 --> 00:21:49 imperfect, but it's administrable. And it's flexible
00:21:49 --> 00:21:51 enough that it can require almost nothing in
00:21:51 --> 00:21:54 Goss and a near trial in Goldberg. So the flexibility
00:21:54 --> 00:21:57 is really the point and the problem. Yes. Well
00:21:57 --> 00:22:01 put. Yeah, I've prided myself in my ability to
00:22:01 --> 00:22:04 turn a phrase. OK, let's get a concrete example
00:22:04 --> 00:22:06 to explain this. All right. We can think about
00:22:06 --> 00:22:09 what happened after 9 -11. The government created
00:22:09 --> 00:22:12 the no -fly list. People barred from commercial
00:22:12 --> 00:22:14 air travel based on suspected terrorist connections.
00:22:15 --> 00:22:17 And so how much process did you get? For years,
00:22:17 --> 00:22:19 there was basically nothing. You found out you
00:22:19 --> 00:22:21 were on the list when you showed up at the airport
00:22:21 --> 00:22:23 and you couldn't board. There was no notice of
00:22:23 --> 00:22:26 why, no way to challenge it, no hearing of any
00:22:26 --> 00:22:30 kind. OK, let's run the Matthew factors on this
00:22:30 --> 00:22:32 no -fly list. Excellent. What about the first
00:22:32 --> 00:22:35 one, the private interest? OK, it's pretty significant.
00:22:35 --> 00:22:37 You can't fly commercially. That affects your
00:22:37 --> 00:22:39 job. It affects your family, your ability to
00:22:39 --> 00:22:41 function in modern life. And you've probably
00:22:41 --> 00:22:44 already spent at least the money for one airline
00:22:44 --> 00:22:46 ticket. We'll find out. Yes, and potentially
00:22:46 --> 00:22:48 money for other people you were traveling with
00:22:48 --> 00:22:51 that you can't use. And there's a real stigma.
00:22:51 --> 00:22:53 Being on the no -fly list is going to imply you're
00:22:53 --> 00:22:57 a terrorism suspect. Okay. Factor two, risk of
00:22:57 --> 00:22:59 error. It has to be significant. These lists
00:22:59 --> 00:23:01 are compiled from intelligence tips, database
00:23:01 --> 00:23:04 matching. The potential for mistakes is real.
00:23:04 --> 00:23:07 People with similar names can get flagged. And
00:23:07 --> 00:23:09 without any real process at all, there's no mechanism
00:23:09 --> 00:23:12 to catch those errors. Right. There are a lot
00:23:12 --> 00:23:15 of names that are super common. So in those instances,
00:23:15 --> 00:23:18 the marginal value of even basic procedures would
00:23:18 --> 00:23:21 be high. Okay. Factor three, the government interest
00:23:21 --> 00:23:23 in preventing terrorism is about as high as you
00:23:23 --> 00:23:26 can get. Right. But that doesn't mean we need
00:23:26 --> 00:23:29 zero process. You can protect intelligence sources
00:23:29 --> 00:23:32 without denying people opportunity to respond
00:23:32 --> 00:23:35 at all. Right. That's what happened. Lawsuits
00:23:35 --> 00:23:37 forced the government to create DHS TRIP, the
00:23:37 --> 00:23:40 Department of Homeland Security Traveler Redress
00:23:40 --> 00:23:43 Inquiry Program. And that created a process to
00:23:43 --> 00:23:45 challenge your inclusion on the list. So did
00:23:45 --> 00:23:49 travelers excluded get a full hearing? No, they
00:23:49 --> 00:23:51 definitely didn't. The government is still not
00:23:51 --> 00:23:52 going to be revealing classified information,
00:23:53 --> 00:23:56 but you at least get notice you're on the list
00:23:56 --> 00:23:58 and you can submit evidence and there's a review.
00:23:59 --> 00:24:01 OK, Matthews told the government your interest
00:24:01 --> 00:24:04 is real, but zero process isn't the answer. Find
00:24:04 --> 00:24:07 some middle ground. Right. Notice the pattern.
00:24:07 --> 00:24:10 We have a national security crisis. The government
00:24:10 --> 00:24:12 built a system with almost no process. Court
00:24:12 --> 00:24:14 said we take the government interest seriously,
00:24:14 --> 00:24:16 but. you still have to give people something.
00:24:17 --> 00:24:19 When the system doesn't give people any meaningful
00:24:19 --> 00:24:21 way to contest what's happening to them, math
00:24:21 --> 00:24:24 uses the tool that courts use to force a redesign.
00:24:24 --> 00:24:26 That's true whether we're talking about a no
00:24:26 --> 00:24:29 -fly list or benefits termination. And the framework
00:24:29 --> 00:24:31 is the same. Right. The framework's the same.
00:24:31 --> 00:24:34 The factors shift and the results change. And
00:24:34 --> 00:24:35 this happens constantly across the administrative
00:24:35 --> 00:24:38 state. When the VA built its new benefit appeals
00:24:38 --> 00:24:41 program, or when any agency is creating procedures,
00:24:41 --> 00:24:44 they should be thinking about Matthews. And so
00:24:44 --> 00:24:47 it becomes both a constitutional floor and a
00:24:47 --> 00:24:49 design framework. Right. So remember those tornado
00:24:49 --> 00:24:51 sirens? Yeah, the government could do a better
00:24:51 --> 00:24:55 job designing the system to target alerts instead
00:24:55 --> 00:24:58 of doing so on a countywide basis. Right. So
00:24:58 --> 00:25:01 same thing here. Matthews doesn't just tell agencies
00:25:01 --> 00:25:04 the constitutional minimum. It tells them that
00:25:04 --> 00:25:06 if their procedures are poorly designed, if they're
00:25:06 --> 00:25:09 the equivalent of blasting sirens for the entire
00:25:09 --> 00:25:12 county, and the system stops working. Roth tells
00:25:12 --> 00:25:14 you whether you've triggered due process protections.
00:25:14 --> 00:25:17 Matthews tells you what those protections look
00:25:17 --> 00:25:19 like. So together they form this operating system
00:25:19 --> 00:25:22 for how agencies handle individual adjudication.
00:25:22 --> 00:25:25 Most people think that due process is one thing,
00:25:25 --> 00:25:28 either you got it or you didn't. But it's actually
00:25:28 --> 00:25:30 a scale, a spectrum. Right. So it depends on
00:25:30 --> 00:25:32 what you're losing, on how likely mistakes are,
00:25:33 --> 00:25:34 and what additional procedures would actually
00:25:34 --> 00:25:37 cost and the difference they would make. Okay,
00:25:37 --> 00:25:39 let's try a new one. Your car gets towed for
00:25:39 --> 00:25:41 illegal parking. Do you get a hearing before
00:25:41 --> 00:25:44 they tow it? Well, what do you think? No, there's
00:25:44 --> 00:25:47 moderate private interest. It's your car, but
00:25:47 --> 00:25:50 you'll get it back. Low risk of error. Either
00:25:50 --> 00:25:52 you were parked illegally or you weren't. The
00:25:52 --> 00:25:53 high government interest in clearing illegal
00:25:53 --> 00:25:57 parked cars quickly. So post -deprivation hearing
00:25:57 --> 00:25:59 is usually fine. All right. So let's go back
00:25:59 --> 00:26:01 to what we mentioned at the very beginning. These
00:26:01 --> 00:26:04 emergency situations, like a pilot who might
00:26:04 --> 00:26:06 be drunk. The government interest now spikes.
00:26:07 --> 00:26:10 Hundreds of lives per flight. You suspend first
00:26:10 --> 00:26:12 and then have a hearing afterwards. That's been
00:26:12 --> 00:26:14 the law for almost 50 years. It is baked into
00:26:14 --> 00:26:17 how every agency designs its procedures or how
00:26:17 --> 00:26:20 every agency should be designing its procedures.
00:26:20 --> 00:26:22 And it means that the amount of process you get
00:26:22 --> 00:26:24 before the government acts depends on what you're
00:26:24 --> 00:26:26 losing, how likely it is that they will get it
00:26:26 --> 00:26:29 wrong. and what it would cost to do more. Which
00:26:29 --> 00:26:32 brings us to the next question. All right. Everything
00:26:32 --> 00:26:34 we've covered so far, talking about hearings
00:26:34 --> 00:26:37 and the procedural safeguards that the Matthews
00:26:37 --> 00:26:39 balancing test puts in place, this assumes that
00:26:39 --> 00:26:42 agencies are the right venue for these proceedings,
00:26:42 --> 00:26:44 that the government can adjudicate claims and
00:26:44 --> 00:26:47 impose penalties to their own tribunals rather
00:26:47 --> 00:26:49 than through the federal courts. And the Article
00:26:49 --> 00:26:52 III courts and the right to a trial by jury are
00:26:52 --> 00:26:55 implicated in the resolution process. Right.
00:26:55 --> 00:26:57 The courts are actually important in a couple
00:26:57 --> 00:26:59 ways. We're going to do a mini episode in a couple
00:26:59 --> 00:27:02 days on the immigration stuff that we're facing
00:27:02 --> 00:27:06 right now and how Matthews works there. And the
00:27:06 --> 00:27:08 next full episode, we will be talking about SEC
00:27:08 --> 00:27:12 v. Jarcosy. The Supreme Court said that some
00:27:12 --> 00:27:14 agency enforcement actions can't stay in -house.
00:27:14 --> 00:27:16 They have to go to a federal court with a jury.
00:27:16 --> 00:27:19 And that's not just a procedural tweak. That's
00:27:19 --> 00:27:21 a challenge to the basic premise that agencies
00:27:21 --> 00:27:24 can adjudicate it all. Right. We'll find out
00:27:24 --> 00:27:26 what darkening means for the agencies that have
00:27:26 --> 00:27:28 relied on in -house adjudication for decades
00:27:28 --> 00:27:32 and why this is yet another crack in the foundation
00:27:32 --> 00:27:35 of the administrative state. So that does it
00:27:35 --> 00:27:37 for today's episode of Administrative Remedies.
00:27:37 --> 00:27:40 Thank you for joining us today. Please, if you
00:27:40 --> 00:27:42 enjoy this podcast, enjoy this episode, give
00:27:42 --> 00:27:46 us a like on Spotify, iTunes or whatever platform
00:27:46 --> 00:27:49 you're listening on. And be sure to tune in next
00:27:49 --> 00:27:51 time where we'll continue to dive into the contours
00:27:51 --> 00:27:53 of administrative law, because remember, you
00:27:53 --> 00:27:55 can't fix what you don't understand.