In this episode, Gwen and Marc step back from Chevron to examine the older doctrine that both preceded it and now survives it: Skidmore deference. They begin with a medical analogy that contrasts expert judgment grounded in examination and experience with advice that merely sounds confident—setting up the central question of Skidmore: when agencies lack the power to control, how much weight should courts give to what they say?
They walk through Skidmore v. Swift & Co., a 1944 case involving firemen at a meatpacking plant and whether on-call waiting time counts as compensable work. The key issue was not disagreement over expertise, but authority: the agency administering the Fair Labor Standards Act had issued extensive guidance but lacked formal rulemaking power. The Court’s response—giving agency interpretations weight proportional to their persuasiveness—created a doctrine that applied specifically when agencies lacked the force of law.
From there, Gwen and Marc situate Skidmore in historical context, contrasting it with highly deferential cases like Yakus v. United States and showing that Skidmore was never meant to replace strong deference where Congress had delegated rulemaking authority. They then explain how Skidmore operates in practice through its familiar factors—thoroughness, reasoning, consistency, and expertise—and why it creates flexibility at the cost of predictability.
The episode closes by tracing Skidmore’s resurgence after United States v. Mead Corp., which limited Chevron to actions with the force of law and returned Skidmore to center stage for guidance documents, opinion letters, and other informal agency actions. This episode sets up the post-Chevron world by explaining what kind of deference remains—and what has been lost.
What They Cover in This Episode
- The facts and holding of Skidmore v. Swift & Co.
- Why lack of rulemaking authority mattered
- “Power to persuade, if lacking power to control”
- The historical baseline of strong agency deference
- The Skidmore factors and how courts apply them
- Why Skidmore is flexible but unpredictable
- United States v. Mead Corp. and the return of Skidmore
- How much of modern governance operates under Skidmore
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 --> 00:00:50 this series. Last time, we explored Chevron,
00:00:50 --> 00:00:52 the doctrine that set up an explicit two -part
00:00:52 --> 00:00:55 test that ruled administrative law for 40 years.
00:00:56 --> 00:01:01 Today, we're going back further to what, as we'll
00:01:01 --> 00:01:03 see next episode, is the only thing left standing,
00:01:03 --> 00:01:06 and that's Skidmore. To try and put this in perspective,
00:01:07 --> 00:01:09 let's think about two different patients. They're
00:01:09 --> 00:01:11 sitting in the same doctor's office. First, we
00:01:11 --> 00:01:14 have Chevelle. Okay. So Chevelle's been having
00:01:14 --> 00:01:18 headaches. The doctor does a complete exam. He
00:01:18 --> 00:01:20 checks her sinuses, her blood pressure, her sleep
00:01:20 --> 00:01:23 patterns, allergies, medication interactions,
00:01:23 --> 00:01:26 the whole deal. After all that, he says, take
00:01:26 --> 00:01:29 Zyvel twice a day for a week. This looks like
00:01:29 --> 00:01:31 allergy -related headaches. They should go away
00:01:31 --> 00:01:34 if we can clear up that congestion. So thorough
00:01:34 --> 00:01:37 evaluation, clear recommendation. And Chevelle
00:01:37 --> 00:01:39 thinks, sure, that makes sense, picks up Zyvel
00:01:39 --> 00:01:43 and goes home. Okay. She's sensible. And then
00:01:43 --> 00:01:46 we have Skye. So Skye walks out of the same office
00:01:46 --> 00:01:48 after the exact same meeting, same examination,
00:01:49 --> 00:01:51 same issue, and says, you know what? I'd really
00:01:51 --> 00:01:54 like a second opinion. So she pulls out her phone,
00:01:54 --> 00:01:58 opens ChatGBT, and says, what should I do for
00:01:58 --> 00:02:01 headaches? ChatGBT replies. Headaches frequently
00:02:01 --> 00:02:04 arise from subtle shifts in your body's mineral
00:02:04 --> 00:02:07 equilibrium, little electrolyte imbalances that
00:02:07 --> 00:02:09 accumulate through stress, green exposure, and
00:02:09 --> 00:02:11 irregular hydration. A simple reset is to take
00:02:11 --> 00:02:14 300 milligrams of magnesium glyconate, then drink
00:02:14 --> 00:02:16 a full glass of water with a pinch of mineral
00:02:16 --> 00:02:19 -rich sea salt. This combination supports neuromuscular
00:02:19 --> 00:02:22 relaxation, stabilizes cellular hydration, and
00:02:22 --> 00:02:24 helps your nervous system downshift out of the
00:02:24 --> 00:02:26 tension cycle that often produces headache symptoms.
00:02:27 --> 00:02:29 Okay, I think I see where we're heading. Sky
00:02:29 --> 00:02:32 thinks. Wow, this makes so much more sense than
00:02:32 --> 00:02:34 what the doctor was saying. I knew it was something
00:02:34 --> 00:02:36 deeper than allergies. Sea salt water. That sounds
00:02:36 --> 00:02:40 so natural, so holistic. Why didn't the doctor
00:02:40 --> 00:02:42 mention this? Probably because doctors just don't
00:02:42 --> 00:02:44 think about the whole body. We know that it's
00:02:44 --> 00:02:47 a particular source of water that makes you healthier.
00:02:48 --> 00:02:51 Clearly Skye knows that too. She's going with
00:02:51 --> 00:02:54 whichever sounds better to her. Yeah, she's considering
00:02:54 --> 00:02:57 them both and going with what sounds more authoritative.
00:02:58 --> 00:03:01 The problem and the reason why we're using this
00:03:01 --> 00:03:04 hypothetical situation is that ChatTPT sounds
00:03:04 --> 00:03:07 like that regardless of its veracity. Confident,
00:03:07 --> 00:03:09 like it's reading from the book of all knowledge,
00:03:09 --> 00:03:12 even when it is completely wrong. Every doctor
00:03:12 --> 00:03:14 in the country has patients coming in saying
00:03:14 --> 00:03:17 stuff like, ChatTPT says I have a brain tumor
00:03:17 --> 00:03:19 because my eyelid twitched. It's like Googling
00:03:19 --> 00:03:21 symptoms, which was a problem for a really long
00:03:21 --> 00:03:24 time. If you could also repeatedly ask Google
00:03:24 --> 00:03:26 if it was right, and it could reassure you that
00:03:26 --> 00:03:28 it is 100 % confident in what it's telling you.
00:03:28 --> 00:03:31 And we lawyers have the equivalent. Lawyers asking
00:03:31 --> 00:03:34 whether a made -up case is real, and ChatGPT
00:03:34 --> 00:03:38 says, yes, absolutely 100 % yes, and people trust
00:03:38 --> 00:03:43 that enough to turn it into a court. Yes, and
00:03:43 --> 00:03:44 at that point, they get sanctioned for using
00:03:44 --> 00:03:49 fake cases. ChatGPT and other AIs, we're just
00:03:49 --> 00:03:51 picking out one here, is great at generating
00:03:51 --> 00:03:54 stuff that sounds right. It's not great at reliably
00:03:54 --> 00:03:57 diagnosing medical conditions, particularly when
00:03:57 --> 00:03:59 people had just put in the symptoms themselves
00:03:59 --> 00:04:02 and we haven't had a lot of back and forth to
00:04:02 --> 00:04:04 try and clarify stuff. It's also really bad at
00:04:04 --> 00:04:07 verifying legal precedent, two fields where sounds
00:04:07 --> 00:04:10 right is not enough. But Skye sees that confident
00:04:10 --> 00:04:12 tone and its reasoning and thinks, wow, this
00:04:12 --> 00:04:15 must be the correct answer, even though the doctor's
00:04:15 --> 00:04:17 recommendation is based on an actual physical
00:04:17 --> 00:04:21 exam of her and his own medical reasoning. Yes.
00:04:21 --> 00:04:24 Chevelle is deferring to the trained expert who
00:04:24 --> 00:04:26 examined her. Skye is following the chat bot
00:04:26 --> 00:04:30 because the advice sounds better to her. So Chevelle
00:04:30 --> 00:04:34 is Chevron. Yes. And under Chevron. If Congress
00:04:34 --> 00:04:36 writes something ambiguous and the agency, the
00:04:36 --> 00:04:39 doctor here, does an analysis based on its expertise,
00:04:39 --> 00:04:42 courts defer to that interpretation. Because
00:04:42 --> 00:04:45 the agency is the expert in that case. Yes. Agencies
00:04:45 --> 00:04:48 are applying decades of institutional expertise.
00:04:49 --> 00:04:52 They're like Chabelle's doctor. And Sky is Skidmore.
00:04:52 --> 00:04:56 Yep. Under Skidmore, the court doesn't give any
00:04:56 --> 00:04:58 deference. It doesn't have to. Agencies under
00:04:58 --> 00:05:00 Skidmore get respect only to the extent that
00:05:00 --> 00:05:04 their reasoning is persuasive. So like Skye reading
00:05:04 --> 00:05:07 both answers and saying, I'll go with whichever
00:05:07 --> 00:05:10 one I like better. Yes, this is kind of a convincing
00:05:10 --> 00:05:13 this test. The agency has to earn deference.
00:05:13 --> 00:05:16 It won't receive it automatically. It has to
00:05:16 --> 00:05:18 persuade the court that its reading is the correct
00:05:18 --> 00:05:22 reading. But that also opens the door to Skye's
00:05:22 --> 00:05:24 mistake, giving equal or greater weight to an
00:05:24 --> 00:05:27 answer that just sounds confident or like what
00:05:27 --> 00:05:30 she wanted to hear, even if it's uninformed.
00:05:31 --> 00:05:34 So in theory, Skidmore lets courts pick the better
00:05:34 --> 00:05:37 reasoning. In practice, courts pick the explanation
00:05:37 --> 00:05:39 that often sounds more appealing. This is like
00:05:39 --> 00:05:42 Skye's chatty BT answer instead of going with
00:05:42 --> 00:05:45 the expert's detailed but complex reasoning that
00:05:45 --> 00:05:47 she doesn't totally understand. Which is how
00:05:47 --> 00:05:49 we end up with courts saying things like this
00:05:49 --> 00:05:52 EPA interpretation is unreasonable, even when
00:05:52 --> 00:05:55 the EPA's experts spent years studying the science
00:05:55 --> 00:05:59 and the courts and the court spent. maybe an
00:05:59 --> 00:06:02 afternoon with briefs. We can think about it
00:06:02 --> 00:06:05 as Chevron is Chevelle. She's trusting the expert
00:06:05 --> 00:06:08 who actually examined the problem. And Skidmore
00:06:08 --> 00:06:11 is Skye, deciding based on what seems persuasive.
00:06:12 --> 00:06:14 And that means that the non -expert's answer
00:06:14 --> 00:06:17 could just sound better, regardless of its veracity.
00:06:18 --> 00:06:20 And relying on the wrong one can cause headaches
00:06:20 --> 00:06:24 in medicine and in administrative law. Wasn't
00:06:24 --> 00:06:26 that such a good setup? That was nice. Okay,
00:06:26 --> 00:06:30 so where does Skidmore come from? 1944, this
00:06:30 --> 00:06:32 is Skidmore v. Swifton Company. What happened
00:06:32 --> 00:06:35 in the case? We had seven employees. They're
00:06:35 --> 00:06:38 at a meatpacking plant in Texas, in Fort Worth,
00:06:38 --> 00:06:40 and they're required to stay on the company premises
00:06:40 --> 00:06:43 waiting for fires to break out. Wait, they're
00:06:43 --> 00:06:45 just waiting for the fires to break out? Yes,
00:06:45 --> 00:06:49 they're just auxiliary firemen at this meatpacking
00:06:49 --> 00:06:51 plant. They have to be on site in case of an
00:06:51 --> 00:06:53 emergency, but while they're there waiting, they
00:06:53 --> 00:06:56 can sleep, they can play cards, they can eat.
00:06:56 --> 00:06:58 They can leave briefly if they're just running
00:06:58 --> 00:07:00 a local errand, as long as they're ready to respond
00:07:00 --> 00:07:02 if an alarm goes off. That sounds like a pretty
00:07:02 --> 00:07:05 good gig. Maybe. But the question was, is this
00:07:05 --> 00:07:07 time where they're just stuck there waiting,
00:07:07 --> 00:07:10 is this work under the Fair Labor Standards Act?
00:07:10 --> 00:07:12 Because if it is, they're owed overtime pay.
00:07:13 --> 00:07:16 And the statute didn't define that. No, the Fair
00:07:16 --> 00:07:18 Labor Standards Act required overtime pay for
00:07:18 --> 00:07:21 work over 40 hours a week. But is sleeping in
00:07:21 --> 00:07:25 the firehouse work if you are contractually required
00:07:25 --> 00:07:27 to do it? Is eating dinner while you're on call
00:07:27 --> 00:07:29 work? Is playing in checkers while you're waiting
00:07:29 --> 00:07:32 for the alarm work? I could see arguments both
00:07:32 --> 00:07:35 ways. They're not actively doing anything, but
00:07:35 --> 00:07:38 they're also not free to leave. Yeah, and that's
00:07:38 --> 00:07:40 what makes this case significant. The wage and
00:07:40 --> 00:07:42 hour division of the Department of Labor, that
00:07:42 --> 00:07:45 was the agency officially in charge of the FLSA,
00:07:45 --> 00:07:48 had developed extensive guidelines on exactly
00:07:48 --> 00:07:50 this kind of situation. They'd issued bulletins.
00:07:50 --> 00:07:53 They'd issued opinion letters. interpretive guides
00:07:53 --> 00:07:55 all trying to work out when waiting time counted
00:07:55 --> 00:07:58 as working time. The agency had thought about
00:07:58 --> 00:08:01 this and created a set of rules around what it
00:08:01 --> 00:08:03 means to work. Yeah, they had thought about it
00:08:03 --> 00:08:06 extensively. But the crucial detail here is that
00:08:06 --> 00:08:08 they didn't have any formal rulemaking authority.
00:08:08 --> 00:08:10 Congress hadn't given them the power to issue
00:08:10 --> 00:08:13 binding regulations defining work. All they could
00:08:13 --> 00:08:15 do was offer these different types of guidances
00:08:15 --> 00:08:18 and interpretations. So their opinion was just
00:08:18 --> 00:08:21 an opinion. That is what the company argued.
00:08:21 --> 00:08:23 And the court could have agreed. They don't have
00:08:23 --> 00:08:26 the rulemaking authority. We don't need to give
00:08:26 --> 00:08:28 them any deference. But instead, here's what
00:08:28 --> 00:08:31 the court said. So they said, we consider that
00:08:31 --> 00:08:33 the rulings, interpretations and opinions of
00:08:33 --> 00:08:35 the administrator. Under this act, while not
00:08:35 --> 00:08:37 controlling upon the courts by reason of their
00:08:37 --> 00:08:40 authority, do constitute a body of experience
00:08:40 --> 00:08:43 and informed judgment to which courts and litigants
00:08:43 --> 00:08:46 may properly resort for guidance. The weight
00:08:46 --> 00:08:48 of such a judgment in a particular case will
00:08:48 --> 00:08:51 depend upon the thoroughness evident in its consideration,
00:08:51 --> 00:08:54 the validity of its reasoning, its consistency
00:08:54 --> 00:08:57 with earlier and later pronouncements, and all
00:08:57 --> 00:08:59 those factors which give it power to persuade
00:08:59 --> 00:09:03 if lacking power to control. So power to persuade
00:09:03 --> 00:09:05 if lacking power to control. That's the heart
00:09:05 --> 00:09:08 of Skidmore. The agency couldn't force courts
00:09:08 --> 00:09:10 to accept its interpretation because, again,
00:09:10 --> 00:09:13 they had no formal rulemaking authority. But
00:09:13 --> 00:09:15 if the reasoning was good, courts should still
00:09:15 --> 00:09:19 listen. Here's the part that people miss. Skidmore
00:09:19 --> 00:09:21 wasn't significant because it created this new
00:09:21 --> 00:09:25 big deference test. It mattered because the relevant
00:09:25 --> 00:09:28 agency didn't have rulemaking authority. They
00:09:28 --> 00:09:30 couldn't issue binding regulations. All they
00:09:30 --> 00:09:32 could do was publish these interpretive guidances
00:09:32 --> 00:09:34 and memos. So courts weren't asking, should we
00:09:34 --> 00:09:37 defer to expert regulations? They were asking,
00:09:37 --> 00:09:40 should we defer to what an agency has to say,
00:09:40 --> 00:09:43 even when Congress didn't give the agency rulemaking
00:09:43 --> 00:09:46 power? Yes, that's why Skidmore says agency interpretations
00:09:46 --> 00:09:48 get respect only to the extent they're persuasive.
00:09:49 --> 00:09:50 They don't have the force of law behind them.
00:09:51 --> 00:09:53 But, and this is the key point, like we talked
00:09:53 --> 00:09:56 about last episode, courts have long been extremely
00:09:56 --> 00:09:59 deferential to agencies when they did have rulemaking
00:09:59 --> 00:10:02 authority. Skidmore was the exception, not the
00:10:02 --> 00:10:06 rule at that time. To try and explain this, we're
00:10:06 --> 00:10:09 going to look at Yakusby, United States. This
00:10:09 --> 00:10:11 was decided just a few months before Skidmore.
00:10:11 --> 00:10:14 This is a case about war time price controls,
00:10:14 --> 00:10:17 which had been granted an incredibly broad delegation
00:10:17 --> 00:10:19 of authority. And the Supreme Court basically
00:10:19 --> 00:10:21 said when Congress gives an agency the power
00:10:21 --> 00:10:23 to make binding rules, courts are not to micromanage
00:10:23 --> 00:10:26 that. Here's what they said. It's no objection
00:10:26 --> 00:10:29 that the determination of facts and the inferences
00:10:29 --> 00:10:31 to be drawn from them in the light of the statutory
00:10:31 --> 00:10:35 standards and declaration of policy call for
00:10:35 --> 00:10:37 the exercise of judgment. and for the formulation
00:10:37 --> 00:10:40 of subsidiary administrative policy within the
00:10:40 --> 00:10:43 prescribed statutory framework. If we translate
00:10:43 --> 00:10:46 that, they're saying, yeah, agencies are going
00:10:46 --> 00:10:48 to have to use judgment, and that's fine. That's
00:10:48 --> 00:10:51 their job. We don't need Congress to tell the
00:10:51 --> 00:10:53 agency exactly what to do. Then they go further.
00:11:11 --> 00:11:15 So this is incredibly deferential. The court
00:11:15 --> 00:11:17 is saying it doesn't matter whether Congress
00:11:17 --> 00:11:19 was specific or vague. They could have absolutely
00:11:19 --> 00:11:22 been clearer. This was... giving this agency
00:11:22 --> 00:11:24 the authority to determine what prices should
00:11:24 --> 00:11:27 be. Congress could have just said prices should
00:11:27 --> 00:11:29 be exactly what they were on this date. They
00:11:29 --> 00:11:31 didn't do that. They gave the agency authority
00:11:31 --> 00:11:33 to determine what date they would be using for
00:11:33 --> 00:11:36 the prices. And that meant that the role of the
00:11:36 --> 00:11:38 courts was to just check whether the agency had
00:11:38 --> 00:11:41 stayed within the lane Congress drew. And then
00:11:41 --> 00:11:43 we come to a line that really kind of crystallizes
00:11:43 --> 00:11:46 the constitutional philosophy of the era. So
00:11:46 --> 00:11:47 they said, as we have said, the Constitution
00:11:47 --> 00:11:51 has never been regarded as denying to Congress
00:11:51 --> 00:11:55 the necessary resources of flexibility and practicality
00:11:55 --> 00:11:59 to perform its function. Hence, it is relevant
00:11:59 --> 00:12:02 that Congress might itself have prescribed the
00:12:02 --> 00:12:04 maximum prices or provided a more rigid standard.
00:12:05 --> 00:12:08 Congress is not complying to that method of executing
00:12:08 --> 00:12:11 its policy, which involves the least possible
00:12:11 --> 00:12:14 delegation of discretion to administrative officers.
00:12:15 --> 00:12:18 This helps show why Skidmore looks so modest
00:12:18 --> 00:12:21 in comparison. At this time, courts were operating
00:12:21 --> 00:12:24 in a world where broad delegation and agency
00:12:24 --> 00:12:26 rulemaking authority meant strong judicial deference.
00:12:27 --> 00:12:30 Skidmore's respect proportional to persuasiveness
00:12:30 --> 00:12:33 wasn't the baseline. It was what courts did when
00:12:33 --> 00:12:36 agencies didn't have actual legal rulemaking
00:12:36 --> 00:12:39 power. Meanwhile, in cases like this, the court
00:12:39 --> 00:12:41 was saying, if Congress gave you rulemaking power,
00:12:41 --> 00:12:43 we're going to defer unless you're wildly off
00:12:43 --> 00:12:45 course. This is the contrast that the modern
00:12:45 --> 00:12:47 court tends to blur and why it's so important
00:12:47 --> 00:12:50 to understand Skidmore properly. It is worth
00:12:50 --> 00:12:53 noting that this was a wartime powers issue.
00:12:54 --> 00:12:56 So we're going to be particularly deferential
00:12:56 --> 00:12:58 to the executive then. But this was also a criminal
00:12:58 --> 00:13:01 case. So they could have very easily said it
00:13:01 --> 00:13:03 needed to be clearer than it was. So Skidmore
00:13:03 --> 00:13:07 wasn't for everything. Then what affects the
00:13:07 --> 00:13:10 way courts give to agency interpretations? Under
00:13:10 --> 00:13:12 Skidmore, we're looking at the thoroughness of
00:13:12 --> 00:13:15 the consideration. Did the agency think really
00:13:15 --> 00:13:17 hard about this, or is this just some quick opinion
00:13:17 --> 00:13:20 they dashed off before lunch? Okay, like whether
00:13:20 --> 00:13:23 my doctor actually examined someone or just took
00:13:23 --> 00:13:26 a brief look and dashed off a prescription. Yes,
00:13:26 --> 00:13:28 like the thorough examination the doctor was
00:13:28 --> 00:13:30 giving in the hypo at the beginning here. The
00:13:30 --> 00:13:33 next one is the validity of the reasoning. Does
00:13:33 --> 00:13:35 the interpretation make sense? Is there a logical
00:13:35 --> 00:13:37 connection between the statutory text and the
00:13:37 --> 00:13:39 conclusion? So not just the agency saying, trust
00:13:39 --> 00:13:41 us, we're the experts, but actually showing it's
00:13:41 --> 00:13:45 math. Exactly. And next, consistency with earlier
00:13:45 --> 00:13:48 and later pronouncements. Has the agency been
00:13:48 --> 00:13:50 saying the same thing over and over again, or
00:13:50 --> 00:13:52 has their position shifted depending on who's
00:13:52 --> 00:13:55 asking? I thought that with Chevron, agencies
00:13:55 --> 00:13:57 could change. Yes, and we'll get back to that.
00:13:58 --> 00:14:01 So next is this catch -all, all those factors
00:14:01 --> 00:14:03 which give it power to persuade if lacking power
00:14:03 --> 00:14:06 to control. That's things like, is this the area
00:14:06 --> 00:14:09 of the agency's core expertise? How long have
00:14:09 --> 00:14:11 they been interpreting the statute? Have they
00:14:11 --> 00:14:13 consulted with different parties? Is there evidence
00:14:13 --> 00:14:15 Congress was aware of their interpretation and
00:14:15 --> 00:14:18 endorsed it? So when you add all this up, what
00:14:18 --> 00:14:21 makes an interpretation highly persuasive versus
00:14:21 --> 00:14:24 not persuasive at all? Highly persuasive is a
00:14:24 --> 00:14:26 longstanding, consistent position from the agency,
00:14:26 --> 00:14:29 an agency that administers this exact statute
00:14:29 --> 00:14:32 that has been developing through careful study
00:14:32 --> 00:14:34 over years with detailed analyses explaining
00:14:34 --> 00:14:36 how it connects to the statutory text and purposes.
00:14:37 --> 00:14:40 which I'll point out is what the court was functionally
00:14:40 --> 00:14:42 saying the agency did in Skidmore. That sounds
00:14:42 --> 00:14:44 like it would get almost as much respect as a
00:14:44 --> 00:14:48 binding regulation. And they often did. But we
00:14:48 --> 00:14:49 can contrast this with something that would get
00:14:49 --> 00:14:53 low persuasiveness, a litigation position that
00:14:53 --> 00:14:55 the agency was only presenting after it had been
00:14:55 --> 00:14:58 sued, contradicting what they'd said earlier,
00:14:58 --> 00:15:02 some sort of agency document with minimal analysis,
00:15:02 --> 00:15:04 something that was outside their main expertise.
00:15:05 --> 00:15:08 And that basically gets no weight. Yeah. Courts
00:15:08 --> 00:15:11 are really skeptical of these ad hoc rationalizations.
00:15:12 --> 00:15:14 If you didn't think about the reasoning until
00:15:14 --> 00:15:16 you got sued, it's hard to claim it's your considered
00:15:16 --> 00:15:20 expert judgment. So it's genuinely a spectrum,
00:15:20 --> 00:15:23 not a yes or no question. Yeah. And that makes
00:15:23 --> 00:15:27 Skidmore flexible and unpredictable. Judges have
00:15:27 --> 00:15:29 real discretion in weighing these factors. Okay.
00:15:29 --> 00:15:31 So let's go back to Chevron for a second. Remind
00:15:31 --> 00:15:34 me the difference between Chevron and Skidmore.
00:15:35 --> 00:15:38 Chevron was this binary test. Step one, is the
00:15:38 --> 00:15:41 statute clear? If yes, that's the answer. If
00:15:41 --> 00:15:44 no, we go to step two. Is the agency's interpretation
00:15:44 --> 00:15:47 reasonable? If it is, the agency wins. Period.
00:15:48 --> 00:15:50 The court doesn't weigh factors. It doesn't consider
00:15:50 --> 00:15:53 persuasiveness. As long as it's reasonable, the
00:15:53 --> 00:15:56 agency interpretation controls. So Chevron was
00:15:56 --> 00:16:01 an on -off switch and Skidmore was a dimmer.
00:16:01 --> 00:16:04 Yes, and agencies loved Chevron because it was
00:16:04 --> 00:16:06 predictable. If you went through notice and comment
00:16:06 --> 00:16:09 rulemaking and your interpretation was reasonable,
00:16:09 --> 00:16:11 assuming the statute was ambiguous, you won.
00:16:11 --> 00:16:13 You didn't need to worry about whether a particular
00:16:13 --> 00:16:16 judge would find your reasoning persuasive enough.
00:16:16 --> 00:16:19 Okay, so then what happened to Skidmore? For
00:16:19 --> 00:16:21 a while, scholars were starting to think that
00:16:21 --> 00:16:23 Chevron was really all there was and Skidmore
00:16:23 --> 00:16:27 was basically irrelevant. I'm sensing that that
00:16:27 --> 00:16:30 has changed. It certainly has now, but it first
00:16:30 --> 00:16:33 changed in 2001 when the Supreme Court decided
00:16:33 --> 00:16:35 a case that functionally resurrected Skidmore.
00:16:35 --> 00:16:38 What happened in 2001? The United States beat
00:16:38 --> 00:16:42 Meade. Meade imported day planners, like spiral
00:16:42 --> 00:16:45 notebooks for scheduling appointments, and U
00:16:45 --> 00:16:47 .S. Customs had to classify those for tariff
00:16:47 --> 00:16:49 purposes. Were these going to count as bound
00:16:49 --> 00:16:52 diaries, which had a 4 % tariff, or were these
00:16:52 --> 00:16:56 other notebooks, which had a 0 % tariff? The
00:16:56 --> 00:16:59 fate of empires rests on whether a day planner
00:16:59 --> 00:17:02 is a diary. The legal issues here were crucial.
00:17:02 --> 00:17:05 The Customs Service issued thousands of these
00:17:05 --> 00:17:07 tariff classification rulings, and they were
00:17:07 --> 00:17:09 claiming that every single one of them deserved
00:17:09 --> 00:17:12 Chevron deference. Did they get it? No. And this
00:17:12 --> 00:17:15 is where Meade became really important. The Supreme
00:17:15 --> 00:17:18 Court said Chevron deference only applies when
00:17:18 --> 00:17:20 Congress has delegated authority to the agency
00:17:20 --> 00:17:23 to make rules with the force of law. So what
00:17:23 --> 00:17:25 does that mean? Generally, they're saying we're
00:17:25 --> 00:17:28 only going to give Chevron deference for notice
00:17:28 --> 00:17:30 and comment rulemaking or formal rulemaking.
00:17:30 --> 00:17:31 If they wanted to do that, we'll have to deal
00:17:31 --> 00:17:34 with formal rulemaking in another episode or
00:17:34 --> 00:17:36 formal adjudication. That's also going to come
00:17:36 --> 00:17:41 up in another episode. But these are elaborate
00:17:41 --> 00:17:43 processes like what we talked about in the notice
00:17:43 --> 00:17:46 and comment rulemaking episode. Customs ruling
00:17:46 --> 00:17:48 letters, in contrast, these are issued by field
00:17:48 --> 00:17:51 offices. There's thousands of them every year.
00:17:51 --> 00:17:53 There's no public comment, no elaborate process.
00:17:54 --> 00:17:58 So no Chevron for basic ruling letters by agencies.
00:17:58 --> 00:18:01 Yes. And here's the crucial part. The court didn't
00:18:01 --> 00:18:03 say those letters get no deference. Here's what
00:18:03 --> 00:18:06 the court wrote. There's room at least to raise
00:18:06 --> 00:18:09 a Skidmore claim here where the regulatory scheme
00:18:09 --> 00:18:11 is highly detailed and customs can bring the
00:18:11 --> 00:18:14 benefit of specialized experience to bear on
00:18:14 --> 00:18:17 this case's questions. The classification ruling
00:18:17 --> 00:18:21 may be at least a respect proportional to its
00:18:21 --> 00:18:24 power to persuade and may claim the merit of
00:18:24 --> 00:18:27 its writer's thoroughness, logic, and expertness.
00:18:27 --> 00:18:31 It's fit with prior interpretations and any other
00:18:31 --> 00:18:33 sources of weight. So Mead said that Skidmore
00:18:33 --> 00:18:36 was absolutely not dead, and now it suddenly
00:18:36 --> 00:18:39 applied to a huge category of agency actions.
00:18:39 --> 00:18:42 Like what? Like the exact stuff we've been talking
00:18:42 --> 00:18:45 about in Skidmore. Opinion letters, policy statements,
00:18:46 --> 00:18:49 guidance documents, agency manuals, interpretive
00:18:49 --> 00:18:51 rules, enforcement policies. All of the different
00:18:51 --> 00:18:54 informal ways that agencies communicate their
00:18:54 --> 00:18:56 understanding of the law. That's mostly what
00:18:56 --> 00:19:00 agencies do. Yes. It is. Notice and comment rulemaking
00:19:00 --> 00:19:04 is really slow and expensive. Agencies do far
00:19:04 --> 00:19:07 more non -formal ways of letting the public know
00:19:07 --> 00:19:10 what they think. And after Mead, all of these
00:19:10 --> 00:19:12 other methods are getting skid more deference.
00:19:13 --> 00:19:16 Okay. So then after Mead, we had two deference
00:19:16 --> 00:19:18 regimes running parallel to each other. Yeah.
00:19:19 --> 00:19:21 Here's the hierarchy we had in place. The very
00:19:21 --> 00:19:24 top level is Chevron deference. This applies
00:19:24 --> 00:19:29 to notice and comment regulations. Formal adjudications,
00:19:29 --> 00:19:32 things coming from a very high level of the agency
00:19:32 --> 00:19:35 that were illegally binding. Agency's reasonable
00:19:35 --> 00:19:37 interpretation controls in these situations.
00:19:38 --> 00:19:40 Then at the middle level, we have Skidmore deference.
00:19:41 --> 00:19:43 And this was opinion letters, guidance documents,
00:19:43 --> 00:19:46 policy statements. In these cases, the agency
00:19:46 --> 00:19:48 interpretation got weight based on its persuasiveness.
00:19:49 --> 00:19:52 And then the bottom tier, there was absolutely
00:19:52 --> 00:19:55 no deference to a position the agency had developed
00:19:55 --> 00:19:58 for the first time in litigation or other arbitrary
00:19:58 --> 00:20:00 actions the agency took. Okay, so let's get back
00:20:00 --> 00:20:04 to some real -world examples. Okay, the EEOC
00:20:04 --> 00:20:07 issues a guidance letter talking about when certain
00:20:07 --> 00:20:10 conduct constitutes workplace harassment. It's
00:20:10 --> 00:20:12 a guidance letter, so that gets Skidmore, not
00:20:12 --> 00:20:15 Chevron. The IRS issues a revenue ruling interpreting
00:20:15 --> 00:20:18 the tax code. That gets Skidmore, not Chevron.
00:20:18 --> 00:20:20 The Department of Labor sends an opinion letter
00:20:20 --> 00:20:23 about overtime rules. That gets Skidmore. The
00:20:23 --> 00:20:25 FDA issues a warning letter about drug labeling.
00:20:26 --> 00:20:28 That gets Skidmore. But if the EPA goes through
00:20:28 --> 00:20:31 full notice and comment to issue a new regulation
00:20:31 --> 00:20:34 about emissions... Then that gets Chevron. This
00:20:34 --> 00:20:37 has created an interesting strategic choice for
00:20:37 --> 00:20:40 agencies. It did. Do you want the maximum deference
00:20:40 --> 00:20:42 courts give? Great. Go through the elaborate
00:20:42 --> 00:20:45 notice and comment process. Do you prefer flexibility
00:20:45 --> 00:20:48 and speed? Then go ahead and issue guidance documents,
00:20:48 --> 00:20:50 but accept that courts are really going to scrutinize
00:20:50 --> 00:20:54 your reasoning. So the process itself was a signal
00:20:54 --> 00:20:57 about how much deference you'd get. Yes, and
00:20:57 --> 00:20:59 this created really useful incentives. Agencies
00:20:59 --> 00:21:02 couldn't just issue informal guidance and demand
00:21:02 --> 00:21:04 that courts blindly defer to them. If the agency
00:21:04 --> 00:21:06 wanted strong deference, they had to do the work
00:21:06 --> 00:21:09 to earn it. It also meant that agencies that
00:21:09 --> 00:21:11 were issuing huge numbers of documents could
00:21:11 --> 00:21:13 choose how much time they wanted to spend on
00:21:13 --> 00:21:16 each one of those documents, since the explanation
00:21:16 --> 00:21:18 that would matter when we're reviewing Skidmore
00:21:18 --> 00:21:20 is the explanation the agency gave at the time,
00:21:20 --> 00:21:23 not an explanation it came up with in court later
00:21:23 --> 00:21:27 when its actions were challenged. Okay. I'm hearing
00:21:27 --> 00:21:30 a lot of past tense in this description. Yeah.
00:21:30 --> 00:21:33 As we mentioned in the last episode, Chevron
00:21:33 --> 00:21:36 is dead. Next episode, we're going to be talking
00:21:36 --> 00:21:38 about... Loper Bright, the case that actually
00:21:38 --> 00:21:41 kills Chevron. We'll talk about what the Supreme
00:21:41 --> 00:21:44 Court said, why they decided to overturn 40 years
00:21:44 --> 00:21:47 of Chevron precedent, and the implications for
00:21:47 --> 00:21:49 the administrative state going forward. Because
00:21:49 --> 00:21:52 all of this constitutes a seismic shift for the
00:21:52 --> 00:21:55 way administrative law has operated. Yes, this
00:21:55 --> 00:21:58 is huge. Losing Chevron would have been bad if
00:21:58 --> 00:22:00 this had been the defining feature of the last
00:22:00 --> 00:22:03 four decades. But hopefully what's been shown
00:22:03 --> 00:22:05 over this episode and the last one... is that
00:22:05 --> 00:22:08 Chevron itself was just a codification of what
00:22:08 --> 00:22:10 courts had been doing for decades, since before
00:22:10 --> 00:22:13 the APA was even adopted. And now that it's gone,
00:22:13 --> 00:22:16 we are in uncharted territory. So that does it
00:22:16 --> 00:22:17 for today's episode on administrative remedies.
00:22:18 --> 00:22:21 Thank you for joining us today. Please, if you
00:22:21 --> 00:22:23 enjoy this podcast, enjoy this episode, give
00:22:23 --> 00:22:27 us a like on Spotify, iTunes, or whatever platform
00:22:27 --> 00:22:30 you're listening on. And be sure to tune in next
00:22:30 --> 00:22:32 time where we'll continue to dive into the contours
00:22:32 --> 00:22:34 of administrative law, because remember, you
00:22:34 --> 00:22:36 can't fix what you don't understand.