facts, the applicable legal standards, and the arguments made by both parties. This ruling refers to the claimant by their initials, B. Z., for anonymity purposes. B.Z. uses the pronoun “they.” The agency has conceded that the ALJ erred in his decision to deny benefits. The only question is whether it is appropriate to remand for an award of benefits or to remand for further proceedings. All three elements of the credit-as-true rule are met. The agency has not “point[ed] to anything in the record that the ALJ overlooked and explain[ed] how that evidence casts into serious doubt” B. Z.’s assertion that they are disabled. Garrison v. Colvin, 759 F.3d 995, 1022 (9th Cir. 2014). The Court has also independently reviewed the entire record and similarly finds no evidence casting serious doubt on B. Z.’s entitlement to benefits during the relevant time period. Thus, remand for award of benefits is appropriate. 1. ALJ committed legal error. Although the agency concedes legal error, a discussion of the error is needed to understand why benefits should be awarded. The ALJ improperly discounted the opinions of clinical counselor Erin Carnahan, Dr. Demetry Apostle, and Dr. Katherine Wiebe. B. Z. saw Carnahan for regular therapy sessions from June 2018 to December 2021. In a report dated March 10, 2021, Carnahan summarized her view of B. Z.’s diagnoses, symptoms, and impairments. Carnahan concluded that B. Z. met the criteria for severe depression and for anxiety—she also noted that B. Z. likely had other mental health conditions. Carnahan found that B. Z.had marked limitations in their ability to understand, remember, or apply information and to concentrate, persist or maintain pace. Carnahan also found that B. Z. had extreme limitations in their ability to interact with others and adapt or manage themselves. She found corresponding impairments in B. Z.’s ability to retain and apply information, complete tasks, maintain personal hygiene, make decisions, regulate emotions, and work a full day or an entire shift. Moreover, Carnahan noted that B. Z.’s mental health symptoms and disorders are “long-standing” and that they persist “despite [B. Z.] receiving treatment when they have had access to services.” Importantly, Carnahan’s notes from each of her 163 counseling sessions with B. Z.are also in the record, and they are consistent with Carnahan’s summary report. Apostle and Wiebe each conducted an evaluation of B. Z.during the relevant time period. Apostle conducted a two-and-a-half-hour telephonic assessment in August 2020. Apostle found moderate impairment in sensory motor and long-term memory functioning, mild to moderate impairment in attention/concentration, and mild impairment in language abilities and short-term memory function. Apostle also found severe depressive symptoms and anxiety symptoms. Beyond that, Apostle took a step back from the “largely transitory symptoms that make up clinical symptoms” to focus on and describe “those enduring and pervasive personality traits that underlie their emotional, cognitive, and interpersonal difficulties.” Apostle described B. Z. as characterized by “a lack of internal cohesion” and a tendency to “waiver unpredictably in their behaviors and relationships, seeming to respond more readily to transient inner cues than a realistic sense of others and their circumstances.” Apostle also noted that “simple tasks may demand more energy than they are capable of producing,” such that “what few efforts they can make give way to emotional outbursts under the slightest of family or social pressures.” This is consistent with Apostle’s observation that, during the assessment, B. Z. “occasionally became irritable and at one point began to yell at another individual who they felt was violating their personal space.” Ultimately, Apostle concluded that B. Z. would have serious difficulty performing a regular job for a year. Wiebe conducted a similar examination in March 2022 and came away with similar conclusions. Wiebe found severe depression and anxiety and mild or moderate impairment in the areas of overall intellectual functioning, executive functioning, attention/concentration, memory, and sensory/motor abilities. There is no need to repeat Wiebe’s analysis of B. Z.’s enduring personality traits—her conclusions were remarkably consistent with the conclusions expressed by Apostle. But Wiebe provided more detail about how she thought B. Z.’s psychological conditions would affect their ability to perform in the workplace. Wiebe determined that B. Z. had marked impairment—meaning that they are unable to perform the activity on a sustained basis in a five-day per week, eight-hour per day, normal work setting—in the following areas: understanding, remembering, and carrying out detailed instructions; getting along and working with others without excessive irritability, sensitivity, argumentativeness, or suspiciousness; accepting instructions and responding appropriately to criticism from supervisors; responding appropriately to changes in routine work setting and dealing with normal work stressors; completing a normal workday and workweek without interruptions from psychologically based symptoms. Wiebe also concluded that B. Z.had moderate impairment—meaning that they have a significant loss in their ability to perform the activity in a normal workweek—in almost every other category considered, including regular attention and punctuality; maintaining attention and concentration for a two-hour period; and remembering and carrying out very short and simple instructions. These opinions are consistent with other key facts in the record. For example, B. Z. was chronically homeless between 2015 (when their mother died) and August 2020. B. Z.’s only reported work history is as an Uber driver for about six months in 2015 and as a packager of marijuana products for about three months in 2019. B. Z. lost the packaging job due to an altercation with their manager. There are also descriptions of B. Z. from people in their life that match the conclusions of Carnahan, Apostle, and Wiebe. In November 2017, right at the beginning of the benefits period, B. Z.’s friend completed a third-party function report describing similar impairments. The friend wrote that B. Z. had serious difficulty performing simple tasks, did not follow even simple instructions, did not respond well to stress (“either completely shuts down or blows up”), and had difficulty paying attention (“sometimes no attention span”). In August 2020, around the time of Apostle’s examination, their sister completed a third-party function report, also describing similar impairments. Their sister gave examples that help contextualize the more abstract language used by the examining doctors: On a “bad day,” B. Z. would find even a simple recipe for cooking a meal to be “incomprehensible.” B. Z. did not “go anywhere regularly or irregularly,” and in the few instances that their sister remembers them going out, it “caused [them] great anxiety and [they] returned early.” When B. Z. did go out, they usually called their sister and kept her on the phone. Their sister also noted that she did not believe B. Z. drove anymore, but that when B. Z. did drive, “there were frequent accidents.” When describing B. Z.’s overall abilities, their sister said that B. Z. could not “make logical conversation or say much more than what she is feeling—often afraid and confused.” She also underscored that B. Z. can be “extremely difficult to get along with, often snapping at a person, or think if the wors[t] of them, that they are in some way against [them] from the get go.” This evidence is consistent with the opinions of Carnahan, Apostle, and Wiebe. The ALJ discarded all three opinions, pointing to a short period of time, between June 2019 and March 2020, in which B. Z.was on psychiatric medication and appears to have been largely medication compliant. There is some evidence that B. Z.’s symptoms responded well to medication. The notes from Community Psychiatry (where B. Z.was prescribed medication) reflect that B. Z. self-reported substantial improvement in symptoms while on the medication, using phrases like “60 percent improvement.” B. Z. also made a few comments about their mood being better—for example, B. Z. said in July 2019 that they were “occasionally feeling happy” and in August 2019 that their “mood was really great.” B. Z. also reported sleeping eight hours per night and being more engaged with their roommates. On the basis of these notes, and basically these notes alone, the ALJ found the opinions of Carnahan, Apostle, and Wiebe unpersuasive. The ALJ concluded that B. Z.’s mental impairment was non-severe until April 3, 2020, when they ceased using medication. Having found Carnahan, Apostle, and Wiebe’s opinions unpersuasive, the ALJ then concluded that— even after April 3, 2020— B. Z. could and can perform “a full range of work at all exertional levels,” that B. Z. “can tolerate frequent interaction with supervisors and coworkers,” and that B. Z. “can tolerate occasional public contact.” The ALJ’s residual functional capacity determination’s only real limitation is that B. Z. “should work with objects rather than people. The ALJ erred in reaching these conclusions—they are not supported by substantial evidence. First, Carnahan was regularly interacting with B. Z. before, during, and after the entire period that the claimant was on medication, and Carnahan’s notes and conclusions do not reflect a period of dramatic improvement and stability during that time. Carnahan acknowledged some improvement in symptoms. But she did not go nearly as far as the ALJ did—Carnahan never said anything like the ALJ’s conclusion that, for the time that B. Z. was taking medication, B. Z. was no longer severely limited by their mental impairments. The ALJ did not even mention Carnahan’s individual session notes. Second, the ALJ overread the Community Psychiatry notes. B. Z.described their symptoms as greatly improved, but that provides little detail about which symptoms or how it would impact B. Z.’s ability to work, and it says even less about whether the improvement brought them to a functional capacity that would render them not disabled (as opposed to improved but still disabled). That is underscored by the fact that the Community Psychiatry notes reflect ongoing changes to and increases in the B. Z.’s medication levels. Also, as B. Z.’s counsel notes, B. Z.’s condition is marked by periods of energy and enthusiasm coupled with severe downturns. B. Z. self-reporting improvement to a clinician over the course of several appointments does not justify a conclusion, flying in the face of other compelling evidence, that B. Z. was well enough to work during that time period. Third, B. Z.’s condition makes it very difficult for them to maintain medication compliance. That should be apparent from the evidence already discussed: they have difficulty with completing simple tasks, difficulty with memory, and difficulty doing things consistently. That inconsistency is reflected in the Community Psychiatry notes themselves—in February 2020, the clinician wrote that B. Z.had “returned” and had been “off medication for two months for no particular reason,” but that B. Z. requested to restart their medication. Carnahan made this point expressly: “[B. Z.] does have difficulty remaining medication-compliant though, despite stated desire to do so, due to their executive functioning deficits that impair their memory and organization, their severely dysregulating symptoms, and their social anxiety, which can impede their ability to fill and pick up prescriptions at pharmacy.” Carnahan also noted that “negative past experiences with the healthcare system” have left B. Z. “wary of psychotropic medication and distrustful of the healthcare system in general.” That distrust is in line with the paranoia and social anxiety that is a symptom of B. Z.’s mental health conditions. So even if the record justified a conclusion that B. Z.’s condition could be improved to the point of them not being disabled through medication, B. Z.’s condition also interferes with their ability to maintain medication compliance long enough for that to happen. Fourth, the ALJ ignores evidence that B. Z. has struggled with these symptoms for many years. The benefits period begins in mid-2017, so there is no need to affirmatively determine whether they were disabled before then. But to the extent that the ALJ concluded B. Z. was not severely limited until 2020, it is worth looking at the evidence from before that time. B. Z.was chronically homeless. They were unable to keep a job, worked with objects in a packaging role and lost that job quickly due to interpersonal conflict, worked as an Uber driver years before that despite their sister remembering “frequent accidents” from back when B. Z. drove. B. Z.’s friend described severe limitations in 2017, and B. Z.’s sister said that B. Z. had some of these limitations in high school and serious mental health issues in college but that their parents were able to help provide some support for them, until their parents passed away. In the face of that evidence, it is not clear how several months of some symptom improvement starting in late 2019 could support a conclusion that B. Z.was not severely limited until April 2020. The ALJ grappled with none of this. Instead, the ALJ’s opinion stated that the opinions of Carnahan, Apostle, and Wiebe were unpersuasive because they are inconsistent with the Community Psychiatry notes. The ALJ also found the sister and friend’s third-party function reports only persuasive as far as they are consistent with the residual function capacity that the ALJ had already laid out—it is not clear how evidence can be persuasive if it is only considered to the extent that it supports a conclusion already reached on the basis of other evidence. It is worth briefly addressing the few other reasons the ALJ briefly gave, at various points in the opinion, for discounting evidence of B. Z.’s disability. The ALJ pointed to various tasks that B. Z. acknowledged being able to do: house cleaning, yoga, and their brief stint as an Uber driver years before the relevant time period. But this just underscores the ALJ’s cherry-picking of the record. B. Z.’s ability to handle a few tasks for various periods of time does not meaningfully contradict the evidence of their marked limitations. And, again, the ALJ overread evidence of B. Z.’s capacity. Take the yoga practice as an example. B. Z. sometimes does yoga to manage their symptoms, and to prevent themselves from engaging in self-harm, saying that if they practice yoga for many hours a day, then sometimes their anxiety becomes manageable enough that they can complete basic tasks for themselves. The ALJ took that as evidence that B. Z. was not severely limited, which makes little sense. Cf. Samers v. Commissioner of Social Security, 357 F. Supp. 3d 1005 (N.D. Cal. 2019). Next, consider the house cleaning. There are some notes that reflect B. Z. telling a clinician that they have been able to clean up around their own home. And there is one note that reflects B. Z. telling a clinician in August 2020 that they have been working as a house cleaner. But the ALJ acknowledged that there was no other evidence of B.Z.’s work as a house cleaner, no income reported from it, and no mention of it in B.Z.’s testimony. Yet the ALJ based his residual functional capacity determination in substantial part on this vague mention in the record of house cleaning. Additionally, the ALJ repeatedly invokes mentions by clinicians or examiners that B. Z. was cooperative or pleasant. The ALJ seems to draw from that a conclusion that B. Z. does not, in fact, have the difficulties interacting with others that leap out from the rest of the record. But the opinions finding severe limitations did not say that B. Z.could never be cooperative or pleasant—they noted fluctuations in B. Z.’s condition, that B. Z. would struggle with authority figures especially, and that B. Z. would have serious difficulties interacting with people in their life because of distrust, paranoia, anxiety, irritability, anger, and fear. There is no true conflict between those pieces of evidence. That the ALJ discarded the opinions of Carnahan, Apostle, and Wiebe is especially striking given the three opinions that the ALJ did credit. The ALJ found the opinions of Dr. Aparna Dixit, Dr. Norman Zukowsky, and Dr. S. Nair persuasive. Dixit conducted her examination in December 2016, prior to the claimed benefits period and prior to much of the rest of the evidence. Dixit did not look at B. Z.’s treatment record. The ALJ found Dixit’s opinion partially persuasive because it was consistent with other therapy notes from 2016. Then the ALJ’s opinion simply says that symptoms worsened after April 3, 2020—skipping about three- and-a half contested years. Zukowsky and Nair were both state agency consultants. Zukowksy reviewed the record in January 2018 and concluded that B. Z. had no severe mental impairment; as the ALJ’s opinion notes, this was based on Dixit’s report and the fact that B. Z. was not then undergoing mental health treatment. Similarly, in April 2020, Nair reviewed the evidence and agreed with Zukowsky. Both reports were essentially box-checking exercises, with little explanation or analysis. The only examination conducted, in all of that, is the brief examination by Dixit almost a year prior to the relevant time period. These three opinions are entitled to substantially less weight than those of Carnahan, Apostle, and Wiebe: they were perfunctory in explanation, reliant on a single examination from before the benefits period, and Dixit and Zukowsky reached their opinions before the development of much of the evidence that is now before the agency and the Court. The opinions of Dixit, Zukowsky, and Nair are also inconsistent with the evidence of B.Z.’s disability and serious impairments that pervades the record, including the opinions of Apostle and Wiebe. Moreover, they are inconsistent with the detailed opinion and treatment notes of Carnahan, who had an ongoing treatment relationship with B.Z. that involved frequent, documented appointments over a long period of time. See Woods v. Kijakazi, 32 F.4th 785, 792 (9th Cir. 2022). So, these consultants’ opinions do not provide substantial evidence for rejecting the opinions of Carnahan, Apostle, and Wiebe, all of whom found severe limitations. Finally, the ALJ gave the same poor reasons for discrediting B. Z.’s own testimony: the brief period of medication compliance, to the opinions of Zukowsky and Nair, and B. Z.’s occasional ability to perform yoga and clean their home. B. Z.’s testimony is generally consistent with the opinions of Carnahan, Wiebe, and Apostle, consistent with the third-party function reports, and supports a disability finding. In all, the ALJ has clearly erred. The ALJ read the record selectively, cherry-picking evidence to reach her conclusion. The residual functional capacity determination is not based on substantial evidence. The ALJ improperly discredited the opinions of examining and treating doctors and B. Z.’s own testimony. Perhaps this is why the agency made no effort to defend the ALJ’s opinion in this case. The first step of the credit-as-true rule, that the “ALJ has failed to provide legally sufficient reasons for rejecting evidence, whether claimant testimony or medical opinion,” is clearly met in this case. Treichler v. Commissioner of Social Security Admin., 775 F.3d 1090, 1100–01 (9th Cir. 2014). 2. Fully developed record. This case has been pending for over six years. The ALJ held multiple hearings before issuing the decision on appeal. The record includes assessments by at least seven professionals regarding B. Z.’s disability and functional limitations. More than half of those professionals either treated or examined B. Z. The record also contains hundreds of pages of treatment notes from therapy sessions with B. Z., multiple statements by B. Z.across the period, and statements from B. Z.’s friend and sister about how their disability impacts their life. The agency has not identified a single essential factual issue that further proceedings would shed additional light on. Indeed, the agency’s motion to remand cherry-picks the record just as the ALJ’s opinion did. The agency points to the opinions of Dixit, Zukowksy, and Nair, to the Community Psychiatry improvement notes, to B. Z.’s ability to do yoga, meditate, cook, and clean, and to a couple of instances in which people who interacted with B. Z. described them as “amicable” or “appropriately dressed.” But for the reasons described above, these are not true conflicts or ambiguities. Taking all these pieces of evidence into account, the record still compels a conclusion that B. Z.is disabled, and that they were disabled throughout the relevant period. Moreover, the agency has not explained how additional proceedings would be useful. To the extent there are questions about B. Z.’s capacity during the time that they were medication compliant, the record is fully developed. To the extent that there are questions about the period after medication compliance, the record is already voluminous. The agency makes no suggestion of what further proceedings would reveal or what facts have yet to be discussed. And Ninth Circuit “precedent and the objectives of the credit-as-true rule foreclose the argument that a remand for the purpose of allowing the ALJ to have a mulligan qualifies as a remand for a ‘useful purpose.’” Garrison v. Colvin, 759 F.3d 995, 1021 (9th Cir. 2014). 3. Record requires a conclusion of disability. There is extensive evidence of B. Z.’s mental impairments. If the improperly discredited evidence were credited as true, the ALJ would have no choice but to conclude that B. Z. is disabled and that they were disabled for the relevant period. Garrison, 759 F.3d at 1022. This case represents one of those “rare circumstances” in which remand for award of benefits, rather than remand for “additional investigation or explanation” is appropriate. Dominguez v. Colvin, 808 F.3d 403, 407 (9th. Cir. 2015). 4. Other source of “serious doubt.” Even where all three elements of the credit-as-true rule are satisfied, courts may still remand for further proceedings if the record as a whole creates “serious doubt” that the claimant is disabled. Burrell v. Colvin, 775 F.3d 1133, 1141 (9th Cir. 2014). The agency’s primary argument for remand—made in roughly one page of its motion—is that B. Z.’s drug use renders them ineligible for benefits. That is because, where a claimant’s drug use is “material” to the determination of disability, there can be no disability determination. More specifically, substance use is material if B. Z. would not meet the definition of disability were they not using drugs or alcohol. The agency claims that further proceedings are needed to answer this question. But again, the agency has cherrypicked the record in an effort to stave off the award of benefits. B. Z. has a history of substance use. They have used marijuana regularly for over a decade, and in 2020 they began micro-dosing psilocybin. But nothing in the record suggests that B. Z. would not meet the definition of disability absent their substance use. In fact, the evidence in the record points in the other direction. The experts who examined and treated B. Z. all expressed versions of the same conclusion: that B. Z.’s substance use is co-occurring with their mental health conditions, and that B. Z. is using substances to cope with their symptoms. For example, Apostle noted that B. Z.’s regular marijuana use began back in 2012 as part of B. Z.’s effort to sleep and reduce their irritability. Apostle also said that “their past and current use of substances has likely been primarily to self-medicate their psychiatric and personality disorder symptoms.” Wiebe noted the same about B. Z.’s micro-dosing with mushrooms. Both Apostle and Wiebe stated explicitly that B. Z. would have difficulty maintaining a regular job for a year, even if they stopped using substances entirely—they based this conclusion on B. Z.’s long-term cognitive, psychiatric, and personality disorder symptoms. These conclusions are supported by the third-party function reports: B. Z.’s sister discusses some symptoms that occurred before the first mention of regular drug use. And B. Z.’s friend describes serious symptoms in 2017, before any use of mushrooms. The agency points to no evidence that B. Z. would not be disabled if they stopped using substances. The agency tries to describe it as an “open question” by parsing Apostle and Wiebe’s language very closely—those doctors said B. Z. would still have “difficulty” absent any drug use, but they did not explicitly state that B. Z. would still have “marked limitations.” But the Court must consider the entire record to see if there is serious doubt. That includes the context in which Apostle and Wiebe used the word “difficulty,” the severe limitations that they both found B. Z. to have, the evidence of the long-term nature of B. Z.’s impairments, the description of the substance use as co-occurring, and the conclusion that the substance use began and continues on as a coping mechanism for the underlying conditions. Given all of that, the evidence creates no serious doubt that B. Z. would still be disabled absent the substance use. And, again, it is not apparent what use further proceedings would serve to cast light on this question. The ALJ held a second hearing before issuing his opinion, and that hearing was dedicated to the question of whether B. Z.’s substance use was material. A medical expert, Dr. Chukwuemeka Efobi, who reviewed the record testified at that hearing that he could not offer an opinion about whether B. Z. would be disabled absent the substance use because he could not identify a totally clean period. Efobi did testify some about what a hypothetical person with similar functional limitations might experience if they stopped using substances. But that fails to capture all the details and nuance of B. Z.’s condition, history, and substance use that both Apostle and Wiebe considered. There is no true conflict between Efobi’s testimony about the hypothetical person and the evidence about B. Z. Further proceedings would not remove the obstacle to reaching a perfect answer to the materiality question: B. Z. has been suffering with these conditions, and self-medicating with substances, for such a long time that an expert cannot say for certain what would happen if the substance use ceased. But two examining doctors concluded that the substance use is not the cause of the mental health conditions that compel the disability finding. TK The case is remanded to the ALJ for a direct award of benefits.

IT IS SO ORDERED.

Dated: July 2, 2024

VINCE CHHABRIA

United States District Judge