Michael Hedlund v. the Educational Resources Inst

OPINION TASHIMA, Circuit Judge: Michael Hedlund is a law school graduate who asserts that he cannot pay off his student loans. After filing for bankruptcy, he sought a discharge of his student loans under 11 U.S.C. § 523(a)(8). The bankruptcy court granted a partial discharge, but, on appeal, the district court reinstated the student loan- debt in full as non-dis-chargeable. Specifically, the district court ruled that Hedlund had not acted in good faith, which is one of three prerequisites for relief under § 523(a)(8). We hold that the district court erred in reviewing the bankruptcy court’s good faith finding de novo. In a § 523(a)(8) proceeding, the good faith finding should be reviewed for clear error. Under the proper standard of review, we affirm the bankruptcy court’s ruling. I. Hedlund was thirty-three years old at the time of the bankruptcy proceedings. He had earned a bachelor’s degree in business administration from the University of Oregon and a law degree from Willamette Law School. Hedlund financed his education with Stafford loans, which were held in part by The Education Resources Institute (“TERI”) and in part by the Pennsylvania Higher Education Assistance Agency (“PHEAA”). After law school, Hedlund took a bar preparation course for the Oregon bar and then took the bar examination in July 1997. While awaiting the results, he worked as an intern for the Klamath *850 County District Attorney. He failed the exam, re-sat in February 1998, and failed again. He lost his job at the District Attorney’s office for failure to pass the bar exam on his second try. He then obtained full-time employment as a Juvenile Counselor with the Klamath County Juvenile Department. While employed full time as a Juvenile Counselor, he enrolled in another bar preparation course and took two months off to study. En route to the exam, however, when he stopped for coffee, he inadvertently locked his keys in his car. He missed the exam. Hedlund married in 2000 and became a father in 2001. Hedlund’s loans went into repayment in January 1999, 1 while he was working as an intern at the District Attorney’s office. He owed PHEAA over $85,000, on which the monthly payments exceeded $800. Because he was making only $10 per hour, he sought and obtained various hardship for-bearances. After the extensions ended and in an effort to reduce his monthly payments, Hedlund applied to consolidate his loans. When he later called to verify the status of his consolidation application, he was told that it had never been received and that, because he was now in default, he was ineligible for consolidation. Hedlund then researched his potential eligibility for the Income Contingent Repayment Plan (“ICRP”). 2 Based on his online research — and on the loan provider’s representation that he was ineligible for consolidation due to the default — Hedlund concluded that he would not qualify for the ICRP. In September 1999, Hedlund received a $5,000 inheritance. He paid $954.72 to PHEAA, and the rest went to other creditors. Still unable to make his monthly payments, Hedlund tried to negotiate a less onerous payment schedule. According to Hedlund, PHEAA offered two options: (1) pay $10,000 up front, then $1,300 a month for ten months, and then an adjusted monthly payment; or (2) pay a lump sum of approximately $80,000. Neither option was feasible for Hedlund, but he did offer to make a $5,000 payment— which he would have borrowed from his parents — in exchange for a more lenient payment schedule. PHEAA declined Hed-lund’s offer. 3 PHEAA began garnishing Hedlund’s wages in January 2002 at the rate of about $250 per month. These garnishments continued uncontested until May 2003 and amounted to $4,272.52. At that time, Hed-lund’s other student loan creditor, TERI, obtained a collection action judgment against Hedlund and garnished $1,100 directly from Hedlund’s bank account. On May 7, 2003, Hedlund filed a Chapter 7 bankruptcy petition. On June 16, 2003, Hedlund commenced an adversary proceeding against PHEAA and TERI, seeking partial discharge of his loans under 11 U.S.C. § 523(a)(8). He settled with TERI before trial, agreeing to pay down $17,718.15 at a rate of $50 per month. In other pretrial negotiations, PHEAA offered three potential repayment plans “if the Loans [were] determined not *851 to be dischargeable.All three options called for payment of the entire loan balance over the course of 30 years. The first option called for monthly payments of approximately $420; the remaining options began with monthly payments of $307 and rose to $430 and $440 respectively. Hed-lund did not, and has not, pursued any of these options. After trial, the bankruptcy court granted a partial discharge of all but $30,000 of the PHEAA debt. On appeal, the Bankruptcy Appellate Panel (“BAP”) reversed and reinstated the debt in its entirety. Hedlund appealed to this Court, and we vacated the BAP decision and remanded for further proceedings. We held that the bankruptcy court failed to consider all of the evidence and properly to apply the three factors from Brunner v. New York State Higher Education Services Corp., 831 F.2d 395 (2d Cir.1987). 4 See Hedlund v. Penn. Higher Educ. Assistance Agency (In re Hedlund), 368 Fed.Appx. 819 (9th Cir.2010). On remand to the bankruptcy court, the parties agreed to proceed on the original 2003 record and the case was reargued and submitted. After the case was submitted for decision, however, the originally assigned judge, Judge Albert Radcliffe, passed away, and the ease was reassigned to Judge Philip Brandt. Judge Brandt ruled in Hedlund’s favor and discharged all but $32,080 of his debt to PHEAA. Applying the three-factor Brunner test, Judge Brandt found that: (1) Hedlund could not have maintained a minimal standard of living, if required to repay the full loans; (2) “additional circumstances” indicated that Hedlund’s inability to repay his loans would persist into the future; and (3) Hed-lund had made good faith efforts to repay his loans. PHEAA appealed and the district court reversed, finding no error under the first two prongs, but concluding that the bankruptcy court’s good faith ruling was erroneous. Accordingly, the district court reinstated the entirety of the PHEAA loan. Hedlund timely appeals. We have jurisdiction under 28 U.S.C. § 158(d), and we reverse the district court. II. A. Student loan obligations are presumptively nondischargeable in bankruptcy absent a showing of “undue hardship.” 11 U.S.C. § 523(a)(8). To determine if a debtor has shown undue hardship, we follow the three-part test from Brunner. See In re Pena, 155 F.3d at 1111-12. Under Brunner, the debtor must prove that: (1) he cannot maintain, based on current income and expenses, a “minimal” standard of living for himself and his dependents if required to repay the loans; (2) additional circumstances exist indicating that this state of affairs is likely to persist for a significant portion of the repayment period; and (3) the debtor has made good faith efforts to repay the loans. Educ. Credit Mgmt. Corp. v. Mason (In re Mason), 464 F.3d 878, 882 (9th Cir.2006). “[T]he burden of proving undue hardship is on the debtor, and the debtor must prove all three elements before discharge can be granted.” Rifino v. United States (In re Rifino), 245 F.3d 1083, 1087-88 (9th Cir.2001). *852 This appeal concerns only the good faith prong of Brunner. The bankruptcy court ruled in Hedlund’s favor on all three prongs, and the district court found error only with respect to the last prong, good faith. 5 Before addressing the proper standard of review, we begin with a summary of the bankruptcy court and district court rulings. B. “Good faith is measured by the debtor’s efforts to obtain employment, maximize income, and minimize expenses.” Penn. Higher Educ. Assistance Agency v. Birrane (In re Birrane), 287 B.R. 490, 499 (9th Cir. BAP 2002) (internal quotation marks and citation omitted). “Courts will also consider a debtor’s effort — or lack thereof — to negotiate a repayment plan, although a history of making or not making payments is, by itself, not dispositive.” In re Mason, 464 F.3d at 884 (internal quotation marks and citations omitted). The bankruptcy court considered each of these factors. 1. Efforts to obtain employment, maximize income, and minimize expenses. The bankruptcy court found that Hed-lund was “well-placed for his skills” and that there were no higher paying jobs available to him in the Klamath Falls area. It also noted that Hedlund had unsuccessfully applied for two higher-paying jobs. Finally, the court cited expert testimony showing that, although higher paying jobs might be available outside of Klamath Falls, the potential increase in salary would be offset by increased living expenses. Noting that Hedlund had tried three times to take the bar exam, the bankruptcy court also found that Hedlund’s failure to pass was not “within his control.” In any event, the court found no evidence suggesting that Hedlund could make a higher wage as a licensed attorney. The court also rejected PHEAA’s argument that Hedlund should seek an additional part time job, although the court did find that his wife could be expected to work three days per week rather than one. Thus, the court found that Hedlund had sufficiently maximized his income. The bankruptcy court then reviewed Hedlund’s personal budget, and concluded that certain expenses exceeded what was reasonably necessary to maintain a minimal standard of living. 6 Specifically, the court found that Hedlund’s clothing, recreation (including cable and internet), and “miscellaneous” budgets (including childcare and haircuts) could all be reduced. By contrast, the court found no fault with Hedlund’s budget for two cell phones (given that the Hedlunds had a small child) and for the lease of a second car (given that the Hedlund’s other car was older and unreliable). Taken together, the court found that the failure fully to minimize *853 expenses did not “tip the balance away from a good faith finding” because Hed-lund and his family “have always lived frugally.” 2. Efforts to negotiate a repayment plan; history of payments; and timing of the attempt to discharge the loan. Under these factors, the court noted with approval that Hedlund had waited four years before filing for bankruptcy and that he had, in that time, made a voluntary payment of approximately $950. The court also credited Hedlund for “enduring],” without challenge, sixteen and a half months of wage garnishments. With regard to alternative repayment plans, the court found that Hedlund had made adequate efforts to pursue one. The court noted that Hedlund had sought to consolidate loans, but that the lender had lost his application. The court also took into account Hedlund’s offer to make an immediate payment of $5,000 in exchange for more lenient repayment terms. The court did not fault Hedlund for failing to apply for the ICRP. Hedlund had investigated the ICRP option online but concluded that he was not eligible because he was in default. Moreover, PHEAA had conceded that payments under the ICRP would have been more per month than the three options PHEAA had offered just before trial. Finally, the court found that Hedlund was justified in rejecting the three repayment options PHEAA had offered. The court stressed that the three options all had monthly payments over $300, which was more than Hedlund could afford without undue hardship. 7 The court also noted that each option called for a thirty-year plan, and thus held that Hed-lund was justified in refusing to obligate himself “into his mid-60s” when his children would likely be seeking to attend college. Considering all of this evidence together, the bankruptcy court found that Hed-lund’s situation was not “self-inflicted” and that he had carried his burden of showing good faith. C. The district court reviewed the good faith ruling de novo and reversed. Although the district court agreed that Hedlund had made sufficient efforts to obtain employment, it concluded that Hedlund “ha[d] not used his best efforts to maximize his income or minimize his expenses.” Hedlund v. Educ. Res. Inst., Inc., 468 B.R. 901, 914 (D.Or.2012). That conclusion was apparently based on the court’s findings under prong one that some of Hedlund’s expenses were “immoderate.” Id. at 910. The court then found Hedlund’s lack of effort in negotiating a repayment plan “even more vexatious.” Id. at 910, 915. In that regard, the court felt that Hedlund was “less than diligent” in exploring the ICRP option, and it viewed Hedlund’s $5,000 payment offer as unrealistic. Id. at 915. The court also faulted Hedlund for rejecting the three pre-trial repayment plans offered by PHEAA. Id. Finally, the court observed that Hedlund and his wife had chosen to live as a single-income family, “a lifestyle that few today can afford.” Id. at 916. Thus, the district court reversed the bankruptcy court’s good faith finding. III. A. “Because this court is in as good a position as the district court to review the *854 findings of the bankruptcy court, it independently reviews the bankruptcy court’s decision.” Ragsdale v. Haller, 780 F.2d 794, 795 (9th Cir.1986). As a threshold matter, we must resolve a dispute over the proper standard of review of the bankruptcy court rulings. The district court reviewed the good faith determination de novo, but Hedlund contends that it should have applied clear error review. We agree with Hedlund. Although we review “the bankruptcy court’s interpretation of the Bankruptcy Code de novo and its factual findings for clear error,” Miller v. Cardinale (In re DeVille), 361 F.3d 539, 547 (9th Cir.2004) (internal quotation marks and citations omitted), we have not expressly stated which standard applies to the good faith prong of Brunner. Nevertheless, we have consistently reviewed the good faith prong for clear error. See In re Mason, 464 F.3d at 885 (“[T]he bankruptcy court clearly erred in finding that Mason demonstrated good faith efforts to repay his loans.”); In re Pena, 155 F.3d at 1114 (“[T]he bankruptcy court did not clearly err in finding that the Penas exhibited good faith in attempting to pay back the student loans.”). 8 We have also applied the deferential clear error standard of review to good faith inquiries in other settings. See Figter Ltd. v. Teachers Ins. & Annuity Assoc. of Am. (In re Figter Ltd.), 118 F.3d 635, 638 (9th Cir.1997) (collecting cases and noting that because good faith is “an essentially factual inquiry” this Court has “in various contexts, declared that [it] will review good faith determinations for clear error” (internal quotation marks and citations omitted)). Accordingly, we now confirm that a good faith finding under Brunner should be reviewed for clear error. 9 This directive does not preclude a reviewing court from correcting errors of law that may arise in the midst of a good faith analysis. For example, in In re Birrane the BAP held that “the bankruptcy court erred as a matter of law in finding that Birrane met the good faith prong.” 287 B.R. at 500. The error in that case was one of law, reviewed de novo, because the bankruptcy court had failed to consider all factors relevant to good faith. Instead, it had found good faith based solely on the evidence of voluntary payments. Id. at 499. Thus, although good faith is primarily a question of fact reviewed for clear error, it can encompass questions of law that must be reviewed de novo. 10 *855 We now consider the good faith ruling in question. B. As an initial matter, the bankruptcy court properly applied all three Brunner prongs, and it considered the various factors that are relevant to good faith. Thus, its ruling withstands our de novo review of the legal questions involved. What remains are the bankruptcy court’s factual findings. As discussed below, those were not clearly erroneous. 11 There was considerable evidence showing that Hedlund had maximized his income, and the court properly declined to attribute Mrs. Hedlund’s underemployment to Hedlund’s bad faith. Although Hedlund had not fully minimized his expenses, the court permissibly interpreted the excess expenses as marginal. And although we might have viewed certain expenses more skeptically — such as the new car lease and the two cell phones- — -the court’s view of the expenses was not clearly erroneous. The record regarding efforts to negotiate and to make voluntary payments is less favorable to Hedlund. Although he did submit a consolidation application, his efforts thereafter were minimal. His offer to pay $5,000 in exchange for a more lenient plan was at best unrealistic, and his research into ICRP eligibility could have been more searching. Hedlund has also declined to pursue the three revised repayment plans that PHEAA offered just before trial. Finally, in the four years prior to bankruptcy, Hedlund made only a single voluntary payment of approximately $950. 12 Although this evidence could be interpreted to support a finding of lack of good faith, it was not so strong as to demand such a finding. Indeed, the evidence of Hedlund’s good faith is more substantial than in the two primary cases relied upon by PHEAA. In In re Birrane, the debtor had “failed to take any steps towards renegotiating a repayment schedule under the ICRP program.” 287 B.R. at 500. In contrast, Hedlund at least made an effort to research his eligibility. Moreover, unlike Hedlund, the debtor in In re Birrane had maintained only part-time employment and thus had failed to maximize her income. Id. at 499-500. In In re Mason, the debtor had failed to pursue the ICRP option “with diligence.” 464 F.3d at 885. Although the same might be said of Hedlund’s efforts, other factors present in In re Mason are not present here. Specifically, Mason had not pursued full time employment and had only taken and failed the bar exam once. In contrast, Hedlund has maximized employment, made three attempts at the bar exam and, in any event, submitted evidence that a law license would not materially improve his financial situation. Also weighing in Hedlund’s favor is the fact that he waited four years from the beginning of his re *856 payment obligations, during which period he was subject to wage garnishments, before filing for bankruptcy. See Brunner, 831 F.2d at 397 (finding bad faith in part because debtor filed for discharge one month after first payment date). In sum, even though some might disagree with the bankruptcy court’s good faith finding, it was not clearly erroneous. The court relied on substantial evidence in the record, and its factual inferences were permissible. IV. The bankruptcy court’s good faith finding was not clearly erroneous; we, therefore, reverse the district court’s contrary holding. We remand to the district court with instructions to reinstate the partial discharge ordered by the bankruptcy court. REVERSED and REMANDED with directions. Footnotes [1] Student loan recipients typically are not; required to begin making payments to pay back their loans until some point after the borrower has completed his or her educational program. [2] Under the ICRP, the debtor pays the lesser of: (1) payments based on a 12-year amortization derived by application of an annually adjusted percentage of the debtor’s adjusted gross income; or (2) 20% of the debtor’s annually adjusted discretionary income, defined as adjusted gross income less applicable federal poverty guidelines. See 34 C.F.R. § 685.209. [3] The record is unclear on when precisely 3 these negotiations took place. [4] We adopted the Brunner test in United Student Aid Funds, Inc. v. Pena (In re Pena), 155 F.3d 1108, 1111-12 (9th Cir.1998). [5] PHEAA makes a passing argument that the bankruptcy court’s rulings on the first two Brunner prongs were erroneous. Without making any cogent argument, PHEAA simply asks us to rely on the BAP’s 2004 reversal of the bankruptcy court during the first round of litigation, which reviewed an entirely different bankruptcy court order and was vacated by this Court’s remand order. Accordingly, PHEAA has waived these arguments. See Leer v. Murphy, 844 F.2d 628, 634 (9th Cir.1988) ("Issues raised in a brief which are not supported by argument are deemed abandoned.”). [6] The court's analysis of Hedlund’s expenses relied in part on its findings under the first Brunner prong. There, the court had ruled that the majority of Hedlund’s expenses were reasonably necessary, and thus that he could not maintain a minimal standard of living if required to repay the loans in full. [7] $300 per month exceeded the bankruptcy court's finding, on the first Brunner prong, of how much income Hedlund could devote to student loan payments. [8] At least one other circuit has expressly applied clear error to the § 523 good faith inquiry. See Krieger v. Educ. Credit Mgt. Corp., 713 F.3d 882, 884 (7th Cir.2013) ("[The good faith] standard combines a state of mind (a fact) with a legal characterization (a mixed question of law and fact). Findings of fact must stand unless clearly erroneous, and... .mixed questions likewise are treated as factual in nature.”). [9] We also note that under our established standard of review of bankruptcy court rulings, after review by the district court or by the BAP, see Ragsdale, 780 F.2d at 795, the district court’s ruling on the good faith issue would not have been entitled to any deference, but would be reviewed “independently.” [10] Similarly, clear error review is consistent with our observation that "undue hardship requires a determination of the legal effect of the bankruptcy court’s findings regarding the student’s circumstances, a question of law which we review de novo." In re Mason, 464 F.3d at 881 (internal quotation marks omitted). This is simply another way of stating that we review the application of Brunner de novo, i.e., whether the bankruptcy court properly applied the three-prong test. For instance, if the bankruptcy court had granted a discharge based only on the first two prongs of Brunner, we would reverse under de novo review. [11] Clear error applies even though Judge Brandt based his findings on the trial transcript and other documentary evidence. See Anderson v. City of Bessemer City, N.C., 470 U.S. 564, 574, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985) ("Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous. This is so even when the [trial] court's findings do not rest on credibility determinations, but are based instead on physical or documentary evidence or inferences from other facts.” (internal citations omitted)); see also Fed.R.Civ.P. 52(a)(6). [12] We need not determine if, as the bankruptcy court ruled, uncontested garnishments demonstrate good faith. Even if this view of the evidence was improper, the court's interpretation of the record as a whole was not clearly erroneous.