Emory Bankruptcy Developments Journal

Student-Loan Discharge—An Empirical Study of the Undue Hardship Provision of § 523(a)(8) imder Appellate Review
Ryan Freeman Notes and Comments Editor, Emory Bankruptcy Developments Journal; J.D. Candidate, Emory University School of Law (2014); B.A., University of Tennessee, Knoxville; Recipient, Keith J. Shapiro Excellence in Consumer Bankruptcy Writing Award. This comment benefited significantly from the suggestions and guidance of Rafael I. Pardo, Robert T. Thompson Professor of Law at Emory School of Law. Professor Pardo has blazed the academic path in this field, and his willingness to aid me in my study was integral. Without his help, the statistical findings of this study could not have been fleshed out so fully. Thanks also belong to Alexandra Ragan, whose wonderful feedback and ideas helped hone the content of this comment. Finally, I owe a special thanks to my wife, Katie, and my colleagues, Joel M. Langdon, Micah J. Revell, and Evan T. Hoole, for their unfailing grace in listening to me talk on and on about each and every one of the 289 cases I read for this study.

Abstract

Prior to the enactment of the Bankruptcy Code, student-loan debtors could receive an automatic discharge of their debts in bankruptcy. Now, they cannot. Since the Code’s enactment, Congress has pursued progressively harsher standards, continually narrowing the scope of when a student-loan debtor could obtain discharge. Following the enactment of the Bankruptcy Abuse Prevention and Consumer Protection Act in 2005, student-loan debtors now encounter the toughest obstacles to discharge they have ever faced. By extending the protection of the discharge exception of 11 U.S.C. § 523(a)(8) to private lenders, Congress effectively placed all students who take out loans to pay for their education at the mercy of a harsh system whose narrow exceptions for discharge force debtors to prove that they face a “certainty of hopelessness” in their future.

The harshness of this system is well documented by empirical studies analyzing the results of bankruptcy courts applying the undue hardship provision in § 523(a)(8). These studies paint a portrait of inconsistency and subjectivity across the many judicial districts. Attempted application of the undefined term “undue hardship” has resulted in multiple judicially-created tests, the most predominant of which is the Brunner test, requiring a finding of non-dischargeability if the debtor fails any of the test’s three prongs.

This Comment presents findings from an empirical study of ten years of bankruptcy appellate decisions dealing with the undue hardship provision, an area yet unexplored. Its findings reveal an enormous inequity in the treatment of debtors and creditors in these cases. These findings suggest a reconsideration of the current approach of the bankruptcy system toward student-loan debt and whether the judicially-created tests have narrowed this exception beyond what the Code intended. Furthermore, it recommends careful assessment for practitioners considering an appeal on behalf of a debtor of an unfavorable bankruptcy court decision.

Introduction

Thomas Edison once said of his electric light, “None of my inventions has cost me as much time, labor and study.” 1The Quotable Edison 27 (Michele Wehrwein Albion ed., 2011). When Edison developed his light bulb, the idea of electric light was nothing new. In fact, the idea had become prevalent in scientific communities more than seventy-five years prior to Edison’s invention. 2See Martin V. Melosi, Thomas A. Edison and the Modernization of America 65–66 (Mark C. Carnes ed., 2008). The first public demonstration of electric light was made in 1808 when Sir Humphrey Davy unveiled his arc lighting invention, which had little commercial application until the late 1870s. See id. at 65. Furthermore, Davy produced incandescence in 1802, but hundreds of experiments following this finding failed to produce the type of lasting filament that Edison achieved. See id. at 66. The effects of electricity had been studied for years, but as the new age of science was ushered into being, 3See id. at 65–66. Edison found purpose and commercial application in the harnessing of electric light through a long-lasting filament. 4See id. at 70. He looked behind the veil of data and previous experiments and discovered a way to make light a practical option for millions across America. Speaking about his methods, Edison stated:

It has been said of me that my methods are empirical. That is true . . . . So, when I am after a chemical result that I have in mind I may make hundreds or thousands of experiments out of which there may be one that promises results in the right direction. This I follow up to its legitimate conclusion, discarding the others, and usually get what I am after. There is no doubt about this being empirical; but when it comes to problems of a mechanical nature, I want to tell you that all I’ve ever tacked and solved have been by hard, logical thinking. 5The Quotable Edison, supra note 1, at 10 .

This response showed the time and effort it took Edison to really understand and apply his knowledge to craft an invention that would revolutionize the world.

The herculean effort Edison applied to the light bulb is not so different from what is required to fully understand the Bankruptcy Code (“Code”) and its application by courts. Indeed, Edison’s methods of empirical investigation and hard, logical thinking should be applied to study and decipher the effects of the Code’s application. Much like electric light, many of the effects of the Code’s application have already been studied and reported. Specifically, researchers and authors have produced significant data and results as it relates to how bankruptcy courts have applied the “undue hardship” provision from § 523(a)(8)’s exception to discharge. 6See Jason Iuliano, An Empirical Assessment of Student Loan Discharges and the Undue Hardship Standard, 86 Am. Bankr. L.J. 495 (2012); Rafael I. Pardo, Illness and Inability to Repay: The Role of Debtor Health in the Discharge of Educational Debt, 35 Fla. St. U. L. Rev. 505 (2008); Rafael I. Pardo & Michelle R. Lacey, The Real Student-Loan Scandal: Undue Hardship Discharge Litigation, 83 Am. Bankr. L.J. 179 (2009) [hereinafter The Real Student-Loan Scandal]; Rafael I. Pardo & Michelle R. Lacey, Undue Hardship in the Bankruptcy Courts: An Empirical Assessment of the Discharge of Educational Debt, 74 U. Cin. L. Rev. 405 (2005) [hereinafter Undue Hardship]. The challenge moving forward is to find, much like Edison, a way to use and expand on that data to provide practitioners and judges guidance and information that adds value to their decision-making processes in this area of the law. This Comment, through an empirical study of ten years of appellate decisions governing undue hardship determinations, seeks to pull back the veil and provide an analysis of those courts whose precedents have set the tone for student-loan discharge.

A look at recent news articles from around the United States reveals a predominantly negative perspective on the financial outlook of a student-loan debtor. 7See, e.g., Peter Coy, Student Loans: Debt for Life, Bloomberg Businessweek (Sept. 18, 2012), http://www.businessweek.com/articles/2012-09-06/student-loans-debt-for-life#p1; Ron Lieber, Last Plea on School Loans: Proving a Hopeless Future, N.Y. Times (Aug. 31, 2012), http://www.nytimes.com/2012/09/01/business/shedding-student-loans-in-bankruptcy-is-an-uphill-battle.html?pagewanted=all; Steve Rhode, American Households Sinking in Student Loan Debt Says Pew Research Center, Huffington Post (Oct. 7, 2012), http://www.huffingtonpost.com/steve-rhode/american-households-sinki_b_1933143.html. Noting that student-loan debt has surpassed the $1 trillion mark, one writer for the Chicago Sun-Times speculated that “student loans are about to become a larger financial crisis than the mortgage disaster.” 8Terry Savage, Terry Savage: Federal Student Loans Next Big Crisis, Chicago Sun-Times, (Oct. 13, 2012), http://www.suntimes.com/business/savage/15755472-452/terry-savage-federal-student-loans-next-big-crisis.html. However, this “sky is falling” perspective on student loans is nothing new. Since the Code’s enactment in 1978, debtors have faced ever-tightening standards making student loans harder to discharge. Prior to the Code’s enactment, student-loan debt could be discharged automatically like many other loans in the bankruptcy context. 9Brendan Baker, Deeper Debt, Denial of Discharge: The Harsh Treatment of Student Loan Debt in Bankruptcy, Recent Developments, and Proposed Reforms, 14 U. Pa. J. Bus. L. 1213, 1214 (2012). The Code imposed a new conditional discharge standard 10See 11 U.S.C. § 523(a)(8) (2012) (requiring the debtor prove that the student-loan debt would impose an undue hardship on the debtor if not discharged). upon student-loan debt, theoretically targeting rampant abuse of the bankruptcy system. Legislators conjured images of college graduates freshly emerging from their universities with advanced degrees in hand and bright futures who might then seek to discharge their significant debt before accepting a high-paying job. 11See H.R. Rep. No. 95-595, at 536–38 (1977), reprinted in 1978 U.S.C.C.A.N. 5963, 6425 (insisting that the average student debtor is trying to take advantage of a loophole in the system). Professors Rafael Pardo and Michelle Lacey, in a 2005 study, addressed this abusive student-loan debtor stereotype thoroughly and concluded that the justification of the change was built on bombastic and factually incorrect assertions. See Undue Hardship, supra note 6, at 419–32. Even in the face of a study proving the contrary, Congress enacted the new provision and has never looked back. 12H.R. Rep. No. 95-595, at 133.

An unfair debtor stereotype based on assumptions directly contradicted by empirical evidence 13Id. was only the beginning of the difficulty faced by student-loan debtors under the new undue hardship standard. 1411 U.S.C. § 523(a)(8). The Code does not define “undue hardship,” 15Section 523(a)(8) states that educational debt is excepted from discharge unless it “would impose an undue hardship,” but the following subparagraphs and § 101 do not define what qualifies as an undue hardship. See Id. §§ 101, 523(a)(8) (omitting any elaboration on what undue hardship means). and the task has now fallen to the bankruptcy courts to determine how to apply this standard. In response, bankruptcy courts developed two predominant tests to assess whether a debtor’s student-loan debt imposed an undue hardship: the totality of circumstances test and the three-factor Brunner test. 16See Brunner v. N.Y. State Higher Educ. Servs. Corp., 831 F.2d 395, 396 (2nd Cir. 1987) (per curiam) (creating the Brunner test); Andrews v. S.D. Student Loan Assistance Corp. (In re Andrews), 661 F.2d 702, 704 (8th Cir. 1981) (creating the totality of the circumstances test). Both tests analyze the same core considerations, 17See Cheney v. Educ. Credit Mgmt. Corp. (In re Cheney), 280 B.R. 648, 659 (N.D. Iowa 2002). and likewise, both suffer from the same maladies—inconsistency and subjectivity. The result is that these tests, which lean heavily on the court’s ability to predict the future as to the debtor’s potential future income or wage earning ability, 18The fact that this inquiry has been problematic is buttressed by an expansive 2005 study that looked at numerous cases to try to find factual differences that underlay judicial determinations. See Undue Hardship, supra note 6, at 503. The results of this study suggest that there is a great amount of inconsistency in judicial determinations when applying different facts to these vague standards. See id. (finding few statistically significant differences between debtors found to have a future ability to repay and those with a future inability to repay). have created a system with an unsettling amount of judicial discretion and subjectivity. 19Professor Douglass Boshkoff stated in his article that “experience with the conditional discharge of educational debts in our country suggests that the bankruptcy judge will be given almost unlimited power to determine the lifestyle of a debtor who seeks a discharge.” Douglass G. Boshkoff, Limited, Conditional and Suspended Discharges in Anglo-American Bankruptcy Proceedings, 131 U. Pa. L. Rev. 69, 125 (1982).

Despite the broad-sweeping negative public outlook on student-loan debt, the prejudicial stereotype faced by student-loan debtors, and the inconsistency produced by judicially created tests, a 2009 study of 115 bankruptcy filings in the Western District of Washington indicated that 57% of those who filed adversary proceedings seeking discharge of their student loans were able to get some or all of their loans discharged. 20The Real Student-Loan Scandal, supra note 6, at 184. So, why does the negativity associated with attempting to discharge student loans persist? The same study further concluded that, even in light of the data showing that more than half of the debtors studied received discharge of at least some of their student-loan debt, the Code’s undue hardship discharge provision has failed to optimize the financial health of debtors in distress. 21Id. at 235. Furthermore, because the study found that courts relied more heavily on extralegal factors than legal factors to make relief determinations, it also asserted that the discharge determination process exhibited “hallmarks of a system that has run amok.” 22Id. This article also postulated that the “higher education finance system suffers from schizophrenia—namely, a public-oriented approach to student-loan origination but a business-oriented approach to student-loan collection.” Id. This view echoes the concerns voiced at the time of the Code’s enactment:[I]f [student] loans are granted too freely and that is what is causing the increase in bankruptcies, then the problem is a general problem, not a bankruptcy problem. The loan program should be tightened, or collection efforts should be increased. If neither of those alternatives is acceptable, then the loan programs should be viewed as general social legislation that has an associated cost. It is inappropriate to view the program as social legislation when granting the loans, but strictly as business when attempting to collect. Such inconsistency does not square with general bankruptcy policy.H.R. Rep. No. 95-595, at 134 (1977), reprinted in 1978 U.S.C.C.A.N. 5963, 6095. Thus, even with some measureable degree of success for student-loan debtors in dire need of financial relief, the bankruptcy system remains plagued by an inconsistent message from the bankruptcy courts and, ultimately, Congress—anyone can receive educational loans, but it takes a very special person to discharge them.

This Comment will build on research that has documented the symptomatic problems that bankruptcy courts have had in applying the Code’s undue hardship provision. Ultimately, this Comment aims to provide practitioners with new information about the approach appellate courts have taken when considering lower court determinations of undue hardship. To accomplish this goal, this Comment will present findings from an empirical analysis of ten years of bankruptcy appellate opinions spanning from January 1, 2002, to December 31, 2011.

This Comment is organized into four parts. Part I explains the background and development of law surrounding the application of the undue hardship provision. Part I.A discusses discharge in bankruptcy proceedings generally, the undue hardship discharge provision and the legislative history prior to its enactment, and the rapid expansion of the exception to discharge. Part I.B examines the two judicially-created tests that courts use to determine whether a debtor has met his or her undue hardship burden and the problems these tests present. Part I.C analyzes the procedural requirements of the undue hardship provision and how they comport with the Code’s overarching principles concerning the scope of a debtor’s opportunity for a fresh start.

Part II addresses the bankruptcy appellate system as a whole. Part II.A outlines the function of this Comment’s study. Part II.B provides an overview of the unique structure of the bankruptcy appellate system. Part II.C delineates the appellate process. Finally, Part II.D discusses the reasons why parties seek to appeal determinations of undue hardship.

Part III then presents the empirical study of bankruptcy appellate opinions. Part III.A discusses the methodology by which opinions were selected for analysis, while Part III.B presents the findings of this study and its conclusions. Finally, this Comment concludes in Part IV with reflections on the results of the study and how these results relate to the principles of bankruptcy. More specifically, this Comment will discuss the study’s implications for practitioners as they make choices about how to effectively maneuver through the bankruptcy appellate system and for judges as they consider undue hardship determination on appeal.

I. The Undue Hardship Discharge Provision—A Look at Its History, Application, and Difficulties at the Bankruptcy Court Level

This Part begins with a discussion of the underlying policies of bankruptcy effectuated by the discharge of debt and describes the history behind the undue hardship provision’s enactment and subsequent legislation. Part I.B will discuss the application of the two judicially-created tests to determine whether a debtor has proven undue hardship and provide a critique of their results. Finally, Part I.C concludes with a critique of the procedural difficulties presented by the undue hardship discharge provision.

A. The Development and Progression of § 523(a)(8)’s Undue Hardship Provision

The bankruptcy system has two foundational principles that guide the interpretations of the Code’s law and policy: (1) a fresh start for the debtor (the “fresh start principle”) and (2) equal treatment of similarly situated creditors (the “equality principle”). 23See S. Rep. No. 95-989, at 7 (1978), reprinted in 1978 U.S.C.C.A.N. 5787, 5793; H.R. Rep. No. 95-595, at 177–78. The fresh start principle is based on “the notion that substantive relief should be afforded in the form of forgiveness of existing debt, with relinquishment by the debtor of . . . existing nonexempt assets.” 24Undue Hardship, supra note 6, at 414–15; see Local Loan Co. v. Hunt, 292 U.S. 234, 244 (1934) (stating that the fresh start policy of bankruptcy law gives a debtor who surrenders property for distribution to creditors freedom from pre-existing debt). The equality principle provides a basis for the procedural processes of the Code that seek to maximize collection efforts of creditors when the debtor has insufficient assets to repay. 25See Undue Hardship, supra note 6, at 415 (noting that the bankruptcy system solves the “common pool problem” that can arise if creditors only pursue their own interests at the cost of other creditors) (citing Thomas H. Jackson, The Logic and Limits of Bankruptcy Law 7–19 (1986) (giving a comprehensive analysis of the role of bankruptcy law as a “collective debt-collection device”)).

Within the context of chapter 7, debt discharge is generally automatic. 26See 11 U.S.C. § 727(a) (2012) (“The court shall grant the debtor a discharge, unless . . . .”) (emphasis added). After a debtor has filed a petition for bankruptcy, a creditor must file a proof of claim to receive any compensation that might result from the liquidation of the debtor’s assets. 27See id. § 501(a). But, note that creditors do not file proof of claims in no-asset chapter 7 cases. Once the creditor’s allowable claim is calculated 28See id. §§ 502, 506. and the distribution of the debtor’s assets is completed, 29See id. §§ 725–26. the debtor’s remaining debt is automatically discharged, 30See id. § 727(a) (“The court shall grant the debtor a discharge, unless . . .”) (emphasis added). relieving the debtor of personal liability for his or her outstanding debts. 31See id. § 727(b) (“[T]his section discharges the debtor from . . . any liability on a claim.”). The automatic discharge is subject to certain exceptions 32See id. §§ 523, 727. where Congress deemed the interest in full recovery of debt to outweigh the debtor’s interest in a fresh start. 33See Patel v. Shamrock Floorcovering Servs., Inc. (In re Patel), 565 F.3d 963, 967 (6th Cir. 2009) (noting that the statutory exceptions to discharge recognize Congressional intent to protect certain creditors from harm); Miller v. Lewis, 391 B.R. 380, 384 (E.D. Tex. 2008) (stating that certain debts, for public policy reasons, have been decided by Congress to override policies favoring a fresh start); Bolen v. Sallie Mae Servicing Corp. (In re Bolen), 287 B.R. 127, 129 (D. Vt. 2002) (noting that there are certain circumstances where giving the debtor a fresh start is not the paramount concern and that protection of the creditor is more important).

Unlike other discharge exceptions, Congress deemed that the student-loan debt discharge exception needed additional protection from discharge. 34See 11 U.S.C. § 523(a)(8). Curiously, the student-loan debt exception is not particularly straightforward. Rather, student-loan debt is unique because its exception from dischargeability creates a very specific test or condition upon which it can be discharged. 35See id. (creating the condition of “undue hardship” to be able to discharge student-loan debt). This has not always been the case. Prior to the Code’s enactment, student-loan debt was automatically dischargeable. 36See id. § 35(a) (1976) (omitting educational debt from the list of debts unaffected by discharge) (repealed 1978). However, fear of abuse by student-loan debtors led to the addition of the undue hardship test to determine whether a debtor could obtain relief. 37See H.R. Rep. No. 95-595, at 536–38 (1977), reprinted in 1978 U.S.C.C.A.N. 5963, 6425. At its inception, then, § 523(a)(8)’s undue hardship provision was unduly slanted against the student-loan debtor.

Analyzing this supposed slant, Rafael Pardo, a prominent bankruptcy law professor, and Michelle Lacey, an accomplished mathematics professor, performed an empirical study to examine this supposed rampant abuse and the effects of applying the undue hardship provision. 38Undue Hardship, supra note 6. Their study found an unearned, negative stereotype of the student-loan debtor presented in the legislative history of the enactment of the Code. 39Id. at 419–28. Arguing against the enactment of the undue hardship provision, Representative James O’Hara stated that the provision was “a discriminatory remedy for a ‘scandal’ which exists primarily in the imagination.” See H.R. Rep No. 94-1232 (1976), reprinted in H.R. Rep. No. 95-595, at 148, and in 1978 U.S.C.C.A.N. 5963, 6109. Representative O’Hara decried the provision for likening the Code’s treatment of student-loan debtors to those who obtain loans by “fraud, felony, and alimony-dodging.” Id. (asserting that the new provision subjects educational loans to an assumption of criminality). Further substantiating Representative O’Hara’s claims, a GAO study performed before the provision officially became law “indicated that less than one percent of all federally insured and guaranteed educational loans were discharged in bankruptcy.” Undue Hardship, supra note 6, at 423; H.R. Rep. No. 95-595, at 133.In the face of such convincing evidence, Representative Allen E. Ertel opposed the removal of the undue hardship provision by resorting to unsubstantiated and unrealistic characterizations of student-loan debtors as bad faith actors who would inevitably abuse the bankruptcy system. See Undue Hardship, supra note 6, at 424; H.R. Rep. No. 95-595, at 536–38. Representative Ertel insisted that, without the undue hardship provision, the law would “encourage fraud,” and, furthermore, student-loan debtors would simply take advantage of a free education, immediately discharge their debts upon graduation, and begin their promising careers with a “clean slate” and the “excellent credit rating that accompanies a bankruptcy.” See H.R. Rep. No. 95-595, at 536–38. Contra Undue Hardship, supra note 6, at 424 (disputing Representative Ertels’s assertion that bankruptcy results in excellent credit ratings). Siding ultimately with Representative Ertel, Congress enacted the provision, which therefore, at its inception, vilified the student-loan debtor. Sadly, courts have embraced this lopsided stereotype when applying the law. See Educ. Credit Mgmt. Corp. v. Mersmann (In re Mersmann), 505 F.3d 1033, 1042 (10th Cir. 2007) (“Limitations on the discharge ability [sic] of student loans serve [to] ‘prevent[] abuses of the educational loan system by restricting the ability to discharge a student loan shortly after a student’s graduation . . . .’”); Undue Hardship, supra note 6, at 427 n.112. Furthermore, their study concluded that the landscape of student-loan debtors is not so inundated with abusers of the bankruptcy system, as Congress had feared upon enactment of the Code. 40Undue Hardship, supra note 6, at 459.

However, since its inception, the Code’s student-loan exception has progressively become more protective of student-loan lenders and harsher in its treatment of student-loan debtors. In its initial iteration, the exception provided the option of discharge both under the undue hardship test and when a debt “first became due before five years before the date of the filing of the petition.” 41Bankruptcy Reform Act of 1978, Pub. L. No. 95-598, § 101, 92 Stat. 2549, 2591 (1978) (formerly codified at 11 U.S.C. § 523(a)(8)(A) (1988)). In 1990, Congress extended the five-year exception to seven years. 42See Robert C. Cloud, When Does Repaying a Student Loan Become an Undue Hardship?, 185 Educ. L. Rep. 783, 786 n.20 (2004). In 1998, Congress eliminated the seven-year exception altogether 43See Higher Education Amendments of 1998, Pub. L. No. 105-244, § 971(a), 112 Stat. 1581 (1998). Interestingly, this amendment came in the wake of a 1997 report by the National Bankruptcy Review Commission recommending that Congress eliminate § 523(a)(8) altogether. See Nat’l Bankr. Review Comm’n, Final Report, § 1.4.5 (October 20, 1997). The report went on to say that student loans should be “treated like all other unsecured debts. In so doing, the dischargeability provisions would be consistent with federal policy to encourage educational endeavors. . . . Litigation over ‘undue hardship’ would be eliminated, so that the discharge of student loans no longer would be denied to those who need it most.” Id. and left student-loan debtors with only the undue hardship test for relief. Most recently, the Bankruptcy Abuse Prevention and Consumer Protection Act of 2005 (“BAPCPA”) expanded the coverage of § 523(a)(8)’s protection to private, for-profit lenders. 44See Bankruptcy Abuse Prevention and Consumer Protection Act of 2005, Pub. L. No. 109-8, § 220, 119 Stat. 23, 59 (codified at 11 U.S.C. § 523(a)(8)(B) (2012)). The BAPCPA amendment further reflects Congress’s pro-lender stance and represents the latest step in the progression of § 523(a)(8) and its attempt to curtail the bankruptcy relief available to student-loan debtors. 45See The Real Student-Loan Scandal, supra note 6, at 180–82 (expounding on the impact of the 2005 amendments, the unregulated nature of private student loans, and the creditor-friendly policies of Congress).

B. The Judicially Created Tests to Assess a Debtor’s Claim of Undue Hardship

The application of the undue hardship provision in bankruptcy courts has been plagued with inconsistency, 46Despite Congress’ repeated re-drafting of § 523(a)(8), incrementally limiting the ability to discharge student-loan debt, they never attempted to define the core provision of exception—undue hardship. See 11 U.S.C. §§ 101, 523(a)(8) (omitting any elaboration on what undue hardship means); supra Part I.A. which has primarily resulted from significant judicial subjectivity in undue hardship determinations. 47See Undue Hardship, supra note 6, at 503 (faulting judicial discretion for the haphazard results in undue hardship litigation and for compromising the fresh start principle). Subjectivity here refers to the overall application of the undue hardship provision across all districts. Although courts find themselves simply limited by precedents today, the precedents differ from district to district and are the result of an initial judge’s determination that they would allow or disallow some expense when determining dischargeability. Because the precedents differ, the overall result is indicative of the initial subjective determinations. Congress failed to define what undue hardship meant when § 523(a)(8) was created, 48Section 523(a)(8) states that educational debt is excepted from discharge unless it “would impose an undue hardship,” but the following subparagraphs and § 101 of the Code do not define what qualifies as an undue hardship. See 11 U.S.C. §§ 101, 523(a)(8) (omitting any elaboration on what undue hardship means). and in the wake of its vague wording, courts created two principal tests to assess a debtor’s claim of undue hardship: (1) the Brunner test 49 See Brunner v. N.Y. State Higher Educ. Servs. Corp., 831 F.2d 395 (2nd Cir. 1987) (per curiam). and (2) the totality of the circumstances test. 50 See Andrews v. S.D. Student Loan Assistance Corp. (In re Andrews), 661 F.2d 702, 704 (8th Cir. 1981). The Brunner test is the predominant test used by most circuits, but this section will address both tests in turn and critique their shortcomings in litigation at the bankruptcy court level.

1. The Brunner Test

The Second Circuit’s opinion in Brunner v. New York State Higher Education Services Corp. has become the principal standard for judging whether a debtor has shown that he or she will suffer undue hardship without the discharge of his or her student-loan debt. 51Brunner, 831 F.2d at 396. The court in Brunner, recognizing that there was little appellate guidance concerning the undue hardship provision, took the opportunity to weigh in and formally adopt the test spelled out in the District Court from which this appeal originated. Id. Quickly analyzing each of the three prongs in turn, the court noted that prongs one and three comported with common sense and the legislative intent behind § 523(a)(8). Id. Though it recognized the second prong was “problematic,” the court ultimately held that requiring evidence that a hardship would continue into the future “more reliably guarantees that the hardship presented is ‘undue.’” Id. The Brunner test requires:

(1) that the debtor cannot maintain, based on current income and expenses, a “minimal” standard of living for herself and her dependents if forced to repay the loans; (2) that additional circumstances exist indicating that this state of affairs is likely to persist for a significant portion of the repayment period of the student loans; and (3) that the debtor has made good faith efforts to repay the loans. 52Id.

A debtor must establish that he or she meets the standard of each prong to satisfy the undue hardship provision. 53See id. (requiring each prong be met in order to afford relief); Pobiner v. Educ. Credit Mgmt. Corp. (In re Pobiner), 309 B.R. 405 (Bankr. E.D.N.Y. 2004); Williams v. EFG Tech/Rutgers (In re Williams), 296 B.R. 128 (Bankr. D.N.J. 2003); Shankwiler v. Nat’l Student Loan Mktg. (In re Shankwiler), 208 B.R. 701 (Bankr. C.D. Cal. 1997). It also bears noting that once the debtor establishes that he or she would suffer undue hardship, courts have differed in their approach as to whether this entitles the debtor to the full discharge of his or her debt or a partial discharge that brings the student-loan debt level down to a judicially determined manageable level. See Amanda M. Foster, All or Nothing: Partial Discharge of Student Loans Is Not the Answer to Perceived Unfairness of the Undue Hardship Exception, 16 Widener L.J. 1053, 1072–83 (2007) The circuit courts are split over this issue, and scholars on both sides of the argument have advocated strongly for congressional clarification. See Frank T. Bayuk, The Superiority of Partial Discharge for Student Loans under 11 U.S.C. § 523(a)(8): Ensuring a Meaningful Existence for the Undue Hardship Exception, 31 Fla. St. U. L. Rev. 1091, 1116 (2004) (arguing that partial discharge better comports with ideals of economic and fundamental fairness for both the creditor and debtor); Foster, supra, at 1073 (arguing that the Code, based on plain meaning and statutory construction, legislative history, and principles of fairness, requires a strict, all-or-nothing approach, awarding the debtor full discharge if he or she meets the undue hardship standard).

Judicial application of the three prongs of the Brunner test has generated many difficulties for student-loan debtors. To prevail under the first prong—the current inability to repay—a debtor must establish that, based on his or her current income and expenses, repayment of the student-loan debt would force the debtor to fall beneath a “minimal standard of living.” 54See generally Brunner, 831 F.2d 395. Courts have interpreted this first prong to call for an analysis of the debtor’s income and expenses, finding that when expenses exceed income, “the debtor will have established an inability to maintain a standard of living in absence of an undue hardship discharge.” 55See The Real Student-Loan Scandal, supra note 6, at 196–97 n.76. However, analyzing income and expenses to determine what constitutes a minimal standard of living and whether the debtor may fall beneath it is a fact-intensive inquiry that allows for considerable amounts of judicial subjectivity because of an ambiguous standard and differing interpretations of what constitutes “minimal living.” 56See Hart v. ECMC (In re Hart), 438 B.R. 406, 410 (E.D. Mich. 2010) (requiring the debtor to show he is currently minimizing his expenses); Innes v. Kansas (In re Innes), 284 B.R. 496, 509 (D. Kan. 2002) (determining that when a debtor’s expenses exceed their income they cannot maintain a minimal standard of living and repay their loan); Pincus v. Graduate Loan Ctr. (In re Pincus), 280 B.R. 303, 317 (Bankr. S.D.N.Y. 2002) (holding that a debtor is not required to live below the poverty line but must show that repayment will require more than a restricted budget); The Real Student-Loan Scandal, supra note 6, at 196–97 n.76 (highlighting the Ninth Circuit’s approach requiring that the debtor fall somewhere between “‘temporary financial adversity’ and ‘utter hopelessness’” to satisfy the first prong). It is in applying the facts of a debtor’s expenses to these vague standards that a judge is able to impose a significant amount of subjectivity. See Pincus, 280 B.R. at 317 (finding a debtor’s expenses excessive in the areas of “(1) communication, (2) cable, (3) food, (4) clothing, (5) laundry and dry-cleaning, (6) transportation, (7) recreation, and (8) certain medical expenses” and further suggesting the debtor could have used more inexpensive cell phone contract plans, eaten out less, and washed his own clothes). Thus, what should be a straightforward test to determine if expenses exceed income may morph into an opportunity for a judge to mandate what kind of cable TV or wireless phone plan the debtor should have. 57See Pincus, 280 B.R. at 317.

Should the debtor establish his or her current inability to repay, he or she must then demonstrate that the less than minimal standard of living will persist if the student-loan debt is not discharged. 58See Brunner, 831 F.2d at 396. Put more simply, the debtor’s current inability to repay is not enough; it must be clear that the debtor will be unable to repay in the future. 59See id.; Sanborn v. Educ. Credit Mgmt. Corp. (In re Sanborn), 431 B.R. 5 (Bankr. D. Mass. 2010) (holding that the debtor must show that their condition of undue hardship will continue into the future). Bankruptcy courts have found this prong to be the most important, insisting that a debtor “must show that circumstances indicate a certainty of hopelessness, not merely a present inability to fulfill financial commitment.” 60Wallace v. Educ. Credit Mgmt. Corp. (In re Wallace), 443 B.R. 781, 789 (Bankr. S.D. Ohio 2010) (quoting Barrett v. Educ. Credit Mgmt. Corp. (In re Barrett), 487 F.3d 353, 359 (6th Cir. 2007)). Courts have attempted to analyze the debtor’s future inability to repay upon certain objective factors, 61See id. (asserting that the court may consider such circumstances as illness, disability, a lack of useable job skills, or the existence of a large number of dependents); Barrett, 487 F.3d at 359 (holding that the overarching principle that should determine whether the debtor satisfies this prong must be that the additional circumstances be “beyond the debtor’s control, not borne of free choice”). The Ninth Circuit has attempted to identify a list of factors that might be useful in determining whether the debtor will have a future inability to pay. Educ. Credit Mgmt. Corp. v. Nys (In re Nys), 446 F.3d 938, 946 (9th Cir. 2006). These factors, otherwise known as the Nys factors, include:(1) Serious mental or physical disability of the debtor or the debtor’s dependents which prevents employment or advancement; (2) The debtor’s obligations to care for dependents; (3) Lack of, or severely limited education; (4) Poor quality of education; (5) Lack of usable or marketable job skills; (6) Underemployment; (7) Maximized income potential in the chosen educational field, and no other more lucrative job skills; (8) Limited number of years remaining in [the debtor’s] work life to allow payment of the loan; (9) Age or other factors that prevent retraining or relocation as a means for payment of the loan; (10) Lack of assets, whether or not exempt, which could be used to pay the loan; (11) Potentially increasing expenses that outweigh any potential appreciation in the value of the debtor’s assets and/or likely increases in the debtor’s income; (12) Lack of better financial options elsewhere.Id. but even these factors allow for significant judicial interpretation and speculation. 62See The Real Student-Loan Scandal, supra note 6, at 196–97 n.76 (finding, in a study of cases in the Ninth Circuit, reliance on extralegal factors, such as what judge decided the case, to be more significant in the outcome of whether a debtor satisfied the second prong of Brunner). However, speculation on the judge’s part is inherent in an inquiry that essentially asks the judge to predict the future. 63Roundtree-Crawley v. Educ Credit Mgmt. Corp., (In re Crawley), 460 B.R. 421, 439 (Bankr. E.D. Pa. 2011) (noting that the inquiry is forward-looking in nature); Berry v. Educ. Credit Mgmt. Corp. (In re Berry), 266 B.R. 359, 364 (Bankr. N.D. Ohio 2000) (recognizing that any analysis under the second prong of Brunner is, by its very nature, speculative). This prong is extremely problematic because it is unclear exactly how a student-loan debtor can prove that his or her situation falls into the area envisioned by the Brunner test as sufficiently hopeless to warrant a grant of discharge. Evidence from a 2007 study suggests a debtor’s health status is likely to have the most statistically significant association with discharge determination in a student-loan debtor’s case. 64See Pardo, supra note 6, at 510–12 (finding in a study of a decade of bankruptcy decisions that the debtor’s health played the largest role, even eclipsing a debtor’s monthly household income and expenses, in a bankruptcy court’s determination of undue hardship). Ultimately, though, this crystal-ball-gazing exercise performed by courts is fraught with subjectivity and guesswork.

Finally, debtors must establish that they have made a good faith effort to repay their loans, which the court assesses by looking at the debtor’s prepetition conduct. 65Brunner, 831 F.2d at 396 (finding that the debtor did not act in good faith by filing for discharge only within a month from when the first payment was due and by failing to take less drastic remedies before filing for discharge). Courts have recognized that good faith is “both an intangible and subjective standard,” 66E.g., Grine v. Tex. Guar. Student Loan Corp. (In re Grine), 254 B.R. 191, 193 (Bankr. N.D. Ohio 2000). which has resulted in even more inconsistent application of the Brunner test. 67See The Real Student-Loan Scandal, supra note 6, at 200 & n.98 (2009) (noting the “inherent subjectivity in an amorphous standard such as good faith”). Factors vary from circuit to circuit, but most circuits analyze some form of the following four factors to determine whether a debtor has made a good faith effort to repay:

(1) whether the debtor attempts to repay the debt; (2) the length of time after the student loan becomes due that the debtor seeks to discharge the debt; (3) the percentage of the student loan debt in relation to the debtor’s total indebtedness; (4) the debtor’s attempts to find suitable employment. 68Green v. Sallie Mae Servicing Corp. (In re Green), 238 B.R. 727, 736 (Bankr. N.D. Ohio 1999) (citations omitted); see Hart v. ECMC (In re Hart), 438 B.R. 406, 413 (E.D. Mich. 2010); Dep’t of Educ. v. Wallace (In re Wallace), 259 B.R. 170, 185 (C.D. Cal. 2000) (citations omitted).

Basically, a court must determine whether a debtor has tried to find work, maximize income, and minimize expenses. 69Afflitto v. United States (In re Afflitto), 273 B.R. 162, 171–72 (Bankr. W.D. Tenn. 2001) (citing Goulet v. Educ. Mgmt. Corp. (In re Goulet), 264 B.R. 527, 531 (W.D. Wis. 2001)) (finding that not making repayments does not preclude a finding of good faith if the debtor has sought to maximize income potential and minimize expenses); see also Downey v. Sallie Mae, Inc. (In re Downey), 255 B.R. 72, 77 (Bankr. N.D. Fla. 2000) (deciding good faith was not established because debtor failed to show she minimized her expenses). Holding consistent with the other prongs in the Brunner test, this prong opens the door to a fact intensive inquiry by judges, who have historically applied the standard from a “less forgiving stance.” 70The Real Student-Loan Scandal, supra note 6, at 200; see Katheryn E. Hancock, A Certainty of Hopelessness: Debt, Depression, and the Discharge of Student Loans under the Bankruptcy Code, 33 Law & Psychol. Rev. 151, 156 (2009) (arguing that application of the good faith prong is unduly harsh based upon the strict approach originally taken when § 523(a)(8) was enacted and debtors still had the option to come back and have their loans discharged automatically without having to prove undue hardship). All in all, the Brunner test reflects the general attitude of the Code’s progressively harsher stance toward student-loan debtors, resulting in a test that is extremely technical and leaves very little room for a debtor to succeed. 71See, e.g., H. Amend. 939 to H.R. 4137 (offered Feb. 7, 2008) (striking down a proposed amendment to repeal protection afforded to private student-loan lenders).

2. The Totality of the Circumstances Test

Offering an alternative to the Brunner test, the Eighth Circuit has officially adopted the totality of the circumstances test. 72See Andrews v. S.D. Student Loan Assistance Corp. (In re Andrews), 661 F.2d 702, 704 (8th Cir. 1981) (expressing a preference for a less restrictive test); see also Long v. Educ. Credit Mgmt. Corp. (In re Long), 322 F.3d 549, 553 (8th Cir. 2003) (formally embracing the Andrews opinion). This approach has been endorsed by not only the Eighth Circuit, but also by courts within the First Circuit. Bronsdon v. Educ. Credit Mgmt. Corp. (In re Bronsdon), 435 B.R. 791, 800 (B.A.P. 1st Cir. 2010). Analysis under this test requires the court to consider: “(1) the debtor’s past, present, and reasonably reliable future financial resources; (2) a calculation of the debtor’s and her dependent’s reasonable necessary living expenses; and (3) any other relevant facts and circumstances surrounding each particular bankruptcy case.” 73Long, 322 F.3d at 554 (citing Andrews, 661 F.2d at 704); see also Ackley v. Sallie Mae Student Loans (In re Ackley), 463 B.R. 146, 149 (Bankr. D. Me. 2011) (implementing a similar three-prong test). An expanded list of facts and circumstances to which bankruptcy courts may look to determine whether the debtor is entitled to an undue hardship discharge include:(1) debtor’s total present and future incapacity to pay debts for reasons not within control of debtor; (2) whether debtor has made good faith effort to negotiate a deferment or forbearance of payment; (3) whether hardship will be long-term; (4) whether debtor has made payments on student loans; (5) whether debtor suffers from permanent or long-term disability; (6) ability of debtor to obtain gainful employment in area of study; (7) whether debtor has made good faith effort to maximize income and minimize expenses; (8) whether dominant purpose of bankruptcy petition is to discharge student loans; and (9) ratio of student loan debt to debtor’s total indebtedness.D’Ettore v. Devry Inst. of Tech. (In re D’Ettore), 106 B.R. 715, 718 (Bankr. M.D. Fla. 1989) (citations omitted); see also Holmes v. NCO Fin. Sys. (In re Holmes), 319 B.R. 620, 622–23 (Bankr. W.D. Mo. 2004); Houshmand v. Mo. Student Loan Program (In re Houshmand), 320 B.R. 917, 920 (Bankr. W.D. Mo. 2004); Morgan v. Dep’t of Higher Educ. (In re Morgan), 247 B.R. 776, 782 (Bankr. E.D. Ark. 2000) (citing In re D’Ettore, 106 B.R. at 718). Ultimately, in adopting the totality of the circumstances test, courts look at many of the same facts as those that apply the Brunner test to determine whether the debtor has proven undue hardship. 74See Pardo, supra note 6, at 515 nn.41–42 (showing different tests analyzing similar factors, with debtor illness emerging as a prominent consideration under both tests).

Much like the analysis under the Brunner test, a multifactor test like the totality of the circumstances allows for considerable amounts of judicial discretion and subjectivity. “The ‘totality of the circumstances’ is obviously a very broad test, giving [courts] considerable flexibility.” 75Houshmand, 320 B.R. at 920; see Dep’t of Educ. v. Meling (In re Meling), 263 B.R. 275, 278–79 (Bankr. N.D. Iowa 2001), aff’d, No. C01-2027, 2002 WL 32107248 (N.D. Iowa Jan. 22, 2002) (noting that courts have broad latitude to consider any relevant factor). This test is generally viewed as less restrictive than the Brunner test because it allows consideration of a wider range of factors. 76See Houshmand, 320 B.R. at 920. Notably, these two tests differ in that under the totality of the circumstances test, no factor is dispositive; a finding against the debtor on a particular factor does not necessitate automatic denial of discharge. 77Morgan, 247 B.R. at 782. However, both tests focus on the same core considerations 78Weir v. Paige (In re Weir), 296 B.R. 710, 716 (Bankr. E.D. Va. 2002) (“Regardless of the test used in determining whether repayment of student loans constitutes undue hardship under § 523(a)(8), at a minimum the court must focus on two issues: (1) the economic prospects of the debtor and (2) whether the conduct of the debtor disqualifies the debtor from taking advantage of the exception.”). —the debtor’s financial ability to repay and the debtor’s conduct—which has led to similar issues with inconsistency and judicial subjectivity. Thus, here too, student-loan debtors face the same systemic message that the law is slanted against them with little hope of escaping the unfair stereotype.

C. Section 523(a)(8)’s Function as a Procedural Bar to Student-Loan Debt Relief

Student-loan debtors must contend not only with the substantive difficulties in proving they meet the vague standard of § 523(a)(8), but they must also overcome the difficulty presented by the procedural requirements of the Bankruptcy Rules. Foremost among these hurdles is the necessity of an adversary proceeding to determine dischargeability. 79Fed. R. Bankr. P. 7001(6). Once a student-loan creditor has established the existence of a valid claim, the debtor has the burden to prove he or she will suffer undue hardship without discharge relief. 80See In re Betz, 31 B.R. 565, 566 (Bankr. W.D. N.Y. 1983); Conn. Student Loan Found. v. Keenan (In re Keenan), 53 B.R. 913, 918 (Bankr. D. Conn. 1985) (finding that defendant sustained the burden of proof in showing undue hardship, thus allowing for the defendant’s debts to be discharged); Garneau v. N.Y. State Higher Educ. Servs. Corp. (In re Garneau), 122 B.R. 178, 179 (Bankr. W.D.N.Y. 1990) (holding that undue hardship must be proved by a preponderance of the evidence). To accomplish this, the debtor must file an adversary proceeding to “determine the dischargeability of a debt.” 81Fed. R. Bankr. P. 7001(6). Unfortunately, litigation concerning an undue hardship discharge is, in some sense, almost antithetical to the Code’s approach to litigation generally. As it is most likely to occur in a situation where the debtor is in dire financial straits, such litigation threatens to encroach upon a debtor’s fresh start. 82The Real Student-Loan Scandal, supra note 6, at 189 (acknowledging four categories of cases concerning the creditor and debtor’s assessment of the nondischargeability status of a debt and finding that litigation is most likely to occur “where the creditor believes the debt to be nondischargeable and the debtor believes the debt to be dischargeable”).

The Code’s approach to the protection of a debtor’s fresh start generally discourages excess litigation during bankruptcy proceedings. Upon filing for bankruptcy, the debtor is afforded protection through the automatic stay provision, which prevents the debtor from facing any proceedings concerning prebankruptcy debt, including arbitration, license revocation, and administrative and judicial proceedings. 83See 11 U.S.C. § 362 (2012); S. Rep. No. 95-989, at 50 (1978), reprinted in 1978 U.S.C.C.A.N. 5787, 5836; H.R. Rep. No. 95-595, at 340 (1977), reprinted in 1978 U.S.C.C.A.N. 5963, 6296–97. The automatic stay is a fundamental function of the Code to protect debtors from overzealous and encroaching creditors. See 11 U.S.C. § 362; S. Rep. No. 95-989, at 50; H.R. Rep. No. 95-595, at 340. Thus, from the debtor’s perspective, this protection evidences the Code’s inherent hostility toward litigation over claims that arose before the filing of bankruptcy.

By allowing a creditor’s claim to simply be a right to payment, which is presumptively valid in the absence of an objection, the Code proclaims its preference for efficiency and expediency in bankruptcy. 84See 11 U.S.C. §§ 101(5)(A), 502(a); see also Fed R. Bankr. P. 3001(f) (proscribing that a properly filed proof of claim is prima facie evidence of the validity and amount of that claim). Furthermore, the equality principle, which seeks to maximize a debtor’s assets through a common pool system of treatment of creditors, discourages individual creditor litigation as it will necessarily limit the recovery of all creditors by causing the debtor to expend resources that would normally be distributed amongst all creditors. 85The Real Student-Loan Scandal, supra note 6, at 186–87. Thus, the procedural requirements put in place by Congress seem adverse to the Code’s general attitude toward the litigation of pre-bankruptcy debts. 86See id. at 188 (putting forth the counterarguments that undue hardship litigation was the type of litigation the Code seeks to prevent).

Beyond the seeming paradox of undue hardship litigation in light of the Code’s protective automatic stay approach, a debtor’s fresh start is also gravely threatened by the requirement of an adversary proceeding, which places the debtor at a significant disadvantage. First, to litigate a claim, the debtor will be forced to expend vital resources at a time when resources are scarce. 87See Katherine Porter & Deborah Thorne, The Failure of Bankruptcy’s Fresh Start, 92 Cornell L. Rev. 67, 124 (2006) (finding that the fresh start may be “more myth than magic bullet” and that many chapter 7 debtors continue to have significant financial trouble post-bankruptcy). Next, there is generally a power imbalance between litigants, “[a]s debtors with student loans are likely to owe debts to large institutional creditors that are well funded, legally sophisticated, and repeat players.” 88The Real Student-Loan Scandal, supra note 6, at 191 (alteration in original). Finally, a debtor may attempt to prove his or her claim pro se, which may pose an almost insurmountable task because of the complexity of the undue hardship tests. 89See supra Part I.B. Worse yet, if a debtor is more likely to succeed with the aid of a lawyer, who will presumably only be available to those debtors with more resources, then this system imposes the most unwarranted hardship on the debtors in the worst financial condition. 90See The Real Student-Loan Scandal, supra note 6, at 191–92. Debtors with little to no resources, those who really have the most sympathetic cases, may be more likely to simply not bring a claim because they fear failure without the aid of a lawyer. Interestingly, debtors with the necessary resources to effectively litigate their claim are thus rewarded for declaring bankruptcy with resources in the bank, so to speak, although creditors can point to these resources as evidence that the debtor should not receive discharge. Id. Thus, an adversary proceeding may place a unique burden on the student-loan debtor. In most cases, it is a creditor who files an adversary proceeding to have a debt deemed nondischargeable, but in the context of student loans, it is the debtor who must initiate the proceeding to have the debts deemed dischargeable. 91See 11 U.S.C. § 523(a) (2012) (establishing different categories of debt that, with the exception of § 523(a)(8), can be excepted from discharge and provide motivation only to creditors to initiate proceedings to preserve their rights to payment). The procedural requirements of § 7001(6) may threaten the debtor’s chance at a fresh start and seem to offend the general spirit of the Code toward prebankruptcy debt litigation, thereby sending the message to debtors that they face an uphill battle in their fight to overcome the exception to student-loan discharges.

II. The Bankruptcy Appellate System

Part II.A begins with an overview of the intended function of this Comment. Next, Part II.B addresses the unique structure of the bankruptcy appellate system, followed in Part II.C by an overview of the appellate process. Finally, Part II.D sets out theories of why parties choose to appeal the determination of whether a debtor has met the standard required for a student-loan discharge under the undue hardship provision.

A. The Function of an Appellate Study of Bankruptcy Decisions Concerning § 523(a)(8) Undue Hardship Determinations

Studying data from ten years of bankruptcy appellate cases that relate to the undue hardship provision can hopefully provide some systematical clarity to the application of an ambiguous standard and how well it comports with the general goals of bankruptcy. However, the unique nature of each individual’s bankruptcy case and of § 523(a)(8)’s conditional provision make this data exclusive and difficult to apply to other areas of bankruptcy law. But, there is a gap in the literature that has thus far failed to address the undue hardship determinations. Years ago, the first empirical data was catalogued to study the effects of the application of § 523(a)(8)’s harsh standard. 92See generally Undue Hardship, supra note 6. Follow-up studies have explored the effects of the presence of specific debtor characteristics and other factors that have influenced decisions. 93See generally Pardo, supra note 6 (analyzing the effect of a medical condition in an undue hardship proceeding); The Real Student-Loan Scandal, supra note 6 (analyzing the effects of specific demographic, economic, and other variables in undue hardship proceedings); Iuliano, supra note 6 (analyzing the effects of many variables, including the presence of a medical condition, age, employment, income, etc., in undue hardship proceedings). Each of these studies, though, has only focused on the bankruptcy courts themselves; none have addressed the application of the undue hardship standard in the bankruptcy appellate system. This Comment will attempt to address this next logical step in the study of §523(a)(8) by extrapolating conclusions from the appellate data in this study for three reasons: (1) to make practitioners aware of the prospects they face in each appellate option in the unique structure of the bankruptcy appellate system; (2) to provide the appellate courts with empirical data to show the effects that have been produced by the tests they have adopted and the attitudes and predispositions with which they approach each student-loan debt case they receive; and (3) to address a systemic issue that the data indicates is preventing debtors from accessing the justice that bankruptcy law is intended to provide.

B. The Structure of the Bankruptcy Appellate System

The bankruptcy appellate structure is unique in that it is a two-tiered system. 947 William L. Norton, Jr. & William L. Norton III, Norton Bankr. L. & Prac. § 170:04 (3d ed. 2008). The unique structure of the bankruptcy appellate process “can be traced to congressional reform efforts during the 1970s that sought to improve the quality of the bankruptcy court while simultaneously maintaining it in a subordinate relationship to the district court.” Jonathan Remy Nash & Rafael I. Pardo, An Empirical Investigation into Appellate Structure and the Perceived Quality of Appellate Review, 61 Vand. L. Rev. 1745, 1753 (2008) (footnote omitted). Under the Bankruptcy Act of 1898, many bankruptcy cases were delegated to bankruptcy referees. Id. at 1754. When the Supreme Court created the rules of bankruptcy procedure in 1973 these bankruptcy referees became bankruptcy judges. Id. With the enactment of the Code in 1978, Congress rejected the possibility of vesting bankruptcy judges with Article III status, yet gave them broad jurisdictional powers. Id. at 1755. This choice caused disputes regarding the jurisdictional power of bankruptcy judges to oversee non-core proceedings, which came to a head in Northern Pipeline Construction Co. v. Marathon Pipe Line Co., 458 U.S. 50, 87 (1982). Nash & Pardo, supra, at 1755. In Northern Pipeline, the Supreme Court determined that to the extent that Article III powers had been granted to non-Article III courts (here, the bankruptcy courts), “[s]uch a grant of jurisdiction cannot be sustained as an exercise of Congress’ power to create adjuncts to Art. III courts.” N. Pipeline Constr. Co., 458 U.S. at 87 (alteration in original). In response, the Bankruptcy Amendments and Federal Judgeship Act of 1984 established bankruptcy courts as units of the district court and thereby created a system where a bankruptcy must be filed with a district court, who will then likely have a standing order which refers the case to a bankruptcy judge. Nash & Pardo, supra, at 1755–56. The bankruptcy judge is then empowered to hear and definitively rule upon any of the core proceedings of the bankruptcy case but may only make recommendations as to non-core proceedings, which are then subject to de novo review by the district court if any party objects to such recommendations. Id. at 1756. Final decisions upon these core proceedings by the bankruptcy court are then subject to review by either the district court or BAP with possible review by the court of appeals to follow. Id. Under this system, a debtor may appeal a decision from the bankruptcy court to either the district court or bankruptcy appellate panel (“BAP”). 9528 U.S.C. § 158(a)–(b)(1) (2012). Note, that an appellant may also appeal directly to the circuit court if certain conditions are met. See id. § 158(d). Both the district court and BAP are authorized to independently consider appeals, as a matter of right, from “final judgments, orders and decrees of . . . bankruptcy judges.” 96Id. § 158(a)–(b)(1). Appeals from either of these courts are made to the appropriate circuit court of appeals, 977 Norton & Norton, supra note 94, § 170:04 (2008). after which parties may seek discretionary review by the Supreme Court. 9828 U.S.C. § 1254(1). The first level of appeal is unique to the bankruptcy system, and the savvy appellant must be familiar with the difference between the two options before making the initial decision of choosing an appellate court. The options set before the appellant are depicted below in Figure 1: 99Nash & Pardo, supra note 94, at 1759.

Figure 1: 

Federal Bankruptcy Appellate Structure

Figure 1: Federal Bankruptcy Appellate Structure

Under 28 U.S.C. § 158(b)(1), a judicial council of each circuit has the option to establish a BAP comprised of bankruptcy judges from that circuit, though not all circuits have chosen to do so. 10028 U.S.C. § 158(b)(1). To date, only five circuits (1st, 6th, 8th, 9th, and 10th) have elected to create a BAP. Ben L. Mesches, Bankruptcy Appeals, 45 Tex. J. Bus. L. 107, 128 (2013). Furthermore, Congress has attempted to encourage the creation of more BAPs. See 28 U.S.C. § 158(b)(1) (stating that “[t]he judicial council of a circuit shall establish a bankruptcy appellate panel”). Furthermore, a majority of district judges in the district must vote in the affirmative to empower a BAP to hear appeals from bankruptcy courts. 10128 U.S.C. § 158(b)(6). Where they exist, BAPs are the default forum for an appeal from the bankruptcy court, 102Id. § 158(c)(1). and the panel is comprised of three bankruptcy judges. 103Id. § 158(b)(5) (stating that the members of the panel cannot be from the same district from which the appeal originates).

The stare decisis effects of decisions by district court and BAPs have been an issue of contention for some time. There are four predominant theories as to how to approach the precedential effect of these first-tier appellate decisions. The first approach, endorsed by Coyne v. Westinghouse Credit Corp. (In re Globe Illumination Co.), posits that because the BAPs are a creation of the circuit courts, their decisions have the same binding authority on all bankruptcy courts within their respective circuit as the court of appeals. 104Coyne v. Westinghouse Credit Corp. (In re Globe Illumination Co.), 149 B.R. 614, 621 (Bankr. C.D. Cal. 1993) (“The bankruptcy courts throughout the circuit certainly must be bound by a BAP decision.”). Dicta within this case also asserted that in matters of bankruptcy, the BAP may even have superiority over the district court, although this view was rejected by many district courts asserting that their power as an Art. III court is not trumped by a BAP. See Bank of Maui v. Estate Analysis, Inc., 904 F.2d 470, 472 (9th Cir. 1990); Far W. Fed. Bank v. Vanasen (In re Vanasen), 81 B.R. 59, 62 (D. Or. 1987). The second approach views both district courts and BAPs as intermediate appellate courts separating the bankruptcy court and court of appeals, giving BAP decisions binding precedential control over all bankruptcy courts in the circuit and giving district courts binding precedential control over the bankruptcy courts within their district. 105See In re Windmill Farms, Inc., 70 B.R. 618, 622 (B.A.P. 9th Cir. 1987), rev’d on other grounds, 841 F.2d 1467 (9th Cir. 1988); In re Proudfoot, 144 B.R. 876, 878–79 (B.A.P. 9th Cir. 1992) (citing In re Windmill Farms, 70 B.R. at 622); In re Gen. Associated Investors Ltd. P’ship, 150 B.R. 756, 760–61 (Bankr. D. Ariz. 1993). The third approach holds that both district court and BAP decisions are binding precedent only within the district from which the appeal arose. 106See State Comp. Ins. Fund v. Zamora (In re Silverman), 616 F.3d 1001, 1005 (9th Cir. 2010), cert. denied, 131 S. Ct. 1679 (2011); In re Warren, 91 B.R. 930 (Bankr. D. Or. 1988). The final, and most restrictive, approach takes the position that the district court and BAP decisions have no precedential effect on any court concerning these matters. 107See In re Rheuban 128 B.R. 551 (Bankr. C.D. Cal. 1991). The reasoning behind such a view is that the BAP and district court are on the same level and thus one does not control the other, and because bankruptcy courts are units of the district courts and district court decisions do not bind other district court judges, the bankruptcy courts are not controlled by either appellate court. See Rinard v. Positive Invs., Inc. (In re Rinard), 451 B.R. 12, 21 (Bankr. C.D. Cal. 2011). These different approaches are further evidence of the complexity of the bankruptcy system and the need for a practitioner to be well versed in and aware of the local practices in his or her district.

With the general bankruptcy appellate structure and the precedential nature of both first-tier appellate options in mind, the next important step for an appellant is to consider both the qualitative and quantitative factors that should affect an appellant’s choice at the first-tier appellate level. The results of a 2008 empirical study by Professors Jonathan Nash and Rafael Pardo (the “Nash/Pardo study”) of three years of bankruptcy appellate opinions from both the first- and second-tier appellate levels suggest that: (1) “courts of appeals are more likely to uphold upon review the conclusions of BAPs than district courts” and (2) “BAP decisions are . . . cited more frequently . . . than are district court decisions.” 108Nash & Pardo, supra note 94, at 1747.

Professors Nash and Pardo posited that BAP decisions have a higher level of quality based on two primary observations: (1) objective characteristics of quality appellate review and (2) opinion affirmance and citation rates. 109Id. at 1769–74. The results of their study confirmed that the BAP may indeed provide higher quality appellate review than district courts. 110See id. at 1775 n.112. The Nash/Pardo study based this conclusion on both a comparison of the traditional factors that might indicate quality appellate review and the statistical results from three years of bankruptcy decisions. 111Id. at 1791.

Recognizing the difficulty in addressing every possible factor that might contribute to quality appellate review, Pardo and Nash stated that “the academic literature does suggest several attributes that will tend to contribute to better review,” 112Id. at 1748. and they are as follows: (1) “panels of judges;” (2) “expertise of the appellate decisionmaking body in the subject matter of the appeals it hears;” (3) “general ‘lawfinding ability;’” (4) “the extent to which an appellate court conforms to traditional appellate hierarchy;” and (5) “judicial independence.” 113Id. at 1748–52 (footnotes omitted). The Nash/Pardo study can be consulted for more information concerning the rationale of why these factors are considered to be indicators of quality appellate review.

The Nash/Pardo study concluded that BAPs have more of the abovementioned features of quality appellate review than do district courts. 114Id. at 1759. First, BAPs decide cases in three-judge panels as opposed to a single district court judge deciding a case alone, giving BAPs the inherent advantage of reviewing a case with three judges instead of one. 115Id. Second, because BAP judges are, “by virtue of their appointment as bankruptcy judges,” presumably experts in bankruptcy law, they are better suited to resolve issues in core bankruptcy proceedings than their district court counterparts. 116 Id. at 1759–60. With regard to the third factor—general lawfinding ability—the Nash/Pardo study concluded that neither court has any innate advantage. 117Id. at 1760. Fourth, the study found BAPs conform more to the traditional notion of appellate review than do district courts. 118Id. at 1761–62 (footnote omitted) (stating that BAP “rulings are generally seen to be binding on future bankruptcy appellate panels drawn from the same circuit . . . . In contrast, one district judge is generally not bound to follow the ruling of another district judge . . . .”). Finally, the study concluded that, concerning the factor of judicial independence, district court judges have an advantage over BAPs. 119Id. at 1765. However, the Nash/Pardo study went on to establish that this advantage held by district court judges was slim at best. On the one hand, district court judges are Article III judges, appointed with life tenure and a guarantee of nonreduction in salary. Id. On the other hand, bankruptcy judges receive neither of these benefits as they are not Article III judges. Id. But, Professors Nash and Pardo go on to demonstrate that this perceived inequality is only marginal. Id. First, although bankruptcy judges only serve a fourteen-year term, their appointments can be and often are renewed; few bankruptcy judges leave for lack of reappointment. Id. at 1765–66. Second, the removal process for bankruptcy judges is, in reality, very similar to that of Article III judges and has not produced results that would likely prevent a bankruptcy judge from acting independently. Id. at 1766–67. Third, because the bankruptcy judges’ salaries have been pegged at 92% of the salary of a district court judge since 1987, the bankruptcy judges enjoy a substantially similar fixed compensation as a district court judge. Id. at 1767. Finally, Professors Nash and Pardo surmise that, given the appointment process of bankruptcy judges, they “may be better situated than district judges to resist the political influence that would threaten to compromise an independent judiciary.” Id. In reflection, the BAPs clearly possess more of the academically suggested features of quality appellate review than do their district court counterparts.

The Nash/Pardo study also found that BAPs have a higher perceived level of quality appellate review based on two broad categories of their findings related to affirmances and citations. 120Id. at 1804–05. Findings in both of these categories sought to test the “correctness” of decisions originating from the BAP and the district courts. 121Id. at 1770. Confirming the authors’ hypotheses, the data indicated that BAP decisions were more often affirmed and positively cited, which “tends to validate theoreticians’ claims about the ideal attributes of appellate review.” 122Id. at 1807. This Comment will seek to build on these results and provide further considerations of the Nash/Pardo findings in light of the study’s results.

C. The Bankruptcy Appellate Process

The unique structure of the bankruptcy appellate process can be a difficult hurdle for the student-loan debtor to overcome, but the procedural requirements of this process also present challenges for a debtor. At the inception of a case, original bankruptcy jurisdiction is vested in the U.S. district courts, 12328 U.S.C. § 1334(a)–(b) (2012). which may create standing orders to refer bankruptcy cases to bankruptcy judges within the district. 124Id. § 157(a). These bankruptcy judges are Article I judges appointed by the court of appeals 125Id. § 152(a)(1). and “may hear and determine all cases under title 11 and all core proceedings arising under title 11, or arising in a case under title 11.” 126Id. § 157(b)(1). As such, these judges are empowered to make “determinations as to the dischargeability of particular debts.” 127Id. § 157(b)(2)(I). Bankruptcy courts have the statutory authority to rule definitively on these proceedings, subject to review by an appellate court. 128Id. § 157(b)(1). To obtain a ruling on the dischargeability of a student-loan debt, a debtor must initiate an adversary proceeding at this initial stage of the bankruptcy process by filing a complaint. 129Fed. R. Bankr. P. 7001(6), 7003.

From the final judgment of a bankruptcy court regarding the dischargeability of a debt, an appellant may appeal to either the district court or the BAP. 13028 U.S.C. § 158(a), (b)(1) (authorizing the creation of the BAP and empowering it to hear appeals with consent of the parties). Although the BAP, where one has been established, is the default appellate option from a bankruptcy court, an appellant or any other party may elect to have the appeal heard by the district court. 131Id. § 158(c)(1) (allowing for an election of the district court as long as the party does so within 30 days of notice of appeal). The district court has jurisdiction to hear such appeals under § 158(a). From the final ruling of the BAP or district court, a party may appeal to the appropriate court of appeal. 132Id. § 158(d)(1).

The procedural requirements for an appeal are rigorous for any debtor, but they can be especially difficult for a pro se debtor. 133See Lark v. Bd. of Trs. (In re Lark), No. CC-09-1239, 2010 WL 6451889, at *4 (B.A.P. 9th Cir. Feb. 4, 2010) (dismissing debtor’s case because she failed to provide a transcript for review). An appeal from the final judgment, order, or decree of the bankruptcy court must be filed with the clerk within fourteen days of entry of the ruling. 134Fed. R. Bankr. P. 8001(f)(3)(D). Failure to do so will result in the dismissal of the appeal. 135Id. 8001(a) (“An appellant’s failure to take any step other than timely filing a notice of appeal does not affect the validity of the appeal . . . .”) (emphasis added). Put another way, while other errors in perfecting one’s appeal, such as an untimely designation of the record, do not affect the appeal’s validity, failure to timely file the notice of appeal does affect the appeal’s validity. In addition, within the fourteen day window the appellant must file “a designation of the items to be included in the record on appeal and a statement of the issues to be presented.” 136Id. 8006. This record includes the “items so designated by the parties, the notice of appeal, the judgment, order, or decree appealed from, and any opinion, findings of fact, and conclusions of law of the court,” and, if needed, “a written request for the transcript,” all of which the appellant bears the costs of providing. 137Id. It should be noted from the outset that these requirements can be difficult for a pro se debtor who is unfamiliar with not only the law in general, but the specific practices of the bankruptcy appellate process. In just the cases reviewed for this Comment’s study, many were found to be dismissed because of a procedural error of the debtor-appellant. See Lark, 2010 WL 6451889, at *6 (dismissing debtor’s case because she failed to provide a transcript for review); Hough v. Pa. Higher Educ. Assistance Agency (In re Hough), 128 F. App’x 369 (5th Cir. 2005) (finding the case unable to be examined because pro se debtor failed to provide transcripts); Looper v. Educ. Credit Mgmt. Corp., No. 3:07-cv-306, 2008 WL 2965887, at *4 (E.D. Tenn. July 30, 2008) (dismissing debtor’s case for failure to file a motion to extend the time to file a brief within the fourteen day timeframe).

Having jumped through the procedural hoops to initiate an appeal, a debtor-appellant faces a difficult task to overturn a bankruptcy court’s findings of fact. An appellate court reviews the bankruptcy court’s findings of fact under a clearly erroneous standard. 138Rifino v. United States (In re Rifino), 245 F.3d 1083, 1086 (9th Cir. 2001) (citing United Student Aid Funds (In re Pena), 155 F.3d 1108, 1110 (9th Cir. 1998)); Educ. Credit Mgmt. Corp. v. Frushour (In re Frushour), 433 F.3d 393, 398 (4th Cir. 2005). Furthermore, “where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous.” 139Pa. Higher Educ. Assistance Agency v. Birrane (In re Birrane), 287 B.R. 490, 494 (B.A.P. 9th Cir. 2002) (quoting Anderson v. Bessemer City, N.C., 470 U.S. 564, 574 (1985)). Applying this lenient standard, appellate judges often do not overturn factual findings by the bankruptcy court, even if they would have chosen differently. 140See Pa. Higher Educ. Assistance Agency v. McDay, No. Civ.A.AW 04 2630, 2005 WL 5327774, at *3 (D. Md. Feb. 9, 2005) (noting that the bankruptcy court, as the trial court, is in the best position to judge the credibility of those who testify at the trial). Notably, in the majority of the cases analyzed for this Comment’s study, either the parties did not contest the bankruptcy court’s factual findings or the appellate court quickly stated that the factual findings of the bankruptcy court were not clearly erroneous and thus accepted. See infra Part III.A. The debtor-appellant then must show that the bankruptcy court somehow erred in the calculation of income, expenses, or any other factual bases that might be used to determine whether the debtor met the standard required for a discharge under the undue hardship provision. 141See Walker v. Sallie Mae Servicing Corp. (In re Walker), 650 F.3d 1227, 1230 (8th Cir. 2011) (“We will not upset those findings of fact unless, after reviewing the entire record, we are left with the definite and firm conviction that a mistake has been made.”).

Because the majority of appellate cases do not involve a challenge to the lower court’s factual findings, 142Only 32% of cases in this Comment’s study contained issues of fact. More often than not, the parties did not contest the underlying court’s factual findings and the court then simply accepted the lower court’s findings of facts. See infra Part III.A. the focus at the appellate level is often on the question of whether the debtor satisfied the applicable test to prove that he or she would have an undue hardship without a discharge. Both appellate tiers in the bankruptcy system “review directly the bankruptcy court’s decision.” 143Frushour, 433 F.3d at 398 (stating that the court of appeals reviews directly). As well, both the BAP and district court review the bankruptcy court’s decision directly because they are the first appellate level. Because the determination of whether the facts of a particular debtor’s case meet the requisite standard to prove undue hardship involves applying facts to legal standards, the standard of review, as shown below, must be clearly defined to afford an effective appellate procedure:

Whether a debtor has met the undue hardship standard is a legal conclusion that is based on the debtor’s individual factual circumstances. It is thus a mixed question of law and fact. As we have explained in a related context, these types of questions are reviewed “under a hybrid standard, applying to the factual portion of each inquiry the same standard applied to questions of pure fact and examining de novo the legal conclusions derived from those facts.” . . . We therefore review de novo the determination of whether a debtor has met the undue hardship standard, and we review the factual underpinnings of that legal conclusion for clear error. 144Id. at 398–99 (citations omitted).

The de novo review applied by an appellate court is problematic, though, for an appellant because it places the appellate judges in the same positions as the bankruptcy judge, equally prone to the inherent subjectivity and inconsistency seen in the lower court. 145See discussion supra Part I.B. Because appellate judges review the determination de novo, they might apply a harsher standard to the facts of the case simply because they were not present at the bankruptcy trial. 146See generally Tirch v. Pa. Higher Educ. Assistance Agency (In re Tirch), 409 F.3d 677, 681 (6th Cir. 2005) (declining to affirm a discharge of student-loan debt based partly on the determination that testimony from the debtor at trial as to her medical condition, without supporting documentation or testimony by doctors, was insufficient to satisfy the standard). In addition, where the debtor may feel that they have a decent chance of relief at the appellate level because the case was close at the bankruptcy court level, in reality, appellate courts often cite to congressional intent to make student-loan discharge difficult, to presumptive nondischargeability, and to negative student-loan debtor stereotypes to justify a strict approach to an undue hardship determination. 147See Educ. Credit Mgmt. Corp. v. Jesperson, 571 F.3d 775, 780, 782 (8th Cir. 2009); Educ. Credit Mgmt. Corp. v. Polleys, 356 F.3d 1302, 1306 (10th Cir. 2005). In sum, debtors face many of the same difficulties in appellate review that they first confronted at the bankruptcy court level.

D. Why Appeal?

In light of the difficulties noted above, debtors may be hesitant to appeal decisions of the bankruptcy courts. Theoretically, one might imagine four scenarios in which an appeal may be made based on whether the debtor or creditor is willing to appeal:

Creditor not willing to appeal

Creditor willing to appeal

Debtor not willing to appeal

No appeal occurs

Appeal occurs

Debtor willing to appeal

Appeal occurs

Appeal occurs

Of course, where neither the creditor nor the debtor believes they have a good case, neither party will appeal, thus accepting the court’s determination. However, assuming that a debtor does not want to appeal because they have won in the bankruptcy court, a creditor may appeal for the obvious reason that they believe they should have won in the lower court. Evidence from the cases used in this Comment’s study suggests another possible reason the creditor may appeal. In some of the first tier cases, repeat players, like Educational Credit Management Corp., were observed focusing primarily on trying to get the appellate court to create per se rules, creating a narrower exception for debtors by which they might gain a student-loan discharge. 148See Barrett v. Educ. Credit Mgmt. Corp. (In re Barrett), 487 F.3d 353, 363–64 (6th Cir. 2007) (showing ECMC pushing for a per se rule that not pursuing an Income Contingent Repayment Plan must mean that the debtor fails the good faith prong); Frushour, 433 F.3d at 400 (showing ECMC asking for per se rule that have Internet and cable connections requires the conclusions that a debtor is maintaining more than a minimal standard of living); Polleys, 356 F.3d at 1311 (showing ECMC pushing for per se rule that medical disability is required in order to satisfy the future inability prong); Saxman v. Educ. Credit Mgmt. Corp. (In re Saxman), 325 F.3d 1168 (9th Cir. 2003) (showing ECMC pushing for per se rule that partial discharge through use of 11 U.S.C. § 105(a) is unacceptable); Educ. Credit Mgmt. Corp. v. Smith, No. H-11-57, 2011 WL 4625397 (S.D. Tex. Sept. 30, 2011) (showing ECMC pushing for per se rule that a low paying job alone is not enough to satisfy the requirements of the future inability prong); McLaney v. Ky. Higher Educ. Assistance Auth. (In re McLaney), 375 B.R. 666 (M.D. Ala. 2007) (showing ECMC pushing for a per se rule that tithing is not an allowable expense). Similarly, when the creditor does not want to appeal but the debtor does, it is likely the result that the creditor won in the bankruptcy court, but the debtor believes they have a strong case on which to appeal. Evidence of this can be seen in some of the cases from this Comment’s study where a debtor with a medical condition appealed an unfavorable decision at the bankruptcy court level. 149Prevailing studies suggest that a medical condition significantly increases the chances of satisfying the undue hardship standard, and thus those with medical conditions that lost in the initial case might think they have a greater chance to succeed on appeal; see Pardo, supra note 6, at 511; Roe v. Coll. Access Network (In re Roe), 295 F. App’x 927 (10th Cir. 2008); Nash v. Conn. Student Loan Found. (In re Nash), 446 F.3d 188 (1st Cir. 2006); Educ. Credit Mgmt. Corp. v. Barrett (In re Barrett), 337 B.R. 896 (B.A.P. 6th 2006); State Univ. N.Y.-Student Loan Serv. Ctr. v. Menezes, 352 B.R. 8 (D. Mass. 2006).

However, if the debtor does have a strong case, why might the creditor not choose to simply settle? The likely explanation is that the creditor feels, having already won in the bankruptcy court, that their side is the stronger of the two, but, once again, they may simply move forward with an appeal because of desire to create per se rules with binding precedential authority. Finally, there is the situation where both sides might be willing to appeal, and indeed both sides may cross-appeal. As an appeal requires further financial outlay for already cash-strapped debtors, they will likely only be willing to appeal if they believe they have a strong overall case with the best chance of winning on appeal. Creditors, as repeat players, have significantly more financial resources and are invested in the process as a whole with a view that goes beyond the case-at-hand. 150This means they may make choices that, though they may not prove fruitful in this particular case, may result in favorable rules they can use in later cases in which they know they will take part.

Given the existence of an affirmance bias in appellate courts, 151Nash & Pardo, supra note 94, at 960 (noting that “appellate review standards create an affirmance bias”) (citing Frank B. Cross, Decision Making in the U.S. Courts of Appeals 48–49 (2007)). the reason must indeed be strong for the creditor or debtor to choose to appeal. However, the debtor may proceed simply because they have little to lose. Already in the bankruptcy process, the debtor may just hope for a better result from the de novo review of the bankruptcy court’s undue hardship determination. 152See Frushour, 433 F.3d at 398–99 (explaining the appellate review de novo standard). The creditor, for similar reasons, may also seek to take advantage of the de novo review standard to overturn a bankruptcy court decision that was unfavorable to them. In addition, creditors, as repeat players, may know of the appellate courts’ reputation for strict review of the undue hardship determination 153See Rifino v. United States (In re Rifino), 245 F.3d 1083, 1087 (9th Cir. 2001); McLaney, 375 B.R. at 673; Geyer v. Dep’t of Educ. (In re Geyer), 344 B.R. 129, 131 (S.D. Cal. 2006); Educ. Credit Mgmt. Corp. v. DeGroot, 339 B.R. 201, 206 (D. Or. 2006). and may also proceed in an attempt to create per se rules that would aid them in future cases.

III. The Appellate Study

In this Part, this Comment presents the results from the study performed. This Part will begin first with a discussion of the methodology implemented in the selection of the opinions chosen for analysis in this study. Then, the findings from data gleaned from those opinions will be presented to show the harsh reality that debtors face in the appellate system.

A. Selection Criteria and Methodology

Before presenting the findings of this study, it is important to explain the methodology used to select the opinions that made up the group from which the data for this study was obtained. The purpose of this study was to survey the entire appellate landscape as it related to review of bankruptcy court determinations of whether a debtor had satisfied the undue hardship requirement of § 523(a)(8). Thus, the study attempted to encompass opinions from every circuit and as many districts as possible. This was done in an attempt to provide data that would be representative of judicial doctrine developed across the country. 154It is important to note that by studying only judicial opinions, this study fails to encompass the many cases that are settled or disposed of by other means. As such, this data is a starting point for future studies to better inform practitioners and judges in specific circuits and judicial districts.

In light of these goals, the study began with a search of Westlaw’s FBKR-CS database 155Westlaw is a commercial electronic database that can be accessed via internet at http://www.westlaw.com. with the following terms and connectors: 523(A)(8) & PR(“COURT OF APPEALS” “BANKRUPTCY APPELLATE PANEL” “DISTRICT COURT”) & da(aft 12/31/2001 & bef 01/01/2012). This restricted the results to those opinions issued by courts of appeals, BAPs, and district courts who used the phrase “523(a)(8)” within their opinion, and whose opinions were issued in a ten year date range beginning on January 1, 2002 and ending on December 31, 2011. Theoretically, this narrowed the study’s focus to appellate opinions that addressed the provisions of § 523(a)(8), although it is important to note that an opinion that dealt with these provision but did not cite to the Code would have fallen by the wayside. However, it is unlikely that many opinions of this nature exist. This initial search produced 289 results, all of which were studied regardless of whether the decisions had been published in the Federal Reporter, Federal Appendix, or the Bankruptcy Reporter. The rationale for including these unpublished results was that they still had significant value as they documented the appellate court decision-making process and yielded valuable data to form a more well-rounded view of the average debtor on appeal. After pouring through each of these opinions and coding for the variables used in the study, the result yielded 144 opinions that dealt directly with an undue hardship determination on appeal, while the remaining 145 opinions merely cited to § 523(a)(8) for other purposes.

The opinions that dealt directly with an undue hardship determination represented a fair cross-section of judicial circuits as illustrated in Bar Charts 1 and 2 below:

Bar Chart 1

Bar Chart 2

These charts reflect the representative nature of this study, which allows the following findings to have a wider degree of applicability across the United States.

Each of these opinions was coded for key variables to provide data for analysis. Many of these variables have been previously identified as influential on the decision-making processes for judges. 156See Undue Hardship, supra note 6; Pardo, supra note 6; Rafael I. Pardo & Michelle R. Lacey, The Real Student-Loan Scandal, supra note 6; Iuliano, supra note 6. These variables included: (1) whether the debtor was an appellant; (2) whether there was an issue of law or fact in the case; (3) the debtor’s filing status as joint or individual; (4) whether the debtor was pro se; (5) the debtor’s household income and expenses; (6) the debtor’s employment status; (7) whether the debtor or a dependent had a medical condition; (8) the debtor’s age; (9) the amount of debt at stake in the case; (10) the creditor’s identity; and (11) the basis of the appellate court’s decision. These variables, combined with key identification information, 157This simply includes the case name, appellate court identification, citation, location, and disposition of the case. provide the data necessary to make some key findings as to how the debtor fares in the appellate arena.

B. The Findings from a Study of Undue Hardship Appellate Court Determinations

Appellate courts applying the undue hardship provision have substantially ruled against discharge of student-loan debt. Theoretically, the provision in § 523(a)(8) allows the honest but unfortunate debtor the opportunity to discharge onerous student-loan debt. The Brunner court presumably had this goal in mind when it formally adopted the test to determine dischargeability. 158Brunner v. New York State Higher Education Services Corp., 831 F.2d 395, 396 (2nd Cir. 1987) (per curiam). “Many subsequent courts employing the Brunner analysis, however, appear to have constrained the three Brunner requirements to deny discharge under even the most dire circumstances.” 159Educ. Credit Mgmt. Corp. v. Polleys, 356 F.3d 1302, 1308 (10th Cir. 2005) (emphasis added) (noting that Congress clearly enacted § 523(a)(8) only in an attempt to curb the abuse of the bankruptcy system by these stereotypical debtors with lots of earning potential but high debts they want to get rid of). This raises the question of what message the appellate courts are sending concerning the availability of student-loan debt discharge.

The primary focus of this Comment is to ascertain, based on empirical data, the message being sent from the appellate opinions for a student-loan debtor, especially as it relates to concerns about access to justice. The quote above brings to the forefront this concern about student-loan debtors and the difficulties they face trying to access the protection that the bankruptcy system is intended to provide. Figure 1 illustrates this key concern by showing the lopsided results in favor of the creditor at the appellate level:

Figure 1: Success On Appeal by Appellant Type

Appellant

Success on Appeal

Total

No

Yes

Creditor

41

48

89

46.07%

53.93%

100.00%

Debtor

43

10

53

81.13%

18.87%

100.00%

84

58

142

Total

59.15%

40.85%

100.00%

Pearson chi2 (1) = 16.9039

Pr. = 0.0001

Applying a chi-squared test 160See generally Alan Agresti, Categorical Data Analysis (1990). with one degree of freedom to compare whether there is a correlation between the appellant and their success on appeal, the results revealed a statistically significant finding 161Here, there was less than a 0.0001 probability that random chance alone would have yielded a difference as large as that witnessed between the observed and expected values. It is generally accepted that where the probability is less than .05, or 5%, the finding is statistically significant and thus indicates that there is a correlative relationship between the two variables. that the student-loan debtor often failed to achieve success. Absent a relationship between the appellant and their success on appeal, one would expect to see a debtor-appellant succeed in approximately 40.85% of the cases and fail in 59.15% of the cases. The data reveals that debtors actually succeed in approximately 18.87% of cases and fail in 81.13% of cases. This shocking deviation from the expected results shows that debtors face substantial challenges on appeal.

Seeing that debtors are often unsuccessful in appealing their cases, the question then becomes “why are they unsuccessful?” 162To fully answer this would require expositions of every case in the study, but the purpose of this study was to see if there were any indicators in the empirical data that could answer this question. A major indicator that appellate courts were engaging in an overly harsh view to start is the fact that many cases began their analysis by noting that student loans are “presumptively nondischargeable.” 163See Rifino v. United States (In re Rifino), 245 F.3d 1083, 1087 (9th Cir. 2001); McLaney v. Ky. Higher Educ. Assistance Authority (In re McLaney), 375 B.R. 666, 673 (M.D. Ala. 2007); Educ. Credit Mgmt. Corp. v. DeGroot, 339 B.R. 201, 206 (D. Or. 2006); Geyer v. Dep’t of Educ. (In re Geyer), 344 B.R. 129, 131 (S.D. Cal. 2006). But, using the empirical data, this study attempted to ascertain whether any single factor gave the debtor a better chance on appeal. To accomplish this, the study considered the effects of: (1) debtor representation; (2) the presence of a medical condition; (3) other variables related to household characteristics; (4) geographic considerations; (5) repeat, institutional creditors; and (6) district v. BAP courts. This Subsection will discuss each of these in turn.

1. Debtor Representation

The first consideration was whether debtors were represented by counsel. Hoping to find a correlation between a debtor’s representation and success on appeal, the data was analyzed using chi squared tests. Figures 2 and 3 show the results when the debtor proceeded pro se or represented, respectively:

Figure 2: Pro Se Debtor Success On Appeal

Appellant

Success on Appeal

Total

No

Yes

Creditor

34

39

73

46.58%

53.42%

100.00%

Debtor

22

6

28

78.57%

21.43%

100.00%

Total

56

45

101

55.45%

44.55%

100.00%

Pearson chi2 (1) = 8.3868

Pr. = 0.004

Figure 3: Represented Debtor Success On Appeal

Appellant

Success on Appeal

Total

No

Yes

Creditor

6

7

13

46.15%

53.85%

100.00%

Debtor

21

4

25

84.00%

16.00%

100.00%

Total

27

11

38

71.05%

28.95%

100.00%

Pearson chi2 (1) = 5.9560

Pr. = 0.015

These charts show a statistically significant correlation between the appellant and whether the debtor was pro se or represented. Interestingly, the debtor had substantially more overall success, 39.60% compared to 26.32%, 164Total success rate = (number of cases appealed by creditors that they lost + number of cases appealed by debtors that they won) / total number of cases observed. when he or she was pro se. However, the debtor had only marginally better success as a pro se litigant when the appellant was the debtor, 21.43% compared to 16.00%.

There are two possible explanations for these findings. First, in cases where the debtor was represented, the court might have more carefully reviewed the debtor’s ability to pay because the debtor had the money to pay for counsel. And second, when the pro se debtor was the appellant, they were much more likely to have a graduate-level degree, 165Many times a law degree. which might indicate better earning potential. This earning potential often affected the judge’s perception of whether a particular hardship was likely to endure and, therefore, whether the debtor would fail under the second Brunner prong, which could explain why the findings between pro se and represented debtor-appellants were so similar.

2. Medical Condition

Studies have indicated that the presence of a medical condition plays a large role in a bankruptcy court’s dischargeability finding. 166Pardo, supra note 6, at 509–12 (finding in a study of a decade of bankruptcy decisions that the debtor’s health played the largest role, even eclipsing a debtor’s monthly household income and expenses, in a bankruptcy court’s determination of undue hardship). Theorizing that a medical condition might also play prominently in appellate courts, the above tests were duplicated to see if the presence of a medical condition indicated a higher chance of success for the debtor. Figures 4 167Analysis includes cases that did not mention the debtor or the health of the debtor’s depenedent. and 5 show this analysis:

Figure 4: Success Rate With No Medical Condition Present

Appellant

Success on Appeal

Total

No

Yes

Creditor

17

33

50

34.00%

66.00%

100.00%

Debtor

29

6

35

82.86%

17.14%

100.00%

Total

46

39

85

54.12%

45.88%

100.00%

Pearson chi2 (1) = 19.7920

Pr. = 0.0001

In the absence of a relationship between debtors with no medical condition and their success on appeal, one would expect to see the debtor-appellant to succeed in approximately 45.88% of the cases. The data reveals that debtors with no medical condition succeed in only 17.14% of cases where they are the appellant. Analysis pursuant to a chi-square test with one degree of freedom indicates that there is less than a 0.0001 probability that random chance alone would have yielded a difference as large as that witnessed between the observed and expected values. Debtors with no medical condition also had less success when the creditor appealed compared to the overall results when a creditor appealed in Figure 1. 16834.00% success with no medical condition compared to the overall debtor success rate when creditors appeal, 46.07%.

Figure 5: Success Rate With Medical Condition Present

Appellant

Success on Appeal

Total

No

Yes

Creditor

24

15

39

61.54%

38.46%

100.00%

Debtor

14

4

18

77.78%

22.22%

100.00%

Total

38

19

57

66.67%

33.33%

100.00%

Pearson chi2 (1) = 1.4615

Pr. = 0.227

While the results of Figure 5 did not reveal a statistically significant relationship between the appellant and success rate when the debtor had a medical condition, a comparison of these results with other data reveal a noteworthy finding. In Figure 1, the results showed that when debtors appealed they succeeded in 18.87% of cases, and when creditors appealed, they succeeded in 53.93% of cases. As well, in Figure 4, the results showed that when debtors have no medical condition, they succeed in only 17.14% of cases, whereas when creditors appealed, they succeeded in 66.00% of cases. In Figure 5, though, when debtors have a medical condition, they succeeded in 22.22% of cases in which they were the appellant, and creditors succeeded in only 38.46% of cases in which they were the appellant.

These results indicate that, while the debtor is not guaranteed to succeed, the presence of a medical condition seemed to level the playing field. In this situation, creditors lose the substantial advantage they hold in most other circumstances. 169In each situation, the observed results were much closer to the expected values. Furthermore, there is evidence from the cases studied that suggests that some debtors with medical conditions could have succeeded with relatively small changes in their approach. For example, some debtors with medical conditions failed on appeal because the only evidence of their condition and its effect was their own testimony, something that could have been easily remedied by the production of supportive documentation by doctors. 170These debtors only provided prognoses of the outlook of their condition based on their own testimony and did not produce medical experts, or doctors, to testify on their behalf as to the prospects of their conditions going forward. See Roe v. Coll. Access Network (In re Roe), 295 F. App’x 927 (10th Cir. 2008); Nash v. Conn. Student Loan Found. (In re Nash), 446 F.3d 188 (1st Cir. 2006); State Univ. N.Y.-Student Loan Serv. Ctr. v. Menezes, 352 B.R. 8 (D. Mass. 2006); Barrett v. Educ. Credit Mgmt. Corp. (In re Barrett), 337 B.R. 896 (6th B.A.P. 2006).

3. Household Characteristics

Theorizing that household characteristics might provide explanation for the negative student-loan debtor appellate results, this study focused on income, expenses, and average debt as a possible predictor of appellate outcomes. However, using summary statistics and logistic regressions, the data showed there was no statistically significant relationship between any of these variables and success on appeal. 171See infra Appendix, Tables A1–15. The fact that there was no statistically significant relationship is important in and of itself in that it shows that debtors across the spectrum received relatively consistent treatment. This is notable because it reveals that this area of bankruptcy law has attained consistency in one aspect—its equally harsh treatment of student-loan debtors, no matter their financial status.

4. Geography

Next, the study catalogued and analyzed geographical information to see if there was any correlation between where the debtor’s case originated and success on appeal. Tables A16–26 in the Appendix show the results of applying two different bivariate statistical tests 172Fischer’s exact and chi squared tests. See generally Agresti, supra note 160. These tests are performing the same analysis as originally shown in Figure 1 for overall success rates on appeal. to assess the debtor’s success rate by circuit. Though debtors seemed to fare better in some circuits than in others, 173It is noteworthy that in the 8th Circuit, which applies the more lenient totality of circumstances test, debtors in the cases studied actually performed terribly when they were the appellant. See infra Appendix, Table A23 (showing that debtors succeeded in 0.00% of cases). However, when the creditor was the appellant, the debtor succeeded in 71.43% of cases studied, possibly indicating that under the totality of the circumstances test, courts may be more likely to uphold decisions that favor the debtor. Id. none of the above tests revealed any statistical significance to the success rate findings, thus indicating that geography is not necessarily important in determining whether or not the debtor will find success. In sum, the student-loan debtor—no matter his or her income expenses, or location—faces a system that consistently rules in favor of the creditor more often and is more willing to overturn decisions favorable to the debtor.

5. Institutional Creditors

The study next sought to determine the effect of key repeat creditors and how their presence in a case affected the debtor’s prospects on appeal. An analysis of the creditors in each case with full dispositions 174There are a total of 142 cases where there is either a “full affirm” or “no affirm” disposition. revealed two standouts: Educational Credit Management Corporation (“ECMC”), 175“ECMC is a national guaranty agency, a national student loan bankruptcy servicer, and a contractor for the U.S. Department of Education, providing collection services and document management services.” Educ. Credit Mgmt. Corp., ECMC–FAQ–General, ECMC, http://www.ecmc.org/details/faqGeneral.html#question3. who made up 67.61% of the cases, and the U.S. Department of Education (“USDOE”), who comprised 11.27% of the cases. Figures 6 and 7 show the effect of the presence of ECMC in a case:

Figure 6: Appeals by Debtors

Appellee

Success on Appeal

Total

No

Yes

Other Creditor

20

2

22

90.91%

9.09%

100.00%

ECMC

23

8

31

74.19%

25.81%

100.00%

Total

4381.13%

1018.87%

53 100.00%

Fisher’s exact = 0.166 1-sided Fisher’s exact = 0.118

Figure 7: Appeals by Creditors

Appellant

Success on Appeal

Total

No

Yes

Other Creditor

13

11

24

54.17%

45.83%

100.00%

ECMC

28

37

65

43.08%

56.92%

100.00%

Total

41

48

89

46.07%

53.93%

100.00%

Pearson chi2(1) = 0.8676

Pr = 0.352

The above chi squared and Fisher’s exact tests reveal that whether in debtor appeals or creditor appeals, ECMC’s success rate is not statistically significantly different than the success rates of other creditors. Based on its experience as a significant repeat player, I expected to see ECMC’s performance to be statistically significantly better than other creditors, but the results above show that while ECMC fared better than other creditors when the creditor in a case was the appellant, the difference in performance had no statistical significance.

Figures 8 and 9 reflect the application of a Fisher’s exact test to determine the effects when the USDOE was the creditor in a case:

Figure 8: Appeals by Debtors

Appellee

Success on Appeal

Total

No

Yes

Other Creditor

36

8

44

81.82%

18.18%

100.00%

USDOE

7

2

9

77.78%

22.22%

100.00%

Total

4381.13%

1018.87%

53 100.00%

Fisher’s exact = 1.000 1-sided Fisher’s exact = 0.546 

Figure 9: Appeals by Creditors

Appellant

Success on Appeal

Total

No

Yes

Other Creditor

35

47

82

42.68%

57.32%

100.00%

USDOE

6

1

7

85.71%

14.29%

100.00%

Total

41

48

89

46.07%

53.93%

100.00%

Fisher’s exact = 0.045 1-sided Fisher’s exact = 0.546

The above Fisher’s exact tests reveal that, in debtor appeals, although the debtor succeeds more often against the USDOE than other creditors, the result is not statistically significant. On the other hand, in creditor appeals, other creditors succeed more often than the USDOE (57.3% compared to 14.29%), and the difference is statistically significant. Thus, when the USDOE appeals, the debtor is likely to fare better than when another creditor appeals. As noted in Figure 9, the USDOE was the appellant in only seven cases during the ten-year period covered in this study. If this small number is indicative of the general role of the USDOE in student-loan discharge cases, it would seem that, with less experience, the USDOE would be likely to fare worse in litigation when compared to private entities. Put another way, the USDOE’s poor performance when it appeals might be the product of an inferior ability to assess appellate opportunities relative to other creditors who may be more familiar with the system or have other driving incentives. 176For example, consider that a major subset of the eighty-two creditor appeals in the “Other Creditor” row in Figure 9 involved ECMC. See supra Figure 9. More specifically, ECMC was involved in sixty-five of the eighty-nine creditor appeals, and ECMC prevailed in approximately 57% of those appeals. See supra Figure 7. Thus, the statistically significant result in Figure 9—that is, that other creditors fare better than the USDOE in creditor appeals—possibly suggests that repeat players are more likely to succeed than non-repeat players. Of course, this sample of seven cases is relatively small and begs the question of whether the statistically significant relationship has substantive significance. While this question is beyond the scope of this Comment, this area would appear to be fertile ground for future research.

6. District Court v. BAP

Finally, the last major focus of the study is to compare the success rates of student-loan debtors in the District Courts and BAPs. Based on the conclusions of the Nash/Pardo study, 177See supra Part II.B. I hypothesized that because the BAPs are courts that deal specifically with bankruptcy cases 178This experience means that bankruptcy judges are likely to have significantly more expertise in this area. and District Courts only deal sporadically with bankruptcy cases, 179Bankruptcy cases only encompass a small percentage of the types of cases heard by district courts. debtors will more likely be on a level playing field in BAPs. Two reasons suggest this outcome. First, judges on the BAPs have repeated exposure to the common players, like ECMC, in student-loan debt discharge proceedings. This familiarity allows these judges to be more sensitive to ECMC’s and others’ litigation techniques. Essentially, their familiarity with common creditor tactics helps them to more effectively hear the case without being swayed by extra-legal factors. Second, because judges on BAPs are bankruptcy court judges themselves, they have greater experience applying the undue hardship provision firsthand. Therefore, because the playing field is more level in the BAP, the likely result is that BAPs will produce results more favorable to the debtor than the results observed in the overall success on appeal data from Figure 1. Figures 10 and 11 show the different success rates of the debtor in both the District Courts and BAPs:

Figure 10: Success on Appeal in the District Courts

Appellant

Success on Appeal

Total

No

Yes

Creditor

20

27

47

42.55%

57.45%

100.00%

Debtor

21

2

23

91.30%

8.70%

100.00%

Total

41

29

70

58.57%

41.43%

100.00%

Pearson chi2(1) = 15.1256

Pr = 0.0001

Figure 11: Success on Appeal in the BAPs

Appellant

Success on Appeal

Total

No

Yes

Creditor

12

11

23

52.17%

47.83%

100.00%

Debtor

5

2

7

71.43%

28.57%

100.00%

Total

17

13

30

56.67%

43.33%

100.00%

Pearson chi2(1) = 0.8103

Pr = 0.368

These results show the district court is a place of little hope for the student-loan debtor. This data indicates a statistically significant finding in the correlation between the appellant and their success on appeal with less than a 0.0001 probability that such finding was based on random chance alone. In the district court, debtors’ success, whether or not they were the appellant, was well below the overall averages shown in Figure 1. 180As debtor-appellee, compare 42.55% in the district court to 46.07% in the overall findings from Figure 1. As debtor-appellant, compare 8.70% in the district court to 18.87% in the overall findings from Figure 1. Although the findings for debtors’ success on appeal in BAPs did not fall within the statistically significant range, they do show that in the cases analyzed, debtors fared much better in the BAP than in the district court.

Finally, Figure 12 shows the success rate of the debtor at the court of appeals level:

Figure 12: Success on Appeal in the Courts of Appeals

Appellant

Success on Appeal

Total

No

Yes

Creditor

9

10

19

47.37%

52.63%

100.00%

Debtor

17

6

23

73.91%

26.09%

100.00%

Total

26

16

42

61.90%

38.10%

100.00%

Pearson chi2(1) = 3.1088

Pr = 0.078

Although the results of this analysis approached statistical significance (p = 0.078), it is interesting to note that in both the courts of appeals and the BAPs the debtor fared considerably better than in the district court. Assuming that results more favorable to the student-loan debtor than the unbalanced results from Figure 1 are theoretically more in line with the Code’s intent, this analysis bolsters the claim of the Nash/Pardo study that appellate panel courts like the court of appeals and the BAP provide a higher level of quality appellate review. 181See Nash & Pardo, supra note 94, at 1747.

Looking to further compare the results of this study to the findings from the Nash/Pardo study of appellate review, Figure 13 illustrates the success rates of debtors on appeal to circuit courts. Here, success for the debtor is defined as either a case where the debtor appealed and the lower court’s decision was overturned or when a creditor appealed and the lower court’s decision was either fully or partly affirmed. Figure 13 is a compilation of findings from forty-two observed cases in the different courts of appeals, thirty of which were appealed from the district courts and twelve of which were appealed from BAPs:

Figure 13: Success Rate for Debtors in Circuit Courts

Court of Appeals

Creditor Appeals from BAPs

37.50%

Debtor Appeals from BAPs

25.00%

Total Appeals from BAPs

33.33%

Overall BAP Affirmance

50.00%

Creditor Appeals from DCs

54.55%

Debtor Appeals from DCs

26.32%

Total Appeals from DCs

36.67%

Overall DC Affirmance

66.67%

Overall Affirmance

61.90%

This table shows that the courts of appeals affirmed 61.90% of all undue hardship decision brought under its purview. However, the affirmance rate was considerably higher on decisions appealed from district courts. 182Compare 66.67% affirmance rate in appeals from district courts to 50% affirmance rate in appeals from BAPs. Interestingly though, the debtor seemed to fare better in the courts of appeals when the appeal originated from a district court. These figures, though they do not directly bolster the results of the Nash/Pardo study, possibly indicate support for the value of panel review. Assuming once more that decisions which favor the student-loan debtor more than the unbalanced results from Figure 1 are theoretically more in line with what the Code intended, then higher success rates for debtors in appeals from district courts are evidence of the circuit courts leveling out the playing field, or at least bringing the results of the district court, where debtors fared poorly, 183See supra Figure 10. more on par with the results from the BAPs. 184See supra Figure 11. This assertion is reinforced by the fact that the results in Figures 11 and 12 are substantially the same.

Conclusion

The observations presented here evidence a trend in appellate courts of applying a progressively harsher standard of review in undue hardship determinations. Increasingly limited by precedent, courts applying the two tests find they have little choice but to find student-loan debts nondischargeable. The findings in this Comment show that many of the characteristics presumed to forecast success have little impact as to the debtor’s success on appeal. The presence of a medical condition, though, appears to still be a differentiating factor, often allowing the debtor a more equitable opportunity to win his case. It still seems, though, that the appellate courts have chosen to take a consistent, but harsh approach to interpreting whether or not a debtor has satisfied the undue hardship test.

For practitioners, this finding means taking a careful look to see whether it is wise to move forward on an appeal, or if appeal is made, whether it might be wise to take a settlement offer. Furthermore, this study may indicate that if an appeal is chosen, then it might be wise for a debtor to go to a BAP, if possible. 185Note that there is still an opt out choice for either party to choose to go to a district court instead of a BAP. 11 U.S.C. § 158(c)(1) (2012). For courts, this study calls for some reflection on why such a harsh standard is being applied at the appellate level. Though student loans are unique in their difficulty to discharge, one court, at least, has noted:

Despite Congress’ proven willingness to amend § 523(a)(8) to further restrict the dischargeability of student loan debt, and in spite of the availability of the ICRP option for repayment of student loans, Congress chose to continue the undue hardship exception to nondischargeability. Indeed, provisions in the House and Senate bills designed to make student loans conclusively nondischargeable were stripped from the bills in favor of maintaining § 523(a)(8)’s undue hardship exception to nondischargeability of student loan debt. 186Educ. Credit Mgmt. Corp. v. Jesperson, 571 F.3d 775, 787–88 (8th Cir. 2009).

The continued presence of the availability of discharge for student-loan debtors suggests that courts should take a second look at their approach to undue hardship determinations to see if they are really conforming to the Code’s intent. If Congress has not been willing to do away with the option of dischargeability for student-loan debt, then perhaps the lopsided results from this study indicate a system in need of reexamination, and, ultimately, a more lenient approach to student-loan debt discharge.

Appendix

Table A1

Debtor-Appellant Income Distribution When Appeal Outcome Was Full/Part Affirm

Percentiles

Smallest

1%

124.17

124.17

5%

124.17

847.36

10%

846.36

1000

Obs

15

25%

1120

1120

Sum of Wgt.

15

50%

1680.47

Mean

2064.193

Largest

Std. Dev.

1137.134

75%

3300

3300

90%

3642.48

3419

Variance

1293074

95%

3666.67

3642.48

Skewness

0.13167

99%

3666.67

3666.67

Kurtosis

1.79294

Table A2

Debtor-Appellant Income Distribution When Appeal Outcome Was Full Reversal

Percentiles

Smallest

1%

891

891

5%

891

1063.21

10%

891

1401

Obs

8

25%

1232.105

1459

Sum of Wgt.

8

50%

1529.5

Mean

1586.589

Largest

Std. Dev.

517.1514

75%

1974.665

1600

90%

2329.17

1650

Variance

267445.6

95%

2329.17

2299.33

Skewness

0.31294

99%

2329.17

2329.17

Kurtosis

1.99209

The two tables above set forth summary statistics for the distribution of the debtor’s income based on the outcome of the appeal for those cases where the debtor was the appellant (i.e., debtorapp==1). Thus, for cases where the debtor was the appellant and the court fully or partly affirmed the judgment below, the median income was $1,680.47 and the mean income $2,064.19. For cases where the debtor was the appellant and the court fully reversed the judgment below (i.e., a complete win for the debtor), the median income was $1,529.50 and the mean income was $1,586.59.

Table A3

Creditor-Appellant Income Distribution When Appeal Outcome Was Full/Part Affirm

Percentiles

Smallest

1%

405

405

5%

868

868

10%

946

946

Obs

26

25%

1132

991

Sum of Wgt.

26

50%

1827.5

Mean

2194.355

Largest

Std. Dev.

1402.654

75%

3300

3842.04

90%

4355.48

4355.48

Variance

1967439

95%

5264.58

5264.58

Skewness

0.968103

99%

5500

5500

Kurtosis

2.926138

Table A4

Creditor-Appellant Income Distribution When Appeal Outcome Was Full Reversal

Percentiles

Smallest

1%

700

700

5%

760

760

10%

946

918

Obs

33

25%

1143

946

Sum of Wgt.

33

50%

2030.72

Mean

2087.118

Largest

Std. Dev.

853.8859

75%

2524

3208

90%

3208

3300

Variance

729121.1

95%

3574.56

3574.56

Skewness

0.292728

99%

4033.33

4033.33

Kurtosis

2.389843

The two tables above set forth summary statistics for the distribution of the debtor’s monthly income based on the outcome of the appeal for those cases where the creditor was the appellant (i.e., debtorapp==0). Thus, for cases where the creditor was the appellant and the court fully or partly affirmed the judgment below, the median income was $1,827.50 and the mean income $2,194.36. For cases where the creditor was the appellant and the court fully reversed the judgment below (i.e., a complete win for the creditor), the median income was $2,030.72 and the mean income was $2,087.12.

Table A5

Logistic Regression Model for Effect of Income on Appellate Decision

Number of Obs =

82

LR chi2 (1) =

0.44

Prob > chi2 =

0.507

Log Likelihood = -56.618 

Pseudo R2 =

0.0039

appoutcome

Coef.

Std. Err.

Z

P>|z|

[95% Conf. Interval]

Income

-0.0001

0.00021

-0.66

0.509

-0.0005

0.00027

_cons

0.283

0.48182

0.59

0.557

-0.6613

1.22736

The table above indicates that the debtor’s income (income) is not a statistically significant predictor of the outcome on appeal (appoutcome), z = -0.66, p = 0.509. The table results are from a logistic regression, with the appeal outcome (appoutcome) as the dependent variable and the income of the debtor (income) as the independent (i.e., predictor) variable.

Table A6

Debtor-Appellant Expenses Distribution When Appeal Outcome Was Full/Part Affirm

Percentiles

Smallest

1%

492

492

5%

492

1100

10%

1100

1300

Obs 13

25%

1619.27

1619.27

Sum of Wgt. 13

50%

1829.39

Mean

2024.025

Largest

Std. Dev.

859.4745

75%

2600

2600

90%

3387

2771

Variance

738696.3

95%

3416.67

3387

Skewness

0.138138

99%

3416.67

3416.67

Kurtosis

2.348373

Table A7

Debtor-Appellant Expenses Distribution When Appeal Outcome Was Full Reversal

Percentiles

Smallest

1%

1074

1074

5%

1074

1161

10%

1074

1575

Obs

6

25%

1161

1785

Sum of Wgt.

6

50%

1680

Mean

1686.537

Largest

Std. Dev.

517.3519

75%

2229.17

1575

90%

2229.05

1785

Variance

267653

95%

2229.05

2229.17

Skewness

0.015347

99%

2229.05

2229.05

Kurtosis

1.477827

The two tables above set forth summary statistics for the distribution of the debtor’s expenses based on the outcome of the appeal for those cases where the debtor was the appellant (i.e., debtorapp==1). Thus, for cases where the debtor was the appellant and the court fully or partly affirmed the judgment below, the median debtor had monthly expenses of $1,829.39 and the mean debtor had monthly expenses of $2,024.03. For cases where the debtor was the appellant and the court fully reversed the judgment below (i.e., a complete win for the debtor), the median debtor had monthly expenses of $1,680 and the mean debtor had monthly expenses of $1,686.54.

Table A8

Creditor-Appellant Expenses Distribution When Appeal Outcome Was Full/Part Affirm

Percentiles

Smallest

1%

830

830

5%

948

948

10%

1172

1172

Obs

21

25%

1404

1234

Sum of Wgt.

21

50%

2111

Mean

2675.894

Largest

Std. Dev.

1591.528

75%

3575

4327.47

90%

5171

5171

Variance

2532962

95%

5913

5913

Skewness

0.832385

99%

5913

5913

Kurtosis

2.526385

Table A9

Creditor-Appellant Expenses Distribution When Appeal Outcome Was Full Reversal

Percentiles

Smallest

1%

750

750

5%

830

830

10%

1022

1018.66

Obs

34

25%

1686.31

1022

Sum of Wgt.

34

50%

2328

Mean

2217.661

Largest

Std. Dev.

804.9414

75%

2619

3259

90%

3259

3626

Variance

647930.6

95%

3657

3657

Skewness

0.1090852

99%

3942

3942

Kurtosis

2.531794

The two tables above set forth summary statistics for the distribution of the debtor’s expenses based on the outcome of the appeal for those cases where the creditor was the appellant (i.e., debtorapp==0). Thus, for cases where the creditor was the appellant and the court fully or partly affirmed the judgment below, the median debtor had expenses of $2,111 and the mean debtor had monthly expenses of $2,675.89 For cases where the creditor was the appellant and the court fully reversed the judgment below (i.e., a complete win for the creditor), the median debtor had expenses of $2,328 and the mean debtor had expenses of $2,217.66.

Table A10

Logistic Regression Model for Effect of Expenses on Appellate Decision

Number of Obs =

74

LR chi2 (1) =

1.29

Prob > chi2 =

0.2568

Log Likelihood = -50.406495

Pseudo R2 =

0.0126

Coef.

Std. Err.

z

P>|z|

[95% Conf. Interval]

Expenses

-0.0002448

0.0002195

-1.11

0.265

-0.0006751

0.0001855

_cons

0.719345

0.5514051

1.3

0.192

-0.3613891

1.800079

The table results are from a logistic regression, with the appeal outcome (appoutcome) as the dependent variable and the expenses of the debtor (expenses) as the independent (i.e., predictor) variable. The table above indicates that the amount of the debtor’s expenses (expenses) is not a statistically significant predictor of the outcome on appeal (appoutcome), z = -1.11, p = 0.265.

Table A11

Debtor-Appellant Average Debt When Appeal Outcome Was Full/Part Affirm

Percentiles

Smallest

1%

3000

3000

5%

11183.4

11183.4

10%

21865.3

16312.34

Obs

37

25%

44500

21865.3

Sum of Wgt.

37

50%

68945.3

Mean

79759.5

Largest

Std. Dev.

56498.73

75%

100000

142000

90%

142000

161000

Variance

3.19E+09

95%

181743.4

181743.4

Skewness

1.714819

99%

297953

297953

Kurtosis

7.217371

Table A12

Debtor-Appellant Average Debt When Appeal Outcome Was Full Reversal

Percentiles

Smallest

1%

27781

27781

5%

27781

45000

10%

27781

70000

Obs

8

25%

57500

73000

Sum of Wgt.

8

50%

77287.5

Mean

108708.5

Largest

Std. Dev.

105596.8

75%

106283.5

81575

90%

359745

85000

Variance

1.12E+10

95%

359745

127567

Skewness

1.932742

99%

359745

359745

Kurtosis

5.29914

The two tables above set forth summary statistics for the distribution of the debtor’s debt based on the outcome of the appeal for those cases where the debtor was the appellant (i.e., debtorapp==1). Thus, for cases where the debtor was the appellant and the court fully or partly affirmed the judgment below, the median debtor had average debt of $68,945.30 and the mean debtor had average debt of $79,759.50. For cases where the debtor was the appellant and the court fully reversed the judgment below (i.e., a complete win for the debtor), the median debtor had average debt of $77,287.50 and the mean debtor had average debt of $108,708.50.

Table A13

Creditor-Appellant Average Debt When Appeal Outcome Was Full/Part Affirm

Percentiles

Smallest

1%

15662

15662

5%

16215.08

16215.08

10%

20000

18500

Obs

37

25%

38978.2

20000

Sum of Wgt.

37

50%

63109.87

Mean

86133.69

Largest

Std. Dev.

75888.02

75%

100000

240000

90%

240000

283354.5

Variance

5.76E+09

95%

283354.5

283354.5

Skewness

1.733684

99%

304463.6

304463.6

Kurtosis

5.205519

Table A14

Creditor-Appellant Average Debt When Appeal Outcome Was Full Reversal

Percentiles

Smallest

1%

7038.21

7038.21

5%

11715.25

11293.68

10%

18263.4

11715.25

Obs

45

25%

33000

12148.7

Sum of Wgt.

45

50%

61800

Mean

75266.94

Largest

Std. Dev.

61293.43

75%

89832.16

16000

90%

150782.9

217920

Variance

3.76E+09

95%

217920

235000

Skewness

1.803723

99%

304463.6

304463.6

Kurtosis

6.671375

The two tables above set forth summary statistics for the distribution of the debtor’s debt based on the outcome of the appeal for those cases where the creditor was the appellant (i.e., debtorapp==0). Thus, for cases where the creditor was the appellant and the court fully or partly affirmed the judgment below, the median debtor had average debt of $63,109.87 and the mean debtor had average debt of $86,133.69. For cases where the creditor was the appellant and the court fully reversed the judgment below (i.e., a complete win for the creditor), the median debtor had average debt of $61,000 and the mean debtor had average debt of $75,266.94.

Table A15

Logistic Regression Model for Effect of Average Debt on Appellate Decision

Number of Obs =

127

LR chi2 (1) =

0.05

Prob > chi2 =

0.8274

Log Likelihood = -86.261702

Pseudo R2 =

0.0003

Coef.

Std. Err.

z

P>|z|

[95% Conf. Interval]

Debt

-5.86E-07

2.70E-06

-0.22

0.828

-5.87E-06

4.70E-06

_cons

-2.859136

0.2838156

-1.01

0.314

-0.8421819

0.2703548

The table results are from a logistic regression, with the appeal outcome (appoutcome) as the dependent variable and the amount of debt (debt) as the independent (i.e., predictor) variable. The amount of debt is not a statistically significant predictor of the outcome on appeal, z = -0.22, p = 0.828.

Table A16

Success on Appeal, 1st Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

3

6

9

33.33%

66.67%

100.00%

Debtor

3

0

3

100.00%

0.00%

100.00%

Total

6

6

12

50.00%

50.00%

100.00%

Fisher’s exact = 0.182 1-sided Fisher’s exact = 0.091

Table A17

Success On Appeal, 2nd Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

1

1

2

50.00%

50.00%

100.00%

Debtor

4

0

4

100.00%

0.00%

100.00%

Total

5

1

6

83.33%

16.67%

100.00%

Fisher’s exact = 0.333 1-sided Fisher’s exact = 0.333

Table A18

Success on Appeal, 3rd Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

2

0

2

100.00%

0.00%

100.00%

Debtor

3

1

4

75.00%

25.00%

100.00%

Total

5

1

6

83.33%

16.67%

100.00%

Fisher’s exact = 1.000 1-sided Fisher’s exact = 0.667

Table A19

Success on Appeal, 4th Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

3

10

13

23.08%

76.92%

100.00%

Debtor

5

4

9

55.56%

44.44%

100.00%

Total

8

14

22

36.36%

63.64%

100.00%

Fisher’s exact = 0.187 1-sided Fisher’s exact = 0.135

Table A20

Success on Appeal, 5th Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

1

3

4

25.00%

75.00%

100.00%

Debtor

3

0

3

100.00%

0.00%

100.00%

Total

4

3

7

57.14%

42.86%

100.00%

Fisher’s exact = 0.143 1-sided Fisher’s exact = 0.114

Table A21

Success on Appeal, 6th Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

8

5

13

61.54%

38.46%

100.00%

Debtor

3

0

3

100.00%

0.00%

100.00%

Total

11

5

16

68.75%

31.25%

100.00%

Fisher’s exact = 0.509 1-sided Fisher’s exact = 0.295

Table A22

Success on Appeal, 7th Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

2

4

6

33.33%

66.67%

100.00%

Debtor

4

1

5

80.00%

20.00%

100.00%

Total

6

5

11

54.55%

45.45%

100.00%

Fisher’s exact = 0.242 1-sided Fisher’s exact = 0.175

Table A23

Success on Appeal, 8th Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

10

4

14

71.43%

28.57%

100.00%

Debtor

3

0

3

100.00%

0.00%

100.00%

Total

13

4

17

76.47%

23.53%

100.00%

Fisher’s exact = 0.541 1-sided Fisher’s exact = 0.421

Table A24

Success on Appeal, 9th Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

5

8

13

38.46%

61.54%

100.00%

Debtor

11

4

15

73.33%

26.67%

100.00%

Total

16

12

28

57.14%

42.86%

100.00%

Fisher’s exact = 0.125 1-sided Fisher’s exact = 0.069

Table A25

Success on Appeal, 10th Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

4

2

6

66.67%

33.33%

100.00%

Debtor

3

0

3

100.00%

0.00%

100.00%

Total

7

2

9

77.78%

22.22%

100.00%

Fisher’s exact = 0.500 1-sided Fisher’s exact = 0.417

Table A26

Success on Appeal, 11th Circuit

Appellant

Success on Appeal

Total

No

Yes

Creditor

2

5

7

28.57%

71.43%

100.00%

Debtor

1

0

1

100.00%

0.00%

100.00%

Total

3

5

8

37.50%

62.50%

100.00%

Fisher’s exact = 0.375 1-sided Fisher’s exact = 0.375

Footnotes

1The Quotable Edison 27 (Michele Wehrwein Albion ed., 2011).

2See Martin V. Melosi, Thomas A. Edison and the Modernization of America 65–66 (Mark C. Carnes ed., 2008). The first public demonstration of electric light was made in 1808 when Sir Humphrey Davy unveiled his arc lighting invention, which had little commercial application until the late 1870s. See id. at 65. Furthermore, Davy produced incandescence in 1802, but hundreds of experiments following this finding failed to produce the type of lasting filament that Edison achieved. See id. at 66.

3See id. at 65–66.

4See id. at 70.

5The Quotable Edison, supra note 1, at 10 .

6See Jason Iuliano, An Empirical Assessment of Student Loan Discharges and the Undue Hardship Standard, 86 Am. Bankr. L.J. 495 (2012); Rafael I. Pardo, Illness and Inability to Repay: The Role of Debtor Health in the Discharge of Educational Debt, 35 Fla. St. U. L. Rev. 505 (2008); Rafael I. Pardo & Michelle R. Lacey, The Real Student-Loan Scandal: Undue Hardship Discharge Litigation, 83 Am. Bankr. L.J. 179 (2009) [hereinafter The Real Student-Loan Scandal]; Rafael I. Pardo & Michelle R. Lacey, Undue Hardship in the Bankruptcy Courts: An Empirical Assessment of the Discharge of Educational Debt, 74 U. Cin. L. Rev. 405 (2005) [hereinafter Undue Hardship].

7See, e.g., Peter Coy, Student Loans: Debt for Life, Bloomberg Businessweek (Sept. 18, 2012), http://www.businessweek.com/articles/2012-09-06/student-loans-debt-for-life#p1; Ron Lieber, Last Plea on School Loans: Proving a Hopeless Future, N.Y. Times (Aug. 31, 2012), http://www.nytimes.com/2012/09/01/business/shedding-student-loans-in-bankruptcy-is-an-uphill-battle.html?pagewanted=all; Steve Rhode, American Households Sinking in Student Loan Debt Says Pew Research Center, Huffington Post (Oct. 7, 2012), http://www.huffingtonpost.com/steve-rhode/american-households-sinki_b_1933143.html.

8Terry Savage, Terry Savage: Federal Student Loans Next Big Crisis, Chicago Sun-Times, (Oct. 13, 2012), http://www.suntimes.com/business/savage/15755472-452/terry-savage-federal-student-loans-next-big-crisis.html.

9Brendan Baker, Deeper Debt, Denial of Discharge: The Harsh Treatment of Student Loan Debt in Bankruptcy, Recent Developments, and Proposed Reforms, 14 U. Pa. J. Bus. L. 1213, 1214 (2012).

10See 11 U.S.C. § 523(a)(8) (2012) (requiring the debtor prove that the student-loan debt would impose an undue hardship on the debtor if not discharged).

11See H.R. Rep. No. 95-595, at 536–38 (1977), reprinted in 1978 U.S.C.C.A.N. 5963, 6425 (insisting that the average student debtor is trying to take advantage of a loophole in the system). Professors Rafael Pardo and Michelle Lacey, in a 2005 study, addressed this abusive student-loan debtor stereotype thoroughly and concluded that the justification of the change was built on bombastic and factually incorrect assertions. See Undue Hardship, supra note 6, at 419–32.

12H.R. Rep. No. 95-595, at 133.

13Id.

1411 U.S.C. § 523(a)(8).

15Section 523(a)(8) states that educational debt is excepted from discharge unless it “would impose an undue hardship,” but the following subparagraphs and § 101 do not define what qualifies as an undue hardship. See Id. §§ 101, 523(a)(8) (omitting any elaboration on what undue hardship means).

16See Brunner v. N.Y. State Higher Educ. Servs. Corp., 831 F.2d 395, 396 (2nd Cir. 1987) (per curiam) (creating the Brunner test); Andrews v. S.D. Student Loan Assistance Corp. (In re Andrews), 661 F.2d 702, 704 (8th Cir. 1981) (creating the totality of the circumstances test).

17See Cheney v. Educ. Credit Mgmt. Corp. (In re Cheney), 280 B.R. 648, 659 (N.D. Iowa 2002).

18The fact that this inquiry has been problematic is buttressed by an expansive 2005 study that looked at numerous cases to try to find factual differences that underlay judicial determinations. See Undue Hardship, supra note 6, at 503. The results of this study suggest that there is a great amount of inconsistency in judicial determinations when applying different facts to these vague standards. See id. (finding few statistically significant differences between debtors found to have a future ability to repay and those with a future inability to repay).

19Professor Douglass Boshkoff stated in his article that “experience with the conditional discharge of educational debts in our country suggests that the bankruptcy judge will be given almost unlimited power to determine the lifestyle of a debtor who seeks a discharge.” Douglass G. Boshkoff, Limited, Conditional and Suspended Discharges in Anglo-American Bankruptcy Proceedings, 131 U. Pa. L. Rev. 69, 125 (1982).

20The Real Student-Loan Scandal, supra note 6, at 184.

21Id. at 235.

22Id. This article also postulated that the “higher education finance system suffers from schizophrenia—namely, a public-oriented approach to student-loan origination but a business-oriented approach to student-loan collection.” Id. This view echoes the concerns voiced at the time of the Code’s enactment:[I]f [student] loans are granted too freely and that is what is causing the increase in bankruptcies, then the problem is a general problem, not a bankruptcy problem. The loan program should be tightened, or collection efforts should be increased. If neither of those alternatives is acceptable, then the loan programs should be viewed as general social legislation that has an associated cost. It is inappropriate to view the program as social legislation when granting the loans, but strictly as business when attempting to collect. Such inconsistency does not square with general bankruptcy policy.H.R. Rep. No. 95-595, at 134 (1977), reprinted in 1978 U.S.C.C.A.N. 5963, 6095.

23See S. Rep. No. 95-989, at 7 (1978), reprinted in 1978 U.S.C.C.A.N. 5787, 5793; H.R. Rep. No. 95-595, at 177–78.

24Undue Hardship, supra note 6, at 414–15; see Local Loan Co. v. Hunt, 292 U.S. 234, 244 (1934) (stating that the fresh start policy of bankruptcy law gives a debtor who surrenders property for distribution to creditors freedom from pre-existing debt).

25See Undue Hardship, supra note 6, at 415 (noting that the bankruptcy system solves the “common pool problem” that can arise if creditors only pursue their own interests at the cost of other creditors) (citing Thomas H. Jackson, The Logic and Limits of Bankruptcy Law 7–19 (1986) (giving a comprehensive analysis of the role of bankruptcy law as a “collective debt-collection device”)).

26See 11 U.S.C. § 727(a) (2012) (“The court shall grant the debtor a discharge, unless . . . .”) (emphasis added).

27See id. § 501(a). But, note that creditors do not file proof of claims in no-asset chapter 7 cases.

28See id. §§ 502, 506.

29See id. §§ 725–26.

30See id. § 727(a) (“The court shall grant the debtor a discharge, unless . . .”) (emphasis added).

31See id. § 727(b) (“[T]his section discharges the debtor from . . . any liability on a claim.”).

32See id. §§ 523, 727.

33See Patel v. Shamrock Floorcovering Servs., Inc. (In re Patel), 565 F.3d 963, 967 (6th Cir. 2009) (noting that the statutory exceptions to discharge recognize Congressional intent to protect certain creditors from harm); Miller v. Lewis, 391 B.R. 380, 384 (E.D. Tex. 2008) (stating that certain debts, for public policy reasons, have been decided by Congress to override policies favoring a fresh start); Bolen v. Sallie Mae Servicing Corp. (In re Bolen), 287 B.R. 127, 129 (D. Vt. 2002) (noting that there are certain circumstances where giving the debtor a fresh start is not the paramount concern and that protection of the creditor is more important).

34See 11 U.S.C. § 523(a)(8).

35See id. (creating the condition of “undue hardship” to be able to discharge student-loan debt).

36See id. § 35(a) (1976) (omitting educational debt from the list of debts unaffected by discharge) (repealed 1978).

37See H.R. Rep. No. 95-595, at 536–38 (1977), reprinted in 1978 U.S.C.C.A.N. 5963, 6425.

38Undue Hardship, supra note 6.

39Id. at 419–28. Arguing against the enactment of the undue hardship provision, Representative James O’Hara stated that the provision was “a discriminatory remedy for a ‘scandal’ which exists primarily in the imagination.” See H.R. Rep No. 94-1232 (1976), reprinted in H.R. Rep. No. 95-595, at 148, and in 1978 U.S.C.C.A.N. 5963, 6109. Representative O’Hara decried the provision for likening the Code’s treatment of student-loan debtors to those who obtain loans by “fraud, felony, and alimony-dodging.” Id. (asserting that the new provision subjects educational loans to an assumption of criminality). Further substantiating Representative O’Hara’s claims, a GAO study performed before the provision officially became law “indicated that less than one percent of all federally insured and guaranteed educational loans were discharged in bankruptcy.” Undue Hardship, supra note 6, at 423; H.R. Rep. No. 95-595, at 133.In the face of such convincing evidence, Representative Allen E. Ertel opposed the removal of the undue hardship provision by resorting to unsubstantiated and unrealistic characterizations of student-loan debtors as bad faith actors who would inevitably abuse the bankruptcy system. See Undue Hardship, supra note 6, at 424; H.R. Rep. No. 95-595, at 536–38. Representative Ertel insisted that, without the undue hardship provision, the law would “encourage fraud,” and, furthermore, student-loan debtors would simply take advantage of a free education, immediately discharge their debts upon graduation, and begin their promising careers with a “clean slate” and the “excellent credit rating that accompanies a bankruptcy.” See H.R. Rep. No. 95-595, at 536–38. Contra Undue Hardship, supra note 6, at 424 (disputing Representative Ertels’s assertion that bankruptcy results in excellent credit ratings). Siding ultimately with Representative Ertel, Congress enacted the provision, which therefore, at its inception, vilified the student-loan debtor. Sadly, courts have embraced this lopsided stereotype when applying the law. See Educ. Credit Mgmt. Corp. v. Mersmann (In re Mersmann), 505 F.3d 1033, 1042 (10th Cir. 2007) (“Limitations on the discharge ability [sic] of student loans serve [to] ‘prevent[] abuses of the educational loan system by restricting the ability to discharge a student loan shortly after a student’s graduation . . . .’”); Undue Hardship, supra note 6, at 427 n.112.

40Undue Hardship, supra note 6, at 459.

41Bankruptcy Reform Act of 1978, Pub. L. No. 95-598, § 101, 92 Stat. 2549, 2591 (1978) (formerly codified at 11 U.S.C. § 523(a)(8)(A) (1988)).

42See Robert C. Cloud, When Does Repaying a Student Loan Become an Undue Hardship?, 185 Educ. L. Rep. 783, 786 n.20 (2004).

43See Higher Education Amendments of 1998, Pub. L. No. 105-244, § 971(a), 112 Stat. 1581 (1998). Interestingly, this amendment came in the wake of a 1997 report by the National Bankruptcy Review Commission recommending that Congress eliminate § 523(a)(8) altogether. See Nat’l Bankr. Review Comm’n, Final Report, § 1.4.5 (October 20, 1997). The report went on to say that student loans should be “treated like all other unsecured debts. In so doing, the dischargeability provisions would be consistent with federal policy to encourage educational endeavors. . . . Litigation over ‘undue hardship’ would be eliminated, so that the discharge of student loans no longer would be denied to those who need it most.” Id.

44See Bankruptcy Abuse Prevention and Consumer Protection Act of 2005, Pub. L. No. 109-8, § 220, 119 Stat. 23, 59 (codified at 11 U.S.C. § 523(a)(8)(B) (2012)).

45See The Real Student-Loan Scandal, supra note 6, at 180–82 (expounding on the impact of the 2005 amendments, the unregulated nature of private student loans, and the creditor-friendly policies of Congress).

46Despite Congress’ repeated re-drafting of § 523(a)(8), incrementally limiting the ability to discharge student-loan debt, they never attempted to define the core provision of exception—undue hardship. See 11 U.S.C. §§ 101, 523(a)(8) (omitting any elaboration on what undue hardship means); supra Part I.A.

47See Undue Hardship, supra note 6, at 503 (faulting judicial discretion for the haphazard results in undue hardship litigation and for compromising the fresh start principle). Subjectivity here refers to the overall application of the undue hardship provision across all districts. Although courts find themselves simply limited by precedents today, the precedents differ from district to district and are the result of an initial judge’s determination that they would allow or disallow some expense when determining dischargeability. Because the precedents differ, the overall result is indicative of the initial subjective determinations.

48Section 523(a)(8) states that educational debt is excepted from discharge unless it “would impose an undue hardship,” but the following subparagraphs and § 101 of the Code do not define what qualifies as an undue hardship. See 11 U.S.C. §§ 101, 523(a)(8) (omitting any elaboration on what undue hardship means).

49 See Brunner v. N.Y. State Higher Educ. Servs. Corp., 831 F.2d 395 (2nd Cir. 1987) (per curiam).

50 See Andrews v. S.D. Student Loan Assistance Corp. (In re Andrews), 661 F.2d 702, 704 (8th Cir. 1981).

51Brunner, 831 F.2d at 396. The court in Brunner, recognizing that there was little appellate guidance concerning the undue hardship provision, took the opportunity to weigh in and formally adopt the test spelled out in the District Court from which this appeal originated. Id. Quickly analyzing each of the three prongs in turn, the court noted that prongs one and three comported with common sense and the legislative intent behind § 523(a)(8). Id. Though it recognized the second prong was “problematic,” the court ultimately held that requiring evidence that a hardship would continue into the future “more reliably guarantees that the hardship presented is ‘undue.’” Id.

52Id.

53See id. (requiring each prong be met in order to afford relief); Pobiner v. Educ. Credit Mgmt. Corp. (In re Pobiner), 309 B.R. 405 (Bankr. E.D.N.Y. 2004); Williams v. EFG Tech/Rutgers (In re Williams), 296 B.R. 128 (Bankr. D.N.J. 2003); Shankwiler v. Nat’l Student Loan Mktg. (In re Shankwiler), 208 B.R. 701 (Bankr. C.D. Cal. 1997). It also bears noting that once the debtor establishes that he or she would suffer undue hardship, courts have differed in their approach as to whether this entitles the debtor to the full discharge of his or her debt or a partial discharge that brings the student-loan debt level down to a judicially determined manageable level. See Amanda M. Foster, All or Nothing: Partial Discharge of Student Loans Is Not the Answer to Perceived Unfairness of the Undue Hardship Exception, 16 Widener L.J. 1053, 1072–83 (2007) The circuit courts are split over this issue, and scholars on both sides of the argument have advocated strongly for congressional clarification. See Frank T. Bayuk, The Superiority of Partial Discharge for Student Loans under 11 U.S.C. § 523(a)(8): Ensuring a Meaningful Existence for the Undue Hardship Exception, 31 Fla. St. U. L. Rev. 1091, 1116 (2004) (arguing that partial discharge better comports with ideals of economic and fundamental fairness for both the creditor and debtor); Foster, supra, at 1073 (arguing that the Code, based on plain meaning and statutory construction, legislative history, and principles of fairness, requires a strict, all-or-nothing approach, awarding the debtor full discharge if he or she meets the undue hardship standard).

54See generally Brunner, 831 F.2d 395.

55See The Real Student-Loan Scandal, supra note 6, at 196–97 n.76.

56See Hart v. ECMC (In re Hart), 438 B.R. 406, 410 (E.D. Mich. 2010) (requiring the debtor to show he is currently minimizing his expenses); Innes v. Kansas (In re Innes), 284 B.R. 496, 509 (D. Kan. 2002) (determining that when a debtor’s expenses exceed their income they cannot maintain a minimal standard of living and repay their loan); Pincus v. Graduate Loan Ctr. (In re Pincus), 280 B.R. 303, 317 (Bankr. S.D.N.Y. 2002) (holding that a debtor is not required to live below the poverty line but must show that repayment will require more than a restricted budget); The Real Student-Loan Scandal, supra note 6, at 196–97 n.76 (highlighting the Ninth Circuit’s approach requiring that the debtor fall somewhere between “‘temporary financial adversity’ and ‘utter hopelessness’” to satisfy the first prong). It is in applying the facts of a debtor’s expenses to these vague standards that a judge is able to impose a significant amount of subjectivity. See Pincus, 280 B.R. at 317 (finding a debtor’s expenses excessive in the areas of “(1) communication, (2) cable, (3) food, (4) clothing, (5) laundry and dry-cleaning, (6) transportation, (7) recreation, and (8) certain medical expenses” and further suggesting the debtor could have used more inexpensive cell phone contract plans, eaten out less, and washed his own clothes).

57See Pincus, 280 B.R. at 317.

58See Brunner, 831 F.2d at 396.

59See id.; Sanborn v. Educ. Credit Mgmt. Corp. (In re Sanborn), 431 B.R. 5 (Bankr. D. Mass. 2010) (holding that the debtor must show that their condition of undue hardship will continue into the future).

60Wallace v. Educ. Credit Mgmt. Corp. (In re Wallace), 443 B.R. 781, 789 (Bankr. S.D. Ohio 2010) (quoting Barrett v. Educ. Credit Mgmt. Corp. (In re Barrett), 487 F.3d 353, 359 (6th Cir. 2007)).

61See id. (asserting that the court may consider such circumstances as illness, disability, a lack of useable job skills, or the existence of a large number of dependents); Barrett, 487 F.3d at 359 (holding that the overarching principle that should determine whether the debtor satisfies this prong must be that the additional circumstances be “beyond the debtor’s control, not borne of free choice”). The Ninth Circuit has attempted to identify a list of factors that might be useful in determining whether the debtor will have a future inability to pay. Educ. Credit Mgmt. Corp. v. Nys (In re Nys), 446 F.3d 938, 946 (9th Cir. 2006). These factors, otherwise known as the Nys factors, include:(1) Serious mental or physical disability of the debtor or the debtor’s dependents which prevents employment or advancement; (2) The debtor’s obligations to care for dependents; (3) Lack of, or severely limited education; (4) Poor quality of education; (5) Lack of usable or marketable job skills; (6) Underemployment; (7) Maximized income potential in the chosen educational field, and no other more lucrative job skills; (8) Limited number of years remaining in [the debtor’s] work life to allow payment of the loan; (9) Age or other factors that prevent retraining or relocation as a means for payment of the loan; (10) Lack of assets, whether or not exempt, which could be used to pay the loan; (11) Potentially increasing expenses that outweigh any potential appreciation in the value of the debtor’s assets and/or likely increases in the debtor’s income; (12) Lack of better financial options elsewhere.Id.

62See The Real Student-Loan Scandal, supra note 6, at 196–97 n.76 (finding, in a study of cases in the Ninth Circuit, reliance on extralegal factors, such as what judge decided the case, to be more significant in the outcome of whether a debtor satisfied the second prong of Brunner).

63Roundtree-Crawley v. Educ Credit Mgmt. Corp., (In re Crawley), 460 B.R. 421, 439 (Bankr. E.D. Pa. 2011) (noting that the inquiry is forward-looking in nature); Berry v. Educ. Credit Mgmt. Corp. (In re Berry), 266 B.R. 359, 364 (Bankr. N.D. Ohio 2000) (recognizing that any analysis under the second prong of Brunner is, by its very nature, speculative).

64See Pardo, supra note 6, at 510–12 (finding in a study of a decade of bankruptcy decisions that the debtor’s health played the largest role, even eclipsing a debtor’s monthly household income and expenses, in a bankruptcy court’s determination of undue hardship).

65Brunner, 831 F.2d at 396 (finding that the debtor did not act in good faith by filing for discharge only within a month from when the first payment was due and by failing to take less drastic remedies before filing for discharge).

66E.g., Grine v. Tex. Guar. Student Loan Corp. (In re Grine), 254 B.R. 191, 193 (Bankr. N.D. Ohio 2000).

67See The Real Student-Loan Scandal, supra note 6, at 200 & n.98 (2009) (noting the “inherent subjectivity in an amorphous standard such as good faith”).

68Green v. Sallie Mae Servicing Corp. (In re Green), 238 B.R. 727, 736 (Bankr. N.D. Ohio 1999) (citations omitted); see Hart v. ECMC (In re Hart), 438 B.R. 406, 413 (E.D. Mich. 2010); Dep’t of Educ. v. Wallace (In re Wallace), 259 B.R. 170, 185 (C.D. Cal. 2000) (citations omitted).

69Afflitto v. United States (In re Afflitto), 273 B.R. 162, 171–72 (Bankr. W.D. Tenn. 2001) (citing Goulet v. Educ. Mgmt. Corp. (In re Goulet), 264 B.R. 527, 531 (W.D. Wis. 2001)) (finding that not making repayments does not preclude a finding of good faith if the debtor has sought to maximize income potential and minimize expenses); see also Downey v. Sallie Mae, Inc. (In re Downey), 255 B.R. 72, 77 (Bankr. N.D. Fla. 2000) (deciding good faith was not established because debtor failed to show she minimized her expenses).

70The Real Student-Loan Scandal, supra note 6, at 200; see Katheryn E. Hancock, A Certainty of Hopelessness: Debt, Depression, and the Discharge of Student Loans under the Bankruptcy Code, 33 Law & Psychol. Rev. 151, 156 (2009) (arguing that application of the good faith prong is unduly harsh based upon the strict approach originally taken when § 523(a)(8) was enacted and debtors still had the option to come back and have their loans discharged automatically without having to prove undue hardship).

71See, e.g., H. Amend. 939 to H.R. 4137 (offered Feb. 7, 2008) (striking down a proposed amendment to repeal protection afforded to private student-loan lenders).

72See Andrews v. S.D. Student Loan Assistance Corp. (In re Andrews), 661 F.2d 702, 704 (8th Cir. 1981) (expressing a preference for a less restrictive test); see also Long v. Educ. Credit Mgmt. Corp. (In re Long), 322 F.3d 549, 553 (8th Cir. 2003) (formally embracing the Andrews opinion). This approach has been endorsed by not only the Eighth Circuit, but also by courts within the First Circuit. Bronsdon v. Educ. Credit Mgmt. Corp. (In re Bronsdon), 435 B.R. 791, 800 (B.A.P. 1st Cir. 2010).

73Long, 322 F.3d at 554 (citing Andrews, 661 F.2d at 704); see also Ackley v. Sallie Mae Student Loans (In re Ackley), 463 B.R. 146, 149 (Bankr. D. Me. 2011) (implementing a similar three-prong test). An expanded list of facts and circumstances to which bankruptcy courts may look to determine whether the debtor is entitled to an undue hardship discharge include:(1) debtor’s total present and future incapacity to pay debts for reasons not within control of debtor; (2) whether debtor has made good faith effort to negotiate a deferment or forbearance of payment; (3) whether hardship will be long-term; (4) whether debtor has made payments on student loans; (5) whether debtor suffers from permanent or long-term disability; (6) ability of debtor to obtain gainful employment in area of study; (7) whether debtor has made good faith effort to maximize income and minimize expenses; (8) whether dominant purpose of bankruptcy petition is to discharge student loans; and (9) ratio of student loan debt to debtor’s total indebtedness.D’Ettore v. Devry Inst. of Tech. (In re D’Ettore), 106 B.R. 715, 718 (Bankr. M.D. Fla. 1989) (citations omitted); see also Holmes v. NCO Fin. Sys. (In re Holmes), 319 B.R. 620, 622–23 (Bankr. W.D. Mo. 2004); Houshmand v. Mo. Student Loan Program (In re Houshmand), 320 B.R. 917, 920 (Bankr. W.D. Mo. 2004); Morgan v. Dep’t of Higher Educ. (In re Morgan), 247 B.R. 776, 782 (Bankr. E.D. Ark. 2000) (citing In re D’Ettore, 106 B.R. at 718).

74See Pardo, supra note 6, at 515 nn.41–42 (showing different tests analyzing similar factors, with debtor illness emerging as a prominent consideration under both tests).

75Houshmand, 320 B.R. at 920; see Dep’t of Educ. v. Meling (In re Meling), 263 B.R. 275, 278–79 (Bankr. N.D. Iowa 2001), aff’d, No. C01-2027, 2002 WL 32107248 (N.D. Iowa Jan. 22, 2002) (noting that courts have broad latitude to consider any relevant factor).

76See Houshmand, 320 B.R. at 920.

77Morgan, 247 B.R. at 782.

78Weir v. Paige (In re Weir), 296 B.R. 710, 716 (Bankr. E.D. Va. 2002) (“Regardless of the test used in determining whether repayment of student loans constitutes undue hardship under § 523(a)(8), at a minimum the court must focus on two issues: (1) the economic prospects of the debtor and (2) whether the conduct of the debtor disqualifies the debtor from taking advantage of the exception.”).

79Fed. R. Bankr. P. 7001(6).

80See In re Betz, 31 B.R. 565, 566 (Bankr. W.D. N.Y. 1983); Conn. Student Loan Found. v. Keenan (In re Keenan), 53 B.R. 913, 918 (Bankr. D. Conn. 1985) (finding that defendant sustained the burden of proof in showing undue hardship, thus allowing for the defendant’s debts to be discharged); Garneau v. N.Y. State Higher Educ. Servs. Corp. (In re Garneau), 122 B.R. 178, 179 (Bankr. W.D.N.Y. 1990) (holding that undue hardship must be proved by a preponderance of the evidence).

81Fed. R. Bankr. P. 7001(6).

82The Real Student-Loan Scandal, supra note 6, at 189 (acknowledging four categories of cases concerning the creditor and debtor’s assessment of the nondischargeability status of a debt and finding that litigation is most likely to occur “where the creditor believes the debt to be nondischargeable and the debtor believes the debt to be dischargeable”).

83See 11 U.S.C. § 362 (2012); S. Rep. No. 95-989, at 50 (1978), reprinted in 1978 U.S.C.C.A.N. 5787, 5836; H.R. Rep. No. 95-595, at 340 (1977), reprinted in 1978 U.S.C.C.A.N. 5963, 6296–97. The automatic stay is a fundamental function of the Code to protect debtors from overzealous and encroaching creditors. See 11 U.S.C. § 362; S. Rep. No. 95-989, at 50; H.R. Rep. No. 95-595, at 340.

84See 11 U.S.C. §§ 101(5)(A), 502(a); see also Fed R. Bankr. P. 3001(f) (proscribing that a properly filed proof of claim is prima facie evidence of the validity and amount of that claim).

85The Real Student-Loan Scandal, supra note 6, at 186–87.

86See id. at 188 (putting forth the counterarguments that undue hardship litigation was the type of litigation the Code seeks to prevent).

87See Katherine Porter & Deborah Thorne, The Failure of Bankruptcy’s Fresh Start, 92 Cornell L. Rev. 67, 124 (2006) (finding that the fresh start may be “more myth than magic bullet” and that many chapter 7 debtors continue to have significant financial trouble post-bankruptcy).

88The Real Student-Loan Scandal, supra note 6, at 191 (alteration in original).

89See supra Part I.B.

90See The Real Student-Loan Scandal, supra note 6, at 191–92. Debtors with little to no resources, those who really have the most sympathetic cases, may be more likely to simply not bring a claim because they fear failure without the aid of a lawyer. Interestingly, debtors with the necessary resources to effectively litigate their claim are thus rewarded for declaring bankruptcy with resources in the bank, so to speak, although creditors can point to these resources as evidence that the debtor should not receive discharge. Id.

91See 11 U.S.C. § 523(a) (2012) (establishing different categories of debt that, with the exception of § 523(a)(8), can be excepted from discharge and provide motivation only to creditors to initiate proceedings to preserve their rights to payment).

92See generally Undue Hardship, supra note 6.

93See generally Pardo, supra note 6 (analyzing the effect of a medical condition in an undue hardship proceeding); The Real Student-Loan Scandal, supra note 6 (analyzing the effects of specific demographic, economic, and other variables in undue hardship proceedings); Iuliano, supra note 6 (analyzing the effects of many variables, including the presence of a medical condition, age, employment, income, etc., in undue hardship proceedings).

947 William L. Norton, Jr. & William L. Norton III, Norton Bankr. L. & Prac. § 170:04 (3d ed. 2008). The unique structure of the bankruptcy appellate process “can be traced to congressional reform efforts during the 1970s that sought to improve the quality of the bankruptcy court while simultaneously maintaining it in a subordinate relationship to the district court.” Jonathan Remy Nash & Rafael I. Pardo, An Empirical Investigation into Appellate Structure and the Perceived Quality of Appellate Review, 61 Vand. L. Rev. 1745, 1753 (2008) (footnote omitted). Under the Bankruptcy Act of 1898, many bankruptcy cases were delegated to bankruptcy referees. Id. at 1754. When the Supreme Court created the rules of bankruptcy procedure in 1973 these bankruptcy referees became bankruptcy judges. Id. With the enactment of the Code in 1978, Congress rejected the possibility of vesting bankruptcy judges with Article III status, yet gave them broad jurisdictional powers. Id. at 1755. This choice caused disputes regarding the jurisdictional power of bankruptcy judges to oversee non-core proceedings, which came to a head in Northern Pipeline Construction Co. v. Marathon Pipe Line Co., 458 U.S. 50, 87 (1982). Nash & Pardo, supra, at 1755. In Northern Pipeline, the Supreme Court determined that to the extent that Article III powers had been granted to non-Article III courts (here, the bankruptcy courts), “[s]uch a grant of jurisdiction cannot be sustained as an exercise of Congress’ power to create adjuncts to Art. III courts.” N. Pipeline Constr. Co., 458 U.S. at 87 (alteration in original). In response, the Bankruptcy Amendments and Federal Judgeship Act of 1984 established bankruptcy courts as units of the district court and thereby created a system where a bankruptcy must be filed with a district court, who will then likely have a standing order which refers the case to a bankruptcy judge. Nash & Pardo, supra, at 1755–56. The bankruptcy judge is then empowered to hear and definitively rule upon any of the core proceedings of the bankruptcy case but may only make recommendations as to non-core proceedings, which are then subject to de novo review by the district court if any party objects to such recommendations. Id. at 1756. Final decisions upon these core proceedings by the bankruptcy court are then subject to review by either the district court or BAP with possible review by the court of appeals to follow. Id.

9528 U.S.C. § 158(a)–(b)(1) (2012). Note, that an appellant may also appeal directly to the circuit court if certain conditions are met. See id. § 158(d).

96Id. § 158(a)–(b)(1).

977 Norton & Norton, supra note 94, § 170:04 (2008).

9828 U.S.C. § 1254(1).

99Nash & Pardo, supra note 94, at 1759.

10028 U.S.C. § 158(b)(1). To date, only five circuits (1st, 6th, 8th, 9th, and 10th) have elected to create a BAP. Ben L. Mesches, Bankruptcy Appeals, 45 Tex. J. Bus. L. 107, 128 (2013). Furthermore, Congress has attempted to encourage the creation of more BAPs. See 28 U.S.C. § 158(b)(1) (stating that “[t]he judicial council of a circuit shall establish a bankruptcy appellate panel”).

10128 U.S.C. § 158(b)(6).

102Id. § 158(c)(1).

103Id. § 158(b)(5) (stating that the members of the panel cannot be from the same district from which the appeal originates).

104Coyne v. Westinghouse Credit Corp. (In re Globe Illumination Co.), 149 B.R. 614, 621 (Bankr. C.D. Cal. 1993) (“The bankruptcy courts throughout the circuit certainly must be bound by a BAP decision.”). Dicta within this case also asserted that in matters of bankruptcy, the BAP may even have superiority over the district court, although this view was rejected by many district courts asserting that their power as an Art. III court is not trumped by a BAP. See Bank of Maui v. Estate Analysis, Inc., 904 F.2d 470, 472 (9th Cir. 1990); Far W. Fed. Bank v. Vanasen (In re Vanasen), 81 B.R. 59, 62 (D. Or. 1987).

105See In re Windmill Farms, Inc., 70 B.R. 618, 622 (B.A.P. 9th Cir. 1987), rev’d on other grounds, 841 F.2d 1467 (9th Cir. 1988); In re Proudfoot, 144 B.R. 876, 878–79 (B.A.P. 9th Cir. 1992) (citing In re Windmill Farms, 70 B.R. at 622); In re Gen. Associated Investors Ltd. P’ship, 150 B.R. 756, 760–61 (Bankr. D. Ariz. 1993).

106See State Comp. Ins. Fund v. Zamora (In re Silverman), 616 F.3d 1001, 1005 (9th Cir. 2010), cert. denied, 131 S. Ct. 1679 (2011); In re Warren, 91 B.R. 930 (Bankr. D. Or. 1988).

107See In re Rheuban 128 B.R. 551 (Bankr. C.D. Cal. 1991). The reasoning behind such a view is that the BAP and district court are on the same level and thus one does not control the other, and because bankruptcy courts are units of the district courts and district court decisions do not bind other district court judges, the bankruptcy courts are not controlled by either appellate court. See Rinard v. Positive Invs., Inc. (In re Rinard), 451 B.R. 12, 21 (Bankr. C.D. Cal. 2011).

108Nash & Pardo, supra note 94, at 1747.

109Id. at 1769–74.

110See id. at 1775 n.112.

111Id. at 1791.

112Id. at 1748.

113Id. at 1748–52 (footnotes omitted). The Nash/Pardo study can be consulted for more information concerning the rationale of why these factors are considered to be indicators of quality appellate review.

114Id. at 1759.

115Id.

116 Id. at 1759–60.

117Id. at 1760.

118Id. at 1761–62 (footnote omitted) (stating that BAP “rulings are generally seen to be binding on future bankruptcy appellate panels drawn from the same circuit . . . . In contrast, one district judge is generally not bound to follow the ruling of another district judge . . . .”).

119Id. at 1765. However, the Nash/Pardo study went on to establish that this advantage held by district court judges was slim at best. On the one hand, district court judges are Article III judges, appointed with life tenure and a guarantee of nonreduction in salary. Id. On the other hand, bankruptcy judges receive neither of these benefits as they are not Article III judges. Id. But, Professors Nash and Pardo go on to demonstrate that this perceived inequality is only marginal. Id. First, although bankruptcy judges only serve a fourteen-year term, their appointments can be and often are renewed; few bankruptcy judges leave for lack of reappointment. Id. at 1765–66. Second, the removal process for bankruptcy judges is, in reality, very similar to that of Article III judges and has not produced results that would likely prevent a bankruptcy judge from acting independently. Id. at 1766–67. Third, because the bankruptcy judges’ salaries have been pegged at 92% of the salary of a district court judge since 1987, the bankruptcy judges enjoy a substantially similar fixed compensation as a district court judge. Id. at 1767. Finally, Professors Nash and Pardo surmise that, given the appointment process of bankruptcy judges, they “may be better situated than district judges to resist the political influence that would threaten to compromise an independent judiciary.” Id.

120Id. at 1804–05.

121Id. at 1770.

122Id. at 1807.

12328 U.S.C. § 1334(a)–(b) (2012).

124Id. § 157(a).

125Id. § 152(a)(1).

126Id. § 157(b)(1).

127Id. § 157(b)(2)(I).

128Id. § 157(b)(1).

129Fed. R. Bankr. P. 7001(6), 7003.

13028 U.S.C. § 158(a), (b)(1) (authorizing the creation of the BAP and empowering it to hear appeals with consent of the parties).

131Id. § 158(c)(1) (allowing for an election of the district court as long as the party does so within 30 days of notice of appeal). The district court has jurisdiction to hear such appeals under § 158(a).

132Id. § 158(d)(1).

133See Lark v. Bd. of Trs. (In re Lark), No. CC-09-1239, 2010 WL 6451889, at *4 (B.A.P. 9th Cir. Feb. 4, 2010) (dismissing debtor’s case because she failed to provide a transcript for review).

134Fed. R. Bankr. P. 8001(f)(3)(D).

135Id. 8001(a) (“An appellant’s failure to take any step other than timely filing a notice of appeal does not affect the validity of the appeal . . . .”) (emphasis added). Put another way, while other errors in perfecting one’s appeal, such as an untimely designation of the record, do not affect the appeal’s validity, failure to timely file the notice of appeal does affect the appeal’s validity.

136Id. 8006.

137Id. It should be noted from the outset that these requirements can be difficult for a pro se debtor who is unfamiliar with not only the law in general, but the specific practices of the bankruptcy appellate process. In just the cases reviewed for this Comment’s study, many were found to be dismissed because of a procedural error of the debtor-appellant. See Lark, 2010 WL 6451889, at *6 (dismissing debtor’s case because she failed to provide a transcript for review); Hough v. Pa. Higher Educ. Assistance Agency (In re Hough), 128 F. App’x 369 (5th Cir. 2005) (finding the case unable to be examined because pro se debtor failed to provide transcripts); Looper v. Educ. Credit Mgmt. Corp., No. 3:07-cv-306, 2008 WL 2965887, at *4 (E.D. Tenn. July 30, 2008) (dismissing debtor’s case for failure to file a motion to extend the time to file a brief within the fourteen day timeframe).

138Rifino v. United States (In re Rifino), 245 F.3d 1083, 1086 (9th Cir. 2001) (citing United Student Aid Funds (In re Pena), 155 F.3d 1108, 1110 (9th Cir. 1998)); Educ. Credit Mgmt. Corp. v. Frushour (In re Frushour), 433 F.3d 393, 398 (4th Cir. 2005).

139Pa. Higher Educ. Assistance Agency v. Birrane (In re Birrane), 287 B.R. 490, 494 (B.A.P. 9th Cir. 2002) (quoting Anderson v. Bessemer City, N.C., 470 U.S. 564, 574 (1985)).

140See Pa. Higher Educ. Assistance Agency v. McDay, No. Civ.A.AW 04 2630, 2005 WL 5327774, at *3 (D. Md. Feb. 9, 2005) (noting that the bankruptcy court, as the trial court, is in the best position to judge the credibility of those who testify at the trial). Notably, in the majority of the cases analyzed for this Comment’s study, either the parties did not contest the bankruptcy court’s factual findings or the appellate court quickly stated that the factual findings of the bankruptcy court were not clearly erroneous and thus accepted. See infra Part III.A.

141See Walker v. Sallie Mae Servicing Corp. (In re Walker), 650 F.3d 1227, 1230 (8th Cir. 2011) (“We will not upset those findings of fact unless, after reviewing the entire record, we are left with the definite and firm conviction that a mistake has been made.”).

142Only 32% of cases in this Comment’s study contained issues of fact. More often than not, the parties did not contest the underlying court’s factual findings and the court then simply accepted the lower court’s findings of facts. See infra Part III.A.

143Frushour, 433 F.3d at 398 (stating that the court of appeals reviews directly). As well, both the BAP and district court review the bankruptcy court’s decision directly because they are the first appellate level.

144Id. at 398–99 (citations omitted).

145See discussion supra Part I.B.

146See generally Tirch v. Pa. Higher Educ. Assistance Agency (In re Tirch), 409 F.3d 677, 681 (6th Cir. 2005) (declining to affirm a discharge of student-loan debt based partly on the determination that testimony from the debtor at trial as to her medical condition, without supporting documentation or testimony by doctors, was insufficient to satisfy the standard).

147See Educ. Credit Mgmt. Corp. v. Jesperson, 571 F.3d 775, 780, 782 (8th Cir. 2009); Educ. Credit Mgmt. Corp. v. Polleys, 356 F.3d 1302, 1306 (10th Cir. 2005).

148See Barrett v. Educ. Credit Mgmt. Corp. (In re Barrett), 487 F.3d 353, 363–64 (6th Cir. 2007) (showing ECMC pushing for a per se rule that not pursuing an Income Contingent Repayment Plan must mean that the debtor fails the good faith prong); Frushour, 433 F.3d at 400 (showing ECMC asking for per se rule that have Internet and cable connections requires the conclusions that a debtor is maintaining more than a minimal standard of living); Polleys, 356 F.3d at 1311 (showing ECMC pushing for per se rule that medical disability is required in order to satisfy the future inability prong); Saxman v. Educ. Credit Mgmt. Corp. (In re Saxman), 325 F.3d 1168 (9th Cir. 2003) (showing ECMC pushing for per se rule that partial discharge through use of 11 U.S.C. § 105(a) is unacceptable); Educ. Credit Mgmt. Corp. v. Smith, No. H-11-57, 2011 WL 4625397 (S.D. Tex. Sept. 30, 2011) (showing ECMC pushing for per se rule that a low paying job alone is not enough to satisfy the requirements of the future inability prong); McLaney v. Ky. Higher Educ. Assistance Auth. (In re McLaney), 375 B.R. 666 (M.D. Ala. 2007) (showing ECMC pushing for a per se rule that tithing is not an allowable expense).

149Prevailing studies suggest that a medical condition significantly increases the chances of satisfying the undue hardship standard, and thus those with medical conditions that lost in the initial case might think they have a greater chance to succeed on appeal; see Pardo, supra note 6, at 511; Roe v. Coll. Access Network (In re Roe), 295 F. App’x 927 (10th Cir. 2008); Nash v. Conn. Student Loan Found. (In re Nash), 446 F.3d 188 (1st Cir. 2006); Educ. Credit Mgmt. Corp. v. Barrett (In re Barrett), 337 B.R. 896 (B.A.P. 6th 2006); State Univ. N.Y.-Student Loan Serv. Ctr. v. Menezes, 352 B.R. 8 (D. Mass. 2006).

150This means they may make choices that, though they may not prove fruitful in this particular case, may result in favorable rules they can use in later cases in which they know they will take part.

151Nash & Pardo, supra note 94, at 960 (noting that “appellate review standards create an affirmance bias”) (citing Frank B. Cross, Decision Making in the U.S. Courts of Appeals 48–49 (2007)).

152See Frushour, 433 F.3d at 398–99 (explaining the appellate review de novo standard).

153See Rifino v. United States (In re Rifino), 245 F.3d 1083, 1087 (9th Cir. 2001); McLaney, 375 B.R. at 673; Geyer v. Dep’t of Educ. (In re Geyer), 344 B.R. 129, 131 (S.D. Cal. 2006); Educ. Credit Mgmt. Corp. v. DeGroot, 339 B.R. 201, 206 (D. Or. 2006).

154It is important to note that by studying only judicial opinions, this study fails to encompass the many cases that are settled or disposed of by other means.

155Westlaw is a commercial electronic database that can be accessed via internet at http://www.westlaw.com.

156See Undue Hardship, supra note 6; Pardo, supra note 6; Rafael I. Pardo & Michelle R. Lacey, The Real Student-Loan Scandal, supra note 6; Iuliano, supra note 6.

157This simply includes the case name, appellate court identification, citation, location, and disposition of the case.

158Brunner v. New York State Higher Education Services Corp., 831 F.2d 395, 396 (2nd Cir. 1987) (per curiam).

159Educ. Credit Mgmt. Corp. v. Polleys, 356 F.3d 1302, 1308 (10th Cir. 2005) (emphasis added) (noting that Congress clearly enacted § 523(a)(8) only in an attempt to curb the abuse of the bankruptcy system by these stereotypical debtors with lots of earning potential but high debts they want to get rid of).

160See generally Alan Agresti, Categorical Data Analysis (1990).

161Here, there was less than a 0.0001 probability that random chance alone would have yielded a difference as large as that witnessed between the observed and expected values. It is generally accepted that where the probability is less than .05, or 5%, the finding is statistically significant and thus indicates that there is a correlative relationship between the two variables.

162To fully answer this would require expositions of every case in the study, but the purpose of this study was to see if there were any indicators in the empirical data that could answer this question.

163See Rifino v. United States (In re Rifino), 245 F.3d 1083, 1087 (9th Cir. 2001); McLaney v. Ky. Higher Educ. Assistance Authority (In re McLaney), 375 B.R. 666, 673 (M.D. Ala. 2007); Educ. Credit Mgmt. Corp. v. DeGroot, 339 B.R. 201, 206 (D. Or. 2006); Geyer v. Dep’t of Educ. (In re Geyer), 344 B.R. 129, 131 (S.D. Cal. 2006).

164Total success rate = (number of cases appealed by creditors that they lost + number of cases appealed by debtors that they won) / total number of cases observed.

165Many times a law degree.

166Pardo, supra note 6, at 509–12 (finding in a study of a decade of bankruptcy decisions that the debtor’s health played the largest role, even eclipsing a debtor’s monthly household income and expenses, in a bankruptcy court’s determination of undue hardship).

167Analysis includes cases that did not mention the debtor or the health of the debtor’s depenedent.

16834.00% success with no medical condition compared to the overall debtor success rate when creditors appeal, 46.07%.

169In each situation, the observed results were much closer to the expected values.

170These debtors only provided prognoses of the outlook of their condition based on their own testimony and did not produce medical experts, or doctors, to testify on their behalf as to the prospects of their conditions going forward. See Roe v. Coll. Access Network (In re Roe), 295 F. App’x 927 (10th Cir. 2008); Nash v. Conn. Student Loan Found. (In re Nash), 446 F.3d 188 (1st Cir. 2006); State Univ. N.Y.-Student Loan Serv. Ctr. v. Menezes, 352 B.R. 8 (D. Mass. 2006); Barrett v. Educ. Credit Mgmt. Corp. (In re Barrett), 337 B.R. 896 (6th B.A.P. 2006).

171See infra Appendix, Tables A1–15.

172Fischer’s exact and chi squared tests. See generally Agresti, supra note 160. These tests are performing the same analysis as originally shown in Figure 1 for overall success rates on appeal.

173It is noteworthy that in the 8th Circuit, which applies the more lenient totality of circumstances test, debtors in the cases studied actually performed terribly when they were the appellant. See infra Appendix, Table A23 (showing that debtors succeeded in 0.00% of cases). However, when the creditor was the appellant, the debtor succeeded in 71.43% of cases studied, possibly indicating that under the totality of the circumstances test, courts may be more likely to uphold decisions that favor the debtor. Id.

174There are a total of 142 cases where there is either a “full affirm” or “no affirm” disposition.

175“ECMC is a national guaranty agency, a national student loan bankruptcy servicer, and a contractor for the U.S. Department of Education, providing collection services and document management services.” Educ. Credit Mgmt. Corp., ECMC–FAQ–General, ECMC, http://www.ecmc.org/details/faqGeneral.html#question3.

176For example, consider that a major subset of the eighty-two creditor appeals in the “Other Creditor” row in Figure 9 involved ECMC. See supra Figure 9. More specifically, ECMC was involved in sixty-five of the eighty-nine creditor appeals, and ECMC prevailed in approximately 57% of those appeals. See supra Figure 7. Thus, the statistically significant result in Figure 9—that is, that other creditors fare better than the USDOE in creditor appeals—possibly suggests that repeat players are more likely to succeed than non-repeat players.

177See supra Part II.B.

178This experience means that bankruptcy judges are likely to have significantly more expertise in this area.

179Bankruptcy cases only encompass a small percentage of the types of cases heard by district courts.

180As debtor-appellee, compare 42.55% in the district court to 46.07% in the overall findings from Figure 1. As debtor-appellant, compare 8.70% in the district court to 18.87% in the overall findings from Figure 1.

181See Nash & Pardo, supra note 94, at 1747.

182Compare 66.67% affirmance rate in appeals from district courts to 50% affirmance rate in appeals from BAPs.

183See supra Figure 10.

184See supra Figure 11. This assertion is reinforced by the fact that the results in Figures 11 and 12 are substantially the same.

185Note that there is still an opt out choice for either party to choose to go to a district court instead of a BAP. 11 U.S.C. § 158(c)(1) (2012).

186Educ. Credit Mgmt. Corp. v. Jesperson, 571 F.3d 775, 787–88 (8th Cir. 2009).

Notes and Comments Editor, Emory Bankruptcy Developments Journal; J.D. Candidate, Emory University School of Law (2014); B.A., University of Tennessee, Knoxville; Recipient, Keith J. Shapiro Excellence in Consumer Bankruptcy Writing Award. This comment benefited significantly from the suggestions and guidance of Rafael I. Pardo, Robert T. Thompson Professor of Law at Emory School of Law. Professor Pardo has blazed the academic path in this field, and his willingness to aid me in my study was integral. Without his help, the statistical findings of this study could not have been fleshed out so fully. Thanks also belong to Alexandra Ragan, whose wonderful feedback and ideas helped hone the content of this comment. Finally, I owe a special thanks to my wife, Katie, and my colleagues, Joel M. Langdon, Micah J. Revell, and Evan T. Hoole, for their unfailing grace in listening to me talk on and on about each and every one of the 289 cases I read for this study.