Tanya Johnson v. William Zimmer

PUBLISHED UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT TANYA RENE JOHNSON, f/k/a Tanya  Rene Zimmer, Debtor-Appellant, v. WILLIAM H. ZIMMER,  No. 11-2034 Creditor-Appellee, ROBERT R. BROWNING, Trustee-Intervenor.  Appeal from the United States Bankruptcy Court for the Eastern District of North Carolina, at Raleigh. J. Rich Leonard, Bankruptcy Judge. (10-07244-8-JRL) Argued: March 20, 2012 Decided: July 11, 2012 Before WILKINSON, KING, and AGEE, Circuit Judges. Affirmed by published opinion. Judge Agee wrote the opin- ion, in which Judge King concurred. Judge Wilkinson wrote a dissenting opinion. 2 JOHNSON v. ZIMMER COUNSEL ARGUED: Cortney I. Walker, SASSER LAW FIRM, Cary, North Carolina, for Appellant. Douglas Wickham, HATCH, LITTLE & BUNN, Raleigh, North Carolina, for Appellee. Christopher Scott Kirk, OFFICE OF THE BANKRUPTCY ADMINISTRATOR, Wilson, North Carolina, for Intervenor. ON BRIEF: Travis Sasser, SASSER LAW FIRM, Cary, North Carolina, for Appellant. Trawick H. Stubbs, Jr., Chap- ter 13 Trustee, OFFICE OF THE CHAPTER 13 TRUSTEE, New Bern, North Carolina, for Intervenor. OPINION AGEE, Circuit Judge: In this direct appeal from the United States Bankruptcy Court for the Eastern District of North Carolina, we address a question of first impression in the circuit courts of appeal: in light of the 2005 amendments to the Bankruptcy Code, 11 U.S.C. § 101 et seq. ("the Code"), codified by the Bankruptcy Abuse Prevention and Consumer Protection Act ("BAPCPA"), Pub. L. No. 109-8, 119 Stat. 23 (2005), how is the "household" size of a debtor seeking bankruptcy relief to be calculated under Chapter 13. Finding no error in the bank- ruptcy court’s method of calculating the Debtor’s household size based on how many individuals operate as an "economic unit" with the Debtor, we affirm the order of the bankruptcy court denying the Debtor’s motion for confirmation with leave to amend the Debtor’s "disposable income calculation and plan to reflect the household size [of five]." (J.A. 100.) See In re Johnson, No. 10-07244-8-JRL, 2011 WL 5902883 (Bankr. E.D.N.C. July 21, 2011). I. The facts are not in dispute. Tanya Rene Johnson ("the Debtor") filed a voluntary petition for Chapter 13 bankruptcy JOHNSON v. ZIMMER 3 in September 2010. Robert R. Browning was appointed as Trustee. Upon receiving notice of the Debtor’s motion for confirmation of a plan, the Debtor’s ex-husband, William H. Zimmer ("the Creditor"), objected. The basis for the Credi- tor’s objection was that the proposed plan overstated the Debtor’s household size, resulting in an inaccurate calculation of her monthly expenses. The Creditor maintained that as a result of this alleged error, the Debtor’s proposed Chapter 13 plan improperly showed a "disposable monthly income" insufficient to make payments on two unsecured loans for which the Creditor was jointly liable with the Debtor. Prior to the bankruptcy court’s consideration of the objec- tions, the parties stipulated to the following facts: the Debtor and Creditor share joint custody of their two minor sons. Nei- ther party pays child support; they share "expenses for cloth- ing, school supplies, and other incidental expenses for their sons based on where the sons live when an expense is neces- sary." (J.A. 92-93.) Out-of-pocket medical expenses are divided equally. By oral agreement, the Debtor’s sons reside with her and are in her care and custody for 204 days each year. The Debtor’s current husband has joint custody of three children from his previous marriage: two minor sons and a nineteen-year-old daughter. The Debtor’s step-children reside with her and her husband approximately 180 days per year.1 The Debtor’s proposed Chapter 13 plan claimed a house- hold of seven members, counting individually each person who resided in her home for any period of time within the past six months (i.e., the Debtor, her husband, her two chil- dren, and her three step-children). The Creditor asserted that the Debtor did not actually have seven members of her house- hold because the five children and step-children did not live at her residence full-time. He contended that rather than sim- ply counting the number of "heads on the bed" to determine 1 It is not apparent in the record before us who claims the Debtor’s chil- dren or step-children as dependents for federal income tax purposes. 4 JOHNSON v. ZIMMER household size, the Debtor’s plan should use a method that better approximated the actual economic impact of each indi- vidual on the Debtor’s expenses. He asserted that such an approach would result in a lower calculation of her monthly expenses such that she would have income available with which to pay toward her unsecured debts as part of a proper Chapter 13 plan. In examining the parties’ dispute, the bankruptcy court observed that the Code does not define "household," there was no binding precedent on point, and that other bankruptcy courts followed three different approaches to define that term. In re Johnson, 2011 WL 5902883, at *1-*2. As described in greater detail below, those three approaches are: the "heads- on-beds" approach that follows the Census Bureau’s broad definition of a household as "all the people who occupy a housing unit," without regard to relationship, financial contri- butions, or financial dependency; the "income tax dependent" method derived from the Internal Revenue Manual’s ("IRM") definition that examines which individuals either are or could be "included on the debtor’s tax return as dependents"; and the "economic unit" approach that "assesses the number of individuals in the household who act as a single economic unit by including those who are financially dependent on the debtor, those who financially support the debtor, and those whose income and expenses are inter-mingled with the debt- or’s." (J.A. 96-98.) The bankruptcy court adopted a variation of the "economic unit" approach, first assessing the number of individuals whose income and expenses are intermingled with the Debt- or’s, and then calculating how much time any part-time resi- dents were members of the Debtor’s household. In adopting the "economic unit" approach, the bankruptcy court noted that the other two definitions were inconsistent with the purpose of the Code and were the least flexible in terms of adapting to an individual debtor’s circumstances. JOHNSON v. ZIMMER 5 In deciding that part-time residents should count as part- time members of the Debtor’s "household," the bankruptcy court acknowledged that "[d]ividing children into fractions is not ideal," but concluded that this additional step in applying the economic unit approach best "capture[d] the nuances of familial support and bonds" and enabled the court to "account for dependents who reside with the debtor on a part-time basis . . . in calculating variable costs such as food, utilities, and out-of-pocket health care expenses." (J.A. 98, 99.) In re John- son, 2011 WL 5902883, at *2-*3. Accordingly, the court relied on the parties’ stipulated facts to determine that each of the Debtor’s two sons constituted .56 members of the Debt- or’s household (residing with her 204 days out of a possible 365), and that each of the Debtor’s three step-children consti- tuted .49 members of her household (residing with her 180 days out of a possible 365).2 Id. at *3. Implementing this fractional economic unit approach thus resulted in the Debtor having a total of 2.59 children in her household full-time, which the court then rounded up to three children. Thus, the Debtor, her husband, and the deemed three children yielded a "household" of five persons. The bank- ruptcy court also noted that the Debtor could claim "any par- ticular expenses . . . that the debtor must meet given the family’s total size of seven" as itemized costs in an amended proposed plan. (J.A. 99-100.) Id. Consequently, it denied the Debtor’s motion for confirmation of a plan, but granted leave to amend the plan based on re-calculation of the Debtor’s dis- posable income based on a household size of five. The bankruptcy court certified the issue of the determina- tion of the "household" size for direct interlocutory appeal. We granted the Debtor’s petition for permission to appeal, 2 The bankruptcy court further noted that although the Debtor’s step- daughter was over the age of 19, it appeared from the stipulated facts that she was "financially dependent" on the Debtor. (J.A. 99 n.5.) 6 JOHNSON v. ZIMMER thus satisfying the requirements for the appeal under 28 U.S.C. § 158(a)(3), (d), and Fed. R. Bankr. P. 8003(d).3 II. A. Statutory Framework We review "the appropriate statutory interpretation of the Bankruptcy Code . . . de novo." Botkin v. DuPont Cmty. Credit Union, 650 F.3d 396, 398 (4th Cir. 2011). In undertak- ing this review, we first set forth the relevant statutory provi- sions, and then describe the approaches bankruptcy courts have adopted as well as the parties’ arguments relating to each. For the reasons then articulated, we conclude that there is no statutorily mandated approach and that the Code’s pur- poses were best served in this case by the "economic unit" approach as applied by the bankruptcy court to calculate the Debtor’s "household" size. Accordingly, we affirm the order of the bankruptcy court. In Hamilton v. Lanning, 130 S. Ct. 2464 (2010), the Supreme Court summarized the key features of a Chapter 13 bankruptcy proceeding: Chapter 13 . . . provides bankruptcy protection to "individual[s] with regular income" whose debts fall within statutory limits. 11 U.S.C. §§ 101(30), 109(e). Unlike debtors who file under Chapter 7 and must liquidate their nonexempt assets in order to pay cred- itors, see §§ 704(a)(1), 7126, Chapter 13 debtors are permitted to keep their property, but they must agree to a court-approved plan under which they pay credi- tors out of their future income, see §§ 1306(b), 1321, 1322(a)(1), 1328(a). A bankruptcy trustee oversees 3 The Trustee is an intervenor in this appeal, on the side of the Creditor. For purposes of simplicity, we address their joint arguments as those of the Creditor. JOHNSON v. ZIMMER 7 the filing and execution of a Chapter 13 debtor’s plan. § 1322(a)(1); see also 28 U.S.C. § 586(a)(3). ... [I]f a trustee or an unsecured creditor objects to a Chapter 13 debtor’s plan, a bankruptcy court may not approve the plan unless it provides for the full repayment of unsecured claims or "provides that all of the debtor’s projected disposable income to be received" over the duration of the plan "will be applied to make payments" in accordance with the terms of the plan. 11 U.S.C. § 1325(b)(1); see also § 1326(b)(1) (2000 ed.). Id. at 2468-69. One of the many changes to the Code arising from the BAPCPA was the tightening of how a Chapter 13 debtor’s "projected disposable income" is calculated. Although "pro- jected disposable income" remained an undefined term in the Code, the BAPCPA "specified in some detail how ‘disposable income’ is to be calculated." Id. at 2469. "‘Disposable income’ is now defined as ‘current monthly income received by the debtor’ less ‘amounts reasonably necessary to be expended’ for the debtor’s maintenance and support, for qual- ifying charitable contributions, and for business expenditures. § 1325(b)(2)(A)(i) and (ii) (2006 ed.)." Id. "The phrase ‘amounts reasonably necessary to be expended’ in § 1325(b)(2) is . . . newly defined" and how it is calculated depends on whether the debtor’s current monthly income is above or below the median for his or her state. Id. at 2470. The determination of how to calculate a debtor’s "amounts reasonably necessary to be expended" is based, in part, on the size of a debtor’s "household." See § 1325(b).4 "Household" is not defined in the Code. 4 Specifically, a debtor’s "household" size is relevant to determining how much the debtor can deduct for two portions of his "amounts reason- 8 JOHNSON v. ZIMMER Rather than including "the full amount for ‘maintenance or support’" (the calculation for below-median-income debtors), an above-median-income debtor can only include "certain specified expenses." Hamilton, 130 S. Ct. at 2470 (citing §§ 707(b)(2), 1325(b)(3)(A)). Section 1325 directs above- median-income debtors to use the "means test" set forth in § 707(b)(2) as the basis for calculating those expenses. The "means test" formulation "is reflected in a schedule (Form 22C) that a Chapter 13 debtor must file." Id. at 2470 n.2. Sec- tion 707(b)’s means test uses standardized expenditure figures in lieu of an above-median income debtor’s actual monthly living expenses. Those amounts are derived from the Internal Revenue Service (IRS) National Standards and Local Stan- dards tables for the geographic area in which the debtor resides. Debtors are also permitted to include their actual monthly expenses for certain categories of expenditures fall- ing under the IRS’s "Other Necessary Expenses." The calcu- lations under § 707(b) are based on a fixed number of individuals: "the debtor, the dependents of the debtor, and the spouse of the debtor in a joint case, if the spouse is not other- wise a dependent." § 707(b)(2)(A)(ii)(I). The term "depen- dent" as used in § 707(b) is not defined in the Code either. The means test "supplants the pre-BAPCPA practice of cal- culating debtors’ reasonable expenses on a case-by-case basis, which led to varying and often inconsistent determinations." Ransom v. FIA Card Services, N.A., 131 S. Ct. 716, 722 (2011). Adopting the means test was "‘[t]he heart of [BAPC- PA’s] consumer bankruptcy reforms,’" Id. at 721 (quoting H.R. Rep. No. 109-31, pt. 1, p.2 (2005)), and was designed ably necessary to be expended": the amounts "for the maintenance or sup- port of the debtor or a dependent of the debtor, or for a domestic support obligation . . ." and, "if the debtor is engaged in business, for the payment of expenditures necessary for the continuation, preservation, and operation of such business." § 1325(b)(2)(A)(i), (iii); § 1325(b)(3). "Dependent," as used in this subsection is not defined. JOHNSON v. ZIMMER 9 "to help ensure that debtors who can pay creditors do pay them." Id. (emphasis in original). B. Statutory Analysis 1. Overview As an initial matter, the Debtor briefly posits that resolving which method should be used to calculate her household size may not ultimately be determinative of her case. This is so, in her view, because under either parties’ method of calculat- ing household size (i.e., five or seven members), the Debtor remains an above-median-income debtor under § 1325. Con- sequently, she must use the means test set forth in § 707(b)(2), which directs a debtor to include herself, her "de- pendents," and in appropriate cases her spouse, in the required calculations. And she observes that the sum of those individu- als may not be the same as her "household" size, depending on how "dependents" and "household" are each defined. The Debtor contends that the "dispositive" issue in her case will be the application of § 707(b)’s means test, which requires a determination of who are her "dependents." On this matter, the Debtor contends the Court should "adopt an ordinary and common meaning of the word dependent" and that the mean- ing of "household" is irrelevant. (Br. for Appellant at 14-15.) The Creditor agrees that "[t]he Court could recognize that the definition of ‘dependent’ [in § 707(b)(2)] is the real issue," but does not develop this idea further. (Br. for Appel- lee at 24.) Instead, he focuses on why § 707(b)(2)’s use of the term "dependent" should not be used as a basis for defining "household" according to the IRS definition of "dependents." While the Debtor is correct that her "household" size may not be the "dispositive" inquiry in determining her disposable income, the determination of "household" remains a signifi- cant component of the § 1325(b) scheme. The purpose of § 1325(b) is to determine a debtor’s ability to pay creditors by 10 JOHNSON v. ZIMMER determining his or her projected disposable income. The com- plex formula used for that calculation deducts the debtor’s expenses ("amounts reasonably necessary to be expended") from his or her monthly income. As an initial step in deter- mining the debtor’s "amounts reasonably necessary to be expended," § 1325(b)(3) directs debtors to calculate the rele- vant median family income for his or her geographic area based on the debtor’s "household" size. Accordingly, we decline the Debtor’s invitation to ignore how the bankruptcy court calculated her "household" size. The bankruptcy court’s order is limited to the § 1325(b) determination of how the Debtor should calculate her "house- hold" size. In re Johnson, 2011 WL 5902883, *1-*3. It resolved that issue and ordered the Debtor to file an amended Form 22C and proposed plan based on the court’s conclusion that she has a household of five. In so doing, it did not specifi- cally address how the § 707(b) means test calculation should be performed. Accordingly, what the Debtor identifies as the "dispositive" means test calculations have not yet been per- formed or ruled upon in the bankruptcy proceedings. It would be premature for us to consider that issue now.5 5 Although the Debtor discounts the significance of the § 1325(b) "household" size in her case, the calculation is an important preliminary step in assessing the "amounts reasonably necessary to be expended" for purposes of determining her projected disposable income. As discussed in greater detail below, the "amounts reasonably necessary to be expended" deduction from the Debtor’s monthly income includes an amount "for the maintenance or support of the debtor or a dependent of the debtor." § 1325(b)(2)(A)(i). This amount, in turn is determined by the means test if the Debtor is an above-median income Debtor. § 1325(b)(3). In deciding whether the Debtor is an above-median income debtor, she must identify whether her "current monthly income, when multiplied by 12, [is] greater than" "the highest median family income of the applicable State for a family of 4 or fewer individuals, plus $625 per month for each individual in excess of 4." In making this calculation, the Debtor must use either a household size of five or seven. Whether she uses five or seven will alter her final calculation by $1,250 per month, or a total of $15,000 JOHNSON v. ZIMMER 11 We also observe that the language in § 707(b)(2)(A)(ii)(I) referring to "the debtor, the dependents of the debtor, and the spouse of the debtor in a joint case, if the spouse is not a dependent," is not the totality of the means test calculation.6 In addition to the word "household" being used in other parts of § 707(b), see, e.g., § 707(b)(2)(A)(ii)(II), a debtor is also permitted to claim certain itemized deductions based on actual expenses for items such as "housing and utilities" if they are "reasonable and necessary." E.g., § 707(b)(2)(A)(ii)(V). At various times, both §§ 1325(b) and 707(b) refer alternatively to a debtor’s "dependents," "dependent child[ren]," "house- hold," "family," and "immediate family." As discussed infra pp. 16-17, in order to understand what Congress meant by each of these undefined terms, they must be understood in context and in light of each other. Our interpretation of "household"—one of the first of these terms used in a Chapter 13 debtor’s calculation of his or her disposable income—will impact how these other terms are understood as well. In sum, a debtor’s entire § 707(b) means test calculation will be affected by the threshold determination of how many people are part of her "household," as determined for pur- poses of § 1325(b). Moreover, the bankruptcy court’s review of the appropriateness of a debtor’s means test calculations, e.g., whether the debtor is claiming "reasonable and neces- sary" additional sums for his or her "housing and utilities," will likely also be informed by the debtor’s § 1325(b) "house- for the twelve-month period. While in either case, the result may be that she is an above-median income debtor, she still must go through the required calculations to reach the § 707(b) means test alternative method of calculating her "amounts reasonably necessary to be expended." In turn, as discussed above, the Debtor’s "household" size under § 1325(b) will impact how she calculates her "amounts reasonably necessary to be expended" under the § 707(b) means test. 6 Because it is beyond the scope of the bankruptcy court’s order, we do not address what "dependents" means for purposes of § 707(b)’s means test calculations. 12 JOHNSON v. ZIMMER hold" size. The bankruptcy court appears to have recognized such nuances to the means test calculations when noting that the Debtor could include "any particular expenses [that] exist that the [D]ebtor must meet given the family’s total size of seven" "in an amended Form 22C and provide documentation of those costs." (J.A. 100.) In re Johnson, 2011 WL 5902883, at *3. For all of these reasons, we conclude that it is necessary for us to resolve whether the bankruptcy court properly calcu- lated the Debtor’s "household" size under § 1325(b). 2. Calculating "Household" Size As noted, § 1325(b)(2) contains the provisions for calculat- ing a debtor’s "disposable income" in order to assess his or her ability to pay debts as part of a Chapter 13 plan. At its most basic level, the relevant formula is that the debtor’s dis- posable income equals the debtor’s "current monthly income" "less amounts reasonably necessary to be expended." § 1325(b)(2). A debtor’s "household" size is relevant to deter- mining how to calculate certain parts of the debtor’s "amounts reasonably necessary to be expended." § 1325(b)(3). Despite the centrality of the term to the requisite analysis, the Code does not state how the size of a debtor’s "house- hold" under § 1325(b) is to be calculated for determining his or her disposable income. From this void comes the primary issue before us: whether the bankruptcy court erred in using a fractional economic unit approach to determine the Debtor’s "household" size for purposes of conducting its review of the disposable income calculation. a. The Party’s Arguments The Debtor would have us answer this query in the affirma- tive, contending that the bankruptcy court erred in looking past the "ordinary and common meaning of statutory words that Congress declined to define." (Br. for Appellant 16.) She asserts that her proffered methodology, the heads-on-beds JOHNSON v. ZIMMER 13 approach, is consistent with "[t]he expansive [dictionary] def- inition of household" and that it is "logical and ‘fair’" to rely on the Census Bureau’s correspondingly broad definition of "household" because § 1325 refers debtors to Census Bureau statistics to determine a "median family income."7 (Id. at 18.) She also posits that if Congress had intended to require famil- ial or economic ties to limit the definition of "household," it could have used a word that connotes such limits, such as the terms "family" or "dependents," which are used elsewhere in the Code. (Id. at 17-19.) She thus concludes that the heads-on- beds approach best reflects what Congress intended by "household," and that the bankruptcy court usurped Congress’ legislative function by utilizing any other definition of "household." (Id. at 20-23.) In response, the Creditor asserts that undefined terms are construed by the terms that surround it, and that the bank- ruptcy court correctly rejected the heads-on-beds approach and adopted the economic unit approach. He contends that the Code’s purpose and text are best served by using a definition of "household" that is based on the financial interdependence of the debtor and those persons comprising the debtor’s "household." (Br. for Appellee 14-16.) The Creditor argues that the heads-on-beds approach is too broad and "can lead to anomalies that are inconsistent with the purpose of the means test" and the Code because it merely counts individuals based on presence in the debtor’s residence without examining whether they are part of the debtor’s financial situation. (Id. at 16.) Furthermore, he submits that the Census Bureau’s expansive definition of "household" is not incorporated into the Code and that adopting it does not make sense given the different objectives of the Census and bankruptcy proceed- ings. In conclusion, the Creditor posits that the bankruptcy court’s approach "provides an accurate snapshot of the Debt- 7 The Code defines "median family income" as "the median family income both calculated and reported by the Bureau of the Census." 11 U.S.C. § 101(39A)(A). 14 JOHNSON v. ZIMMER or’s household size when considered over a period of time," and thus was an appropriate method to calculate her house- hold "given the complexities of today’s family structure." (Id. at 19.) Neither party advocates that the bankruptcy court should have used the income tax dependent method of calculating household size. Indeed, other than defining the approach, the Debtor does not address it at all. The Creditor posits that this approach is inconsistent with the purpose of the Code, but submitted during oral argument that it was at least a better approach than the "heads on beds" definition, should the court not adopt the "economic unit" approach. b. Analysis As always, we begin with the statute, "bearing in mind that we should give effect to the legislative will as expressed in the language." United States v. Murphy, 35 F.3d 143, 145 (4th Cir. 1994); see also Pennsylvania Dep’t of Public Welfare v. Davenport, 495 U.S. 552, 557-58 (1990) (stating that con- struction of terms in the Bankruptcy Code "is guided by the fundamental canon that statutory interpretation begins with the language of the statute itself"). Where statutory language is "facially clear and ‘within the constitutional authority of [Congress], the sole function of this [C]ourt[ ] is to enforce it according to its terms.’" Murphy, 35 F.3d at 145 (internal quotation marks and citation omitted). In order "[t]o deter- mine whether the language is plain, we consider ‘the language itself, the specific context in which that language is used, and the broader context of the statute as a whole.’" Newport News Shipbuilding & Dry Dock Co. v. Brown, 376 F.3d 245, 248 (4th Cir. 2004) (quoting Robinson v. Shell Oil Co., 519 U.S. 337, 341 (1997)). In undertaking this inquiry, words that are not defined in the relevant statutory provisions are typically "interpreted as taking their ordinary, contemporary, common meaning." JOHNSON v. ZIMMER 15 United States v. Lehman, 225 F.3d 426, 428 (4th Cir. 2000) (quoting Perrin v. United States, 444 U.S. 37, 42 (1979)). The Court "customarily turn[s] to dictionaries for help in deter- mining whether a word in a statute has a plain or common meaning." Nat. Coalition for Students v. Allen, 152 F.3d 283, 289 (4th Cir. 1998). In this instance, however, reviewing the dictionary definition of "household" does not resolve the mat- ter because this term has multiple definitions of varying scope and consequence. For example, Black’s defines the noun "household" as "1. A family living together. 2. A group of people who dwell under the same roof." Black’s Law Dictio- nary 744 (7th ed. 1999). Webster’s defines the term as "those who dwell under the same roof and compose a family : a domestic establishment; specif[ically] : a social unit com- prised of those living together in the same dwelling place." Webster’s Third New International Dictionary (Unabridged) 1096 (2002 ed.). Although the Debtor relies on the expansive definition of the word to support her argument, by doing so, she necessar- ily overlooks other aspects of these dictionary definitions that are narrower, and would cut against her interpretation. Thus while the heads-on-beds approach looks solely to how many individuals reside under one roof, that definition ignores other components of the dictionary definitions of "household" that limit it to individuals who comprise a "family" or "a domestic establishment." These components of the "ordinary" dictio- nary definition suggest that a "household" also consists of something beyond co-residency. It is therefore not evident from the word "household" alone which common and ordi- nary definition Congress intended to apply in the § 1325(b) analysis. A statutory interpretation of "household" could be crafted in several different ways that would yield different household sizes depending on which dictionary definition is used. As noted, it is a "cardinal rule" of statutory interpretation that "statutory language must be read in context [because] a 16 JOHNSON v. ZIMMER phrase gathers meaning from the words around it." Gen. Dynamics Land Sys., Inc. v. Cline, 540 U.S. 581, 596 (2004) (citation and internal quotation marks omitted). In some cases, "[t]he meaning of a word that appears ambiguous if viewed in isolation may become clear when [it] is analyzed in light of the terms that surround it." Smith v. United States, 508 U.S. 223, 229 (1993). Context provides some guidance in this case, but ultimately does not resolve the fundamental uncertainty of what Con- gress intended "household" to mean. On the one hand, Con- gress used the word "household" as opposed to "family," "dependent child," or "dependent," all of which are used else- where in the surrounding and cross-referenced Code provi- sions. Consistent with the principle that "[t]he use of different terms within related statutes generally implies that different meanings were intended," this would often mean that Con- gress intended the term "household" to mean something other than what those terms mean. See Cunningham v. Scibana, 259 F.3d 303, 308 (4th Cir. 2001) (quoting 2A Norman J. Singer, Sutherland’s Statutes and Statutory Construction, § 46.06, at 195 (6th ed. 2000)). On the other hand, as set forth above, the dictionary definitions of "household" overlap with each of these terms to varying degrees, as "household" may or may not be defined to include the concept of familial connection or domestic interconnectedness. Further muddying the waters is the statute’s own use of dif- ferent terms for aspects of the same calculation. Section 1325(b)(3) states that a debtor should use his or her "house- hold" size and determine the "median family income" for "1 earner" if the household size is one; for "a family of the same number or fewer individuals" for a "household size of 2, 3, or 4 individuals"; or for "a family of 4 or fewer individuals, plus $625 per month for each individual in excess of 4" for a "household exceeding 4 individuals." It is not entirely clear from this language whether Congress intended for the addi- tional references to "family" to serve as qualifiers to the defi- JOHNSON v. ZIMMER 17 nition of "household," or whether "family" simply refers to the cross-referenced "median family income" statistics the debtor is supposed to use when comparing his independently defined "household" size. In either event, it is not immedi- ately clear how to define "family" even as part of the determi- nation of a "household." Similarly, § 1325(b)(3) uses the number in the debtor’s "household" to explain how a debtor is to calculate two com- ponents of the "amounts reasonably necessary to be expended" under § 1325(b)(2)(A)(i) and (B). For example, subsection (A)(i) includes expenses the debtor may have for the "maintenance or support of the debtor or a dependent of the debtor, or for a domestic support obligation." (Emphasis added.) In addition, § 1325(b)(3) instructs above-median income debtors to use § 707(b)’s means test to perform their "amounts reasonably necessary to be expended" calculation, and § 707(b) also refers to a debtor’s "dependents" as opposed to his or her "household." Congress’ use of "dependents" in both of these other statutory provisions that directly relate to the same calculation that is to occur in § 1325(b)(3) would tend to suggest that although Congress used "dependent" and "household" at different points, they may nonetheless have related meanings. In addition to and arising from these immediate contextual conundrums, bankruptcy courts have offered reasoned expla- nations for why one method of defining "household" is more or less appropriate than others based on the context within § 1325(b), the BAPCPA amendments, and the Code as a whole.8 (Text continued on page 19) 8 Most of the bankruptcy courts to address the issue have noted the lack of a definition of "household" in the Code, and have provided some expla- nation for why one or another approach is most consistent with congres- sional intent. As discussed infra p. 19-20, the bankruptcy court in In re Ellringer, 370 B.R. 905 (Bankr. D. Minn. 2007), concluded that the Cen- sus Bureau definition of "household" should be used because that is the definition used in the Census Bureau’s creation of the median family 18 JOHNSON v. ZIMMER income statistics. Id. at 910-11. Other bankruptcy courts to apply this approach have noted that while the "economic unit" approach may have appealing aspects to it, nothing in the word "household" or the relevant statutory text suggests that the individuals in a household be financially interconnected. E.g., In re Epperson, 409 B.R. 503, 506-07 (Bankr. D. Ariz. 2009) ("Congress does not require courts to take into account finan- cial contribution of the household member, relationship to the debtor, and dependency when determining household size. Absent Congressional direction, it is inappropriate to consider a household member’s depen- dency on the [d]ebtor when determining household size; accordingly, household should be understood in the ordinary sense of the word.") (internal quotation marks and citation omitted). The bankruptcy court in In re Law, 2008 Bankr. LEXIS 1198 (Bankr. D. Kan. 2008), concluded that § 1325(b)’s cross-reference to § 707(b)—which only allows debtors to claim "dependents" and refers to the IRS tables—meant that the Code "consistently limits allowable expenses to those of the debtor, the debtor’s dependents, and the spouse of the debtor," and thus the term "household" should be restricted to those individuals a debtor can claim as dependents on his or her income tax fil- ings. Id. at *19. In contrast, the bankruptcy court in In re Jewell, 365 B.R. 796 (Bankr. S.D. Ohio 2007), was "unwilling to accept the broad definitions of [house- hold] suggested by the Debtors of ‘heads on beds’" because it believed that definition was "inconsistent with the methodology and purpose" of the test to "calculate[e] a debtor’s ‘disposable income.’" Id. at 800. It also noted that the income tax dependent method was too narrow "inasmuch as it fails to recognize those instances when a debtor may be actually provid- ing support for a household member, but the circumstances do not fall neatly within the parameters envisioned by the [tax code]." Id. at 801. It also noted that even the IRS recognized "reasonable exceptions" to the principle that "the number of household members allowed should gener- ally be the same as those claimed as dependents on the taxpayer’s tax returns." Id. Consistently, the bankruptcy courts that have applied the "economic unit" approach over the other methods have pointed to its flexi- bility to the financial status of the debtor, and how its results provide the courts with the best tools to achieve the Code’s goal of providing a realis- tic assessment of the debtor’s financial obligations. See In re Herbert, 405 B.R. 165, 169 (Bankr. W.D.N.C. 2008); see also In re Robinson, 449 B.R. 473, 478-82 (Bankr. E.D. Va. 2011) (same, and delineating similar reasons for rejecting the other methods); Morrison, 443 B.R. 378, 384-87 (Bankr. M.D.N.C. 2011) (same). JOHNSON v. ZIMMER 19 However, none of these approaches is directly required by the statutory language and is consequently based on inference and implication to one degree or another. These are the hallmarks of statutory language that is any- thing but plain. Because the term "household" "lends itself to more than one reasonable interpretation," it is ambiguous. See Newport News Shipbuilding & Dry Dock Co., 376 F.3d at 248 (quoting United States ex rel Wilson v. Graham Cnty., 367 F.3d 245, 248 (4th Cir. 2004)).9 Accordingly, "our obligation is to find that interpretation which can most fairly be said to be imbedded in the statute, in the sense of being most harmo- nious with its scheme and the general purposes that Congress manifested." Id. Put differently, where statutory language is ambiguous, we "turn to other evidence to interpret the mean- ing of the provision," interpreting provisions harmoniously, where possible, or by reference to the legislative history, and always with the goal of ascertaining congressional intent. See New Cingular Wireless PCS, LLC v. Finley, 674 F.3d 225, 249 (4th Cir. 2012) (internal quotation marks and citations omitted). Our next step, then, is to determine which approach best corresponds to Congress’ intent despite the ambiguous meaning of the precise text it has used. We begin by examining whether the heads-on-beds approach best reflects Congress’ intent in requiring a debtor to determine his or her "household" size, as the Debtor con- tends. A handful of bankruptcy courts have adopted the heads-on-beds approach, using the Census Bureau definition 9 As one bankruptcy court noted, if § 1325(b) was "so plain, so clear, so unambiguous that the [c]ourt need look no further to determine its mean- ing . . . then how could [so many] courts have interpreted it one way and [so many other] courts have interpreted it in exactly the opposite way? Doesn’t the concept of ‘plain’ meaning carry with it the implication that the same meaning would be ‘plain’—ordinary, literal and obvious—to every reader?" Morrison, 443 B.R. at 387 n.6 (quotation marks and cita- tion omitted). 20 JOHNSON v. ZIMMER of "household," although they use different—and in some cases no—reasons to explain why. Some bankruptcy courts have stated that this definition of "household" is simply its plain meaning. E.g., In re Smith, 396 B.R. 214, 215-18 (Bankr. W.D. Mich. 2008). For the reasons set forth above, we reject that conclusion. Other bankruptcy courts do not spe- cifically identify how they reach that conclusion, but base their use of the heads-on-beds approach upon § 1325(b)’s ref- erence to Census Bureau statistical data. In Ellringer, for example, the bankruptcy court concluded that because § 1325(b)(2) refers to the "median family income," and 11 U.S.C. § 101(39A)(A) defines "median family income" by reference to the Census Bureau’s reports, "the Census Bureau provides the most appropriate definition of ‘household.’" Id. at 910-11; see also, e.g., In re Bostwick, 406 B.R. 867, 872- 73 (Bankr. D. Minn. 2009); In re Fleishman, 372 B.R. 64, 67- 68 (Bankr. D. Or. 2007). We are not persuaded that Congress intended for "house- hold" to be so broadly defined. At the outset, nothing in § 1325(b) directly or indirectly incorporates the Census Bureau’s definition of "household." Although there is a cross- reference to the Census Bureau’s median income tables, that is not sufficient to demonstrate a congressional intent to adopt the usage utilized by the Census Bureau in its wholly separate sphere of government work. Moreover, while the use of "household" as opposed to "family" or "dependents" (terms used elsewhere in § 1325(b)(2)) would indicate that the term means something different from those words, that does not automatically indicate that Congress intended for the term "household" to be synonymous with the Census Bureau’s def- inition and refer to any person living in a particular residence. Nothing in § 1325(b) supports the conclusion that a bank- ruptcy court is required or intended to use such a broad defi- nition. In the absence of clear direction in the Code to use the heads-on-beds approach, the question becomes whether the JOHNSON v. ZIMMER 21 bankruptcy court erred in failing to choose that method over other possible approaches. On this point, we agree with the majority of bankruptcy courts in concluding that the heads- on-beds approach using the Census Bureau’s expansive defi- nition of "household" is inconsistent with the purpose and objectives of the Code. As noted, the Census Bureau defines a "household" as "all of the people, related and unrelated, who occupy a housing unit." Ellringer, 370 B.R. at 911 (internal quotation marks and citation omitted). This definition serves the Census Bureau’s need to compile demographic informa- tion identifying the number of people in a particular geo- graphic region. It is wholly unrelated to any bankruptcy purpose and does not serve the Code’s objective of identify- ing a debtor’s deductible monthly expenses and, ultimately, his or her disposable income. As the Jewell bankruptcy court observed, the Census Bureau definition is at odds with the purpose of § 1325(b) because "[i]f a person lives in the home with the debtor but the debtor does not support that person, then inclusion of that person for purposes of calculating the applicable median fam- ily income and disposable income would give rise to a faulty calculation and would result in an inaccurate figure for both." 365 B.R. at 800; see also Herbert, 405 B.R. at 169 (criticizing the heads-on-beds approach as being "too broad"). The calculation of a debtor’s monthly income and expenses is aimed at ensuring that debtors pay the amount they can rea- sonably afford to pay to creditors. It makes little sense to allow debtors to broadly define their "households" so as to include individuals who have no actual financial impact on the debtor’s expenses. The over-inclusion of individuals in a debtor’s household size would lead to an artificially high cal- culation of the debtor’s "amounts reasonably necessary to be expended" each month, and thus to an incorrect determination of the debtor’s disposable income and ability to pay creditors.10 10 While the heads-on-beds approach runs a substantial risk of being over-inclusive for purposes of § 1325(b)’s calculations, we note that in 22 JOHNSON v. ZIMMER This result would be entirely at odds with the stated purpose of the BAPCPA: "The heart of the bill’s consumer bankruptcy reforms consists of the implementation of an income/expense screening mechanism (‘needs-based bankruptcy relief’ or ‘means testing’), which is intended to ensure that debtors repay creditors the maximum they can afford." H.R. Rep. No. 109-31(I), *89. Because "household" should be interpreted consistent with the context of § 1325(b), the BAPCPA, and the Code as a whole, we conclude the bankruptcy court did not err in rejecting the heads-on-beds approach. Next, we consider whether the bankruptcy court erred in using the "economic unit" approach. As set forth above, because the text does not define "household" and leaves room for different connotations, the economic unit approach does not inherently contradict the language of § 1325(b). And unlike the heads-on-beds approach, the economic unit approach is consistent with § 1325(b), the BAPCPA, and the Code as a whole. By examining the financial interdependence of individuals to determine whether someone is an economic part of the debtor’s household, bankruptcy courts are able to avoid over- and under-inclusive results that would result by artificially defining "household" according to factors unre- lated to which individuals within a residence impact the debt- or’s financial situation. The approach is flexible because it recognizes that a debtor’s "household" may include non- family members and individuals who could not be claimed as dependents on the debtor’s federal income tax return, but who nonetheless directly impact the debtor’s financial situation. See Morrison, 443 B.R. at 388; Herbert, 405 B.R. at 169; Robinson, 449 B.R. at 481-82. Accordingly, it is consistent certain cases it may also be under-inclusive. There may be situations in which a bankruptcy court determines it is appropriate to include within a debtor’s "household" an individual who does not live with the debtor, but is part of the debtor’s reasonably necessary expenses, such as a dependent relative in a nursing home or assisted medical care facility. JOHNSON v. ZIMMER 23 with § 1325(b)’s purpose of identifying a debtor’s disposable income by identifying a debtor’s "amounts reasonably neces- sary to be expended" each month and determining what amount a debtor is able to pay to his or her creditors. Under this method, a debtor’s "household" would include individuals who operate as an "economic unit" with the debtor: those the debtor financially supports and those who financially support the debtor. In other words, those whose income and expenses are interdependent with the debtor’s are part of his or her "household" for purposes of § 1325(b). Such financial intermingling is an appropriate factor in determining "household" size under § 1325(b)(2) because the debtor’s finances are the focal point of the Code. Indeed, § 1325(b) is specifically directed at determining a debtor’s disposable income, which is defined as currently monthly income less "amounts reasonably necessary to be expended." § 1325(b)(2). The "amounts reasonably necessary to be expended" are calculated based, in part, on the relevant median family income for the size of the debtor’s household. Thus, the entire purpose of identifying a debtor’s household size is to use that number to determine his or her financial obligations and ability to pay. A definition of "household" that is also tailored to reflect a debtor’s financial situation focuses directly upon the ultimate purpose of the Code. The "economic unit" approach is also consistent with other components of the § 1325(b)(2) analysis. For example, a debt- or’s disposable income is based, in part, on the debtor’s "cur- rent monthly income," a term that is defined in 11 U.S.C. § 101(10A) to mean "the average monthly income from all sources that the debtor receives (or in a joint case the debtor and the debtor’s spouse receive) without regard to whether such income is taxable income, derived during [a specified period of time]" "and" "any amount paid by any entity other than the debtor (or in a joint case the debtor and the debtor’s spouse), on a regular basis for the household expenses of the debtor or the debtor’s dependents (and in a joint case the debt- 24 JOHNSON v. ZIMMER or’s spouse if not otherwise a dependent) . . . ." Id. (emphases added). Since the "plus" side of the current monthly income part of the equation would include financial contributions from various outside sources, it is also logical to include in the amount deducted therefrom the financial liability of the debtor for the members of the debtor’s "household" on the opposing side of the equation. The relevance on either side of the equation is established by the actual financial connection to the debtor, not mere physical presence or a federal income tax status. Therefore, it makes sense to define "household" with reference to whether the person’s "income or expenses are inter-mingled or interdependent with [the] debtor," and whether the individuals "are acting as a single economic unit." See Morrison, 443 B.R. at 386. The extent to which that household member increases the debtor’s "current monthly income" due to financial contributions set forth under § 101(10A) or increases the debtor’s "amounts reasonably necessary to be expended" under § 1325(b)(2) would then be determined according to each provision. For each of these rea- sons, we conclude that the bankruptcy court did not err in adopting the economic unit approach in its § 1325(b) analysis. As noted, neither party directly advances the third approach—the IRM definition and income tax dependent model—as the best method of defining "household" for § 1325(b) purposes. We agree with the Creditor that there are numerous limitations that this approach would place on the bankruptcy court’s ability to accurately determine a debtor’s household size. While there might be a case where the income tax dependent model could be feasible for § 1325(b) purposes, the case at bar ably demonstrates why that case would be a rare case indeed. In contrast to the "economic unit" approach, the income tax dependent approach fails to match the goals of the BAPCPA and the Code. Bankruptcy courts have articulated the income tax depen- dent method to define a "household" as the same number of individuals as "those allowed as dependents on the taxpayer’s JOHNSON v. ZIMMER 25 tax returns," Jewell, 365 B.R. at 800, or "the people who could be or are included on the debtor’s tax return as depen- dents." Morrison, 443 B.R. at 385.11 The reason some courts have concluded that using this definition is appropriate or even required is because of § 1325(b)’s cross-reference to § 707(b). Above-median income debtors (a status determined, in part, based on the debtor’s "household" size) must use § 707(b)’s means test to determine their "amounts reasonably necessary to be expended," and the means test includes in its formula "the debtor, the debtor’s dependents, and the spouse of the debtor in a joint case, if the spouse is not otherwise a dependent." § 707(b)(2)(A)(ii)(I) (Emphasis added.) The means test specifically refers debtors to use certain IRS tables (the National Standards and Local Standards) as part of their calculations in determining deductible monthly expenses.12 Because these provisions contain two parts of the same calcu- lation, and because § 1325(b) refers to "household" while § 707(b) refers to "dependents," there is some logic to defin- ing the terms consistently. That said, there are three factors that—as with the Census Bureau definition—negate the appropriateness of using this definition for purposes of § 1325(b). First, neither § 707(b) nor § 1325(b) (nor any other Code provision) expressly incor- 11 The IRS defines "dependent" as a qualifying child or relative and takes into consideration numerous factors, unrelated to any bankruptcy purpose, as set forth in 26 U.S.C. § 152 (defining "dependent" for federal income tax purposes). 12 A debtor using the means test is permitted to deduct not only the "ap- plicable monthly expense amounts specified" in the tables, but also "the debtor’s actual monthly expenses for the categories specified as Other Necessary Expenses [by the IRS] for the area in which the debtor resides . . . ." § 707(b)(2)(A)(ii)(I). Thus, in requiring above-median income debt- ors to use the means test to assess their "amounts reasonably necessary to be expended," the BAPCPA moved away from permitting such debtors to claim their actual expenses and toward requiring them to use a more stan- dardized amount. See Ransom, 131 S. Ct. at 721-22; Baud v. Carroll, 634 F.3d 327, 333-34 (6th Cir. 2011). 26 JOHNSON v. ZIMMER porates the IRS definitions. Indeed, the Supreme Court’s deci- sion in Ransom v. FIA Card Services, N.A., 131 S. Ct. 716 (2011), while referencing the IRS’s explanatory guidelines to clarify a point regarding how another aspect of § 1325(b) should be interpreted, expressly "emphasize[d] . . . that the statute does not ‘incorporate’ or otherwise ‘import’ the IRS’s guidance." Id. at 726 n. 7 (internal alterations omitted). Sec- ond, the IRM definition of "dependent" was created with a markedly different purpose from the Code’s definition of "household." As the bankruptcy court in Morrison observed, "[t]he Internal Revenue Code is designed to collect tax reve- nue for the government" by ensuring that each person is accounted for, and that those claimed as dependents are only claimed by one individual for federal income tax purposes. 443 B.R. at 387-88. As demonstrated above, a debtor’s "household" size for bankruptcy purposes is determined in order to assess his disposable income and ability to pay credi- tors: a goal unrelated to tax collection. Third, § 1325(b)(2) refers to "dependents" at one point in the provision and "household" in another, even setting aside § 1325(b)’s subse- quent cross-reference to § 707(b), which in turn uses both "dependents" and "household." If Congress had intended for "household" to refer exclusively to the debtor and his or her dependents as defined for income tax purposes, it could have done so. Congress did not, and instead referred to a debtor’s "household." As the parties recognize, and as we explain further below, the income tax dependent approach tends to be under- inclusive for purposes of ascertaining a debtor’s household size and disposable income. Thus, for example, if narrowly defined as actually claimed dependents on the debtor’s income tax return, this approach would not permit a debtor to include minor children who live with the debtor, but whom by formal or informal agreement the debtor does not claim on his or her tax return. Nor would it necessarily allow a debtor to claim as a member of his or her "household" step-children, a cohabiting fiance, live-in elderly parents, and the like. This JOHNSON v. ZIMMER 27 would be so regardless of the actual financial contributions to the debtor’s monthly income and regardless of the real finan- cial obligations the debtor incurred regarding those individu- als. If the individual did not satisfy the IRS definition of a "dependent," the person would not be included in the debtor’s "household" for § 1325(b) purposes, even though Congress never expressly indicated such an intent, and even though it results in an inaccurate assessment of the debtor’s financial situ- ation.13 Just as the heads-on-beds approach poses the risk of skew- ing the calculation by being over-inclusive and thus risks mis- representing a debtor’s ability to pay, the income tax dependent method poses an unnecessary risk of skewing the calculation by "undercounting legitimate deductions due to a debtor who financially provides for individuals he or she does not claim as dependents" on his or her tax return. Robinson, 449 B.R. at 481. Neither the income tax dependent nor the "heads-on-beds" approach best reflect the purposes of the Code—and specifically the BAPCPA amendments and § 1325—as to how a debtor’s disposable income is to be cal- culated. Rather than the choice between Scylla and Charybdis that these other methods pose, the economic unit approach affords bankruptcy courts the ability to calculate a debtor’s disposable income under § 1325(b) in a way that satisfies the language of the Code, all the while steering clear of creating an unrealistic assessment (in either direction) of the debtor’s disposable income. 13 As with the heads-on-beds approach, although the income tax depen- dent approach tends to skew one way, in some circumstances it may also create a false picture in the other direction. We can readily contemplate situations in which the income tax dependent method would be over- inclusive, such as where divorcing spouses agree who will claim a child for purposes of filing federal income taxes irrespective of the degree of financial support or where the child predominantly lives. 28 JOHNSON v. ZIMMER 3. Part-Time "Household" Members The foregoing analysis does not end our inquiry, however, because the bankruptcy court opted to further refine the eco- nomic unit approach to account for the part-time members of the Debtor’s "household." The Debtor contends that even if the bankruptcy court did not err in using an economic unit analysis to determine her household size, it nonetheless erred in dividing individuals (the Debtor’s children and step- children) into "fractions and percentages" of her "household" when the Code "only speak[s] in terms of whole ‘person[s]’ and ‘individuals.’" She points to the majority of courts that have used the economic unit approach to contend that the bankruptcy court relied on the outlier case of Robinson to "carve[ ] children into fractions" and thus lead to "a contrived result." (Br. for Appellant 23-25.) We find no error in the bankruptcy court’s method of applying the economic unit approach in a manner that accounted for part-time members of the Debtor’s household. The cases apart from Robinson that have used the economic unit approach were asked to determine whether an individual who resided full-time with the debtor should be considered part of his or her "household." E.g., Morrison, 443 B.R. at 388 (using the economic unit approach to determine that the boyfriend of a debtor who lived with that boyfriend full-time was a member of the debtor’s "household" under § 1325(b)); Jewell, 365 B.R. at 802 (using the economic unit approach to determine whether the debtors’ adult daughter and her chil- dren, and the debtors’ adult son, who resided full-time with the debtors were members of the debtors’ "household").14 14 In criticizing the bankruptcy court’s application of the fractional eco- nomic unit approach, the dissent notes that the bankruptcy court in Jewell "did not treat [the debtors’ son] as a fractional member of the household based on the ratio of his parents’ [limited financial support] to his total income." Supra, at 37. Jewell did not ever resolve the issue of whether the debtors’ son should have been part of the debtors’ household because that JOHNSON v. ZIMMER 29 The situation in those cases is not what was presented in Robinson or here, where individuals do not live with the debtor on a full-time basis. See 449 B.R. at 482 (determining whether children who lived part-time with the debtor were part of his "household"). We recognize, as the bankruptcy courts in Robinson and this case have, that dividing individu- als into fractional members of a household is less than ideal. At the same time, we recognize that the Debtor’s situation is increasingly common in modern American life, and that the number of individuals with a financial relationship to a debtor may well vary depending on the day of the week and other cir- cumstances.15 Contrary to the view expressed in the dissenting opinion, this result is consistent with the statutory language precisely was not the overarching issue in the case. 365 B.R. at 802 ("Even if this Court agreed . . . that [the son] should be included as a member of the Debtors’ household . . . ."). Moreover, each of the individuals at issue in Jewell resided with the debtors full-time, so the bankruptcy court was not faced with the circumstance presented here, where the Debtor has individ- uals who reside regularly, but only part-time, with her and for whom she provides primary support only during the periods in which they reside with her. 15 While both we and the bankruptcy court acknowledge this social shift as a factor in why it may be consistent with the intent of the Code to divide individuals into fractional members of a household, this recognition would not justify the approach used if it were inconsistent with the lan- guage Congress used. It is only because we conclude that the statutory text does not preclude this methodology that we recognize that it may be increasingly appropriate to use in light of the rising number of households with part-time residents. Given the language Congress chose to use and has left undefined, it may be that Congress recognized the bankruptcy court would be in the best position to determine a debtor’s "household" size and that a practical, fact-based definition best served the Code’s purposes. In either event, Congress did not provide a clear definition such as it has to define terms in other contexts, so we are tasked with assessing whether the bankruptcy court’s interpretation is an unreasonable interpretation of the language Congress did use. We conclude it is not. 30 JOHNSON v. ZIMMER because Congress did not define "household" or instruct bank- ruptcy courts how to make the determination as to a debtor’s "household" size. Despite this lack of clear guidance, the bankruptcy court carefully considered the specific financial circumstances of the parties involved, and ably recognized that the Debtor’s "household" size fluctuated from two to seven depending on what day or days were used to assess it. Requiring the bankruptcy court to pick between these extremes, which, although conveniently allowing for consid- eration of "whole" dependents, would also result in an inaccu- rate financial snapshot that has no basis in reality or the capacity of a debtor to pay creditors. The Code and the eco- nomic unit approach are flexible enough to permit the bank- ruptcy court’s conclusion that a composite number best reflected the debtor’s true "household" size. To be sure, the bankruptcy court’s task was made easier in this case given the parties’ stipulations as to the amount of time each child and step-child resided with the Debtor and the financial support the Debtor provides to them.16 As the dissent points out, such mathematical precision may be more difficult in other cases. That it was possible here does not mean that every bankruptcy court must undertake the same level of precision in future cases. But it seems to us that the better model of § 1325(b)’s analysis is for the bankruptcy court to try to get to the bottom of the debtor’s true financial circumstances, rather than merely to ballpark a figure based on the approximate number of "whole" dependents. Such a result would lead to similar under- and over-inclusive problems to those addressed above with respect to the heads-on-beds and income tax dependent methods, while simultaneously eliminating the flexibility and advantages of the economic unit approach. Section 1325(b), 16 The dissent may well be correct that the time spent in a debtor’s "household" is not always true reflection of how much financial support a debtor provides. But this is not the case here, and that concern is some- thing bankruptcy courts and the parties can argue in an appropriate case in which the economic unit method is applied. Because of the parties’ stip- ulations in this case, however, the bankruptcy court appropriately deter- mined that the one reflected the other. JOHNSON v. ZIMMER 31 and the Code in general, is focused on assessing a debtor’s financial obligations and ability to pay creditors; in some cases, it may be necessary for the bankruptcy court—as the court did in this case—to recognize that a debtor has part-time members of the household who can neither be wholly excluded or wholly counted, but should be included for the fraction of time during which they are a part of the debtor’s household expenses. Bankruptcy courts are capable of review- ing and making the requisite findings to permit such a deter- mination of a debtor’s household size that reflects a just approximation of the debtor’s true financial situation. This result is consistent with the Supreme Court’s recogni- tion in Hamilton that bankruptcy courts possess flexibility to look beyond a mechanical application of § 1325(b)’s calcula- tions in order to account for variations readily apparent in the record, even after the BAPCPA. See 130 S. Ct. at 2471-78; see also id. at 2478 ("[W]hen a bankruptcy court calculates a debtor’s projected disposable income [under § 1325(b)], the court may account for changes in the debtor’s income or expenses that are known or virtually certain at the time of confirmation."). Consistent with that approach, we conclude that the bankruptcy court did not err in recognizing that the Debtor’s actual household size fluctuated based on the chang- ing number of part-time members of her household. Nor did the bankruptcy court err in exercising its discretion to accom- modate this reality in the debtor’s situation by representing the individuals as fractional full-time members of the house- hold and then rounding to a whole number.17 17 The bankruptcy court did not explain its decision to round from the fraction of 2.59, which resulted from the calculations of how much time each child and step-child lives with the Debtor, to the round number of three "children" for purposes of § 1325(b). The Debtor does not specifi- cally raise any challenge to this specific component of the bankruptcy court’s decision. We find nothing about doing so that conflicts with the Code’s provisions and thus do not find that the court abused its discretion in deciding to round off to the nearest whole individual for the purpose of determining a household size that could be cross-referenced to the median family income tables. 32 JOHNSON v. ZIMMER III. Congress did not make determining a debtor’s "household" size a straightforward component of his or her disposable income calculation. As discussed above, the Census Bureau’s "heads-on-beds" approach is too removed from the purpose of § 1325(b) to be an appropriate measure of a debtor’s "house- hold" size. We further conclude that there are significant con- cerns with the income tax dependent method and the bankruptcy court here was correct not to rely on that approach in this case. Because Congress’ intent will most often be best implemented through a definition of "household" that is based on whether individuals operate as a single economic unit and are financially interdependent, we conclude the bankruptcy court did not err in applying this method to determine the Debtor’s "household" size. Accordingly, we affirm the order of the bankruptcy court. AFFIRMED WILKINSON, Circuit Judge, dissenting: While there is much in the majority’s thoughtful opinion with which I agree, I cannot approve the bankruptcy court’s decision to break a debtor’s children into fractions for pur- poses of Chapter 13’s means test. That approach contravenes statutory text, allows judges to unilaterally update the Bank- ruptcy Code, and subjects debtors to needlessly intrusive and litigious proceedings. Because I do not believe the bankruptcy court adopted a permissible interpretation of the statutory pro- visions at issue, I respectfully dissent. I. The bankruptcy court treated appellant’s children and step- children as fractional units for purposes of Chapter 13’s means test, calculating the fractions based on the number of days per year the children resided with her. The chief problem JOHNSON v. ZIMMER 33 with this approach is that it lacks a foundation in statutory text. Any "interpretation of the Bankruptcy Code . . . must begin . . . with the language of the statute itself," Ransom v. FIA Card Servs., 131 S. Ct. 716, 723 (2011) (internal quota- tion marks and citation omitted), and it is here that the bank- ruptcy court finds itself on shaky ground. According to the Bankruptcy Code, a debtor’s "disposable income" equals "current monthly income received by the debtor . . . less amounts reasonably necessary to be expended" for certain statutorily recognized expenses. 11 U.S.C. § 1325(b)(2) (2006). The Code instructs a debtor to use a sta- tutorily prescribed formula known as the "means test" to cal- culate her deductions for "reasonably necessary" expenditures and her disposable income if her "current monthly income, when multiplied by 12," exceeds the "median family income" in her state for "household[s]" with the same number of "indi- viduals" as hers or fewer. Id. § 1325(b)(3). If the means test applies, as the parties agree it does here, then the amount of the deduction is based in part on the number of "dependents" the debtor has. Id. § 707(b)(2)(A)(ii)(I). The bankruptcy court’s approach embraces the startling conclusion that the meaning of the terms "individuals" and "dependents" in these provisions can encompass fractional human beings. A textual rendering of statutes may seem inconvenient and even incorrect at times, but it has the long-term benefit of pushing Congress to precision and courts to observance of enacted law. The approach below may seem to reflect the eco- nomic realities of modern domestic life where children split time between parents, but it is hardly the only approach capa- ble of doing so. Indeed, as the majority acknowledges, bank- ruptcy courts have a variety of other options available that may suit the circumstances of the case without so grievous an assault upon the statutory text. See In re Kops, No. 11-41153, 2012 WL 438623, at *2-5 (Bankr. D. Idaho Feb. 9, 2012) (canvassing the primary approaches bankruptcy courts have adopted for determining the size of a debtor’s "household"). 34 JOHNSON v. ZIMMER We need not endorse one or the other approach as the abso- lute and invariable best in order to know that the fractional approach is wrong. Whatever the merits or demerits of the bankruptcy court’s fractional view, it is not how we ordinarily interpret statutes. The entire process of determining the size of a debtor’s "household" in order to decide whether the means test applies and applying the means test in order to calculate a debtor’s disposable income is defined by reference to "individuals" and "dependents." See 11 U.S.C. §§ 1325(b)(3), 707(b)(2)(A)(ii)(I). The fact that "Congress did not define ‘household,’" ante, at 30, "individuals," or "dependents" in a definitions section does not entitle bankruptcy courts to take liberties with those terms. On the contrary, "[w]hen terms used in a statute are undefined, we give them their ordinary meaning," Hamilton v. Lanning, 130 S. Ct. 2464, 2471 (2010) (internal quotation marks and citation omitted), and the com- mon meaning of the words "individual" and "dependent" does not include partial people. As the Supreme Court recently observed, the word "‘individual’ ordinarily means ‘[a] human being, a person.’" Mohamad v. Palestinian Auth., 132 S. Ct. 1702, 1707 (2012) (quoting 7 Oxford English Dictionary 880 (2d ed. 1989)). Similarly, the ordinary meaning of "depen- dent" is "a person who relies on another for support." Merriam-Webster’s Collegiate Dictionary 593 (10th ed. 1999). Neither of these definitions authorizes the approach adopted by the bankruptcy court: a fractional view destined for intrusiveness in its effort to capture every sliver of time a child resides with a debtor. The ordinary meaning of these terms is confirmed by their usage in other federal enactments. Cf. Lanning, 130 S. Ct. at 2472 (consulting definitions of the word "projected" in other federal statutes in order to interpret the term in 11 U.S.C. § 1325(b)). To start with, "Congress does not, in the ordinary course, employ the word [individual] any differently" than the "common usage" of the word "to denote a natural person." JOHNSON v. ZIMMER 35 Mohamad, 132 S. Ct. at 1707. Similarly, the word "depen- dent" appears frequently in federal statutes, yet I can find no instance where it is defined to include a fractional unit. For example, as any taxpayer knows, he cannot claim .37 depen- dency on his return, even if his child lived with him for part of the year. See 26 U.S.C. § 152(a)(1), (c)(1)(B) (defining "dependent" for purposes of federal income taxes to include "an individual . . . who has the same principal place of abode as the taxpayer for more than one-half of [the relevant] tax- able year"). Nor is there anything in the Bankruptcy Code that suggests we should assign these terms a contrived fractional definition. In order for us to conclude that Congress has assigned "the word ‘individual’ . . . a broader or different meaning" than a "natural person," "there must be some indication Congress intended such a result." Mohamad, 132 S. Ct. at 1707 (empha- sis in original). There is none here. The Bankruptcy Code speaks only in terms of whole "individuals" and "dependents," 11 U.S.C. §§ 1325(b)(3), 707(b)(2)(A)(ii), and I can find "no textual or other indication in the Code . . . to show that Con- gress intended debtors to calculate ‘fractional dependents,’ or ‘fractional household members,’ in completing means test cal- culations." Kops, 2012 WL 438623, at *5. Nor has the bank- ruptcy court or appellee "given us [any] reason in the statutory text or context to disregard the ordinary meaning of" the terms individual or dependent, FCC v. AT & T, Inc., 131 S. Ct. 1177, 1184 (2011), having chosen instead to rely on policy arguments to advance their position. Because we must "use the ordinary meaning of terms unless context requires a different result," Gonzales v. Carhart, 550 U.S. 124, 152 (2007), this lack of statutory support should have been dispo- sitive. I recognize that these provisions are not models of clarity, but that does not give us the right to overlook those things about the text that are clear. One such thing is the Code’s 36 JOHNSON v. ZIMMER treatment of a debtor’s dependents as whole persons, not per- centages to be rounded up. II. My disagreement with the lower court’s approach does not end with its lack of textual support. I also object to its deci- sion to update the Bankruptcy Code to address the increase in split custody arrangements. From the start, the bankruptcy court made clear that its interpretation would be guided by the need to "address[ ] the growing number of debtors with blended families and joint custody obligations without ignor- ing the economic realities of a debtor’s living situation." In re Johnson, No. 10-07244-8-JRL, 2011 WL 5902883, at *1 (Bankr. E.D.N.C. July 21, 2011). It consequently chose to employ fractional units not because of any textual analysis, but out of a belief that this approach was "the best method the court can employ to ‘adapt to dynamic economic change including various types of family structures regardless of size, shape, or composition.’" Id. at *2 (quoting In re Robinson, 449 B.R. 473, 482 (Bankr. E.D. Va. 2011)). These may be laudable goals, but in our legal system, we leave the updating of statutes to Congress. The Supreme Court has rejected "a dynamic view of statutory interpretation, under which the text might mean one thing when enacted and yet another" if circumstances later change, Harris v. United States, 536 U.S. 545, 556 (2002), choosing instead to adhere to "the general presumption that legislative changes should be left to Congress." State Oil Co. v. Khan, 522 U.S. 3, 20 (1997). Thus, while it may be true that many modern "house- holds look increasingly different from the outmoded images of the ‘traditional’ family," In re Robinson, 449 B.R. 473, 482 (Bankr. E.D. Va. 2011), it is not our job to amend the Bank- ruptcy Code to account for these social changes. Nor is there any reason to think that the legislative branch cannot respond to these challenges. In other areas of the law, JOHNSON v. ZIMMER 37 Congress has proven quite capable of drafting statutes that account for a range of family structures. The Internal Revenue Code, for instance, provides an elaborate "[s]pecial rule for divorced parents" that determines which parent can claim the child as a dependent for income tax purposes in a split cus- tody arrangement. 26 U.S.C. § 152(e). Similarly, Congress has created a detailed method for determining the "[p]arental income and assets for a student whose parents are divorced or separated" for purposes of higher education funding. 20 U.S.C. § 1087oo(f)(1). These statutes offer carefully tailored and democratically enacted mechanisms for addressing a vari- ety of living arrangements. If Congress wishes to do the same in the Bankruptcy Code, it is perfectly able to do so. And even if Congress does not amend the Code, courts can still take economic realities into account without slicing a debtor’s dependents into bits and pieces. Bankruptcy courts can and do apply a version of the economic unit approach that simply considers whether there is an economic relationship between a debtor and a purported dependent rather than trying to measure that relationship in fractional terms. At least one court, for instance, has adopted the economic unit approach while explicitly refusing to treat children in a split custody arrangement as fractions. See Kops, 2012 WL 438623, at *5. Others have applied the economic unit approach without dividing dependents when they easily could have. In In re Jewell, 365 B.R. 796, 802 (Bankr. S.D. Ohio 2007), for exam- ple, the court noted that the debtors provided some limited financial support to their adult son, but it did not treat him as a fractional member of the household based on the ratio of his parents’ support to his total income. See also In re Gaboury, No. 11-10725, 2011 WL 5833972, at *1-2 (Bankr. D.R.I. Nov. 18, 2011) (same). In sum, courts need not engage in leg- islative revisions in order to render decisions that reflect eco- nomic reality. III. Finally, by allowing judges to treat dependents as fractions, today’s decision will require courts to conduct more intrusive 38 JOHNSON v. ZIMMER and more litigious proceedings in order to apply the Chapter 13 means test. Assigning dependents precise percentages will almost always demand a more searching examination of a debtor’s circumstances than an approach that treats them as whole beings. In this case, for instance, the bankruptcy court divided the number of days of the year Johnson’s sons and stepsons lived with her by 365 and then rounded off the results to two deci- mal places. Calculations involving this degree of precision are unlikely to be simple affairs. For while the parties in this case stipulated to the number of days Johnson’s sons and stepsons reside with her each year, that will not always be the case. One can easily envision estranged parents disputing the details of their custody arrangements in bankruptcy court, especially when those details were not settled in or vary from a state court order. See Kops, 2012 WL 438623, at *1 (noting that the debtor’s temporary divorce order granted him "‘su- pervised visitation’ with the children during the pendency of the divorce action, but provided no guidance on when, or for how long, each visit should occur"); Robinson, 449 B.R. at 475 n.1 (noting that the "Debtor testified that visitation sched- ules were originally established by the state court for different days, but the Debtor has arranged schedules with the mothers of his children"). The upshot of this is that courts will often need to scour a debtor’s financial records as well as hear testimony from the debtor and his family in order to calculate household size for purposes of the means test. Such proceedings are apt to be lengthy and intrusive, if not downright litigious. The majority seems to believe that this is an acceptable price to pay for greater economic accuracy, bemoaning the prospect of "under- and over-inclusive" determinations of debtors’ dis- posable income if "individuals" and "dependents" are treated as whole persons. Ante, at 30-31. It contends that such inaccu- racies can be averted if the "part-time members of the [debt- JOHNSON v. ZIMMER 39 or’s] household" are "included for the fraction of time during which they are a part of the debtor’s household expenses." Id. But this exhortation to such fractional determinations suf- fers not only from its inconsistency with the text of the Bank- ruptcy Code, but also from the fact that such perfect accuracy was recognized by Congress as an elusive goal that carried significant litigating costs. According to the Supreme Court, "Congress intended the means test to approximate the debt- or’s reasonable expenditures on essential items," Ransom, 131 S. Ct. at 725 (emphasis added), by replacing "the pre- BAPCPA case-by-case adjudication of above-median-income debtors’ expenses" with "a standardized formula," id. at 729. Indeed, one of the purposes of this formula was to "reduc[e] litigation on the issue of how much a debtor can pay." In re Armstrong, 370 B.R. 323, 328 (Bankr. E.D. Wash. 2007). By permitting the calculation of a debtor’s household size out to two decimal places, today’s decision lets the sort of litigation Congress sought to prevent in through the backdoor. Of course, careful inquiry into a debtor’s financial records is a part of any bankruptcy proceeding, and treating depen- dents as whole individuals may reduce litigation only to a degree. But the decision on where to draw the line between what can and cannot be litigated in this context has already been made by Congress, and we cannot transgress this textual boundary for the sake of what can be a changing and elusive economic accuracy. Indeed, even the bankruptcy court here fell well short in its quest for perfect accuracy, for the number of days a child spends with his parent is only a very rough proxy for the amount of financial support the parent provides the child. To be sure, treating children in joint custody arrangements as whole individuals may lead to some inaccuracies redound- ing to the benefit of either debtors or creditors, depending on the particular case. But that problem "is the inevitable result of a standardized formula like the means test." Ransom, 131 40 JOHNSON v. ZIMMER S. Ct. at 729. The same inaccuracy would occur, for example, if a debtor bore all the costs of child support while remaining frugal, as the means test would allow him to claim standard- ized expenses greater than his actual costs. Kops, 2012 WL 438623, at *5 n.18. In the interest of administrability, "Con- gress chose to tolerate the occasional peculiarity that a brighter-line test produces." Ransom, 131 S. Ct. at 729. We have no authority to rework that formula today. IV. I understand the argument to the contrary. It is contended that fractionalization is simply a fact-finding tool, to be deployed as circumstances dictate. The approaches to such fact finding are several, and, the argument goes, we should not take one off the table as a matter of law. Moreover, "frac- tionalization" is really a misnomer because it is not a matter of fractionalizing human beings, but rather of calculating the percentage of time that whole persons spend in a household. Finally, in rounding off the fraction, the bankruptcy court arrived at three whole "individuals" in the end. This argument, however, understates just how unconven- tional the bankruptcy court’s approach was. The bankruptcy court began by identifying the number of days in a year that each child resided with the debtor. It then divided each of those numbers by 365, yielding five fractions—one for each child. Indeed, the bankruptcy court equated the children with these fractions, holding that "each [of the debtor’s children] constitute[s] .56 members of the household" and that "each [of the debtor’s stepchildren] constitute[s] .49 members." Johnson, 2011 WL 5902883, at *3. After calculating these fractions, the bankruptcy court added them together for a total of 2.59 household members. Finally, without any explanation, the bankruptcy court rounded this total up to three. Such rounding obscures what is really happening: the calculation of fractions for each child, the reduction of the children to these fractions, and the addition of the fractions to yield one larger JOHNSON v. ZIMMER 41 fraction. In other words, the rounding obscures the inconsis- tency between the bankruptcy court’s approach and the text of the Bankruptcy Code. What we really have here is a case of Congress taking an option off the table and the bankruptcy court putting it back into play. I recognize that bankruptcy courts have a degree of discre- tion in applying the Code, and I do not seek to needlessly con- strain their flexibility. But that discretion is not unlimited. "Bankruptcy courts lack authority to . . . depart from [rules] in the Code . . . to implement their own views of wise policy," In re A.G. Fin. Serv. Ctr., Inc., 395 F.3d 410, 413-14 (7th Cir. 2005), and dividing a debtor’s dependents into fractions falls outside those statutory bounds. Nor is such an approach even necessary to preserve flexibility in administering the Code. Eliminating the option of carving up children for purposes of the means test does not prevent bankruptcy courts from choosing from among multiple ways to calculate a debtor’s household size or from taking the economic circumstances of a debtor into account. Judges can get along just fine in this area without embracing interpretations that co-opt the legisla- tive function. With all respect for my friends who see this matter differently, I would reverse the judgment.