GARCIA v. JOHANNS

KAREN LECRAFT HENDERSON, Circuit Judge.

This appeal arises from one of several actions brought against the United States Department of Agriculture (Department or USDA) alleging discrimination in the administration of various federally-funded loan and benefit programs for American farmers. The appellants, individual Hispanic farmers, seek to represent a class of similarly situated Hispanic farmers throughout the nation who claim that the Department discriminated against them in denying them farm loans and other benefits because of their ethnicity and that it failed to investigate the discrimination complaints they subsequently filed with the Department. In the district court, the appellants sought class certification and the USDA moved to dismiss, inter alia, the failure-to-investigate claim. The district court granted the Department's motion to dismiss and denied class certification, concluding that the appellants had failed to meet the requirements of Federal Rule of Civil Procedure 23(a) and 23(b). For the reasons that follow, we affirm in part and remand in part.

See, e.g., Pigford v. Johanns, 416 F.3d 12, 14 (D.C. Cir. 2005) (black farmers); Keepseagle v. Glickman, 194 F.R.D. 1 (D.D.C. 2000) (Native American farmers). A related appeal challenging the district court's denial of class certification to women farmers was heard the same day as this appeal. See Love v. Johanns, 439 F.3d 723 (D.C. Cir. 2006).

I.

The Farm Service Administration (FSA) administers the Department's various loan programs for American farmers through county committees, the members of which are selected locally and are located in over 2,700 counties nationwide. A farmer seeking a loan must first obtain an application from his county committee. 7 C.F.R. § 1910.4(b). He then submits the completed application to the committee which determines whether the farmer meets specific USDA loan criteria, including, inter alia, citizenship, legal capacity to incur debt, education and farming experience, farm size, inability to obtain sufficient credit elsewhere and character. Id. §§ 1941.12 (2006), 1943.12(a) (2006), 1943.12 (1988), 764.4 (2006). If an unsuccessful applicant believes the committee discriminated against him in denying his application, he may lodge a complaint with either the USDA Secretary or the USDA Office of Civil Rights. Id. § 15.6. USDA regulations provide that complaints "shall be investigated in the manner determined by the Assistant Secretary for Civil Rights and such further action taken by the Agency or the Secretary as may be warranted." Id.

In 1994, the Farmers Home Administration (FmHA) was combined with other Department entities to form the FSA. See United States v. Lewis County, 175 F.3d 671, 673 n. 2 (9th Cir. 1999) (citing 7 U.S.C. § 6932 (Supp. 1998)). All references are to the FSA.

On October 13, 2000, ten Hispanic farmers filed this action in the district court. The complaint set forth three counts. Count I sought a declaratory judgment to determine "the rights of plaintiffs and the Class members under the defendant's farm programs including their right to equal credit, and equal participation in farm program, and their right to full and timely enforcement of racial discrimination complaints." 2d Am. Compl. at 56, reprinted in Joint Appendix (JA) 83. The second count alleged a violation of the Equal Credit Opportunity Act (ECOA), 15 U.S.C. §§ 1691 et seq. JA 84. Specifically, the appellants alleged that the "[d]efendant's acts of denying plaintiffs and Class members credit and other benefits and systematically failing to properly process their discrimination complaints was racially discriminatory and contrary to the [ECOA]." JA 84. Finally, the appellants alleged a violation of the Administrative Procedure Act, 5 U.S.C. §§ 551 et seq. JA 84. The appellants sought declaratory relief as well as $20 billion in damages. JA 85. Their complaint also proposed a class of

Although they subsequently amended their original complaint twice, see infra, n. 5, the substantive counts did not change.

ECOA creates a private right of action against a creditor, including the United States, 15 U.S.C. § 1691e(a), who "discriminate[s] against any applicant, with respect to any aspect of a credit transaction" "on the basis of race, color, religion, national origin, sex or marital status, or age" or "because the applicant has in good faith exercised any right under this chapter." Id. § 1691(a). The regulations governing ECOA define a "credit transaction" as "every aspect of an applicant's dealings with a creditor regarding an application for credit or an existing extension of credit (including, but not limited to, information requirements; investigation procedures; standards of creditworthiness; terms of credit; furnishing of credit information; revocation, alteration, or termination of credit; and collection procedures)." 12 C.F.R. § 202.2(m). Although ECOA claims are subject to a two-year statute of limitations, see 15 U.S.C. § 1691e(f), the Congress retroactively extended the limitations period for individuals who had filed administrative complaints with the USDA between January 1, 1981, and July 1, 1997 for alleged acts of discrimination occurring between January 1, 1981 and December 31, 1996. See Pub.L. No. 105-277, § 741, 112 Stat. 2681 ( reprinted in 7 U.S.C.A. § 2279 notes).

all Hispanic participants in FSA farm programs who petitioned or would have petitioned had they not been . . . prevented from timely filing a complaint [against] USDA at any time between January 1, 1981, and the present for relief from . . . racial discrimination . . . and who, because of the failings in the USDA civil rights complaint processing system . . . were denied equal protection . . . and due process in the handling of their . . . complaints.

JA 78 (emphasis in original).

All references are to the appellants' Second Amended Complaint. The appellants eventually moved to file a Third Amended Complaint, which the district court denied. See Garcia v. Veneman, 224 F.R.D. 8, 16 (D.D.C. 2004); see also infra n. 13.

On December 22, 2000, the Department moved to dismiss the complaint pursuant to Federal Rule of Civil Procedure 12(b)(1), contending that the court lacked jurisdiction over the ECOA claim because the appellants had not exhausted their administrative remedies and that, in any event, their claims were time-barred. In addition, the Department moved to dismiss under Federal Rule of Civil Procedure 12(b)(6), arguing that the appellants had failed to state a claim under ECOA, the APA or the Declaratory Judgment Act, 28 U.S.C. §§ 2201 et seq. On March 20, 2002, the district denied the motion in part and granted it in part, relying on its earlier — and similar — order in Love v. Johanns, No. 00-2502 (D.D.C. Dec. 13, 2001). Garcia v. Veneman, No. 00-2445, 2002 WL 33004124 (D.D.C. Mar. 20, 2002), reprinted in JA 95-99. Of relevance here, the court dismissed the failure-to-investigate claim, concluding that the appellants failed to state a claim under ECOA because the investigation of a discrimination complaint is not a "credit transaction" within the meaning of ECOA. JA 97-98. It further held that the claim was not cognizable under the APA because ECOA provides "an adequate remedy." JA 97-98.

On December 2, 2002, the district court denied class certification. Garcia v. Veneman, 211 F.R.D. 15 (D.D.C. 2002) ( Garcia I). It concluded that the appellants failed to show the required "commonality" under Federal Rule of Civil Procedure 23(a)(2) and did not represent a certifiable class under Rule 23(b). They did not show commonality, the court concluded, because they did not demonstrate that the Department operated under a general policy of discrimination nor did they identify a common USDA policy or practice that disparately affected them. Id. at 19-22. The court then considered whether the requested class could be certified under Rule 23(b) and concluded that Rule 23(b)(2) certification was inappropriate because the $20 billion in damages they sought predominated over their request for equitable relief. Id. at 22-23. The court also found Rule 23(b)(3) certification inappropriate because they had not shown that common questions predominated. Id. at 23-24.

After additional discovery, the appellants submitted a supplemental brief on the issue of commonality, which the district court treated as a renewed motion for class certification. Garcia v. Veneman, 224 F.R.D. 8 (D.D.C. 2004) ( Garcia II). They had obtained in discovery 37 USDA loan and disaster benefit files as well as two USDA databases which, they alleged, showed the requisite commonality for both their disparate impact and their disparate treatment allegations of discrimination. Id. at 10. They argued that the files revealed that the USDA had denied Hispanic farmers' applications based on the subjective, rather than the objective, eligibility criteria set forth in 7 C.F.R. § 15.6 and that, as a result, the use of subjective criteria had a disparate impact on them. Id. at 13-15. They also claimed that the USDA as a "single actor" had treated them discriminatorily through a pattern and practice of discrimination. Id. at 10. They listed five sub-patterns of discrimination, including (i) refusal to provide Hispanic farmers with loan applications or assistance in completing applications; (ii) subjecting Hispanic farmers to protracted delays in processing and funding their loans; (iii) using subjective criteria to reject the applications of Hispanic farmers; (iv) unnecessarily subjecting Hispanic farmers to the inconvenience of supervised bank accounts; and (v) delaying or denying loan servicing for Hispanic farmers. Id. at 10. The court nevertheless concluded that, even with their supplementation, they failed to demonstrate commonality.

On September 24, 2004, the appellants moved the district court to certify the order dismissing their failure-to-investigate claim for interlocutory appeal under 28 U.S.C. § 1292(b), which motion the court granted. Garcia v. Veneman, No. 00-2445 (D.D.C. Sept. 27, 2004). In accordance with Federal Rule of Civil Procedure 23(f), the appellants petitioned this court on September 22, 2004 for leave to file an interlocutory appeal of the class certification denial, which petition we granted. In re Garcia, No. 04-8008 (D.C. Cir. Dec. 16, 2004). Before us for review, then, are three orders, namely Garcia, No. 00-2445, 2002 WL 33004124 (D.D.C. Mar. 20, 2002) (granting motion to dismiss), Garcia I, 211 F.R.D. 15 (D.D.C. 2004) (denying class certification), and Garcia II, 224 F.R.D. 8 (D.D.C. 2004) (denying class certification again).

II.

As we have recognized, the district court is "uniquely well situated" to rule on class certification matters. Wagner v. Taylor, 836 F.2d 578, 586 (D.C. Cir. 1987). Accordingly, we review a certification ruling "conservatively only to ensure against abuse of discretion or erroneous application of legal criteria," id., and we will affirm the district court even if we would have ruled differently in the first instance. See McCarthy v. Kleindienst, 741 F.2d 1406, 1410 (D.C. Cir. 1984).

Under Federal Rule of Civil Procedure 23(a), a plaintiff seeking class certification must show that:

(1) the class is so numerous that joinder of all members is impracticable, (2) there are questions of law or fact common to the class, (3) the claims or defenses of the representative parties are typical of the claims or defenses of the class, and (4) the representative parties will fairly and adequately protect the interests of the class.

Failure to adequately demonstrate any of the four is fatal to class certification. See In re Lorazepam Clorazepate Antitrust Litig., 289 F.3d 98, 106 (D.C. Cir. 2002). The district court found that the appellants failed to show "questions of law or fact common to the class" or "commonality" under Rule 23(a)(2). We affirm that ruling.

If a plaintiff meets the requirements of Rule 23(a), he must then establish that class certification is appropriate under one of the three alternatives of Rule 23(b). See Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 613-16, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997). Although the district court found that certification was also inappropriate under subsections (b)(2) and (3), we do not reach that holding because of our affirmance of its subsection (a)(2) holding. See Love v. Johanns, 439 F.3d 723, 730 n. 3 (D.C. Cir. 2006).

To establish commonality under Rule 23(a)(2), a plaintiff must identify at least one question common to all members of the class. See In re Warfarin Sodium Antitrust Litig., 391 F.3d 516, 528 (3d Cir. 2004). Not every common question, however, suffices under subsection (a)(2). As the United States Supreme Court declared of an alleged disparate treatment class in a Title VII action,

there is a wide gap between (a) an individual's claim that he has been denied a promotion on discriminatory grounds, and his otherwise unsupported allegation that the company has a policy of discrimination, and (b) the existence of a class of persons who have suffered the same injury as that individual, such that the individual's claim and the class claims will share common questions of law or fact and that the individual's claim will be typical of the class claims.

Gen. Tel. Co. of Sw. v. Falcon, 457 U.S. 147, 157, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982). Following Falcon, we have required a plaintiff seeking to certify a disparate treatment class under Title VII to "make a significant showing to permit the court to infer that members of the class suffered from a common policy of discrimination that pervaded all of the employer's challenged employment decisions." Hartman v. Duffey, 19 F.3d 1459, 1470 (D.C. Cir. 1994). And in Love v. Johanns, we held that a showing of commonality for a disparate treatment class under ECOA requires the plaintiff to show "(i) discrimination (ii) against a particular group (iii) of which the plaintiff is a member, plus (iv) some additional factor that `permit[s] the court to infer that members of the class suffered from a common policy of discrimination.'" 439 F.3d at 728. (emphasis in original) (citation omitted) (alteration in original).

Other courts have used Title VII precedent in cases involving ECOA. See, e.g., Mays v. Buckeye Rural Elec. Co-op., Inc., 277 F.3d 873, 876 (6th Cir. 2002) ("Given the similar purposes of the ECOA and Title VII, the burden-allocation system of federal employment discrimination law provides an analytical framework for claims of credit discrimination."); Rosa v. Park W. Bank Trust Co., 214 F.3d 213, 215 (1st Cir. 2000) ("In interpreting the ECOA, this court looks to Title VII case law. . . ."); Bhandari v. First Nat'l Bank of Commerce, 808 F.2d 1082, 1100 (5th Cir. 1987) ("The language [of ECOA] is closely related to that of Title VII of the Equal Employment Opportunity Act ("EEOA"), 42 U.S.C. § 2000e-2, and was intended to be interpreted similarly.").

Regarding the appellants' challenge to Department action with an allegedly class-wide discriminatory impact, they must make a showing sufficient to permit the court to infer that members of the class experienced discrimination as a result of the disparate effect of a facially neutral policy. See Cooper v. S. Co., 390 F.3d 695, 716 (11th Cir. 2004). That is, similar to our formulation of the commonality showing necessary for a disparate treatment class set out in Love v. Johanns, the appellants must show for their disparate impact class (i) a discriminatory impact, (ii) affecting a particular group, (iii) of which the plaintiffs are members, (iv) resulting from a common facially neutral policy or practice.

A.

First, the appellants contend that the district court erred in denying class certification of their discriminatory treatment claim based on the geographic spread of the local decisionmakers, labeling it a "pattern and practice" claim, see Appellants' Br. at 40. But see Garcia I, 211 F.R.D. at 22 ("Commonality is defeated . . . by the large numbers and geographic dispersion of the decision-makers. . . ."). As with a Title VII claim, to establish a charge of pattern and practice discrimination under ECOA, a putative class must prove that "discrimination was the company's standard operating procedure — the regular rather than the unusual practice." Bazemore v. Friday, 478 U.S. 385, 398, 106 S.Ct. 3000, 92 L.Ed.2d 315 (1986) (quoting Teamsters v. United States, 431 U.S. 324, 336, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977)). Similarly, to show commonality under Federal Rule of Civil Procedure 23(a)(2), the plaintiff must "make a significant showing to permit the court to infer that members of the class suffered from a common policy of discrimination that pervaded all of the [defendant's] challenged . . . decisions." Hartman, 19 F.3d at 1472.

"As is now well recognized, the class action commonality criteria are, in general, more easily met when a disparate impact rather than a disparate treatment theory underlies a class claim." Stastny v. S. Bell Tel. Tel. Co., 628 F.2d 267, 274 n. 10 (4th Cir. 1980). Establishing commonality for a disparate treatment class is particularly difficult where, as here, multiple decisionmakers with significant local autonomy exist. Id. at 278-80 (reversing class certification because of geographic separation of workforce and autonomy of local decisionmakers); see also Cooper, 390 F.3d at 715. The appellants failed to identify any centralized, uniform policy or practice of discrimination by the USDA that formed the basis for discrimination against Hispanic loan applicants with varied eligibility criteria in over 2,700 counties nationwide over a 20-year period. Rather, despite the appellants' allegation that the USDA's actions are those of a "single actor," their claims arise from multiple individual decisions made by multiple individual committees. Moreover, they do not cite a single reversal of a district court's denial of class certification based on no commonality resulting from the geographic spread of the decisionmakers. Cf. Stastny, 628 F.2d at 278-79 (district court abused discretion in certifying class of employees spread through "great number of geographically dispersed facilities" with "almost complete local autonomy"). Our standard of review is deferential and the appellants have failed to convince us that the district court abused its discretion in denying class certification to the appellants' alleged disparate treatment class.

The appellants contend that we cannot rely on the geographic spread of defendant decisionmakers in deciding whether to certify a disparate treatment class. Appellants' Br. at 40. They are wrong. See, e.g., Bacon v. Honda of Am. Mfg., Inc., 370 F.3d 565, 571 (6th Cir. 2004) (no abuse of discretion in denying certification of class of all black employees at four separate facilities of defendant over 20 year period); Cooper, 390 F.3d at 715 (no abuse of discretion in denying class certification to employees working for different defendants throughout wide geographic area and encompassing range of working environments); Stastny, 628 F.2d at 278-79 (no abuse of discretion in denying class certification to employees working in different plants with local decisionmakers throughout state); Webb v. Merck Co., Inc., 206 F.R.D. 399, 406 (E.D.Pa. 2002) (denying class certification "cut[ting] across employment status, job categories, facilities and geographic regions").

B.

We next consider the appellants' claim that the district court erred in failing to certify a class on whose members the Department's facially neutral action has had a discriminatorily disparate impact. Assuming without deciding that a disparate impact claim is cognizable under ECOA, the claim would require a plaintiff to "identify a specific policy or practice which the defendant has used to discriminate and must also demonstrate with statistical evidence that the practice or policy has an adverse effect on the protected group." Powell v. Am. Gen. Fin., Inc., 310 F.Supp.2d 481, 487 (N.D.N.Y. 2004) (recognizing disparate impact claim under ECOA).

Both Title VII and the Age Discrimination in Employment Act (ADEA) prohibit actions that "otherwise adversely affect" a protected individual. See 42 U.S.C. § 2000e-2(a)(2); 29 U.S.C. § 623(a)(2). The Supreme Court has held that this language gives rise to a cause of action for disparate impact discrimination under Title VII and the ADEA. See Smith v. City of Jackson, 544 U.S. 228, 125 S.Ct. 1536, 1540, 161 L.Ed.2d 410 (2005) (ADEA); Griggs v. Duke Power Co., 401 U.S. 424, 432, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971) (Title VII). ECOA contains no such language. We express no opinion about whether a disparate impact claim can be pursued under ECOA. See Eisen v. Carlisle Jacquelin, 417 U.S. 156, 177-78, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974) (court should not examine whether "plaintiffs have stated a cause of action or will prevail on the merits" in determining class certification vel non).

The appellants press two alternative theories to support their contention that the district court erred in not certifying a disparate impact class. First, they argue that they do not need to specify a facially neutral practice if it is impossible to determine which of the USDA eligibility criteria have had the discriminatory effect, instead borrowing from Title VII's "one employment practice" notion. Alternately, they argue the USDA's subjective decisionmaking process constitutes the common facially neutral practice. We reject both theories and instead affirm the district court's denial of class certification because the appellants failed to show a common facially neutral USDA farm loan policy, resulting in the disparate effect on them and the putative class of Hispanic farmers. See Caridad v. Metro-North Commuter R.R., 191 F.3d 283, 292 (2d Cir. 1999) ("Of course, class certification would not be warranted absent some showing that the challenged practice . . . has a disparate impact on African-American employees at Metro-North."). As the Supreme Court noted in Falcon — where class certification was denied — "[t]he mere fact that an aggrieved private plaintiff is a member of an identifiable class of persons of the same race or national origin is insufficient to establish his standing to litigate on their behalf all possible claims of discrimination against a common employer." Falcon, 457 U.S. at 159 n. 15, 102 S.Ct. 2364.

The appellants cite the "one employment practice" language of Title VII, see 42 U.S.C. § 2000e-2(k)(1)(B)(I), and argue that it relieves them from having to tie a disparate impact to a facially neutral USDA policy. Appellants Br. at 34-38. The Congress added the "one employment policy" language following the Supreme Court's holding in Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 109 S.Ct. 2115, 104 L.Ed.2d 733 (1989). It provides that "if the complaining party can demonstrate to the court that the elements of a respondent's decisionmaking process are not capable of separation for analysis, the decisionmaking process may be analyzed as one employment practice." 42 U.S.C. § 2000e-2(k)(1)(B)(I). Assuming — again, without deciding — the "one employment practice" notion applies to an ECOA disparate impact claim, it does not alter the required commonality showing under Federal Rule of Civil Procedure 23(a)(2). The appellants erroneously confuse the commonality showing with the prima facie case of disparate impact discrimination. See Bacon v. Honda of Am. Mfg., Inc., 370 F.3d 565, 572 (6th Cir. 2004) ("Plaintiffs cannot avoid the heavy lifting of showing eligibility for class certification by conflating two exceptions to separate rules for adjudicating discrimination cases."). Under Rule 23(a)(2), the appellants must show that the putative class members have something in common — they all suffered an adverse effect from the same facially neutral policy, see id. — and their showing must be "significant," see Hartman, 19 F.3d at 1470. On the other hand, courts have set a lower bar for establishing a prima facie discrimination case. See, e.g., Tex. Dep't of Cmty. Affairs v. Burdine, 450 U.S. 248, 253, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981); Bundy v. Jackson, 641 F.2d 934, 950 (D.C. Cir. 1981) (recognizing difficulty plaintiff faces in proving motive behind employer's actions).

In Garcia I, 211 F.R.D. at 21-22, the district court rejected the appellants' disparate impact claim because they did not connect disparate impact with a common facially neutral USDA policy. They had submitted the declaration of Jerry Hausman, an expert in econometrics and microeconomics, in which declaration he concluded that Hispanic farmers received a lower percentage of USDA loans than white farmers received in 1997. JA 123. Hausman, however, analyzed all farmers (white and Hispanic) as opposed to only those farmers (white and Hispanic) who had applied for USDA loans. After further discovery produced USDA loan databases, two of which the appellants used to support their renewed class certification motion, they submitted the declaration of statistician Karl Pavlovic, who found that 72 per cent of white applications were approved in the period from October 1997 to January 2003 while 59 per cent of Hispanic applications were approved in the same period. JA 477. In Garcia II, the district court assumed a disparate impact without discussion of Pavlovic's declaration. Garcia II, 224 F.R.D. at 11. The court, however, again concluded that the appellants had failed to connect the disparate impact to a common facially neutral USDA policy. Id. (rejecting appellants' argument because "[n]ot only does it `leapfrog to the merits,' . . . but it also boils down to the proposition that unexplained discrepancies in the distribution of government benefits satisfy the commonality requirement of Rule 23(a)(2) without more").

The appellants attempted to connect the disparate impact to USDA's subjective loan decisionmaking criteria, relying in part on statistical evidence. But their statistical analyses were analytically flawed because they did not incorporate key relevant variables connecting disparate impact to loan decisionmaking criteria. See Bazemore v. Friday, 478 U.S. 385, 400 n. 10, 106 S.Ct. 3000, 92 L.Ed.2d 315 (1986) ("some regressions [are] so incomplete as to be inadmissible as irrelevant"). It does not suffice under Rule 23(a)(2) to show an ethnic imbalance in the USDA's award of loans to farmers; rather, the appellants must show that a common facially neutral policy caused the imbalance. See Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 656-57, 109 S.Ct. 2115, 104 L.Ed.2d 733 (1989) ("[A] Title VII plaintiff does not make a case of disparate impact simply by showing that, `at the bottom line,' there is racial imbalance in the work force."); Caridad, 191 F.3d at 292. The appellants could have done this, for instance, by employing multiple regression. See Love v. Johanns, 439 F.3d at 731 ("Instead of conducting a relatively simple statistical analysis (such as a multiple regression) to control for any or all of these variables, O'Brien simply reported a series of elementary cross-tabulations, from which it is impossible — as a statistical matter — to draw meaningful conclusions."); see also Segar v. Smith, 738 F.2d 1249, 1261 (D.C. Cir. 1984) ("Multiple regression is a form of statistical analysis used increasingly in Title VII actions that measures the discrete influence independent variables have on a dependent variable such as salary levels."). The appellants' statistics failed to account for variables that affected the analyses such as whether fewer Hispanic farmers were U.S. citizens, whether Hispanic farmers had worse credit and whether Hispanic farmers had less experience. Love, 439 F.3d at 731-32.

The district court thus acted within its discretion in rejecting the appellants' statistical showing as insufficient to infer classwide discrimination arising from the Department's administration of the farmers' loan programs. Its decision to deny class certification "did not constitute a clear error of judgment, nor [was it] otherwise outside the range of choices the district court was allowed to make." Cooper, 390 F.3d at 715. We, of course, do not suggest that statistical evidence alone could never show commonality; we simply believe that the district court did not abuse its discretion in finding the appellants' statistical evidence inadequate. See Hartman, 19 F.3d at 1473 ("While statistics can generally be probative of the question of commonality, we would feel uncomfortable in resting on the trial statistics in the present record for a final determination of commonality.").

We think the class certification issue here is similar to that in Cooper v. Southern Co., 390 F.3d 695 (11th Cir. 2004), in which the Eleventh Circuit affirmed the denial of class certification in a Title VII action. Seven black employees of Southern Company and several of its subsidiaries sought to represent a class alleging disparate impact and disparate treatment claims in connection with promotion opportunities, performance evaluations and compensation. The court found that the "plaintiffs' statistical evidence was insufficient to establish a presumption of discrimination common to the claims of all members of the putative class." Id. at 719 (emphasis in original). "[A]nalytical flaws in the statistical evidence" prevented the Cooper plaintiffs from making a showing sufficient to "`raise a presumption of discrimination arising from the collective whole of Defendant's compensation and promotion policies. Thus, disparate impact analysis produce[d] no evidence common to the claims of all class members.'" Id. at 716 (quoting Cooper v. S. Co., 205 F.R.D. 596, 613 (N.D.Ga. 2001)) (alteration and emphases in original). The statistical evidence there did not account for variables such as an employee's type or level of acquired skills and field of study, the quality, type and relevance of an employee's experience, an employee's job performance, etc., to ensure that black and white employees were similarly situated. Compare id. at 717 with Caridad v. Metro-North Commuter R.R., 191 F.3d 283, 293 (2d Cir. 1999) (reversing denial of class certification because expert "controlled for various factors that one would expect to be relevant to the likelihood of disciplinary action and promotion"). In addition, the statistical evidence did not reference the named plaintiffs or their specific similarly-situated comparators and, accordingly, the court found that they had not established " commonality among these named plaintiffs' claims and the overall affected class." Cooper, 390 F.3d at 718 (emphasis in original).

The other evidence the appellants relied on — namely, the 37 USDA case files — arguably may have come closer to establishing commonality because it showed that the USDA often used the infeasibility of an applicant's farm plan as one reason for denying a loan. See Garcia II, 224 F.R.D. at 14 (farm plan infeasibility given as one reason for almost half of loan rejections). Nonetheless, mindful of our limited scope of review, see supra at 632, we do not believe that the district court abused its discretion in denying class certification. The USDA denied loans for a variety of reasons, including inadequate farm plans and lack of funds. Mem. in Response to the Court's July 15, 2003 Order with Respect to Commonality at app. 7, Garcia v. Veneman, No. 02-2445 (D.D.C. filed Dec. 5, 2003). The case files as well as the anecdotal evidence upon which the appellants relied showed that often the appellants were denied loans based on objective financial data. See id. In sum, the Department used an array of objective — and individual — justifications in denying the appellants loans. Accordingly, we affirm the district court's denial of class certification of the appellants' disparate impact claim.

For instance, Roberto Salinas and his son jointly applied for an ownership loan in 2000 and Roberto Salinas solely applied for an operating loan in the same year. The USDA denied both loans because of the infeasibility of the farm plan as well as inadequate verification of Roberto Salinas's debt. Mem. in Response to the Court's July 15, 2003 Order with Respect to Commonality at app. 7, Garcia v. Veneman, No. 02-2445 (D.D.C. Dec. 5, 2003).

In addition to the disparate impact and treatment classes already discussed, the appellants sought certification of five subclasses. Garcia II, 224 F.R.D. at 15-16. The five subclasses were set forth in their proposed Third Amended Complaint, see JA 512-13, which the district court denied without prejudice. Garcia II, 224 F.R.D. at 16. Their challenge to the district court's denial of their motion to amend is supported by conclusionary assertions only, see Appellants' Br. at 44, and they have therefore waived the issue. See United States v. Yeh, 278 F.3d 9, 16 n. 4 (D.C. Cir. 2002).

III.

We have jurisdiction to review, in our discretion, the district court's dismissal of the appellants' failure-to-investigate claim under ECOA and the APA pursuant to 28 U.S.C. § 1292(b). The appellants must persuade us that exceptional circumstances justify a departure from the ordinary policy of postponing appellate review until after entry of final judgment. See United States. v. Philip Morris USA, Inc., 396 F.3d 1190, 1209 (D.C. Cir. 2005).

We exercise our jurisdiction over the dismissal of the ECOA failure-to-investigate claim, as we did in Love v. Johanns, and affirm the district court's dismissal for the same reason — the failure to investigate a discrimination complaint is not a "credit transaction" within the meaning of ECOA. Love v. Johanns, 439 F.3d at 732-33. We decline, however, to exercise our jurisdiction regarding the appeal of the denial of the appellants' failure-to-investigate claim made under the APA. As in Love, the class certification issues took most of the trial court's and the parties' attention and unlike the straightforward statutory construction issue the appellants' ECOA failure-to-investigate claim presents, we think this claim will benefit from further development in the district court. Id. at 732-33.

Before us, the appellants used slightly more than four pages of their 59-page brief and no time at oral argument addressing the APA failure-to-investigate claim.

For the foregoing reasons, we affirm the district court's denial of class certification as well as its dismissal of the failure-to-investigate claim asserted under ECOA. We dismiss the appeal of the APA failure-to-investigate claim and remand to the district court for further proceedings consistent with this opinion.

So ordered.