Prosecuting Preference Actions Post-BAPCPA Another View Toward a Reliable Statistical Model

Prosecuting Preference Actions Post-BAPCPA Another View Toward a Reliable Statistical Model

Journal Issue: 
Column Name: 
Journal Article: 
In "Defending Preference Actions: Optimal Strategies for Comprehensive Mathematical Analysis," published in the October 2006 issue of the ABI Journal, Kurt A. Winiecki advocated using certain statistical tools to augment a defendant's claim that challenged transactions may be within the subjective course of prior business dealings. While Mr. Winiecki is to be commended for using a matrix approach to formulating a subjective "ordinary course of business" (OCB) defense to preference liability, his methodology, standing alone, fails to enunciate the basis for establishing the critical element of that defense—identifying the proper historical baseline between the debtor and the preference recipient. Instead, Mr. Winiecki merely postulates various potential deviations from the historic average payment time to see what the result will be with respect to remaining liability. His purpose, of course, is to draw a range of normalcy that minimizes the net claim amount.

We take a different approach by assessing ordinariness based on the statutory purpose of avoiding discrimination among creditors while acknowledging the need for debtors to receive normal and consistent payments on credit to avoid premature bankruptcy filings. This is the only way to produce a meaningful and fair result as is necessary to ensure equality of treatment of creditors.

Behind all statistical arrays there also must be a rationale that will move opposing counsel and/or the court to an advocate's position. To accomplish this, we take Mr. Winieck's analysis to the next level. We explore various methodologies from which to establish the parties' baseline of dealings for purposes of the subjective prong of the OCB defense. Also, we explore the rational bases for setting a range of normalcy, including reference to the statistical array of historic payments, the use of deviation percentages and accounting for the credit term as indicative of an acceptable range. Further, we discuss using a "days past due" approach when one desires an objective element in the case snapshot, or alternatively, using a "days to pay" approach when a term change occurs during or just prior to the preference period (a sign indicative of abnormality). Other statistical tools such as age bucketing, use of weighted averages, establishing running balances, evaluation of improvement in position and analysis of individual checks and associated invoices (to establish batching or unusually large payments) are necessary if one is to fully and fairly evaluate the subjective OCB defense. Finally, we consider the changes implemented by the Bankruptcy Abuse Prevention and Consumer Protection Act of 2005 (BAPCPA) on the OCB defense.

Illustrating the "Baseline of Dealings" for Purposes of the Subjective OCB Test

The OCB defense contains three elements. First, the debt at issue must have been incurred in the ordinary course of the parties' business relationship. 11 U.S.C. §547(c)(2)(A). This is rarely at issue and is usually conceded. Second, the payment must accord with the parties' prior course of conduct (the Subjective Test). 11 U.S.C. §547(c)(2)(B). Historically, this has been the most litigated aspect of the defense as it is less expensive to prove than the industry standard test, which often requires expert testimony. The third element of the defense is the so-called Objective Test, which requires that the payment at issue accord with "ordinary business terms." 11 U.S.C. §547(c)(2)(C). As BAPCPA now makes the Subjective Test and Objective Test disjunctive (the law changed the "and" between these elements to an "or"), parties are more likely to litigate this issue, and perhaps the inquiry will be more complex than was the case in the past. (See our discussion below at Part IV.)1 Thus, counsel on both sides of an avoidance action will likely focus first on the Subjective Test as the clearest indicator of whether a preferential transfer has occurred. This inquiry has several steps.

Step One: Choosing the Relevant Historical Date Range

The most common factual inquiry focuses on the difference in relative aging of payments. The greater the deviation from the historical baseline, the less likely a defendant will be able to satisfy the Subjective Test. Determining the relevant historical baseline is a creative statistical endeavor. Most attorneys simply compare the prior 12-month history in its entirety with the payments in the preference period. However, one may choose different universes of historical data (the Selected Historical Range) depending on what best suits the purpose of the inquiry, to wit:

1. the one-year prior history before the preference period;
2. the two-year prior history in toto;
3. the 21-month history (i.e., after excluding the three months before the preference period); or
4. the 18-month history (i.e., after excluding the six months before the preference period).

A 12-month period may be all the data that is available. Or, it may be preferable to using a two-year history due to a previous bankruptcy or intervening change in the debtor's business model. In a case where there has been a relatively stable business relationship, using a subset of the two-year history probably makes the most sense. Excluding the 90- or 180-day pre-preference transaction period is intended to foster a comparison to when the debtor was presumably financially healthy. Case law supports our view that the comparative historical baseline should be based on the timeframe when the debtor was financially healthy.2 Of course, any one of these methods may be chosen for tactical reasons as well.

Step Two: Selecting the Baseline within Each Historical Date Range

Once the Selected Historical Range is chosen, the next step is to compute the optimum historical weighted average (Baseline Average). In so doing, counsel should apply one of the following three variables:

1. Calculate the weighted average days past due based on every invoice paid within the Historical Range;
2. Calculate the weighted average days past due of only those invoices that were paid within the 80 percent middle band (i.e., after excluding the 10 percent of invoices on each extreme of the days past due range); or
3. Calculate the weighted average of days to pay (not by due date) by computing the invoice-to-payment days (ignoring credit terms).

Normally, the weighted average days past due creates the largest deviation, but if there was a contraction in credit terms, the invoice-to-pay method must be used to enable the most accurate comparison to the preference period.

Step Three: Applying a Deviation Percentage Swing to the Baseline Average

After the Historical Range and Baseline Average are selected, the next step in the evaluation process is to compare the selected Baseline Average to the Preference Period to arrive at a difference "by days" and in percentage terms. This is a statistical exercise, and as with all statistics, data can be presented in the manner that best suits the purposes of the inquiry. The Chart 1 demonstrates the advantages of using a days past due approach as the measure of the change:

The Historical Baseline Average (as expressed in weighted days past due) is determined by selecting a particular Historical Date Range. In Chart 2, a 30 percent +/- swing from the Historical Baseline average is applied to arrive at the day range that would be considered "ordinary" as to preference payments made within the 30 percent swing.

How we internally assign an appropriate swing percentage (in the above example 30 percent) is a trade secret. However, in general terms we determine the swing percentage based on a series of algorithms that take into account each creditor's credit term days, the "gap" in days between the credit term and the Historical Weighted Average Days Past Due and the absolute value of the Historical Weighted Average Days Past Due. For example, if the credit term is between 10-30 days, the closer the days past due number is to one credit cycle, the greater the likelihood is that the chosen percentage will approximate one standard deviation of the Historical Weighted Average. Thus, if the credit term was 20 days and the Historical Weighted Average Days Past Due was within a 15-25 day range, the percentage would be at or close to the standard deviation percentage.5 The smaller the Historical Weighted Average Days Past Due, the less meaningful is a percentage-based "swing." After all, a 50 percent swing on a Days Past Due of two days would yield a range of one to three days, irrespective of the credit term. In these scenarios, we apply an algorithm that computes a swing number by days.

Graphically Comparing the Historical and Preference Period Payments

Chart 3 demonstrates the end result of a multifaceted baseline course of dealings analysis. The green shaded areas are the conceded OCB range. We have placed a column for settlement purposes that reflects potential for recovery of preference payments that fall outside the conceded OCB range. Also, we summarize by percentage the baseline as compared with the upper and lower ranges so the viewer can see the appropriateness of the selected band.

Chart 4 is a bar graph representation of the age bucketing that compares the Preference Period dollars per age bucket with the selected Historical Baseline.

In Chart 5, we graphically demonstrate the Deviation Percentage Method and the effects of widening the swing percentage. This chart also shows an improvement in position calculation that can be used to highlight the effects of the preference period payments.6

We also have the tools necessary to explore whether the average check size and the number of invoices paid per check differed between the Preference Period and the Historical Baseline. A significant batching of invoices per check may indicate collection pressure.

Days Past Due vs. Days to Pay: Factoring Terms Changes

Where there has been a terms change in the preference period as compared with the historic baseline, a Days-to-Pay approach must be used. Otherwise, the result will mask the disparate treatment inherent in the varied term. For example, if the historic term was 60 days and the preference period term was contracted to 30 days, and if the debtor because of its computer payment system continued to issue checks based on the due dates entered in its system, a Days-Past-Due analysis would make the predominate payment pattern (in bold) appear unchanged. Using a Days-to-Pay approach demonstrates that the debtor substantially changed its payment pattern with this creditor.

Below we graphically illustrate by use of a running balance comparison how a change in terms can result in an acceleration in payments that allows the creditor to reduce its credit balance at the petition date (in this case Dec. 28, 2000).

Implications of the Objective OCB Test and BAPCPA

By changing the conjunctive "and" between the Subjective and Objective portions of the OCB defense to the disjunctive "or," Congress split the OCB exception into two separate defenses. Now, a preference defendant need only prove that the subject transaction complied with the parties' prior course of dealing, or that it comports with "ordinary business terms." At first blush, this would appear to make proving an OCB defense much simpler, which has been the opinion of many commentators. However, proving the Subjective Test is the same as it always was, and there is some indication that proving the new "Ordinary Business Terms" defense will be more difficult than proving the same element under the pre-BAPCPA OCB statute.

The first reported case to construe the post-BAPCPA ordinary-business-terms defense is In re National Gas Distributors LLC, 346 B.R. 34 (Bankr. E.D.N.C., 2006). In National Gas, the court noted that post-BAPCPA there was a separate, independent ordinary-business-terms defense and that it was required to determine whether the phrase had new import in its new context. In making this determination, the court found a "plain meaning" analysis unhelpful, as the phrase "ordinary business terms" is substantially vague. The court also found the legislative history of the BAPCPA amendment to be equally unhelpful. Thus, Judge Small made the determination based on prior case law now overlaid by the new statutory scheme that "ordinary business terms" has been released from the controlling influence of the "ordinary course of business subsection," and thus the sliding-scale approach used by many courts was no longer relevant. Id. at 404.7

The court also clarified that under BAPCPA, an ordinary-business-terms analysis not only required that both the creditor's and the debtor's industry standards be examined, but also "general business standards" as well. Judge Small stated: "If the 'ordinary business terms' defense only requires examination of the industry standards of the creditor, there would be no review or check on the debtor's conduct." Id. By holding that the terms of the transaction must be viewed in the context of the rest of the debtor's financial affairs, Judge Small has enunciated a subjective element to this otherwise objective test, which may be construed as expanding what need be proven for a creditor to avail itself of this defense. It remains to be seen if other courts follow his lead.


Mark Twain famously commented after reading his own obituary: "Reports of my death were greatly exaggerated." The same could be said regarding the belief that preference litigation will subside due to the bifurcation of the OCB defense. While the preference defendant need only prove one of the two OCB defenses, the subjective course of dealings will continue to be the dominant battle ground in preference litigation. Such battles will continue to be won by the litigant who presents the most compelling statistical analysis.



1 BAPCPA became effective as of Oct. 17, 2005, as to petitions filed after that date. Therefore, we will be living with the pre-BAPCPA statute for some time. 11 U.S.C. §546(a)(1)(A) provides a two-year statute of limitations from the order for relief. Trustees appointed or elected under one of the qualifying sections set forth in §546(a)(1)(B) may bring actions as late as one year from the appointment/election so long as the trustee was elected/appointed before the expiration of two years from the petition date.

2 See In re Carled Inc., 91 F.3d 811 (6th Cir. 1996); In re Molded Acoustical Products Inc., 18 F.3d 217, 227 (3rd Cir. 1994) ("ordinary business terms" for purposes of ordinary course of business preference avoidance exception are those that prevail in healthy, not moribund companies); In re Meridith Hoffman Partners, 12 F.3d 1549, 1553 (10th Cir. 1993) ("ordinary business terms therefore are those used in 'normal financing relations': the kinds of terms that creditors and debtors use in ordinary circumstances, when debtors are healthy"); In re Furrs Supermarkets Inc., 296 BR. 33 (Bankr. D. N.M. 2003).

3 We theorize that the debtor's deteriorating financial condition extended beyond the preference period. Transactions in the 90-180th day before bankruptcy are therefore not reflective of the prior course of dealing. By removing them the effect is to reduce the average days past due, thereby creating a larger discrepancy to the preference period. (See discussion infra).

4 We theorize for purposes of illustration only that the day range comprising 80 percent of all paid invoices constitutes the dominant historical payment practice.

5 A "standard deviation" is a statistical measure of the amount by which a set of values differs from the arithmetical mean. In plain English, when the standard deviation number is applied to both sides of the mean, the numbers that fall within the resulting range will encompass 67.76 percent of all numbers used to arrive at the mean.

6 While improvement in position may occur for nonpreferential reasons such as seasonality of purchasing of product, it can also alert one to the potential of creditor pressure, change of terms or imposition of credit limits.

7 Pre-BAPCPA, the subjective test was given primary importance, except where there was little prior history between the parties. Courts generally considered the parties' history as a better measure of what was normal, rather than some general industry standard and applied a sliding scale approach depending on the length of the parties' course of conduct. Advo-System Inc. v. Maxway Corp., 37 F.3d 1044, 1049 (4th Cir. 1994); Molded Acoustical Prods. Inc., 18 F.3d at 226 (3rd Cir. 1994); see also In re Tolona Pizza Products Corp., 3 F.3d 1029 (7th Cir. 1993).

Bankruptcy Code: 
Topic Tags: 
Journal Date: 
Friday, December 1, 2006