LEXIS-NEXIS® Congressional Universe-Document
Back to Document View

LEXIS-NEXIS® Congressional


Copyright 1999 Federal News Service, Inc.  
Federal News Service

 View Related Topics 

MARCH 17, 1999, WEDNESDAY

SECTION: IN THE NEWS

LENGTH: 9091 words

HEADLINE: PREPARED STATEMENT BY
MARIANNE B. CULHANE
AND MICHAELA M. WHITE
BEFORE THE HOUSE JUDICIARY COMMITTEE
SUBCOMMITTEE ON COMMERCIAL AND ADMINISTRATIVE LAW
SUBJECT - TAKING THE NEW CONSUMER BANKRUPTCY MODEL
FOR A TEST DRIVE: MEANS-TESTING REAL CHAPTER 7 DEBTORS

BODY:

TABLE OF CONTENTS
Introduction
Part I--Our Results, Ernst & Young's Results and Why the Twain Don't Meet
Part II--Applying H.R. 3150's Means-Testing Formula to the Sample
A. The Median Income Test
B. The $50 a Month Test (Projected Monthly Net Income Test)
C. The 20% of Unsecured Debt Test Part III--Projected Net Gain and the Impossible Dreams
Conclusion
Appendix A--Meet the Can-Pay Debtors
Appendix B--Our Sample
Reprinted with Permission from the American Bankruptcy Institute Law Review
INTRODUCTION
The continued climb in personal bankruptcy filings in the late 1990's, a time of general prosperity and low unemployment, has convinced some that debtors are abusing the system, taking an "easy out" when they could, with effort, repay much of their unsecured debt. Access to Chapter 7's fresh start, it is urged, should be restricted. Others argue that most Chapter 7 filers are more burdened by debt than ever, and that major consumer creditors, especially credit card issuers, are hardly blameless victims.(2) Some holders of this view would let the market punish irresponsibility in both camps, imposing losses on lenders and reducing credit to debtors.
The debate on bankruptcy abuse has been accompanied by bankruptcy reform bills in both houses of Congress. Several of these bills include means-testing for Chapter 7 debtors,(3) a concept long advanced by the consumer credit industry.(4) Means-testing would first, develop a mathematical model to predict which Chapter 7 debtors have substantial ability to repay unsecured debt; and second, require those "can-pay debtors"(5) either to repay in Chapter 13 over five to seven years, or to forego a discharge.(6) Means-testing would deny such debtors the relatively quick Chapter 7 fresh start, which does not require use of post-petition income to repay most debts.(7) Some models of means-testing would reduce the discretion bankruptcy judges now have under section 707(b) to dismiss consumer Chapter 7 cases for "substantial abuse."(8)
VISA/U.S.A. Inc., a preeminent unsecured creditor and vigorous advocate of means-testing, has commissioned a series of empirical studies of the repayment capacity of Chapter 7 debtors.(9) The most recent of these, by the accounting firm of Ernst & Young in March, 1998,(10) followed the means-testing formula of H.R. 3150, the bill which appeared to have the greatest prospect for passage in the 105th Congress.(11) Ernst & Young initially concluded that 15% of a national proportional sample of 2,200 Chapter 7 filers were "can-pays" under that bill. They further asserted that H.R. 3150's means-testing would have allowed unsecured creditors, on a national basis, to collect $4 billion more than did the current system.(12) Ernst & Young later reduced the can-pay estimate to 11%, after H.R. 3150 was amended.(13) However, to our knowledge, they did not publicly announce any reduction in the projected $4 billion net gain. The VISA-funded studies were much criticized by academics(14) as well as by the General Accounting Office (GAO). The GAO's chief criticisms were that these studies ignored Chapter 13 administrative expenses and based their projections on two unrealistic assumptions: first, that for the next five years, each affected debtor's income would rise as quickly as debts and expenses; and second, that 100% of these debtors would complete a 60-month Chapter 13 plan. The GAO noted that current voluntary Chapter 13 plans have only a 30% completion rate.(15)
It appeared to us that one more study was needed, one not funded or controlled by creditors with a financial stake in the outcome. In late 1997, we applied for a grant from the non-profit American Bankruptcy Institute's (ABI) Endowment to support an empirical investigation of means-testing; a test drive of this new model of consumer bankruptcy. We had in hand a ready-made database of over 1,000 individual Chapter 7 cases, from seven judicial districts in seven federal circuits across the nation. These had been collected for use in an ongoing study of reaffirmation practices, funded by the National Conference of Bankruptcy Judges.(16)
In April, 1998, the ABI agreed to fund the project. In light of pending legislation and the release of Ernst & Young's March, 1998 report, we decided to apply the version of means-testing in H.R. 3150 passed by the House of Representatives in June, 1998.(17) We also decided to follow many of the steps used by Ernst & Young to facilitate comparison of results.
This article records our results, as well as how and why they might differ from Ernst & Young's. It also sets out insights gained from our test drive of means-testing. Part I is a short summary of means- testing and our conclusions. Part II covers the mechanics of means- testing: steps required, assumptions made and results obtained. Part III projects our sample's repayment capacity to the national level and examines the "impossible dreams" on which VISA's five-year estimates have been based. Appendix A contains short profiles of the can-pay debtors whom H.R. 3150 would dismiss from Chapter 7. Appendix B describes our sample design, data collection and coding.
PART I--OUR RESULTS, ERNST & YOUNG'S RESULTS AND WHY THE TWAIN DON'T MEET.
Our test drive has led us to the following conclusions:
First, abuse of Chapter 7, in the form of filings by debtors who could repay under H.R. 3150's formula, appears minimal. Only 3.6% of sample debtors emerged as apparent can-pays. Ninety-six and four-tenths percent of the sample debtors were rightly in Chapter 7.(18)
Second, sophisticated debtors could avoid can-pay status by taking on more debt or increasing charitable contributions. The 3.6% could drop once debtors adjusted to the new rules.
Third, even under the overly optimistic assumptions of the VISA studies, H.R. 3150 would allow nonpriority unsecured creditors to collect at most an additional $930 million from can-pay debtors across the nation. If our sample results held for the nation as a whole, a more realistic estimate would be $450 million, less than one-eighth of VISA's estimate.

Fourth, in operation, the necessary analysis will be costly and labor- intensive when applied to a million or more Chapter 7 cases a year. In addition to median income testing of all cases, H.R. 3150 required additional individualized scrutiny of 24% of sample Chapter 7 filings (33% in some districts) to find the 3.6% who emerged as apparent can- pays.
We did not attempt to assess the cost to taxpayers of H.R. 3150's means-testing. The Congressional Budget Office (CBO), however, estimated that H.R. 3150's provisions as a whole would have cost $214 million in the first five years, plus another $8-16 million for additional judges needed for means-testing. (19)
All the empirical studies to date agree on one point; the vast majority of Chapter 7 debtors belong in that chapter. They have too little income after necessary expenses to repay unsecured debt. It is vital, therefore, that no undue burdens be thrust on that needy majority in order to flush out a small minority of abusers.
Overview of Means-Testing Under H.R. 3150.
H.R. 3150's version of means-testing asks three questions about each individual or joint Chapter 7 case. In greatly simplified form, those three questions are:
1) Is debtor's annual gross income at least equal to the national median income for households of like size?
2) Would debtor have more than $50 of remaining income each month after living expenses and payments on secured and priority debt?
3) If debtor's remaining monthly income were applied to nonpriority unsecured debt for the next five years, would it repay at least 20% of that debt?(20)
If the answer to all three questions is yes, that debtor is a can-pay, ineligible for Chapter 7. On the other hand, if the answer to any of the questions is "No," then H.R. 3150 treats the debtor as a can't- pay, rightfully in Chapter 7.
Our results for each question are displayed in the chart below.(21)
(Chart not transmittable.)
Table 1 sets forth the results of H.R. 3150's three tests for each of the sample districts.
(Table not transmittable.)
We conclude that only 3.6%(22) of our sample are can-pays under H.R. 3150. Ernst & Young put 11% of their sample into the can-pay category.(23) Both studies assumed, unrealistically, no avoidance behavior by debtors likely to be impacted. If many debtors chose to increase debt or charitable contributions, these percentages would fall. This dramatic difference is due to a variety of factors. First, we interpreted the bill's car ownership expense allowance more broadly than did Ernst & Young. Ernst & Young allowed debt retirement only, while we added major repairs, replacement and leasing costs. Second, we estimated Chapter 13 administrative expenses and deducted these priority claims from amounts available to creditors. Ernst & Young, like earlier creditor studies, ignored these real world costs.(24) Third, we charged interest on secured claims for five years rather than just two years. Fourth, the reports are based on samples of different designs.(25) Fifth, our sample was drawn from 1995 filings, while Ernst & Young's sample was drawn from 1997 filings.
PART II--APPLYING H.R. 3150'S MEANS-TESTING FORMULA TO THE SAMPLE
A. The Median Income Test.
H.R. 3150 uses an income test to make its first cut. If required proof of income is not unduly burdensome, an income test might serve reasonably well.(26) Any means-testing system needs a mechanism that quickly, inexpensively and predictably eliminates all but the most likely candidates for can-pay status, freeing the can't-pay majority of debtors to make a fresh start. Such a front-end screen will reduce the cost and increase the efficiency of means-testing. Using a median income test at the front end to free most debtors to continue in Chapter 7 is far better than the approach taken in the Conference version of H.R. 3150, recently resurrected as the Bankruptcy Reform Act of 1999 in the 106th Congress. In these bills, all Chapter 7 cases are put through the complex Projected Monthly Net Income Test, which requires rigorous examination of individual debts and expenses. A median income test is used only at the end of the process, to determine whether creditors, in addition to judges and trustees, can move to dismiss the Chapter 7 case. (27)
Current Annual Gross Income and Household Size. H.R. 3150 compares the debtor's annual gross income to national median incomes. The test is relatively simple; the only data needed from the debtor are annual gross income and household size. The bill calculates annual gross income by first, determining the debtor's average monthly gross income for the six months preceding the petition; and second, multiplying that figure by 12 to come up with the figure to be compared with the relevant median.(28) Since the necessary data for that calculation are not supplied by the current Official Bankruptcy Forms, we, like Ernst & Young, substituted the debtor's current monthly gross income from Schedule I and multiplied it by 12.(29) Both studies determined household size by adding two to the number of dependents in joint cases and one to the number of dependents in individual cases.(30)
National Median Incomes. The bill requires use of the most recent prior year's medians available from the Census Bureau as of January 1st of the year in which the petition was filed.(31) Because our cases were filed in 1995, we used 1993 medians:
one person(32) $ 25,560
two persons 31,302
three persons 38,727
four or more persons(33) 45,161
Our Median Income Test Results. As Table 2 shows, 24% of our weighted sample(34) had incomes at or above the national medians. These results are not directly comparable to Ernst & Young's, for they followed an earlier version of H.R. 3150 which set the cut at 75% rather than 100% of the medians.(35) Forty-seven percent of their sample had incomes meeting that lower threshold.(36)
TABLE 2 THE MEDIAN INCOME TEST--WEIGHTED BY TOTAL
(Table not transmittable.)
While the majority of cases in our sample and Ernst & Young's had incomes below 75% of the medians, almost one-fourth of our sample had incomes at or above the national medians. This may seem surprising, but one should keep in mind two factors raising the numbers. First, both samples include not just wage-earners and pensioners, but also debtors who are self-employed in whole or part. Gross income for the self-employed means gross revenues from their business, before deduction of ordinary and necessary business expenses. Only after deduction of those expenses is their income comparable to that of wage-earners, because only the adjusted income is available for living expenses and debt repayment. Thus, the higher gross incomes of some self-employed debtors do not necessarily indicate greater capacity to repay.
A and more pervasive influence is the use of national medians. Debtors from relatively high cost-of-living areas, such asrancisco and Boston, are much more likely than Nebraskans or Georgians to have incomes at or above national medians, even though their higher expenses may mean San Franciscans and Bostonians are no more likely to emerge as can-pays after all three tests are done. Use of national medians works both ways, cutting out some debtors in low cost-of-living areas with incomes below national medians, who may nevertheless have repayment capacity. But the overall impact of national medians on our sample, when weighted by cases filed per district, is to increase the percentage with incomes at or above the medians. This occurs because many more Chapter 7 cases were filed in 1995 in the Northern District of California, the District of Massachusetts and the District of Colorado, some of the higher cost- of-living parts of our sample, than in the Northern District of Georgia, the District of Nebraska, the Middle District of North Carolina, and the Western District of Wisconsin.

(38)
We find the median income test to be more efficient at 100% rather than 75% of the relevant median. No doubt, there are debtors with incomes below the 100% level who currently have repayment capacity. However, they will be scarcer than at 100%, so more cases would have to be tested (and more deserved fresh starts delayed) to find each additional can-pay. Further, debtors with lower incomes are less likely to complete a 60-month plan returning any substantial amount to unsecured creditors. Setting the cut at 100% of the median will better balance the need for a quick and efficient cutoff of all but the most likely can-pays against the desire to capture the real can-pays.
We also recommend increasing the cutoff level for larger families. As noted above, H.R. 3150 applies the family of four median to all families of four or more persons.(39) Our sample included 106 cases filed by debtors with families of five or more persons. Twenty-two of these cases passed the median income test, but only one emerged as a can-pay. Thus, the family of four median required individualized scrutiny of 11 large-family cases to locate one can-pay case. A substantial increase in the minimum income figure for families of five or more would increase efficiency.
In sum, H.R. 3150's median income test is not as efficient a front-end screen as it could be. The test will require considerable stream- lining if it is to fulfill its function of identifying only the likely can-pays.
B. The $50 a Month Test (Projected Monthly Net Income Test).
The second stage of means-testing under H.R. 3150 is the Projected Monthly Net Income Test. This test is much more complex than the first, requiring detailed scrutiny of each debtor's debts and expenses. The test asks whether the debtor would have more than $50 of remaining income each month after living expenses and payments on secured and priority debt. If so, the debtor is a potential can-pay who must proceed to the third and final test. If not, the debtor is rightfully in Chapter 7. In essence, the bill presumes for purposes of means-testing that for five years the debtor will enjoy a stable income, keep expenses within an IRS budget, retain all collateral, repay 100% of non-real estate secured and priority debt in 60 equal installments and continue regular payments under the original contract on long-term real estate debt. The No Cram-Down Rule. Anyone familiar with Chapter 13 may ask, "Why assume 100% payment of non-housing secured debt?" After all, current practice is to bifurcate or "cram down" secured claims to the value of the collateral, and treat the balance of the debt in law as it is in fact, unsecured.(40) "Cram down" is one of the principal inducements to debtors to file in Chapter 13.(41)
However, the no-cram-down assumption for means-testing reflects another of H.R. 3150's proposed changes to consumer bankruptcy. One version of the bill would amend Code section 506 to prohibit cram down on purchase-money claims for "personal property acquired by the debtor within five years of the filing of the petition."(42) Instead, the value of the collateral and the allowed secured claim equal "the sum of the unpaid principal balance and interest and charges at the contract rate."(43) Is the no-cram-down provision a payoff to car and furniture lenders for not opposing means-testing?
Imagine the glee of the car financiers when (the bill) fixed the value of five year old used cars at the amount due under the contract, including contractual interest, late payment fees and collection fees. Where were Visa and Mastercard? Every dollar of "fake" value added to a car is lost for unsecured creditors in Chapter 13 cases.(44)
Means-testing, of course, aims to increase returns to unsecured creditors by pushing more debtors into Chapter 13. But Chapter 13 for secured creditors has meant cram down, waiting months for payments to begin, and years under the automatic stay. Secured creditors have fared better in Chapter 7, where the debtor may reaffirm car debt in full, and the quick discharge eases the burden of reaffirmation payments. Chapter 7's automatic stay lasts months, not years, so if the debtor fails to pay, the secured creditor can retake the collateral. The no-cram-down rule would make Chapter 13 more acceptable to secured creditors. However, it will do so at the cost of reducing payments to unsecured creditors in Chapter 13, making Chapter 13 less attractive to debtors and increasing plan failure rates.
H.R. 3150 goes even further for initial means-testing analysis. It presumes that debtors will retain all collateral and repay the full unpaid balance without cram-down. Not surprisingly, very few debtors can pay their non-housing secured debt in full in five years, cover the trustee's fees and priority taxes, and pay 20% of unsecured claims as well.
1. Living Expenses and the IRS Collection Financial Standards.
To find Projected Monthly Net Income, the bill starts with average monthly gross income for the six months prior to filing (the same income figure used in the median income test). From that figure, one deducts living expenses as well as payments on secured and priority debt. H.R. 3150 directs that living expenses for this test be based on the Collection Financial Standards (CFS) used by the Internal Revenue Service in tax payment plans.(45) H.R. 3150 turns to the CFS for guidance on reasonable living expenses, rather than relying exclusively on the debtor's scheduled expenses, which are widely viewed as inaccurate. Some say that debtors often inflate living expenses to avoid dismissal under section 707(b). Others say that when debtors' scheduled expenses are wrong, the errors are often on the low side, because many debtors do not keep track of actual amounts they spend for categories like food, clothing and transportation that require many small expenditures rather than one monthly check.
Many questions will arise with use of the CFS for means-testing. Two general questions will be treated here and others will be addressed in the discussion of the Projected Monthly Net Income Test below. The first is which version of the CFS to use. The IRS issues and updates the CFS on a somewhat irregular schedule. For example, the 1997 CFS were to be effective from and after Apr997, while the 1998 CFS were to be effective from and after October 15, 1998.(46) H.R. 3150 does not state whether one should follow IRS practice, that is, shift allowances in mid-year, or instead use the same set for the full calendar year. We assumed that, for means-testing purposes, the CFS should be used in the same way as the national income medians from the Census Bureau. That is, one should use the most recent CFS that is available January 1st for all cases filed in that calendar year. Administratively, this would be substantially simpler, especially if the IRS continues to adjust the CFS on an irregular schedule. Trustees, attorneys for debtors and creditors, makers of bankruptcy software, and others could adjust their computer programs at the same time once a year for new national medians and new CFS.
A second problem is that the CFS are not, for the most part, designed as allowances. Instead, under IRS practice at least, several CFS categories are maximums and the taxpayer is allowed only the lesser of actual expenses or the CFS limit.(47) Again, it is unclear if H.R. 3150's drafters intended to follow that IRS practice or instead, for purposes of means-testing, to use the CFS as straight allowances. Both we and Ernst & Young(48) assumed that H.R. 3150 intended the CFS to be used in means-testing as allowances, not merely as maximums. We made this assumption for several reasons. First, use as allowances addresses both of the alleged shortcomings in debtors' scheduled expenses. If the debtor's expenses are too high, the CFS will limit them. If they are unrealistically low, use of the CFS as an allowance, rather than reliance on the debtor's actual scheduled expenses, will produce a more realistic budget and, hence, generate a better prediction of ability to repay. Much of the utility of the CFS for means-testing purposes would be lost if debtors and trustees had to collect evidence of and calculate the debtors' actual expenses for categories covered by the CFS. While it is easy to determine a monthly rent or mortgage payment, it is much more difficult to keep track of all monthly transportation or food expenses for a family of five, for example.
What did we do for our sample? We did not resurrect the CFS actually in effect as of January 1, 1995. Instead, we used the CFS version promulgated in 1997 and reduced those allowances to 1994 dollars,(49) because the 1994 CFS would have been in effect as of January 1, 1995. We chose that route in case there had been any structural changes to the CFS, in addition to increased dollar amounts. Since means-testing, if adopted in a version using CFS, would have to use the post-1997 structure, it seemed best for our test drive to use the version closest to that considered for adoption.
The four CFS categories are: 1) Food and Clothing, 2) Housing and Utilities, 3) Transportation and 4) Other Necessary Expenses. The first three categories set specific monthly dollar amounts.

(50) The Other Necessary Expenses category is apparently left to the discretion of IRS agents (or in bankruptcy, perhaps the U.S. Trustee and the bankruptcy judge). It is helpful to have objective guidance on allowable expenses, but if the CFS understates actual and reasonable expenses, projected repayment capacity will be overstated under H.R. 3150.(51)
a. CFS Food and Clothing Allowance. The Food and Clothing Allowance, for "food, clothing and clothing care, housekeeping supplies, personal care products and services, and miscellaneous,"(52) presents no particular interpretive problems, because it is a national standard that does not vary by location (except for Alaska and Hawaii). The situation is not nearly so simple, however, with the rest of the CFS under H.R. 3150.
b. CFS Transportation Allowance. The CFS Transportation Allowance is subdivided into a uniform national standard for Ownership expense and local standards (adjusted for cost-of-living) for Operating expense (or Public Transportation expense if the debtor does not own or lease a car). The Operating Allowance covers "insurance, registration fees, normal maintenance, fuel,...parking and tolls,"(53) for cars owned or leased. We read the CFS to limit Ownership and Operating Allowances to one car for households of one and no more than two cars for households of two or more. Debtors who did not own or lease a car were allowed one Public Transportation Allowance regardless of family size. For larger families, these limits may understate reasonable transportation expenses.
The Ownership Allowance covers lease or purchase of up to two motor vehicles.(54) Integrating this particular CFS allowance into means- testing is complex because H.R. 3150 treats payment of secured debt separately from living expenses, while two CFS categories (Transportation and Housing and Utilities) include repayment of secured debt. Without adjustment, payments for homes and cars would be counted twice. To avoid such double-dipping, H.R. 3150 directs that one "exclud(e) payments for debts" from the CFS allowances. (55)
To follow that mandate, we first calculated a monthly payment that would amortize all secured car debt over 60 months. Next, we subtracted that monthly car payment from the Ownership portion of the Transportation Allowance. If there was a remainder, however, we treated it as an allowable expense. Ernst & Young, on the other hand, disallowed the CFS Ownership component altogether, and merely amortized existing motor vehicle debt over 60 months.
To see the dollar difference this makes over five years, look at the example in Table 3 below. The IRS in 1997 allowed $335 a month for Ownership costs, separate and apart from Operating costs. H.R. 3150 is more generous than the IRS; the bill assumes for means-testing analysis that debtors will repay all prepetition car debt (even payments for a luxury car that greatly exceed the CFS Ownership Allowance). H.R. 3150 directs one to amortize all car debt, however high, over 60 months.
TABLE 3 MEANS-TESTING AND MOTOR VEHICLES
(Table not transmittable.)
Assume Debtor A bought a luxury car just a month before filing. Debtor A's high monthly car payments exceed the IRS allowance, so neither Ernst & Young nor we would give Debtor A any Ownership Allowance. Her car is new, however, and likely to run well for five years, so that may not be unrealistic.
Debtor B, on the other hand, was more frugal; she filed bankruptcy owning an eight-year old car that was paid off. Her eight-year old car is not likely to run another five years (it would then be 13 years old!) without major repairs or replacement. Ernst & Young, however, would deny Debtor B any Ownership Allowance. Our interpretation, on the other hand, affords Debtor B her CFS Ownership Allowance, to cover a transmission transplant or a purchase at some point of a newer model. The difference over 60 months is $20,100. This is how Ernst & Young found much of the money they claim would be available for unsecured creditors--by assuming all cars, no matter how old, would run five more years. That assumption will prove false, and that money will not be available for unsecured creditors, for if the debtor cannot get to work, she cannot pay anyone.
One cannot understand the full impact of Ernst & Young's approach without a closer look at our debtors' cars, especially the age of those cars. This is old car territory, even very old car territory. Debtors in more than 80% of our 1,041 sample cases owned a car;(56) altogether they owned about 1,300 cars. Yet the debtors scheduled only 696 claims secured by motor vehicles. We at first found it hard to believe so many cars were free of debt. Further checking, however, revealed that most cars in the sample were at least five model years old when the debtors filed their Chapter 7 cases. Even older cars were not always debt-free: 267 cars with secured debt were five or more model years old at time of filing; 82 with secured debt were 10 or more years old.(57) Due to the debtors' financial distress, these cars may not have received regular maintenance. Most of these older cars would need major repairs and many might have to be replaced over the next five years.(58) The debtors need reliable transportation to earn the income which H.R. 3150 presumes will continue unabated all that time. Remember that we are forecasting car purchase/lease repair costs for five years into the future under H.R. 3150. To deny the CFS Ownership Allowance altogether, as Ernst & Young do, and amortize preexisting car debt only, seriously understates necessary and foreseeable car- related expenses. First, debtors who lease rather than buy get no allowance under Ernst & Young's method even for current monthly lease payments. Second, the many debtors who come into bankruptcy with older cars get no allowance for major repairs or eventual replacement. This interpretation increases the already substantial incentive to buy a car shortly before filing. After all, it may be the debtor's last chance for five years.(59)
To measure the impact on means-testing outcomes, we tried Ernst & Young's method on the 215 sample cases which passed the median income test. When we held all other variables constant, but denied the Ownership Allowance, the number of can-pays nearly doubled, to 70 cases. We believe that a substantial part of the difference between Ernst & Young's results and our own results is due to the treatment of motor vehicle expense.
The CFS Ownership Allowance, when used for a five-year forecast, must be read to cover not only current car debt, but also leasing, major repairs, and in some cases, eventual replacement of aging or damaged vehicles. Because Ernst & Young omit these other reasonable expenses, their report seriously overstates repayment capacity.
c. CFS Housing and Utilities Allowance. The CFS Housing and Utilities Allowance covers monthly rent, mortgage payments, utilities, property taxes, homeowner's and renter's insurance, maintenance and repairs, homeowner dues and condominium fees. It is adjusted for cost of living to the county level.(60)
H.R. 3150's mandate to exclude debt repayment from the CFS allowances applies to home mortgage payments. As with cars, H.R. 3150 ignores the CFS cap on total monthly housing expenses to allow full payment of housing debt (to the extent due under the original contract within the 60 months after the petition).
Excluding debt repayment is more complex for Housing than for Transportation, because the IRS does not subdivide the Housing and Utilities Allowance into ownership and operating portions. The IRS gives no direction as to how much of this lump-sum allowance should be preserved for utilities, maintenance, property taxes, insurance and other related expenses. We followed the method used by Ernst & Young,(61) giving renters the CFS Housing and Utility Allowance, but treating homeowners differently. For homeowners, both we and Ernst & Young substituted the amounts scheduled by the debtor for monthly mortgage payments, maintenance, property taxes and insurance (if not included in the mortgage payment), and utilities (other than cable television which we disallowed). We, like Ernst & Young, allowed these expenses in full,(62) even if they exceeded the CFS Housing and Utilities Allowance. d. CFS Other Necessary Expense Allowance. The Other Necessary Expense category has no fixed dollar amounts, and is intended to cover a wide variety of actual expenses.(63) Under this heading we, like Ernst & Young, allowed expenses as scheduled by the debtor for tax and social security withholdings, union dues, work uniforms, alimony and child support, child care, regular business expenses, life, health, and disability insurance, taxes other than those withheld by the employer (unless it appeared that the same taxes were listed as priority claims on Schedule E), charitable contributions and medical/dental expenses. We disallowed, on the other hand, debt payments withheld from the paycheck, transfers into savings plans and pension contributions except in the one case where the debtor stated the contribution was mandatory.

We also disallowed all payments for dependents not at home (other than alimony and support) and tuition payments.
As this recitation shows, use of the IRS CFS to budget for means- testing still requires detailed examination of and judgments about each debtor's scheduled expenses. This will be time-consuming, expensive, and, inevitably, the source of much litigation under any means-testing regime.
2. Additional Expenses Required by Extraordinary Circumstances and the Tithing Bill.
In addition to the CFS allowances, H.R. 3150 has a wild card category for "extraordinary circumstances."(64) To claim additional expenses under this heading, the debtor and her attorney must file a sworn statement describing the expense and the circumstances. If the trustee objects, a hearing will be held and the debtor has the burden of proof. Our finding that 3.6% are can-pays assumes that no sample debtors had expenses of this type.(65)
However, extraordinary circumstance expenses may become very ordinary indeed, under the Religious Liberty and Charitable Donation Protection Act of 1998 or "Tithing Act."(66) The Tithing Act allows debtors in Chapters 7 and 13 to donate 15% of their annual gross income (and sometimes more) to qualified churches and charities. Such donations, the Act states, are not constructive fraudulent transfers, nor are they disposable income in Chapter 13 or evidence of substantial abuse in Chapter 7.(67) If, as some commentators urge, no prior history of giving is required, the Tithing Bill gives new meaning to the phrase "eve-of-bankruptcy conversion."
H.R. 3150's section 118(d) would extend the Tithing Bill into means- testing. It directs that qualified donations up to 15% of the debtor's annual gross income "shall be considered ... additional expenses of the debtor required by extraordinary circumstances."(68) One wonders whether the right hand knew what the left hand was doing here. Means- testing is intended to benefit unsecured creditors by pushing can-pay debtors into Chapter 13. The Tithing Act and section 118 of H.R. 3150, however, seem to offer the same debtors a way out. As one commentator put it, "any attorney worth his salt is going to include within projected expenses 15% of...gross income for...contributions....To do otherwise probably would require notification to the E&O carrier of a potential claim."(69) For almost all debtors, charitable donations of 15% of gross income, on top of other expenses and debt payments, would reduce projected monthly net income to zero, enabling them to avoid can-pay status.
Our sample, like Ernst & Young's, was collected before the Tithing Act passed, so we cannot directly measure its impact. None of our sample debtors scheduled charitable contributions even close to 15% of gross income. However, we retested our 37 can-pay debtors, this time assuming they made qualified contributions of 15% of gross income in each of the five years after filing. In that scenario, well under one percent (six of 1041) had enough projected monthly net income to repay 20% of nonpriority unsecured debt.
3. Secured Debt Payments.
After deducting allowable monthly expenses from monthly gross income, we moved on to the second component of the Projected Monthly Net Income Test: monthly debt repayment. As noted above, H.R. 3150's means-testing provisions assume full amortization of priority and non- real estate secured debt, plus maintenance of regular monthly payments on real estate debt, before repayment of general unsecured debt. Thus, one must calculate total monthly payments on secured and priority debt (at least to the extent that payments would have been due within the 60 months after the petition).
a. Secured Debt. There are several problems with using current Official Bankruptcy Forms to project secured debt payments under H.R. 3150. First, neither the term nor the interest rate is disclosed for non-housing claims; only the unpaid balance is available. Second, the schedules do not clearly separate homeowners from renters, yet the distinction is important for several purposes under this analysis. One must make that judgment based on comparison of the debtor's address with real property or mobile homes on Schedules A and B and then examine Schedules C (exempt property) and D (secured claims). Third, one cannot tell whether the debtor is behind on house or car payments. In Chapter 13, such a debtor would have to make cure payments in addition to regular car or mortgage payments until the arrearage is paid.(70) For this study, we assumed "no default, no cure," and therefore only a single level monthly payment on each secured debt. However, this clearly understates required payments for debtors in default.
We followed Ernst & Young's methods for housing debt, allowing deduction of home mortgage payments from Schedule J in the living expense calculations as discussed above, and excluding home mortgage debt from the treatment for other secured debt described below. One complexity in this area is the need to separate home mortgage debt from debt secured by other real property. For home mortgages, we used as the monthly payment whatever amount the debtor entered on Schedule J for "Rent or Home Mortgage Payment."(71) However, some debtors who rent their living quarters (and thus make regular monthly rent payments) also own real estate subject to one or more mortgages. For example, 65 debtors in our sample owned (and had mortgaged) real estate other than their primary residence. They owned apartment buildings, farmland, vacation condos, and unfinished "spec" homes. If one mistakes such a debtor for a homeowner, and treats her rent as a home mortgage payment, one will omit her additional "Other Real Estate" payment, and overstate repayment capacity.
Similar problems occur for homeowners with two or more home mortgages and for owners of mobile homes when the house and lot are separately financed. Schedule J does not reveal whether the "Rent or Home Mortgage" amount is the sum of monthly payments on all the mortgages or not, and without term and interest rate data, it is impossible to distill the truth from the files.
For debts secured by real property or mobile homes other than the debtor's primary residence, only the claim amount was available from the schedules. To follow H.R. 3150's direction to project repayment only of amounts in fact due within 60 months,(72) we had to construct an artificial amortization schedule. Claims of $20,000 or more were amortized over 15 years at 9%; only the resulting level monthly payment was deducted from the debtor's monthly income. Other real estate debts of less than $20,000 were fully amortized over 60 months at 9% like non-real estate secured debt. Ernst & Young's report does not discuss treatment of non-housing real estate debt.
All non-real estate secured debt (and other real estate claims under $20,000) were treated as due within five years.(73) We generally followed Ernst & Young's methods here, grossing up the claim amount to simulate interest, then dividing by 60 to get an assumed monthly payment.
However, Ernst & Young grossed up such debt by 10%, equivalent to interest at 9% for only two years.(74) However, we are amortizing these claims over five years, not two. Therefore, we added 24% to principal, to simulate interest at 9% for five years. We suspect automobile lenders and furniture dealers will not settle for two years' interest on five-year loans. By understating required secured debt payments, Ernst & Young have overstated unsecured debt repayment capacity. We suspect that many Chapter 7 debtors have borrowed in the subprime market where interest rates are significantly higher than 9%. To that extent, we have also overstated unsecured debt repayment capacity.
4. Priority Claims.
a. Prepetition Priority Claims. Prepetition priority claims were taken from Schedule E, with care to include only that portion of the claim entitled to priority if the debtor gave that additional information. We also excluded student loans, which debtors frequently scheduled as priority rather than the general unsecured claims they are. We divided the total by 60 to get one month's prepetition priority debt payment.(75)
b. Priority Expenses of Administration in Chapter 13. H.R. 3150 directs that priority claims be "estimated,"(76) recognizing that post-petition expenses of administration are an important part of total priority debt, entitled to first priority under section 507(77) of the Code in a Chapter 13. Most creditor-funded studies, however, have omitted these expenses. For example, Ernst & Young assumed that the debtors' attorney fees would be paid in full before filing, although that is almost never true in Chapter 13.(78) As for Chapter 13 trustee's fees, their report stated only that "administrative expenses would necessarily rise for the petitions that are converted to Chapter 13."(79) They then proceeded to ignore these expenses in their calculations. We, on the other hand, estimated and deducted these expenses when calculating Projected Monthly Net Income.
Chapter 13 trustees are compensated by a percentage of payments made through Chapter 13 plans.

(80) We used the rate of 5.6% (the 1995 national average Chapter 13 trustee's fee computed as a percentage of disbursements).(81) This is a conservative figure for two reasons. First, the national average Chapter 13 trustee's fee rose after 1995, and this would have impacted these debtors in a five-year plan. Second, under H.R. 3150, Chapter 13 trustees would have additional duties, so the percentage fee might rise to cover their increased expenses. We applied the 5.6% fee to prepetition priority claims and to secured debt (other than home mortgages and other real estate claims of $20,000 or more, which we assumed would be made by the debtor outside the plan).(82)
Fee credit is frequently extended by debtors' counsel in Chapter 13 cases. In Chapter 13, unpaid fees are treated as administrative expenses and paid through the plan.(83) We assumed that attorney fees paid through the plan and the trustee fee thereon would total $800 for our sample debtors. This added $13 a month to monthly expenses in each sample case. This amount is a reasonable estimate of 1995 practice, based on a 1996 study by the National Association of Consumer Bankruptcy Attorneys (NACBA), which found that the average total attorney fee in Chapter 13 was $1,281, with $428 paid up front, and the balance of $853 paid through the plan(84) subject to the trustee's percentage fee.
Final Calculation of Projected Monthly Net Income.
Once living expenses, extraordinary circumstances expenses and payments on secured and priority debt (including administrative expenses) have all been deducted from monthly gross income,(85) the remainder, if any, is called Projected Monthly Net Income by H.R. 3150.(86) If the debtor's Projected Monthly Net Income is $50 a month or less, the debtor is eligible to remain in Chapter 7. On the other hand, if the debtor has more than $50 a month remaining, the debtor must proceed to the third and final portion of H.R. 3150's three-part means-test.(87)
Table 4 shows how our sample fared on this test. Only 45 of our sample debtors had Projected Monthly Net Income over $50, for a weighted pass rate of 4.07%. Ernst & Young, by contrast, found that 17% of their sample had more than $50 Projected Monthly Net Income.(88) We believe the differences are due to use of different medians, different treatment of motor vehicle expenses, interest on secured debt for five rather than two years and our inclusion of Chapter 13 administrative expenses. TABLE 4 THE PROJECTED MONTHLY NET INCOME TEST-- WEIGHTED BY TOTAL CHAPTER 7 NON-BUSINESS FILINGS
(Table not transmittable.)
We have serious questions as to the feasibility and cost-effectiveness of applying this complex test to hundreds of thousands of Chapter 7 filers each year. As has been shown, H.R. 3150's Projected Monthly Net Income test is labor intensive, and heavily dependent upon the accuracy of debtor's schedules regarding debts and expenses. Sophisticated debtors can manipulate the outcome by increasing their secured debt and charitable contributions.
C. The 20% of Unsecured Debt Test.
The third, final and mercifully simpler test under H.R. 3150 is whether the debtor's Projected Monthly Net Income is sufficient to repay at least 20% of her nonpriority unsecured debt over five years.(90) The steps we took to apply this test follow.
First, we multiplied Projected Monthly Net Income by 60 to get the total amount available for general unsecured claims. Second, because Chapter 13 trustees collect fees on payments to nonpriority unsecured creditors, we multiplied that total by .9469 (in effect reducing it to reflect the trustee's 5.6% fee). Finally, we divided this adjusted Projected Five-Year Net Income by total nonpriority unsecured debt(91) to project the percentage the debtor could repay over 60 months.
As Table 5 shows, only 37 of our sample debtors, or 3.55% on a weighted basis, had sufficient income to meet or exceed H.R. 3150's 20% threshold. Ernst & Young reported that 15% of their sample emerged as can-pays when it used 75% of the median. At 100% of the median, 11% of their sample were can-pays.(92)
TABLE 5 THE 20% OF UNSECURED DEBT TEST-- WEIGHTED BY TOTAL CHAPTER 7 NON-BUSINESS FILINGS
(Table not transmittable.)
The Can-Pay Debtors
The 3.6% of our sample identified as can-pays have median annual gross incomes of $52,080, almost two and one-half times the median income of the can't-pays ($20,688), and more than twice that of the whole sample ($21,264). The can-pays' median incomes are also well above the 1995 national median for all families ($40,611). See Tables 6 and 7 below, which use the term "impacted filers" for can-pays.
(Tables not transmittable.)
Median nonpriority unsecured debt among the can-pay debtors was $33,526, while for the rest of the sample, the median was $20,303. Who are these can-pays? In Appendix A, we profile each can-pay case. These debtors are truck drivers and engineers, mechanics and restaurant managers, a young couple expecting twins, retirees with part-time jobs, one man with a serious gambling problem, a married couple who ended up deeply in debt by caring for and eventually burying their parents. Several got into financial trouble when their small business failed. At least one family matches the stereotypical abuser, with income of $90,000, homes in California and Hawaii and $72,000 in credit card debt.
Many of the can-pays had the apparent capacity to repay well over 20%; in fact, average repayment capacity for nonpriority unsecured debt was 74.6%, and 14 of the 37 could repay 100%. However, these repayment percentages are merely forecasts based on the same shaky foundation of the VISA report: that for five years, these debtors' income will rise as quickly as expenses and debts. We critically examine that premise in the next section.
PART III--PROJECTED NET GAIN AND THE IMPOSSIBLE DREAMS.
How much more might unsecured creditors collect under means-testing? VISA estimated that number at "over $4 billion,"(94) then qualified that estimate by stating "(t)his assumes that income remained unchanged relative to expenses and liabilities during the 60 month repayment period."(95)
In 1997, some 926,000 nonbusiness Chapter 7 cases were filed. If we assume that our sample holds for the nation as a whole and that 3.55% or 32,873 were can-pays, each with nonpriority unsecured debt of $35,303,(96) and that each could repay 75% or $26,477 of that unsecured debt. If all 32,873 debtors repaid $26,477 apiece, it would total about $870 million, nowhere close to VISA's estimate. However, even that $870 million is based on at least four unrealistic assumptions:
First, that well-counseled debtors will not evade can-pay status by increasing debt or charitable contributions;
Second, that the debtors' incomes, expenses and debts will remain relatively unchanged for five years;
Third, that 100% of the can-pays will file and complete five-year Chapter 13 plans; and
Fourth, that unsecured creditors will bear no part of the cost to sort all Chapter 7 cases for means-testing and to monitor the 30,000 + can- pays over five years in Chapter 13.
Fifth, that unsecured creditors collect nothing from Chapter 7 debtors at present.
None of these assumptions is well-founded. The first four are impossible dreams and the last is simply untrue. Let's start with the first, the assumption that well-counseled debtors will not choose to avoid can-pay status by increasing debt or charitable contributions prior to filing. As discussed above, the Tithing Bill allows debtors to divert up to 15% of annual gross income into H.R. 3150's extraordinary expense category. We also pointed out the incentives to increase secured debt, especially by buying a car, prior to filing. Many debtors with above-median income could and some would use these routes to escape five years of payments in Chapter 13. If we conservatively assume that only one-quarter of the can-pays would do so, that reduces the can-pay pool to 25,000 debtors and reduces net gain to about $700 million.

Now let's consider the likelihood that for five years, the incomes, expenses and debts of the can-pay debtors will remain relatively stable. Truly, this is an impossible dream. Over five years, of course, some will die. The rest will get five years older and many will be deeper in debt. Like other human beings, the can-pays will suffer disease, divorce, disability and downsizing that for some will substantially reduce income and increase expenses. Even delightful events like new babies entail years of additional expenses and, eventually, student loans. None of these events is stopped by the automatic stay. As the GAO remarked with reference to VISA's studies, there is "no empirical basis for assuming five years of stable income and expenses."(97)
Another impossible dream is that the can-pays would file and complete a five-year Chapter 13 plan. It is common knowledge that current voluntary Chapter 13 plans have only a 30% completion rate, and many of these were just three-year plans to begin with.(98) To be sure, the can-pays would be a select group, with above-average ability to repay at the outset, so one could expect a better completion rate, perhaps as high as 50%. But for all the reasons set forth above, plus sheer bad luck and for some, bad habits, many will not finish their plans.(99) If only half of these 25,000 plans fail, after making payments for an average of two and one-half years, that would reduce net gain to about $525 million.
The fourth impossible dream is that it will cost unsecured creditors nothing to sort a million Chapter 7 cases a year into can- and can't- pays. Perhaps all these costs can be shifted to taxpayers, but there will be real costs, as the CBO has shown,(100) and unsecured creditors are taxpayers too. But let's be conservative here too. The CBO estimates that H.R. 3150's provisions as a whole would cost $214 million in the first five years, plus another $8-16 million a year for additional judges needed for means-testing.(101) Assume that only $25 million of these costs are born by unsecured creditors either directly due to higher Chapter 13 administrative costs, or indirectly through increased taxes. That brings us down to $500 million.
Finally, the VISA studies also assume that unsecured creditors collect nothing from Chapter 7 debtors at present. This is false. Chapter 7 debtors frequently reaffirm debts scheduled as unsecured.(102) They must also repay student loans and other nondischargeable unsecured debts. Let's assume that all such payments by 1997 Chapter 7 filers amount to $50 million. This $50 million is not new money dependent on means-testing; unsecured creditors already collect that much. So we subtract $50 million from our $500 million, giving us a projected net gain of $450 million.
In seeking that pot of gold from the can-pay minority, however, it is vital that we not bar the many can't-pays from a fresh start. Means- testing would be mean indeed if it hurt the many to capture the few.
CONCLUSION
Our intent in this article is to assess, by means of a test drive of sorts, whether one suggested legislative response to perceived bankruptcy abuse would fulfill its apparent aims. That is, would H.R. 3150's formula find debtors with ability to repay and divert them into Chapter 13 plans that would produce a net gain to unsecured creditors without undue cost to other debtors and taxpayers?
As we have shown, H.R. 3150's formula produces relatively few apparent abusers for diversion into Chapter 13.(103) Second, because the sample cases were filed in 1995, when the only sanction for abuse was section 707(b) dismissal, debtors and their attorneys had less reason to take evasive maneuvers. At this writing, of course, the Religious Liberty and Charitable Donation Protection Act of 1998 is in force, providing an escape route for all but the wealthiest few. Further, incentives to load up on secured debt, especially by purchasing a new car shortly before filing, are substantially increased. Thus, there is reason to believe that the number of can-pays in our sample is substantially higher than would be found today if means-testing were adopted. Third, the assumed benefits to unsecured creditors from means-testing depend very substantially on whether those can-pay debtors in fact convert to and complete lengthy Chapter 13 plans. Chapter 13's recent history provides little reason to believe that even a bare majority of the debtors would in fact complete their plans. The net gains to unsecured creditors, in sum, appear small relative to the costs likely to be imposed on the great majority of Chapter 7 debtors, as well as trustees, judges and taxpayers. In sum, we conclude that means-testing as enshrined in H.R. 3150 will not go the distance.
END


LOAD-DATE: March 20, 1999