AIRA Journal

Volume

33
No. 3

WHAT’S INSIDE
Cryptocurrencies & The Dark
Web: Insolvency Considerations
Unified Loss Rules
How Can Lessons Learned in
Asbestos Help the Opioid Crisis?
Third-Party Releases?– Not So
Fast! An Update on Releases and
Warnings on Common Pitfalls
From Delaware to the Caymans
New Frontier for Fair Value Share
Appraisal Opportunities
Human Resources Levers to Drive
Transaction Value
The Altman Z Score Does Not
Predict Bankruptcy
VALCON2020 Papers

19th Annual Advanced Restructuring and Plan of Reorganization Conference

Online, NovElectronic
9 & 16,copy
2020 – Registration
available open now at www.aira.org
at: https://ssrn.com/abstract=3570149
 CONTENTS
PAGE# VOL.33: NO.3

06 Cryptocurrencies & The Dark Web:

Insolvency Considerations
Regina Lee, CIRA, and David White

15 Unified Loss Rules

Michael Barton

20
How Can Lessons Learned in Asbestos Help the Opioid
Crisis?
Jessica Horewitz, Marc Scoppettone, and Holland Sullivan

23 Third-Party Releases?– Not So Fast! An Update on

Releases and Warnings on Common Related Pitfalls
Michael A. Kaplan, Nicole Fulfree, and Colleen M. Maker

27 From Delaware to the Cayman Islands: The New Frontier

for Fair Value Share Appraisal Opportunities
Paul Madden

29 Human Resources Levers to Drive Transaction Value

Rehan Farooq and John Schultz

32 The Altman Z Score Does Not Predict Bankruptcy

J.B. Heaton

35
VALCON2020 Papers
— Expert Pitfalls and Ways to Avoid Them
— Valuing Contigent Assets and Liabilities

37 Association News

AIRA Journal is published four times a year by the Association of Insolvency and Restructuring Advisors, 221 Stewart Avenue, Suite 207, Medford, OR 97501. Copyright 2020 by the Association of Insolvency and Re-
structuring Advisors. All rights reserved. No part of this Journal may be reproduced in any form, by xerography or otherwise, or incorporated into any information retrieval systems, without written permission of the
copyright owner. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold with the understanding that the publisher is not engaged in rendering
legal, accounting or other professional service. If legal or accounting advice or other expert assistance is required, the services of a competent professional should be sought.

2 Vol. 33 No. 3 - 2020 AIRA Journal

Electronic copy available at: https://ssrn.com/abstract=3570149
 B AN KRU PT CY

THE ALTMAN Z SCORE taxes/total assets; (4) market value equity/book value of
total debt; and (5) sales/total assets.

DOES NOT PREDICT From the start, however, accounting and financial

BANKRUPTCY
researchers recognized that the Altman method did not
actually predict bankruptcy. As just one critic, Johnson
(1970),5 put it, Altman Z-scores are only “largely
J.B. HEATON descriptive statements devoid of predictive content ...
One Hat Research LLC Altman demonstrates that failed and non-failed firms
Among insolvency practitioners, the Altman Z Score have dissimilar ratios, not that ratios have predictive
(“AZS”) enjoys a reputation as a valuable indication of power. But the crucial problem is to make an inference
bankruptcy risk. That influence has allowed the AZS in the reverse direction, i.e., from ratios to failures.”
to play a role in court decisions on financial distress.1 Altman had not developed a way of determining whether
This article shows that the AZS does not work. In a a company was likely to go bankrupt, but instead
large sample of U.S. publicly traded firms from 2000- whether a certain kind of statistical analysis known as
2019, the AZS has no ability to classify firms accurately discriminant analysis could, when presented with a set
into those that will and will not file for bankruptcy in of bankrupt firms and a set of “matched” firms known
the next one or two years. The measured correlation not to go bankrupt, use financial ratios to sort the firms
between AZS and bankruptcy incidence is zero. into their correct categories of bankrupt or not. For this
The false positive rate of AZS is 98%. That is, a purpose, the statistical analysis is powerful. The error
firm with a “bad” AZS has less than a 2% chance of noted by early academic critics was the inference that
going bankrupt. This makes the AZS unreliable for Altman suggested: that the same ratios can be used to
predicting bankruptcy. predict whether a given firm not already known to be
Altman’s work was part of an effort in the late 1960s to bankrupt or not would go bankrupt.
use financial ratios to predict bankruptcy. Researchers That is, while discriminant analysis can distinguish
hypothesized that financial ratios that provided bankrupt from non-bankrupt firms when the number
advanced warning of financial distress ought to, if of each kind of firms is approximately the same and
effective, predict when firms will file for bankruptcy the non-bankrupt firms have been selected because
and when they will not. Early work by Tamari (1966),2 they were healthy and not distressed, this is practically
Beaver (1966, 1968a, 1968b)3 and Altman (1968)4 found never the question of interest in a real-world bankruptcy
that firms that filed for bankruptcy protection had worse prediction problem. In the real world, the question is
financial ratios, on average, than those that did not.
whether a firm that appears to be in some amount of
Altman’s (1968) analysis became the best-known ratio-
financial distress is likely to go bankrupt. Users of the
based bankruptcy prediction model. Altman found
AZS fall into Altman’s error, confusing the probability
that five financial ratios could, in a particular statistical
of having bad financial ratios given that a firm filed for
model, separate bankrupt from nonbankrupt firms
bankruptcy versus the probability of filing for bankruptcy
in a pre-selected and balanced sample with an equal
given that the firm has bad financial ratios. As shown
number of bankrupt and nonbankrupt firms. Altman’s
below, predicting that a given firm with bad financial
ratios are (1) working capital/total assets; (2) retained
ratios or a bad Altman “Z” score will go bankrupt is
earnings/total assets; (3) earnings before interest and
a bad prediction rule. In the real world, most all firms
1
  See, for example, Gernandt v. SandRidge Energy Inc., No. CIV-15-1001-D, 2017 with bad financial ratios (and Altman scores that predict
WL 3219490, at *2 (W.D. Okla. July 28, 2017) (“During this time, SandRidge’s
Z-Score was 0.12[.]”); Power Integrations, Inc. v. Fairchild Semiconductor Int’l,
bankruptcy) do not go bankrupt.
Inc., No. 09-CV-05235-MMC, 2017 WL 2311249, at *2 (N.D. Cal. May 26, 2017)
Data and Findings
(“Further, Power Integrations has submitted evidence that ONs ‘Altman
Z-Score,’ a ‘standard method by which to predict future bankruptcies’, was, as All data is from Bloomberg Terminal. Bloomberg
of April 2017, 1.60, which places ON ‘in Distress Zones’ and ’implies bankruptcy
possibility in the next two years’”); In re Empresas Inabon, Inc., 358 B.R. 487, 506 provides calculations of the AZS in the function
(Bankr. D.P.R. 2006) (“Robert Morris parameters and ZScore reflected a ‘highly “AZS.”6 For each year 2000 to 2017, I download data
real risk’ operation. The Altman ZScore was negative, minus 3.3”). for all Russell 3000 firms as of the last trading day
2
  Tamari, M. 1966. Financial Ratios as a Means of Forecasting Bankruptcy.
Management International Review, 6(4), 15-21. While not directly addressing of each calendar year. I remove financial companies
Altman’s work, Tamari’s work showed sensitivity to the problem addressed from the sample since Altman has described Z Score
here.
3 
Beaver, W.H. 1966. Financial Ratios as Predictors of Failure. Journal of 5
  Johnson, C.G. 1970. Ratio Analysis and the Prediction of Firm Failure. Journal
Accounting Research, 1966, 71-111; Beaver, W.H. 1968a. Alternative Accounting of Finance, 25(5), 1166-1168. See also Moyer, R.C. 1977. Forecasting Financial
Measures as Predictors of Failure. Accounting Review, 43(1), 113-222; Beaver, Failure. Financial Management, 6(1), 11-17.
W.H. 1968b. Market Prices, Financial Ratios, and the Prediction of Failure. Journal 6
  Inclusion as a function in Bloomberg Terminal is evidence of the continuing
of Accounting Research, 6(2), 179-192. demand for the AZS among even the most sophisticated financial market
4
  Altman, E.I. 1968. Financial Ratios, Discriminant Analysis and the Prediction participants. This article shows that demand to be unjustified, at least on
of Corporate Bankruptcy. Journal of Finance, 23(4), 589-609. predictive ability.

32 Vol. 33 No. 3 - 2020 AIRA Journal

Electronic copy available at: https://ssrn.com/abstract=3570149
 Table 1
Occurrence AZS < 1.81 1.81 < AZS < 2.99 2.99 < AZS
Full Data Set 4,621 4,320 16,501
Bankrupt w/in 1 Year 46 8 5
Not Bankrupt w/in 1 Year 4,575 4,312 16,496
Bankrupt w/in 2 Years 115 26 24
Not Bankrupt w/in 2 Years 4,506 4,294 16,477

Table 2
Occurrence Min AZS Median AZS Mean AZS Max AZS
Full Data Set -117.6 3.9 5.0 8,050.9
Bankrupt w/in 1 Year -8.9 0.9 1.0 22.0
Not Bankrupt w/in 1 Year -117.6 3.9 5.0 8,050.9
Bankrupt w/in 2 Years -13.8 1.2 1.1 22.0
Not Bankrupt w/in 2 Years -117.6 3.9 5.0 8,050.9

as not appropriate for such firms and since some 18% (4,621/25,442) of the dataset has AZS below the
firms are considered “too big to fail.” I also remove distress threshold, but 98% of those firms (4,506/4,621)
public utilities because, with rare exceptions mostly do not file bankruptcy within 2 years; 99% of those
in California, these are not subject to bankruptcy. firms (4,575/4,621) do not file bankruptcy within 1 year.
I remove real estate companies as well, including Table 2 presents additional data on the distribution of
real estate investment trusts, because they hold AZS in the sample. While the median, mean, and max
substantial debt against assets. I also remove firms AZS are lower for firms that go bankrupt, the minimum
with no reported debt. All of these deletions should AZS are associated with firms that do not go bankrupt
increase the power of the AZS to predict bankruptcy within one or two years. The correlation coefficient for
if it has predictive power. AZS and a 0-1 indicator for bankruptcy within 2 years is
-0.0055 and for 1 year is -0.0034. Both coefficients are
I obtain bankruptcy dates for 2001 through 2019
effectively zero.
from three sources: (1) Bloomberg Terminal; (2) the
UCLA-LoPucki Bankruptcy Research Database,7 and The likeliest explanation for the inability of AZS to
(3) searches on sec.gov. I also obtain total debt, predict bankruptcy is that financial ratios based on
price, return, index return, and equity market cap as accounting data ignore market evidence of financial
of the year end date for each firm from Bloomberg distress. Only one AZS variable - (4) market value equity/
Terminal. The final dataset contains 25,442 firm-year book value of total debt - contains any direct market
observations. There are 165 bankruptcies within two evidence. By contrast, published finance research
years of a given firm-year and 59 within one year. That (see, for example, Campbell, Hillscher, and Szilagyi
is, just 0.65% of the firm-years are associated with (2008))8 identifies several market-based variables that
a bankruptcy within two years and just 0.23% within are related to financial distress: (1) the ratio of debt to
one year, reflecting the serious “class imbalance” the sum of market value of equity and total debt (an
problem in bankruptcy detection: far more firms approximation, potentially poor however, to the ratio
do not go bankrupt than do. This class imbalance of book debt over the market value of assets); (2) a
problem is at the heart of the practical uselessness of measure of the firm’s year-to-date deviation from the
AZS for predicting bankruptcy. market return; (3) the natural log of stock price; and (4)
volatility of stock returns. In a statistical (probit model)
Table 1 sets out the distribution of AZS in the sample. estimation for one- and two-year bankruptcy, the AZS
The AZS thresholds are AZS < 1.81 for “distressed” is statistically insignificant, while (1) the ratio of debt
firms predicted to go bankrupt, 1.81 < AZS < 2.99 as to the sum of market value of equity and total debt,
a gray zone where prediction is said to be difficult, and (2) a measure of the firm’s year-to-date deviation from
2.99 < AZS which classifies firms safe from bankruptcy. the market return, and (3) the natural log of stock price
Inspection of Table 1 is sufficient to prove that the AZS are all reliably significant (volatility is not). Significantly
lacks the ability to distinguish firms that will go bankrupt greater accuracy is possible with variables that are more
within one or two years and those that will not. Fully
8
  Campbell, J.Y., J. Hillscher, and J. Szilagyi. 2008. In Search of Distress Risk.
  See https://lopucki.law.ucla.edu/.
7
Journal of Finance, 63(6), 2899-2939.

AIRA Journal Vol. 33 No. 3 - 2020 33

Electronic copy available at: https://ssrn.com/abstract=3570149
 Cont i n ue d fr o m p.33

informative of market indications of deep insolvency.9 otherwise believe. Others with exposure to a company
But the AZS is useless as a predictor. – vendors, tort claimants (such as those wronged by
Conclusion dangerous chemicals or utility-caused wildfires), and
employees and communities – all have good reasons
Despite its difficulty, accurate bankruptcy prediction
to want early and accurate warnings of company
would be of value in a number of contexts. Potential
bankruptcies that impact them. The Altman Z-Score,
creditors do not want to lend money they will not be
able to collect, and their prospects for collection - both however, is unreliable.
as to amount and timing - will depend in part on whether ABOUT THE AUTHOR
the company is likely to file for bankruptcy around the
time the debt is due. Existing creditors want notice J.B. Heaton
as early as possible that a borrower is at significant One Hat Research LLC
risk of filing for bankruptcy. Credit analysts and credit J.B. Heaton is the founder and
managing member of One Hat Research
rating agencies must have the ability to employ reliable LLC, a financial analytics developer and
bankruptcy predictors, lest they deem a company consulting firm. He received his law
creditworthy that files for bankruptcy embarrassingly degree, M.B.A. and Ph.D. (finance),
quickly after their pronouncement. Shareholders of all each from the University of Chicago. He
is widely published in law and financial
varieties, from long-term investors to high-frequency economics. Prior to founding One Hat
traders, and from longs to shorts, take positions that Research, Dr. Heaton was a litigation
may be bear more or less bankruptcy risk than they partner at Bartlit Beck LLP where he focused on financial
litigation and large damages cases. He is also a member of the
9
  See, for example, the variables identified in Heaton, J.B. 2019. Simple bars of Illinois and New York. Dr. Heaton testifies as a financial
Insolvency Detection for Publicly-Traded Firms. Business Lawyer, Vol 74(3), 723- expert in cases related to his published research.
734.

NEW SELF STUDY COURSES

If your three-year CPE reporting period ends in December and you’re a bit short, we have some options
for you to catch up. Four new self-study courses are now available, each offering one CPE credit upon
completion. Each course was a session presented during AC20 earlier this year.

NEW – Farmers in Peril

NEW – Liability Management Transactions
NEW – Industry: Franchising
NEW – Tax Attributes in Reorganization: Is There a There There?

More self study courses coming soon before the end of the year!

For more information or to purchase visit the CPE Resources section of the AIRA website at
www.aira.org/self_study

34 Vol. 33 No. 3 - 2020 AIRA Journal

Electronic copy available at: https://ssrn.com/abstract=3570149