Zero Leverage Firms

The Mystery of Zero Leverage Firms by Ilya A. Strebulaev and Baozhong Yang

Theme: The authors have evaluated various factors that could possibly explain the zero leverage behavior of the firms.

Theory Reference :

A) Pecking Order Theory
B) Trade off Theory
C) Information Assymetry argument
D) Market timing theory
E) Free Cash flow Argument
F) Divergence of beliefs conjecture

Data Description

A) Data sources : a) Accounting Information : COMPUSTAT
b) Equity Prices : CRSP

B) Time period : 1962-2003 split in to 1962-86 and 1987 – 2003

C) Firms Excluded : Finance companies, utilities, non US companies, non publicly traded firms and subsidiaries & firm years with book value of assets of less than $ 10 million in inflation adjusted year 2000 dollars

Initial Empirical Evidence

A) Frequency of Zero Levered Firms (Table I)

Definitions used :

D9it + D34it
BLit = __
D9it+D34it +D6it +D181it +D35it - D10it


a) Dx is a COMPUSTAT annual data item
b) D9 is the amount of long-term debt exceeding maturity of one year
c) D34 is debt in current liabilities, including the portion of long-term due within one year
d) D181 is the book value of total liabilities
e) D35 is deferred taxes and
f) D10 is preferred stock
g) D6 is total book assets

Reasons for choosing QML as the definition of leverage:

a) in line with the definition used by others
b) indicates true debt in the enumerator of leverage calculation
c) includes those firms that adopted a zero leverage policy but still have debt outstanding on their balance sheet.

a) Zero Leverage (ZL) Firm: A firm i is a ZL firm if in year t the outstanding amount of both short term and long term debt equal zero.

b) Almost Zero Leverage (AZL) Firm: Is a firm which has quasi market leverage or book leverage ratio less that 5%. ( although choice of 5% is adhoc, it represents a firm which is debt averse)

I. Column 1 represents the fraction of ZL firms relative to total sample size in each year from 1962 to 2003.
a) Average number : 8.9% firm years have zero leverage.
b) Max : 18.2% in 2003
c) Min : 4.4% in 1980
d) Implying significant variation across the years.

II. Column 2 represents the fraction of firms with zero long term debt
a) Average : 13.3% have no long term debt
b) Implying 34% of the firms with ZLTD carry short term debt ( as classified by Compustat).

III. Column 3 & 4 represent the fraction of AZL firms defined by book leverage and quasi market leverage, respectively.
a) Time consistent moderate difference between fractions of AZL firms based on book value and that on quasi market
b) Average of 22% of the firms have AZL (as defined by QML).

IV. Column 5 represents the fraction of firms that have Non Positive Net Debt
a) Between 1962 to 2003, 29.5% of the firms have NPND
b) Substantial variation noticed :
Max : 46.8% in 2003
Min : 16.9% in 1974

V. Column 1 and 5 together imply that firms consider cash to be a more important part of their balance sheet than debt
VI. Large and time consistent proportion of firms with ZL + AZL leads to question whether similar firms have different leverage
structures. Thus, proxy firm years are determined.


Descriptive Statistics of Zero Leverage and Almost Zero Leverage Firms (Table II)

Key Observations:

I. Proxy firms have higher leverage
II. Industry and size alone cannot be responsible for zero leverage behavior.
III. While descriptive statistics explains the ZL/ AZL firm characteristics with respect to their proxies quiet well, it ignores the
possibility of different firms in different industries going through different stages of life as indicated by market / book ratios
amongst others (which ones?)
IV. Since ZL/AZL firms pay higher dividends and have higher cash balances, it is possible that they are either high growth
firms / cash cows. To distinguish between the two, they are further classified in to zero dividend (high growth) and
dividend paying (cash cows).
V. Theoretically, zero leverage dividend paying firms are effectively replacing payout to debt holders with that to equity
VI. But authors cite major differences in dividend paying and non dividend firms, and hence redefine proxies to include dividend
payment / lack of it and restate descriptive statistics for each category.
VII. Theory Impact : violate trade off propositions and support Graham’s (2000) assertion that “large, profitable, liquid, in stable
industries, and face low ex ante costs of distress" (p. 1902) are under levered.

Thereafter, authors focus only on dividend paying ZL/ AZL firms.


B. Understanding Zero Leverage Behavior:
Authors have evaluated some of the reasons to try and understand the zero leverage behavior of the firms.

I) Value of potential tax benefits
Authors try to understand that despite the marginal tax rate of ZL/AZL firms being higher than that of comparable firms, why do
they not try to increase their leverage.

Authors use following assumptions in deriving potential tax benefits, which are largely on the conservative side:
a) Tax savings from carry-forward/-back provisions are largely excluded;
b) By using the minimum cash flow in the past five years as the liquidity threshold in the benchmark case, the chances that
the firm goes into liquidity crisis and incurs significant costs of distress are much lower;
c) Marginal tax rate function linearly decreases tax benefits relative to the case when it is concave (which is more likely to be
the case in practice);
d) Also, the use of average proxy leverage as the target leverage accounts for unobserved industry factors.

Results in Table III show that

a) ZL – DP firms forego, on an average, 5.7% of their market value by not using leverage.
b) If optimal levels of target leverage were to be considered (i.e. L* =min (Lm , 70%)), Column II shows that ZL / AZL – DP
give up 15.3% / 15.2% of their market value by not changing over to optimal leverage capital structure.
c) However, tax savings of ZL / AZL – ZD firms are not found to be very significant.
d) Similar results are proved in columns 3 through 8 by using some variations in definition of cash flow and interest rate.

Therefore, the fact that dividend-paying zero-leverage firms give up substantial amounts of tax benefits reinforces the mystery about their extreme debt aversion.

II) Relation between Zero Leverage and Low Leverage Puzzles

Table IV

This table signifies that even if ZL and AZL firms were to be removed from the sample set of firms, the leverage ratios of the remaining firms would not change significantly from the levels predicted by other researchers, thus indicating / suggesting that it is the “very low levered firms” rather than “normally” levered firms which would need to be studied further to explain the ‘low leverage puzzle”.

III) Size and Zero Leverage

Following from table II and the intuition that size of the firm could be an important determinant of leverage decision, size of the firm is evaluated as below. (ZL / AZL DP and ZD firms are considered separately).

Following figures explain the distribution


IV Persistence of Zero Leverage (Table IV , V and VI)

Using Monte Carlo simulation along with the regular probability, it is established that zero leverage policy is a persistent phenomena.

V) Industry and Zero Leverage Firms

Using following distribution function of the fraction of ZL firms (compared to Nj,t, the number of total firms), authors try to establish if there is any impact of industry features on zero leverage decision of the firm:

DFZLj,t={ Nj,t , if x is an element of [ ZL/N - 1/2N - ZL/N + 1/ 2n] AND [0,1]}
{ 0, otherwise}

(More weight is assigned to larger industries to maintain the sample precision of ZL firms)

a) Substantial variance noticed in zero leverage behavior across various industries
b) Prevalence of significant number of ZL/AZL firms in each industry indicates that debt aversion is a widespread rather than a
specialized phenomenon.
c) 247 industries with about 13 firms each in any given year was evaluated using Monte Carlo economy simulation is used to
suggest that there is concentration of ZL policy across various industries.
d) 75% of industries have at least 3.5% (8.5%) of ZL (AZL) firms, compared to 4.5% (12.5%) observed in the simulated

The following diagrams explain the above phenomenon:


Whether the industry proportions are stable over time, is evaluated through a persistence measure.

The mean persistence measure over all industries is 3.1% (6.5%) for ZL (AZL) firms in the COMPUSTAT sample, compared with the 7.7% (12.7%) in the simulated economy. It follows that the ZL/AZL behavior in individual industries is much more stable over time than if it were caused by random effects.

C) Testing Explanation of the puzzle

I) Long term performance of zero leverage firms (table VII, VIII and IX)

Manager of a ZL/ AZL firm believes that the firm is undervalued while capital market views the firm to be over valued if M/ B ratio of ZL/AZL firm is higher than that of the proxy. The, divergence in beliefs could be driven by both rationality and irrationality. But it leads to an empirical implication that in the long run, the value of the will be get corrected. Thus, in this subsection authors test the implication on the long-run performance to explain the ZL / AZL behavior .
The testable hypothesis is as follows: conditional long-term stock returns of ZL/AZL firms are lower than those of proxy firms.

Testing this hypothesis is complicated for at least two reasons:

First, even if the conjecture is correct, there are no statements on how long it will take the market to revert, i.e.long term" is vaguely defined.

Second, the claim is made only about the conditional returns after risk characteristics are taken into account.
First test is to compare the buy-and-hold returns for the portfolio of ZL/AZL firms and for
the portfolio of proxies for ZL/AZL firms in every cohort year in the sample. Properties of the returns on these portfolios over the horizons of three and five years is investigated.

If the difference between buy-and-hold returns of the portfolio of ZL/AZL firms and the portfolio of proxy firms is negative over a chosen period, conditional on the factor characteristics of the portfolios, it indicates support for divergence in beliefs, either rational or irrational.

Implication from table VII: the capital market overvaluation phenomenon is proved

II) Equity Issues of Zero Leverage Firms:

Two conflicting theoretical views emerge on the question whether ZL firms are expected to raise more equity than the proxy firms.
First, according to the divergence of beliefs conjecture discussed above, zero- leverage firms would prefer issuing external equity. Second, if the rationale for being zero-leverage is the presence of severe information asymmetry, firms will also avoid issuing equity-like securities, for debt is a cheaper external financing tool as implied by the pecking order argument (Myers and Majluf (1984)). Any evidence that ZL/AZL firms proceed with equity issuance, even though their debt capacity is not exhausted, is likely to be inconsistent with the pecking order conjecture. Here we investigate whether, conditional on external issuance, ZL/AZL firms issue relatively more equity-like securities than their proxy firms.

Table X compares the equity issues of DP-ZL/AZL firms with those of their proxies. IPOs are excluded from the sample set.

It proves the above conjecture wrong as the equity issues by ZL firms are lower than that by proxy firms. The second belief of information asymmetry is also proved wrong by taking equity net of share repurchases.

The robustness of the result is also proved through regression analysis of equity issues to control for differences in zero – leverage firms and proxies. Table XI shows that despite controlling for firm characteristics, zero leverage firms issue less total and net equity than proxy firms.

III) Firms with jumps in leverage (Table XII)

Authors analyze the impact of “jump” in leverage ratios i.e. either the firm stopped following ZL policy (jump up of leverage) or adopted ZL policy (jump down of leverage).

An observation is classified as a ‘jump-down’ in year t if the firm's (quasi-)market leverage ratio decreases by at least 10% from year t ¡ 1 to year t and is below 5% in years t and t + 1. Analogously, a ‘jump-up’ is said to occur in year t if the market leverage ratio is below 5% in years t ¡ 2 and t ¡ 1 and increases by at least 10% between years t ¡ 1 and t. For example, if f20%; 15%; 1%; 0%; 0%; 2%; 15%g is the sequence of market leverage ratios of a firm, then a ‘jump-down’ occurs only in year 3 and a ‘jump-up’ only in year 7. To exclude external causes, firms that file for bankruptcy, go through sizable merger and acquisition activities, or have unusually large extraordinary income and income from discontinued operations are excluded from the sample of jump-down/up firms.

Panel A reports change in firm characteristics between the jump year and previous year for jump firms and their proxies.

Panel B explains the results of the Logit model.

Implications :
Jump Down behavior does not hold with
• Asymmetric information argument
• Free cash flow explanation
• But is consistent with market timing theory
Jump up behavior is consistent with
• Trade off Theory
• Pecking Order Theoty

D) Other factors that are likely to explain the Zero Leverage behavior:
Other explanations for the presence of zero-leverage firms are as below :

a) these firms may not need to issue debt to receive tax benefits for two (possibly related) reasons.

First, these firms may in fact be substantially less profitable than the analysis of the COMPUSTAT data suggests, since the accepted GAAP measures, on which this data set is based, and the actual taxes reported to the IRS can be substantially different. It may be the case that the larger the difference between these two estimates, the more likely the firm is to follow a very conservative debt policy.

Second, these firms may have non-debt tax shields by having issued substitutes for debt such as leases or executive option compensation packages , or having entered into derivatives contracts many of which are off-balance sheet items and thus are intrinsically difficult to measure.

b) Higher cost of accessing capital markets relative to those of proxy firms.

c) Agency conflicts between managers and shareholders, large owner shareholders and marginal shareholders, institutional
holding and poor managerial compensation structures.

d) Over confidence and previous experience.

Conclusion: The empirical study fails to explain the reason for zero / low leverage cross sectional and inter temporal firms.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License