A PECKING ORDER APPROACH TO LEASING: THE AIRLINE INDUSTRY CASE

Journal Of Financial And Strategic Decisions
Volume 7 Number 3 Fall 1994
A PECKING ORDER APPROACH TO LEASING:
THE AIRLINE INDUSTRY CASE
Suzanne M. Erickson* and Ruben Trevino*
Abstract
This paper investigates the determinants of both short-term and long-term leasing in the airline industry.
By examining leasing within a pecking order framework, profitability and growth are introduced as
potentially important determinants of leasing. Financial leases are found to substitute for debt and to be
used relatively more by firms with higher credit risk. On the other hand, short-term operating leases do not
substitute for debt. Operating leases are used by smaller firms, non-tax paying firms and firms experiencing
more rapid sales growth. By confining the sample to one industry, asset factors which are potential
determinants of lease use can be controlled for.
INTRODUCTION
This paper examines lease use in the airline industry. The purpose is to gain insight into the determinants of
leasing and to determine whether leases are substitutes or complements for debt. Capital structure theory has
traditionally focused on the optimal levels of debt and equity. Firms trade off the tax benefits of debt for increasing
costs of financial distress as debt is added to the capital structure. However, the empirical support for the static
trade-off theory has been limited. Profitable firms, with high needs for tax shields and low probabilities of default,
have actually been shown to borrow less not more.(e.g. Titman [22], Long and Malitz [11]).
Recently, a more dynamic theory of capital structure has gained renewed interest. Myers and Majluf [17] argue
that in the presence of asymmetric information, firm's will follow a "pecking order" in raising funds; financing
first with retained earnings, then with debt and with external equity only as a last resort.
The pecking order has received strong empirical support. Baskin [3] and Toy [23] find debt ratios to be
positively related to the need for funds (growth) and negatively related to the availability of internally generated
funds (profitability).
As yet, a pecking order approach has not been applied to leasing. As a debtlike instrument, leasing is expected
to be positively related to growth and negatively related to profitability. Previous empirical studies of the
determinants of leasing have omitted profitability and growth from their models, resulting in potentially serious
misspecification problems. When the lease choice is framed within the financial pecking order, it is shown that for
firms with similar profitability and growth, leases and debt are indeed substitutes.
In this paper, the determinants of short-term operating leases (rentals) are also examined. Most studies have
focused on long-term capitalized leases. Recognizing that the motivations for the two may differ, operating leases
are not combined with financial leases but are studied separately.
This study extends the previous literature by focusing on a single industry, the airline industry, in an attempt to
control for potentially significant asset factors. The airline industry was chosen because the leased assets are
homogeneous and because prior studies have shown the airline industry to be significant in cross sectional studies
of lease use.

*Seattle University
71

---

72
Journal Of Financial And Strategic Decisions
REVIEW OF THE LITERATURE
Leasing
Traditionally, asset leasing has been considered strictly a function of a firm's tax status (Schall [19], Myers, Dill
and Bautista [16], Lewellen, Long and McConnell [9], and Miller and Upton [15]). Firms with low or zero
marginal tax brackets were assumed to favor leasing as a method of transferring unusable tax shields to tax paying
lessors. Within a given capital structure, leases were assumed to substitute for debt, although the substitution ratio
was not agreed upon.
Lewis and Schallheim [10] drop the maintained assumption of these early models, that leases and debt are
substitutes, and frame the lease choice within the optimal capital structure choice. They show that leasing can
actually increase a firm's debt capacity by selling excess non-debt tax shields. They conclude that leasing and
borrowing can be complementary within the firm's optimal capital structure.
Consistent with Lewis and Schallheim's predictions, Ang & Peterson [2] and Finucane [7] find leasing to be
positively related to the firm's debt ratio. Marston and Harris [12], on the other hand, find empirical evidence of
substitutability between leases and debt by focusing on changes in lease ratios rather than on levels.
Recently, other factors which may be important in the lease decision have been suggested. Ang and Peterson
[2], and SJL&M [20] find firm size to be important. Financial health of the lessee has been found to be important
by Ang and Peterson, SJL&M and Finucane [8]. Finucane also finds lease use to be negatively related to the
lessee's level of subordinated debt.
It is hoped that by framing leasing within the confines of the pecking order, and controlling for several of the
factors mentioned above, the issue of substitutability and the role of leasing in financing will be better understood.
The Pecking Order
Myers and Majluf [17] demonstrate that information asymmetries may cause firms to follow a pecking order
approach to financing. Due to asymmetries in the information available to managers relative to outsiders,
managers may find it optimal to maintain reserve borrowing capacity and avoid external equity markets. Their
arguments imply that firms will choose retained earnings before debt and use new stock offerings only as a last
resort. The implication of the pecking order for capital structure is that individual capital structures will reflect
historical profitability and growth rather than a predetermined optimal mix of debt and equity.
Baskin [3] provides empirical support for the pecking order among a sample of large U.S. firms.1 He finds debt
ratios to be negatively related to profitability and positively related to growth in assets. If historic profitability and
growth influence lease use as well, they must be incorporated into the leasing models. The previous leasing
literature ignores the effects of profitability and growth on leasing which results in model misspecification and
makes significance tests questionable.
THE MODEL
Within the pecking order leasing is predicted to be negatively related to profitability over time and positively
related to asset growth as debt is. Although there are no generally accepted models of the determinants of lease use,
most researchers agree on the importance of certain factors. One factor is the tax bracket of the lessee. Leasing
allows firms with low or zero marginal tax rates to transfer unusable tax shields to tax paying lessors in exchange
for lower lease payments. Thus, tax bracket is predicted to be negatively related to leasing.
A more subtle tax effect has been suggested by Erickson [6] and Lewis and Schallheim [10]. For firms with
positive tax rates, but high levels of non-debt tax shields (e.g. depreciation and amortization), additional debt may
displace existing non-debt tax shields and render them redundant due to statutory limitations on the amount of
income that can be sheltered through tax shields. Leasing provides a way to acquire an asset with debt-like
financing, while preserving the usability of the firm's existing non-debt tax shields. Thus, firms with high levels of
non-debt tax shields are predicted to lease more.
Alternatively, a negative relation is possible for non-debt tax shields when profitability is considered. Following
Baskin [3], in the model below profitability is measured as return on assets. However, if firms have high levels of

---

A Pecking Order Approach To Leasing: The Airline Industry Case
73
non-debt tax shields, ROA may be a misleading measure of available cash flows. Other things held constant, firms
with high levels of non-debt tax shields will have higher levels of internally generated cash flows and less need to
borrow or lease. Thus, leasing will be negatively related to non-debt tax shields.
A factor which both Ang and Peterson [2] and SJL&M [20] find to be important is firm size. Lessor/lenders
may choose to reduce the uncertainty surrounding their claims by leasing rather than lending to small firms.
Leasing is preferred because the lessor's security is tied to the asset itself rather than to the general credit of the
lessee.2 Thus, other things held constant, smaller firms are predicted to lease relatively more.
Like information, financial health of the lessee is likely to impact the supply of leases versus debt. Finucane [7],
Ang and Peterson [2] and SJL&M [20] all suggest that leasing should be negatively related to financial health,
although the significance of their findings vary. Since lessors are better protected than lenders in the event of
default by the lessee, leasing may be the only way to obtain asset services for firms with a significant probability of
experiencing financial distress. Leasing is therefore predicted to be positively related to the lessee's probability of
experiencing financial distress.
One factor which may explain the choice between short and long-term leasing is earning variability. For a firm
faced with a highly variable expected future cash flow stream, a short-term, cancelable lease may be preferred to a
long-term, noncancelable obligation. A positive relation is therefore predicted between earning variability and the
use of short-term rentals. Conversely, a negative relation is predicted between long-term leasing and earning
variability.
The firm's current debt ratio is included in the model to test for the substitutability of leases for debt. Other
things held constant, leasing is predicted to be negatively related to the firm's debt ratio. Finally, the beginning
total debt ratio, which includes financial leases, is included to control for historical debt levels. This variable would
be unnecessary if the data included a complete history of profitability and growth.
Most empirical leasing models assume linearity. Since there is no compelling evidence to the contrary, a linear
specification of the preceding relationships is assumed:
Equation 1
LR
π
i = ß0 + ß1 i + ß2GROWTHi + ß3SIZEi + ß4TRi +
ß5NDTSi +
-
+
-
-
+
ß6VARi + ß7PRi + ß8DRi + ß9BEGDRi + ei
-
+
-
+
where LRi is the ratio of leases to total assets for firm i, πi denotes profitability over time, TRi stands for the lessee's
marginal tax rate, NDTSi denotes non-debt tax shields, VARi denotes earnings variability, PRi represents the lessee's
probability of experiencing financial distress, DRi is the current ratio of debt to assets, BEGDRi is the ratio of debt plus
leases to assets at the beginning of the period and ei is a normally distributed, mean zero error term. Short-term leases will
be analyzed similarly but separately.
THE DATA
Erickson [6] and Finucane [7] find industry to be highly significant in cross sectional regressions of lessee
firms. The significance is undoubtedly the result of asset specific factors. Smith and Wakeman [21] and Klein,
Crawford and Alchian [8] point to asset specificity, sensitivity to use and maintenance decisions and residual value
as being potentially important asset factors in the lease versus buy decision.
The sample used here is drawn from a single industry to control for asset factors. While restricting the sample
to a single industry results in relatively small sample sizes, the restriction was necessary to control for the effects of
asset specific factors on the lease decision.
The sample was taken from the 1990 Compustat Annual Industrial File, SIC code 4500, Air Transportation.
Airlines were chosen because they are consistently heavy users of lease financing and because the assets under
lease are homogeneous. The industry includes air freight companies as well as passenger airlines, and small as well
as very large companies.

---

74
Journal Of Financial And Strategic Decisions
The sample period, 1985 - 1990, encompasses only the period since deregulation. While the post-deregulation
shake-out in the airline industry has left airline earnings depressed, the industry is not dissimilar to many other
industries reacting to increased international competition or a need for downsizing.
In constructing the sample, firms were deleted if data were missing for any of the variables, or if data was not
available for four or six consecutive years, depending on the regression. In addition, the individual variables were
screened for outliers and outliers were deleted from the sample.3
A summary of mean values over time for key variables is presented in Table 1. Forty three firms had a six year
history from which to compute mean values.
From 1985 to 1990 Total Assets and Sales increased steadily, approximately doubling. Over the same period,
the Debt/Asset ratio remained relatively constant. On the other hand, the Equity/Asset ratio and the Lease/Asset
ratio both declined, which indicates an increased reliance on short-term liabilities. Consistent with this increased
reliance on short-term liabilities firms also increased their use of rentals over the sample period.
TABLE 1
History Of Variable Means
N = 43
1990
1989
1988
1987
1986
1985
Assets
2269*
2054
1918
1657
1293
1079
Sales
2347
2176
1865
1507
1347
1222
Debt/Assets
.32
.36
.35
.35
.32
.33
Equity/Asset
.16
.23
.23
.26
.27
.22
Lease/Asset
.076
.076
.086
.085
.090
.098
Rent/Assets
.114
.113
.093
.078
.073
.070
ROA
-.013
.054
.048
.060
.075
.033
*Assets and Sales are in millions.
RESULTS
Debt And The Pecking Order
Baskin [3] uses a cross section of very large firms to test the pecking order theory, by regressing debt to asset
ratios on profitability and growth. His findings support the pecking order. Firms with higher profits (high
internally generated funds) borrow less while high growth firms (firms with high need for funds) borrow more.
Before presenting the results with leasing, it is interesting to examine the extent to which Baskin's results can
be supported in the airline industry. Table 2 contains the results of regressions of debt to asset ratios in the airline
industry on measures of profitability, growth, and beginning total debt ratios.
The pecking order predicts that a firm's use of debt will be inversely related to historical profitability. Baskin
uses ROAs at time t, t-2 and t-7 to capture the effect of historic profitability on the firm's use of debt. Exact
implementation of Baskin's methodology is problematic in this sample, however. In small samples, collinearity
between annual ROAs and degrees of freedom are significant concerns. Additionally, the post deregulation shake

---

A Pecking Order Approach To Leasing: The Airline Industry Case
75
out in the airline industry means that very few firms in this sample have a seven year history as independent
companies.
TABLE 2
Regressions Of Debt To Asset Ratios
3 YEAR
CONST
ROA3a
GROWA3
GROWS3
TDAR87b
N = 37
(1)
-.015
.233
.122***
-.092**
.649***
R 2 =.30c
(-.132)d
(.512)
(2.682)
(-2.086)
(4.023)
5 YEAR
CONST
ROA5
GROWA5
GROWS5
TDAR85
N = 30
(2)
-.077
.694
.073**
-.059*
.723***
R 2 =.36
(-.710)
(1.06)
(2.201)
(-1.821)
(4.193)
a. ROA3 = (ROA89 + ROA88 + ROA87)/3, where ROA = EBIT/TOTAL ASSETS.
b. TDAR87 = TOTAL DEBT/TOTAL ASSETS in 1987.
c. R 2 are adjusted R2 s.
d. t statistics in parentheses.
***Significant at the 1% level.
**Significant at the 5% level.
*Significant at the 10% level.
To deal with these problems average ROA over three and five year periods are substituted for annual ROAs.
Two time frames are used because the three year period provides the benefit of a larger sample size, while the five
year period provides a longer run framework more consistent with Baskin.
Average ROAs are calculated using average earnings before interest and taxes divided by average total assets.
The current period (1990) ROA is not included in the average ROA calculation to avoid injecting spurious
correlation into the model.
Recognizing that ROA may not completely capture a firm's internally generated funds, growth in sales was also
computed. Sales growth is included to capture increases in revenue and cash flow which are not accompanied by
increased profits. GROWS represents growth in sales and is the ratio of ending average sales to beginning average
sales. For example, GROWS3, the three year sales growth rate is calculated as:
((SALES90 + SALES89)/2)/((SALES87+SALES86)/2)
and similarly for the five year case.
GROWA is the proxy for growth in assets, and is calculated analogously to GROWS.
The total debt to asset ratio at the beginning of the period is included in the regression to control for the firm's
initial level of debt.4 In effect, this focuses the regression on additions to debt as they relate to growth and
profitability over the period of the regression.5 The beginning total debt ratio reflects differences in historic
profitability and growth prior to the sample period.
The results of the debt to asset regressions are presented in Table 2. The coefficients for the average ROAs are
insignificant in both regressions. A year by year look at ROAs shows that the airlines have been plagued by
fluctuating earnings.6 Over time, the positive years have been canceled by the negative years making average ROA
an insignificant source of funds.

---

76
Journal Of Financial And Strategic Decisions
Sales growth, on the other hand, is significant and negative. Consistent with the pecking order, firms with
declining revenues were found to borrow more. Finally, growth in assets is positive and significant as predicted by
the pecking order.
Overall the regressions for the airline industry in Table 2 are consistent with Baskin's results indicating that
airlines, despite their troubles, have tended to follow the prescriptions of the pecking order.
Leasing And The Pecking Order
The determinants of the ratio of leases to total assets are now discussed. Leases includes only those leases
reported in the body of the balance sheet.
Three regressions were estimated and the results are shown in Table 3. In the first regression only the "pecking
order" variables are included for comparability with Table 2. The independent variables are: average ROA, growth
in sales, growth in assets, the current debt ratio and the beginning total debt ratio. The coefficients of average ROA
and growth in assets are of the right sign but not significant at conventional levels. Growth in sales is significant
and negative as in the debt regressions. Consistent with the predictions of the pecking order, firms with higher
sales growth lease less.
Perhaps the most interesting result in equation (1) centers on the debt ratio. Finucane [7] and Ang and Peterson
[2] find leasing to be positively related to the firm's debt ratio, concluding that they are complements. The results
in Table 3 indicate that leasing is significantly and negatively related to the current debt to asset ratio. In other
words, leases and debt are substitutes.
Finally, the coefficient on the beginning total debt ratio is significant and positive. This indicates that firms
which have higher starting levels of total debt, continue to use more leasing.
In the second regression, the additional leasing determinants discussed in the previous section are included as
independent variables. The addition of these "leasing factors" more than doubles the adjusted R2 from 0.23 to 0.52.
The coefficient on average ROA remains insignificant and sales growth is still negative and significant as in
equation (1). In addition, in this more complete specification of the model, growth in assets becomes significant
and positive, as predicted by the pecking order. The current debt ratio remains negative and significant, consistent
with the substitutability hypothesis. As in equation (1) the beginning debt ratio is significant and positive.
The firm's Z score is used as a proxy for financial distress.7 A high Z score indicates a low probability of
experiencing financial distress. The Z score was chosen because it is a more comprehensive variable than single
ratio measures of liquidity and financial health that have been used previously. The Z score is negative and
significantly related to leasing as predicted. Firms in poor financial health lease relatively more.
The ratio of non-debt tax shields to total assets is significant and negative, indicating that firms with high levels
of non-debt tax shields actually lease less. Since non-debt tax shields are the sum of depreciation, amortization,
ITC and tax loss carryforwards,8 firms with high levels of non-debt tax shields will have, other things held
constant, higher levels of internally generated cash flows and less need to lease or borrow.
Size, measured as the log of total assets, is negative but not significant at conventional levels. In addition,
earning variability and the tax rate are not significantly related to the lease ratio.
The third regression in Table 3 is the same as the second regression except that sales growth, asset growth, and
profitability are estimated over five years rather than three. Likewise, the beginning total debt ratio is at the
beginning of the five year period.
The results over five years are generally consistent with the three year case although overall significance is
reduced. Increasing the sample period to five years reduces the sample size to 25, which in turn reduces the
adjusted R2 (from 0.52 to 0.23) as well as the individual t statistics.
Rents And The Pecking Order
In this section the regression results using the ratio of rents to assets as the dependant variable are discussed.
While financial leases and long-term debt represent long-term commitments by the firm, rents represent short-term
commitments. Because of this, it is reasonable to expect that the determinants of lease use may differ from the
determinants of rents. Most previous studies either ignore rents or combine them with leases. In the regressions the
rent ratio is defined as the sum of known rental commitments over the next five years as disclosed in the footnotes
to the financial statements, divided by total assets. The results are presented in Table 4.

---

A Pecking Order Approach To Leasing: The Airline Industry Case
77
TABLE 3
Regressions On Lease Ratios
Panel A
3 YEAR
CONST
ROA3
GROWA3
GROWS3
DEBTAR
TDAR87
(1)
.106*
-.162
.030
-.062***
-.240***
.271***
N=37
(1.823)
(-.674)
(1.139)
(-2.491)
(-2.589)
(2.607)
R 2 =.23
(2)
.235**
-.088
.074**
-.086***
-.492***
.367***
N=32
(2.193)
(-.301)
(2.148)
(-3.631)
(-3.868)
(3.549
R 2 =.52
5 YEAR
CONST
ROA5
GROWA5
GROWS5
DEBTAR
TDAR85
(3)
.355**
-.326
.037
-.042*
-.422**
.309
N=25
(2.210)
(-.482)
(1.471)
(-1.786)
(-2.357)
(1.702)
R 2 =.23
Panel B
3 YEAR
SIZEa
VARb
TAXc
NDTSARd
Z90
(1)
N=37
R 2 =.23
(2)
-.009
.105
.054
-.073*
-.043***
N=32
(-1.249)
(.347)
(1.49)
(-1.943)
(-3.88)
R 2 =.52
5 YEAR
SIZE
VAR
TAX
NDTSAR
Z90
(3)
-.020
-.559
.059
-.095
-.040*
N=25
(-1.532)
(-.91)
(.80)
(-1.63)
(-1.82)
R 2 =.23
a. SIZE = Log of total assets.
b. VAR = Standard deviation of EBIT from 1987 to 1990/average sales.
c. TAX = 1 if the average tax rate > 0, = 0 otherwise.
d. NDTSAR = (Depreciation + Amortization + ITC + Tax Loss Carryforwards) / Total Assets
***Significant at the 1% level.
**Significant at the 5% level.
*Significant at the 10% level.

---

78
Journal Of Financial And Strategic Decisions
TABLE 4
Regressions On Rent To Asset Ratios
Panel A
3 YEAR
CONST
ROA3
GROWA3
GROWS3
DEBTAR
TDAR87
(1)
-.045
-.014
-.245*
.521***
-.052
.123
N=35
(-.132)
(-0.012)
(-1.825)
(3.146)
(-.114)
(.233)
R 2 =.15
(2)
1.496**
-.410
-.258
.452***
.228
.020
N=30
(2.568)
(-.266)
(-1.511)
(3.053)
(.355)
(.038)
R 2 =.56
5 YEAR
CONST
ROA5
GROWA5
GROWS5
DEBTAR
TDAR85
(3)
1.644***
.370
-.216***
.240***
.422
-.927*
N=24
(4.044)
(.195)
(-3.138)
(3.856)
(.931)
(-1.876)
R 2 =.65
Panel B
3 YEAR
SIZE
VAR
TAX
NDTSAR
Z90
(1)
N=35
R 2 =.15
(2)
-.144***
-3.289**
-.324*
-.465**
.014
N=30
(-4.107)
(-2.211)
(-1.763)
(-2.434)
(.244)
R 2 =.56
5 YEAR
SIZE
VAR
TAX
NDTSAR
Z90
(3)
-.097***
-.202
-.300
-.297*
-.047
N=24
(-2.974)
(-.128)
(-1.397)
(-1.923)
(-.779)
R 2 =.65
***Significant at the 1% level.
**Significant at the 5% level.
*Significant at the 10% level.

---

A Pecking Order Approach To Leasing: The Airline Industry Case
79
As in the lease case, average ROA is insignificant in the rent regressions. Growth in assets and growth in sales
are both significant but the signs are opposite the signs found in the lease regressions. Rentals are negatively
associated with asset growth. As the rate of growth in assets reported in the balance sheet declines, rents become
relatively more important, perhaps suggesting that firms are substituting off balance sheet financing for debt or
leasing. On the other hand, rentals are positively associated with sales growth. This finding may represent the
airlines unwillingness to take on long-term financial commitments in the face of highly fluctuating revenues.
There is a significant size effect with rentals that was not evident with leasing. Small firms are making
relatively more use of short-term rentals than large firms, supporting the hypothesis that small firms may be
perceived as being riskier.
In marked contrast to the lease regressions, the debt ratio is insignificant in influencing the use of short-term
rentals. It seems that rentals are neither substitutes nor complements for debt. These results underscore the
importance of treating leases and rents separately in empirical tests.
The result regarding earning variability is somewhat puzzling. Earning variability, which was insignificant in
the lease regressions, is significant in the decision to rent but the sign is opposite that predicted. Low variability
firms actually are renting more of their assets.
The tax variable is negative and significant as predicted. Again, non-debt tax shields are significant and, as
with leases, negatively related to the rent ratio. The Z score, which was significant in the lease regressions, is
insignificant in the rent regressions.
The five year case is shown in equation (3) of Table 4. The results over five years are generally consistent with
the results over three years, although as with leases, the R2 decreases. Growth in assets is negative and significant
and sales growth is positive and significant as predicted by the pecking order. Rents are significantly negatively
related to the beginning debt ratio and size of the firm.
CONCLUSIONS
The results of the statistical tests performed here support the pecking order approach to leasing. Leases increase
with growth in assets and are inversely related to sales growth. In addition, even though lease ratios are statistically
unrelated to profitability, they are related to non-debt tax shields which may proxy for higher cash flows.
An important result of this study is that when profitability and growth are controlled for, lease ratios are
inversely related to debt ratios, consistent with the substitutability hypothesis.
In addition, other factors were found to be significant determinants of lease ratios. Faced with a choice between
a long-term lease and debt, the results indicate that leasing is preferred by firms for which the probability of
financial distress is higher (low Z score). Consistent with other empirical studies, the tax rate itself was found to be
insignificant.
On the other hand, the determinants of rentals are found to be different from the determinants of financial
leases. These results indicate that in empirical research the two should not be combined.
Rentals were found to be positively related to sales growth, indicating that rentals are supplying off balance
sheet financing for firms with growing sales. Unlike leasing, which was found to substitute for debt, rents are not
related to the current debt ratio. Furthermore, rents are not associated with poorer credit risks as leases are.
However, rentals are negatively related to size, indicating that they may be a more important source of finance for
smaller firms.
Contrary to expectations, rents are negatively related to earning variability, high variability firms are actually
renting less. Finally, rentals are negatively related to the firm's tax status, providing rare evidence consistent with
the often mentioned tax motivation for leasing.
It is clear from the results in Tables 3 and 4 that financial and operating leases are quite different instruments.
Previous leasing studies have either combined leases and rents, or ignored rental commitments completely. The
results indicate that better insights are gained by treating leases and rents separately.
Previous capital structure studies have used the pecking order approach to further understand the debt versus
equity trade-off. These results indicate that the pecking order also applies to leasing. Further research is necessary
to see if the results are supported in other industries.

---

80
Journal Of Financial And Strategic Decisions
ENDNOTES
1.
See Baskin for a review of the pecking order literature.
2.
An additional size related factor is access to capital. Small firms may have limited access to capital, or prefer to
deploy limited capital in other directions, making leasing a preferred method of asset acquisition. Both size effects
indicate that leasing should be negatively related to firm size.
3.
Criteria used for deleting outliers for each variable are available from the authors upon request.
4.
Total debt is the sum of long-term debt, including financial leases, plus short-term notes and the current portion of
long-term debt.
5.
Regressions were also run using change in lease ratios, however, since a large number of the changes were small, the
results were weaker and are not shown here.
6.
In fact, the situation in the airline industry is not so different from the auto industry, the steel industry or any of the
other basic U.S. industries faced with modernizing plants in the face of declining demand for their products.
7.
See Altman, et. al. [1] for a discussion of the Z score.
8.
Mackie-Mason distinguish between investment tax credits, which are tax shields associated with growth, and tax loss
carry forwards, tax shields associated with losses. The regressions were run with the tax shields separated by type
with no material difference.
REFERENCES
[1]
Altman, Edward I., Financial Ratios, "Discriminant Analysis and the Perception of Corporate Bankruptcy,"
Journal of Finance, September 1968, pp. 589-609.
[2]
Ang, James and Pamela P. Peterson, "The Leasing Puzzle," Journal of Finance, September 1984, pp. 1055-
1065.
[3]
Baskin, Jonathan, "An Empirical Investigation of the Pecking Order Hypothesis,” Financial Management,
Spring 1989, pp. 26-35.
[4]
Collins, Daniel W., S. P. Kothari and Judy Dawson Rayburn, "Firm Size and the Information Content of
Prices with Respect to Earnings," Journal of Accounting and Economics, 1987, Vol. 9, No. 2, pp. 111-138.
[5]
DeAngelo, Harry and Ronald Masulis, "Optimal Capital Structure Under Corporate and Personal Taxation,"
Journal of Financial Economics, 1980, Vol. 8, pp. 3-29.
[6]
Erickson, Suzanne M., "An Empirical Investigation of the Cross Sectional Determinants of Lease Use," July
1991, Seattle University, Working Paper.
[7]
Finucane, Thomas J., "Some Empirical Evidence on the Use of Financial Leases," The Journal of Financial
Research, Winter 1988, pp. 321-333.

---