This study uses panel data constructed from the Standard and Poor’s COMPUSTAT of North American Database. Included in this sample are annual observations from 1996-2005 for the firms with SIC code between 0100 and 5999 on the active files. We exclude the financial industry because the observations are unavailable in the database. We also exclude the public administration industry because this industry is controlled by the government. The full sample contains 2710 observations according to above collection criteria.
Based on the accounting standard of operating leases3, we defined “shorter” debt as debt mature shorter than five years and “longer” debt as debt mature longer than five years. Our definition of the short-term debt is measured as the ratio of book value of total shorter debt excluding capital leases to the market value of the firm. The market value of the firm is the book value of total debt plus the market value of equity and plus operating leases. The short-term debt (STDR) is as follows,
leases operating equity
of value market debt
total of value book
leases capital debt
shorter
STDR + +
= − (3.1.1)
Long-term debt is measured as the ratio of book value of total longer debt excluding
3 The accounting standard is presented in Appendix B.
capital leases to the market value of the firm. Based on the accounting standard of operating leases, we calculate the below five years debt as the short-term debt and above five years debt as the long-term debt. The long-term debt (LTDR) is as follows,
leases operating equity
of value market debt
total of value book
leases capital debt
longer
LTDR + +
= − (3.1.2)
As to security provision, followed by Stulz and Johnson (1985), the secured debt is measured as the ratio of book value of total secured debt excluding capital leases to the market value of the firm. The secured debt (SDR) is as follows,
leases operating
equity of value market debt
total of value book
leases capital debt
SDR ured
+ +
= sec −
(3.1.3)
Unsecured debt is measured as the ratio of book value of total unsecured debt to the market value of the firm. The unsecured debt (USDR) is as follows,
leases operating equity
of value market debt
total of value book
debt ured USDR un
+
= + sec
(3.1.4) Operating leases is measured as the ratio of current year rental expense plus present value of rental commitments over the next five years (discounted at 6.6 percent)4 to the market value of the firm. In the remainder of this section, we describe the variables which are used to explain the relationship between debt and lease financing.
3.2 Explanatory variables to Debt and Lease Financing
In the following section, we describe the measurement of the explanatory variables to debt and lease financing.
(1) Non-debt tax shield (NDT)
To measure the non-debt tax shield, we include the expense which could create the non-debt tax shield, such as depreciation expenses. The measurement of NDT is as
4 Because there are many missing values of short-term borrowing rate, we use the average borrowing rate as the short-term borrowing rate for all firms.
follows,
TA
TAX IE ITAX EBDIT
NDT
−
−
=
, (3.2.1) where IE is the interest expenses, ITAX is the income tax amounts of the specific year, TAX is the tax rate, and TA is the total asset.
(2) Financial distress (Z_score)
We use the modified version of Altman’s Z-score to measure the ex post financial distress. Our measurement of Z-score5 is as follows,
TA
WC TA
RE TA
Sales TA
score EBIT
Z − =3.3 +1.0 +1.4 +1.2
, (3.2.2) where RE is the retained earnings, WC is the working capital, and TA is the total asset.
A higher value of Z_score indicates a low level of financial distress.
(3) Collateral
Since fixed assets are more valuable in liquidation and support a higher external obligation capacity, we measures collateral as net property, plants, and equipment divided by total assets. The measurement of collateral is as follows,
TA E Collataral PP&
= (3.2.3) (4) Uniqueness
The lessor and debtholders probably suffer relatively high cost in the event that liquidate the firms of produce unique or specialized products. Hence, we measure uniqueness as research and development expenses divided by total assets. The measurement of uniqueness is as follows,
TA enses D
Uniqueness R& exp
= (3.2.4) (5) Firm size (Size)
Firm size is a proxy for the quality of outsider’s information about a firm’s
5 The calculation of Z_score is the same to the calculation of Graham, Lemmon and Schallheim(1998).
operation and prospect. Thus, we define firm size as the natural logarithm of the total assets. The measurement of firm size is as follows,
Size=ln TA( ) (3.2.5) (6) Profitability (prof)
According to Pecking-Order theory, a firm may reduce the external funds if the firm has sufficient retained earnings for new investment. We use the ratio of operating income over total assets to measure profitability. The measurement of profitability is given as follows,
TA
income operating
prof = (3.2.6) (7) Cash Dividend (Dv)
Since firms that do not pay cash dividends are likely to be burdened by asymmetric information, we construct a dummy variable to investigate the effect of dividend policy. The dummy variable is set to one if the firms paid no cash dividend and zero otherwise. The measurement of cash dividend is as follows,
Dv = 1 if the firm does not pay cash dividend
0 otherwise (3.2.7) (8) Industry Effect
Several researchers have documented industry effect associated with debt and leasing policy. To control for industry effect in the regression, we include three dummy variables for each one-digit SIC code grouping. The first dummy is set to one for firms with SIC code between 0100 and 1999 and zero otherwise; the second dummy is set to one for firms with SIC code between 2000 and 3999 and zero otherwise; the third dummy is set to one for firms with SIC code between 4000 and 4999 and zero otherwise. The measurement of profitability Dvis given as follows,
Dummy1= 1 if 0≦SIC≦1999
0 otherwise Dummy2= 1 if 2000≦SIC≦3999
0 otherwise (3.2.7) Dummy3= 1 if 4000≦SIC≦4999
0 otherwise
Table 1 reports descriptive statistics for all the variables. Operating leases account for 2.1 percent of the firm value on average; short-term debt, long-term debt, secured debt, and unsecured debt account for 19.7 percent, 5.4 percent, 6.7 percent, and 1.1 percent of firm value, respectively. There are 245 missing numbers of operating leases. These numerous missing numbers represent that operating leases are censored data.
Table 2 shows the Pearson correlation coefficients of all variables. These positive coefficients between operating leases and various debts as short-term debt, long-term debt, and secured debt seem to suggest that leases and debt are complements. And the negative coefficient between operating leases and unsecured debt seems to imply that the lease and debt are substitutes. Later, we further investigate the relationship with appropriate control for the heteroskedasticity of disturbance term’s variance and the endogeneity of debt. We also find that both leases and debt are significantly correlated with most of control variables. These correlations exhibit that proper control of control variables is potentially important for an accurate measurement of the relationship between leases and debt.