Tải bản đầy đủ

Determinants of capital structure an empirical study of american companies

DETERMINANTS OF CAPITAL STRUCTURE
AN EMPIRICAL STUDY OF AMERICAN COMPANIES

Supervisor:
Author:

Dr. Cormac Mac Fhionnlaoich
Nguyen Thi Dieu Hong

August 31st 2016


Abstract
This study investigates the factors that potentially have impact on capital
structure decisions of American firms, and identify the key determinants of the
capital structures of these firms. The paper also explores the capital structure
theories and how they explain capital structure decisions of firms worldwide
and in the U.S. The sample of includes 1.500 U.S firms, which covers 90%
publicly-traded companies in the U.S during post-financial crisis time, from
2010 to 2016. Using panel data techniques with fixed-effects model and
random-effects model, firms ‘characteristics are tested if they explain for

leverage ratios. The explanatory variables represent the factors that potentially
determine capital structure: business risk, profitability, firm size, growth
opportunities, tangibility of assets, non-debt tax shields.
This study finds that the most reliable and important factors that
determine the use of debt by American listed firms are firm size (+), tangibility
of assets (+), profitability (–). Besides, the moderately influential factors of
leverage includes: business risk (+/–), non-debt tax shield (+/–) and growth
opportunities (+/–). The study finds evidences which are consistent with
pecking-order theory’ prediction of a positive relationship between asset
tangibility and financial leverage and a negative relationship between
profitability and financial leverage. The finding moderately supports trade-off
theory’s prediction of negative relationship between non-debt tax shield and
leverage, business risk and leverage. The trade-off suggestion of a positive
relationship between asset tangibility and financial leverage are also confirmed
by this study. Finally, agency’s prediction of a negative relationship between
growth opportunities and leverage is moderately supported by a negative and
insignificant relationship found in this study.


Acknowledgement
In the last three months, I have been lucky to receive great support from
many people who have helped to make this study possible.
At first, I would like to extend the most sincere gratitude to Dr.Cormac
Mac Fhionnlaoich, UCD Michael Smurfit Graduate Business School, for
providing supporting during the process of conducting this research.
Dr.Cormac Mac Fhionnlaoich has shown great support in guiding me through
this project, providing detailed comments and thoughtful suggestions in my
completion of the project.
Furthermore, I would like to send my appreciation to Irish Aid, The UCD
Michael Smurfit Graduate Business School staff and lectures, and ICOS for
giving the scholarship and support me during my academic year in Ireland.
I also would like to place on the record my sincere gratitude to my
IDEAS Fellows, in particular, Pham Khanh Linh, for his guide on data solving
and Nguyen Thi Phuong Thao for her company and encouragement.
Last but not least, I would like to show my great appreciation family and
friends for their support and constant care for me on the way to my completion
of the master program.


Contents

o
1. INTRODUCTION ........................................................................................... 1
2. LITERATURE REVIEWS .............................................................................. 3
2.1. Modigliani-Miller theorem ........................................................................ 3
2.2 Tradeoff theory ........................................................................................ 3
2.3. Pecking order theory ............................................................................... 7
2.5. Determinants of capital structure ............................................................ 8
2.5.1. Size .................................................................................................. 8
2.5.2. Tangible assets ................................................................................ 9
2.5.3. Profitability ...................................................................................... 10
2.5.4. Growth............................................................................................ 10
2.5.5. Non-debt taxed shield .................................................................... 11
2.5.6. Risk ................................................................................................ 12
3. DATA AND METHOLOGY .......................................................................... 15
3.1 Data Description .................................................................................... 15
3.2 Panel data regression model ................................................................. 15
3.2.1 Panel data ....................................................................................... 15
3.2.2. Definition of variables ..................................................................... 17
3.2.3. Model ............................................................................................. 20
4. RESULT ...................................................................................................... 21
4.1 Data descriptive ..................................................................................... 21
4.2 Correlation Test ..................................................................................... 22
4.3. Test of determinants of capital structure ............................................... 24
5. CONCLUSION, LIMITATION AND SUGGESTED FUTURE WORKS ........ 34
5.1 Conclusion ............................................................................................. 34
5.2 Limitation ............................................................................................... 35
5.3 Suggestions for future research ............................................................. 37
REFERENCES ................................................................................................ 38


List of tables
Table 1: Predicted effects on leverage based on capital structure theories .... 13
Table 2: Determinants of capital structure in previous researches .................. 14
Table 3: Measures of leverage ........................................................................ 17
Table 4: Measures of capital structures determinants ..................................... 19
Table 5: Descriptive statistics of the sample ................................................... 21
Table 6: Correlation of variables...................................................................... 22
Table 7: Fixed-effects regression results......................................................... 24
Table 8: Random effects regression results .................................................... 25
Table 9: Hausman test results ......................................................................... 25
Table 10: Regression with selected factors ..................................................... 27
Table 11: Modified Wald test results for heteroskedasticity ............................. 27
Table 12: Wooldridge test results for autocorrelation ...................................... 28
Table 13: Fixed-effects regression with Driscoll-Kraay standard errors .......... 29


1. INTRODUCTION
Capital structure is one of the most important decisions of every
company. A false decision in capital structure can lead a firm to severe
difficulties. Managers always want to find a suitable capital structure policy to
meet their goals. Researchers, from another aspect, are curious to know how
firms choose of sources of financing, do they have target structure and what
factors affect firms’ decisions. That is the reason makes the capital structure to
become one of the most important fields of corporate finance. Modern
corporate capital structure theory originated by Modigliani and Miller’s (1985)
with irrelevance theorem. The idea of the theory is that in an efficient market
with the absence of taxes, agency costs, bankruptcy, and asymmetric
information, how a firm is financed does not affect its value. Since the value of
the company depends neither on its dividend policy nor its source of capital,
the Modigliani, and Miller theorem is often called capital irrelevance principle.
This theory is considered to lay stones for many followers to study capital
structure.
However, it is clear that Modigliani and Miller’s assumptions are
unrealistic and hard to happen in the real market. Following Modigliani and
Miller (1958), several other theories have been developed on the topic of
capital structure. The trade-off theory states that companies choose an amount
of debt finance and equity finance base on the balance of benefit and financial
cost. The pecking order theory focusses on asymmetric information with
considering that the cost of finance increases with asymmetric. Therefore, firms
will choose to internal financing as the first priority, then debt and equity as a
‘’last resort’’. Another stream of research was initiated by Ross (1977) on how
the choice of a firm’s capital structure can signal information to outside
investors about the company, i.e. issuing large debt levels is a signal of higher
quality of the firm.
With regards to empirical work, many studies were done to find an
answer to the capital structure puzzle. Concerning to U.S firms, one of the
earliest attempts to extend empirical study on capital structure was conducted
1


by Titman and Wessels (1988). A large set of data of U.S. companies between
1974 and 1982 was used to examine theoretical determinants of capital
structure. Following this, Rajan and Zingales (1995) investigated the
determinants of capital structure decisions on a broader scope in G-7 countries
with more focus on U.S firms.
It can be clearly seen that both theoretical and empirical work has made
progress in investigating which factors influence capital structure decisions.
Yet, Titman and Wessels (1988), Rajan and Zingales (1995), and Harris and
Raviv (1991) agreed on the fact that, while progress has been made from the
initial work of Modigliani and Miller in 1958, the empirical work was lagging
behind and doing very little to identify empirical findings of capital structure in
practice. While theoretical work had identified a large number of potential
determinants of capital structure, empirical studies have not frequently
considered various contexts outside the G-7 countries.
In recent years, empirical studies on capital structure determinants have
been largely extended to different developed and developing countries
including Malaisia, (Pandey, 2001), India (Joy Pathak, 2010), Portugal (Vergas,
Cerqueira, Brandão 2015), Sweden (Han-Suck Song, 2005). They pointed out
both similarities and discrepancies in what factors influence firm financing
decisions across different contexts.
By updating data, applying the methodology used for the panel data that
has been improved upon and updated with a thorough analysis of different
models and using four kinds of leverage ratios, the study wants to find out
which factors are important in the capital structure decisions of U.S in recent
time, especially after the financial crisis time.

2


2. LITERATURE REVIEWS
2.1. Modigliani-Miller theorem
Fifty-eight years ago (1958), two economists Franco Modigliani and
Merton Miller proposed Modigliani-Miller theorem on the capital structure which
plays an essential role in modern corporate finance. Before them, no widelyaccepted theory of capital structure has existed. The theorem states that in a
perfect market without taxes, asymmetric information, bankruptcy cost, and
agency costs, the way in which a firm raise its capital makes no influence on its
value. This suggests that the valuation of a firm is irrelevant to its capital
structure, so the theorem is also called capital structure irrelevance principle.
The assumptions of the theory are based on an efficient market. First,
there are no taxes. Second, there is no transactions costs and bankruptcy
costs. Third, the information is symmetric, all investors are rational and have
the same access to information. Fourth, the costs of debt are the same for
everyone and last, debt financing do not affect firms
The fact is that it is very hard to find a perfect market. In reality,
corporations do business in a market containing transaction costs, borrowings
costs, taxes, asymmetric information and agency costs. By relaxing some
assumption in Modigliani-Miller theory, some alternative theories were
proposed to address these imperfections.
2.2 Tradeoff theory
The Static tradeoff theory
The basic idea of trade-off theory is that that firms follow an optimal
capital structure to maximize value by offsetting the cost of the additional unit
of debt by its benefit. Baxter (1967) and Kraus and Litzenberger (1973) stated
that a taxable corporation should consider an increase in its debt level until
there is a balance between the marginal value of tax shield and the present
value of any financial distress costs occurred.
Trade-off model with bankruptcy costs
3


When firms borrow, they get a tax advantage as interest is deductible for
income tax. Besides, firms have to incur bankruptcy cost of debt. Companies
using leverage need to pay interest on their borrowings. This changes
companies’ earnings and cash flow. The more firms borrow, the higher
probability of bankruptcy increases. The trade-off theory predicts that firms
choose an optimal capital structure to balance tax benefits and cost of debt.
Companies substitute debt with equity or versus while maximizing the
company’s value.

Trade-off model with agency costs
Jensen and Meckling (1976) notice that debt had been used widely
before the appearance of subsidies tax on interest payment, given that there
must be other important factors of capital structure that have not been
recognized. Two kinds of agency costs were suggested that is the gap
between shareholders and managers and conflict between shareholders and
creditors.
Agency cost between shareholders and managers: This type of agency
costs appear when there is a separation between ownership and management.
When shareholders lose control, sometimes managers have opportunities to
put their benefits above shareholders. Instead of always making decisions to
maximize the market value of the firms, managers may make the inefficient
allocation of capital. For example, managers may pursue growth and size at
4


the expense of profitability and value by investing in unprofitable projects.
Some managers prefer managing a bigger, and more influential firms have less
incentive to act for benefits of shareholders. If increasing using debt, firms have
to pay interest payment and by that reduce free cash flow within the firms. As a
result, there is a deterioration of liquidity that allows managers to take part in
projects that the profit maximization (Jensen, 1986).
Agency cost between shareholders and creditors: The shareholder's
attempt to engage in new projects that generate more benefits for shareholder
while posing higher risks to the firm’s creditors If the risky capital investment
project is successful, shareholders will gain more rate of return. Lender’s
benefit does not change because the interest rate is fixed. If the project fails,
the creditors are forced to share in the loss. The reduction in value of debt due
to risky projects is called agency cost of debt financing.
Jensen and Meckling (1976) suggested that firms can find optimal
capital structure point where the total cost of agency is minimized.

Dynamic Trade-off theory
There are abundant studies supports this static trade-off theory such as
Myers (1993); Andrade and Kaplan (1998); Graham (2000); Hovakimian,
Kayhan, and Titman (2012). Graham and Harvey (2001) find that 81% of CFOs
5


claims that they apply target debt-equity or something like that on debt
decisions. However, the evidence of this theory is still widely debatable.
In respond to this debate, especially static trade-off theory, academics
have turned to dynamic versions of the trade-off theory.
Studies on dynamic adjustment share a similar view that the optimal and
real capital structure are not able to be the same all the time. The imperfection
of market such as transaction costs can prevent the instantaneous adjustment
of the real debts to the desired level. Heinkel and Zechner (1989) propose that
small recapitalization costs have the ability to give significant changes in debt
ratio of a firm while the role of agency costs of debt is emphasized by Leland
(1998) through the determination of the optimal debts.
From a dynamic model perspective, a suitable capital decision often
depends on the financing margin that the firm anticipates in the next period.
Some companies expect to pay our funds in next period, while others want to
raise funds in the form of debt or equity. Firms will consider all these factors
while making decisions.
Kane (1984), Brenman and Schwartz (1984) propose first dynamic
models which take the tax savings and bankruptcy cost into consideration.
Both examined continuous time models with uncertainty, taxes, and bankruptcy
costs without transaction costs. Firms react to adverse shocks immediately by
rebalancing; firms maintain high levels of debt to get benefits of tax.
Heider (2003), suggests that companies specify a

target capital

structure. By estimating a partial adjustment model, companies quickly adjust
to that target when they are spread further to changes in share prices or
changes in firm characteristics.
Many studies of dynamic trade-off models are done rather recently however
they bring us new look of profits, retained earnings, taxes, mean reversion.
Further studies in line with trade-off theories appear to be very promising in the
future.

6


2.3. Pecking order theory
This theory is often set up as a competitor theory to trade-off theory. The
first seed of pecking order theory was from Donaldson’s book in 1961 when he
analyzed a large sample of corporations. He found that: “Management strongly
favored internal generation as a source of new funds even to the exclusion of
external funds except for occasional unavoidable ‘bulges’ in the need for
funds”. When external finance is needed, managers are reluctant to issue
share. “Several financial officers showed that they were well aware that this
had been a good time to sell common, but the reluctance still persisted.” The
primary focus of this theory is not optimal capital structure but more about the
way managers make use of various sources of funds to finance their
operations. This is really meaningful to any firm.
The theory is then modified and popularized by Stewart C.Mayers and
Nicolas Majluf (1984) and Mayer (1984). In “Corporate financing and
investment decisions when firms have information that investors do not have”,
the authors assumed that the financial market is perfect except there are
information asymmetries. The investors do not have all information of operation
and prospect of the firms as well as the managers. When managers believe
that their shares are overvalued by the market, they have more incentive to
issue new common shares rather than debt. Investors are rational when
interpreting this signal; they compensate for their own informational
disadvantage by decreasing issuer’s share price, especially when they see no
clearly valuable opportunities. The equilibrium model of the issue-invest
decision shows that firms tend to refuse to issue stock in this circumstance.
The managers act in the interesting of existing shareholders refuse to issue
new share when their equity is on the verge of being undervalued even though
they may miss opportunities for projects with positive NPV.
Mayer (2001) summation of Pecking Order is that:
“1. Internal finance is preferable to external finance for which firms must
take information asymmetries into account while making decision) 2. Dividend
is “sticky,”. Dividend cuts, therefore, are not spent for financing capital
7


expenditure. In addition, changes in cash requirements are not attributable to
short-run dividend changes.

That being said, changes in net cash display

changes in external financing.
3. If firms require external funds for capital investment, they will first
issue debt, which is the safest security, followed by the equity. If cash flow that
is internally generated is greater than capital investment, the excess is used to
repay debt rather than repurchasing and retiring equity. As additional external
financing is necessary, the pecking order is utilized; the firm moves from safe
to riskier debt such as convertible securities or preferred stock, and finally to
equity as last alternative
4. Each firm’s debt ratio therefore reflects its cumulative requirement for
external financing.”
There have been numerous empirical studies are in line with Pecking
Order theory. Many studies show that price of stock goes down after the
announcement of new issues. According to Asquith and Mullins (1986), the
price of stocks falls averagely 3 percent after new issue. .
2.5. Determinants of capital structure
2.5.1. Size
The size of a firm is one of the most recognized determinants in
previous studies of capital structure. Small firms have to face many difficulties
to raise long-term debt. They have pay with a higher rate of interest and
inconvenient terms. Therefore, small firms tend to rely on owned capital and
retained earnings in designing its capital structure. Large companies with
advantages of economies of scale, bargaining power can approach loans much
easier. They have more flexibility to choose sources of financing. They can
issue shares, bonds, obtain loans, and retained earnings. Trade-off theory,
therefore suggest that a positive relationship between firms size and leverage
exists Marsh (1982) found that large firms often use long-term debt while small
firms tend to use short-term debt. By analyzing financing decisions of public
firms in the major industrial countries, Rajan and Zingales (1995) found that

8


firms size are positively correlated with leverage except for Germany. In
Germany, larger firms tend to borrow less.
Firm size is also function as a proxy of information that outside investors
can have.

Fama and Jensen (1983) said that large firms disclose more

information to our side investors than small firms. Large firms are more able to
issue equity than small firms. Small firms are managed by a small group of
managers. Those managers tend to avoid the intrusion from outside entities.
They often use retained earnings as the first priority of financing. When internal
funds are not enough for new projects, they prefer to use debt than issuing. As
a result, pecking order theory are more often used by small firms. Data from
Hussan and Matlay (2007) asserts that small firms only look for outside
resources when internal funds are exhausted. Daskalakis and Psillaki (2008)
shows that the motivation of small firms to buy is to meet financial needs rather
than achieve an optimal capital structure.
2.5.2. Tangible assets
Assets structure is an important factor of capital structure. A larger
portion of tangible assets increases the liquidation value of firms. While
intangible assets are difficult to identify, value, separate and utilize, tangible
assets are easy to identify and measure. Therefore, tangible assets can easily
be collateral to obtain debt. Companies are more able to approach creditors
when they have large tangible assets as a guarantee in case of bankruptcy.
Therefore, trade-off theory indicates that firms will be more leveraged when
they have more tangible assets.
Firms having more tangible assets have less impact of asymmetric
information and agency costs. Thus, pecking order theory and agency theory
predict that firms with more tangible assets have more capability to borrow
debt. Previous studies provided different results. Some studies prove that this
is true by revealing that the capital structure is positive with firms’ assets
structure such as Allen (1995), Michealas (1999), Amidu (2007). Daskalakis
and Thanou (2010) by studying Greek small and medium companies from
2003 to 2007 found an inverse impact of tangibility on debt ratio. They
9


explained that firms owning a larger portion of tangible assets already have a
stable source of returns, which provides internally stable funds and discourage
it to use external financing. Thus, pecking order theory also predicts a negative
relationship between leverage and tangibility.
Degryse et al. (2010) found a positive relationship between debt and
tangibility, however, positive effect on total debt came entirely from long-term
debt, as short-term debt is negatively affected by the collateral. It is possible
that banks charge relatively higher interest rates on short-term debt than longterm debt. This finding follows the maturity matching principle that long-term
assets are financed with long-term financing and short-term assets are
financed with short-term funds.
2.5.3. Profitability
Many researchers have conducted the investigation on the determinants
of capital structure, but the result of the relationship between profitability and
capital structure is still a conflict. According to Pecking Order theory, a firm just
uses external finance when internal finance exhausted. When firms make
profits, they tend to use retained earnings more than borrowings. There is
abundance evidence of this. Many studies such as Al-Sakran (2001), Fama
and French (2002), Frank and Goyal (2003), Daskalakis and Psillaki (2008),
Kayo and Kimura (2010), etc. are consistent with Pecking Order theory by
showing that there is a negative relationship between profitability and leverage.
Although firms cannot enjoy the tax shield benefits which could be availed
because of increased leverage, firms with high profitability incorporate less
debt while forming their capital structure decisions.
2.5.4. Growth
The relationship between growth and capital structure is still
ambiguous. Major previous studies generally favor a negative relationship
between the growth opportunities and leverage of firms. Rajian and Zingles
(1988) while using market-to-book as a proxy of growth found a negative
relationship between growth opportunities and leverage. One reason for the

10


market-to-book ratio to be negatively correlated with leverage is that firms often
issue stocks when their market value is high in comparison with book value.
In contrast, another school of thought supports a positive relationship
between growth and leverage. Modigliani and Miller (1958) explained that
managers have to pay costs for debt. In the begin, a firm may use debt as a
source of capital but when the firm realize that projects will become profitable
they reduce debt. Modigliani and Miller found a positive relationship between
leverage and growth. Myers (1977) said that companies with highly potential
growth opportunities do like risky debt and the firms using risky debt, in some
circumstances, will pass up valuable investment opportunities. Pandey (2001)
proved that firms whose sales grow rapidly are often in need of expanding their
fixed assets. Thus firms with high growth are in more demand of funds and
also retain more earnings. When examining the determinants of capital
structure of Malaysian during 1994 to 1999, Pandey found a significant relation
between growth and all types of book and market value debt ratios. Because
short-term debt ratios are higher than long-term debt ratios, Malaysian firms
tend to employ short-term debt to finance their growth.
Titman and Wessels (1988) using data of U.S firms from 1974 to 1982
reported no relationship between leverage and growth.
2.5.5. Non-debt taxed shield
DeAngelo and Massulis (1980) proposed a model of optimal capital
structure that combines the impact of corporate taxes, personal taxes, and
non-debt tax shields. They said that tax deduction for depreciation and
investment tax credits are tax benefits of debt financing. Firms with a large
amount of non-debt tax shield will enjoy more benefit, as a result, borrow more.
According to trade-off theory, when a firm is forced to pay higher taxes
on its earnings it has higher leverage. The more firms borrow, the more value
of debt tax shields they can gain, as suggested by Myers (1984). Besides, nondebt tax shields such as accounting depreciation, depletion allowances, and
investment tax credits have been found to have a negative influence on

11


leverage because they act as substitutes for the benefit of debt financing
coming from interest tax shields (DeAngelo & Masulis, 1980).
Empirical research has shown support for the above predictions
regarding the relationship between taxes and leverage, although it is not one of
the most popular factors. Bauer (2004) and Chi (2013) found that taxes have a
positive influence on the use of debt, yet, the impact of taxes does not always
hold.
Non-debt tax shields, often relating to depreciation and other operating
expenses, are observed in the empirical work of Bauer (2004), Huang and
Song (2002), and Cortez and Susanto (2012) to have a negative relationship
with leverage. This is consistent with what trade-off theory suggest.
2.5.6. Risk
Firms risk can be defined as financial risk and business risk. Here,
business risk is, as volatility in earnings, is tested to see if it determinate capital
structure. The more risk firms have, the higher the cost of the financial distress
firm face. Trade-off theory predicts that riskier firms reduce leverage to avoid
high bankruptcy cost. Titman and Wessels (1988) confirmed that optimal debt
level to be a decreasing function of earnings volatility. A similar argument is
introduced by Frank and Goyal (2009). More risky cash flows resulting from
cyclicality or seasonality of business lines will make firms not able to gain full
benefit of tax advantage of debt; thus, trade-off theory would support a
negative relation between volatility and leverage. Moreover, firms with volatile
cash flows will also want to avoid making large fixed commitments for debt
holders.
On the other hand, Schoubben and Hulle (2004) explained that when
risk increases, the cost of debt increase simultaneously. Creditors set higher
fee of debt to protect themselves from a bankruptcy of firms. Therefore,
according to pecking order theory, to deal with higher cost of debt, firms will
use internal funds as a priority. Frank and Goyal (2009) argued that firms with
high volatility in earnings can be regarded in the financial markets as having
poor management or problems in business lines, resulting in volatile stock
12


prices. Because the impact of information asymmetry is higher in riskier firms,
the pecking order theory would also suggest that these firms have higher
leverage.
Table 1 presented predicted effects on leverage based on capital
structure theories and table 2 presented the results of some previous
rehearses on capital structure.
Table 1: Predicted effects on leverage based on capital structure theories
Predicted effects on leverage
Trade-off
theory

Pecking
-order
theory

Business risk



+

Profitability

+



Firm size

+



Growth opportunities



Tangibility

+

Non-debt tax shields



Factors

Agency
Theory

Free cash
flow
theory

+

+

+





–/+

+

Market
timing
theory

13


Table 2: Determinants of capital structure in previous researches
Study
Titman and
Wessel (1988)
Pandey(2001)
Joy Pathak
(2010)
Vergas,
Cerqueira,
Brandão
(2015)
Frank and
Goyal (2009)
Rajan and
Zingales(1995)

Sample
period

Country

Relationship between leverage and firm’s
characteristics
Size Tang
Grow Profit
NDTS
Risk

19741982
19941999
19902009

U.S

-

-

+/-

-

-

-

Malaysia

+

-

+

-

x

+/-

India

+

+

+

-

x

-

20052012

Portugal

+

+

-

-

-

x

U.S

+

+

-

-

+/-

-

U.S

+

+

-

-

x

x

19502003
19871991

14


3. DATA AND METHOLOGY
3.1 Data Description
This study analyzes a sample of 1500 U.S companies of S&P
Composite 1500 index which cover 90% of the market capitalization of U.S
public stock. In order to ensure the generality, the sample includes three
groups of firms. The first group is S&P 500 with 500 large companies having
common stock listed on the NYSE or NASDAQ. The second group is S&P 400
(mid-cap firm) which includes investors with a benchmark for mid-sized
companies to reflect the distinctive risk and return characteristics of this market
segment. The third gruop is S&P 600 which includes small-cap segment of the
U.S equity market. However, among those 1.500 companies, only companies
meeting all requirements are included. For example, the model specifications
include lagged variables so companies do not have data of at least two
consecutive years will not be used.
The data sample includes data of those U.S companies in post-financial
crisis time, Q1/2010 to Q2/2016, to exclude the effect of financial crisis on
capital structure. The final sample is a panel data comprising 39.000
observations.
3.2 Panel data regression model
3.2.1 Panel data
Panel data is used in this study to analyze the factors affecting capital
structure for numerous reasons. According to Hsiao (2006):
First, that panel data combines cross-sectional data and time series
data can help to identify individual and time effects which cannot be identified
by pure cross-sectional or time series data. This is especially necessary as we
need to analyse effects of many factors on companies’ capital structure during
25 periods.
Second, panel data provides rich information about cross-sectional and
makes it possible to analyze dynamic properties of data as well as capture the
15


complexity of capital structure puzzle. Panel data provides information on both
inter-temporal dynamics and the individuality of the entities which will help
explain determinants of capital structure.
Last but not least, panel data can control unobservable individual
heterogeneity by allowing for entity-specific variables and avoid other problems
in time series data such as multicollinearity.
First, descriptive statistics were used to demonstrate characteristics of
U.S firms, as presented previous part. Following this, correlation analysis was
conducted to derive an overview of the relationship between each pair of
variables. Subsequently, linear regression was performed as the main analysis
to identify the best factors to explain the capital structure.
Since there are many companies lacking data in several years, the data
set is constructed as an unbalanced panel. There are several potential
estimating techniques to deal with panel data. In this paper, the most two
popular estimators, the random effects estimators, and fixed effects estimators
will be used.
The fixed-effects model explores the relationship between explanatory
variables and dependent variables within an entity (i.e. a company in this case)
and removes the effect of time-invariant characteristics pertaining to the entity
in order to assess the net effect of the explanatory variables on the dependent
variable. Unlike the fixed-effects model, the random-effects model assumes
that the variation across entities is random and not correlated to explanatory
variables in the model, allowing for time-invariant variables to have an effect on
the dependent variable.
Hence, the regressions were run with both the fixed-effects model
(FEM) and random-effects model (REM) on only independent variables without
dummies variables, the Hausman test will be used to examine whether the
estimated coefficients from the fixed effects estimation and the random effects
estimation is statistically significant, as suggested by both Wooldridge (2010)
and Baltagi (2005).
16


In this study, panel data regressions using both the fixed-effects model
and random-effects model were conducted using Stata 13 software.
3.2.2. Definition of variables
Measures of capital structure
There are many types of securities a firm can issue to raise funds such
as short-term debt, long-term debt, convertible debt, preferred stock, and
common stock. The term “leverage” considered in capital structure research
refers to financial leverage, not operating leverage or total leverage After
reviewing various leverage measures used in previous research, Rajan and
Zingales (1995, p. 1429) concluded that “the effects of past financing decisions
is probably best represented by the ratio of total debt to capital.” Hence, the
most relevant definitions of leverage to use in this study include the ratios of
total debt to capital – either in book value or market value, the ratios of longterm debt and short-term debt to book value of capital. These measures are
identified below.
Table 3: Measures of leverage
Den
otation
TD
M
TD
B
LD
B
SD
B

Dependent Variable

Calculation

Total Debt to Market Value

= Total debt/

of Capital Ratio

(Total Debt + MV of Equity)

Total Debt to Book Value

= Total debt /

of Capital Ratio

(Total Debt + BV of Equity)

Long-term Debt to Market Value = Long-term debt /
of Capital Ratio

(Total Debt + BV of Equity)

Long-term Debt to Book Value

= Long-term debt /

of Capital Ratio

(Total Debt + BV of Equity)

Measures of capital structure determinants
Profitability
There exist many different measures of profitability. Different studies use
different measurements. Titman and Wessels (1988) used the ratios of
17


operating income over sales and operating income over total assets. Frank and
Goyal (2009) used operating income before depreciation. Michaelas,
Chittenden and Poutziouris (1998), Han-Suck Song (2005) and Sogorb-Mira
(2005) used the ratio of earnings before interest and taxes (EBIT) to total
assets as indicators of profitability. Hall, Hutchinson and Michaelas (2000)
used ratio of EBT over sales. Klapper, Sarria-Allende and Zaidi (2006) and
Degryse, Goeji and Kappert (2009) used returns on assets as proxy of
profitability. In this study, returns on assets is used as indicator of profitability.
Size
There have many different measurement of size, for example sales,
number of employee, total assets. This study follows Frank and Goyal (2009),
Flannery and Rangan (2006) to use the log of total assets is used as proxy for
firm size.
Growth
Growth can be seen as change business operation. Business growth
has been considered an influential determinants to capital structure in many
empirical studies. The most commonly used proxy for growth is market-to-book
asset ratio. Capital expenditure and a change in log assets are also popular
proxies for growth opportunities.

In a study by Han-Suck Song (2005),

percentage change in total assets is the measurement of growth. Michaelas,
Chittenden and Poutziouris (1998), Hall, Hutchinson and Michaelas (2000)
used percentage change of sales over previous 3 years. Degryse, Goeij and
Kappert (2009) used change of total assets from previous year.
This study aims to investigate the realized values as proxy for growth
when the capital structure decision was made not the potential opportunities of
growth. Therefore, growth is not measured by market-to-book asset ratio but
the percentage change of sales from previous year.
Tangibility
Tangibility assets affect firm’s ability to borrow. Therefore, it influences
policy debt leverage of firms. In previous studies, tangibility is often calculated
18


as percentage of fixed assets over total assets. Here, tangibility is calculated
as gross property, plant and equipment taken from Compustat over Total
Asset.
Non-debt tax shield
Non-debt tax shield which is tax deduction for depreciation and
investment tax credits, is considered to be tax benefit of debt financing.
Indicators for non-debt tax shield often includes the ratios of depreciation over
total assets, investment tax credits over total assets and a direct estimate of
non-debt tax shields over total assets. In this study, the ratios of depreciation
over total assets will be used.
Risk
To measure influence of risk on capital structure, volatility of income is often
used. In this study, Three-year rolling standard deviation of returns on assets.
Business risk of each period equal standard deviation of ROA of the last
period, this period and the next period.
A complete set of the potential determinants of capital structure, the
proxies to measure them is presented in table 4
Table 4: Measures of capital structures determinants
Proxy

Factors

Calculation

ROA

Profitability

Returns on Assets

SIZE

Size

Ln (Total Assets)

GROWTH

Growth

Growth n= (Growth n– Growthn-1)/ Growthn-1

TANG

Tangibility

Net Property, Plant, and Equipment / Total Assets

NDTS

Non-debt tax shields

Depreciation Expense / Total Assets
Three-year rolling Standard Deviation of ROA:

RISK

Business risk

RISKn = Standard Deviation of ROA in quarter n-1,
n, n+1

19


3.2.3. Model
Determinants of capital structure
The regression model showing how the determinants capital structure
decisions can be presented as follows:
Li,t =
+

+

ROAit +

SIZEit +

GROWTHit +

TANGit +

NDTSit +

RISKit

+ eit
: the firm effect assumed constant for firm i over time t
: the time effect assumed constant for given t over i
eit: error term

20


Tài liệu bạn tìm kiếm đã sẵn sàng tải về

Tải bản đầy đủ ngay

×