the dominance of the agency model on financing decisions

157

Gadjah Mada International Journal of Business – May-August, Vol. 17, No. 2, 2015

Gadjah Mada International Journal of BusinessVol. 17, No. 2 (May-August 2015): 157-178

* Corresponding author’s e-mail: [email protected]

ISSN: 1141-1128http://journal.ugm.ac.id/gamaijb

The Dominance of the Agency Modelon Financing Decisions

Bramantyo Djohanputro

PPM School of Management, Indonesia

Abstract: There are some issues about how companies consider their financing. These issues are relatedto the amount, source, type, and the structure of such financing. So far, there is no uniform model that isable to explain how companies deal with these issues. There are three competing, dominant theories offinancing decision making, i.e. the Pecking Order Theory, the Static Trade-off Theory, and the AgencyModel Theory. This study attempts to explore which theory explains the best way for companies in theconsumer industry to decide their financing method. There are five hypotheses to be tested in this study.Using data from public listed companies on the Indonesian Stock Exchange from 2008 to 2011, it seemsthat the Agency Model Theory is more dominant than the other two theories in explaining the waycompanies fulfill their financing needs.

Abstrak: Terdapat beberapa isu terkait bagaimana perusahaan mempertimbangkan kebutuhan pendanaanmereka. Isu-isu yang terkait dengan pertimbangan tersebut adalah mengenai jumlah, sumber, jenis, danstruktur pendanaan. Sejauh ini, tidak ada model yang seragam yang mampu menjelaskan bagaimanaperusahaan menangani isu tersebut. Terdapat tiga teori yang saling bersaing dan dominan terkait keputusanpendanaan, yaitu teori pecking order, static trade-off, dan agency model. Studi ini berusaha menggali teori manayang terbaik menerangkan cara perusahaan di dalam industry konsumsi memutuskan pendanaan mereka.Terdapat lima hipotesis yang diuji di dalam studi ini. Dengan menggunakan data perusahaan terbuka diBursa Efek Indonesia dari tahun 2008 sampai 2011, tampak bahwa agency model theory lebih dominandalam menjelaskan cara perusahaan memenuhi kebutuhan pendanaan mereka dibandingkan dengan keduateori lainnya.

Keywords: agency model theory; capital structure; corporate finance; pecking order theory(POT); trade-off theory

JEL classification: G00

Djohanputro

158

Introduction

There are some issues about how com-panies decide their financing methods. Sucha decision is not only concerned with theamount that can be obtained but also withthe source, type, and the structure of the fi-nancing. In dealing with these issues, compa-nies may consider some of the underlyingfactors in financing, such as the availabilityof funds, the cost of capital, and the abilityto control the company.

A theoretical point of view is neededto explain how and why practitioners behavein a certain way in fulfilling their financingneeds. For example, some Chief FinancialOfficers (CFOs) and boards of directors maydecide their financing needs on the basis of abenchmarking approach, i.e. by followinghow their peers determine their financingneeds, especially for the capital structure.This may be one reason why companies in anindustrial group tend to have similar capitalstructures because they adjust their capitalstructures to reflect those of their peers.Other CFOs determine the capital structureon the basis on their capacity to generate in-come or cash flows to fulfill their obligations.The more the companies have the capacityto generate an operating cash flow, the greateris their capacity to service their borrowing.

Academicians provide several alterna-tives to explain practitioners’ behavior in suchcases. There are three competing, dominanttheories that explain how financing needs andcapital structures should be determined. Thetheories are the Pecking Order Theory(POT), the Agency Model Theory (AMT) ofcapital structure, and the Static Trade-offTheory (STT). According to the (POT), acompany attempts to obtain capital on thebasis of the ease with which it can be ac-cessed, starting with the easiest source. The

CEO and executive directors put some ef-fort into negotiating with the non-executivedirectors or with the board of commission-ers regarding how much net profit should beretained, as retained earnings, and how muchis to be distributed as dividends. If retainedearnings, which are the easiest source of fi-nancing that can be obtained, are exhausted,then they must start searching for othersources of financing.

The AMT on Capital Structure empha-sizes the idea of improving the discipline ofthe board of directors in managing their com-pany. A company with a higher profit needsto be closely controlled to assure that the di-rectors put the maximum effort into protect-ing the interests of the shareholders by maxi-mizing the value of the company and by us-ing the company’s resources wisely. One wayof improving their discipline is by askingcreditors to provide loans to the companyand, as a consequence, closely control it. Thecreditors need to assure that the company iswell managed and, as a result, their loans aresafe.

The STT, on the other hand, assumesthat a company attempts to maintain a cer-tain Debt to Equity Ratio (DER). Compa-nies who apply this theory believe that a cer-tain ratio contributes to the efforts to maxi-mizing the value of the company, as long assuch a ratio results in the minimum level of

the Weighted Average Cost of Capital (WACCor simply COC) or the maximum value ofthe firm.

The use of these models by practitio-ners may vary from one region to another,and across time (Rafiq et al. 2008). As an im-plication, some researchers have focusedtheir studies on specific industries(Shanmugasundaram 2008; Jahan 2014;Kühnhausen and Stieber 2014; Mwangi et al.

159


2014). In the case of Indonesia, choosing abasic industry sector for this study was basedon the fact that the chosen industry was notheavily regulated and played in a free, com-petitive, market. No single company used inthis study belongs to the government. This isone indication that the industry works undermarket-mechanisms. Therefore, the CFOs orBoard of Directors have the freedom to de-cide the sources of financing and their capi-tal structure based on the companies inter-ests.

Based on the aforementioned argument,this study attempts to answer the followingquestion: is there any tendency for compa-nies that operate in the basic industry sectorto decide their sources of financing and theircapital structure? This study indicates thatthe AMT seems to be more dominant thanthe other two theories.

This paper is organized as follows. Thefirst section is the introduction. It is followedby a brief description on the previous stud-ies on the Pecking Order, Agency Model, andSTT of Capital Structure. This is then fol-lowed by the proposed models and hypoth-eses. The next section elaborates the data em-ployed in this study and their analysis. Thispaper closes with the conclusion.

Previous Studies

Papers on testing the existence of onlytwo theories, i.e. the POT and STT, are quitecommon. Myer (1984) tested whether com-panies tended to use the Pecking Order orStatic Trade-off Theories. Myer argues thatthe STT emphasizes the existence of an op-timal capital structure between debt and eq-uity by emphasizing the optimum benefit andcosts of various sources of financing. Theabsence of an optimal capital structure isassumed to follow the POT.

Graham and Harvey (2001), also testedthe existence of those theories through a sur-vey sent to CFOs. They emphasize the waythat CFOs consider certain factors when theylook for capital, whether they strictly set atarget ratio, moderately set a target ratio, ordo not have a target ratio. They reveal thatnot all CFOs have a target capital structure.While only ten percent of CFOs have a verystrict capital structure, at least nineteen per-cent of the CFOs surveyed did not have atarget ratio at all. Their research also indi-cates the existence of both the Pecking Or-der as well as the STT in the actual decisionmaking process for financing.

Graham and Harvey (2002) andAcaravci (2015) also conducted research toprove the Pecking Order and Static Trade-off Theories by designing one model that wasable to explain the existence of both theoriesby employing several explanatory factors.Some of those factors explain the existenceof the POT, while others explain the exist-ence of the STT. Some other studies, such asthe one conducted by Rafiq et al. (2008),employed the same factors to indicate theexistence of either the Pecking Order orStatic Trade-off Theories, depending on thesigns of those factors, i.e. positive or nega-tive signs. The models may have employedparametric as well as nonparametric models.

Some studies indicate the tendency to-ward a certain DER. Frank and Goyal (2004)found that a company tends to have similari-ties in the growth of debt and equity at thesame pace. In other words, CFOs of thosecompanies behave as if they use the STT asthe basis of their financing decisions by main-taining the DER at a constant level.

However, following Wurgler (2002) andWelch (2003), the DER may deviate from itslong target ratio, because of stock price move-ment, the timing of equity issuance, and the

Djohanputro

160

timing of debt issuance. Therefore, the DERbased on market value as the dependent vari-able needs to consider the market-to-bookratio and the interest rate as the explainingfactors of DER changes.

Myer and Majluf (1984), were followedby some fellow researchers, such as Frankand Goyal (2008); Neus and Walter (2008);and Getzmann et al. (2010) who placed em-phasis on the Asymmetric Information Con-dition. This condition leads to the possibilityof moral hazards and adverse selections. Thismeans that firms generally give priority tointernal rather than external sources of fi-nancing because internally generated capitalis easier to negotiate and acquired than capi-tal from outside, regardless of the quality ofthe companies. Based on this argument, com-panies tend to follow the POT in decidingtheir capital needs.

The studies aforementioned have simi-larities in terms of the variables and modelsapplied. Most of them employed the changein debt as the dependent variable. Cash flow,profit, capital expenditure, and changes inworking capital were the explanatory vari-ables. A few of those studies added other in-dependent variables in order to improve therobustness of the studies.

AMT, as an alternative to the CapitalStructure Theory, stands by the principle thata company needs to incur costs to reduce it’sagency problem (Sen and Oruc 2009). Ac-cording to Vo and Nguyen (2014), this agencyproblem can be dealt with by two mecha-nisms. The first mechanism is by allowing topmanagement to hold a portion of shares inthe company to make them take proper careof the company (see also Morck et al. 1988).The second mechanism is by increasing thelevel of debt in order to improve the marketdiscipline (see also Jensen, 1986). Sharehold-

ers expect the senior management team tomaximize the firm’s value and shareholders’wealth. However, the senior managementmay behave in such a way as to exploit theopportunity to maximize their own interests.Therefore, shareholders need a model to con-trol them in such a way that they are forcedto behave on behalf of the shareholders’ in-terests. This is the role of debt in a company.

The AMT puts the role of debt differ-ently from that in the Pecking Order andStatic Trade-off Theories. Under the AMT,debt becomes an instrument to control man-agement. Under the POT, debt becomes asources of capital after internal sources ofcapital are exhausted. Under the STT, debtbecomes the balancing source of capital tokeep the DER constant at a certain targetratio.

The AMT argues that the larger, moreprofitable companies tend to increase theirdebt. By doing so, creditors, as external par-ties, put a lot of effort into controlling theway senior management manage the compa-nies. Even though the interests of the credi-tors are mainly for the safe and timely repay-ment of their loan, the way management con-trol and manage certain areas of the companyare forced to be in accordance with the inter-ests of shareholders such as the companies’financial performance, the role of fiduciary,and the sustainability of the companies. Theinvolvement of creditors may be costly butthe costs may not exceed the benefits ex-pected by the shareholders.

Hypotheses and Models

Pecking Order Theory (POT)

To prove the existence of the POT, thisstudy employed three models. The first model,

161


i.e. the simple one, used the financing defi-ciency on the change in debt. The financingdeficiency is defined as the condition in whichthe investment required by the company ex-ceeds the cash flow generated internally bythe company. The source of the internal cashflow is profit. As the profit is not sufficientto fulfill the capital need, it implies that thecompany needs to receive a financial injec-tion from outside. The source of such financ-ing could be in terms of either debt or eq-uity.

The way a company fills the deficiency,or the gap between profit and investment re-quirement, depends on several factors. In thefirst case, the deficiency of a company ex-ists, i.e. earnings after tax or the net profit ofcompany i at year t (EAT

i,t) is positive, but

less that the investment required at time t.Under this circumstance, the company needsto acquire debt to fulfill the gap. At a certainlevel of investment need, the lower theEAT

i,t, i.e. the wider the gap, the larger the

amount of debt is needed, Di,t, to fill the

gap. In other words, the lower the EATi,t, the

higher the Di,t.

In the second case, the EATi,t exceeds

its investment requirement at year t. Accord-ing to POT, the company puts EAT

i,t as the

first priority to fulfill the investment required.The excess of EAT

i,t would be idle if it is

kept in the company. Therefore, idle EATi,t

will be used to repay a part of the debt in thefirst place. Therefore, the larger the EAT

i,t

the lower is the debt. In other words, EATi,t

has the opposite movement in the change indebt at time t (D

i,t). The higher the EAT

i,t,

the lower the Di,t.

In the third case, EATi,t is negative, i.e.

the company suffers losses and does not gen-erate an internal cash flow to fulfill the in-vestment required. POT argues that the com-

pany will attempt to obtain new debt to fillthe investment required. Therefore, the EAT

t,i

is negatively related to Di,t.

The relationship of EATt,i and D

i,t, can

be expressed mathematically as follows:

Di,t= a + bEAT

i,t+

t...................... (1)

(-)

where

Di,t= the change in debt of company i at

period t;

EATt,I

= the profit after tax of company i atperiod t;

t= the error term at time t.

Hypothesis1 (H1): The earnings after tax of a com-

pany has a negative relationshipwith the change in debt.

Equation (1) can be slightly extendedby putting the dividend under the followingargument. Suppose a company has a positiveEAT

i,t, i.e. the company generates a profit at

time t. Under most terms of credit, the com-pany has to fulfill its repayment obligationof the existing debt as its first priority. In thiscase, the total debt decreases or D

i,t is nega-

tive. The remained profit will be reinvestedin the company, and the residual of the profitwill be distributed as dividend. Under thesecircumstances, the larger EAT

i,t, may result

in a larger Divi,t and at the same time lower

the debt.

The relationship of EATi,t

andDividend

i,t toward D

i,t can be expressed as

follows:

Di,t= a + bEAT

i,t + cDiv

i,t +

t...........(2)

(-) (-)

The aforementioned leads to the hy-pothesis formulation to prove the existenceof POT as follows.

Djohanputro

162

Hypothesis2 (H2): The earnings after tax and pay-

ment of dividends by a com-pany have a negative relation-ship with the change in debt.

The above model can be extended byputting in three other independent variables.Rafiq et al. (2008) applied the portion of tan-gible assets as an explanatory variable. Theyargued that more tangible assets providedbetter opportunities for companies to borrowbecause they can use those tangible assets ascollateral. This study does not employ the tan-gibility of assets as an independent variable.Instead, the change of fixed and current as-sets are employed to explain the behavior ofthe change in debt. The increase in either fixedor current assets can be fulfilled, partially orwholly, by debt because those assets can beused as collateral.

Rafiq et al (2008) also applied the loga-rithm of asset as an explanatory variable. Thepurpose of using this variable was mainly toaccommodate the size effect. Large compa-nies tend to have more opportunities to bor-row. They also proposed the impact of asym-metric information on the amount of debt.In a large company, insiders have more infor-mation than outsiders. Moreover, the priceof equity in large firms tends not to be un-dervalued. Under this condition, managementprefer to finance companies by using equity.As a consequence, the increase in debt is notas much as the increase in firm size.

Di,t= a + b

1EAT

i,t + b

2Div

i,t + b

3FA

i,t +

b4WC

i,t + b

5LA

i,t +

t ..................(3)

where,

Di,t= the change in debt,

FAt,i= the change in fixed asset,

WCt,i= the change in working capital,

Divt,i= the dividend, and

LAt,i= the logarithm of the total book value

of assets of company i at period t.

The last variable is to control the inducementof the size of the company.

POT expects the following results.FA

t,i and WC

t,i are expected to have a posi-

tive relationship with Dt,i, in the sense that

the demands for large investment tend topush the company to increase its borrowing.The increase in fixed and current assets pro-vides a larger amount of assets that are avail-able as collateral. Therefore, the increase inboth types of assets may induce the increaseof debt.

Dividend, Dt,i is expected to have a

strong negative correlation with Divt,i be-

cause a large dividend indicates a largeamount of excess funds. Therefore, the com-pany does not need to borrow money. Asmaller dividend, or the absence of such, in-dicates a shortage of funds and therefore thecompany needs to borrow money and hencethe D

t,i increases.

The coefficient of LAt,i only indicates

the accessibility of the creditor based on thesize of the company. Larger companies tendto exploit equity financing, instead of debt.They use debt to fill any lack of equity fi-nancing. Therefore, LA

i,t is expected to have

a weak, positive relationship with the Dt,i.

In summary, based on Equation (3),POT expects the following results:

Di,t= a + b

1EAT

i,t + b

2FA

i,t + b

3WC

i,t

(-) (+) (+)

b4Div

i,t + b

5LA

i,t +

t ..................(4)

(-) (+)

Based on the Equation (4), the hypoth-esis to explore the evidence of POT is as fol-lows:

163


Hypothesis3 (H3): The change in debt has a strong

negative relationship with thedividend but a moderate or lowpositive relationship with thechange in fixed assets and thechange in working capital.

Agency Model Theory (AMT)

As aforementioned, the AMT arguesthat a company with a higher profit encour-ages the management to borrow funds fromcreditors. In this case, shareholders feel moresecured in terms of the safety of their wealthheld by the management team because theyare under the scrutiny of the creditors. Thehigher portion of debt in their capital, thetighter the creditors supervise and monitorthe management.

The above argument can be expressedstatistically as follows:

Di,t= a + b

1EAT

i,t +

t.........................(5)

(+)

whereD

i,t= is the change in debt; and

EATi,t= the earnings after tax or net profit of

company i at time t.

As mentioned earlier, under AMT, sharehold-ers attempt to control the company throughcreditors. The larger the profit gained by thecompany, the larger the debt of the company.Creditors always put efforts into monitoringthe company closely to assure that the com-pany maintains its creditworthiness throughit’s financial, operational, strategic, and mana-gerial healthiness.

The hypothesis related to Equation (5)is as follows:


positive relationship with theearnings after tax.

Following Dudley (2007), Leland(1994), and Breman and Schwartz (1978), amore complete form of AMT is modeled asfollows:

Di,t= a + b

1EAT

i,t + b

2FAR

i,t +

(+) (+)

b3LA

i,t +

i,t .................................(6)

(+)

FARi,t

is the fixed asset ratio of companyi at year t, i.e. the ratio of fixed assets to totalassets. This variable has two meanings.Firstly, a higher ratio indicates the higher capi-tal requested. Under AMT, a company tendsto finance the investment in fixed assets fromdebt in order to enhance the control. Dudley(2007), Leland (1994), and Brennan andSchwartz (1978) put another meaning to thisratio. According to them, FAR

i,t represented

bankruptcy costs. The higher the ratio, thelower the bankruptcy costs. It is becausecreditors have collateral in case of bank-ruptcy. Therefore, a higher FAR

i,t makes

creditors feel comfortable to provide loansand, as a result, increase debt.

The variable LAi,t in AMT has a similar

meaning to that in POT. Large companiestend to have better access to creditors. There-fore, the amount of debt is expected to havea positive relationship with the size of thecompany.


positive relationship with theearnings after tax, and a moder-ate, positive relationship with thefixed asset ratio.

Static Trade-off Theory (STT)

The hypotheses for the STT are devel-oped as follows. A company with a hugeEAT

i,t may use this net profit in different

Djohanputro

164

ways, depending on the opportunities avail-able to it. Under the condition of the absenceof investment opportunities, the companytends to distribute the money as dividends orkeep it in the company as a reserve. How-ever, if the company has an opportunity toinvest, it will deploy the money to be investedin fixed as well as current assets. The largerthe investment opportunity to be captured,the larger the EAT

i,t, portion will be to be

allocated for investment. According to theSTT, every investment must be financed byequity and debt to keep the capital ratiostable. In other words, the larger the EAT

i,t,

the larger the debt to fulfill the demand forinvestment.

A company that suffers losses can donothing to grasp an investment opportunity.The only possibility it has is to divest or sellsome assets to fulfill its need for cash. Undera loss condition, the total book value of eq-uity decreases. Under STT, the company hasto reduce its debt to keep the DER constant.Therefore, the sale of assets will provide thecash needed for operations and, at the sametime, to repay a part of the debt. This leadsto the following conditions. The first condi-tion, negative EAT

i,t, is accompanied by nega-

tive Di,t when the company suffers losses

and, at the same time, reduces its debt. Inother words, the EAT

i,t has a positive rela-

tionship with Di,t. The second condition, the

EATi,t is negative but the company does not

want to reduce its size, i.e. the company doesnot want to sell a part of its assets and doesnot reduce its debt. This is unlikely to hap-pen under STT because this decision leadsto the deviation of the DER from its target.

The above arguments can be expressedmathematically as follows:

Di,t= a +bEAT

i,t +

t .......................... (7)

(+)

Equation (7) proposes that a higherEAT

i,t provides more equity for investment.

Therefore, the debt must be increased to ful-fill the investment in order to keep the DERat a certain or targeted level. The lower EAT

i,t

also leads to a lower need for additional debt.Therefore, the positive relationship is ex-pected to be considerably strong.

The above explanation leads to the fol-lowing hypothesis under STT

Hypothesis 6 (H6): The change in debt has a moder-

ate, positive relationship withearnings after tax.

A company may suffer losses but findsan investment opportunity and needs to in-vest in fixed or current assets. Under STT,the company may seize the investment op-portunity by issuing shares to obtain new eq-uity and, at the same time, obtain a new loanto accompany the new shares to keep theDER constant. This decision leads to a posi-tive relationship between the change in debtand the change in equity. This relationshipcan be modeled in a statistical Equation 8 asfollows:

Di,t= a +bE

i,t +

t ............................. (8)

(+)

With Ei,t being the change in book value of

equity of company i at time

t. The hypoth-

esis, then, is as follows:

Hypothesis 7 (H7): The change in debt has a strong,

positive relationship with thechange in equity.

Ei,t depends on internal equity financ-

ing and external equity financing. Internalequity financing is reflected by the EAT

i,t and

DIVi,t. External equity financing is reflected

by NEFi,t, the amount of new equity financ-

ing from external sources by the issue of newshares. Hence, E

i,t in model (8) can be re-

165


placed by EATi,t, DIV

i,t,and NEF

i,t. Further-

more, other variables employed in model (4)are also applicable to test STT. Therefore,Equations (7) and (8) can be combined andextended as Equation 9.

Di,t= a + b

1EAT + b

2Div

i,t + b

3NEF

i,t +

(+) (-) (+)

b4FA

t,i + b

3WC

t,i +

(+) (-)

b5LA

t,i +

t .................................(9)

(+)

Under STT, the dependent variable Di,t

is expected to have a strong relationship withthe first three independent variables. The in-crease in debt is needed if the company hasretained earnings, i .e. earnings after tax(EAT

i,t) minus dividends (Div

i,t). It is because

if the company does not increase its debt,the capital structure will change and this situ-ation violates (STT). At the same time, theincrease in new equity financing (NEF

i,t),

must be responded to by an increase in debt,to keep the DER stable.

The change in debt is also expected tohave a positive relationship with the changein fixed assets as well as the change in work-ing capital. Every penny invested in fixed andcurrent assets must be financed by both eq-uity and debt. Therefore, the increase in as-sets must be responded to by the increase indebt. This argument also applies the otherway around. The decrease in assets needs tobe responded to by reducing both equity anddebt to keep their ratios as constant as pos-sible.

The relationship between the change indebt and the size of the company under STTis similar to that under POT. The hypothesisis, then, as follows.

Hypothesis8 (H8): The change in debt has a strong,

positive relationship with earn-ings after tax, new equity financ-ing, changes in fixed assets, andchanges in working capital; astrong , negative relationshipwith dividends, and a moder-ate, positive relationship withthe size of company.

Data and Analysis

Following Rafiq et al. (2008),Shanmugasundaram (2008), Jahan (2014),Kühnhausen and Stieber (2014), and Mwangiet al (2014) aforementioned, this study fo-cused on testing the existence of POT, AMT,and STT in an industry or strategic group in acertain time period. A strategic group con-sists of companies that have similarity interms of their business aspects, mainly in theirtypes of products, markets, and businessmodels. Some studies indicate that compa-nies within a strategic group tend to have asimilarity in their capital structures due toseveral factors, such as the behavior of credi-tors (Sen and Oruc 2009). According to them,the deviation from the industry average ofcapital structure is overcome by adding orreducing the debt. As a result, this action hasan effect on the market’s reaction (see, forexample, Siegfried 1984; Harris and Raviv1991; and Hatfield et al. 1994). However,some studies also indicate that certain stra-tegic groups tend to have diverse financingresources (Almazan and Molina 2005; Kim2008; Shanmugasundaram 2008). Variousfactors influence the divergence of capitalstructure (see, for example, Lev 1974;Mandelker and Rhee 1984; and Bradley et al.1984).

Djohanputro

166

This study focused on the consumergoods industry in Indonesia due to two mainreasons. Firstly, all the companies are inde-pendent companies, especially for decidingthe amount, sources, and structure of theircapital. This means that each of the boardsof directors of these companies has the free-dom to decide the capital required becauseof very limited regulations stipulated by thegovernment. There are an appropriate num-ber of listed companies in this industry andnone of them are state-owned companies.Therefore, their funding decisions are ex-pected to follow market mechanisms and, asa result, the behavior of the board of direc-tors may reflect the theories to be tested.Moreover, the industry is in a competitiveenvironment, meaning that their top manage-ment tend to make the best decisions for thecompanies. This study attempted to use forits data source as many public companies pro-ducing consumer goods which are listed onthe Indonesian Stock Exchange as possible.After being screened, based on the availabil-ity of the data, there were 24 companies suit-able for inclusion in this study.1

This study also specified the data onlyfrom a four years period, i.e. from 2008 to2011. The period chosen was based on thefact that Indonesia faced an economic crisisin 2008, caused by the impact from thesubprime mortgage crisis that started in theUnited States. Therefore, companies wereexpected to start preparing new financingstrategies between 2008 and 2009, as part oftheir corporate strategies, to return the com-panies to a stable condition, i.e. years 2010and 2011. Since some variables were in theprocess of change, such as the change in eq-uity, in the end there was only a period of

three years for the data employed in this study,i.e. the years 2009, 2010, and 2011.

Based on the industry and period cho-sen in this study, there are 72 items of dataavailable from 24 companies. However, theabsence of a time lag in the model leads tothe treatment of the data as being similar tocross section data. Research with these 72items of data and five independent variablesfulfills the statistical requirement. As thepanel data have the characteristics of timeseries data, the robustness of the interferenceis considered after regressions, i.e. whetherthe error fulfills the requirement of the ro-bustness.

Description of the Data

The description of the data is shown inAppendix 1. The table shows the huge fluc-tuation of the changes in debt, D

i,t, from

negative to positive. This means that the com-panies may have reduced their debt at cer-tain times and increased their debt at othertimes. EAT

i,t, one of the most important in-

dependent variables in this study, is alwayspositive. However, the companies sometimesdo not distribute dividends. In other words,they tend to accumulatec their profits as re-tained earnings.

The change of fixed and current assetsalso f luctuates considerably. Negativechanges in fixed assets indicate that at leastone company reduced its size of operations.This reduction leads to the reduction of theneed for current assets too. It needs to beproven whether the reduction in its opera-tions is balanced by the reduction of its debtor equity.

1 The companies finally selected as the samples in this study are ADES, AISA, CEKA, DLTA, DVLA, GGRM,HMSP, INAF, INDF, KAEF, KDSI, KLBF, LMPI, MERK, MLBI, MRAT, MYOR, PYFA, RMBA, SKLT, STTP,TCID, ULTJ, UNVR.

167


Appendix 1 also shows that Ei,t and

NEFi,t fluctuate strongly, from negative to

positive. While negative Ei,t shows the re-

duction in the total book value of equity,negative NEF

i,t shows the reduction in the

external equity. The external equity refers topaid-in capital and capital in excess. A nega-tive NEF

i,t may indicate that a company takes

a corporate action, such as stock buyback.This study did not intend to conduct researchinto the issue of corporate action. Instead,this study focused on the relationships be-tween the change in debt, on one side, andthe change in equity or net equity financing,on the other. Suffice to say here, that thosefluctuations are interesting to explore in rela-tion to the hypotheses under study.

This study also ran the Hausman testsfor all the hypotheses. As the first step, boththe fixed and random effects models of theHausman tests were conducted. For compari-son, this study reported the results for boththe fixed and random effects models for ev-ery model tested, as can be seen in Appendix1, Appendix 2, and Appendix 3. Appendix 1shows the regression results to test the hy-pothesis of POT, Appendix 2 to test the hy-pothesis of AMT, and Appendix 3 to test thehypothesis of STT.

The majority of the regressions fit withthe fixed effect model. Only two regressionsfit with the random effect model. The Equa-tion 1, i.e. the regression with EAT

i,t as the

only independent variable in the model, indi-cated the fitness of the random effect modelas well as the fixed effect model, as shown inAppendix 1. The regression result showed thatEAT

i,t had a positive relationship with D

i,t

at a 5 percent significance level. Therefore,both the random and fixed effect models maybe used to test the hypothesis.

The random effect model also appliesto the last model for testing STT. The lasttwo columns of Appendix 3 show the regres-sion results of Equation (9) using the ran-dom and fixed effect models. The regressionresults clearly indicate that the model underthe random effect model has the significantlevel of less than 1 percent while the modelunder fixed effect model is not significant toexplain the relationships between the inde-pendent variables and the dependent vari-able. Therefore, the random effect model isused to prove the existence of the last hy-pothesis.

Hypothesis Testing

As formulated in the hypotheses, the re-gression results were used to test each hy-pothesis by looking at the signs and signifi-cant levels of the coefficients of the inde-pendent variables. We started with testing H

1

to H3, to test for the existence of POT. The

regression results are shown in Appendix 2.The test for H

1 is shown in column (2) and

(3). H1 says that earnings after tax have a

negative relationship with the change indebt. However, the coefficient of EAT

i,t was

positive and significant at a 5 percent signifi-cance level. Since the sign and the signifi-cance level of the coefficient were not asexpected by POT, H

1 is not proven. Hence,

this Equation does not support the existenceof POT.

Columns (4) and (5) of Appendix 2show the results of regression to test the ex-istence of H

2. Adding DDiv

i,t as another ex-

planatory variable, together with EATi,t, was

consistent with the above result, i.e. the non-existence of POT. Both coefficients of theindependent variables were expected to benegative, while the results were both posi-

Djohanputro

168

tive. The coefficient of EATi,t was positive

and significant at the 5 percent significancelevel while the coefficient of Div

i,t was posi-

tive but not significant.

Columns (6) and (7) of Appendix 2 pro-vide the regression results to test H

3. The

coefficient of WCi,t was the only one that

was strongly significant, i.e. at the 1 percentsignificance level. However, the sign was theopposite to the expected sign for POT. Thismeans that the test to H

3 was also consistent

with the above two results, i.e. the nonexist-ence of POT.

The above tests clearly indicate thatPOT was not evident. In other words, it seemsthat POT does not fit in with the way themanagement of the companies within theconsumer goods industry decided their capi-tal needs. Even though a company has a highprofit, the companies do not merely rely onthe profit to fill the gap between the capitalrequired and the capital available.

The tests for the existence of AMT wereconducted by running the regressions thatwere in relation to hypotheses (4) and (5), asshown in Appendix 3. Columns (2) and (3)of Appendix 3 show the results of the re-gression of Equation (5) to prove H

4. Hy-

pothesis4 (H4)says that the change of debt

has a strong positive relationship with theearnings after tax. The coefficient of EAT

i,t

was positive and significant at the 5 percentsignificance level. Based on this regression,it seems that AMT exists.

Columns (4) and (5) show the regres-sion results to prove H

5, saying that the

change in debt has a strong positive relation-ship with the earnings after tax, and a posi-tive relationship with the fixed asset ratio.Column (4) shows that the coefficients ofEAT

i,t and FAR

i,t were both positive but none

of them were significant. Column (5) shows

slightly different results from column (4). Thecoefficient of EAT

i,t was positive and had a

1 percent significance level, i.e. very strong,while the coefficient of FAR

i,t was negative

but not significant. In this case, it seems thatH

5 is moderately accepted. Hence, H

5 sup-

ports the existence of AMT but not very con-clusively. Based on the results of testing bothH

4 and H

5, it can be concluded that AMT

seems to exist.

Appendix 4 shows the regression resultsto test for the existence of STT. Columns (2)and (3) show the regression results to test H

6,

saying that the change in debt has a moder-ate, positive relationship with earnings aftertax. The results of the regression support thishypothesis, i.e. the coefficient of EAT

i,t was

positive and had a 5 percent significance level.This means that higher profit is followed byhigher debt.

Columns (4) and (5) of Appendix 4 pro-vide different evidence. H

7 says that the

change in debt has a strong, positive relation-ship with the change in equity. Column (4)indicates that the coefficient of E

i,t was sig-

nificantly negative, while column (5) indicatesthat the coefficient of E

i,t was positive but

not significant. From the sign and significancelevel point of view, it is clear that H

7 is not

accepted. This also means that STT is notsupported by this model.

Columns (6) and (7) of Appendix 4show the regression results in relation to H

8.

This hypothesis says that the change in debthas a strong, positive relationship with earn-ings after tax, new equity financing, changesin fixed assets, and changes in working capi-tal; a strong, negative relationship with divi-dends, and a moderate, positive relationshipwith the size of the company. The coefficientsthat support the hypothesis were the coeffi-cients of EAT

i,t, and LA

i,t. The other coeffi-

169


cients were not as expected by STT. More-over, the coefficient of NEF

i,t was negative

and significant under the random effectmodel. This was the opposite to the expectedsign of the coefficient because under STT,higher debt has to be accompanied by higherequity financing. Therefore, this hypothesiswas not fully accepted and, therefore, the re-gression result only weakly supported theexistence of STT.

Based on the above analysis on the ex-istence of POT, AMT, and STT, it is clearthat AMT is superior to STT and POT in theway companies in the consumer industry inIndonesia deal with their financing and capi-tal structure issues. The application of AMTin explaining the behavior of companies’ fi-nancing and capital structure could be as fol-lows. Companies tend to increase their loanswhen they earn profits. They do not intendto exploit the internally generated funds tobecome retained earnings. Instead, they ob-tain new loans to fulfill their financing needsto complement their retained earnings. Thisbehavior seems to be in line with STT. How-ever, companies keep increasing their loans,instead of increasing both loans and new ex-ternal equity at a certain set proportion. Theyprefer to increase their loans by keeping paidin capital at a certain level. Based on AMT,the increase in loans when the companies aregrowing shows the aim of using their credi-tors’ ability to control the companies.

The last important point to note fromthis study is the relationships between D

i,t

and FAi,t and WC

i,t. Large companies have

more opportunity and access to acquire debt.These loans need collateral. There are vari-ous types of collateral used in Indonesia, suchas fixed assets, deposits in terms of back-to-back loans, and personal guarantees. Theweak relationship between the changes infixed assets as well as current assets and the

increase in debt may indicate that some com-panies are able to use something other thanfixed and current assets as collateral.

Conclusion

Based on the aforementioned analysis,it may be concluded that: firstly, the POTdoes not work for companies within the con-sumer industry. In other words, their manag-ers are not tempted to exploit profits to ful-fill their financing needs. Supposing the POTworked, this would indicate that the manage-ment tended to exploit the sources of fundsbased on how easily the source of funds wasaccessible. As the POT does not apply to theconsumer industry, that kind of behavior isnot implemented in the industry. The man-agement do not only put their efforts into ob-taining capital based on how easily the capi-tal is accessed, but there is at least one otherconsideration.

Another consideration, based on theevidence from the STT, is that the manage-ment needs to exploit other sources of capi-tal, i.e. debt. The purpose of applying the STTis to minimize the cost of capital and, at thesame time, to maximize the value of the firmby keeping the composition of debt and eq-uity, or the DER at the most optimum level.Based on interviews with some financialmanagers, the increase in the use of the STTis probably due to the increase in the num-bers of financial managers and CFOs withan educational background from financialacademies. The increases in equity are bal-anced by the increased debt. This study doesnot involve itself in the study of the opti-mum DER. However, the significance of theincrease in equity as well as the increase innet profits in relation to the increase in debtindicates such an effort.

Djohanputro

170

However, there is still another consid-eration, even more important than the valuemaximizing consideration, i.e. the control ofcompanies by outsiders, mainly by their credi-tors. The AMT emphasizes the governancemodel, in the sense that the separation be-tween principals and agents creates difficul-ties for principals to control the behavior ofthe agents. Principals have limited access tocompanies, therefore they encourage compa-nies to borrow funds. Creditors always askfor as much information as they need beforeapproving and disbursing any loans. Creditorswant to make sure their funds are safe byevaluating the past performance of borrow-ers, evaluating the future or expected perfor-mance of the borrowers, and by designing acontract to assure that the borrowers’ man-agement behave properly to make sure thefunds are secured and the companies aremanaged according to the investors’ criteria.In this sense, shareholders are indirectly pro-tected by the presence of creditors in thecompanies.

Therefore, the order that the theoriesare implemented in is as follows: The mostimportant is the AMT, followed by the STT,and the least important is the POT.

Note that this order may be generalizedfor companies within the consumer industryin the Indonesian market. However, furthergeneralization still needs other studies to beconducted in order to assure the validity andapplicability of the findings when applied toother areas. There are some research alterna-tives, the first is to conduct similar studiesbased on different industries in the Indone-sian market. If the findings of such studiesare similar to the findings in this study, thegeneralization may be applied to the Indone-sian market. If those findings diverge anddiffer, there may be a uniqueness to the fi-nancing decisions within different industrygroups.

Secondly, studies could be conducted insimilar industries in other markets, such asSingapore, Malaysia, and other neighboringcountries. The third is to conduct studies inthose countries in different industries. It mayalso be interesting to conduct similar studiesin developed countries, in order to comparethe results of such studies against the resultsof studies in Indonesia as a developing coun-try. The purpose is to compare whether thereare similarities and differences in financingdecisions between developed vis a vis a de-veloping country or countries.

References

Acaravci, S. K. 2015. The determinants of capital structure: Evidence from the Turkish manufacturingsector. International Journal of Economics and Financial Issues 5 (1): 158-171.

Ahmed, H. J. A., and N. Hisham, 2009. Revisiting capital structure theory: A test of pecking order andstatic order trade-of model from Malaysian Capital Market. International Research Journal of Financeand Economics 30: 58 – 65.

Almazan, A., and C. A. Molina. 2005. Intra-industry capital structure dispersion. Journal of Economics andManagement Strategy 14 (2): 263-297.

Bradley, M., G. Jarrell, and E. H. Kim. 1984. On the existence of an optimal capital structure: Theory andevidence. Journal of Finance 39: 857-878.

171


Brennan, M., and E. Schwartz. 1978. Corporate income taxes, valuation, and the problem of optimalcapital structure. Journal of Business 51 (1): 103-114.

Capozza, D. R., and P. J. Seguin. 1999. Leverage and value in apartment REITs. The National Multi HousingCouncil (the paper was accessed on 23 July 2011 at http://www.umich.edu/~reecon/restate/fac-ulty/Capozza/Deb0899.pdf).

Frank, M. Z., and V. K. Goyal. 2004. The effect of market conditions on capital structure, adjustment.Financial Research Letter 1: 47-55.

Frank, M. Z., and V. K. Goyal. 2008. Trade off and pecking order theories of debt. In Eckbo, E. 2008,Handbook of Corporate Finance: Empirical Corporate Finance 2 (Chapter 12).

Getzmann, A., L. Sebastian, and K. Spremann. 2010. Determinants of target capital structure and adjust-ment speed – Evidence from Asian Capital Markets. Asian Finance Symposium Paper.

Graham, J. R., and C. R. Harvey. 2001. The theory and practice of corporate finance: Evidence from thefield. Journal of Financial Economics 60: 187-243.

Graham, J. R., and C. R. Harvey. 2002. How do CFOs make capital budgeting and capital structuredecisions? Journal of Applied Corporate Finance 15 (1): 8 – 23.

Harris, M., and A. Raviv. 1991. The theory of capital structure. Journal of Finance 46: 297- 355.

Hatfield, G. B., L. T. W. Cheng, and W. N. Davidson III. 1994. The Determination of optimal capitalstructure: The effect of firm and industry debt ratios on market value. Journal of Financial andStrategic Decisions 7.

Jahan, N. O. 2014. Determinants of capital structure of listed textile enterprises in Bangladesh. ResearchJournal of Finance and Accounting 5 (20): 11-20.

Jensen, M. C. 1986. Agency costs of free cash flow, corporate finance and takeovers. The American Eco-nomic Review 76: 323-329.

Kim, J-E. 2008. Strategic choice and financial structure in casual themed restaurants. Thesis, the Faculty ofthe Virginia Polytechnic Institute and State University (accessed on July 23 2011 on http://scholar.lib.vt.edu/theses/available/etd-10292008-202447/unrestricted/joungeun.pdf).

Kühnhausen, F., and H. W. Stieber. 2014. Determinants of capital structure in non-financial companies.Discussion Paper of Department of Economics, University of Munich (accessed 8 April 2015, available ononline at http://epub.ub.uni-muenchen.de/21167/).

Leland, H. 1994. Corporate debt value, bond convenants, and optimal capital structure. Journal of Finance49 (4): 1213-1252.

Lev, B. 1974. On the association between operating leverage and risk. Journal of Financial and QuantitativeAnalysis: 627-641.

Mandelker, G. N., and S. G. Rhee. 1984. The impact of the degrees of operating and financial leverage onsystematic risk of common stock. Journal of Financial and Quantitative Analysis 19 (1): 45-57.

Morck, R., A. Shleifer, and R. Vishny. 1988. Management ownership and market valuation. Journal ofFinancial Economics 20: 293–315

Mwangi, L. W., M. S. Makau, and G. Kosimbei. 2015. Relationship between capital structure and perfor-mance of non-financial companies listed in the Nairobi Securities Exchange, Kenya. Global Journalof Contemporary Research in Accounting, Auditing and Business Ethics 1 (2): 72-90.

Myers, S. C. 1984. The capital structure puzzle. Journal of Finance 39 (3): 575-592.

Djohanputro

172

Rafiq, M., A. Iqbal, and M. Atiq. 2008. The determinants of capital structure of the chemical industry inPakistan. The Lahore Journal of Economics 13 (1): 139-158.

Shanmugasundaram, G. 2008. Intra-industry variations of capital structure in pharmaceutical industry inIndia. International Research Journal of Finance and Economics 16: 1450-2887.

Siegfried, J. J. 1984. Effective average U.S. corporation income tax rates. National Tax Journal 27: 245-259.

Vo, D. H., and V. T. Nguyen. 2014. Managerial ownership, leverage and dividend policies: Empiricalevidence from Vietnam’s listed firms. International Journal of Economics and Finance 6: 274-284.

Welch, I. 2004. Capital structure and stock returns. Journal of Political Economy 112 (1).

173


APPENDIX 1. Descriptive Statistics of the Variables Under Study

Di,t

EATi,t

Divi,t

Ei,t

FAi,t

(Rp Millins) (Rp Millins) (Rp Millins) (Rp millions) (Rp Millins)

Mean 168,547 924,427 907 599,152 188,896

Minimum (2,463,664) 2,126 0 (364,482) (236,673)

Maximum 5,116,374 8,051,057 16,150 14,825,554 4,080,407

Standard 897,383 1,727,430 2,900 1,986,117 549,494Deviation

WCi,t

FARi,t

L Ai,t

NEFi,t

(Rp Millins) (percentage) (Rp Millins) (Rp Millins)

Mean 337,243 31.08% 13 264,322

Minimum (1,538,147) 0.91% 11 (176,614)

Maximum 8,423,025 62.03% 16 12,916,171

Standard 1,244,192 13.81% 2 1,608,241Deviation

Djohanputro

174

AP

PE

ND

IX 2

. T

he

Reg

ress

ion

Res

ult

s to

Tes

t th

e H

ypo

thes

es o

f P

eck

ing

Ord

er T

heo

ry

Co

mp

on

ent

PO

T M

od

el 1

PO

T M

od

el 2

PO

T M

od

el 3

Ran

do

m E

ffec

tF

ixed

Eff

ect

Ran

do

m E

ffec

tF

ixed

Eff

ect

Ran

do

m E

ffec

tF

ixed

Eff

ect

Co

nst

ant

3955

0.3

-360

876.

229

769.

79-3

8518

1-1

9129

76-3

9293

33

EA

Ti,

t0.

140.

570.

140.

570.

130.

26

`(2.

0625

)**

(2.4

241)

**(2

.035

8)**

(2.4

036)

**(1

.560

4)(0

.964

5)

Ex

pec

ted

sig

n(-

)(-

)(-

)(-

)(-

)(-

)

Div

i,t

9.56

27.0

88.

1122

.26

(0.2

539)

(0.5

014)

(0.2

266)

(0.4

195)

Ex

pec

ted

sig

n(-

)(-

)(-

)(-

)

F

Ai,t

0.01

-0.1

3

(0.0

306)

(-0.

3790

)

Ex

pec

ted

sig

n(+

)(+

)

W

Ci,t

-0.3

7-0

.30

(-3.

9229

)***

(-1.

8897

)*

Ex

pec

ted

sig

n(+

)(+

)

LA

i,t

1599

32.2

3052

26.1

(1.4

506)

(1.0

380)

Ex

pec

ted

sig

n(+

)(+

)

R2

5.53

%48

.50%

5.60

%48

.78%

24.8

2%54

.18%

Ad

just

ed R

24.

18%

22.2

1%2.

87%

20.9

5%19

.13%

24.3

5%

Hau

sman

Tes

t (p

val

ue)

0.05

570.

1494

0.97

51

175


APPENDIX 2. Continued

Notes:

Appendix 2 provides the regression results in relation to H1, shown under columns POT Model 1, hypothesis 2, under

columns POT Model 2, and H3, under columns POT Model 3. The equations of POT model 1 is

(-)

,, ttiti bEATaD , POT model 2 is

(-) (-)

,,, ttititi cDivbEATaD , and POT model 3 is

)( (-) )( )( (-)

,5,4,3,21,

tititititti LAbDivbWCbFAbEATbaD . D

i,t, the dependent variable, is the change in

debt, EATi,t, earnings after tax of company i at time t, DIV

i,t, dividend of company i at time t, FA

i,t change in fixed

assets of company i at time t, WCi,t the change in working capital of company i at time t, and LA

i,t natural logarithm

of Assets of company i at time t. The table shows regressions of every equation using both random and fixed effectmodels. The figures in the table are the coefficients, and the figures in the parentheses are t-statistics, with (***)significance at 1 percent, (**) significance at 5 percent, (*) significance at 10 percent. The signs (-) and (+) in the table areexpected signs according the hypotheses.

Djohanputro

176

APPENDIX 3: The Regression Results to Test the Hypotheses of Agency Model Theory

Component AMT Model 1 AMT Model 2

Random Effect Fixed Effect Random Effect Fixed Effect

Constant 39550.3 -360876.2 -432529 -5477774

EATi,t

0.14(2.0625)** 0.57(2.4241)** 0.13(1.3088) 0.48(1.9716)*

Expected sign (+) (+) (+) (+)

FARi,t

240630 -1823820

(0.2767) (-0.9682)

Expected sign (+) (+)

LAi,t

31444.04 444750.10

(0.2949) (1.5097)

Expected sign (+) (+)

R2 5.53% 48.50% 5.78% 50.99%

Adjusted R2 4.18% 22.21% 1.62% 22.68%

Hausman test (p value) 0.0557 0.1233

Notes:

Appendix 3 provides the regression results in relation to H4, shown under columns AMT Model 1, and H

5, under

columns AMT Model 2. The equations of AMT model 1 is )(

,,

ttiti bEATaD , and AMT model 2 is

)( )( )(

,3,2,1,

ttitititi LAbFARbEATbaD . D

i,t, the dependent variable, is the change in debt, EAT

i,t, earn-

ings after tax of company i at time t, FARi,t the ratio of fixed assets to total assets of company i at time t, and LA

i,t natural

logarithm of Assets of company i at time t.The table shows regressions of every equation using both random and fixedeffect models. The figures in the table are the coefficients, and the figures in the parentheses are t-statistics, with (***)significance at 1 percent, (**) significance at 5 percent, (*) significance at 10 percent. The signs (-) and (+) in the table areexpected signs according the hypotheses.

177

Gadjah Mada International Journal of Business – May-August, Vol. 17, No. 2, 2015A

PP

EN

DIX

4.

Th

e R

egre

ssio

n R

esu

lts

to T

est

the

Hyp

oth

eses

of

Sta

tic

Tra

de-

off

Th

eory

Co

mp

on

ent

ST

T M

od

el 1

ST

T M

od

el 2

ST

T M

od

el 3

Ran

do

m E

ffec

tF

ixed

Eff

ect

Ran

do

m E

ffec

tF

ixed

Eff

ect

Ran

do

m E

ffec

tF

ixed

Eff

ect

Co

nst

ant

3955

0.3

-360

876.

218

9990

.80

1309

25.0

0-1

9430

96-3

7544

73

EA

Ti,

t0.

140.

570.

150.

33

(2.0

625)

**(2

.424

1)**

(1.9

546)

*(0

.984

5)

E

xp

ecte

d s

ign

(+)

(+)

(+)

(+)

Div

i,t

5.73

22.0

0

(0.1

753)

(0.4

105)

E

xp

ecte

d s

ign

(-)

(-)

E

i,t

-0.0

40.

06

(-0.

6425

)(0

.732

4)

Ex

pec

ted

sig

n(+

)(+

)

F

Ai,

t0.

07-0

.12

(0.3

411)

(-0.

3362

)

W

Ci,t

-0.3

3-0

.30

(3.6

983)

***

(-1.

9000

)(*)

Ex

pec

ted

sig

n(+

)(+

)

LA

i,t

1619

68.2

028

7506

.90

(1.6

806)

*(0

.954

7)

Ex

pec

ted

sig

n(+

)

NE

Fi,

t-0

.12

-0.0

4(-1

.785

9)*

Ex

pec

ted

sig

n(+

)(+

)

Djohanputro

178

AP

PE

ND

IX 4

. C

onti

nued

Co

mp

on

ent

ST

T M

od

el 1

ST

T M

od

el 2

ST

T M

od

el 3

Ran

do

m E

ffec

tF

ixed

Eff

ect

Ran

do

m E

ffec

tF

ixed

Eff

ect

Ran

do

m E

ffec

tF

ixed

Eff

ect

R2

5.53

%48

.50%

0.58

%42

.72%

29.5

6%54

.32%

Ad

just

ed R

24.

18%

22.2

1%-0

.84%

13.4

7%23

.06%

22.7

9%

Hau

sman

tes

t (p

val

ue)

0.05

570.

1303

0.00

23

Not

es:

Ap

pen

dix

4 p

rovi

des

the

regr

essi

on

resu

lts in

rel

atio

n to

H6, s

how

n u

nde

r co

lum

ns S

TT

Mo

del 1

, H7,

un

der

colu

mn

s ST

T M

odel

2, a

nd H

8, u

nde

r co

lum

ns

STT

Mo

del

3. T

he e

qua

tions

of

STT

mo

del

1 is

, ST

T m

od

el 2

is ,

and

ST

T m

ode

l 3 is

.

Di,t, t

he

dep

ende

nt

vari

able

, is

the

chan

ge in

deb

t, E

AT

i,t, e

arni

ngs

afte

r ta

x o

f co

mp

any

i at t

ime

t,, D

IVi,t, d

ivid

end o

f co

mpa

ny i

at ti

me

t,

Ei,t th

e ch

ange

in b

ook

val

ue o

f eq

uity

of

com

pan

y i a

t tim

e t,

“F

Ai,t c

han

ge in

fix

ed a

sset

s o

f co

mp

any

i at t

ime

t,

WC

i,t th

e ch

ange

in w

ork

ing

cap

ital o

f co

mp

any

i at t

ime

t, L

Ai,t n

atur

al lo

gari

thm

of A

sset

s o

f co

mp

any

i at t

ime

t, a

nd N

EF

i,t n

ew e

qui

ty f

inan

cin

g o

f co

mp

any

i at

tim

e t.

Th

e ta

ble

sho

ws

regr

essi

ons

of

ever

y eq

uatio

n us

ing

bot

h r

ando

m a

nd f

ixed

eff

ect

mo

dels

. Th

e fi

gure

s in

th

e ta

ble

are

th

e co

effi

cien

ts, a

nd t

he

figu

res

in t

he

par

enth

eses

are

t-st

atis

tics

, with

(***

) sig

nifi

can

ce a

t 1 p

erce

nt,

(**)

sig

nif

ican

ce a

t 5 p

ercc

ent,

(*) s

igni

fica

nce

at 1

0 p

erce

nt. T

he s

ign

s (-

) an

d (+

) in

the

tabl

e ar

e ex

pec

ted

sign

s ac

cord

ing

the

hyp

oth

eses

.

the dominance of the agency model on financing decisions

Documents