TheImpactofFirmSizeonDividendBehaviour: AStudyWithReferencetoCorporateFirms acrossIndustriesinIndia AzhagaiahRamachandran VeeramuthuPackkirisamy The objective of this paper is to examine the association between the Corporate Leverage (cl ) and the Dividend Policy (dp )offirmsacross industries in India in respect of Size of Corporate Firms. The inves- tigation is conducted on a panel sample of 73 firms across industries [Cement, Chemical and Fertilizer, it , Oil and Gas, Pharmaceutical, Shipping, and Textiles], which listed their shares in National Stock Ex- change(nse )inIndiafortheperiod1996 –2007 .TheimpactsofCapital Structure (cs ) variables (leverage) on dp measures – dividend payout (Net dividend paid/net income) in the presence of some basic funda- mental variables are considered to be the determinants of dp , using the Multiple Regression Technique (ols method). The results of the cross-sectional ols Model for the selected sample firms under various sectors show that there is a significant effect of selected independent variables,dpo α+β 1 dpo t−1 +β 2 pat +β 3 tde +β 4 cf +β 5 size +β 6 inv + β 1 ltd +β 8 std + e. Therefore, this study proves that the dp of Small Size, Medium Size, Large Size, and Overall Corporate Firms across in- dustriesinIndiaisdependentonthelevelofdebtincs . KeyWords:capitalstructure,dividendpolicy,corporateleverage, longtermdebt,shorttermdebt,totaldebt jel Classification:g30 ,g32 ,g35 Introduction Fromthepractitioners’viewpoint,dividendpolicy(dp )ofafirmhasim- plications for stakeholders. For investors, dividends – whether declared oraccumulatedandpaidatalaterdate–arenotonlyameansofregular income,butalsoanimportantinputinvaluationofafirm.Thisimplies AzhagaiahRamachandranisanAssociateProfessoratKanchi MamunivarCentreforPostGraduateStudies,Pondicherry Central University, India. VeeramuthuPackkirisamyisaResearchScholaratKanchi MamunivarCentreforPostGraduateStudies,Pondicherry Central University, India. ManagingGlobalTransitions8 (1):49 –78 50 AzhagaiahRamachandranandVeeramuthuPackkirisamy that dividends may have negative consequences too for investors. Simi- larly, the cost of raising funds is not insignificant and may well lead to lower payout, particularly when positive net present value projects are available. Apart from flotation costs, information asymmetry between managers and outside investors may also have implications for dp .Fur- ther, in the presence of information asymmetry and flotation costs, in- vestmentdecisionsmadebymanagersaresubjecttothepeckingorderof financingchoicesavailable. One of the mechanisms of reducing expropriation of outside share- holders by agents is high payout, which will result in reduction of free cash flow available to managers. The presence of information asymme- try may also mean that managers need to signal their ability to generate higher earnings in future with the help of high dividend payouts (Bhat- tacharya (1979 ), Kose and Williams (1985 ), and Miller and Rock (1985 )). Rozeff (1982 ) model payout ratios are presented as a function of three factors:flotationcostsofexternalfunding,agencycostofoutsideowner- ship,andfinancingconstraintsasaresultofhigheroperatingandfinan- cialleverage. StatementoftheProblemsandSignificance The study mainly focuses on the effect of cs on dp of corporate firms across industries in India, and seeks to answer whether the size of the firmwouldappeartobeoneoftheimportantfactorsindeterminingthe dividendbehaviorofcorporatefirmsinIndia. ReviewoftheLiterature Sincestrategiesareaimedatacquiringcompetitivestrength,thisrequires considerable funding, firms need to adopt appropriate financial policies to mobilize risk capital. The cs and dp is a complex set of analysis, as the investment decision and financing decision are important decisions a firm should take in the course of its operation. Gordon (1959 ) exam- ined the three possible hypotheses with respect to what an investor pays forwhenheacquiresashareorcommonstockthatheisbuying:(1)both the dividend and the earnings, (2)thedividends,and( 3) the earning. It maybearguedthatmostcommonlyheisbuyingforthepriceatsomefu- turedate,butifthefuturepricewillberelatedtotheexpecteddividends. Wilson (1967 )arguedthatitshouldnotbepossibletoincreasetheex- pected utility of one member without decreasing the expected utility of someothermember.Hagen(1973 )discussedtheproblemofdetermining ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 51 anoptimaldp forafirmhavingasetofshareholderswithspecifiedpref- erences. An optimal dp will consequently mean a dividend payout rule, which maximizes some utility criterion as defined by the shareholders’ preferences. Michel (1979 ) examined the extent to which industry divi- dendfiguresaffectdeterminationofaparticularfirm’sdp . Woolridge (1983 )analyzedthee ffectofunexpecteddividendchanges of common stock, preferred stock, and bonds. Two potential effects are identified: a wealth transfer effect, and a signaling effect. Kane, Young, and Marcus (1984 ) found that there is a statistically significant interac- tion effect, i. e. that the abnormal return corresponding to any earn- ings or dividend announcement depends upon the value of the other announcement. Miller, and Rock (1985 ) examined the standard finance modelofthefirm’sdividenddecisions.Theextensionendogamieswhich the dividend announcement affects are amply documented in recent re- search.GhoshandWoolridge(1989 )examinedthefocusesonsharehold- ers’reactiontogrowth-motivatedcutsandomissionsandstatedthat,al- thoughgrowthannouncementsmitigatethecapitallossinducedbydiv- idenddecreases,thestock-marketresponsetogrowth-orienteddividend cutsisstillstronglynegative.Lim(1989 )observedthatdividenddepends in part on the firm’s current earnings and in part on the dividend of the previousyear. Lambert, Lanen, and Larcker (1989 ) found the association between the initial adoption of stock options for senior-level executives and sub- sequent changes in corporate dp , and suggested that dividends be re- duced relative to expected dividends. Brennan and Thakor (1990 )ex- amined the preferential tax treatment of capital gains for individual in- vestors; it is shown that a majority of a firm’s shareholders may sup- port a dividend payment for small distribution. For larger distribution open market stocks repurchase is likely to be preferred by a majority of shareholders, and for the largest distribution, tender offer repurchases dominate. Deangelo and Deangelo (1990 )analyzedthedp adjustments of firms to protracted financial distress as evidenced by multiple losses during 1980 –1985 , and found that almost all sample firms reduced div- idends, and more than half apparently faced binding debt covenants in theyearstheydidso.HodderandSenbet(1990 )havedevelopedatheory of cs in an international setting with corporate and personal taxes and highlighted the key role that corporate tax arbitrage plays in generating internationalcs equilibrium. Allen (1991 ) examined the financial managers’ perceptions of the Volume 8 · Number 1 ·Spring2010 52 AzhagaiahRamachandranandVeeramuthuPackkirisamy broad determinants of listed Australian company cs decisions. The results are consistent with Donaldson’s previously reported American findings, in that firms appear to follow a pecking order with respect to funding sources and they also report policies of maintaining spare debt capacity. Yener (1991 ) analyzed the Korean securities market’s reliance on debt financing and emphasis on debt financing as one of the ma- jor issues related to corporate financial policy in Korea. Factors such as persistently high international interest rates, foreign exchange rate fluc- tuations, inflation, international competition and indications of a slow- down in the world trade have led to increased pressure on the liquidity of many growth-oriented Korean firms, especially in the Manufacturing Sector.ChunchiandKao(1992 )foundasignificantrelationshipbetween dividend changes and subsequent earnings. Changes in dp are interest- ing because dividends are the focus of agency conflicts between owners andmanagersoffirms(Rozeff1982 ). Akhigbe,Borde,andMadura(1993 )usedaneventstudymethodology and found that the share price response for insurers is significant and positive. The magnitude of the response for life insurers is smaller than that ofthe other types of insurers or industrial firms, but is greater than thatofthebanks.PapaioannouandSavarese(1994 )examined,inpartic- ular,thenewtaxlawwhichloweredthetoppersonalmarginaltaxratefor dividends. Johnson (1995 ) used recent theoretical models and suggested that debt and dividends can serve as substitute free cash flow control or signalingdevices. Collins,Saxena,andWansley(1996 )haverecognizedthepotentialdif- ferencesindp betweenregulatedandunregulatedfirms,andfocusedon agency-costandmonitoringexplanationsfortherelevanceofdividends, revealing that there are fundamental differences in the relationship be- tween insider holdings and dp for unregulated firms and utilities, but suggesting that the regulatory environment enhances rather than miti- gates the importance of the insiders’ role for utilities. Elston (1996 )ana- lyzedtheimportanceofdp andliquidityconstraintsinthecontextofthe firm’sinvestmentbehavior,suggestingthataftercontrollingforthefirm’s dividend payment, liquidity constraints remain an important determi- nantofthefirm’sinvestmentbehavior.GulatiandZantout(1997 )found thatimmunizingthefirm’srealgrowthpotentialagainsttheeffectsofin- flation and interest ratefluctuationsgenerally requires frequent changes initscs . Koch and Shenoy (1999 )consideredabroaderdistinctionamongthree ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 53 types of firms’ value – maximizing firms (Tobin’s q close to 1), over in- vesting firms (q< 1), and under investing firms (q> 1). Using this in- terpretation of the free-cash-flow hypothesis, dividend and cs changes shouldreflectalargerchangeinagencycosts(andthusalargerinforma- tioneffect)forbothlowandhighqfirmsthanthatforfirmswithqvalues close to 1. Feed Back Measures (gfm s) found a distinct U-shaped rela- tion between Tobin’s q and the amount of predictive information con- tainedinafirm’sdividendandcs policies,withaminimumataqvalue near one. Mohanty (1999 ) examined whether the firms offering bonus issue have been able to generate greater returns for their sharehold- ers than those which have not offered any bonus issue but have main- tained a steadily increasing dividend rate, and found that a few firms increased the dividend rate after a bonus issue, while the bonus issuing firms fielded greater returns to their shareholders than those which did notmakeanybonusissuebutmaintainedasteadily increasing dividend rate. Romano, Tanewski, and Smyrnios (2000 ) examined the factors that influence small-medium enterprises (sme ) owner-managers’ financing decisions and found that these processes are influenced by firm own- ers’ attitudes toward the utility of debt as a form of funding as mod- erated by external environmental conditions, in addition to a number of other factors: e. g., culture; entrepreneurial characteristics; entrepre- neurs’ prior experiences in cs ; business goals; business life-cycle issues; preferred ownership structures; views regarding control, debt equity ra- tios,andshort-vs.long-termdebt;ageandsizeofthefirm;sourcesof funding for growth; and attitudes. La Porta et al. (2000 ) examined the ‘outcome model’ and found that dividends are paid because minority shareholderspressurecorporateinsiderstodisgorgecash;the‘substitute model’ reveals that insiders are interested in issuing equity in the future paydividends toestablish areputationfordecent treatment ofminority shareholders. Ooi(2001 )analyzedbyemployingpaneldatamethodology;thedp of property firms quoted on the London Stock Exchange (lse )showsthat the dividend payout ratio of the average real estate corporation is dic- tated,toalargeextent,bythefirm’stotalassetholdingandleverageratio. Propertyinvestmentfirmspaysignificantlyhigherdividendswhencom- paredtopropertytradingfirms.Boothetal.(2001 )examinedwhethercs theoryisportableacrosscountrieswithdifferentinstitutionalstructures, andprovidedevidencethatthesedecisionsareaffectedbythesamevari- Volume 8 · Number 1 ·Spring2010 54 AzhagaiahRamachandranandVeeramuthuPackkirisamy ables as exist in developed countries. Kumar and Lee (2001 ) examined how to develop an empirically dynamic model of discrete dp based on an inter-temporal signaling framework, in which dividend adjustments signal only substantialvariations inthepermanent earningsofthefirm, and showed that dividend smoothing is positively associated with fac- tors such as, earnings variance, low liquidity, and high probability of bankruptcy, as well as the expected return on capital investment by the firm. Goldstein, Ju, and Leland (2001 )foundthatmostcs models assume that the decision on how much debt to issue is a static choice; however firms adjust outstanding debt levels in response to changes in the firm’s value.Ahmedetal.(2002 )examinedthisusingbothamarket-basedand anaccrual-basedmeasureofconservatism,andfoundthatthefirmsfac- ingmoresevereconflictsoverdp tendtousemoreconservativeaccount- ing; they document that accounting conservatism is associated with a lower cost of debt after controlling for other determinants of the firm’s debtcosts.JohnGrahamandHarvey(2002 )examinedfinancetheory,as well as aspects that are hard to reconcile and found systematic relation- shipsbetweencorporatefinancialchoicesandmanagerialfactors,suchas theextentoftopmanagement’sstockownership,andtheage,tenure,and educationofthechiefexecutiveofficer(ceo ).BakerandWurgler(2002 ) found that firms are more likely to issue equity when their market val- ues are high, relative to book and past market values, and to repurchase equitywhentheirmarketvaluesarelow.Asaconsequence,currentcs is strongly related to historical market values. The results suggest the the- orythatcs isthecumulativeoutcomeofpastattemptstotimetheequity market. Stenbacka and Tombak (2002 ) analyzed the simultaneous investment and financing decisions made by incumbent owners in the presence of capitalmarketimperfections,representingatheoryforhowtheoptimal combinationofdebtandequityfinancingdependsonthefirm’sinternal funds,andidentitycomplementaritiesbetweenthetwofinancialinstru- ments. Mao (2003 ) presented a unified analysis that accounts for both riskshiftingandunder-investmentdebtagencyproblems.Forfirmswith positive marginal volatility of investment, equity holders’ risk-shifting incentive will mitigate the under-investment problem, which implies that contrary to conventional views, the total agency cost of debt does not uniformly increase with leverage, and predicts that for high growth firms in which the under-investment problem is severe, the optimal ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 55 debt ratio is positively related to the marginal volatility of investment. Gugler (2003 ) analyzed the relationship between dividends and the ownershipandcontrolstructureofthefirmforapanelofAustrianfirms over the 1991 –1999 period, and found that state-controlled firms engage in dividend smoothing, while family-controlled firms do not. Campello (2003 ) examined firm- and industry-level evidence of the effects of cs on product market outcomes for a large cross-section of industries over a number of years, and found that debt financing has a negative impact on firm’s (relative- to-industry) sales growth in industries in which ri- vals are relatively un-levered during recessions, but not during booms. Graham, Lang, and Shackelford (2004 ) found that employee stock op- tion deductions lead to large aggregate tax savings for Nasdaq. For s&p firms, in contrast, option deductions do not affect marginal tax rates to alargedegree. Anand (2004 ) analyzed most valuable public sector undertakings (psu s) in India to find out the determinants of the dp decisions of the corporatefirmsinIndia,andrevealedthatthefindingsareinagreement with Lintner’s study on dp .Thedp is used as a signaling mechanism to convey information on the present and future prospects of the firm, and thus affects its market value. Mihir et al. (2004 ) examined the cs of foreign affiliates and internal capital markets of multi-national cor- porations (mnc s) and found that the mnc s appear to employ internal capital markets opportunistically to overcome imperfections in external capitalmarkets.NishiokaandBaba(2004 ),whoinvestigatedthedynam- ics of cs of Japanese firms, found that the trade-off theory provides an appropriate framework to assess this issue after controlling for various variables as proxies for other hypotheses, including governance struc- ture, the pecking order theory, and market-timing hypothesis. Among such variables, profitability as a proxy for the pecking order theory has significantexplanatorypower. Sharma (2006 ), who examined the focuses on the dividend trends of selectedIndianfirms,foundastrongconfirmationforthesignalingthe- ories of Bhattacharya (1979 ), and Miller and Rock (1985 ), which gives inconclusive results about the tax-effect theory. Graham and Tuckerb (2006 ) investigated the magnitude of tax shelter activity to analyze whether participating in a shelter is related to corporate debt policy, and found that the average annual deduction produced by the shelters in their sample is very large, equalling approximately nine per cent of asset value. Faulkender, Milbourn, and Thakor (2006 ) presented an in- Volume 8 · Number 1 ·Spring2010 56 AzhagaiahRamachandranandVeeramuthuPackkirisamy tegrated theory ofcs anddp , in which both financial policy choices are drivenbythesameunderlyingfactorsandjointlydeterminedasimplicit governance mechanisms to allocate control over real decisions between managersandinvestors.Singhania(2006 )examinedthedividendtrends of manufacturing, non-government, non-financial, and non-banking companies listed on Bombay Stock Exchange (bse ) and found that the tax preferences theory does appear to hold true in the Indian context in the Indian case of both the categories of firms, i.e., regular payer and non-regular payer, and also found that there is a significant difference in average dividend payout ratio in the two different tax regimes, and also that there are wide industry-wise variations in empirical findings. Sharma(2007 )offersmixedandinconclusiveresultsaboutthetax-effect theory,whichisnotapplicabletotheselectedIndianfirms,thusindicat- ingthatthechangeinthetaxstructuredoesnothaveasubstantialeffect on the dividend behavior of firms. Kalea and Shahrurb (2007 )found that the firm’s leverage is negatively related to the r &d intensities of its suppliers and customers. Anil and Kapoor (2008 ) found that profitabil- ityhasalwaysbeenconsideredasaprimaryindicatorofdividendpayout ratio, while there are numerous factors other than profitability that also affect the dividend decisions of an organization e. g., cash flows, cor- porate tax, sales growth, market to book value ratio, and size. Dividend payoutratioispositivelyrelatedtoprofits,cashflows,andsizeand,ithas an inverse relationship with corporate taxes, sales growth and market to bookvalueratio. Thoughanamplenumberofresearchstudieshasbeenundertakenin the field of cs and dp , very few of them have associated the effect of cs ondp basedonsizeofthefirms.Therefore,tofillthisgapintheliterature and to shed light, the present paper attempts to analyze the effect of cs ondp ,consideringthesizeofthefirmsacrossindustriesinIndia. ScopeoftheStudy The paper is an attempt to provide an empirical support to the hypoth- esizedrelationshipbetweencs anddp inrespectofthesizeofcorporate firmsacrossindustries.Hence,thestudyproposestoseekanswerstothe followingstatedquestions: How far does the corporate firms’ mix of cs policies dynamically interact overtime to influence firms’ performance with respect to dp ? ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 57 Isthereasignificantimpactofcs ondp basedonsize? Howfararethecs anddp inter-related? ObjectivesoftheStudy Toanalyzetheimpactofcs ondp inrespectofthesizeofcorporate firmsacrossindustriesinIndia. To suggest appropriate measures with respect to the inter-depend- ence of cs and dp in respect of the size of corporate firms across industries. h 1 0 Thereisnosignificantrelationshipbetweenthelevelofdebtincs and levelofequitydividendinCementIndustry,ChemicalandFertilizer Industry, Information Technology Industry, Oil and Gas Industry, PharmaceuticalIndustry,ShippingIndustry,andTextileIndustry. h 2 0 There is no significant relationship between the level of debt in cs and level of equity dividend in small size firms, medium size firms, largesizefirms,andallselectedfirmsacrossallselectedsectors. h 3 0 Thereisnosignificanteffectofselectedindependentvariablesondpo of Cement Industry, Chemical and Fertilizer Industry, Information Technology Industry, Oil and Gas Industry, Pharmaceutical Indus- try,ShippingIndustry,andTextileIndustry. h 4 0 Thereisnosignificanteffectofselectedindependentvariablesondpo of small size firms, medium size firms, large size firms, and all firms underselectedsectors. Methodology sourcesofdata The study used only secondary data, which are collected from cmie [CenterforMonitoringIndianEconomyPrivateLimited]ProwessPack- age.Thedatacollectedfromthissourcehavebeencompiledandusedas pertheobjectivesofthestudy. samplingdesign Thestudyhasbeenmadeonasampleof73 corporatefirmsacrossseven industriesinIndia.Theseindustrieshavebeenchosenbasedonastratifi- cationprocessinrespectofdividendhighyieldingsectors.Thestratifica- tion process for the choice of corporate firms across industries has been adopted based on the asset value of firms, i.e., corporate firms whose total assets value has been significantly increasing over the period have Volume 8 · Number 1 ·Spring2010 58 AzhagaiahRamachandranandVeeramuthuPackkirisamy table1 Numberofcorporatefirmschosenforthestudy Industry Numberofcorporatefirms Cement 12 ChemicalandFertilizer 10 it8 OilandGas 10 Pharmaceutical 15 Shipping 10 Textile 8 Total corporate firms 73 been included in the sample of corporate firms, in this way of stratifi- cation the sample of 73 corporate firms has been arrived at, after giving due consideration to the parameters, i. e., proper and regular dividend payment to shareholders, and availability of required data for the study period. Further, the sample corporate firms are classified into three groups basedonassetsvalueviz.,smallsizefirms–firmswhosetotalassetsvalue isuptoRs.100 crore(tenmillion);mediumsizefirms–firmswhosetotal assets value is between Rs.100 crore (ten million) and Rs.500 crore (50 million); large size firms – firms whose total assets value is larger than Rs.500 crore(50 million). ToolsUsedforAnalysis ratios std _ta = Short Term Debt Total Assets ·100 (1) ltd _ta = LongTerm Debt Total Assets ·100 (2) td _ta = Total Debt Total Assets ·100 (3) correlationcoefficient (karlpearson ’scoefficient ofcorrelation ) Thesignificanceofthecorrelationcoefficient= r (1−r) 2 (n−2) (4) Degreesoffreedom= (n−2)( 5) Correlationanalysisiscarriedouttofindouttheexistenceofmulti-co ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 59 linearityamongindependentvariables,inordertodecidewhatvariables can be used in the ols regression model, or how the regression model withallindependentvariablescanbeused. olsregressions Here, the impacts of cs variables (leverage) on dividend policy mea- sures – dividend payout (net dividend paid/net income) in the pres- ence of some basic fundamental variables–are considered to be the de- terminants of dividend policy using the multiple regression technique (ols method).Beforeusingtheols method,thedegreesofrelationship among independent variables as well as between independent and de- pendentvariableswereanalysedwithPearson-productmomentcorrela- tion. It is appropriate to use the regression technique with the step-wise procedure,ifthereisanycollinearityamongsomeindependentvariables. Thespecificationoftheregressionmodelisgivenbelow: dpo =α+β 1 dpo t−1 +β 2 pat +β 3 tde +β 4 cf +β 5 size +β 6 inv +β 7 ltd +β 8 std +e,( 6 ) wheredpo =DividendPayoutRatio,dpo t−1 =Laggeddpo ,pat =Profit After Tax (Net Income), tde = Total Distributable Earnings, cf =Cash Flow,size =FirmSize(naturallogarithmofTotalAssets),inv =Capital Expenditure,ltdta = Long-Term Debt to Total Assets, stdta =Short- Term Debt to Total Assets,β=estimatedcoefficients,α=interceptterm, e=error. chowtest TheChowTestformulais: F(k,n 1 +n 2 −2k)= [sse p −(sse 1 +sse 2 )]:k (sse 1 +sse 2 ):(n 1 +n 2 −2k) ,( 7) where sse p = sum of squared error term for pooled model, sse 1 =sum of squared error term for group 1,sse 2 = sum of squared error term for group2,k=numberofestimatedparameters(includingconstant), n 1 + n 2 = n ’sforeachofgroup1 andgroup2,respectively. PeriodoftheStudy The data used for the study relate to the selected corporate firms across industries in India for the period of ten years, on a yearly basis ranging from1996 –1997 to2006 –2007 . Volume 8 · Number 1 ·Spring2010 60 AzhagaiahRamachandranandVeeramuthuPackkirisamy LimitationsoftheStudyandScopeforFurtherStudy The study is limited to only 7 industries. Therefore, this comprises the trend of only a few numbers of industries, which would not be sufficient, totally, to generalize the inferences to the whole of a country,India. The data used for thestudyare secondary in nature. Therefore, the accuracyoftheresultsofanalysisisdependent,too,upontherelia- bilityandaccuracyofthecompiledsecondarydata. Further studies could be undertaken by future researchers in the fol- lowingaspectsandareas: by undertaking studies in other industries, new and interesting in- ferencescouldbefound; by categorizing the firms into various classes based on other bases, proportionofcapitalelements,e.g.,debtandequitystudies,could alsobeconducted. MajorFindingsonAcross-IndustryAnalysis Tables 2 to 9 present the results of regression analysis for sample firms undersevensectors. Itisevident(seetable2)thatlaggeddividendpayouthasasignificant positive effect on dpo (β= 0 .2605 , t= 2.77 , p< 0 .01)andinv has a significantnegativeimpactonthedependentvariable(β=−0 .0697 , t= −2.03 , p< 0 .05 ). This shows that the dividend payout in the previous year plays a vital role in determining the current year dividend payout of sample firms under Cement Sector. However, the increase in capital expenditure decreases the level of dividend payout significantly. In the full model with addition of cs variables, only the said variable is found tohaveasignificantcoefficientwithdpo . None of the debt variables has a significant impact on dpo ,astheir coefficients are insignificant. However, coefficient of determination is found to have increased by two per cent. So, [Chow-test] F is calculated to find whether there is a collective impact of those two cs variables on dpo .TheF is found to be insignificant, providing strong evidence that cs does not have any effect on dp in terms of distribution of dividend payout relative to net income of the sample firms under Cement Sector. Therefore, h 1 0 in respect of Cement Sector is rejected, as dpo t−1 has a positive impact at 1%level( 0 .2605 ), and inv has a negative impact at ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 61 table2 ResultsofCross-Sectionalols ModelforsamplefirmsunderCementSector (meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 9 .3640 ** 2.11 25 .7511 ** 2.04 dpo t−1 0 .2605 *** 2.77 0 .2342 ** 2.46 pat0 .0170 1 .50 0 .0107 0 .88 tde 0 .0298 1 .63 0 .0329 1 .79 cf ———— S i z e ———— inv –0 .0697 ** –2.03 –0 .0697 ** –2.03 ltd _ta ——– 0 .2628 –1.56 std _ta ——– 0 .0577 –0 .19 R 2 0 .1462 0 .1680 Adjusted R 2 0 .1131 0 .1185 F-value 4.41 *** 3.340 *** 4.103 6 .101 ChowTestF-value 1.32 2.101 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. 5%level(– 0 .0697 ) in the reduced model; with respect to the full model dpo t−1 (0 .2342 ),andinv (–0 .0697 )at5%level. The regression analysis (see table 3) shows that both, the reduced model and the full model, are fitted significantly (F-value = 3.70 , p < 0 .01 and F = 2.70 , p < 0 .05 ). The reduced model fitted with only lagged dpo , pat , tde and cf , and together explaining 14 .84 per cent of the variationindpo (R 2 =0 .1484 ).Thecoefficientsofpat withpositivesign andoftde withnegativesignaresignificant.Thatis,netprofitincreases thedpo ,andanyincreaseintde decreasesit.Thenegativerelationship betweentde anddpo indicatesthatthesamplefirmsunderthisindus- try have reduced dividend payout when a portion of net income is held for future investments (Reserve and Surplus). With the best-fitted (re- duced)model,leveragevariablesareaddedandthefullmodelisrun.R 2 valuehasincreasedbutthesignificanceofpat andtde hasdisappeared. Further, none of the coefficients of the debt variable is found to be sig- nificant, revealing that they do not have a unique impact on dpo after Volume 8 · Number 1 ·Spring2010 62 AzhagaiahRamachandranandVeeramuthuPackkirisamy table3 ResultsofCross-Sectionalols ModelforSampleFirmsunderChemicaland FertilizerSector(meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 26 .2192 *** 4.81 35 .0989 *** 3.14 dpo t−1 0 .1954 1 .65 0 .2300 * 1.82 pat0 .0676 * 1.92 0 .0452 1 .12 tde –0 .0188 ** –2.23 –0 .0150 –1.61 cf 0 .0132 1 .57 0 .0150 1 .76 inv ———— ltd _ta ——– 0 .1525 –0 .75 std _ta ——– 0 .4550 –1.08 R 2 0 .1484 0 .1632 Adjusted R 2 0 .1084 0 .1028 F-value 3.70 *** 2.70 ** 4.85 6 .83 ChowTestF-value 0 .73ns 2.83 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. partiallingouttheeffectofsomecharacteristicsoffirms.Butthenegative sign of the coefficient has shown that increase in debt financing in cs is likely to reduce the dpo . F (Chow) for both ltd and std reveals that there has been an increase in R 2 value of the full model. But, F (Chow) is found to be insignificant, providing evidence that cs does not play a vital role in determining the dividend payout relative to net income of sample firms under the Chemical and Fertilizer Sector. Hence, h 1 0 is re- jected in respect of pat at 10 %level( 0 .0676 ), and tde negatively at 5% level (–0 .0188 ) in reduced model; with respect to the full model dpo t−1 (0 .2300 )at10 %levelforChemicalandFertilizerSectorinrespectof pat andtde . The analysis shows (see table 4) that the reduced model, even after step-wise procedure, is not fitted significantly. But coefficients of ex- planatory variables in the model are significant at a level of 10 per cent. This may be due to the existence of high collinearity between pat and tde . The significant negative coefficient of lagged dpo indicates that payout of dividend from net income is reduced if dividend payout in ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 63 table4 ResultsofCross-Sectionalols ModelforSampleFirmsunderInformation TechnologySector(meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 70 .8656 2 .76 80 .2616 3 .00 dpo t−1 –0 .0737*– 1.81 –0 .0603 –1.41 pat0 .0237 ** 2.12 0 .0196 * 1.66 tde –8.7622*– 1.85 –9 .3753 ** –1.96 cf ———— S i z e ———— inv ———— ltd _ta ——– 0 .2897 –1.20 std _ta ——– 0 .0894 –0 .18 R 2 0 .0736 0 .0983 Adjusted R 2 0 .0327 0 .03 F-value 1.80ns 1 .44 *** 3.68 5 .66 ChowTestF-value 0 .90 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. the previous year is high. The full model with leverage variables is also notfittedsignificantly,andexplanatorypowerofthemodel(R 2 values)is higherwhencomparedtothatofthereducedmodel.Butthesignificance of lagged dpo has disappeared in the presence of leverage. The negative sign of leverage variables reveals that there are chances of reduction in dpo with increase in debt fund. On the whole, it is found that sample firmsbelongingtoInformationTechnologySectorhavekeptgivingdivi- dends,irrespectiveoftheirperformance.Hence,h 1 0 isrejectedinrespect ofdpo t−1 negativelyat10 %level(–0 .0737 )andpat at5%level(0 .0237 ), and tde negatively at 10%level(– 8.7622 ) in reduced model for Infor- mationTechnologySector;withrespecttothefullmodelpat (0 .0196 )at 10 %level,andtde negativelyat5%level(–9 .3753 ). Theanalysis(seetable5)showsthatboththereduced(F-value=10 .40 , p<0 .01 )andfullmodels(F-value=7.21 ,p<0 .01 )forOilandGasSector arefittedsignificantlyat1 percentlevel. Theexplanatoryvariableinthereducedmodelexplains26 .63 percent ofthevariableindpo .Also,thecoefficientsoflaggeddpo (β=0 .4485 ,p Volume 8 · Number 1 ·Spring2010 64 AzhagaiahRamachandranandVeeramuthuPackkirisamy table5 ResultsofCross-Sectionalols ModelforsamplefirmsunderOilAndGas Sector(meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 14 .0944 *** 4.72 20 .8602 *** 3.69 dpo t−1 0 .4485 *** 4.53 0 .3920 *** 3.86 pat0 .0032 * 1.69 0 .0030 1 .56 tde –0 .0005 –1.39 –0 .0005 –1.49 cf ———— S i z e ———— inv ———— ltd _ta ——– 0 .1809*– 1.87 std _ta ——– 0 .0710 –0 .57 R 2 0 .2663 0 .3002 Adjusted R 2 0 .2407 0 .2585 F-value 10 .40 *** 7.21 *** 3.86 5 .84 ChowTestF-value 2.03 2.84 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. <0 .01 )andpat (β=0 .0032 ,p<0 .10 )aresignificantwithapositivesign. From the significant positive co-efficient of these variables, it is inferred that the increase in net profit kept increasing the dividend payout every yearamongthesamplefirmsunderOilandGasSector.Theexplanatory power of the full model with addition of leverage variables is found to have increased by 3.39 per cent (R 2 = 0 .3002 when compared to R 2 = 0 .2663 for the reduced model). Also the coefficient of leverage variable, ltd _ta is significant positively (β=– 0 .1809 , p < 0 .10 ), indicating that the dividend payout has come down to a significant level when there hasbeenaconsiderableincreaseindebtfinancingincs fromlong-term sources. However, both the leverage measures together failed to explain vari- ation in dpo ,asF (Chow) (F = 2.03 , p > 0 .10 ) is insignificant, i.e., the sample firms under Oil and Gas Sector have not considered the status of debt financing in cs before distributing a part of their net income as dividend.Hence,h 1 0 isrejectedinrespectofdpo t−1 positivelyat1%level ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 65 table6 ResultsofCross-Sectionalols ModelforSampleFirmsunder PharmaceuticalSector(meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 30 .5227 ** 2.35 27 .3356 ** 2.10 dpo t−1 0 .4011 *** 4.86 0 .3999 *** 4.90 pat –0 .0396 –1.28 –0 .0276 –0 .88 tde 0 .0180 ** 2.38 0 .0190 ** 2.44 cf –0 .0044 –0 .94 –0 .0099*– 1.87 size –3.4379 –1.41 –3.6663 –1.48 inv ———— ltd _ta —— 0 .0018 0 .02 std _ta —— 0 .4469 ** 2.15 R 2 0 .2001 0 .2285 Adjusted R 2 0 .1691 0 .186 F-value 6 .45 *** 5.37 *** 5.129 7 .127 ChowTestF-value 2.34 * 2.127 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. (0 .4485 )andpat at10 %level(0 .0032 )inthereducedmodel;withrespect tothefullmodeldpo t−1 positively(0 .3920 )at1%level,andltd _ta neg- atively at 10%level(– 0 .1809 )inrespectofdpo t−1 , pat and ltd _ta for OilandGasSector. The full model with inclusion of leverage variables of sample firms underPharmaceuticalSectorisfittedsignificantlywiththecoefficientof determination to the extent of 22 .85 per cent (R 2 = 0 .2285 , F = 5.37 , p < 0 .01 ). Further, in the full model the effect of cf becomes significant with a negative sign (β=–0 .0099 , p <0 .01 ). Between leverage variables, the coefficient of std _ta is significant positively at 5 per cent level (β = 0 .4469 , p < 0 .01 ). Therefore, the sample firms kept the dividend pay- out on the positive side when they have sizeable fund in reserves and surpluses through borrowing from short-term sources even if there has been a marginal decline in pat as well as a notable decline in cf .The F result (Chow) (F = 2.34 , p < 0 .01 ) [significant] reveals that cs with a sizeablelevelofshorttermfundandmeagreleveloflong-termfundhave Volume 8 · Number 1 ·Spring2010 66 AzhagaiahRamachandranandVeeramuthuPackkirisamy table7 ResultsofCross-Sectionalols ModelforSampleFirmsunderShipping Sector(meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 7.2854 ** 2.31 1 .2477 0 .26 dpo t−1 0 .4928 *** 5.54 0 .4392 *** 4.70 pat –0 .0382 –1.38 –0 .0327 –1.17 tde 0 .0066 1 .04 0 .0057 0 .90 cf ———— S i z e ———— inv 0 .0203 1 .48 0 .0161 1 .16 ltd _ta —— 0 .2320 * 1.77 std _ta —— 0 .0205 0 .04 R 2 0 .3033 0 .3292 Adjusted R 2 0 .2705 0 .2807 F-value 9 .25 *** 6 .79 *** 4.85 6 .83 ChowTestF-value 1.60 2.83 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. significant explanatory power on the dpo of sample firms under Phar- maceutical Sector. Hence, h 1 0 is rejected in respect of dpo t−1 [positively at1%level(0 .4011 )] andtde at5%level(0 .0180 )inthereducedmodel; with respect to the full model dpo t−1 [positively (0 .3999 )] at 1%level, and tde at 5%level( 0 .0190 ), cf negatively at 10%level(– 0 .0099 ), and std _ta positivelyat5%level(0 .4469 )inrespectofdpo t−1 ,tde ,cf ,and std _ta ofPharmaceuticalSector. Thefirms underShipping Sectorare significantly positivelyrelated to lagged dpo (r = 0 .5243 , p < 0 .01)andltd _ta (r = 0 .3546 , p < 0 .01 ). Only pat andtde (r =0 .9196 ,p<0 .01 )andtde andinv (r =0 .8412 ,p <0 .01 )arecollinearwitheachother.Inordertoknowwhichissuperior over the other in explaining dpo when otherwise held constant, pat , tde and inv are included in the reduced model, and step-wise proce- dureiscarriedtogetthemodelofbestfit.Theresults(seetable7)show that the degree of collinearity between pat and tde has come down to marginal level in the presence of lagged dpo and inv (capital expendi- ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 67 ture), because these two variables are found in the reduced model even after the step-wise process. All the four explanatory variables in the re- duced model could explain to the extent of 30 .33 per cent of variation significantlyindpo (F-value=9 .25 ,p<0 .01 ). Asfarastheestimatedco-efficientoftheexplanatoryvariables,inthe reduced model, are concerned, only the co-efficient of lagged dpo (β = 0 .4928 , p < 0 .01 ) is significant positively, which reveals that the sample firms under Shipping Sector have kept increasing the dividend payouts overtheperiodwhenallothersareheldconstant.Thefullmodelwithcs proxies,ltd _ta andstd _ta isalsofittedsignificantlywithaco-efficient of determination at 32 .92 per cent in dpo (R 2 = 0 .3292 , F-value = 6 .79 , p < 0 .01 ). Also, the co-efficient of ltd _ta is significant positively at 10 percentlevel(β=0 .2320 ,t =1.77 ,p<0 .01 ).Thisrevealsthatthesample firms under Shipping Sector have kept paying dividend irrespective of the level of increase in debt fund in cs through long-term financing. Thedifferenceinexplainedvariance(R 2 )betweenthetwomodelsisnot significant (Chow F =1.63 is insignificant), indicating that the influence of long-term debt financing on dpo has disappeared with a marginal increaseinshort-termfundincs .Hence,thereisnoimpactofcs ondp of the sample firms under Shipping Sector, and therefore h 1 0 is rejected inrespectofdpo t−1 positivelyat1%level(0 .4928 )inthereducedmodel; withrespecttothefullmodeldpo t−1 positively(0 .4392 )at1%level,and ltd _ta positivelyat10 %level(0 .2320 )ofShippingSector. In respect of sample firms under Textile Sector, no multi-collinearity amongtheindependentvariablesisfound,andthedpo isnotcorrelated withallthevariables,providingevidencethatdistributionofpartofthe net income as dividend is independent of earnings, cash flow, and debt fund in cs . However, the analysis shows that the reduced model with coefficient of determination to an extent of 7.75 per cent (see table 8)is fitted significantly (R 2 = 0775 ,F-value =2.90 , p <0 .10 ), whereas the full modelisnot,thoughtherehasbeenamarginalincreaseintheexplained variationwithinclusionofcs variables(R 2 =0 .0810 ,F =1.47 ,p>0 .10 – insignificant).Butthesignificantco-efficientof pat (β=1.9594 ,t =2.11, p<0 .05 )withapositivesignandthatoftde (β=–0 .3547 ,t=–2.13 ,p< 0 .05 )withanegativesigninboththemodelsindicatesthatdpo ismore thanthatofthenetprofit,anddividendpaidfromthetotaldistributable earningswithoutconsideringthedebtfinancingincs forTextileSector. Hence, h 1 0 is rejected in respect of pat positively at 5%level( 1.9594 ), and tde negatively at 5%level(– 0 .3547 ) in the reduced model; with re- Volume 8 · Number 1 ·Spring2010 68 AzhagaiahRamachandranandVeeramuthuPackkirisamy table8 ResultsofCross-Sectionalols ModelforSampleFirmsunderTextileSector (meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 20 .3188 *** 2.98 21 .4650 0 .80 dpo t−1 ———— pat1 .9594 ** 2.11 1 .9488 ** 2.01 tde –0 .3547 ** –2.13 –0 .3929*– 1.97 cf ———— S i z e ———— inv ———— ltd _ta —— 0 .0265 0 .07 std _ta ——– 0 .1521 –0 .31 R 2 0 .0775 0 .0810 Adjusted R 2 0 .0508 0 .0261 F-value 2.90 * 1.47ns 2.69 4 .67 ChowTestF-value 0 .13 2.67 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. specttothefullmodel,pat positively(1.9488 )at5%level,andtde nega- tivelyat10 %level(–0 .3929 ).Hence,H 1 0 isrejectedinrespectoftheeffect of pat andtde ondpo ofTextileSector. FirmSize-WiseAnalysis Therelationshipbetweendp measuredasdpo (dividendpayoutpaid/net income)andcs forsamplefirmswithsmall,mediumandlargesizetotal assets is analyzed and the results are shown in tables 9 –12 . dpo is pos- itively related with lagged dpo (r = 0 .1953 , p < 0 .05 ), pat (r = 0 .1950 , p < 0 .05 ), tde (r = 0 .1680 , p < 0 .10 ), and negatively associated with cf (r=– 0 .1793 , p < 0 .10 )andinv (r=– 0 .3634 , p < 0 .01 ). From significant correlation coefficients, it is found that the current year dpo has im- pactedonthepreviousyeardpo ,increaseinpat anddecreaseincapital expenditure(inv )amongsmallsizesamplefirms. With regard to the unique impact of control and debt variables, the results of regression analysis using reduced and full model for small size ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 69 table9 ResultsofCross-Sectionalols Modelforsmallsizefirmsunderallselected sectors(meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 23 .6180 *** 5.61 18 .5859 *** 2.73 dpo t−1 0 .2350 ** 2.04 0 .2323 ** 2.00 pat0 .4259 1 .64 0 .4732 * 1.77 tde ———— cf ———— S i z e ———— inv –1.1126 *** –3.78 –1.1996 *** –3.81 ltd _ta ——0 .1216 0 .89 std _ta ——0 .1182 0 .47 R 2 0 .1927 0 .2003 Adjusted R 2 0 .1675 0 .1578 F-value 7.64 *** 4.71 *** 3.96 5 .94 ChowTestF-value 0 .45 2.94 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. samplefirmsacrosstheselectedsectorsreveal(seetable9 )thatboththe reduced model (R 2 = 0 .1927 , F = 7.64 , p < 0 .01)andfullmodel( R 2 = 0 .2003 , F =4.71 , p <0 .01 ) are fitted significantly, explaining to an extent of 19 .27 per cent and 20 .03 per cent of the variation respectively for the reducedmodelandfullmodelinrespectofdpo .Theco-efficientof pat issignificantatthelevelof10 percent,whichshowsthatthelaggeddpo and pat have a positive effect on dpo when there has been a decline in capitalexpenditure(inv ). Whereas, the co-efficient of ltd _ta and std _ta is not significant, revealing the fact that the dpo is independent of the debt level in cs of small size sample firms, which supports the significance of both the leveragevariablesinexplainingthedpo (F ofChowtestisinsignificant). Hence, h 2 0 isrejectedinrespectofdpo t−1 positivelyat5%level(0 .2350 ), and inv negatively at 1%level(– 1.1126 )inthereducedmodel;withre- specttothefullmodeldpo t−1 positively(0 .2323 )at5%level,pat at10 % level(0 .4732 ),andinv negativelyat1%level(–1.1996 )forSmallsizefirms Volume 8 · Number 1 ·Spring2010 70 AzhagaiahRamachandranandVeeramuthuPackkirisamy table10 ResultsofCross-Sectionalols Modelformediumsizesamplefirmsunder allselectedsectors(meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 36 .8846 1 .64 34 .5672 1 .46 dpo t−1 0 .1634 ** 2.55 0 .1592 ** 2.47 pat –0 .1142*– 1.78 –0 .1196*– 1.67 tde 0 .0582 ** 2.13 0 .0479 1 .22 cf 0 .0418 ** 2.04 0 .0442 ** 2.14 Size –5.2592 –1.20 –4.2581 –0 .82 inv –0 .0660 –1.26 –0 .0688 –1.31 ltd _ta ——– 0 .0664 –0 .66 std _ta —— 0 .1230 0 .73 R 2 0 .0577 0 .0628 Adjusted R 2 0 .0354 0 .0329 F-value 2.58 ** 2.10 ** 6 .253 8 .251 ChowTestF-value 0 .68 2.251 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. under all selected sectors. Therefore, the dp of small size sample firms acrossallselectedsectorsisindependentofthelevelofdebtincs . Theregressionmodelformediumsizefirmsisfittedsignificantly(see table 10 ) with all the control variables, but together they explain to an extent of5.77 percentofthevariationindpo (R 2 =0 .0577 ,F =2.58 ,p< 0 .05 ).Theco-efficientoflaggeddpo (β=0 .1634 ,t =2.55 ,p<0 .05 ),tde (β = 0 .0582 , t = 2.13 , p < 0 .05 )andcf (β = 0 .0418 , t = 2.04 , p < 0 .05 )is significant positively, and that of pat (β=–0 .1142 , t=– 1.78 , p < 0 .10 )is significantnegativelyinthereducedmodel. Inthepresenceofleveragevariables(fullmodel),thestatisticalsignif- icance of tde has disappeared. Therefore, there has been a continuous increase in dpo when there has been an increase in tde and cf ,evenif there is a decline in pat when the status of debt is not taken into con- sideration.However,withmarginaldecreaseinltd andconsiderablein- crease in std , the medium size firms have not considered tde before distributing dividend payout to shareholders. Hence,cs anddp are un- ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 71 table11 ResultsofCross-Sectionalols Modelforlargesizefirmsunderallselected sectors(meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 14 .8011 *** 7.37 18 .9381 *** 5.58 dpo t−1 0 .4380 *** 7.89 0 .4390 *** 7.90 pat ———— tde ———— cf ———— S i z e ———— inv ———— ltd _ta ——– 0 .1238*– 1.78 std _ta ——– 0 .0455 –0 .37 R 2 0 .1741 0 .1830 Adjusted R 2 0 .1713 0 .1747 F-value 62 .21 *** 21 .88 *** 1.295 3 .293 ChowTestF-value 1.60 2.293 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. related with each other in the case of medium size sample firms of all selected sectors. (F of Chow test is insignificant), and therefore reject- ingh 3 0 inrespectofdpo t−1 positivelyat5%level(0 .1634 ),pat negatively at 10%level(– 0 .1142 ), tde positively at 5%level( 0 .0582 ), and cf posi- tively at 5%level( 0 .0418 )inthereducedmodel;withrespecttothefull modeldpo t−1 positively(0 .1592 )at5%level,pat negatively at10 %level (–0 .1196 ),andcf positivelyat5%level(0 .0442 ).Hence,thereisasignif- icant effectofdpo t−1 , pat ,tde ,andcf ondpo inmediumsizesample firmsunderallselectedsectors. The full model regression analysis with step-wise approach for large size sample firms, in which leverage variables as proxy for cs are in- cluded,isfittedsignificantly(seetable11)explainingtotheextentof18 .30 percentofthevariationindpo (R 2 =0 .1830 ,F =21 .88 ,p<0 .01 ). Besides,ltd _ta hasauniquesignificantnegativeimpactondpo (β= –0 .1238 , t=–1.75 , p <0 .10 ). However, there is a lack of collective impact ofbothleveragevariablesondpo (F ofChowtestissignificant).Hence, Volume 8 · Number 1 ·Spring2010 72 AzhagaiahRamachandranandVeeramuthuPackkirisamy table12 ResultsofCross-Sectionalols Modelforallsamplefirmsunderallselected sectors(meanvaluesin%) Independentvariable ReducedModel FullModel Coeff. t-value Coeff. t-value Intercept 15 .3441 *** 11.17 19 .6463 *** 8.00 dpo t−1 0 .3094 *** 7.79 0 .3010 *** 7.56 pat0 .0013 1 .36 0 .0006 0 .61 tde ———— cf ———— S i z e ———— inv ———— ltd _ta ——– 0 .1117 *** –2.43 std _ta ——– 0 .0070 –0 .08 R 2 0 .0889 0 .0972 Adjusted R 2 0 .0861 0 .0917 F-value 31 .91 *** 17.55 *** 2.654 4 .652 ChowTestF-value 3.00 ** 2.652 notes ***Significantat1%level,**Significantat5%level,*Significantat10 %level. h 4 0 is rejected in respect of dpo t−1 positively at 1%level( 0 .4380 )inthe reducedmodel;withrespecttothefullmodeldpo t−1 positively(0 .4390 ) at 1%level,ltd _ta negatively at 10 %level(– 0 .1238 ) for large size firms under all selected sectors. The distribution of dividend payout is inde- pendentofthecs betweendebtandequityofthelargesizesamplefirms acrossallselectedsectors. It is evident from the result of regression analysis for all sample firms pooled together (see table 12 ) that both, the reduced model as well as thefullmodelarefittedsignificantly,explainingtotheextentof8.89 per cent (R 2 = 0 .0889 , F = 31 .91 , p < 0 .01 )and9 .72 (R 2 = 0 .0972 , F = 17.55 , p <0 .01 ) variation indpo respectively. In the reduced model, only lagged dpo andpat areretainedbystep-wiseprocedure.Betweentheestimated co-efficient, itis significant positively only for laggeddpo (β=0 .3094 ,t = 7.79 , p < 0 .01 ). The co-efficient, though positive, is not significant for pat ,indicatingthatthe pat hasauniquenegligibleeffectondpo inthe presence of the previous year’s dpo status. However, in the presence of ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 73 leveragevariables,thoughpresentwithinsignificantco-efficient,thede- greeofuniquerelationshipofpat withdpo hasdecreasedheavily.Atthe sametime,theco-efficientoflaggeddpo issignificantat1 percentlevel (β=0 .3010 ,t =7.56 ,p<0 .01 ).Besides,betweenltd _ta andstd _ta ,the coefficient ofltd _ta is significant negatively (β=–0 .1117 , t=–2.43 , p < 0 .05 )at5 percentlevel.Thedifferenceinexplainedvariancebetweenthe fullmodelandthereducedmodel(differenceinR 2 )isalsosignificantat 5 percentlevel(ChowF =3.00 , p <0 .05 ). Hence, it is found that all the sample firms across all selected sectors keep distributing dividend if it is done so in the previous years, and increase/decrease in dpo is based on the level of debt fund in cs .Hence,H 5 0 is rejected in respect of dpo t−1 positively at 1%level( 0 .3094 )inthereducedmodel;withrespectto the full model dpo t−1 positively (0 .3010 )at1%level,ltd _ta negatively at 1%level(– 0 .1117 ) as there is significant impact of dpo t−1 , ltd _ta on dpo in all sample firms under all selected sectors, therefore dp in terms of dividend payout is significantly influenced by the cs of firms inIndia. ConcludingRemarks This study examines the impact of firm size on the dividend behaviour ofcorporatefirmsinIndia,andhasbeencarriedouton73 firmsbyem- pirically analysing the determinants of dp over a wider testing period from 1996 /1997 to 2006 /2007 . Dividend behaviour was tested using the full-Britainmodelanditsvariantsonthepooledcrosssectional/timese- riesdataforthesampleofobservationsfrom1996 /1997 to2006 /2007 .The modelsareestimatedusingtheOrdinaryLeastSquare(ols )method. Dividend stocks are expected to provide a combination of dividend cash flows and capital gains from the investors’ view. The preference of shareholders for one or the other should have a powerful influence on decisions regarding dividend payment, which leads one to examine the extenttowhichdividendpaymentsanddividendyieldsvarysignificantly acrossfirms,industriesandtime.Firmscomeinvarioussizesandshapes, and they could be single-owner enterprises or large mnc s with many shareholders cutting across geographical boundaries. The management of each firm normally makes dp , but the nature of the ownership can playanimportantroleindp decision. While investors concentrate their attention on dividend yield, man- agement pays more attention to the impact of dividend payouts on the firm’s capital needs. A high dividend pay out reduces firm’s access to re- Volume 8 · Number 1 ·Spring2010 74 AzhagaiahRamachandranandVeeramuthuPackkirisamy tainedearnings,whichisoftenviewedasthelowestcostsourceofcapital. For that reason, management may prefer lower dividend payout ratios, but must recognize therealities imposed byshareholders’ preference for atleastsomepaymentofdividends.Accordingtothetraditionalviewof cs ,whenafirm’sleverageexceedstheoptimumcs ,itscostofcapitalin- creases.Shareholderswillwantmoredividendsandmoremoneywillbe needed to meet interest obligations to debt holders, the weighted aver- age cost of capital will be high and the firm’s liquidity position may be affectedandgrowthretarded. The purpose of this study was to empirically analyse the extent to whichtheperceivedtheoryabouttheconventionaldeterminantsofdiv- idend behaviour of corporate firms explains the dividend behaviour of quoted firms across industries in India with respect to size. The effects of firm size, growth prospect and the level of gearing on dividend be- haviouroffirmshavebeenanalysed.Thedividendbehaviourofcorpo- ratefirmsinIndia’semergingmarketseemstobesignificantlyinfluenced byanumberoffactors,whichsubstantiallydifferfromwhatiscommon indevelopedcountries. Irrespective of the sector, the relationship between the cs and dp re- mains same, i. e in most of the cases the impact of cs measures, viz., ltd _ta , std _ta ,andtd _ta on dp is unique. The hypothesis which formulated that ‘there is no significant relationship between the level of debt in capital structure and level of equity dividend’ has been rejected in almost all the sectors. The inter-correlation matrix among variables in the regression models for various sectors also supports the conclu- sion that there is impact among the independent variables chosen for thestudy.Theresultsofthecross-sectionalols Modelregressionforthe selectedsamplefirmsundervarioussectorsalsoshowthatthereisasig- nificant effect of selected independent variables [dpo =α+β 1 dpo t−1 + β 2 pat +β 3 tde +β 4 cf +β 5 size +β 6 inv +β 7 ltd +β 8 std + e] on the dividendpayout. The study proves that the equity dividend percentage and the debt financingincs areinverselyrelatedtoeachotherinmostofthesectors. Besides, the cs of sample firms significantly influences dividend payout acrossallselectedsectorswhenpooledtogether,butsector-wiseandsize- wise,thereisaninsignificantrelationshipbetweendp andcs .Therefore, it is concluded that the dp of small size, medium size, large size, and overall corporate firms across industries in India is independent of the levelofdebtincs . ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 75 References Aharony, J., and I. Swary. 1980 . Quarterly dividend and earnings an- nouncements and stockholders’ returns: An empirical analysis. Jour- nalofFinance35 (1):1–12 . Ahmed, A., B. K. Billings, R. M. Morton, and M. Stanford-Harris. 2002 . Theroleofaccountingconservatisminmitigatingbondholder-share- holder conflicts over dividend policy and in reducing debt costs. The Accounting Review 77 (4):867 –90 . Akhigbe,A.,S.F.Borde,andJ.Madura.1993 .Dividendpolicyandsignal- ing by insurance companies. The Journal of Risk and Insurance 60 (3): 413 –28 . Allen, D. E. 1991 . The determinants of the capital structure of listed Aus- tralian companies: The financial manager’s perspective. Australian JournalofManagement 16 (2):103 –28 . Anand, M. 2004 . Factors influencing dividend policy decisions of corpo- rateIndia.Theicfai JournalofAppliedFinance10 (2):5–16 . Anil, K., and S. Kapoor. 2008 . Determinants of dividend payout ratios: A study of Indian information technology sector. International Research JournalofFinanceandEconomics15 (may):63 –71 . Baker, M., and J. Wurgler. 2002 . Market timing and capital structure. The Journal of Finance 57 (1):1–32 . Bhattacharya, S. 1979 . Imperfect information, dividend policy, and the ‘birdinthehand’fallacy.BellJournalofEconomics10 (1):259 –70 . Booth,L.,V.Aivazian,A.Demirguc-Kunt,andV.Maksimovic.2001 .Cap- ital structures in developing countries. The Journal of Finance 56 (1): 87 –130 . Brennan,M.J.,andA.V.Thakor.1990 .Shareholderpreferencesanddivi- dendpolicy.TheJournalofFinance45 (4):993 –1018 . Campello, M. 2002 . Internal capital markets in financial conglomerates: Evidence from small bank responses to monetary policy. The Journal ofFinance5 (6 ):2773 –805 . Chunchi,W.,andC.Kao.1992 .Theadjustmentofdividendstopermanent earnings.SouthernEconomicJournal58 (4):1058 –71 . Collins, M. C., A. K. Saxena, and J. W. Wansley. 1996 . The role of insid- ers and dividend policy: A comparison of regulated and unregulated firms.JournalofFinancialandStrategicDecisions9 (2):1–9 . Deangelo, H., and L. Deangelo. 1990 . Dividend policy and financial dis- tress: An empirical investigation of troubled nyse firms. The Journal ofFinance14 (5):1415 –31 . Elston,J.A.1996 . Dividend policy and investment: Theory and evidence fromus paneldata.ManagerialandDecisionEconomics17 (3):267 –75 . Volume 8 · Number 1 ·Spring2010 76 AzhagaiahRamachandranandVeeramuthuPackkirisamy Faulkender, M., T. Milbourn, and A. Thakor. 2006 . Capital structure and dividend policy: Two sides of same coin. Http://apps.olin.wustl.edu /faculty/milbourn/flexibility.pdf. Ghosh, C., and J. R. Woolridge. 1989 . Stock-market reaction to growth- induceddividendcuts:Areinvestorsmyopic?ManagerialandDecision Economic 10 (1):25 –35 . Gordon, M. J. 1959 . Dividends, earnings, and stock prices. The Review of EconomicsandStatistics41 (2):99 –105 . Graham,J.R.,andC.Harvey.2002 .Howdocfo smakecapitalbudgeting and capital structure decisions? Journal of Applied Corporate Finance 15 (1):8–23 . Graham, J. R., M. H. Lang, and D. A. Shackelford. 2004 .Employeestock options,corporatetaxes,anddebtpolicy.TheJournalofFinance59 (4): 1585 –618 . Graham, J. R., and A. L. Tuckerb. 2006 . Tax shelters and corporate debt policy.JournalofFinancialEconomics81 (3):563 –94 . Goldstein, R., N. Ju, and H. Leland. 2001 .Anebit -based model of dy- namiccapitalstructure.TheJournalofBusiness 74 (4):483 –512 . Gulati, D., and Z. Zantout. 1997 . Inflation, capital structure, and immu- nizationofthefirm’sgrowthpotential.JournalofFinancialandStrate- gicDecisions10 (1):77 –90 . Gugler, K. 2003 . Corporate governance, dividend payout policy, and the interrelationbetweendividends, r &d andcapitalinvestment. Journal ofBankingandFinance27 (7):1297 –321 . Hagen, K. P. 1973 . Optimal dividend policies and corporate growth. The SwedishJournalofEconomics75 (3):238 –48 . Hodder,J.E.,andL.W.Senbet.1990 .Internationalcapitalstructureequi- librium.TheJournalofFinance45 (5):1495 –516 . Johnson, S. A. 1995 . Dividend payout and the valuation effects of bond announcements.TheJournalofFinancialandQuantitativeAnalysis30 (3):407 –23 . Kalea, J. R., and H. Shahrurb. 2007 . Corporate capital structure and the characteristics of suppliers and customers. Journal of Financial Eco- nomics83 (2):321 –65 . Kane, A., K. L. Young, and A. Marcus. 1984 . Earnings and dividend an- nouncements: Is there a corroboration effect? The Journal of Finance 39 (4):1091 –99 . Koch,P.D.,andC.Shenoy.1999 .Theinformationcontentofdividendand capitalstructurepolicies.FinancialManagement 28 (4):16 –35 . Kose, J., and J. Williams. 1985 . Dividends, dilution and taxes: A signaling equilibrium.Journal of Finance 40 (4):1053 –70 . ManagingGlobalTransitions TheImpactofFirmSizeonDividendBehaviour 77 Kumar, P., and B. S. Lee. 2001 . Discrete dividend policy with permanent earnings.FinancialManagement 30 (3):55 –76 . La Porta, R., K. Lopez-de-Silanes, A. Shleifer, and R. W. Vishny. 2000 . Agency problems and dividend policies around the world. The Jour- nalofFinance55 (1):1–33 . Lambert, R. A., W. N. Lanen, and D. F. Larcker. 1989 .Executivestockop- tionplansandcorporatedividendpolicy.TheJournalofFinancialand QuantitativeAnalysis24 (4):409 –25 . Lim,K.G.1989 .Dividendpolicyandtaxstructure.EconomicsLetter31 (3): 269 –72 . Mao, C. X.2003 .Interactionofdebtagencyproblemsandoptimalcapital structure: Theory and evidence. The Journal of Financial and Quanti- tativeAnalysis38 (2):399 –423 . Michel,A.1979 .Industryinfluenceondividendpolicy.FinancialManage- ment 8 (3):22 –6 . Mihir, A., C. Desai, F. Foley, and J. R. Hines, Jr. 2004 . A multinational perspective on capital structure choice and internal capital markets. TheJournalofFinance59 (6 ):2451 –87 . Miller, M. H., and K. Rock. 1985 . Dividend policy under asymmetric in- formation.TheJournalofFinance40 (4):1031 –51 . Mohanty, P. 1999 . Dividend and bonus policies of Indian companies: An analysis.Vikalpa24 (4):35 –42 . Nishioka, S., and N. Baba. 2004 . Dynamic capital structure of Japanese firms: How far has the reduction of excess leverage progressed in Japan?BankofJapanWorkingPaperSeries04 -e -16 . Ooi,T.L.J.2001 .Dividendpayoutcharacteristicsofuk property compa- nies.JournalofRealEstatePortfolioManagement 7 (2):133 –42 . Papaioannou, G. J., and C. M. Savarese. 1994 . Corporate dividend policy responsetotheTaxReformActof1986 .FinancialManagement 23 (1): 56 –63 . Romano, C. A., G. A. Tanewski, and K. X. Smyrnios. 2000 . Capital struc- turedecisionmaking:Amodelforfamilybusiness.JournalofBusiness Venturing 16 (3):285 –310 . Rozeff,M.S.1982 .Growth,betaandagencycostsasdeterminantsofdivi- dendpayoutratios.JournalofFinancialEconomics5 (3):249 –59 . Singhania,M.2006 .Taxationandcorporatepayoutpolicy. Vikalpa 31 (4): 47 –64 . Sharma,D.2006 .Corporatedividendtrends:AnempiricalstudyofSensex companies.IndianJournalofAccounting 37 (1):14 –21 . ———.2007 .AredividendsinvogueinIndia?AnempiricalstudyofSen- sexcompanies.TheIcfaianJournalofManagementResearch6 (3):22 –9 . Stenbacka, R., and M. Tombak. 2002 . Investment, capital structure, and Volume 8 · Number 1 ·Spring2010 78 AzhagaiahRamachandranandVeeramuthuPackkirisamy complementaritiesbetweendebtandnewequity.ManagementScience 48 (2):257 –72 . Wilson,D.R.1967 .APareto-optimaldividendpolicy.ManagementScience 13 (9 ):756 –64 . Woolridge,R.J.1983 .Dividendchangesandsecurityprices.TheJournalof Finance38 (5):1607 –15 . Yener, D. 1991 .CorporatecapitalstructureandtaxesintheKoreanmanu- facturingsector.GlobalFinanceJournal2 (3–4):207 –19 . ManagingGlobalTransitions