Economic Interferences

Transcription

1 Economic Inerferences Zélia Serrasqueiro Managemen and Economics Deparmen, Beira Inerior Universiy, Covilhã, Porugal and CEFAGE Research Cener Évora Universiy, Porugal Absrac In his aricle, using dynamic panel esimaors, we es empirically he relaionship beween he growh of Poruguese companies and heir profiabiliy. The empirical evidence obained indicaes ha growh in Poruguese companies means increased profiabiliy. Growh in Poruguese companies is a caalysing facor of profiabiliy, wih he moivaional effec on employees, given he expecaion of greaer gains in he fuure, being more relevan han he possible negaive effecs of growh on profiabiliy, as a consequence of he need for new more formal labour relaions in companies. Keywords: company growh, dynamic esimaors, profiabiliy. Jel Classificaions: D21, G32, L25 Inroducion According o Greiner (1972), he relaionship beween company growh and profiabiliy can be posiive or negaive. On he one hand, increased growh can conribue o a breakdown of informal relaionships esablished over ime in companies, greaer growh requiring greaer formaliy in relaionships a work, which in he shor-erm can be difficul o achieve efficienly, his leading o diminished company profiabiliy. On he oher hand, greaer growh can resul in greaer profiabiliy, as a resul of increased moivaion among employees who expec greaer gains in he fuure, gains resuling from greaer company size. Greiner (1972) concludes ha he effec of company growh on profiabiliy will above all be dependen on he owners abiliy o moivae employees. If he posiive effec of employee moivaion has greaer magniude han he negaive effec resuling from he change in he usual working relaionships, growh can mean increased company profiabiliy. Oherwise, growh can mean reduced profiabiliy. Empirical evidence concerning he relaionship beween growh and profiabiliy is scarce and does no clarify he relaionship beween he variables. Roper (1999), in he conex of Irish companies, and Gschwandner (2005), for American companies, find saisically insignifican relaionships beween growh and profiabiliy. Based on a sample of Poruguese companies, his aricle conribues o he lieraure showing evidence of a posiive and saisically significan relaionship beween growh and profiabiliy. Therefore, we can conclude ha he employee moivaion effec, a Vol XI Nr. 26 June

2 Growh and Profiabiliy in Poruguese Companies: a Dynamic Panel Daa Approach consequence of he possibiliy for greaer gains in he fuure, can be relevan for companies, so as o overcome he negaive effecs arising from emporary inefficiency of he new more formal labour relaions. Mehodologically, o esimae he relaionship beween growh and profiabiliy in Poruguese companies, we use dynamic panel esimaors, namely he dynamic esimaors: 1) GMM (1991); GMM sysem (1998) and LSDVC (2005). Iniially, we only esimae he relaionship beween growh and profiabiliy. Aferwards, we add oher possible explanaory variables of profiabiliy, wih he simulaneous objecives of esing robusness of he relaionship obained beween growh and profiabiliy and enhancing he analysis made in his sudy. As oher explanaory variables of profiabiliy, we consider: 1) size (Adams and Buckle; 2003; Goddard e al.; 2005); 2) deb (Adams and Buckle, 2003; Goddard e al.; 2005); and 3) liquidiy (Adams and Buckle, 2003; Goddard e al., 2005). Apar from wha was saed above, using dynamic panel esimaors allows us o deermine he level of persisence of profiabiliy, ha is o say, i les us esimae he relaionship beween profiabiliy in he previous period and profiabiliy in he curren period. In order o aain he objecives of his sudy, we divide i as follows, afer his inroducion: 1) secion 2 presens he daabase, he variables and he esimaion mehod used; 2) secion 3 presens he resuls which we go on o discuss; and 4) finally, secion 4, presens he conclusion of his sudy. 1. Daabase, Variables and Esimaion Mehod 1.1. Daabase This sudy uses he daabase from he Exame journal, he Poruguese branch of Dun & Bradsree Consulans, concerning he 500 larges Poruguese companies. The companies forming he daabase are seleced annually based on sales volume. The period of sudy is he years beween 1999 and Given he use of dynamic esimaors as esimaion mehod, we oped for a uniform panel. Choosing a non-uniform panel could lead o he exclusion of companies ha were no presen for all he years under analysis, prevening correc inerpreaion of effecs over ime. Wih he aim of considering a uniform panel, we selec companies ha belong o he daabase for he oal number of years of analysis ( ). Based on his crierion, he panel is made up of 162 companies, wih a oal of 810 observaions Variables Our cenral aim is o sudy he relaionship beween growh and profiabiliy in Poruguese companies. In addiion, wih he aims of checking he robusness of he relaionship obained beween growh and profiabiliy, and simulaneously enhancing he analysis made, we add oher possible deerminans of he profiabiliy of Poruguese companies, which serve as conrol variables. This sudy considers as conrol variables: 1) size (Adams and Buckle, 2003; Goddard e al., 2005); 2) deb (Adams and Buckle, 2003; Goddard e al., 2005); and 3) liquidiy (Adams and Buckle, 2003; Goddard e al.). The variables used, ogeher wih heir corresponding measures are presened in he following able. 566 Amfiearu Economic

3 Economic Inerferences Variables Dependen variables Profiabiliy ( PROF ) Independen variables Growh ( GROW ) Size ( SIZE ) Leverage ( LEV ) Liquidiy ( LIQ ) Measuremen of variables Table 1 Measuremen Raio Beween Operaional Resuls and Toal Asses Growh of Sales Logarihm of Sales Raio Beween Toal Liabiliies and Toal Asses Raio Beween Curren Asses and Shor Term Deb 1.3. Esimaion Mehod The advanages of using dynamic esimaors menioned by Arellano and Bond (1991) are: 1) effecive conrol of endogeny; 2) greaer conrol of possible collineariy beween he independen variables; 3) conrol of he effecs of possible omission of independen variables in explaining he dependen variable; and 4) eliminaion of non-observable individual effecs. Use of dynamic esimaors allows correc esimaion of he relaionship beween he dependen variable in he previous and curren periods. Therefore, in his sudy we op o use dynamic panel esimaors which, besides he advanages menioned above by Arellano and Bond (1991), allow correc esimaion of he relaionship beween profiabiliy in he previous period and profiabiliy in he curren period, i.e. hey allow us o es for he possible persisence of profiabiliy in Poruguese companies, a procedure also used by Goddard e al. (2005), in he conex of Belgian, Spanish, French, German and Briish firms. Iniially, we es he relaionship beween growh and profiabiliy in Poruguese companies (Regression I). Aferwards, we add he conrol variables used in his sudy (Regression II), in order o es robusness of he previously idenified relaionship beween growh and profiabiliy, and simulaneously enhance he analysis carried ou. Therefore, we es he following regressions. Regression I PROF i, = 0 + δprofi, 1 + β1growi, + d + ei, β ; (1) Regression II PROF i, = β0 + δprofi, 1 + β1growi, + β2sizei, + β3levi, + β4liqi, + d +, (2) in which: previous period, PROF is profiabiliy in he curren period, i, 1 GROW is company growh in he curren period, PROF is profiabiliy in he e SIZE is size in he i, Vol XI Nr. 26 June

4 Growh and Profiabiliy in Poruguese Companies: a Dynamic Panel Daa Approach curren period, LEV is deb in he curren period, LIQ is liquidiy in he curren period, d are annual dummy variables measuring he effecs of possible macroeconomic aleraions on profiabiliy and e is he random error which is assumed o have normal disribuion. Arellano and Bond (1991) propose esimaion of equaions (1) and (2) in firs differences, and use of lags of he dependen and independen variables, a levels, as insrumens. This esimaor became known as GMM (1991). However, in cases where we find persisence of he dependen variable, wih high correlaion beween he dependen variable in he curren and previous periods, and he number of cross-secions is no paricularly high, Blundell and Bond (1998) conclude ha he GMM (1991) esimaor may no be very efficien, given ha he insrumens may no be valid. To solve his problem, Blundell and Bond (1998) propose a new esimaor, considering a sysem of variables a levels and in firs differences. The dynamic esimaor proposed by Blundell and Bond (1998), became known as GMM sysem (1998). In he equaion o esimae, he variables a levels are used as insrumens in firs differences, and on he conrary when he variables appear ransformed in firs differences in he equaion o esimae, hey are used a levels as insrumens. Neverheless, he resuls obained wih he GMM (1991) and GMM sysem (1998) dynamic esimaors can only be considered robus on wo condiions: 1) validiy of he insrumens used; 2) non-exisence of second order auocorrelaion. Wih he aim of esing validiy of he resricions, we use he Sargan es, in he case of he GMM (1991) dynamic esimaor, and he Hansen es, in he case of he GMM sysem (1998) dynamic esimaor. The null hypohesis is ha he insrumens are valid, he alernaive hypohesis being non-validiy of he insrumens used. If we rejec he null hypohesis, we can conclude ha he resuls obained are no robus. This sudy ess for he exisence of firs and second order auocorrelaion. The null hypohesis is non-exisence of auocorrelaion, agains he alernaive hypohesis of exisence of auocorrelaion. In he case of rejecion of he null hypohesis of non-exisence of second order auocorrelaion, we conclude ha he resuls of he dynamic esimaors canno be considered robus. Taking advanage of quie recen progress concerning dynamic esimaors, we presen he resuls obained wih he LSDVC (Leas Squares Dummy Variable Correced) dynamic esimaor, proposed by Bruno (2005). Bruno (2005) concludes ha in cases where he number of observaions is no very high, using GMM (1991) and GMM sysem (1998) dynamic esimaors, given he high number of insrumens generaed compared o he number of observaions may lead o biased esimaes of he parameers. Due o he raher low number of observaions in his sudy, we presen he resuls of he LSDVC (2005) esimaor, so as o es robusness of he resuls obained wih he GMM (1991) and GMM sysem (1998) dynamic esimaors. 568 Amfiearu Economic

5 Economic Inerferences 2. Resuls and Discussion 2.1. Descripive Saisics The following able presens he resuls of he descripive saisics referring o he variables used in his sudy. Descripive Saisics Table 2 VAR. OBSER. MEAN S.D. MIN. MAX. PROF, i GROW SIZE LEV LIQ We can see ha profiabiliy and growh in Poruguese companies presen some volailiy, since sandard deviaions of he variables are above he respecive means. Regarding he remaining variables used in his sudy: 1) size; 2) deb; and 3) liquidiy, volailiy is no pronounced, since sandard deviaions are under he means Dynamic Esimaors Table 3 presens resuls referring o he relaionship beween growh and profiabiliy in Poruguese companies (Regression I), wihou adding conrol variables. In able 4, we add o he analysis he conrol variables used in his sudy (Regression II). Dynamic Esimaors - Regression I Table 3 Dependen Variable: PROF, Independen Variables GMM (1991) GMM Sysem (1998) LSDVC (2005) Iniial (AB) LSDVC (2005) Iniial (BB) PROF *** *** *** *** i, 1 ( ) ( ) ( ) ( ) GROW *** *** *** *** ( ) ( ) ( ) ( ) CONS *** ( ) ( ) Insrumens GMM GMM sysem Observaions Sargan (χ 2 ) 2.23 Hansen (N(0,1)) m1 (N(0,1)) -5.60*** -3.98*** m2 (N(0,1)) Noes: 1 Heeroskedasiciy consisen and asympoic robus sandard deviaions are repored in brackes. 2. Time dummy is included in esimaed equaions, bu no shown. 3. *** indicaes significance a 1% level, ** indicaes significance a 5% level, and * indicaes significance a 10% level. i Vol XI Nr. 26 June

6 Growh and Profiabiliy in Poruguese Companies: a Dynamic Panel Daa Approach Dynamic Esimaors - Regression II Dependen Variable: PROF Independen Variables GMM (1991) GMM Sysem (1998) LSDVC (2005) Iniial (AB) PROF *** *** *** i, 1 ( ) ( ) ( ) GROW *** *** *** ( ) ( ) ( ) SIZE ** *** *** ( ) ( ) ( ) LEV *** *** *** ( ) ( ) ( ) LIQ ( ) ( ) ( ) CONS * ** ( ) ( ) Insrumens GMM GMM sysem Observaions Sargan (χ 2 ) 5.02 Hansen Table 4 LSDVC (2005) Iniial (BB) *** ( ) *** ( ) *** ( ) *** ( ) ( ) (N(0,1)) m1 (N(0,1)) -3.11*** -3.71*** m2 (N(0,1)) Noes: 1 Heeroskedasiciy consisen and asympoic robus sandard deviaions are repored in brackes. 2. Time dummy is included in esimaed equaions, bu no shown. 3. *** indicaes significance a 1% level, ** indicaes significance a 5% level, and * indicaes significance a 10% level. Firsly, wheher adding conrol variables (Regression II), or no (Regression I), he resuls of he Sargan and Hansen ess indicae we canno rejec he null hypohesis of validiy of he insrumens used, wih he GMM (1991) and GMM sysem (1998) dynamic esimaors respecively. Secondly, in no circumsances do we rejec he null hypohesis of absence of second order auocorrelaion. Based on he resuls of he Sargan/Hansen and second order auocorrelaion ess, wheher adding conrol variables (Regression II), or no (Regression I), we can consider he resuls obained wih he GMM (1991) and GMM sysem (1998) dynamic esimaors o be robus. The resuls obained wih he LSDVC (2005) esimaor, wheher considering iniial correcion of he values obained wih he GMM (1991) esimaor or he GMM sysem (1998) esimaor and adding (Regression II), or no (Regression I), conrol variables, corroborae hose obained when using he GMM (1991) and GMM sysem (1998) dynamic esimaors. Based on he empirical evidence obained in his sudy, we can conclude ha: wheher adding conrol variables (Regression II) or no (Regression I), he relaionship beween growh and profiabiliy in Poruguese companies is posiive and saisically significan; 570 Amfiearu Economic

7 Economic Inerferences wheher adding conrol variables (Regression II) or no (Regression I), he relaionship beween profiabiliy in he previous and curren periods is posiive and saisically significan, and so profiabiliy of Poruguese companies is persisen; he relaionship beween size and profiabiliy in Poruguese companies is posiive and saisically significan; he relaionship beween deb and profiabiliy in Poruguese companies is negaive and saisically significan; he relaionship beween liquidiy and profiabiliy in Poruguese companies is no saisically significan Discussion of he Resuls Based on he empirical resuls obained in his sudy, we can conclude ha greaer growh of Poruguese companies means increased profiabiliy. Therefore, in he conex of Poruguese companies, we can conclude ha he posiive effecs of growh on profiabiliy are greaer han he negaive effecs. Growh in Poruguese companies is a caalysing facor of profiabiliy, he employee moivaion effec, due o expeced fuure gains, being more relevan for profiabiliy han he possible negaive effecs of growh on profiabiliy, arising from he need for new more formal labour relaions in companies. The resuls obained show ha owners of Poruguese firms will be able o moivae employees appropriaely, so ha he posiive effecs of growh on profiabiliy are more relevan han he negaive effecs, corroboraing he conclusions of Greiner (1972). The posiive relaionship deeced beween growh and profiabiliy in Poruguese companies does no corroborae he empirical evidence of previous sudies, since Roper (1999), in he conex of Irish companies, and Gschwandner (2005), for American companies, find saisically insignifican relaionships beween company growh and profiabiliy. While in he conex of Irish and American companies, he posiive and negaive effecs of growh on profiabiliy seem o be equally relevan, his does no happen wih Poruguese companies where he posiive effecs of growh on profiabiliy appear o be clearly more relevan han he negaive effecs. Concerning he oher empirical evidence obained in his sudy, we find ha profiabiliy is persisen. The values obained for persisence of profiabiliy vary beween , esimaing regression II wih he LSDVC (2005) dynamic esimaor correcing he values obained wih he GMM (1991) dynamic esimaor, and , esimaing regression I wih he GMM sysem (1998) dynamic esimaor. Goddard e al. (2005), obain profiabiliy persisence of for Belgian companies, for French companies, for Ialian companies, for Spanish companies, and for Briish companies. Persisence of profiabiliy in Poruguese companies is generally above ha of companies in oher European counries, his resul indicaing ha disurbance arising from companies enering and leaving he marke is less relevan in he conex of Poruguese companies. We find a posiive and saisically significan relaionship beween size and profiabiliy in Poruguese companies. Adams and Buckle (2003), in he conex of companies in Bermuda, Vol XI Nr. 26 June

8 Growh and Profiabiliy in Poruguese Companies: a Dynamic Panel Daa Approach find a saisically insignifican relaionship beween size and profiabiliy, and Goddard e al. (2005), obain negaive relaionships beween size and profiabiliy in Belgian, French, Ialian, Spanish and Briish companies. The greaer possibiliy of aking advanage of economies of scale and diversificaion of aciviies and producs, as a consequence of greaer size, seems o be more relevan for increased profiabiliy in Poruguese companies han for companies in Bermuda, Belgium, France, Ialy, Spain and he Unied Kingdom. In hese counries, here seems o be more relevance of reduced owner conrol of managers acions, as a consequence of greaer size, compared o wha happens in he Poruguese business siuaion. The relaionship beween deb and profiabiliy in Poruguese companies is negaive and saisically significan, unlike he case for companies in Bermuda, for which Adams and Buckle (2003) find a posiive and saisically significan relaionship beween deb and profiabiliy, bu in agreemen wih ha found for Belgian, French, Ialian, Spanish and Briish companies, for which Goddard e al. (2005) idenify a negaive and saisically significan relaionship beween deb and profiabiliy. Jus as in he conex of companies in oher European counries, recourse o deb seems o be a resricive facor of profiabiliy in Poruguese companies, possibly due o he need o make periodic paymens of he deb charges diminishing he capaciy for aking advanage of good opporuniies for company growh, opporuniies which could mean increased profiabiliy. The relaionship beween liquidiy and profiabiliy in Poruguese companies is no saisically significan. Adams and Buckle (2003) obain a negaive and saisically significan relaionship beween liquidiy and profiabiliy for firms in Bermuda, while Goddard e al. (2005), in he conex of Belgian, French, Ialian, Spanish and Briish companies, find posiive relaionships beween liquidiy and profiabiliy. On one hand, he possible caalysing effec of liquidiy on profiabiliy, as a consequence of he greaer possibiliy of meeing shor-erm commimens, seems no o be sufficienly relevan in he case of Poruguese companies for greaer liquidiy o mean increased profiabiliy. On he oher hand, he possible resricive effec of liquidiy on profiabiliy, as a consequence of managers invesing in unprofiable projecs also seems insufficienly relevan for greaer liquidiy in Poruguese companies o mean diminished profiabiliy. Conclusion In his sudy, using various dynamic panel esimaors, we deermine he relaionship beween growh and profiabiliy in Poruguese companies. Irrespecive of he dynamic esimaor used, and adding oher possible deerminans of profiabiliy or no, he empirical evidence obained indicaes ha he relaionship beween Poruguese company growh and profiabiliy is posiive and saisically significan. In he conex of Poruguese companies, employees expecaions of greaer fuure gains, as a consequence of greaer growh, seem o be paricularly relevan for increased profiabiliy. The moivaion ransmied by he owners of Poruguese companies in heir ineracion wih employees could also conribue o his sae of affairs. I is worh noing paricularly ha he possible negaive effecs of company growh on profiabiliy, as a consequence of he need o esablish new formal relaions in companies, seem no o be sufficienly relevan for 572 Amfiearu Economic

9 Economic Inerferences Poruguese company growh o mean diminished profiabiliy, unlike wha appears o happen in companies in oher European counries. Concerning he oher empirical evidence obained in his sudy, we can conclude ha: 1) here is considerable persisence of profiabiliy in Poruguese companies; 2) greaer size of Poruguese companies means increased profiabiliy; 3) greaer recourse o deb by Poruguese companies means diminished profiabiliy; and 4) liquidiy appears no o be relevan in explaining he profiabiliy of Poruguese companies. References 1. Adams, M. and Buckle, M., (2003), The Deerminans of Corporae Financial Performance in he Bermuda Insurance Marke, Applied Financial Economics, 13, pp Arellano, M. and Bond, S., (1991), Some Tess of Specificaion for Panel Daa: Mone Carlo Evidence and an Applicaion o Employmen Equaions, Review of Economic Sudies, 58, pp Blundell, M. and Bond, S., (1998), Iniial Condiions and Momen Resricions in Dynamic Panel Daa Models, Journal of Economerics, 87, pp Bruno, G., (2005), Approximaing he Bias of LSDV Esimaor for Dynamic Unbalanced Panel Daa Models, Economic Leers, 87, pp Goddard, J, Tavakoli, M. and Wilson, J., (2005), Deerminans of Profiabiliy in European Manufacuring and Services: Evidence From a Dynamic Panel Daa, Applied Financial Economics, 15, pp Greiner, L., (1972), Evoluions and Revoluions as Organizaions Grow, Harvard Business Review, 50, pp Gschwandner, A., (2005), Profi Persisence in he Very` Long Run: Evidence From Survivors and Exiers, Applied Economics, 37, pp Roper, S., (1999), Modelling Small Business Growh and Profiabiliy, Small Business Economics, 13, pp Vol XI Nr. 26 June