Volume 11 Issue 2 Article 2 12-31-2009 Planned growth as a determinant of the markup: the case of Slovenian manufacturing Nina Ponikvar Maks Tajnikar Follow this and additional works at: https://www.ebrjournal.net/home Recommended Citation Ponikvar, N., & Tajnikar, M. (2009). Planned growth as a determinant of the markup: the case of Slovenian manufacturing. Economic and Business Review, 11(2). https://doi.org/10.15458/2335-4216.1263 This Original Article is brought to you for free and open access by Economic and Business Review. It has been accepted for inclusion in Economic and Business Review by an authorized editor of Economic and Business Review. 119 ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 2 | 2009 | 119–135 PLANNED GROWTH AS A DETERMINANT OF THE MARKUP: THE CASE OF SLOVENIAN MANUFACTURING NINA PONIKVAR* MAKS TAJNIKAR** ABSTRACT: Th e paper follows the idea of heterodox economists that a cost-plus price is above all a reproductive price and growth price. Th e authors apply a fi rm-level model of markup determination which, in line with theory and empirical evidence, contains pro- posed fi rm-specifi c determinants of the markup, including the fi rm’s planned growth. Th e positive fi rm-level relationship between growth and markup that is found in data for Slov- enian manufacturing fi rms implies that retained profi ts gathered via the markup are an important source of growth fi nancing and that the investment decisions of Slovenian man- ufacturing fi rms aff ect their pricing policy and decisions on the markup size as proposed by Post-Keynesian theory. Th e authors thus conclude that at least a partial trade-off between a fi rm’s growth and competitive outcome exists in Slovenian manufacturing. Key words: Manufacturing; Markup; Firm’s growth; Slovenia UDC: 338.3:330.35(497.4) JEL classification: B50; C23; D21; L21; L6 1. INTRODUCTION It is generally acknowledged by theory and empirical facts that the formation of price in manufacturing fi rms is largely achieved by adding the markup to some sort of average unit cost. Although a lively discussion about the determinants of the markup has been underway for several decades, various authors still list and investigate quite mixed fac- tors infl uencing the size of the markup. Th e reason behind these diff erences in opinion is clearly the complexity of the pricing decision-making process regarding the markup size. A concern about the factors causing the markup to diff er across countries, indus- tries and fi rms can be found in neoclassical economic theory (Lerner, 1934; Oliver, 1947), in a more applied and empirically-based branch of mainstream economics, namely in- * University of Ljubljana, Faculty of Economics, Kardeljeva pl. 17, 1000 Ljubljana, Slovenia, Email: nina. ponikvar@ef.uni-lj.si ** University of Ljubljana, Faculty of Economics, Kardeljeva pl. 17, 1000 Ljubljana, Slovenia, Email: maks. tajnikar@ef.uni-lj.si ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 2 | 2009120 dustrial economics (Bain, 1956; Schmalensee, 1989; Martin, 2001), within the strategic management school (Porter, 1980; Barney, 1991) and especially within Post-Keynesian price theory and its pricing hypotheses (Hall and Hitch, 1939; Kalecki, 1954 and An- drews, 1955). Th e paper follows the idea of heterodox economists, especially Kalecki (1954) and Eich- ner (1973), who argue that a cost-plus price is above all a reproductive price and growth price and accordingly tests the hypothesis that the fi rm’s investment plans impact on its pricing decisions and the size of the markup. Th e study thus focuses on the fi rm’s planned growth as one of the fi rm-specifi c markup determinants. It is hypothesised that fi rms with larger growth ambitions incorporate higher markups into the prices of their products compared to their rivals. In order to test this hypothesised relationship, the authors apply a fi rm-level model of markup determination which in line with theory and empirical evidence contains proposed fi rm-specifi c markup determinants, includ- ing the fi rm’s planned growth. Th e model also controls for industry membership and environmental factors and is based on data on Slovenian manufacturing fi rms for an 11-year period. 2. LITERATURE REVIEW Th e factors determining the markup size can be classifi ed in three general groups. Th e fi rst group of determinants includes the characteristics of the fi rm, usually called fi rm- specifi c factors. Th ese factors are connected to the fi rm’s market power (Kalecki, 1954; Demsetz, 1973), its cost effi ciency and/or the productivity of its production factors and to the technological characteristics of the fi rm’s production process and which are chiefl y a result of strategies accepted and pursued by the fi rm in order to achieve its goal, i.e. long- run profi t maximisation and growth (Eichner, 1973). Industry-specifi c factors represent the characteristics of a particular industry with regard to the concentration of fi rms, entry barriers, product diff erentiation, technological characteristics of the industry’s produc- tion and the demand dynamics (Kalecki, 1954; Schmalensee, 1989). Industry-level factors determine the average power that fi rms within a particular industry exert over the price and the markup of their products. Consequently, these factors determine the average in- dustry markup, while fi rm-level factors determine the deviations of a fi rm’s markups from the industry average. Environmental and institutional factors represent the third group of markup determinants and consist of governmental anti-trust policy, the role of workers’ and employers’ organisations as well as general economic trends (Motta, 2004; Konings et al., 2001). Th e environmental factors are time-specifi c since they infl uence all fi rms in a particular economy in a similar fashion. While for the most part the industry and envi- ronmental characteristics set limits on the markup size, the internal, fi rm-specifi c factors mainly comprise the fi rm’s activities aimed at realisation of the business plan and achiev- ing the fi rm’s goal and as such determine the required markup (Shapiro, 1981). For Kalecki (1954), in cost-determined oligopolistic markets prices are set at the fi rm level with reference to average costs and the prices of other fi rms producing similar prod- NINA PONIKVAR, MAKS TAJNIKAR | PLANNED GROWTH AS A DETERMINANT OF THE MARKUP: ... 121 ucts. Th us, the fi rm’s price p is p = mu + np, where u are average prime costs, p is the aver- age price charged by all fi rms in the industry and m and n are parameters characterising the price-fi xing policy of the fi rm refl ected in the ‘degree of monopoly’1. Accordingly, prices are expected to vary directly with the level of average direct cost but to be con- strained by the price level in the industry, namely the competing group of fi rms (Kalecki, 1954, p. 13). Th e fi rm must ensure that the price does not become too high in relation to the prices of other fi rms, for this would drastically reduce sales, and that the price does not become too low in relation to its average prime cost, for this would drastically reduce its profi t margin. Th e average price and average degree of monopoly of the industry was defi ned by Kalecki (1954, p. 16) with the equation p = u(m/(1 — n)), where p is the average price of the industry, u are average unit prime costs, m and n are weighted averages of the coeffi cients m and n and the expression m/(1 — n) represents the degree of monopoly of the industry. Th e fi rm’s decision on the price and markup is thus subject to various fac- tors deriving from the fi rm itself or from the fi rm’s environment. Th e idea of a strong linkage between the markup and investment fi nance has mostly been developed in theoretical works of non-neoclassical economists, especially within the Post-Keynesian school such as Eichner (1973, 1976), Eichner and Kregel (1975), Harcourt and Kenyon (1976), Shapiro (1981), Wood (1975) etc. Th ese authors present a variant of a model of a price-setting fi rm facing relatively stable marginal costs, assuming that the fi rm’s main objective is its growth and thus the preservation and/or improvement of its market position. Th eir models show that investment decisions infl uence fi rms’ pricing decisions, more specifi cally; they at least partially determine the markup size. Price and the markup size are thus determined by the fi rm, not solely by current demand but also by expected future demand and investment requirements (Kalecki, 1971). Th e latter also derives from the empirically confi rmed fact that fi rms gather a large part of the funds they need for investing from their retained profi ts in both developing and developed countries, although institutional and historical factors must be taken into ac- count to explain some of the variation seen across countries (see Hubbard (1998) for a re- view of empirical studies). Athey and Lumas (1994) along with Athey and Reeser (2000) in their empirical work using panel data from developed countries fi nd that the availabil- ity of internal funds is an important determinant of fi rms’ capital spending. Similarly, the amount of corporate investment is aff ected by internal resources in OECD countries (Kadapakkan et al., 1998). For the USA, for example, Carpenter and Petersen (2002) test a panel of small fi rms and Worthington (1995) a panel of manufacturing fi rms. Th ey fi nd that the growth of most of these fi rms is constrained by internal fi nance and that the cash fl ow and investment spending are positively correlated. Similarly, evidence of the existence of a liquidity constraint on investment in the Dutch manufacturing sector can be found in Van Ees et al. (1997). When comparing the dependence of fi rms on inter- nal fi nance, Bond et al. (2003) report that there is less dependence on internal fi nancial sources in the countries of continental Europe, while the external fi nancial constraints on investment are more serious in the more market-oriented UK fi nancial system. 1 Th e fi rst important use of the concept of the degree of monopoly was made by Lerner (1934) as a measure of the welfare loss of a monopoly and not as a measure of market imperfections. ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 2 | 2009122 For transition economies, Konings et al. (2005) confi rm that fi rms’ investment levels in these countries are sensitive to internal fi nance, although the sensitivity is not equally strong in all the investigated countries. Th ey ascertain that the strength of sensitivity to internal funds depends on the strength of the persistence of soft budget constraints. Sim- ilar results are obtained for Slovenia for which empirical studies that are based on aggre- gate or sectoral data confi rm that fi nancial factors, the availability of internal fi nancial funds as well as the accessibility of external fi nancial sources signifi cantly impact on the business sector’s investment decisions (Tajnikar and Ogrin, 2001). Based on extensive survey data for Slovenian fi rms, Bartlett and Bukvič (2001) similarly identify external fi nancial constraints, including the high cost of capital, as one of the biggest obstacles to the growth of small- and medium-sized fi rms in Slovenia. Accordingly, funds for growth fi nancing have to be gathered internally. Financial factors are hence closely related to investment decisions in which internally generated earnings play a crucial role in investment fi nancing and where external fi nan- cial funds are a complementary fi nancial source, not a substitute for internal fi nancial sources. Because the markup size determines the profi ts and is thus a source of retained internal fi nancial funds, the pricing decisions of the fi rm and the size of its markup also depend on the fi rm’s aspirations for investing and growth (Harcourt and Kenyon, 1976). Th e fi rm’s growth ambitions, and the factors expressing it, are therefore important fi rm- specifi c markup determinants. 3. DATA AND METHODOLOGY 3.1 Data Th e primary data source for the empirical investigation of markup determinants is the database of fi rms’ fi nancial statements collected by the Agency of the Republic of Slov- enia for Public Legal Records and Related Services, which covers the whole population of Slovenian manufacturing fi rms and is extended with some of the internal databases of the Statistical Offi ce of the Republic of Slovenia. Th e database employed in our analysis contains 35,371 observations for 6,987 manufacturing fi rms for the 1994-2004 period2. However, only 20,466 observations on 4,470 fi rms are without missing values and are thus regarded as our sample. A fi rm’s industry membership is defi ned according to the fi ve-digit NACE classifi cation of industries and all fi nancial data are in fi xed prices from the year 2000 in Slovenian tolars. Th e panel nature of the fi rm-level data allows us to combine inter-temporal (within units) as well as inter-fi rm (between-unit) information effi ciently and to control for unobservable fi rm-specifi c variables by focusing on diff er- ences over time (Schmalensee, 1989) and to effi ciently overcome the problems. In addi- tion, it enables us to test the time persistence of the markup and to study the variability of markups over time. 2 Firms with missing values and with the highest and lowest 5 percent of markup values were excluded from the analysis using the method of removing excessive outliers from the dataset introduced by Hadi (1992) since the excessive outliers could have biased the subsequent results and conclusions. NINA PONIKVAR, MAKS TAJNIKAR | PLANNED GROWTH AS A DETERMINANT OF THE MARKUP: ... 123 3.2 Model and description of the variables In line with the various theoretical approaches there is a theoretical disagreement about the list of markup determinants and especially the relative importance of these deter- minants. Yet, irrespective of the theoretical foundations underpinning the empiricism, the list of markup factors becomes very similar in the case of empirical investigations (Porter, 1981). Accordingly, the markup of fi rm i from industry j in year t is determined by general economic trends and the economic environment γt, industry-specifi c factors ηjt and fi rm-specifi c factors εijt. We can thus formulate the most general model of markup determination as: (1) where subscript i refers to a fi rm, j to industries according to the fi ve-digit NACE clas- sifi cation of industries and t to a particular year, respectively. Th us, the markup of fi rm i operating in industry j in year t is modelled as a function of fi rm i’s contemporaneous characteristics, industry j’s contemporaneous characteristics and the characteristics of the economic environment in year t (X’it, X’jt and X’t respectively) with unknown weights β, γ and θ and a lagged dependent variable with an unknown weight δ. yit = σyit-1 + X'itβ + X'jtγ + X't θ + uit i = 1,…,N; j = 1,…,J; t = 1,...,T (2) where yit is the markup for fi rm i in time period t, δ is a scalar, X’ it, X’jt and X’t are 1 x K vectors of explanatory variables with unknown K x 1 coeffi cient vectors β, γ and θ. Further, a dynamic relationship can be characterised by the presence of a lagged de- pendent variable among the regressors3. uit is composed of μit = μi + λi + νit, where μi is an unobserved individual-specifi c time-invariant eff ect which allows for heterogeneity in the means of the average markup across individual fi rms, λt is a time-specifi c individual- invariant eff ect and νit is a disturbance term. Because the aim of this study is to test the hypothesised impact of a fi rm’s investment plans on its markup policy, we estimate a fi rm-level model with a specifi cation that in- cludes a fi rm’s planned growth according to fi rm-specifi c markup determinants sug- gested by theory and empirical evidence. Th e model also controls for industry member- ship and changes in the economic and institutional environment. Th e model allows us fi rst of all to explain the deviations of a fi rm’s markups from the industry average, and especially to investigate whether a fi rm’s growth plans are aff ecting its pricing decisions. Th e variables of the model are specifi ed as follows. Th e appropriate empirical measurement of the markup that arises from theory is a con- tentious issue and empirical results have been shown to be sensitive to the measure of the margins that is used. In the Industrial Organisation tradition, the diff erence be- 3 For the purpose of clarity, the lags and expected values of some variables as well as some interaction terms between regressors are not explicitly included in the general model, but are considered in detail in the speci- fi cations of the empirical model. ),,( ijtjttijt fmarkup εηγ= ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 2 | 2009124 tween price and production costs is usually defi ned as the price-cost margin (herein- aft er ‘PCM’) as proposed by Collins and Preston (1969) and improved by Domowitz, Hubbard and Petersen (1986), which is a good proxy for Lerner’s degree of monopoly (Lerner, 1934) on the proposition that MC = AVC. A similar, but theoretically diff er- ently based idea of the structure of price is that of Kalecki (1954), according to which the degree of monopoly (markup μ) is derived from the price equation, where the price is a product of the markup and the variable unit cost of production. By using sales, inventories and costs in a similar manner as Domowitz, Hubbard and Petersen (1986), Kalecki’s version of the markup defi nition as the ratio between price and unit direct cost of production can be constructed. When multiplied by the quantity produced, the fi rm’s markupijt is thus defi ned as the ratio between a fi rm’s revenues and direct (vari- able) costs: (3) and the average industry markup INDmarkupjt as (4) We proxy the fi rm’s planned growth GRijt, which is the dependent variable that we are focussing on, by the growth of a fi rm’s fi xed assets. A one-year lead of asset growth is included among the regressors as it is presumed that all of the fi rm’s plans are fully car- ried into eff ect in the next year, as proposed by Blecker (1989). We have already argued that when a product’s price set by an oligopolistic fi rm is seen fi rst of all as ‘a reproduc- tive price and growth price’ (Lee, 1998), the fi rm’s growth ambitions become one of the most important markup determinants (Eichner, 1973) since the markup is the source of generating profi ts to fi nance growth. Other fi rm-specifi c markup determinants are defi ned as proposed by existing empirical investigations. Market share MSijt is defi ned as the share of a fi rm’s domestic market sales in the fi ve-digit NACE industry annual sales (the home sales of domestic fi rms in an in- dustry plus imports in industry j). Th e simple oligopoly model of fi rm performance im- plies a positive relationship between market share and markup size because a fi rm with a bigger market share is able to charge higher prices (and therefore achieve a superior level of markup) due to its stronger markup power (Stigler, 1968). Th e empirical literature also shows that the relationship is very likely not to be linear and that a certain threshold market power (market share) oft en exists (Feeny and Rogers, 1999; Bennenbroek and Haris, 1995). However, an opposing hypothesis is that as market share increases com- petitive pressures are weakened, suggesting that profi tability (and the markup) is lower because the incentive to minimise costs is no longer important. Equally, it is possible that fi rms with lower market shares are smaller and more fl exible, allowing for lower costs and higher profi tability. INDmarkupjt = Σ value of salesijt + Σ Δ inventoriesijt Σ payrollijt + Σ cost of materialijt markupijt = value of salesijt + Δ inventoriesijt payrollijt + cost of materialijt NINA PONIKVAR, MAKS TAJNIKAR | PLANNED GROWTH AS A DETERMINANT OF THE MARKUP: ... 125 Th e criterion for the fi rm’s sizeijt is the number of employees. Firms are regarded as small when they have less than 50 employees and as large when they have 250 or more em- ployees. All other fi rms are regarded as being of a medium size. Th e impact of size on the markup can be twofold. On one hand, larger fi rms have larger market power (Bain, 1956) and/or are more effi cient (Penrose, 1972; Demsetz, 1973) and can therefore achieve higher markups. On the other hand, when the markup is measured in gross form (see Kalecki, 1954), as in our empirical study, larger fi rms theoretically have lower overhead unit costs and can therefore charge lower markups. Th e utilisation of production capacities CUijt of fi rm i from industry j in year t is defi ned as a ratio between the actual and potential volume of sales of a fi rm, where the potential sales of fi rm i are a product of the highest existing ratio between sales and production capacities (fi xed assets) in the period 1994 to 2004 and the production capacities of fi rm i from industry j in year t. Production capacities are measured in terms of fi xed assets. In a short-term analysis, the production capacities of a fi rm and its capacity costs are given. However, a fi rm can produce various quantities of output with the same produc- tion capacities. Th ere are three possible eff ects of a fi rm’s production capacity utilisation on the markup size. Th e fi rst is the negative eff ect of capacity utilisation on the markup size in the case of target return pricing (Lanzillotti, 1958). Th e second is a positive eff ect due to the higher technical effi ciency of a fi rm, which utilises its production capacities better (Blecker, 1989). Th e more the fi rm utilises its capacities, the higher the output it produces. At given unit variable costs and at a given price, fi xed unit costs are lower at a higher production capacity utilisation and consequently the markup can be higher. Th e third source of a possible positive relationship between capacity utilisation and the markup level is the incentive of an oligopolistic fi rm to keep some level of reserve capaci- ties, allowing the exploitation of any chance increase in selling power and acting as a competitive weapon (Sylos-Labini, 1969). Th e higher capacity utilisation of a fi rm also indicates that fewer reserve capacities are available and that a fi rm is moving closer to full capacity utilisation. Th e latter forces the fi rm to plan its investments in additional production capacity in order to be able to adapt to changing demand conditions with some level of reserve capacity. Th e last two reasons speak in favour of a positive capacity utilisation-markup level linkage. A fi rm’s labour productivity Lprodijt is defi ned as the value added per employee in real terms, the price of labour on fi rm level wijt is calculated by dividing real annual gross wages by the average number of employees for each fi rm, while the price of capital rijt is defi ned as the ratio of the sum of depreciation and the cost of fi nancing to the sum of fi xed assets and inventory. More productive fi rms are able to charge higher markups due to their lower unit costs at given prices of inputs. It is therefore expected that labour productivity explains the variability of the fi rm-level markups of fi rms within the same industry since these fi rms compete with each other. In addition, the price set for a par- ticular product by a fi rm is the sum of the unit production cost and the markup. Higher production factor prices on the fi rm-level, leading to higher production costs and also to higher unit costs, do not always result in higher prices. How much of the higher costs will be spilled over into higher prices depends on the strength of the competition within ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 2 | 2009126 a particular industry. When the competition among fi rms within the same industry is strong enough, it is able to prevent the complete (or even any) transformation of the higher unit cost into higher prices of fi nal products. In such cases, higher production costs, especially the cost of labour, which are not covered by the fi rm’s markup, lead to lower markups. Because the markup is defi ned in a gross form, it is composed of one part for profi t and another part for covering overhead costs (including capital costs). It is thus expected that a higher capital price leads to a higher cost of capital and therefore to a higher markup at a given level of sales. A fi rm’s export orientation EXorijt is measured by the share of revenues from exports in the total annual sales of fi rm i. Empirical literature and theory suggests that fi rms that sell their products in domestic and foreign markets are disciplined by foreign competi- tion and thus charge lower markups within their price (Bughin, 1996; Caves and Porter, 1980), although the direction of the impact of an export orientation on markups also depends on the market structure and institutional framework. Th e capital intensity of a fi rm’s production KIijt is calculated as the ratio of total fi xed assets to the number of employees. Firms with a higher capital intensity of production compared to the industry average are expected to have a lower fi rm-level markup com- pared to the average markup in the industry due to their inferior cost effi ciency (Coelli et al., 1998). Th is implies a higher capital cost per unit of output and, at a given output’s price, it leads to a lower markup. Further, studies using price-cost margins or markup as a dependent variable generally employ the capital/revenue ratio as a control because the margins and markup are used in their gross form (see the overview in Schmalen- see, 1989). A positive relationship between the capital intensity of production and the markup size should only appear on the industry level where more capital demanding technology leads to higher industry-level markups. Th e broad literature on profi t persistence (Mueller, 1977; Mueller and Cubbin, 1990) sug- gests that current markups will be heavily infl uenced by the past realisation of such. Econometrically, this necessitates the additional inclusion of a lagged dependent variable in the basic specifi cation. In addition, a serial correlation of markups and profi t margins is empirically observed in time series (Machin and Van Reenen, 1993). Both issues sug- gest that current output conjectures may depend on previous performance. Th e model is thus specifi ed as: (5) where subscript i refers to a fi rm, subscript j to industries according to the fi ve-digit NACE classifi cation of industries and subscript t to a particular year, respectively. Th e average industry-level markup is included in the model in order to control for the infl u- ence of industry and market characteristics as well as the impact of environmental and markupijt = α + β1 markupij,t-1 + β2INDmarkupjt + β3Grai(t+1) + β4MSit + β5MS2it + β6EXorit + + β7Lprodit + β8wit + β9rit + β10KIit + β11CUit + uit NINA PONIKVAR, MAKS TAJNIKAR | PLANNED GROWTH AS A DETERMINANT OF THE MARKUP: ... 127 institutional factors, which infl uence all fi rms in a particular industry in the same year in the same fashion. In Table 1 the characteristics of an average fi rm in the sample according to the year and fi rm size are described. TABLE 1: Characteristics of the database 1994 to 2004 period average All fi rms Small fi rms Medium fi rms Large fi rms Markup 1.125 1.134 1.078 1.095 Annual growth of fi xed assets 1.07 6.88 1.41 1.02 Average number of employees 55 7 117 658 Market share 2.4 1.0 6.5 15.1 In the 1994–2004 period the average Slovenian manufacturing fi rm from the sample employed 55 people and had a market share of 2.4 percent. Th is average fi rm set its prices 12.5 percent above its variable unit costs although the average markup varied from 11 to 14 percent in the 1994-2004 period. According to the size class, small fi rms achieved the highest and medium-sized fi rms the lowest markups on average, while large fi rms remained in the middle during the whole investigated period. Th e average fi rm had an average annual growth of fi xed assets (in real terms) of 1 percent, while on average the growth in the value of a fi rm’s fi xed assets was negative at the beginning of the 1994– 2004 period and positive in later years. Th e model is estimated in three specifi cations, denoted I, II and III respectively. Specifi - cation I includes all fi rm-level markup determinants proposed by theory and empirical literature as well as the average industry-level markup as a control variable for the indus- try-specifi c factors and changes in the economic environment. In specifi cation II a set of year time dummies is additionally tested as a measure of the impact of the changes in the environment, while a fi rm’s size is added in the form of a set of size dummies in the third specifi cation (III). Th e latter specifi cation is included because the descriptive statistics of the data (see Ponikvar, 2008 and Table 1) show that in Slovenian manufacturing the characteristics regarding a fi rm’s capital intensity, export orientation, price of produc- tion factors and productivity diff er a great deal among small, medium and large fi rms. 3.3 Method Th e lagged dependent variable among the regressors complicates the application of the markup dynamic panel since yit is a function of μi and it thus immediately follows that yi,t-1 is also a function of μi. Th erefore, yi,t-1, the right-hand side regressor in the model is correlated with the error term and the OLS estimator is thus biased. Further, the usual panel data techniques cannot be used for the above equation since they are biased and inconsistent as N→∝ and fi nite T in a dynamic setting (Nickell, 1981). In addition, the fact that the specifi cation of models includes fi rm-specifi c variables can also imply the ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 2 | 2009128 possibility of endogeneity arising from individual eff ects, that is from the fact that fi rm- level variables are likely to be correlated with unobserved fi rm-specifi c eff ects μi. Be- sides, the possibility of simultaneity bias should also be considered since, according to the theoretical origins of the Structure-Conduct-Performance paradigm, some funda- mental variables in the model of fi rm performance (e.g. markup, concentration, product diff erentiation) are jointly determined (Hay and Morris, 1991) and as such do not satisfy the zero-conditional-mean assumption. In our case, the most apparent possible source of endogeneity among the regressors are sellers’ concentration, market share, import penetration, export orientation etc. Th ese issues prevent the standard procedures for estimating panel data models from being consistent and/or effi cient. Arellano and Bond (1991) propose the Generalised Method of Moments procedure (hereinaft er ‘AB GMM’) which off ers a large feasible instrument set by exploring instruments motivated by moment conditions, compared to Anderson and Hsiao (1982). Th e instruments include suitable lags of the levels of en- dogenous variables, which enter the equation in diff erenced form, as well as strictly ex- ogenous regressors and any others that might be specifi ed. It is argued that all the xi,t are also valid instruments for the fi rst diff erenced equation if xi,t are correlated with μi. Th is permits us to exploit both the cross-section and time-series elements of the data in constructing instruments and hence yields effi ciency gains relative to other estimation methods for panel data. In the case of the presence of heteroscedasticity in the model, a two-step procedure should be used where the fi rst-step residuals are used to compute the variance covariance matrix in a second step. In other words, Δ itν need to be replaced by diff erenced residuals obtained from the one-step estimator and the resulting estimator becomes the Arellano-Bond two-step estimator. Th e consistency of the two Arellano Bond GMM estimators hinges heavily upon the assumption that the 0)( )2( =−tiitE νν , where E is the mathematical expectation. 0)( )1( =−tiitE νν need not be zero since the itν are diff erences of serially uncorrelated errors. Arellano and Bond (1991) therefore propose a test of hypothesis H0 that there is no second-order serial correlation for the disturbances of the fi rst-diff erenced equation with the test statistic m2 for second-order serial correlation based on residuals from the fi rst-diff erenced equation. A further aspect of interest concerns the validity of the chosen instruments above the minimum set nec- essary for econometric identifi cation. Although we cannot test the validity of the instru- ments directly, we can assess the adequacy of instruments in an over-identifi ed context with a test of over-identifying restrictions. If we reject the null hypothesis of such test, we cast doubt on the suitability of the instrument set and establish that one or more of the applied instruments do not appear to be uncorrelated with the disturbance process (Baum, 2006). In our case, a test of over-identifying restrictions as advised by Sargan (1958) is used. 4. RESULTS Th e null hypothesis of the Wald test that the estimated coeffi cients of all regressors are all zero is rejected in all of the tested specifi cations. Th e moment conditions in the model are NINA PONIKVAR, MAKS TAJNIKAR | PLANNED GROWTH AS A DETERMINANT OF THE MARKUP: ... 129 appropriate since the null hypothesis of the Sargan test of over-identifying restrictions cannot be rejected, which is also our case. Crucial for dynamic models is the absence of autocorrelation of diff erenced model residuals of order 2. It is evident from the test sta- tistics m2 that in none of the model specifi cations does such an autocorrelation exist. On the other hand, the null hypothesis that average autocorrelation in residuals of order 1 is 0 is rejected, which is also what was expected with regard to the estimation technique applied. TABLE 2: Firm-specifi c markup determinants Dependent variable: Firm-level markup Specifi cation I Specifi cation II Specifi cation III Markup(-1) 0.1306 (7.46)** 0.1285 (7.32)** 0.1275 (7.27)** INDmarkup 0.2082 (8.60)** 0.2075 (7.91) ** 0.2076 (7.94)** žMS -0.6633 (-6.97)** -0.5337 (-5.86)** -0.5297 (-5.82)** MS2 0.5926 (5.94)** 0.5005 (5.18)** 0.4978 (5.18)** EXor -0.1285 (-3.35)** -0.1507 (-3.68)** -0.1496 (-3.62)** GRa (+1) 0.0096 (4.73)** 0.0091 (4.50)** 0.0091 (4.49)** Lprod 0.0000029 (5.51)** 0.0000031 (5.97)** 0.0000031 (5.95)** W -0.000055 (-9.84)** -0.00006 (-10.69)** -0.00006 (-10.59)** R 0.1221 (2.09)* 0.0806 (1.36) 0.0800 (1.35) KI 0.0000052 (5.48)** 0.0000055 (5.78)** 0.0000055 (5.77)** CU 0.0222 (3.88)** 0.0251 (4.34)** 0.0253 (4.36)** Medium size fi rm -0.0032 (-0.35) Large size fi rm -0.0147 (-1.14) Year dummy NO YES YES Constant 0.0051 (4.97)** 0.0080 (6.77)** 0.0080 (1.99)* No. of observations 20466 20466 20466 No. of fi rms (i) 4470 4470 4470 Instrumented MS, EXor, CU MS, EXor, CU MS, EXor, CU (df) Wald χ2 (11) 322.05** (19) 416.52** (21) 424.72** (df) Sargan χ2 (140) 187.33 (140) 143.42 (140) 143.51 m1 -15.24** -15.23** -15.21** m2 0.10 0.03 0.02 Notes: - t-statistics are in parentheses - **,* denote signifi cance at 1% and 5%, respectively Th e size and signs of the estimated regression coeffi cients are mostly in accordance with the theoretical expectations, with their size remaining relatively stable regardless of changes in the model specifi cation, which is an indicator of the model’s robustness. Th e fi rm’s planned growth positively impacts on the markup size, as proposed by Post- Keynesian theory. Th e size of the estimated coeffi cient of planned growth on the markup size is stable in all specifi cations. Various economic reasons can be provided to explain this link. First, retained profi ts are a prime source of capital for a fi rm seeking expansion ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 2 | 2009130 and/or, alternatively, if a fi rm seeks to raise capital externally, adequate profi tability is likely to be viewed by lenders as an important prerequisite. Second, according to the in- stitutional theory of the fi rm the main objective of the ‘megacorp’ is to grow and expand its market share (Eichner, 1973). In addition, according to well-known managerial theo- ries whereby managers have a discretion to pursue their own objectives as well, growth as well as profi t may enter the fi rm’s objective function and are thus positively correlated. Our results show that profi ts accumulated from markups are needed to fi nance growth (a high markup today is a precursor of high growth in the next period). Similar results are reported by Goddard et al., (2005) for a panel of manufacturing fi rms from the EU. Th e direction of the impact of other regressors in the model is in accordance with the ex- pectations. Firm-level markups relative to the industry average are higher when labour is more productive and when the price of labour is lower. On the other hand, the fi rm-level price of capital measured on the does not aff ect the markup size in a signifi cant way. Th e linear relationship between a fi rm’s market share and fi rm-level markup is signifi cant and negative. Such a result is not predicted by the oligopoly models and is in contrast to some similar studies for developed European countries (see Goddard et al., 2005 for a review). One possible explanation is that being small is more advantageous in a small economy such as Slovenia. However, empirical literature has provided evidence of a U- shaped relationship between market share and profi tability. In our case, when the pos- sibility of both the linear and non-linear impact of the market share size on the markup is incorporated in the model the linear link remains statistically signifi cantly negative, while the quadratic link is signifi cantly positive. It is thus possible to identify the ‘thresh- old’ market share size above which the market share starts to increase markups and it is surprisingly high, amounting to a 53 percent market share. Other studies for larger economies (e.g. Fenny et al., 2005 for Australia) fi nd this threshold market share to be much smaller. An acceptable explanation is that a fi rm must have a relatively large mar- ket share in the small Slovenian market to obtain enough market power in a general product market to be able to achieve higher prices and increase its markups. Exposure to competition in foreign markets obviously decreases fi rms’ markup sizes in Slovenian manufacturing. Th is is also confi rmed in studies for manufacturing in other countries (e.g. Bennenbroek and Harris, 1995; Kambhampati and Parikh, 2003). Evidently, stronger competitive pressure due to greater exposure to competition abroad and the higher export orientation of fi rms decreases markups. In addition, large Slovenian manufacturing fi rms are more export-oriented and have smaller markups on average (Ponikvar, 2008). Th e result that the higher capital intensity of a fi rm’s production increases a fi rm’s markup relative to the industry average was not expected since a fi rm’s higher capital intensity of production relative to the industry average means inferior cost effi ciency and should therefore result in a fi rm’s lower markup relative to the industry markup. A posi- tive link between the markup and the capital intensity of production is only expected to appear at the industry level. However, our results are in line with some other empirical studies (Bennenbroek and Harris, 1995; Feeny et al., 2005) where the diff erence between the industry- and fi rm-level impact was not accounted for. Firm-level capacity utilisation NINA PONIKVAR, MAKS TAJNIKAR | PLANNED GROWTH AS A DETERMINANT OF THE MARKUP: ... 131 also increases fi rm-level markups. Th e higher utilisation of a fi rm’s production capaci- ties means higher production (and possible revenues) at a given capital cost. Hence the achieved markups of fi rms can be higher at a given price. Similar results can be found in Bennenbroek and Harris (1995) for manufacturing industries in New Zealand. Descriptive statistics of the dataset applied in the analysis (see Ponikvar, 2008) show that larger fi rms achieve lower markups on average, which is a logical consequence of the markup’s gross defi nition. However, this is not confi rmed in our models (III) where the impact of a fi rm’s size (measured by dummy variables for small, medium and large fi rms) on markups is negative but insignifi cant. Th e estimated coeffi cient on the lagged markup is positive and signifi cant. Th ese results are in line with the fi ndings of the persistence of profi tability literature (see the overview in Mueller and Cubbin, 1990). Th e coeffi cient amounts to approximately 0.13. It indicates that a 1 percent increase in the last year’s fi rm-level markup will result in a 0.13 percent increase in this year’s fi rm-level markup. In other words, 87 percent of the total adjust- ment in a fi rm’s markup from a shock will occur in the fi rst year, while 13 percent will not. Th is indicates that in Slovenian manufacturing the return of the markup to some equilibrium level is monotonic (δ<1) and fast. Th e size of the obtained ‘persistence’ coef- fi cient for Slovenian manufacturing fi rms is relatively small compared to studies for oth- er economies. In estimations for other countries the coeffi cients on the lagged dependent variable in performance equations range from 0.2 to almost 0.5 (Fenny et al., 2005 and McDonald, 1999 for manufacturing fi rms in Australia; Machin and Van Reenen, 1993 for UK fi rms; Goddard et al., 2005 for Belgium, France, Italy and the UK). One possible explanation is that the markup defi nition in our study follows Kalecki’s defi nition while the abovementioned studies use PCM as the dependent variable. Another, more content- oriented reason is that the relatively large export orientation of manufacturing fi rms, their exposure to competition abroad, the relatively large import penetration and rela- tive smallness of Slovenian markets force manufacturing fi rms in Slovenia to adapt their markups to changed market conditions faster than in other larger economies. Th e average industry markup included among the regressors shows a positive and sig- nifi cant impact on the markup of fi rms that belong to the industry. An increase in the av- erage markup in an industry by 1 percentage point infl uences fi rms within this industry to increase their own markups on average by 0.2 of a percentage point. Th e positive and statistically signifi cant coeffi cient thus shows the interdependence of the pricing deci- sions of fi rms within a particular industry. It thus empirically confi rms the theoretical markup pricing equation (Kalecki, 1954, p.12), in which the pricing decision of a fi rm is infl uenced by its characteristics as well as the average industry price. 5. CONCLUSIONS Our analysis shows that the diff erences seen in markups among Slovenian manufactur- ing fi rms within the same industry can be explained by diff erences in their ambitions to ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 2 | 2009132 grow, diff erences in their market share, diff erences in the utilisation of their production capacity, diff erences in the price of labour and productivity of their labour, the capital intensity of their production and diff erences in their export orientation. In Slovenian manufacturing investment decisions aff ect the pricing policy and the deci- sions on the markup size, as proposed by Post-Keynesian theory. Our results show that planned growth positively aff ects the size of a fi rm’s markup relative to its rivals, which confi rms our hypothesis. We may conclude that the profi ts accumulating from markups are obviously needed to fi nance growth (a high markup today is a precursor of high growth in the next period). Th is empirical evidence on the positive fi rm-level relationship between growth and markup also has some policy implications. It has generally been acknowledged that com- petitive pressure reduces markups, forces fi rms to organise themselves more effi ciently and, as such, increases economic welfare. It would thus be socially desired for competi- tive pressures to be high and markups to be relatively low. On the other hand, the empiri- cal evidence shows that fi rms which grow faster have higher markups compared to fi rms with lower growth ambitions. Economic growth can therefore be achieved only when markups are not on low competitive levels. Restrictive competition and/or antitrust pol- icy might therefore, although resulting in a more competitive industrial environment and fi rms’ decreased market power, slow down the grow path of Slovenian manufactur- ing industries since they limit the source of internal funds for investment fi nancing via decreased markups. Obviously, at least a partial trade-off between a fi rm’s growth and competitive outcome exists, which is an important issue that has to be considered by competition policy authorities when adopting policy measures. Th e same also holds for restrictive fi scal policy. Th e empirical evidence also reveals the interdependence of the pricing decisions of fi rms within a particular industry since any change in the industry markup is refl ected in the markup size of the fi rm. Th e theoretical markup pricing equation (Kalecki, 1954, p.12), in which a fi rm’s pricing decision is infl uenced by its characteristics as well as by the av- erage industry price, is thus confi rmed in the case of Slovenian manufacturing. RECEIVED: NOVEMBER 2008 REFERENCES: Anderson, T.W. and Hsiao, C. (1982), “Formulation and Estimation of Dynamic Models Using Panel Data”, Journal of Econometrics 18 (1), 47–82. Andrews, P.S. (1955), Manufacturing Business, London: Macmillan. Arellano, M. and Bond, S. (1991), “Some Tests of Specifi cation for Panel Data: Monte Carlo Evidence and an Application to Employment Equations”, Review of Economic Studies 58 (194), 277–297. NINA PONIKVAR, MAKS TAJNIKAR | PLANNED GROWTH AS A DETERMINANT OF THE MARKUP: ... 133 Athey, M. J. and Laumas, P. S (1994), “Internal Funds and Corporate Investment”, Journal of Development Economics 45 (2), 287–303. Athey, M. J. and Reeser, W. D. (2000), “Asymmetric Information, Industrial Policy and Corporate Investment in India”, Oxford Bulletin of Economics and Statistics 62 (2), 267–292. Bain, J.S. (1956), Barriers to New Competition, Cambridge, Mass.: Harvard University Press. Barney, J. (1991), “Firm Resources and Sustained Competitive Advantage”, Journal of Management 17 (1), 99–120. Bartlett, W. and Bukvič, V. (2001), “Barriers to SME Growth in Slovenia”, SME Development 11 (2), 177–195. Baum, F. C. (2006), An Introduction to Modern Econometrics Using Stata, Texas: Stata Press. Bennenbroek, N. and Harris, R.I.D. (1995), “An Investigation of the Determinants of Profi tability in New Zealand Manufacturing Industries in 1986-1987”, Applied Economics 27 (11), 1093–1101. Blecker, R. A. (1989), “Markup Pricing, Import Competition, and the Decline of the American Steel Indus- try”, Journal of Post Keynesian Economics 12 (1), 70–87. Bond, S. R. et al. (2003), “Financial Factors and Investment in Belgium, France, Germany and the United Kingdom: A Comparison Using Company Panel Data”, Review of Economics and Statistics 85 (1), 153–165. Bughin, J. (1996), “Capacity Constraints and Export Performance: Th eory and Evidence from Belgian Manu- facturing”, Journal of Industrial Economics 44 (2), 187–204. Carpenter, R. E. and Petersen, B. C. (2002), “Is the Growth of Small Firms Constrained by Internal Finance”, Review of Economics and Statistics, 84 (2), 2, 298–309. Caves, R. E. and Porter, M. E. (1980), “Th e Dynamics of Changing Seller Concentration”, Journal of Industrial Economics 29 (1), 1–15. Coelli, T.J., Rao, P.D.S. and Battese, G.E. (1998), An Introduction to Effi ciency and Productivity Analysis, Klu- wer Academic Publishers. Collins, N.R. and Preston, L.E. (1969), “Price-Cost Margins and Industry Structure”, Review of Economics and Statistics 51 (3), 271–286. Demsetz, H. (1973) “Industry Structure, Market Rivalry and Public Policy”, Journal of Law and Economics 16 (1), 1–10. Domowitz, I., Hubbard, G. R. and Petersen, B. C. (1986), “Business Cycles and the Relationship between Concentration and Price Cost Margins”, Rand Journal of Economics 17 (1), 1–17. Eichner, A. S. (1973), “A Th eory of the Determinants of the Mark-up under Oligopoly”, Economic Journal 83 (332), 1184–1200. Eichner, A. .S. (1976), Th e Megacorp and Oligopoly, Cambridge: Cambridge University Press. Eichner, A. S. and Kregel, J.A. (1975), “An Essay on Post-Keynesian Th eory: A New Paradigm in Economics”, Journal of Economic Literature 13 (4), 1293–1314. Feeny, S., Harris, M. N. and R., Mark (2005), “A Dynamic Panel Analysis of the Profi tability of Australian Tax Entities”, Empirical Economics 30 (1), 209–233. Feeny, S. and Rogers, M. (1999), Market Share, Concentration and Diversifi cation in Firm Profi tability. Mel- bourne: Melbourne Institute Working Papers, no. 22/99. ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 2 | 2009134 Goddard, J., Tavakoli, M. and Wilson, J.O.S. (2005), “Determinants of Profi tability in European Manufactur- ing and Services: Evidence from Dynamic Panel Model”, Applied Financial Economics 15 (18), 1269–1282. Hadi, A. S. (1992), “Identifying Multiple Outliers in Multivariate Data”, Journal of the Royal Statistical Society, Series B 54, 761–771. Hall, R.L. and Hitch, C.J. (1952), “Price Th eory and Business Behaviour”, In Wilson T. and P.W.S Andrews, eds. Oxford Studies in the Price Mechanism, Oxford: Clarendon Press, 107–138. Harcourt, G.C. and Kenyon, P. (1976), “Pricing and the Investment Decisions”, Kyklos 29 (3), 449–477. Hay, D. A. and Morris, D. J. (1991), Industrial Economics and Organisation: Th eory and Evidence, Oxford: Oxford University Press. Hubbard, G. R. (1998), “Capital-Market Imperfections and Investment”, Journal of Economic Literature 36 (1), 193–225. Kadapakkan, P. R., Kumar, P.C. and Riddick, L. (1998), “Th e Impact of Cash Flow and Firm Size on Invest- ment: Th e International Evidence”, Journal of Banking and Finance 22 (3), 293–320. Kalecki, M. (1954), Th eory of Economic Dynamics: An Essay on Cyclical and Long-Run Changes in Capitalist Economy, London: Allen&Unwin. Kalecki, M. (1971), Cost and Prices, Selected Essays on the Dynamic of the Capitalist Economy, 1933-1970, Cambridge: Cambridge University Press. Kambhampati, U. S. and Parikh, A. (2003) “Disciplining Firms: Th e Impact of Trade Reforms on Profi t Mar- gins in Indian Industry”, Applied Economics 35 (4), 461–470. Konings, J., Van C., Patric and Warzynski, F. (2001), “Th e Dynamics of Industrial Mark-ups in Two Small Open Economies: Does National Competition Matter?”, Journal of Industrial Organization 19 (5), 841–860. Konings, J., Van Cayseele, P. and Warzynski, F. (2005) “Th e Eff ects of Privatization and Competitive Pres- sure on Firms’ Price-Cost Margins: Micro Evidence from Emerging Economies”, Review of Economics and Statistics 87 (1), 124–134. Lanzillotti, R. F. (1958), “Pricing Objectives in Large Corporations”, American Economic Review, 48 (5), 921–940. Lee, F. S. (1998), Post Keynesian Price Th eory, Cambridge: Cambridge University Press. Lerner A. (1934), “Th e Concept of Monopoly and Measurement of Monopoly Power”, Review of Economic Studies 1 (3), 157–175. Machin, S. in Van Reenen, J. (1993), “Profi t Margins and the Business Cycle: Evidence from U.K. Manufactur- ing Firms”, Journal of Industrial Economics 41 (1), 29–50. Martin, S. (2001), Industrial Organisation: A European Perspective, Oxford: University Press. McDonald, J. T. (1999), “Th e Determinants of Firm Profi tability in Australian Manufacturing”, Economic Record 75 (229), 115–126. Motta, M. (2004), Competition Policy: Th eory and Practise, Cambridge: Cambridge University Press. Mueller D. C. (1977), “Th e Persistence of Profi ts above the Norm”, Economica 44 (176), 369–380. Mueller, D. C. and Cubbin, J. (1990), Th e Dynamics of Company Profi ts: An International Comparison, Cam- bridge: Cambridge University Press. NINA PONIKVAR, MAKS TAJNIKAR | PLANNED GROWTH AS A DETERMINANT OF THE MARKUP: ... 135 Nickell, S. (1981), “Biases in Dynamic Models with Fixed Eff ects”, Econometrica 49 (6), 1417–1426. Oliver, H. M. Jr. (1947), “Marginal Th eory and Business Behaviour”, American Economic Review 37 (3), 375– 383. Penrose, E. T. (1972), Th e Th eory of the Growth of the Firm, Oxford: Basil Blackwell. Ponikvar, N. (2008), Markup Determinants in Manufacturing Industries: Th e Case of Slovenia, Doctoral dis- sertation, Ljubljana: Faculty of Economics. Porter, M. E. (1980), “Industry Structure and Competitive Strategy: Keys to Profi tability”, Financial Analysts Journal 36 (4), 30–41. Porter, M. E. (1981), “Th e Contributions of Industrial Organisation to Strategic Management”, Academy of Management Review 6 (4), 609–620. Sargan, J. D. (1958), “Th e Estimation of Economic Relationships Using Instrumental Variables”, Economet- rica 26 (July) 393–415. Schmalensee, R. (1989), “Inter-Industry Empirical Analysis”, In Schmalensee R. and Willig R. D, eds. Th e Handbook of Industrial Organization, Vol. 2, Amsterdam: North-Holland, 951–1009. Shapiro, N. (1981), “Pricing and the Growth of the Firm”, Journal of Post Keynesian Economics 4 (1), 85–100. Stigler, G. J. (1968), Th e Organisation of Industry, Homewood: Richard D. Irwin. Sylos-Labini, P. (1969), Oligopoly and Technical Progress, Cambridge, Mass.: Harvard University Press. Tajnikar, M. and Ogrin, N. (2001), “Post-Keynesian Explanation of the Transitional Adaptation of the In- vestment Function in Slovenia”, In Fourth International Conference on Enterprise in Transition: Proceedings, Split: University of Split, Faculty of Economics. Van Ees, H., Kuper, G. H. and Sterken, E. (1997), “Investment, Finance and the Business Cycle: Evidence form the Dutch Manufacturing Sector”, Cambridge Journal of Economics 21 (3), 395–407. Wood, A. (1975), A Th eory of Profi ts, London: Cambridge University Press. Worthington, P. R. (1995), “Investment, Cash Flows and Sunk Costs”, Journal of Industrial Economics, 43 (1), 49–61.