Bachelor Thesis
The market-share profitability relationship revised.
Florian Pruim
Author: Florian Pruim
Student nr: 10423702
Study: Economie & Bedrijfskunde
Specialization: Finance & Organization
Field: Finance
Supervisor: dhr. drs. P.V. Trietsch
Date: 29-06-2015
ABSTRACT:
This study revises the relationship between market share and profitability. This is done for the top 250 companies based in 31 different countries and operating in 10 different industry categories. The effect of market share is measured by performing an ordinary least squares regression. The only significant effect of market share on profitability found in this study, was negative with a value of 0.99. The conclusion is that market share is not a key to profitability and firms should therefore focus on other strategic pillars to accomplish higher profit rates. It is concluded that market share should be seen as an indicator of firm performance rather than a determinant of profitability.
2 Verklaring eigen werk
Hierbij verklaar ik, Florian Pruim, dat ik deze scriptie zelf geschreven heb en dat ik de volledige verantwoordelijkheid op me neem voor de inhoud ervan.
Ik bevestig dat de tekst en het werk dat in deze scriptie gepresenteerd wordt origineel is en dat ik geen gebruik heb gemaakt van andere bronnen dan die welke in de tekst en in de referenties worden genoemd.
De Faculteit Economie en Bedrijfskunde is alleen verantwoordelijk voor de begeleiding tot het inleveren van de scriptie, niet voor de inhoud.
3 Acknowledgements
There are several people I would like to thank for their help during the process of research and writing my thesis. First of all, I would like to thank my supervisor Mr. Trietsch for supporting me in writing my thesis. By highlighting the critical points and providing me with writing tips, I was able to improve the thesis. Next to this I would like to thank the coordinator of the Thesis seminar, Mr. Versijp. By giving feedback during the tutorials, I didn’t get lost in the process and kept an eye on the final goal. I also want to thank my colleagues of my internship at Ahold, for providing me with additional and important insights in this topic. Finally, I want to thank my girlfriend, friends and family who greatly supported me throughout the whole process of research and thesis writing.
4 Table of Contents 1. Introduction ... 5 2. Literature review ... 6 2.1 Market share ... 7 2.2 Firm characteristics ... 10 2.2.1 Firm size ... 10
2.2.2 Research and Development ... 11
2.2.3 Firm growth ... 11 2.3 Market characteristics ... 12 2.3.1 Market growth ... 12 2.3.2 Industry Concentration... 12 2.3.3. Financial crisis ... 13 2.3.4 Industry category ... 13 2.4 Prior return ... 14 2.5 Non linearity ... 14 3. Method ... 14 3.1 Profitability ... 15 3.2 Market share ... 15 3.3 Firm characteristics ... 15 3.3.1 Firm size ... 15 3.3.2 R&D ... 15 3.3.3 Firm growth ... 16 3.4 Market characteristics ... 16 3.4.1 Market growth ... 16 3.4.2 Industry concentration ... 16 3.4.3 Financial crisis ... 17 3.4.4 Industry group ... 17 3.5 Prior return ... 17
3.6 Quadratic market share ... 18
3.7 Regression model: ... 18
4. Results and Analysis ... 19
5. Discussion and Conclusion ... 22
5 1. Introduction
Understanding the determinants of financial performance is a central issue in economic literature. By recognizing the factors which influence profitability, an organization can optimize the strategy to obtain significant operating results. There is an ongoing debate which determinant is most important in explaining the difference in profit levels between firms. The Profit Impact of Market Strategies (PIMS) project described by Buzell, Gale and Sultan (1975) aimed at identifying and measuring the key determinants of financial performance. In their study, they conclude that market share is one of the main determinants of profitability.
In addition to Buzell et al. (1975), several studies have been done on the effect of market share on profitability, but the sign, the magnitude and the statistical significance of the results differ among these studies. Most of the findings confirm that market share has a positive effect on profitability (Shepherd, 1972; Gale, 1972; Schoeffler, Buzzell and Heany, 1974), but at the same time there is an ongoing debate if market share is a valid predictor of financial performance. While the correlation between the two factors is undeniable according to Laverty (2001), the causal relationship is often questioned. While most outcomes are based on the PIMS database there might exist a possibility of a bias. It is therefore useful to revise the market share profitability relationship and to find out if there is a positive effect of market share on financial performance. For this reason, this paper aims at identifying whether market share is a reliable determinant of financial performance and to find out what the effect of market share on profitability is.
Whether the effect of market share is indeed found to be positive is important from both a business strategic as an empirical point of view. Buzell et al. (1975) illustrates the strategic importance by specifying that an acceptable rate of return can only be earned by attaining some minimum level of market share. Prescott, Kohli and Venkatraman (1986) assert that if a positive and strong relationship exist, the pursuit of gaining market share as a goal may be appropriate to increase profitability. But in the case that the relationship is negative, weak or spurious, then this goal of gaining market share could be detrimental to the long-run continuity of the firm.
While the relationship is extensively described, most of these studies conclude that the effect of market share on profitability depends on both firm as market characteristics (Gale, 1972; Buzell et al. 1975). In this paper the firm characteristics: firm size, research and development expenses and firm growth rate are examined. For the market characteristics, market growth rate, industry concentration rate and industry category are examined. Since analysis by Buzell et al. (1975) points out that the importance of market share in effecting profitability differs considerably among industries, an industry dummy is used to compare whether there are differences between industry categories.
For this paper, information is collected for the top two hundred fifty companies based in thirty-one different countries and operating in ten different industry categories. The time period examined is from the year 2000 till 2013 thereby including the financial crisis of 2007-2011. For this study this comes down to 2600 observations.
6
With a fast changing economic environment, the earlier found results might not be applicable anymore. The purpose of this paper is therefore to find out if the effect of market share on profitability is still found to be positive and if any discrepancies accord during the financial crisis. The effect of the financial crisis will be measured by means of a dummy variable for the period 2007-2011.
The research question for this thesis is: what is the sign and the magnitude of the effect of market-share on profitability? In order to answer that question, a number of sub questions are established. First of all, it is needed to know what the reasoning is behind the market share-profitability relationship. Secondly it needs to be indentified how profitability and market share are measured. Finally, the other variables influencing profitability need to be identified together with their expected effect. After the required variables are determined, an OLS regression will be performed to obtain the results. Based on the literature the resulting hypothesis is that: a bigger market share is associated with higher profitability. When this effect is indeed confirmed to be positive, the pursuit of increasing market share to increase profitability is an appropriate one. It is argued that looking at the relationship in different markets, in a new time period and retrieving the data from another source, could produce additional and important insights.
This paper contains four primary sections. In the first section the conceptual model is discussed together with the theoretical effect of the determinants of profitability. The second section describes the data sources, sample selection and the construction of variables. In the third section the statistical results of the regressions will be discussed. The fourth and final section contains of the conclusions and some remarks.
2. Literature review
In this section the conceptual model will be discussed. Secondly, the reasoning behind the relation between market share and profitability will be discussed together with the main findings and concerns about the reliability. In addition, the other determinants of profitability will be discussed together with their expected effect on both profit levels and the effect of market share on profitability.
In Figure 1 the conceptual model is exposed including the hypothetical relations between the variables described in the previous section. In the chart, for every variable the hypothesized effect on both the profitability and on the effect of market share on profitabilty is decribed.
7
Figure 1: Conceptual model
For every variable, the first +/- indicates the assumed effect on profitability.
The second +/- indicates the assumed influence on the effect of market share on profitability.
2.1 Market share
To get a vision on the market share profitability relationship, the ways in which market share effects profitability will be examined.
From the early 70’s on the market share-profitability relationship is often studied and discussed. In reviewing the market share profitability relationship, it is argued by Prescott et al. (1986) that there is a direct effect of market share on profitability, but in the reversed way there isn’t a direct effect. This finding is widely accepted and studies that examine the market share profitability relationship focus therefore on the effect of market share.
Several studies (Sheperd 1970, Gale 1972, Buzell et al., 1975, Jacobsen, 1988) argue that increasing market share leads to a certain degree of market power which can be used to enforce favorable market conditions. Dominating the market generates a good bargaining position to demand for vendor allowances, set prices and extract concessions from channel members (Szymanski, Bharadwaj and Varadarajan, 1993). The market power allows the higher market share firm to obtain lower input costs by enforcing discounts at his suppliers. Since the suppliers don’t want to lose the most significant firms in the industry to their competitors, they are almost forced to provide the discounts. At the same time, the higher market share allows the firm to set prices rather than acting as a price taker (Boulding and Stealin, 1990). By having market power, the firm is able to raise prices without losing a great deal of its customers, thereby stimulating total revenues. Summing this up means that having market power results on the one hand in lowering total cost and in addition it leads to higher revenues. Together this would result in the fact that a higher market share tend to lead to higher profit levels.
ROI Market share o Firm size ++ o R&D expenses ++ o Firm growth ++ Firm characteristics o Market growth +- o Industry Concentration ++ o Industry category Industry characteristics o Prior return Control variable
8
Secondly, an explanation of the effect of market share on profitability is the product quality assessment theory (Smallwood and Conlisk, 1979). Since the buyer has uncertainty about the product quality, he will look for a way to overcome this problem. It is argued that market share is a way to evaluate product quality. The reasoning behind this is that a firm obtains a higher market share by attracting more costumers. When a product is bought by a lot of different customers, this indicates that it can satisfy consumer preferences really well and therefore indicates superior product quality (Smallwood and Conlisk, 1979). At the same time it is argued that the market leaders are considered as the most reliable in terms of guarantee and assured quality. Since the continuation of these firms is less insecure, consumers are more willing to buy their products (Buzell et al., 1975).
A third factor explaining the influence of market share on financial performance is the quality of management. It is argued by Buzell et al. (1975) that it might be the case that management which is successful in gaining market share is also successful in achieving higher profit levels. This can be done in several ways. High quality management could reduce cost by increasing the efficiency within the work process thereby decreasing cost level or increase profits by making strategic sophisticated choices. It is argued that when a firm achieves a leadership position it is easier to retain and exploit that leading position than for the competitors to catch up (Buzell et al., 1975). High quality management would on the one hand be able to increase market share and at the same time increase profitability.
While the determinants of profitability are extensively studied, also the methods of measurement are discussed. Buzell et al. (1975) use return on investment (ROI) as a measure of profitability. This is conducted as the ratio of pretax operating profits divided by the sum of equity and long-term debt1. The use of ROI as measure seems adequate since prior research (Schoeffler at al. 1974; Gale and Branch, 1982) has indicated a high correlation between ROI, return on sales, and return on equity, which are other commonly used profitability measures. The paper by Laverty (2001) point out that ROI is the most common method used to measure profitability in the market share profitability relationship.
To measure market share, Buzell et al. (1975) use the dollar sales of a firm as a percentage of the total industry sales2. To then measure the effect of share on profitability, the ordinary least squares regression is the most common method used. Szymanski et al. (1993) only included the studies which use regression to estimate the effect, since they argue that this is the most common method that also provides the most useful information.
Where before the seventies, industry concentration was seen as the most important determinant of profitability, the PIMS project indicates that market share is actually more important in explaining profitability (Buzell et al. 1975). Most studies find a positive effect of market share on profitability (Shepherd, 1972, Gale, 1972, Schoeffler, Buzzell and Heany, 1974, Buzzell et al., 1975).
1
Return on investment was measured by relating pretax operating profits to the sum of equity and long-term debt. (Buzell et al., 1975)
2
The market share of each business is simply its dollar sales in a given time period, expressed as a percentage of the total market sales volume. (Buzell et al., 1975)
9
To provide an estimate of the central tendency of the market share effect on profitability, Szymanski, Bharadwaj and Varadarajan (1993) performed a meta-analysis on two hundred seventy-six market share findings from forty-eight studies. They find that on average, market share has a significant positiveeffect of 25.9% on business profitability. But it is said that this effect is moderated by modeling, sample and measurement factors. In their study Jacobsen and Aaker (1985) indicate that the various PIMS studies have suggested that a 1% increase in market share would generate a 0.5% increase in ROI.
Despite the arguments in what way market share could positively influence profitability, there is an ongoing debate if market share can be used as a valid predictor. Especially the explanation of the quality of management is subject to criticism. While the correlation between the two factors is undeniable according to Laverty (2001), the causal relationship is often questioned. According to Rumelt and Wensley (1981) the effect of market share has been overstated and contains of a strong stochastic component. But at the same time they argue that market share is a valid predictor of return on investment. This comes down to the fact that market share is a strong and useful measure of financial performance, but it should not be the sole strategic content.
In their research Fraering and Minor (1994) find on average a weak negative relationship between market share and profitability. The results differ significant among the different industries they study. Where in home building and manufacturing markets a positive effect is found, the opposite effect is found in the banking and financial markets. They argue that the relationship between market share and profitability is too weak to justify a goal of achieving high market share to stimulate financial performance. In addition, the validity of market share as a predictor of profitability is questioned by Prescott, Kohli & Venkatraman (1986). They distinguish three categories of which one category with mature and declining environments, one category including fragmented standard products and a final category containing of exporting and emerging environments. In the first category they mainly find a direct effect of share on profitability, in the second category they find both a direct and spurious effect and the third one has predominantly a spurious relationship. They argue that only in the first category with a significant and strong effect, market share as a goal to improve profitability is appropriate.
Jacobsen and Aaker (1985) find that a large part of the effect is due to the influence of some third factors on both market share and ROI. They argue that management of a firm could be lucky in realizing both a big market share and high profit levels. Boulding and Staelin (1990) confirm these findings and find that industry structure and short term changes in market share strongly influence selling prices and costs. They state that the preliminary observed positive market share profitability relationship is probably due to managerial efficiency, returns to innovation or risk, and other to the researcher unobservable factors that correlate with both market share and firm profits.
10
Laverty (2001) points out that despite the earlier mentioned criticism, the dominant finding of prior research is that there is a positive relationship between the two factors. In his retrospect on the PIMS studies and related positive findings, Buzell (2004) refutes the critical notes of previous studies. By pointing out the model of serial correlation including the lagged dependent variable as an independent variable in studies by Jacobsen (1988; 1990; Jacobsen and Aaker,1985), he argues that the critical notes are based on use of an incorrect model. He argues that the positive effect of market share on profitability found in earlier studies are actually correct. Jacobsen and Aalker (1985) state that including prior return in cross sectional studies facilitate to control for the impact of past factors and in that way reduce the potential for a bias and remove the spurious effect. This would control for the impact of past and unobserved factors.
The magnitude of the effect of market share on profitability will however depend on other variables affecting both profitability and the effect of market share. To get a vision on the determinants of profitability, the meta-analysis on three hundred twenty studies by Capon, Farley and Hoenig (1990) is used. In their paper they sum up the variables which are mostly used in the literature to account for changes in profitability. Since the number of factors they include in the meta-analysis is enormously, it is impossible to include all these factors is a single study. Capon et al. (1990) state that, given the large number of potential explanatory factors and relatively limited data bases, meta-analysis might be the only feasible way to sort through alternative explanations in the existing literature. However, by including the most important variables together with a multiannual return period and variables controlling for a spurious effect, a regression analysis can be performed.
Based on the theory, the hypothesis is that a higher market share would result in higher profitability. This leads to the following hypothesis:
HO: A higher market share will not result in higher profitability H1: A higher market share results in higher profitability
The effect of both the firm and market characteristics on profitability and on the effect of market share on profitability will be discussed hereafter.
2.2 Firm characteristics 2.2.1 Firm size
From Hall & Weiss (1967) it becomes clear that bigger size tends to result in higher profit rates. The reasoning they give is that the bargaining power of large share firms is reinforced by overall firm size. Larger companies can enforce beneficial market conditions in terms of vendor allowances, influencing sales prices and other favorable actions. Buzell et al. (1975) argue that with a bigger market share a firm will attain more efficient methods of operation thereby benefit from economies of scale. Closely related to this is the explanation of the experience curve. According to Buzell et al. (1975) total unit production cost tends to decline by a constant percentage for every doubling of a company’s output. This suggests that by repetitive performing operations a more efficient way of working is introduced thereby reducing total costs. Bradley (1972) argues that the ability for a firm to exploit the benefits of a patent or invention increased with firm size.
11
Hall and Weiss (1967) state that firm size measured in asset terms is superior to measures in sales or employment, since the limitations of a firm are in financing big amount of assets. Next to this, they argue that assets are superior to equity, since it’s the size of the total capital that determines the opportunities available. From Bradley (1972) it becomes clear that the variables by which larger market share leads to higher profitability are intensified when firm size increases. The hypothesis is that bigger firm size will result in higher profit levels and by adding the size variable, the effect of market share of profitability increases.
2.2.2 Research and Development
From Johansson and Lööf (2008) it becomes clear that persistent R&D spending is positively influencing profitability. They argue that by performing R&D, more labor knowledge is generated which can be used to produce in a more efficient manner. At the same time, Szymanski et al. (1993) state that R&D expenditures are made to develop new and innovative products to seduce costumers from competitors. The likelihood that new goods can satisfy consumer needs better, suggest that more revenues are generated by developing a new product or service. So if efficiently used, R&D expenses affects profitability in a positive way by both driving production cost down and increasing revenues.
The positive effect of R&D expenses on market share is argued by Szymanski et al. (1993). The reasoning behind this is that a firm can entice customers to try the new products, thereby convincing them with the superior product quality. Where in some industries R&D acts as an entry barrier, this effect would even be more present. In the meta analysis performed by Capon et al. (1990) only studies which measured R&D as ratio of sales are included as a determinant of profitability. The study by Buzell et al. (1975) also includes research and development expenses relative to sales.
Johansson and Lööf (2008) state that the positive effect of R&D is particularly strong for multinationals with persistent R&D expenditures. Since the reference group in this paper consists of this kind of companies, the hypothesis is that there is a positive effect of R&D spending on profitability and by adding the R&D term, the effect of market share on profitability increases.
2.2.3 Firm growth
The effect of firm growth rate on profitability is expected to be positive. Kurtz and Rhoades (1992) state that a high firm growth rate could indicate a superior product, which in its turn could result in high profits. When a firm grows faster than the market does, it indicates that this firm is able to satisfy the demands of the customers and thereby stimulating revenues. It becomes clear from Cho and Pucik (2005) that there is a path from firm growth to profitability to market value. They conclude thereby that there is a positive effect of firm growth on profitability. To measure firm growth they use the compound annual growth rate of total assets, total revenues and market capitalization.
12
It is argued that by outperforming the market in terms of growth, a firm will increase its market share, since a bigger part of the total revenues will be generated by this specific firm. When a firm is able to keep the cost level constant, it will increase its profitability to. The hypothesis is that a positive firm growth rate will have a positive effect on profitability and by adding this term, the effect of market share on profitability increases.
2.3 Market characteristics 2.3.1 Market growth
Markets experiencing high levels of growth can be characterized with high margins, growing demand, high levels of buyer spending, increased investment and higher productivity levels. At the same time, high growth markets can be associated with higher advertising costs and low or negative cash flow. The net effect of these opposite effects is that market growth rate and profitability are positively related (Szymanski et al. 1993).
The expectation is that the effect of market share on profitability decreases when the growth rate increases. The reasoning behind this is that in a rapidly growing industry, firms are willing to sacrifice current profits to increase their market share. This is based on the assumption that the higher market share will in the future increase profitability which more than offset the current loss (Bradley, 1972). Szymanski et al. (1993) argues that a high growth market has to deal with lots of new entrants, thereby decreasing the incumbents’ market share. In the meta analysis by Capon et al. (1990) all studies included market growth did that by means of a percentage. Some studies based growth rates on total market revenues whereas others did that on assets.
Based on the above reasoning it is assumed that a higher growth rate result in higher profit levels but at the same time market growth rate and market share would be inversely related. The hypothesis is that there is a positive effect of positive market growth on profitability, and by adding the market growth term, the effect of market share on profitability decreases.
2.3.2 Industry Concentration
A higher level of concentration is typically associated with a lower degree of competition, which results in higher profit levels. Domowitz, Hubbard and Petersen (1986) state that high concentration influences a firm’s strategy to eventually affect its profitability. The firm could attempt to control the output by colluding with other firms to drive up prices and profit. The extreme case of monopoly power would result in higher profits then under perfect competition conditions. Kurtz and Rhoades (1992) confirm these findings and argue that high concentration might result in joint profit maximization behavior.
It is argued by Gale (1972) that economies of scale are one of the determinants of industry concentration. The effect of cost reduction should therefore be stronger in highly concentrated industries than it would be in unconcentrated industries where economies of scale are not as contributory. By definition, the combined market share of al firms will be dispersed over fewer firms in a more concentrated industry. So the hypothesis for the model is that a more concentrated market leads to higher profit levels and including the concentration variable will have a positive effect on the market share profitability relationship.
13 2.3.3. Financial crisis
In their survey on more than thousand Chief Financial Officers in thirty-nine countries, Campello, Graham and Harvey (2010) find that the financial crisis caused a significant financial constraint on some firms. For these firms this led to cutting down investment, marketing, employment and technology expenses. Therefore some firm where forced to bypass positive NPV projects, causing the strength of future economic growth to reduce. From Heifetz, Grashow and Lindsky (2009) it becomes clear that by managing critical adaptive strategies, a firm can keep itself from being corralled by the forces that generated the financial crisis. In addition, Gale (1972) argues that small market share firms are more flexible and therefore better to adapt to changing market circumstances than the often more bureaucratic large share firms. The hypothesis is that the financial crisis will have a negative impact on both profitability and on the effect of share on profitability.
2.3.4 Industry category
Analysis by Buzell et al. (1975) points out that the importance of market share in effecting profitability differs considerably among industries. They argue that market share is more important for infrequently purchased products than for frequently purchased ones. The reasoning behind this is that infrequently purchased goods tend to be durable and of higher cost, which are often hard to evaluate by the purchaser. There is a bigger risk in making a wrong choice, so the purchaser is willing to pay a premium for assured quality. Another difference pointed out by Buzell et al. (1975) is that market share is more important when buyers are fragmented instead of concentrated. A likely explanation is that the fragmented buyers have a weak bargaining position thereby allowing profits for the large-share businesses.
Szymanski et al. (1993) state that the markets hare-profitability relationship is likely to be strongest for consumer businesses, weakest for industrial businesses, and moderately strong for a mixed group of businesses. This difference could be capturing variations in such factors as the nature of the selling process and the types of goods sold that can distinguish consumer from industrial businesses. Fraering and Minor (1994) find in the home building and manufacturing markets a positive effect of market share on profitability. But in the banking and financial sector the opposite effect is found, namely a small negative effect of share.
In this study, six different industry categories will be examined. This contains of the Energy, Materials & Utilities, Industrials, Consumer Health & Discretionary, Consumer Staples and Telecom & Information Technology group. Based on the findings of Szymaski et al. (1993) the hypothesis is that the effect is strongest for the Consumer groups. The Materials and Utilities and Industrials group are assumed to have the weakest effect. And a moderately strong effect is expected for the Energy and Telecom and Information Technology group. Since the Energy group is used as the reference group, the hypothesis is that both the Consumer groups will have a stronger effect. This would result in a positive value for these groups in the regression output. The Materials and Utilities and Industrials are expected to have a negative sign. There is no significant difference expected for the Telecom and Information Technology group.
14 2.4 Prior return
Jacobsen and Aaker (1985) argue that the explanation of a positive relationship between market share and profitability ignores factors like firm culture, access to scarce resources and luck, which might influence both market share and profitability. They argue that a way to control for these unobserved factors is trough the inclusion of previous years ROI in the model. This can act as an alternate to firm specific factors occurring in previous periods and influencing ROI and possibly also market share. This can help to reduce the potential bias in estimating the market share effect. At the same time it is argued that a disadvantage of including the prior return term is the relative lack of variation, which may not allow adequate variation to isolate the impact of the different factors. In their study Jacobsen and Aaker (1985) include both a prior return of one and two years before the current year. In the regression performed by Laverty (2001) the prior return over three before the current year is used.
Both studies find that the effect of market share decreases when the prior return term is included. Jacobsen and Aaker (1985) find for both the one and two term a significant effect of 0.54 and 0.06 in there regression on ROI. Laverty (2001) finds a significant effect of 0.55 of the prior return. Both studies find that the magnitude of the market share effect decreases by including the prior return variable. Based on these findings, the hypothesis is that prior return will have a positive effect on profitability, but by adding this term, the magnitude and significance of the market share effect will be decreased.
2.5 Non linearity
It is argued by Gale (1972) that it might be the case that there exist a curvilinear relationship between market share and profitability. This means that the increase in profitability could possibly be at an increasing or decreasing rate as market share increases. To control for the possible non linearity, Gale (1972) uses both a dummy variable of share interacting with itself and the quadratic market share. It is found that profitability increases at an increasing rate as market share increases. But it is argued that this effect could be due to the fact that market share is correlated with concentration and sales.
3. Method
In this section the data sources, sample selection and the construction of variables will be discussed. In addition the key characteristics of these variables will be discussed together with the regression model.
For this study a reference group of the two hundred fifty biggest companies around the world is selected based on operating revenues. The time period examined for this study is from the year 2000 till the year 20133. For every firm the total revenues, profit before taxation, total assets, R&D expenses, total equity and long term debt are collected for every year in the sample period.
15
The data is collected from the Osiris database. This database contains of extensive information on listed companies worldwide, including the key financial information and ratio’s.
3.1 Profitability
As earlier discussed, there are several ways to measure profitability. In this study it is measured in terms of ROI. This is done, because in most previous study this method of measurement is also used. By using the same method it is less complex to compare the outcomes and to make a statement about de validity.
The profit before tax will be used, since tax rates vary along countries. In this way, measurement errors can be prevented. The pretax profit level will be divided by the sum of equity and long-term debt to conduct the ROI term. This will be done for every year in the sample period. For the Financials group no data was available for equity and long term debt, so the ROI term couldn’t be conducted. This resulted in fewer observations for the ROI variable.
3.2 Market share
For every firm, the corresponding Global Industry Classification Standard (GICS) code is used to compare it to the right industry category. For the various industries, the total revenues are gathered by summing up all individual firms revenue. To estimate a firm’s market share in the industry, the firms’ revenues will be divided by the total industry revenues. This method is used for the whole period. Since the GICS code industry is of significant size, the market shares are relatively low. As can be seen in table 1, the maximum market share is less than 15%. But since this measure is used in several previous studies, it is seen as the most appropriate and therefore used.
3.3 Firm characteristics 3.3.1 Firm size
The firm size is based on total assets. The choice to do this is because it is the size of the total capital that determines the opportunities available. Since total assets are values of significant size in comparison with the other factors, there is a risk that these big values will have influence on the residuals. This could possibly lead to biased results. Therefore the natural logarithm of total asset is used to mitigate this risk. For every firm the total assets are collected for the whole sample period. The total assets are transferred into a natural logarithm to come to the firm size variable.
3.3.2 R&D
The research and development variable used in the model is a ratio of gross sales. Both the total R&D expenditures and the gross sales are collected for the whole sample period. To conduct the variable, the R&D expenses are divided by the gross sales per period. Not for every firm, the R&D expenditures where available, resulting in less observation for this term.
16 3.3.3 Firm growth
The firm growth rate is based on total revenues for the specific firm. For every firm the growth rate is calculated by dividing the change in revenues by the old standard. After that, the market growth rate is subtracted to attain the so called superior growth rate.
This is done since it is argued that it is more informative to see how the firm performs in comparison with its competitors, than just solely look at the own situation. If the market growth rate isn’t subtracted, a firm could have a positive growth rate because the market as a whole grew. By having a positive superior growth rate, it indicates that the firm outperformed the market in the context of growth.
3.4 Market characteristics 3.4.1 Market growth
The total market growth rate is based on revenues. For every GICS industry, all firms’ revenues will be gathered. The sum of all these revenues is used as total market revenue. Based on these market totals, a growth rate will be computed. For every year this will be done by dividing the increase in revenues by the old standard.
3.4.2 Industry concentration
The concentration rate will be measured by the Hirfindahl Index. This is the sum of the squared market shares of all the firms in the industry. For every single firm in the GICS industry, the revenues will be divided by the total industry revenues to estimate the firms’ market share. This value is squared and summed up to compute the concentration rate. Table 1 Descriptive statistics
Variable Nr of observation Mean Std. dev Min Max
ROI 2827 0.153 0.209 -1.789 2.798 Market share 3265 0.017 0.018 0 0.149 Firm Size 3262 17.786 1.481 9.068 21.855 R&D 2829 17.013 1.266 7.823 20.357 Firm growth 3016 0.067 0.827 -3.697 21.224 Market growth 3250 0.108 0.149 -0.256 0.974 Industry Concentration 3500 0.208 0.017 0.005 0.101
17 3.4.3 Financial crisis
Since the time period of 2000 till 2013 includes the global financial crisis of 2007 till 2011, a dummy variable is added to find out if the effect of market share on profitability was different during this time period. The crisis dummy contains of 1250 observations, which accounts for almost 36% of the total observations.
3.4.4 Industry group
As mentioned above, the GICS code is used to devide the companies into the six industry peer groups. In the sample, there where ten different GICS industries, which means four categories are merged. Given the fact that the ROI variable was missing for the Financials category, no separate dummy could be created. This resulted in merging the Financials category with the Consumer Staples category, which was the most similar in characteristics. For the Health Care, Information Technology and Utilities categories to less observations where available to conduct a representative dummy. Therefore these categories were merged with the most similar category based on total revenues, growth and concentration. For Health Care this was the Consumer Discretionary, for Information Technology this was Telecommunication Services and for Utilities this was the Material group.
This resulted in the dummy categories included in table 1.2. In performing the regression, the Energy market will be the reference group. Therefore this industry group isn’t labeled with a dummy name.
Table 2 Industry dummy categories
Industry group Dummy label Number of obs
Energy* - 39
Materials and Utilities Material 26
Industrials Industrial 39
Consumer Health and Discretionary Health 48
Consumer Staples Consumer 68
Telecom and Information Technology Telecom 30
Total 250
*reference group
3.5 Prior return
To calculate the prior return term, the average of the three years before the current year is used. To be able to compute the prior returns for the years 2000-2002, the ROI for the years 1997-1999 where needed. Therefore this returns where calculated in advance. As earlier discussed, this would control for the spurious effect of market share on profitability.
18 3.6 Quadratic market share
The relationship between market share and profitability shown in the scatter plot in Figure 2 is complex to interpret. From the figure a linear relationship isn’t obviously clear and the possibility of a curvilinear relationship can’t be excluded. It might be the case that a quadratic form of the market share variable is a better able to estimator to explain the variance in ROI. Therefore an alternative regression model is established where the market share variable is replaced by the squared market share variable.
Figure 2: scatter plot of ROI and market share
3.7 Regression model:
Based on al the variables discussed the following base model is conducted:
Where:
MS = Market share, FS = Firm size, RD = Research & Developtment, FG= Firm growth, MG = Market growth, IC = Industry concentration, PR=Prior Return
7 represents a dummy variable for the crisis period and , , , and are dummy variables
respresenting the industry categories.
To check whether the quadratic market share is a better estimator of the relationship, an alternative model is used. This is model is consistent with the base model except for the quadratic market share variable.
Table 3: Correlation matrix
Variable Label 1 2 3 4 5 6 7 8 9 1 ROI 1.00 2 Market share 0.03 1.00 3 Firm Growth -0.00 -0.02 1.00 4 Firm size -0.00 0.49 -0.06 1.00 5 R&D -0.26 -0.05 -0.02 0.40 1.00 6 Market growth 0.12 0.02 -0.01 -0.06 -0.05 1.00 7 Concentration 0.08 0.26 -0.01 -0.12 -0.09 0.37 1.00 8 Prior return 0.80 0.04 -0.02 0.08 -0.22 0.05 0.04 1.00 9 Quadratic 0.02 0.92 -0.01 0.38 -0.02 0.03 0.27 0.01 1.00 -2 -1 0 1 2 3 0 .05 .1 .15 Marketshare
19
Several tests are performed to justify the use of an OLS regression. The correlation matrix reported in table 3 doesn’t cause a concern of multicollinearity. As can be seen from the matrix, there is no significant correlation between two explanatory variables. This is except for the high correlation between market share and quadratic market share. But since those two variables will not be included at the same time, this will not cause a problem. In the regression, robust standard errors are used to control for heteroskedasticity.
4. Results and Analysis
In this section the results of the OLS regression will be discussed and analyzed on. The outcomes will be discussed in conjunction of the theoretical expectations. The different variables are used in several ways to come to the best model.
The results of the different regression model are shown in table 4. In the first model, next to the market share variable, the firm and market characteristics are included expect for the industry dummies. In contrast to the expectation of a positive effect, there is a significant effect of -0.99 of market share on ROI. With a significant effect of 0.03 for firm size, 0.17 for market growth and 0.70 for industry concentration, these variables have the expected positive effect. With a value of -0.09 the R&D variable is significantly negative, which is contrary to the expectation that R&D would positively influence profitability. In addition, there is no effect found of firm growth on ROI, which was expected to have a positive effect. Note however that this first model is only able to explain 9 percent of the variance in ROI. The difference between the first and the second model is the addition of the prior return variable, which causes the explanatory power to increase from 9 to 66 percent. The prior return variable is highly significant and has a value of 0.82. The positive value of prior return matches the expectation. In this second model, the market share effect becomes 0.12 and as expected the significance decreases. With an effect of 0.13 for market growth and 0.17 for industry concentration both magnitudes decrease by including the prior return, whereby industry concentration isn’t significantly contributing anymore. In addition, the effect of R&D slightly increases to -0.02. The firm growth variable remains zero in the second model. While most of the effect of the explanatory variables decreased, it seems that the prior return term adopt a big part of these effects.
In the third model the alternative model is used, where the market share variable is replaced by the quadratic form. The quadratic market share has with a value of 2.28, a much larger positive effect on ROI, but this effect is, just like the market share variable in the second model, not significantly contributing in explaining ROI. The usage of the quadratic market share has, as expected, little influence on the other variables. The effect of market growth slightly increases to 0.14 and the effect of industry concentration slightly decreases to 0.14. The other variables are not influenced by the usage of the quadratic form. The explanatory power of the model remains constant at 66 percent in comparison with the second model.
20
Table 4: OLS regression on ROI
1 2 3 4 5 64 Market share -0.99*** (0.33) 0.12 (0.20) -0.08 (0.23) -0.04 (0.05) Quadratic market share 2.28 (2.06) 0.76 (2.12) Firm characteristics: Firm Size 0.03*** (0.01) -0.01 (0.01) -0.01 (0.01) 0.00 (0.01) -0.01 (0.01) R&D -0.09*** (0.01) -0.02*** (0.01) -0.02*** (0.01) -0.02*** (0.01) -0.02*** (0.01) -0.00 (0.04) Firm growth 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) Market characteristics: Market growth 0.17*** (0.04) 0.13*** (0.02) 0.14*** (0.02) 0.15*** (0.03) 0.15*** (0.03) 0.07*** (0.03) Industry Concentration 0.70** (0.35) 0.17 (0.24) 0.14 (0.24) -0.40 (0.28) -0.44 (0.28) 0.04 (0.05) Industry dummies: Material -0.02 (0.01) -0.02 (0.01) - Industrial -0.03** (0.01) -0.03** (0.01) - Health -0.02** (0.01) -0.02** (0.01) -0.035 (0.04) Consumer -0.02*** (0.01) -0.02*** (0.01) -0.03 (0.04) Telecom 0.00 (0.01) -0.00 (0.01) 0.10** (0.03) Crisis Dummy -0.03*** (0.00) -0.03*** (0.00) Prior return 0.82*** (0.05) 0.82*** (0.05) 0.80*** (0.05) 0.81*** (0.05) Constant -0.32 0.14 0.14 0.14 0.16 -0.16 0.09 0.66 0.66 0.67 0.67 0.55 N 2585 2506 2506 2506 2506 249
Robust standard errors in parentheses * Significant at 10% level
** Significant at 5% level *** Significant at 1% level
4
Outcomes based on the study by Szymanski et al. (1993)
5 Note that the effect of Health and Consumer are the same, since the study of Szymanski et al. (1993) uses
only three categories. A Industrial, Consumer and mixed category. It is argued that the Industrial and Material industry would belong to the Industrial category. Health and Consumer belong to the Consumer category and Telecom would belong to the mixed category.
21
In the fourth model, in addition to all market and firm characteristics, the financial crisis dummy and industry dummies are included. For market share purposes, the original method of measurement is used. Compared to the second model, the magnitude of the share effect decreases to -0.08 and thus became negative. In addition, the magnitude of the industry concentration also decreases and became -0.40. However, both variables remain not significant. The other variables remain constant in comparison with the third model. The explanatory power slightly improves to 67 percent.
With a value of -0.02 for Materials, -0.03 for Industrials, -0.02 for Health and -0.02 for Consumer for four out of five industry dummies a small negative effect is found compared to the Energy reference group. For the Materials group, this effect is not significant. The remaining Telecom group has a magnitude of zero, which is not significant. The findings for the Telecom group support the fact that no difference was expected. In addition, the negative deviation in accordance to the reference group for the Materials and Industrials is as expected. But the negative effect for Health and Consumer is contrary to the expectation for these groups. The crisis dummy, with an effect of -0.03, has the expected negative sign. For the fifth model, the alternative model including quadratic market share is used again. The magnitude of the share effect increases to 0.76, compared to the fourth model. However, market share is still not significant contributing in the explanation of ROI. By using the quadratic market share, the effect of concentration slightly decreased to -0.44 and firm size decreased to -0.01. However both factors are still not significant. The other variables are not influenced by the usage of the quadratic form. The explanatory power of the model remains constant at 67 percent in comparison with the fourth model.
The outcomes in the sixth model are of the meta-analysis by Szymanski et al. (1993). Since they combine and integrate results from two hundred seventy six findings in their analysis, this is expected to generate a representative result. The results by Szymanski et al. (1993) are closely related to the results of the fourth model. The market share effect is with -0.04, small and negative but not significant. Firm size is not included in their analysis, but is not found significant different from zero in this study neither. In contrast to the -0.02 effect of R&D found in the fourth model, Szymanski et al. (1993) find no effect for R&D. The market growth effect is with 0.07 smaller in the sixth model compared to the 0.15 effect in the fourth. In addition, the industry concentration effect of 0.04 in the sixth model is opposed to the effect of -0.40 found in the fourth model.
Note that Szymanski et al. (1993) divided the industries into three categories. They make use of an Industrial group, a Consumer group and a mixed group. They use the Industrial industry group as the reference category. It is argued that the Industrial and Material dummies in this study would belong to the Industrial group and therefore no effect is measured here. Both the Health and Consumer category would belong to the Consumer group, having an effect of -0.03. Finally, the Telecom industry would belong to the mixed category with an effect of 0.10.
With an explanatory power of 0.55 in the sixth model, less of the variance in ROI is explained than in the models two until five. This might be due to the fact that in the model by Szymanski et al. (1993) the prior return variable is not included.
22 5. Discussion and Conclusion
In this section, the purposes of this study will be summarized, together with the findings. A statement will be made on the hypothesis and the related implications will be discussed. Finally, the limitations of this study will be considered and suggestions for future research will be made.
This study revises the relationship between market share and profitability. While the effect of market share is often found to be positive, it is also subject to criticism. In this study, the top 250 companies based in 31 different countries and operating in 10 different industry categories are examined. To get better understanding on what the effect of market share is, this study examined the question what the sign and the magnitude of the effect of market-share on profitability is. The related alternative hypothesis this study is dealing with is the following:
H1: A higher market share results in higher profitability
The outcomes of the OLS regression models fail to support this hypothesis. The effect of share on profitability was found negative in two of the three models. The only significant effect of market share, of approximately minus one percent, was without including prior return. It could be very well the case that this effect is spurious, given the fact that there is no control for unobserved factors. Both the other effects, of which one was positive, are not significant different from zero.
In the two estimates by the alternative model, where quadratic market share was used to allow for a possible curvilinear relationship, a positive effect was found. However, these effects aren’t significantly contributing to the model and therefore fail to support the hypothesis. Based on this, it cannot be concluded that a higher market share results in higher profitability. These findings support the findings of Fraering and Minor (1994) and Szymanski et al. (1993) who find a weak negative effect of market share on profitability. The addition of the prior return term had the expected positive effect and increased the explanatory power of the model. This supports the findings of Jacobsen and Aaker (1985) who find that the effect of market share on profitability is mainly spurious. They state that there might be variables which both influence profitability and market share, which prior return controls for. It becomes clear from this study that there is indeed a significant effect of prior return, thereby concluding that there are important unobserved factors which influence profitability.
For firm size, the expected positive effect was only found in the first model. In the other models, no significant effect was found. These outcomes are unable to support the findings of Hall & Weis (1967) who find that bigger firm size results in higher profit rates. The same applies to firm growth. No significant effect of firm growth on profitability was found in any of the regressions. These outcomes don’t support the findings by Cho and Pucik (2005), who find that there is a path from firm growth to profitability to market value. This could be due to the fact that in this study the superior firm growth rate is used.
23
For R&D, opposed to the expectation, a significant negative effect was found in all the models. This is contrary to the findings of Szymanski et al. (1993) who find that R&D is positively related to profitability. Next to this, market growth has an expected significant positive effect in all the models. It thereby supports the findings of Szymanski et al. (1993). Industry concentration has the expected positive value in three of the five models, however only significant in the first model. This is insufficient to support the findings of Domowitz et al. (1986) that higher industry concentration would result in higher profits.
For the dummy categories, the sign of the crisis dummy is negative as expected. The Telecom, Materials and Industrials have the expected sign. But the negative effect for Health and Consumer is contrary to the expectation for these groups. These findings partly support the findings of Szymanski et al (1993) who state that the markets hare-profitability relationship is likely to be strongest for consumer businesses, weakest for industrial businesses, and moderately strong for a mixed group of businesses.
Implication for firm strategy:
In contradict to the findings by Buzell et al. (1975), the findings of this study suggest that the pursuit of gaining market share as a goal is not appropriate to increase profitability. It became clear that market share would not appear to be a key to profitability. The findings tend to support the findings by Jacobsen and Aaker (1985) who state that the effect of market share is spurious, which means that both profitability and market share are influenced by some third unobserved factors.
Based on these findings, having a strategic focus on gaining market share, could be detrimental for the continuation of the firm. Efforts to keep up market share, for having a certain rank or being the market leader could lead to a significant cost increase without increasing profitability. More emphasis should therefore be placed on other key areas like product quality, management effectiveness and efficiency.
However these findings do not implicate that market share is not an important factor in business performance. While it shouldn’t be used as a strategic goal, it could still be used as an indicator of performance. High market share together with high ROI implicates that management has a strategy which proved to be successful and should therefore be followed.
24 Limitations and directions for future research
The limitations of this study should be borne in mind when interpreting the findings. The meta-analysis on the determinants of financial performance by Capon et al. (1990) indicates, next to the most important ones used here, several other variables that could possibly influence profitability. Given the fact that the number potential determinants of profitability is enormously, it is almost impossible to include all of them, thereby having the risk of omitted variables. Because the data in this study is gathered from only one database this could lead to biased results. It might very well be the case that the key variables in other databases are conducted in a different way, thereby leading to slightly different results. While there is a significant effect found of prior return on profitability, it is not investigated where this effect comes from. It could be the case that the prior return term controls for important unobserved factors and therefore explains the variance in ROI. However it is not clear if that is the case.
While it becomes clear by this study that market share can’t be used as the main determinant of financial performance, additional research should be done on this topic. Clearly, there are factors which influence the causal structure that are difficult to model. Research should try to overcome this problem by investigating the effect in a different way. While the prior return variable is found to be very important in explaining the variance in profitability, the relationship between the two factors should be further examined.
25 6. Reference list
Boulding, W. & Staelin, R. (1990). Environment, market share, and market power. Management
Science, 36(10), 1160-1177
Buzzel, R. D. (2004). The PIMS program of strategy research. A retrospective appraisal . Journal of
Business Research, 57(5), 478-483
Buzzel, R. D., Gale, B. T., & Sultan, R. G.M. (1975). Market share-a key to profitability. Harvard
Business Review, 53(1), 97-106
Campello, M., Graham, J. R., & Harvey, C. R (2010). The real effect of financial constraints: Evidence from a financial crisis. Journal of Financial Economics, 97(3), 470-487
Capon, N., Farley, U. J., & Hoenig, S. (1990). Determinants of Financial Performance: A Meta-Analysis.
Management Science, 36(10), 1143-1159
Cho, H-J., & Pucik, V. (2005). Relationship between innovativeness, quality, growth, profitability, and market Value. Strategic Management Journal, 26 (6), 555-575
Domowitz, I., Hubbard, G. R., & Petersen, B. C. (1986). Business Cycles and the Relationship between Concentration and Price-Cost Margins. The RAND Journal of Economics, 17(1 ), 1-17
Fraering, M. J., & Minor, M. S. (1994). The industry specific basis of the market share–profitability relationship. Journal of Consumer Marketing, 11(1), 27-37
Gale, B. T. (1972). Market share and rate of return - The Review of Economics and Statistics, 54( 4), 412-423
Hall, M., & Weiss, L. (1967). Firm size and profitability- The Review of Economics and Statistics, 49(3), 319-331
Heifetz, R.,, Grashow, A., & Linsky M. (2009). Leadership in a (permanent) crisis. Harvard Business
Review, 87(7/8), 62-69
Jacobsen, R. (1988). Distinguishing among competing theories of the market share effect. Journal of
Marketing, 52(4), 68-80
Jacobsen, R., & Aaker, D. A. (1985). Is market share all that it's cracked up to be? Journal of
Marketing, 49(4), 11-22
Johansson, B., & Loöf, H. (2008). The Impact of Firm’s R&D Strategy on Profit and Productivity. CECIS
Electronic Working Paper Series
Kurtz, R. D., & Rhoades, S. A (1992). A note on the market share-profitability relationship. Review of
Industrial Organization, 7(1), 39-50
Laverty, K. J. (2001) . Market share, profits and business strategy. Management Decision, 39(8), 607-618
26 Prescott, J. E., Kohli, A. K., & Venkatraman, N. (1986). The market share-profitability relationship: An empirical assessment of major assertions and contradictions. Strategic Management Review, 7(4), 377-394
Rumelt, R. P., & Wensley R. (1981). In search of the market share effect. Academy of Management
Proceedings, 2-6
Schoeffler, S., Buzell, R. R., & Heany, D. F. (1974). Impact of strategic planning on profit performance.
Harvard Business Review, 17(1), 137-145
Sheperd, W. G. (1972). The elements of market structure. The Review of Economics and Statistics, 54(1), 25-37
Smallwood, D. E., & Conlisk, J (1979). Product quality in markets where consumers are imperfectly informed. The Quarterly Journal of Economics, 93(1), 1-23
Szymanski, D. M., Bharadwaj, S. G., & Varadarajan, R. P. (1993). An analysis of the market share- profitability relationship. Journal of Marketing, 57(3), 1-18
