Does the environment influence the employment growth of SMEs?

Janssen, Frank

Introduction

Since the publication of Birch's work in 1979, an impressive number of studies have been devoted to the role of small and medium-sized enterprises (SMEs) in job creation. In a book published in 1987, Birch observed that most employment was created by a tiny proportion of companies (i.e. fast-growing companies). Several other studies from both Europe and America confirm this phenomenon (O'Farrell, 1984; Dunkelberg, William and Cooper, 1982; Storey and Johnson, 1987; McMullan and Vesper, 1987; Gallagher and Miller, 1991; OECD, 1999; Julien, 2000).

Growth has been measured on the basis of an impressive number of variables, but the two indicators that are most widely used in literature are employment and sales. We decided to limit the scope of this paper to employment growth. Apart from the fact that it is a measure of economic growth (Kirchoff, 1991), it can serve as an indicator of the entrepreneur's success and, for society as a whole, it represents a measure of the firm's economic contribution to the common good (Dunkelberg, William and Cooper, 1982). That is why this criterion has been used by many economists and sociologists, even though it seems that companies themselves prefer to measure their success in terms of sales growth (Hughes, 1998; Donckels, 1990), and that the latter criterion has been favoured by researchers in management sciences (Weinzimmer, 1993). Furthermore, according to Child (1973), employment is an appropriate criterion for measuring the size of an organization, as it is primarily human beings that are "organized". Finally, since managers generally wait for demand to stabilize before recruiting personnel, employment is, in theory, a less volatile measure of growth than sales (Delmar, 1997). In some European countries, such as Belgium, the stability of this criterion is reinforced by rigidities on the labour market, linked to restrictive social legislation. (1)

Two competing theoretical approaches designed to explain the causes of performance and growth have been developed in the field of management sciences (Weinzimmer, 1993). The first approach, which we can call the "external" model, studies the influence of environment on organizations. The source of this external perspective goes back to the industrial economy movement, the so-called structuralist movement, which postulates that industrial structures determine the conduct of firms and hence their performance (Julien and Marchesnay, 1997). According to this movement, the performance of a company should tend towards that of the industry as a whole under the effect of competition. The structuralist school has influenced the strategic thought movement referred to by Mintzberg, Ahlstrand and Lampel (1999) as the environmental school, which came to the fore primarily due to the work of population ecologists (Hannan and Freeman, 1977). This school maintains that the conditions found in a company's external environment are the principal determining factors in its survival. The second approach, in other words the "internal" model, is chiefly concerned with studying the internal characteristics of a company and the way in which an organization adapts to its environment and attempts to shape it. The initial source of this internal approach is to be found in the industrial organization movement, also known as the "behaviorist" approach and, ultimately, in strategic management. The industrial organization movement believes that an industry is made up of companies that not only pursue their own personal strategies, but also build up the basic structures of the industry, along with other companies, by means of these strategic decisions on matters such as investment, technology, markets and products (Julien and Marchesnay, 1997). The resource-based theory (Barney, 1986; Wernerfelt, 1984) is consistent with the internal approach movement. According to Lohmann (1998), studies that are concerned with the impact of entrepreneurs' own characteristics should come under this theoretical movement. By the same token, it is possible to say that economic models of human capital (Oi, 1983; Lucas, 1978) and entrepreneurial learning (Audretsch, 1994; Jovanovic, 1982) may be regarded as an "internal" type of approach.

Growth is a complex and multidimensional phenomenon (Weinzimmer, 1993), so it goes without saying that a purely external or internal type of approach will inevitably be over-simplistic. However, within the limited context of this paper, we have decided to concentrate on the external approach and specifically on the influence of environment-related factors on growth. (2) We are aware that this decision ignores the predictive potential of internal variables linked to the manager (Janssen, 2006), (3) the firm and its strategies, and does not take the interactions between different types of variables into account (Janssen, 2002).

The analysis of the relationship between environmental characteristics and company growth has already given rise to numerous empirical studies. However, the vast majority of research on growth has only studied the impact of a limited number of variables. Moreover, most of this work has a relatively weak theoretical basis. The concept of growth is only rarely justified in theoretical terms. This area of research is highly fragmented, accentuated by the fact that too much attention is paid to the manufacturing sector, and that there is considerable variety in terms of the time period studied and of the way in which growth is measured (Janssen, 2002). It is also regrettable that, with the exception of Swedish and British scholars, few European researchers have taken any interest in this issue.

Several authors (Grinyer, McKiernan and Yasai-Ardekani., 1988; Miller and Friesen, 1984) agree that the impact of a large number of variables needs to be tested simultaneously in order to arrive at a global and more realistic picture of the growth phenomenon. To our knowledge, no research project to date has attempted to provide an exhaustive list of all the independent variables examined by previous studies.

Based on a state-of-the-art study of the research on environment-related growth determinants, we formulate 15 hypotheses. These hypotheses use all the determining factors that we have identified in the literature on growth. Rather than making value judgments on the relative importance of particular variables, which would have left us with only a limited number of hypotheses, we will test all these variables, as several recent empirical studies have demonstrated the surprising importance of factors that may, at first sight, seem to be of minor interest (e.g. see Gartner and Bhat, 2000).

Environmental Impact Hypotheses

From an economic point of view, the environment corresponds to a set of exogenous determinants, i.e. pre-determined or set (Pearce, 1997). Mintzberg (1979) classified the environmental aspects identified by research on the basis of four major characteristics: generosity, dynamism, complexity and market integration.

The environment can be generous or hostile. Generosity represents the extent to which an environment is liable to stimulate sustained growth (Starbuck, 1976). A hostile environment, on the other hand, curbs growth. The environment may also be stable or dynamic. Dynamism is related to the degree of instability of a market. Dynamism is the result of unforeseen factors that it is not possible to plan for in advance, such as unpredictable changes in demand or competition, or rapidly changing technology. These factors give rise to uncertainty, which makes it more difficult for managers to process all the information (Dess and Beard, 1984). Moreover, the environment may be simple or complex. Complexity arises from the company's need to gather large quantities of information and facts about its products or customers. If this knowledge can be rationalized, in other words, broken down into sub-groups that are easy to grasp, the environment can be regarded as simple (Mintzberg, Ahlstrand and Lampel, 1999). According to Child (1972), complexity is due to the variety and range of activities carried out by an organization. Finally, a company's markets may be integrated or diversified. This can also be expressed as homogeneity or segmentability. According to Dess and Beard (1984), this aspect of the environment is contained within the notion of complexity. Environmental aspects can therefore be reduced to three categories: generosity, dynamism and complexity (Dess and Beard, 1984).

Generosity

Of all the environmental variables that influence growth and that have been considered by empirical studies, generosity is by far the most widely examined. These studies have particularly looked at variables such as sectoral growth rate, level of concentration, entry barriers, public aid, economic policies, degree of unionization, the crime rate and the appearance of the area in which the company is located, the proximity of university institutions, whether or not the company is based in a science or industrial park and the regional economic environment.

As a general principle, industry growth means that existing companies do not necessarily suffer if newcomers take a share of the market. This also reduces the likelihood of retaliation by these existing companies (Porter, 1980). Theoretically, a growth market offers more opportunities, especially for newcomers since, by definition, demand within the market is growing (Eisenhardt and Schoonhoven, 1990). Weinzimmer (1993) notes the existence of a positive relationship between the generosity of the environment and growth in sales, assets and employment. According to this author, this would tend to demonstrate that the growth of a company could simply be due to the fact that it belongs to a "generous" sector. Such an environment would allow companies to grow without having to acquire resources or market shares at the expense of their competitors. This positive effect of industry growth is confirmed by Audretsch (1995) and Audretsch and Mahmood (1994). Other studies highlight considerable sectoral differences in terms of firm growth rates (Brush and Chaganti, 1999; Storey, 1994b; Siegel, Siegel and MacMillan, 1993; Eisenhardt and Schoonhoven, 1990; Dunkelberg, William and Cooper, 1982). (4) Medium or high-tech sectors appear to have a higher percentage of fast-growing SMEs than other sectors (Calvo and Lorenzo, 2001; Philips and Kirchoff, 1989). High-growth firms would rather be found in growing sectors (Woywode and Lessat, 2001; Davidsson and Delmar, 2001). On the basis of theoretical arguments and empirical results, we would put forward the hypothesis of a positive link:

H1: the sectoral growth rate has a positive influence on the firm's growth.

A high rate of concentration and high entry barriers do, in principle, protect existing firms from the arrival of newcomers and should therefore stimulate their growth (Hamilton and Shergill, 1992). The intensity of competition restricts access to resources for a recently established business or a newcomer (Romanelli, 1989). A considerable degree of competition also implies lower prices and profit margins than in a more concentrated structure, which could have a negative effect on growth. However, less concentration should, in theory, allow more companies to grow (Eisenhardt and Schoonhoven, 1990). An American study observes a positive link between the concentration rate of a sector and growth in sales and employment (Weinzimmer, 1993). This is confirmed by Geroski and Toker (1996). Along the same lines, a Swedish research notes that there is a negative link between competition intensity and growth (Delmar, 1997). (5) We also presuppose a positive influence:

H2: the level of market concentration has a positive influence on growth.

Some researchers studied a specific aspect of concentration, i.e. the impact of entry barriers on company growth. Entry barriers are said to exist when, in a given market, companies are capable of maintaining monopolistic prices and profits without attracting newcomers. These entry barriers may be linked to the capital intensity of a sector or to the promotion or R&D expenditures of existing companies. The promotion expenditures of existing firms create expenditure thresholds below which promotion no longer has any impact (Comanor, Kober and Smiley, 1981) and increases the loyalty of consumers to the existing firms (Scherer, 1980). In order to build up a reputation, a newcomer would have to consider a disproportionately high outlay in order to attract consumers. Considerable R&D expenditure can also be an entry barrier (Comanor, 1967). This actually increases the initial outlay that newcomers have to make and increases the complexity of the knowledge base they have to acquire. In certain industries, R&D is used as a defensive weapon allowing companies to file patents that are not necessarily used. Companies that already have a footing in the market in question should benefit from these entry barriers and grow more rapidly than they would if these barriers did not exist. These barriers actually reduce the numbers of newcomers or prevent their entry to the market and allow these firms to make monopolistic profits. One study notes that there is a positive link between the scale of the entry barriers resulting from R&D and promotion, on the one hand, and sales growth (Weinzimmer, 1993). We formulate the same hypothesis for employment growth:

H3: the existence of entry barriers linked to capital intensity and R&D or promotion expenditure has a positive influence on growth.

Governmental support ought to help stimulate growth. A study conducted in Quebec reports that government subsidies, particularly in the R&D and export fields, have a positive effect on growth (Julien, 2000). Still on the subject of economic policies, high fiscal pressure, restrictive social legislation or difficult industrial relations, which are generally perceived as curbing growth, could have a negative effect on growth (Gibb and Davies, 1990). These variables do, as a general rule, correspond to hostile policies as far as businesses are concerned. We thus put forward the following hypotheses:

H4: obtaining public aid has a positive influence on growth.

H5: restrictive fiscal and social policies have a negative influence on growth.

The degree of unionization of the company or the sector to which the company belongs is also liable to affect growth. Wooden and Hawke (2000) note that a majority of Anglo-Saxon authors with an interest in the impact of unions on employment and growth in employment feel that the degree of unionization generally has a negative effect for three main reasons. First, salaries in highly unionized companies tend to be much higher than in those that are not. As a result, these companies would employ fewer workers than a similar company with less unionization. Second, the effect of unionization on productivity would tend to be negative or fairly low. Third, unions tend to favour the salaries of their members to the detriment of the employment situation. A fourth reason provides a possible explanation for this potentially negative link: unions seek to minimize opportunities for reducing the workforce, which is in turn liable to discourage recruitment. Several studies conducted in the United States, Canada, the United Kingdom and Australia observe that the employment growth rate in unionized companies is two to four percent lower than in non-unionized companies (Wooden and Hawke, 2000; Blanchflower and Burgess, 1996; Long, 1993; Leonard, 1992). Other authors also note that the degree of unionization in a sector has a negative influence on the growth of SMEs (Acs and Audretsch, 1990). Along the same lines, Grinyer, McKiernan and Yasai-Ardekani (1988) report a negative relationship between the manager's perception of pressure from the unions and growth. However, this negative perception must be treated with caution. It may equally well result from the manager's need to justify low growth by putting it down to exogenous factors over which he/she has no control, as from real growth inhibitors such as excessive pay demands, repeated production stoppages or a rigid and aggressive union policy (Grinyer, McKiernan and Yasai-Ardekani, 1988). On the basis of theoretical arguments and corroborative conclusions from empirical studies, we put forward the hypothesis of a negative link:

H6: a high rate of unionization has a negative influence on growth.

The location of a company is also liable to influence growth, be this with regard to the area itself, the immediate vicinity or the region in which it is situated.

As far as the immediate geographical environment of the company is concerned, a study (Gartner and Bhat, 2000) reports that the crime rate in a particular area, as well as its appearance in terms of upkeep and cleanliness of streets, pavements and buildings, have a significant correlation (negative and positive, respectively) with the growth forecasts of the company manager. These results confirm the results of other studies (Bull and Winter, 1991).

H7: the crime rate in the area where the company is located has a negative influence on growth.

H8: the appearance of the area where the company is located has a positive influence on growth.

In addition, the proximity of university institutions is also liable to have a positive effect on growth (Snuif and Zwart, 1994). This proximity allows companies easier access to scientific expertise and to the results of certain research programs, thus making it easier to market these research results (Colombo and Delmastro, 2002). Similarly, the fact that the company is based in a science or industrial park may also have a copycat effect of stimulating growth. These parks may be defined as public and/or private property initiatives that aim to promote the start-up and development of businesses by providing logistical, technical, administrative and/or managerial services, technology transfer and network development between the firms based in the park, as well as between these firms and universities or public bodies. These parks allow businesses to benefit from agglomeration economies associated with interactions between companies that are concentrated within a restricted space (Marshall, 1922). Unlike science parks, industrial parks focus less on high-tech activities or activities that are associated with the use of scientific results and are characterized by weaker or non-existent links with academic institutions or research centres (Colombo and Delmastro, 2002). A British study (Westhead and Storey, 1994) shows that location in a science park has a positive effect on growth. These results are confirmed by an Italian study that examines both science and industrial parks (Colombo and Delmastro, 2002). We have formulated identical hypotheses:

H9: the proximity of university institutions has a positive influence on growth.

H10: the fact that the company is based in a science or industrial park has a positive influence on growth.

As far as the regional geographical environment is concerned, it is possible to make a distinction between rural and urban areas. In a rural area, the resources required for growth, such as specialized production factors, may be harder to find than in an urban area. For example, a rural area is more likely to have shortages of specialized or managerial staff (O'Farrell and Hitchens, 1988). The population density in an urban area also provides a larger potential demand than would be the case in a rural area (Woywode and Lessat, 2001). On the basis of these arguments, we formulate the hypothesis that an urban location has a positive effect:

H11: the fact that the company is located in an urban area has a positive influence on growth.

Being located in a region that is considered more economically developed or more dynamic than another region could also influence growth. The extent of development of the means of communication within a region can have an impact on growth. In principle, the more developed the road infrastructure, means of transport and communication networks, the greater the effect on growth. Macroeconomic conditions can also vary from one region to another and some regions may experience lower economic growth rates and/or lower income levels that might inhibit company growth (Gabe and Kraybill, 2002). Some regions attract more economic activities than others. In addition, regions compete with one another to attract national or foreign investment. However, in Quebec, Julien et al. (1997) do not observe any systematic influence on company growth associated with being located in a particular region. On the basis of theoretical arguments, we put forward the hypotheses of positive links:

H12: a sufficiently developed road infrastructure, means of transport and communication networks have a positive influence on growth.

H13: the dynamism of the region in which the company is located has a positive influence on growth.

Dynamism and Complexity

Very few studies have examined the impact of the dynamism and complexity of environment on growth.

According to Weinzimmer (1993), dynamism, at least if it is the result of innovation, could be a source of growth opportunities for companies that already have a hold in the marketplace. In addition, growth makes it possible to reduce the uncertainties associated with market dynamism. However, empirical studies have found no significant influence of environmental dynamism on growth (Wiklund, 1999; Weinzimmer, 1993). However, on the basis of the theoretical arguments, we will assume that the many market opportunities available in a dynamic environment are likely to stimulate growth and, hence, that a positive effect should prevail.

H14: the dynamism of the environment has a positive influence on growth, where such dynamism is defined as a market that is experiencing rapidly changing technology.

Some researchers have observed that a complex or unstable economic environment and, in more specific terms, the perceived difficulty of predicting economic conditions, has a negative impact on growth (Grinyer, McKiernan and Yasai-Ardekani, 1988). We can assume that this complexity slows down the decision-making process. Based on this empirical result, we put forward the hypothesis of a negative link:

H15: a complex environment has a negative influence on growth, where such an environment is defined as being the degree of complexity associated with collecting data and/or facts regarding a firm's products or customers, or the difficulty of predicting economic conditions that affect the market. (6)

Methodology

Population, Sample and Representativeness

A database compiled by ING Bank (Internationale Nederlandse Groep) that includes all firms established in Belgium that have delivered their annual accounts to the Accounts Central of the Belgian National Bank was used in order to determine the population of SMEs to be analyzed. We have retained all firms that were active over the period studied (1994-2000) for which we have data on employment for 1994 and 2000 and which corresponded in 1994 to the definition of an SME given by the European Commission. (7) Insofar as numerous firms in Belgium have been created purely for fiscal reasons (8) and do not really undertake activities, we have eliminated firms that were already active in 1994, but that still employed less than five people in 2000.

On the basis of these criteria, the population was composed of 11,481 firms. We randomly selected 788 firms, while at the same time ensuring proportions of micro- (less than 10 people), small- (between 10 and 49 people) and medium-sized (between 50 and 249 people) firms identical to those of the total population. In order to allow a dynamic analysis, this size criterion was checked at the beginning of the period studied, in 1994. We also kept the proportions of firms from the three regions of the country (Flanders, Wallonia, Brussels) identical to those of the population.

Out of the 788 firms, 331 refused to participate in the survey and 186 were not available during the interview period. For 121 other firms, the telephone number in the database was incorrect or corresponded to a fax number. Our study therefore focused on a sample of 150 firms. According to Harris (1985), the size of the sample must exceed the number of predictors by at least 50. Our sample of 150 observations respects this rule. (9)

In order to determine the representativeness of our sample in relation to the original population, we compared the average growth (10) of the sample firms to the one of the population (11) using a bilateral t-test. One of the application conditions underlying this test on two independent samples is the homogeneity of variances (Howell, 1998). We first used Levene's test to check that there is no significant difference in the variances (F = 1.476 and sign. = 0.224), and then tested the difference between the averages of employment growth for the two groups. The results of the bilateral t-test (t = -0.823; d.f. = 11.479; sign. = 0.411) indicate that the average employment growth of the firms in our sample is not significantly different from that of the firms of the overall population.

As our sample was composed on the basis of size and regional location constraints characteristic of the population, it is no longer necessary to examine the representativeness of the sample in relation to the population with regard to these two criteria. Finally, we also examined the percentages of independent firms within the population and the sample. These are also identical (66.7%). These particular elements of comparison were chosen because they appear in the initial database.

Data Collection Method and Measure of the Dependent Variable

The data published by Belgian firms do not make it possible to test the vast majority of the hypotheses developed in our research. Hence, we opted for a telephone survey. (12) We first established a questionnaire consisting of closed questions that we had pre-tested on several SME managers. The managers of 150 SMEs were interviewed by phone in November 2001.

The value of the dependent variable was calculated using the initial database. The choice of an appropriate growth index has given rise to a number of theoretical discussions (Wooden and Hawke; 2000; OECD, 1998; Birch, 1986). As none of the proposed measures is neutral (Julien, Morin and Gelinas, 1998), we decided to use a simple measure, namely the relative variation "([E.sub.t] - [E.sub.t-1])/[E.sub.t-1])," as this is the most frequently used index in studies on growth determinants (Delmar, 1997). In our case, this measure reads: ([E.sub.2000] - [E.sub.1994]/[E.sub.1994]).

In order to carry out a logistical regression (see infra), these dependent variables were split into "high growth" (code 1) and "low growth, stagnation or regression" (code 0). We defined "high growth" as being growth above or equal to 50% over the period studied. A total of 34.3% of the firms in our sample can be considered as having undergone high growth. (13)

Previous studies differ enormously in terms of the time period studied. In order to identify irregular short-term tendencies and to allow for a reliable estimation of organizational performances, the time period studied should be at least five years (Weinzimmer, Nystrom and Freeman, 1998). On the basis of the constraints of our database, we have measured growth over a period of seven years, stretching from 1994 to 2000. (14)

So as to avoid static measures, when growth is essentially a dynamic phenomenon, we have excluded firms that were established during the period studied.

Data Processing

In order to test our hypotheses, we have carried out a binomial logistic regression. This method presents certain advantages in comparison to the standard multiple regression that is subject to more restrictive application conditions (Garson, 2001; Howell, 1998). (15) Among these advantages, we could draw particular attention to the fact that, contrary to standard regression, logistic regression does not presuppose a linear relationship between the dependent variable and the independent variables, and does not require a normal distribution of the variables. We had observed that our dependent variable did not present a normal distribution. The logistic regression also made it possible to integrate dichotomous or polytomous and metrical predictors into one single model. Each modality of an original variable gave rise to a dummy variable coded 1 if the characteristic was realized and 0 in the opposite case. In order to avoid a linear relation between the independent variables, for each original variable one of the binary variables created was excluded from the model. In the case of "filtering," in other words when part of the sample is not concerned by a question, we created a dummy variable composed of the firms not concerned.

Results and Discussion

Prior to the regression, we compared the growth averages of firms that had responded to our survey with those of the firms who had refused to respond by using a bilateral t-test. The growth averages for the firms that had responded to the survey were not significantly different from those of the firms that had refused to respond. We then compared the size, independence and regional location of the firms of the two groups using Pearson's [chi square] test. Whether the firm had responded or not to the survey is independent of its size at the start of the period, its independence or dependence and also of its regional situation.

The statistically significant results at the 5% threshold (p < 0.05) of the logistic regression analysis of employment growth against environment-related variables are given in Table 1. Only two out of 15 variable have a significant effect on employment growth.

In accordance with our tenth hypothesis, the fact that the company is located in a science or industrial park increases the chances of employment growth. This result confirms the conclusions of other European studies (Colombo and Delmastro, 2002; Westhead and Storey, 1994). Given a copycat effect, development of networks, easier access to a wide range of resources and/or scientific research, or agglomeration economies, these companies grow more rapidly. However, this result does have a potential double selection bias linked to the auto-selection processes and/or the selection process implemented by the body responsible for managing the park (Storey, 1998). Indeed, it is possible that only companies with reasonable growth prospects might opt to set up business in such a park. Furthermore, it may be that the organization managing the park might only select companies with a certain growth potential. Be that as it may, this type of initiative is the only environmental variable that has a positive influence on growth.

We have put forward the hypothesis that being located in a region that is regarded as more dynamic than another region could have a positive influence on growth. We measured the impact of regional location by means of objective and subjective criteria. The actual region in which companies are located (Wallonia, Flanders, Brussels) does not have any significant influence on their growth prospects, which is in line with the results of other studies (Julien et al., 1997). The macroeconomic conditions that prevail in a region and the measures taken by a region's decision-makers in order to attract capital or businesses do not therefore have a significant impact on growth, in contrast to more local initiatives such as the creation of science or industrial parks. Moreover, looking at Table 1, we are forced to recognize that if a manager perceives that his/her company is part of an economically dynamic region, this has a negative effect on employment growth. It is possible that these companies might tend to "rest on their laurels" on the basis of this perceived regional dynamism at the expense of the real growth of their company.

Both of these variables are linked to the generosity of the environment. Other variables associated with generosity--sectoral growth rate, level of concentration in the market, the existence of entry barriers, obtaining public aid, the perception of fiscal and social pressures, the crime rate and the appearance of the area in which the company is located, the proximity of university institutions, location in an urban area and the extent of development of the road infrastructure, means of transport and communication networks--do not have any significant effect. Finally, neither dynamism nor environmental complexity appears to be statistically significant predictors.

Over and above this observation, this tends to show that environment, at least if it is studied independently of other variables, only has an extremely limited influence on employment growth of SMEs. Other research that simultaneously examines the impact of internal and external variables and the coherence between variables indicates that environment does have a significant influence when studied in interaction with internal variables (Janssen, 2002). Therefore, these results lead us to doubt the ability of purely external approaches to explain employment growth. Our results also argue in favour of adopting an integrative model and an integrative approach in order to study growth. As far as employment growth of SMEs is concerned, this shows the limits of the explanatory potential of the structuralist theories or environmental school theories, such as population ecology of organizations.

Conclusions

The impact of a company's environment on its growth has prompted considerable empirical research. However, most of this work focused on the analysis of one or a few predictors. Furthermore, no research has provided an exhaustive list of all variables previously analyzed. We have tried to fill this gap and have tested the potential influence of 15 environmental variables on employment growth.

Our results show that only two factors linked to the generosity of the environment have a significant effect on employment growth--the fact that the company is located in a science or industrial park (positive) and the way in which the manager perceives the economic dynamism of the region in which the company is located (negative). Other variables associated with generosity and variables that are concerned with the dynamism and complexity of the environment do not represent statistically significant predictors of employment growth. This observation leads us to question the validity of purely external approaches to examining growth and to argue in favour of a method that includes both external and internal determinants and considers the fit between them.

The results of our research are subject to certain limitations. First of all, this study is solely concerned with individual companies. However, some organizations are likely to grow due to creation of franchises. We did not use groups as our analysis unit, so this type of growth was inevitably ignored. Furthermore, we measured growth on the basis of data relating to the start and end of the period. However, growth does not necessarily follow a regular pattern. The development process may in fact be full of ups and downs. Nevertheless, our study does not take this phenomenon into consideration, because it does not take into account intermediate data. Moreover, the type of survey conducted and the questions asked prevented us from obtaining longitudinal data for a number of variables. Finally, the methodology that we chose to follow--to test most of the variables simultaneously and not to exclude factors that might seem to be of minor importance at first sight--does theoretically dilute the potential contribution of the various predictors. Another research option that might help overcome this problem would be adopting a more selective approach based on the results of this research.

As far as future research paths are concerned, subsequent studies on the link between environment and employment growth could explore two different avenues, possibly even at the same time. Firstly, since our research showed that only one more local environmental variable (i.e. location in an industrial or science park) has a significant positive impact on growth, while more macroeconomic variables--at least if these are analyzed in isolation--do not produce significant results, it would seem appropriate to analyze the impact of more microeconomic variables on growth. Secondly, our results show that a purely external approach leads to results that are not particularly convincing. Other relatively isolated studies (Weinzimmer, 1993; Eisenhardt and Schoonhoven, 1990) observe that environment does have an influence on growth when its interaction with certain internal variables is studied. These interactions may, for example, be concerned with the fit between entrepreneurial variables and variables associated with the dynamism of the environment, between the latter and the organizational structure or between the aggressiveness of the company strategy and the generosity of the environment. We therefore feel that it is of paramount importance to adopt a method incorporating these interactions and extending beyond the external model.

Acknowledgements

This study was conducted with the financial support of the ING Bank.

Contact Information

For further information on this article, contact:

Frank Janssen, Holder of the Brederode Chair in Entrepreneurship, Universite catholique de Louvain, Louvain School of Management, CRECIS, 1, Place des Doyens, 1348 Louvain-la-Neuve, Belgium

Tel.: +32 10 47 84 28/Fax: +32 10 47 83 24

E-mail: janssen@poge.ucl.ac.be

References

Acs, Z.J. and D.B. Audretsch. 1990. "The Determinants of Small-Firm Growth in US Manufacturing," Applied Economics 22, no. 2: 143-53.

Audretsch, D.B. 1994. "Small Business in Industrial Economics: The New Learning," Revue d'economie industrielle 67: 21-39.

--. 1995. "Innovation, Growth and Survival," International Journal of Industrial Organization 13: 441-57.

Audretsch, D.B. and T. Mahmood. 1994. "Firm Selection and Industry Evolution: The Post-entry Performance of New Firms," Journal of Evolutionary Economics 4: 243-60.

Barney, J. 1986. "Strategic Factor Markets: Expectations, Luck and Business Strategy," Management Science 32, no. 10: 1231-41.

Bernard, P.-M. 1999. Regression logistique. Cours EPM-64312, Doctorat en epidemiologie, Universite Laval. http://w3.res.ulaval.ca/cours-epm-64312/Default.htm.

Birch, D. 1979. The Job Generation Process. Cambridge, MA: MIT.

--. 1986. "The Job Generation Process and Small Business." In P.A. Julien, J. Chicha et A. Joyal. La P.M.E. dans un monde en mutation. Quebec: Les Presses de l'Universite du Quebec.

--. 1987. Job Creation in America: How Our Smallest Companies Put the Most People to Work. New York, Free Press.

Birley, S. and P. Westhead. 1990. "Growth and Performance Contrasts between 'Types' of Small Firms," Strategic Management Journal 11, no. 7: 535-57.

Blanchflower, D.G. and S. Burgess. 1996. "New Technology and Jobs: Comparative Evidence from a Two-Country Study." Centre for Economic Performance Discussion Paper no 285, London School of Economics and Social Science, quoted by Wooden and Hawke (2000).

Brush, C.G. and R. Chaganti. 1999. "Business without Glamour? An Analysis of Resources on Performance by Size and Age in Small Service and Retail Firms," Journal of Business Venturing 14, no. 3: 233-57.

Bull, I. and F. Winter. 1991. "Community Differences in Business Births and Business Growths," Journal of Business Venturing 6, no. 1: 29-43.

Calvo, J.L. and M.J. Lorenzo. 2001. "Une caracterisation des entreprises manufacturieres espagnoles a forte croissance: 1994-1998," Revue Internationale PME 14 nos. 3-4: 45-66.

Child, J. 1972. "Organizational Structure, Environment and Performance: The Role of Strategic Choice," Sociology 6: 1-22.

--. 1973. "Predicting and Understanding Organizational Structure," Administrative Science Quarterly 18, no. 2: 168-85.

Colombo, M.G. and M. Delmastro. 2002. "How Effective are Technology Incubators? Evidence from Italy," Research Policy 31, no. 7: 1103-22.

Comanor, W.S. 1967. "Market Structure, Product Differentiation and Industrial Research," Quarterly Journal of Economics 81, no. 4: 639-57.

Comanor, W.S., A.J. Kober and R.H. Smiley. 1981. "Advertising and its Consequences." Pp. 429-39 in P.c.

Nystrom and W.H. Starbuck (eds), Handbook of Organizational Design, vol. 2. Oxford: Oxford University Press.

Dahlqvist, J., P. Davidsson and J. Wiklund. 1999. "Initial Conditions as Predictors of New Venture Performance: a Replication and Extension of the Cooper et al. Study." 44th World Conference of the International Council for Small Business, Naples, June 20-23.

Davidsson, P. and F. Delmar. 2001. "Les entreprises a forte croissance et leur contribution a l'emploi: le cas de la Suede 1987-1996," Revue Internationale PME 14, nos. 3-4: 163-87.

Delmar, F. 1999. "Entrepreneurial Growth Motivation and Actual Growth--A Longitudinal Study." In RENT XIII, Research on Entrepreneurship. London, November 25-26.

Delmar, F. 1997. "Measuring Growth: Methodological Considerations and Empirical Results." In R. Donckels and A. Miettinen (eds.), Entrepreneurship and SME Research: On Its Way to the Next Millennium. Aldershot: Ashgate.

Dess, G.G. and D.W. Beard. 1984. "Dimensions of Organizational Task Environments," Administrative Science Quarterly 29, no. 1: 52-73.

Donckels, R. (ed.). 1990. Les leviers de croissance de la P.M.E. Brussels: King Baudouin Foundation/Roularta Books.

Dunkelberg, W.G., C. William and A.C. Cooper. 1982. "Patterns of Small Business Growth," Academy of Management Proceedings: 409-13.

Eisenhardt, K.M. and C.B. Schoonhoven. 1990. "Organizational Growth: Linking the Founding Team Strategy, Environment, and Growth among U.S. Semiconductor Ventures, 1978-1988," Administrative Science Quarterly 35, no. 1: 504-29.

Gabe, T.M. and D.S. Kraybill. 2002. "The Effect of State Economic Development Incentives on Employment Growth of Establishments," Journal of Regional Science 42, no. 4: 703-30.

Gallagher, C.C. and P. Miller. 1991. "New Fast-Growing Companies Create Jobs," Long Range Planning 24, no. 1: 96-101.

Garson, G.D. 2001. "Logistic Regression." North Carolina State University, http://www2.chass.ncsu.edu/garson/ pa765/logistic.htm

Gartner, W.B. and S. Bhat. 2000. "Environmental and Ownership Characteristics of Small Business and their Impact on Development," Journal of Small Business Management 38, no. 3: 14-26.

Geroski, P.A. and S. Toker. 1996. "The Turnover of Market Leaders in UK Manufacturing Industry, 1979-86," International Journal of Industrial Organization 14: 141-58

Gibb, A. and L. Davies. 1990. "In Pursuit of the Frameworks of Growth Models of the Small Business," International Small Business Journal 9, no. 1: 15-31.

Grinyer, P.H., P. McKiernan and M. Yasai-Ardekani. 1988. "Market, Organizational and Managerial Correlates of Economic Performance in the UK Electrical Engineering Industry," Strategic Management Journal 9, no. 4: 297-318.

Hamilton, R.T. and G.S. Shergill. 1992. "The Relationship Between Strategy-Structure Fit and Financial Performance in New Zealand: Evidence of Generality and Validity with Enhanced Controls," Journal of Management Studies 29, no. 1: 95-133.

Hannan, M.T. and J. Freeman. 1977. "The Population Ecology of Organizations," American Journal of Sociology 82, no. 5: 929-64.

Harris, R.J. 1985. A Primer of Multivariate Statistics. New York, 2nd ed. New York: Academic Press.

Howell, D.C. 1998. Methodes statistiques en sciences humaines. Brussels: De Boeck Universite.

Janssen, F. 2002. "Les determinants de la croissance des P.M.E.: analyse theorique et etude empirique aupres d'un echantillon d'entreprises belges" (PhD dissertation, University Jean Moulin-Lyon 3).

--. 2006. "Do Managers' Characteristics Influence the Workforce Growth of SMEs?," Journal of Small Business and Entrepreneurship 19, no. 3: 293-315.

Jovanovic, B. 1982. "Selection and the Evolution of Industry," Econometrica 50, no. 3: 649-70.

Julien, P.-A. 2000. "Les P.M.E. a forte croissance : les facteurs explicatifs," Actes du Congres de l'Association Internationale de Management Strategique, Montpellier, May 24-26.

Julien, P.-A. and M. Marchesnay. 1997. Economie et strategies industrielles. Paris: Economica.

Julien, P.-A., M. Morin and J. Gelinas. 1998. "L'importance des P.M.E. a forte croissance au Quebec de 1990 a 1996," Cahiers de Recherche de l'Institut de recherche sur les P.M.E. Universite du Quebec a Trois Rivieres, 97-15-C.

Kalleberg, A.L. and K.T. Leicht. 1991. "Gender and Organizational Performance: Determinants of Small Business Survival and Success," Academy of Management Journal 34, no. 1: 136-61.

Kirchhoff, B.A. 1991. "Entrepreneur's Contribution to Economics," Entrepreneurship Theory and Practice 16, no. 2: 93-112.

Lambin, J.-J. 1990. La recherche marketing: analyser, mesurer, prevoir. Paris: Ediscience International.

Leonard, J.S. 1992. "Unions and Employment Growth," Industrial Relations 31 (Winter): 80-94.

Lohmann, D. 1998. "Strategies of High Growth Firms in Adverse Public Policy and Economic Environments." Babson Entrepreneurship Research Conference, Ghent.

Long, R.J. 1993. "The Effect of Unionization on Employment Growth of Canadian Companies," Industrial and Labor Relations Review 46 (July): 691-703.

Lucas, R.E. 1978. "On the Size Distribution of Business Firms," Bell Journal of Economics 9 (Autumn): 508-23.

Marshall, A. 1922. Principles of Economics, 8th ed. London: MacMillan.

McMullan, W.E. and K. Vesper. 1987. "New Ventures and Small Business Innovation for Economic Growth," R&D Management 17, no. 1: 3-13.

Miller, D. and P.H. Friesen. 1984. Organizations: A Quantum View. Englewood Cliffs, NJ: Prentice-Hall Mintzberg, H. 1979. The Structuring of Organizations: A Synthesis of the Research. Englewood Cliffs, NJ: Prentice Hall.

Mintzberg, H., B. Ahlstrand and J. Lampel. 1999. Safari en pays strategie, L'exploration des grands courants de la pensee strategique. Paris : Village Mondial.

OECD. 1998. "Les P.M.E. a forte croissance et l'emploi : evaluation des pratiques exemplaires des pouvoirs publics." Report presented to the Athens meeting, Paris, April 26-28. OCDE DSTI/IND/P.M.E.(98)11, quoted by Julien (2000).

--. 1999). "High-growth Firms and Employment." Report prepared for the Helsinki meeting of the SME working group, May 20-21, DSTI/IND/P.M.E.(99)6.

O'Farrell, P.N. 1984. "Small Manufacturing Firms in Ireland: Employment Performance and Implications," International Small Business Journal 2, no. 2: 48-61.

O'Farrell, P.N. and D.M. Hitchens. 1988. "Alternative Theories of Small-Firm Growth: A Critical Review," Environment and Planning 20, no. 10: 1365-83.

Pearce, D.W. (ed.). 1997. The MIT Dictionary of Modern Economics. Cambridge, MA: The MIT Press.

Philips, B.D. and B.A. Kirchoff. 1989. "Innovation and Growth Among New Firms in the US Economy," Frontiers of Entrepreneurship Research.

Porter, M.E. 1980. Competitive Strategy: Techniques for Analyzing Industries and Competitors. New York, The Free Press.

Romanelli, E. 1989. "Environments and Strategies of Organization Start-Up: Effects on Early Survival," Administrative Science Quarterly 34, no. 3: 369-87.

Scherer, F.M. 1980. Industrial Market Structure and Economic Performance. Chicago: Rand McNally.

Siegel, R., E. Siegel and I.C. MacMillan. 1993. "Characteristics Distinguishing High-Growth Ventures," Journal of Business Venturing 8, no. 2: 169-80.

Starbuck, W.H. 1976. "Organizations and Their Environments." Pp. 1069-1123 in M.D. Dunnette (ed.), Handbook of Industrial and Organizational Psychology. Chicago: Rand McNally.

Storey, D.J. 1994a. Understanding the Small Business Sector. London-Boston: International Thomson Business Press.

--. 1994b. "The Role of Legal Status in Influencing Bank Financing and New Firm Growth," Applied Economics 26, no. 2: 129-36.

--. 1998. "Six Steps to Heaven: Evaluating the Impact of Public Policies to Support Small Businesses in Developed Countries." Warwick Business School, Working paper 59.

Storey, D.J. and S.G. Johnson. 1987. Job Creation in Small and Medium Sized Enterprises, Vol. 1. Luxembourg: Office for Official Publications of the European Community.

Weinzimmer, L. 1993. "Organizational Growth of U.S. Corporations: Environmental, Organizational and Managerial Determinants" (PhD dissertation, University of Wisconsin-Milwaukee, Ann Arbor).

Weinzimmer, L.G., P.C. Nystrom and S.J. Freeman. 1998). "Measuring Organizational Growth: Issues, Consequences and Guidelines," Journal of Management Studies 24, no. 2: 235-62.

Wernerfelt, B. 1984. "A Resource-based View of the Firm," Strategic Management Journal 5, no. 2: 171-80.

Westhead, P. and S. Birley. 1995. "Employment Growth in New Independent Owner-managed Firms in Great Britain," International Small Business Journal 13, no. 3: 11-34.

Westhead, P. and D.J. Storey. 1994. An Assessment of Firms located on and off Science Parks in the UK. London: HMSO.

Wijewardena, H. and G.E. Tibbits. 1999. "Factors Contributing to the Growth of Small Manufacturing Firms: Data from Australia," Journal of Small Business Management 37, no. 20: 38-45.

Wiklund, J. 1999. "The Sustainability of the Entrepreneurial Orientation-Performance Relationship," Entrepreneurship Theory and Practice 24, no. 1: 37-48.

Wooden, M. and A. Hawke. 2000. "Unions and Employment Growth: Panel Data Evidence," Industrial Relations 39, no. 1: 88-107.

Woywode, M. and V. Lessat. 2001. "Les facteurs de succes des enterprises a croissance rapide en Allemagne," Revue Internationale PME 14, nos. 3-4: 17-43.

Frank Janssen, Holder of the Brederode Chair in Entrepreneurship, Universite catholique de Louvain, 1, Place des Doyens, 1348 Louvainla-Neuve, Belgium

(1.) Of course, it is possible to increase sales without recruiting additional employees (Delmar, 1999), for instance through subcontracting. However, such a decision would have a positive impact on employment at the macroeconomic level.

(2.) For an analysis of a complete causal growth model, see Janssen (2002).

(3.) In this other paper, published in the same journal, we concentrated on the impact of entrepreneurial variables on employment growth.

(4.) A few studies have found no impact of sectoral growth rate on companies' growth (Wiklund, 1999; Kalleberg and Leicht, 1991).

(5.) Two other studies observe no significant link between these variables (Kalleberg and Leicht, 1991; Einsenhardt and Schoonhoven, 1990).

(6.) See Grinyer, McKierman and Yasai-Ardekani (1988).

(7.) According to the European Commission's Recommendation of April 3, 1996 (OJEC, L 107/4, 1996), the following firms must be considered as SMEs:

--those employing less than 250 people; the number of people employed corresponds to the number of annual work units.

--those whose either turnover does not exceed 40 million EUR., or the annual balance sheet total does not exceed 27 million EUR.

--those that respect an independence criteria. Independent firms are those which are not owned as to 25% or more of the capital or the voting rights by one or several large firms. We have not used this criterion, given that one of our hypotheses presupposed that the fact that a firm is dependent on another firm would have a positive influence on the growth of the former.

The Commission also distinguishes medium-, small- and micro-sized enterprises. The small firm is that which employs less than 50 people. This respects the independence criterion defined above and for which either the turnover does not exceed 7 million EUR, or the annual balance-sheet total does not exceed 5 million EUR. A firm is considered to be micro-sized if it has less than 10 workers.

(8.) The fiscal regime for companies is in fact more advantageous than the regime for physical persons. As a result many people create "empty" firms only in order to deduce expenses. These empty shelves do, of course, not grow.

(9.) According to other authors (Bernard, 1999), a minimum of 10 observations per predictor is necessary. Harris (1985) underlines that this principle is not based on any empirical proof. Others suggest more liberal rules than Harris and consider that the number of observations must only exceed the number of variables by 40 (see Howell, 1998).

(10.) For the measure of this variable, see next section.

(11.) From which we withdrew firms that belonged to the examined sample.

(12.) The major advantage of this method in relation to personal surveys or by post is its rapidity. In comparison with the personal survey, it also presents a lower risk of bias linked to the person of the interviewer (Lambin, 1990). Finally, it allows for the immediate codification of the responses, thus reducing risks of error.

(13.) The cutoff point of 50% has been chosen in order to generate a proportion of high-growth firms (34.3%) important enough to generate significant results.

(14.) During this period, the average annual growth rate of the Belgian GDP in constant prices has been of 2.68% (Banque Nationale de Belgique).

(15.) This also represents certain advantages in comparison to the discriminant analysis that can also be used when the dependent variable is dichotomized. Apart from the fact that the discriminant analysis involves a normal distribution of the variables, it can give rise to "impossible" probabilities of success situated outside the 0-1 range (Howell, 1998).

Table 1. Statistically significant predictors of the binomial logistic
regression of employment growth on environment-related variables

Independent variables              Coeff.    SE     Wald

Hypothesis 10: location in a       0.799    0.451   3.143
  science or industrial park
Hypothesis 13: economically       -1.022    0.503   4.126
  dynamic region

Independent variables              D.F.     Sig.     Exp.
                                                     (b)

Hypothesis 10: location in a         1     0.076    2.224
  science or industrial park
Hypothesis 13: economically          1     0.042    0.360
  dynamic region

[chi square] of the model: 21.316 Sign.: 0.379 Degree of concordance
between the predicted values and the observed values: 69.9%