Volume 11 Issue 4 Article 1 12-31-2009 A longitudinal comparison of the growth factors of Slovenian fast growing enterprises Viljem Pšeničny Follow this and additional works at: https://www.ebrjournal.net/home Recommended Citation Pšeničny, V. (2009). A longitudinal comparison of the growth factors of Slovenian fast growing enterprises. Economic and Business Review, 11(4). https://doi.org/10.15458/2335-4216.1270 This Original Article is brought to you for free and open access by Economic and Business Review. It has been accepted for inclusion in Economic and Business Review by an authorized editor of Economic and Business Review. 265 ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009 | 265–283 * Chamber of Craft and Small Business, Celovška cesta 71, 1000 Ljubljana, Slovenia, Email: viljem.psenicny@ guest.arnes.si A LONGITUDINAL COMPARISON OF THE GROWTH FACTORS OF SLOVENIAN FAST GROWING ENTERPRISES VILJEM PŠENIČNY* ABSTRACT: Th e article presents the main features of Slovenia’s fastest growing companies and compares them with “gazelles” in the EU. Th e longitudinal survey presented connects with three other studies applying the same research method, namely studies employing the same questionnaire on growth factors that aff ect growing companies through to the criteria by which they were selected as growing businesses for the survey. Th e author notes that the growth factors which have an impact on Slovenian businesses and gazelles in the EU mostly do not show any signifi cant diff erences, and that these diff erences also did not change sig- nifi cantly over a 15-year period. Th is hypothesis is verifi ed by both statistical methods and the data mining method called machine learning from examples. Keywords: Entrepreneurship; Dynamic enterprises; Growth; Growth factors; Data mining UDC: 658.01:330.341.1 JEL classification: L26 1. INTRODUCTION Micro, small and medium-sized enterprises constitute the “heart” of the Slovenian econ- omy. Slovenia has over 117,000 micro enterprises with less than EUR 2 million in sales revenue (data for 2008), constituting almost 96% of all economic entities in the Republic of Slovenia. Th ere are only 4,976 small, medium and large fi rms (or 3.8%), while just 774 large companies have more than 250 employees (AJPES, 2009). Th is means that Slovenia is a country where mainly micro and small enterprises operate. Even in the current EU economy, micro-, small- and medium-sized enterprises represent 99.8% of all economic entities, employing 67.1% of the workforce; however, they generate 57.6% of total value added in the EU and, most importantly, represent the most dynamic part of the economy because in the past fi ve years, according to the Commission, they have created over 80% of all new jobs (http://epp.eurostat.ec.europa.eu). ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009266 In Slovenia and the EU, only some (about 5%) of companies grow at an above-average rate employ and generate the largest part of the growth of value added and national economic growth. Five thousand fast-growing, small companies in Slovenia (represent- ing 4.5% of all businesses) in the fi ve-year period from 2003 to the end of 2007 created 22,514 new jobs, i.e. 60% of all new jobs during this period; value added per employee increased almost three-fold while sales revenues rose by more than two and a half times. In fi ve years these 5,000 companies and sole entrepreneurs generated nearly one-fi ft h of the increase in sales revenue in Slovenia (EUR 4.4 billion from the total amount of EUR 26.4 billion) or 23.4% of the total rise in net value added in the country (a EUR 1.2 billion increase from a total of EUR 5 billion) (Pšeničny, 2008). Th e question remains: How do we in Slovenia stimulate the “propelling power”, the “en- gine” of entrepreneurship – the fast-growing dynamic enterprises that are the only ones generating economic growth and added value (as recognised by David Birch, 1987)? How can we create the conditions and opportunities to ensure the prosperity of the most dy- namic part of the economy? Challenged by this issue, we launched a long-term research project into the prerequisite conditions and possibilities of developing dynamic entrepreneurship in Slovenia. Th e examination of fast-growing companies and growth factors in Slovenia has an almost 20-year tradition. Th e fi rst survey was conducted by Jan Žižek in the early 1990s (Žižek & Liechtenstein, 1994) and the second by the author of this contribution (Pšeničny, 2003). Since 2002, growing companies in the context of the GEM research team in Slovenia have also been examined (Rebernik et al., 2008). Further, much research in recent years has been joined by the contribution of Rado Bajt (2008) who reviewed the impact of changes in growth factors over the previous fi ve years. 2. RESEARCH GOALS Th e underlying reasons for researching dynamic entrepreneurship in Slovenia are: (1) we believe that the Slovenian economy vitally depends on the successful growth of the most dynamic part of small enterprises which will manage to overcome the “growth pains”; (2) we wish to ascertain which external (environmental) and internal factors stimulate or impede the growth of dynamic enterprises in Slovenia; and (3) we hope to establish which factors are most relevant in identifying the potential of dynamic enterprises – the so-called gazelles – and their chances of success. Moreover, with this research we also seek to contribute to: (4) improving knowledge of the factors of dynamic entrepreneurship and their eff ects on dynamic entrepreneurs; (5) more successful and effi cient managing of the growth of dynamic enterprises; (6) devel- oping a testing expert system to identify dynamic enterprises and their more successful and effi cient administration and management; and (7) to shape governmental policy in relation to entrepreneurship, or infl uence the planning stage of the policy to promote entrepreneurship, in particular dynamic entrepreneurship, as a relevant creator of jobs and economic development. V. PŠENIČNY | A LONGITUDINAL COMPARISON OF THE GROWTH FACTORS OF SLOVENIAN FAST ... 267 3. DYNAMIC ENTREPRENEURSHIP, GROWTH FACTORS AND FORMATION OF THE RESEARCH HYPOTHESES We have restricted our study of entrepreneurship to dynamic entrepreneurship. Th is has proven to have played an exceptional macro-economic role and the growth of the most dynamic enterprises contributes crucially to the growth of national economies, social prosperity, job creation, and to technological progress and development, as well as creat- ing the highest added value. Dynamic entrepreneurship is defi ned in great detail within the framework of the theory of growth (Penrose, 1995), by models and factors of growth divided into environmental and internal ones (the enterprise and entrepreneur), by the motivation for growth (and harvest), by strategies of growth as well as by management systems and development of the organisation of enterprise. In the long run, growth means profi t – i.e. a harvest for the entrepreneur who has identifi ed and seized a market opportunity and developed, on the basis of his clear vision and harvest expectation, a proactive strategy of growth and organisation throughout all organisational stages up to corporate entrepreneurship (Tajnikar, 2000). Dynamic enterprises are led by dynamic entrepreneurs who create change and have an eff ect on the environment, are innovative and successful in the long run (as can be measured by fi nancial and non-fi nancial indices), and whose business strategies are competitiveness, internationalisation and globalisation. Th e examination of the determinants of growth of enterprises can be divided into three groups. Th e fi rst group mainly concerns the study of the eff ects of the environment on growth of the company, the second examines in detail the internal environment of dy- namic businesses, while the third deals with dynamic entrepreneurs and the entrepre- neurial-managerial team. If we go back to the theory of growth, we see that Penrose set the foundations for this division of the factors of rapid growth which on one hand high- lighted the external, environmental factors of growth (Penrose, 1995; 229) and where, on the other, within the internal factors of growth the emphasis is placed on the role of administrative organisation, which is critical for growth (ibid., 15), and the role of the entrepreneur and entrepreneurial management (ibid., 34-37, 44-47). In previous studies (Pšeničny, 2002, 30-38), we found evidence that the growth of (dy- namic) enterprises mostly depends on certain factors: (1) the business environment; (2) the entrepreneur and/or the entrepreneurial-managerial team and their capability; (3) the attitude of the entrepreneur and the enterprise to innovation, research and development activities, and introducing changes; (4) the strategy or model of growth and harvest; (5) the management system and business model; (6) the employees’ and the management of human resources; and (7) the fi nancing of growth. Th e factors of growth have external environmental (1) and internal components (2–7). Th e similarities and diff erences in the interplay of these factors and individual principles of dynamic enterprises in Slovenia were scrutinised and compared with dynamic enter- prises in the European Union (EU). In Slovenia, we already have dynamic enterprises ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009268 and dynamic entrepreneurs that can be categorised, according to the EU criteria, among the fastest growing dynamic enterprises in Europe. Some of them are listed among Eu- rope’s 500 gazelles. Our thesis is that dynamic enterprises in Slovenia emerge and operate with the same characteristics but face diff erent internal and external conditions that are relevant to the fast growth of enterprises in the EU. In order to accelerate enterprise growth and support dynamic entrepreneurship, we should at least provide conditions in the environment and within fast-growing enterprises similar to those which benefi t dynamic enterprises in Europe. If we identify these diff erences, we can stimulate those activities that should lead to similar conditions for dynamic entrepreneurs in the near future such as those currently enjoyed by European dynamic enterprises. Th erefore, our primary hypothesis is: (H) External and internal factors infl uencing the dynamic growth of Slovenian dynamic enterprises diff er signifi cantly from the factors aff ecting dynamic enterprises in Eu- rope at the start of the 21st century. To allow international comparability at more advanced stages we adopted factors and attributes aff ecting the growth and success of dynamic enterprises from European re- search (Roure et. al., 1999; Mei-Pochtler, 1999). Th e growth and success of dynamic en- terprises were measured according to seven standard criteria: the DaBEG1 index, the total revenue growth rate, the revenue profi t growth rate, the capital profi tability growth rate, the assets profi tability growth rate, and the profi t per employee growth rate. On this basis, we reshaped the primary hypothesis (H), applied it as a basic working hy- pothesis (H1), and analysed it by developing several working hypotheses concerning the diff erences between individual factors of growth. (H1) Th e growth of dynamic enterprises in Slovenia depends on factors of dynamic entre- preneurship that are characteristically diff erent from the factors in the EU. Th e results of verifying this hypothesis (H1) also help verify the primary hypothesis (H) of our research. Th e confi rmation or rejection of hypothesis (H1) is, in fact, relevant to future planning of the business environment and the way entrepreneurs handle business activities; how- ever, it does not provide an answer to a fundamental issue raised as part of the goals of this paper, i.e. how to recognise and identify a dynamic enterprise, or how to establish whether an enterprise has the potential for growth, on the basis of a minimum number 1 David Birch Employment Growth Index (Birch, 1987; 36-37), measuring the employment growth of the company: where z stands for the absolute number of employees in a given year (t) DaBEG = (zt – zt – 5)* ztzt – 5 V. PŠENIČNY | A LONGITUDINAL COMPARISON OF THE GROWTH FACTORS OF SLOVENIAN FAST ... 269 of attributes. Th erefore, we took a further step in our research and tested the following hypothesis: (H2) Some factors aff ecting the faster growth of dynamic enterprises are much more im- portant than others and thus enable a forecast of the success and growth of dynamic enterprises. Verifi cation of this hypothesis is not only useful for entrepreneurs who lead dynamic enterprises and for investors, but also for the policymakers who can establish the condi- tions for the faster growth of dynamic enterprises. 4. RESEARCH MODEL AND METHODOLOGY To verify the diff erences in growth factors between Slovenian and European dynam- ic enterprises, from the database of all enterprises in Slovenia we selected enterprises that met certain criteria and further checked them against the growth criteria specifi ed above. Th e criteria that were applied to select the most dynamic enterprises are same as the criteria applied in the selection of European dynamic enterprises – Europe’s 500 (GrowthPlus, 2001; Europe’s 500, 2008). To examine both hypotheses, we employed: (1) original data sets of three fundamen- tal studies (Žižek & Liechtenstein, 1994; Roure, 2001; Pšeničny, 2003); (2) the research model developed in the previous research (Žižek & Liechtenstein, 1994); and (3) the set of external-environmental and internal attributes identifi ed as signifi cant characteris- tics by researchers of European gazelles (Mei-Pochtler, 1999; 97-104). Th e basic data sets on the dynamic enterprise databases applied in our research are shown in Table 1. Th e six factors with 17 external-environmental attributes and 14 internal-environmental at- tributes are shown in Figure 1. TABLE 1: Basic data on the dynamic enterprise databases in the research SI Dynamic entrepreneurs (1989-1993) SI Dynamic entrepreneurs (1994-1999) EU Dynamic entrepreneurs (1994-1999) Žižek 1994 Pšeničny 2001 Roure and Pšeničny 2001 Average age of enterprises 7,1 9 23 Average age of entrepreneurs 43 41 43 Average volume of total revenues in mEUR 2,2 6,4 53,1 Aver.growth of total revenues % in the appl. term 105 386 318 Average no. of employees in the last year 26 98 754 Av.Total Rev.per employee in 000 EUR, last year 84,6 65,3 70,4 Aver. Employment growth in % in the appl. term 64,6 172 302 Sample size 150 175 93 Sources: Žižek and Liechtenstein (1994), Pšeničny (2003), Roure (2001) ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009270 FIGURE 1: Factors and attributes of the fast growth of dynamic enterprises in the EU Source: Adapted from Mei-Pochtler, 1999. To compile the descriptive data, opinions and points of view by entrepreneurs, we ap- plied a questionnaire developed for research on dynamic enterprises in Central and Eastern Europe in 1993 (Žižek & Liechtenstein, 1994) and in the fi rst research completed on European dynamic enterprises in 1995 (EFER, 1996). As this questionnaire did not cover certain questions and attributes, we amended the underlying questionnaire on the basis of test results obtained from a sample of 94 dynamic enterprises in 1999 by add- ing 14 questions that enabled us to analyse the entrepreneur’s motivation, business and harvest strategy, attitude to hiring consultants, and some others. However, the basic 87 questions were kept. We approached verifi cation of our hypothesis (H) by noting diff erences in the factors aff ecting dynamic enterprises in Slovenia and the EU. We approached verifi cation of the additional hypothesis related to the diff erences in the responses by studying dynamic entrepreneurs in these three research projects, and by an alternative method to establish causal (cause-eff ect) connections between the attributes of the enterprises; i.e., one of the contemporary artifi cial intelligence methods. For statistical analysis, we applied the t-test and the χ2 test to establish diff erences in separate samples, while for the analysis of cause-eff ect relations we applied a data mining method called machine learning from ex- amples, also known as inductive machine learning (Mitchell, 1997). Th e particular form used in our case was the induction of decision trees (Quinlan, 1986; Witten & Frank, 2005). V. PŠENIČNY | A LONGITUDINAL COMPARISON OF THE GROWTH FACTORS OF SLOVENIAN FAST ... 271 Th e analysis of data by machine learning is a fi eld of computer science dealing with the extraction of implicit, previously unknown and potentially useful information from databases (Witten & Frank, 2005). Th e key procedure of this methodology is machine learning which includes the automatic induction of decision trees, classifi cation rules, regression models and other types of models from data. Th e models derived with these techniques represent generalisations of the input data (or cases) and can be used for the classifi cation, prediction and explanation of explored phenomena. Th e best explored and most frequently used machine learning approach is learning from examples, also referred to as inductive machine learning. In this approach, ex- amples of problem situations are submitted to a learning system (a computer program) which induces a general description of the underlying concepts useful for problem solving. Th e resulting concept descriptions can take the form of decision trees or if- then rules. Learning examples can oft en be very naturally described with attributes and classes. Attributes represent features of objects from the considered domain, while class defi nes how an example with given attribute values is treated or classifi ed. A deci- sion tree corresponds to a set of if-then rules relating attributes with classes and can be used for classifi cation and predictions in the problem domain. Similar to this approach is a “what-if” analysis which has already been applied in predicting business develop- ment (Makridakis, 1990; Stevenson, 1998). Machine learning has been used to analyse enterprise growth factors (Filipič & Pšeničny, 2003) and is becoming increasingly use- ful for business forecasting (in CRM, Competitive Intelligence and Knowledge Man- agement) (Zanasi et al., 2007). In our study, we used the Weka machine learning soft - ware (Witten & Frank, 2005) that allows using various methods of machine learning on the same data. We varied the procedure of decision tree induction by changing the parameters so as to obtain several models for each particular decision problem: these models give a diff eren- tially detailed insight into the concrete problem and also diff er according to the accuracy of classifi cation. Th e transparency and interpretability of these models are features that generate a new level of quality compared with the results of statistical processing, which are normally a standard approach when studying the growth of enterprises (such as in Solymossy, 1998; Wiklund, 1998). Our analysis by means of decision trees comprised the 134 most dynamic enterprises in Slovenia in 2002 and 21 test dynamic companies in 2007. Out of 320 descriptive and numerical data items on dynamic enterprises in our database, a subset of data was selected for the analysis. We excluded the attributes not containing information po- tentially relevant to the prediction of enterprise growth, such as the company name, contact information, instructions on fi lling in the questionnaire etc. As a result, 158 data items were selected. However, some of these items were actually questions with more than one possible answer. To obtain clearer results in the data mining stage, these company attributes were transformed into multiple attributes with binary val- ues. ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009272 5. THE FEATURES OF DYNAMIC ENTERPRISES IN SLOVENIA AND DIFFERENCES BETWEEN SLOVENIAN AND EU GAZELLES In this section we briefl y present the results of the statistical analysis of diff erences be- tween Slovenian and EU gazelles and the importance of factors infl uencing fast growth among Slovenian gazelles in given years. We established that Slovenian dynamic enterprises have not changed considerably in the preceding 15 years (internal – environmental factors of growth); on the other hand, busi- ness, fi nancial and tax environments have changed, as has the attitude of the environ- ment towards entrepreneurs with more critical remarks being elicited from the dynamic entrepreneurs involved in our research in 2002 and 2008 than in 1994; however, they remained less critical than their European counterparts. When measuring the impact of individual features of dynamic enterprises on our growth criteria, we established that the DaBEG index of Slovenian dynamic enter- prises in the past strongly depended on favourable governmental regulations, the level of remuneration for a dynamic entrepreneur, the age of the enterprise’s equipment, the knowledge of the habits and behaviour of consumers, and the quality of the entrepre- neurial team. Th e growth of total revenues in dynamic enterprises depended on the company’s activity (the highest being in building and construction), favourable governmental regulations and administration, an orientation to foreign (non-European) markets, the source of suppliers (suppliers from Central and Eastern Europe), and planning of future invest- ments. Th e growth of profi ts from total revenue generated by dynamic enterprises was the high- est in the branch of engineering, and depends on the entrepreneur’s opinion on the level of corporate profi t tax: the profi t can grow from year to year if the entrepreneur consid- ers the tax rates reasonable. Likewise, the profi t increased if the entrepreneur had been receiving the highest compensation for their current work, if the competition in their branch was not strong, and if members of the managerial team contributed to the fi nanc- ing of growth. Higher total capital profi tability growth rates are found in enterprises in which the owner would set up an equivalent enterprise once again if they had the opportunity, the owner pays himself relatively low remuneration for their current work, the owner’s employees are suffi ciently qualifi ed for their work, and where the primary source of start-up capital (not the founding capital) was their own capital. Th e total assets profi tability growth is aff ected by problems in transportation and com- munications, social recognition or recognition by the environment, the origin of the enterprise (if founded by the entrepreneur), the business activity, the remuneration to the management, and the expectation of the harvest; whereby the growth of profi tability is V. PŠENIČNY | A LONGITUDINAL COMPARISON OF THE GROWTH FACTORS OF SLOVENIAN FAST ... 273 adversely aff ected by high remuneration to the management, a neutral attitude to work- ers’ participation in the management, and by the entrepreneur himself if he founded the enterprise merely to implement his idea and provide for his existence. We also found that the responses of Slovenian and European dynamic entrepreneurs dif- fer characteristically in questions concerning a stimulating innovative environment and the transfer of R&D achievements to dynamic enterprises, as well as in the expansion strategies to international markets, the tax bonus for the co-ownership of employees and their participation in the profi ts, and all factors of the fi nancial environment (accessibil- ity of venture capital, the effi ciency of fi nancial markets, and taxation on retained profi ts and re-investments). For other environmental factors, we found either no considerable diff erence or no diff erence at all. In spite of this, we can assume that the diff erences in the environmental impact on the growth of enterprises in Slovenia and Europe are important, which supports our hypoth- esis regarding the diff erences existing in the business, fi nancial and fi scal environments of dynamic enterprises between Slovenia and the EU. For the internal growth factors, we found several characteristic diff erences, mainly in the entrepreneur’s attitude to building up a solid organisation. Dynamic enterprises in Slovenia are in their early developmental stages and most of them have not entered the professionalisation stage. However, due to the large diff erences in the enterprise histories of Slovenian and European gazelles this is quite unlikely to point to typical diff erences in the entrepreneurs as the other three features of the EU gazelles (the attitude to inter- nal entrepreneurship, leadership, and a clear vision) are equally present in Slovenian dynamic enterprises. Th e hypothesis on diff erences emerging with this factor cannot be confi rmed or rejected on the basis of these tests. Signifi cant diff erences between Slovenian and European dynamic entrepreneurs and enterprises were found in the attitude to innovation and in business strategies. In most answers to these two factors the answers diff er greatly, leading us to conclude that the hypotheses on diff erences in these two factors can be confi rmed. In questions related to the management system, there were bigger diff erences with re- spect to the features of the management system that point to an “organisation that pro- motes growth and innovation”, and fewer diff erences in the entrepreneur’s attitude to the remuneration of employees and the management. Th e hypothesis on diff erences in this factor cannot be fully rejected or confi rmed. Th e situation is similar regarding the diff erence in relation to the European dynamic entrepreneurs in the attitude to employees. In particular, diff erences are seen in the re- sponses to questions on the loyalty and commitment of employees to the dynamic en- terprise, while with questions related to work conditions, promotion, and possibilities of participation in a growing enterprise we fi nd more similarities than diff erences. Our hypothesis on diff erent eff ects of this factor can be rejected, with some reservation. ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009274 Th e greatest diff erences were found in the respondents’ opinions on fi nancing the grow- ing business; however, due to the diff erent size and corporate life of these enterprises (and thus diff erent phases in the corporate development and diff erent phases of fi nanc- ing the enterprise), we cannot cogently confi rm the hypothesis on diff erences between EU and Slovenian dynamic enterprises regarding fi nancing and fi nancial manage- ment. Looking at the overall results of our statistical analysis, we may conclude that there are signifi cant diff erences between Europe and Slovenia in factors aff ecting growth, prima- rily in: (1) the business environment; (2) the business strategies; (3) the attitude to inno- vation; and (4) fi nancing growth. On the other hand, there are no important diff erences in the attitude of dynamic enterprises to: (1) the employees in dynamic enterprises and (2) entrepreneurs themselves. However, on the basis of our analysis we cannot assess the diff erences in the scope of management which is, in fact, not developed yet in Slovenian dynamic enterprises. We verifi ed the diff erences with the machine learning method. 6. FINDING DIFFERENCES BETWEEN SLOVENIAN AND EUROPEAN GAZELLES WITH INDUCTIVE MACHINE LEARNING When developing decision trees by means of using the inductive machine learning method on examples, we fi nd that based on the examples of the 134 most dynamic enter- prises we can extract a number of rules by predicting numerically and non-numerically expressed attributes of dynamic enterprises and their growth; these rules can help us defi ne the conditions for the fastest growth of dynamic enterprises. Our predictions will be much more accurate in the future if we “screen” the attributes of dynamic enterprises by using a questionnaire developed on the basis of our own knowledge of the attributes resulting from this research and incorporate it in a study of a still bigger number of suc- cessful dynamic enterprises from several countries. To illustrate the applicability of machine learning from examples, we present a decision tree for predicting the DaBEG index and planning the attitude of entrepreneurs to share- holders’ options in gazelles. Th e fi rst case is explained in full detail; in the second case only the fundamental information based on the decision trees is given. Example 1: Predicting the DaBEG index Th e calculation of the DaBEG index is shown in the footnote on page 4, with the classes for the DaBEG index taking the following ranges: • Class 1: DaBEG > 1000 (10.4% of the enterprises in the database); • Class 2: 200 < DaBEG ≤ 1000 (11.2%); • Class 3: 100 < DaBEG ≤ 200 (23.9%); and • Class 4: DaBEG ≤ 100 (54.5%). V. PŠENIČNY | A LONGITUDINAL COMPARISON OF THE GROWTH FACTORS OF SLOVENIAN FAST ... 275 The classification accuracy is still acceptable when higher than the share of the majority class. An example of such a decision tree to predict the DaBEG index is shown in Figure 2. The inner nodes are labelled with the attributes and the end nodes (leaves) with classes. Paths from the top node (root) to the end nodes (leaves) correspond to if-then rules. The classification accuracy of the decision tree on the 134 training data is 72.4% and on the test data it is 46.3% (transversal testing of the model obtained). FIGURE 2: Th e decision tree to predict the DaBEG index Source: Pšeničny (2003) Th is decision tree allows us to derive several rules to predict the DaBEG index; however, we only list the rules to predict the highest class or value of the DaBEG index above 1000 (such as in Birch’s “gazelles”). Class 1 IF (A99-0 = 0) & (A19 = 0) & (A75-4= 0) & (A97 = 0) OR (A99-0 = 0) & (A19 = 0) & (A75-4= 0) & (A97 = 1) OR (A99-0 = 0) & (A19 = 0) & (A75-4= 0) & (A97 = 4) OR (A99-0 = 0) & (A19 = 0) & (A75-4= 0) & (A97 = 5) THEN DaBEG > 1000 Here A99-0, A19 etc. denote the attributes extracted from the questionnaire. Th is formal representation tells us how to predict the highest values of the DaBEG index (DaBEG>1000). Written in natural language, such values of the DaBEG index can be found in dynamic enterprises that are: ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009276 (1) limited liability companies believing that the business environment could motivate them for higher growth, having effi cient cash management, and which do not plan new investments or to create new jobs; (2) limited liability companies believing that the business environment could motivate them for higher growth, having effi cient cash management, and planning new invest- ments but not creating new jobs; (3) limited liability companies believing that the business environment could motivate them for higher growth, having effi cient cash management, and planning new invest- ments and 50 to 99 new jobs in the coming fi ve years; and (4) limited liability companies believing that the business environment could motivate them for higher growth, having effi cient cash management, and planning new invest- ments and 100 to 199 new jobs in the coming fi ve years. Example 2: Predicting the employees’ stock option plans Th e factors underlying the fast growth of European dynamic enterprises also involve the inclusion of employees as co-owners of a dynamic enterprise. We also checked this attribute in the gazelles in our database. Possible replies (SOP) to the question, “What do you think about the possibility of the workers becoming shareholders in your company?” were: – SOP = 0: no, on no account (22.4% of the enterprises in the database) – SOP = 1: it makes no diff erence to me (4.5%) – SOP = 2: maybe it could work, but I won’t commit myself to it (29.1%) – SOP = 3: maybe it could work; I plan to undertake it (13.4%) – SOP = 4: they are shareholders already; I am satisfi ed (19.4%) – SOP = 5: they are shareholders already; I am not satisfi ed (3.7%) Th e decision tree to predict the dynamic entrepreneur’s attitude to the employees’ stock option plans is shown in Figure 3. It achieves a classifi cation accuracy of 66.1% on the training data and 45.1% on the test data. V. PŠENIČNY | A LONGITUDINAL COMPARISON OF THE GROWTH FACTORS OF SLOVENIAN FAST ... 277 FIGURE 3: Decision tree to predict employees’ stock option plans Source: Pšeničny (2003) Th is decision tree can, in the same way as the tree shown in Figure 2, be interpreted as follows: (1) Employees will not (SOP=0: “on no account”) be included in the shareholding struc- ture of dynamic enterprises where: 1.1 Employees have not become owners yet, and the entrepreneur has a two-year college degree, the prevailing strategy for growth is not globalisation, and in en- terprises where payment collection is causing the greatest diffi culties; 1.2 Employees have not become owners yet, and the entrepreneur has a two-year college degree, the prevailing strategy for growth is not globalisation, and in en- terprises that have not stated the greatest diffi culty in payment collection, but in tough competition with state-owned enterprises. (2) Employees will not (SOP=2: “maybe it could work, but I won’t commit myself to it”) be included in the shareholding structure in those dynamic enterprises in which the employees have not yet become owners, the entrepreneur completed a four-year col- lege or university education, the prevailing strategy for growth is not globalisation, their greatest diffi culty is other than payment collection, the main (5 on the 1 to 5 scale) reason for growth is the customer satisfaction approach of the employees, and their main competitors are those other than state-owned enterprises. (3) Employees will most probably (SOP=3: “maybe it could work, I plan to undertake it”) be included in the shareholding structure in dynamic enterprises in which the employees have not yet become owners, and are led by an entrepreneur with a college or university education, and who did not start up the enterprise due to their dissatis- faction with a previous business. ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009278 (4) In enterprises where employees are shareholders already and the entrepreneurs are satisfi ed with this (SOP=4: “they are shareholders already, I am satisfi ed”), employees will be invited to become shareholders in enterprises that are led by entrepreneurs with a higher education, and who apply the strategy of growth with globalisation or introducing new products into new markets. With more than 70 of such decision trees and on the basis of the data in the dynamic enterprise database described with 158 fi nancial and non-fi nancial attributes, we found that dynamic enterprises in Slovenia have the following characteristics and factors for the growth of dynamic enterprises in Europe: (1) Th e growth of dynamic enterprises in Slovenia depends on the external environment attributes of the enterprise: of the 17 environmental factors that stimulate or hinder growth in dynamic enterprises in Europe, only two in our decision trees remained without any descriptive attribute. Th ese are “social recognition by the environment” and “the protection of intellectual property”. We may therefore conclude that the external environment aff ects the growth of Slovenian enterprises similarly to those in Europe. (2) Th e growth of dynamic enterprises in Slovenia depends on the entrepreneur or the entrepreneurial-management team; however, we did not record the most important attributes from the EU among the factors of growth in Slovenia. Th e fundamental attribute, the vision and strategic management, is the only factor stimulating the growth of dynamic entrepreneurs from the set of European attributes we obtained in our decision trees to predict the growth of dynamic enterprises in Slovenia. Other factors (building-up the organisation, internal entrepreneurship, leadership) did not occur in our results. (3) Th e growth of dynamic enterprises in Slovenia depends on an innovation-friendly at- titude and implementation of the change, like in Europe; however, in the decision trees no attribute appeared that points to the readiness of Slovenian entrepreneurs to assume higher growth-related risk, which is a major characteristic of the European gazelles. (4) Th e growth of dynamic enterprises in Slovenia depends on the selection and imple- mentation of the business strategy (the strategy of growth), similarly as in Europe. Th e results obtained by means of the decision trees reveal that the strategy of interna- tional expansion and a strict customer-centred orientation were the most important features in Slovenian dynamic enterprises as well as in their European counterparts. (5) Th e growth of dynamic enterprises in Slovenia depends on features of the manage- ment system. Th e attributes of the European gazelles were identifi ed in our results as well, although more specifi cally the Slovenian dynamic entrepreneurs are more neu- tral than the European entrepreneurs regarding the relevance of corporate organisa- tion which is innovation-friendly and they do not fi nd the employee remunerating system as important as their European counterparts. (6) Th e growth of dynamic enterprises in Slovenia depends on employees’ working con- ditions, such as promotion and responsibility, the loyalty and commitment of em- ployees to the enterprise, the possibility of participation in the growing concern and the personal growth of the employees, which was similar to dynamic enterprises in the EU. V. PŠENIČNY | A LONGITUDINAL COMPARISON OF THE GROWTH FACTORS OF SLOVENIAN FAST ... 279 (7) Th e growth of dynamic enterprises in Slovenia also depends on fi nancing the growth or the development of fi nancial planning and management in a dynamic enterprise. However, considering the responses from dynamic entrepreneurs in both Europe and Slovenia we may conclude that there are diff erences in fi nancial management and planning in dynamic enterprises. 7. LONGITUDINAL COMPARISON OF GROWTH FACTORS IN SLOVENIAN DYNAMIC COMPANIES Our longitudinal research of growth factors in Slovenian gazelles (fast-growing compa- nies) was enriched in 2008 by comparing the diff erences in the answers in the research by both Pšeničny (2003) and Bajt (2008). Using a statistical χ2 test, we estimated the dif- ferences in answers between the two research works and tried to ascertain if the growth factors had changed in the last fi ve years. In our recent survey, 21 owners and entrepre- neurs of 74 fast-growing companies participated. Th e questionnaire used was the same as those used in 1994 and 2002. In 2002, all gazelles (74) employed 2,589 people (with an average of 35) and, in 2007, al- together 5,252 people (with an average of 71), which means that in the entire period they employed 2,663 people in total (with an average of 36). Th e annual average number of newly employed at Slovenian gazelles is 14.2, which is the same as in Pšeničny’s research (hundreds of Slovenian gazelles created 7,150 new jobs between 1998 and 2002). In Table 2, the number of all answers according to infl uential factors is presented as well as the number and share of the same and diff erent answers. In total, 10% have sta- tistically signifi cant diff erent answers but we can see diff erences in factors referring to fi nancing (25% diff erent answers), innovativeness (15.79%), business strategy (14.29%), management system (10%), external environment (9.09%), then entrepreneur (5.88%) and employees (4.35% diff erences). TABLE 2: Comparison of answers of Pšeničny and Bajt according to growth rate factors Ex te rn al en vi ro nm en t f or co m pa ni es En tr ep re ne ur In no va ti ve ne ss Bu si ne ss st ra te gy M an ag em en t sy st em Em pl oy ee s Fi na nc in g To ta l Answers 66 17 19 28 10 23 8 171 Same 60 90.91% 16 94.12% 16 84.21% 24 85.71% 9 90.00% 22 95.65% 6 75.00% 153 89.47% Diff erent 6 9.09% 1 5.88% 3 15.79% 4 14.29% 1 10.00% 1 4.35% 2 25.00% 18 10.53% Source: Bajt, R. (2008): Growth factors in Slovenian Dynamic Companies ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009280 With the growth factors describing the external environment of the company, we checked 66 answers, 90.91% of which were the same as in 2002; the rest diff er from each other. Th is factor includes characteristics like: the relationship between the risk and award which can be gained by an entrepreneur; what is education and support for entrepreneurship like; the social climate for exclusion; is there any creativeness in the education system; what is protection of intellectual property like; is there any support and co-operation in research and development; are there any barriers to international expansion; what is the climate for internationalisation; what are tax supports in the current income statement; share options and plans for interests; is there enough personnel available; what is the mobility of personnel; is the risk capital accessible; are fi nancial markets eff ective; and what is the taxation of deferred income tax assets and reinvestment. 8. CONCLUSIONS If Slovenia is to become a prosperous country and transform into a more developed European economy and even overtake some of the most developed EU countries in a decade or two (as some years ago Aleš Vahčič “called for” in the Slovenian Economic Periodical (1995; 295–312), we need to follow the example of the entrepreneurially most developed and active countries (Glas, 2000). Knowing that some countries are more en- trepreneurial than others (Reynolds et al., 2001; Bosma et al., 2008), the most advanced and expressly entrepreneurially friendly countries that favour the emergence and growth of enterprises seem to be best suited as our model of development. Our analysis confi rms that the growth of enterprises in Slovenia is aff ected by more or less the same growth factors as in the EU, bearing in mind that as regards some features, mainly related to the business and fi nancial environment, along with some internal- environmental factors, our gazelles are not yet comparable with their European coun- terparts. By applying the method of machine learning to the case of Slovenian dynamic enter- prises as an alternative and complementary method of growth factor analysis, we fi nd that: (1) some attributes are more relevant to the success of dynamic enterprises than other attributes; and (2) such attributes are quite few in number, which facilitates the identifi cation of success- ful dynamic enterprises with growth potential. Our research shows that the growth factors found in the research of European dynamic enterprises can be “trusted” and relied upon: we have identifi ed the vast majority of these factors as key growth factors in Slovenian gazelles as well. Further research on dy- namic entrepreneurship should focus, according to our fi ndings and experience, on the most relevant growth factors and features that have proven successful in research into European, and now Slovenian, dynamic enterprises. Likewise, social eff orts should be directed at setting up the identifi ed conditions for the fast growth of enterprises, whereas V. PŠENIČNY | A LONGITUDINAL COMPARISON OF THE GROWTH FACTORS OF SLOVENIAN FAST ... 281 on the enterprise level the attributes common to the most successful gazelles in Slovenia and the EU should be highlighted. Th e growth of Slovenian gazelles in the last fi ve years is highly correlated with almost nine of the ten environmental factors infl uencing the growth and success of European gazelles, such as those seen in 2002. In contrast, some important factors (e.g. stimu- lating the innovation and internationalisation policy, growth of a supportive taxation system, availability of diff erent fi nancial resources) are still impeding the faster growth of fi rms. A strong entrepreneurial vision and a strategic management approach are the most signifi cant characteristics of dynamic entrepreneurs in both Slovenia and Europe. Sustainable growth depends on a permanent innovative and research-implementing orientation of dynamic enterprises, while the lack of risk taking among Slovenian ga- zelles could be a signifi cant barrier to further sustainable growth. Internationalisation and globalisation, both inexorably customer-oriented, are signifi cant characteristics of the growth strategy of gazelles. Some indicators of winning business models of Euro- pean gazelles (e.g. the importance of logistics, organisation and awarding employees) are less important for Slovenian gazelles, while the loyalty and commitment of employ- ees and their ability for personal growth are not signifi cantly diff erent. Some major diff erences between Slovenia and Europe were found in the fi nancial environment (e.g. taxation on stock option plans and retained earnings) but also for fi nancial planning and cash management. On the other hand, we checked the diff erences in answers between the 2002 and 2007 studies. Answers were grouped to describe several growth factors and the most numer- ous diff erences were found in the “fi nancing” group (25% diff erent answers), whereas answers in the “employee” group remained practically unchanged with diff erent answers only about newly created jobs. It seems as if the fi nancial sector is adapting quickly to the new conditions by off ering new products and services. It is worrying that the answers in the group of factors of “innovativeness” had only changed by 16%, mostly about the growth strategy in the future, the main advantages, and the reason for success. Inno- vation in Slovenian dynamic companies is very poor and only a small share (3/21) of companies owns a patent or a license. Responses regarding the “business strategy” only diff er in 14% of cases. Th is is quite understandable since Slovenia joined the EU relatively recently (2004). As the machine learning model was built in 2002, we also tested its accuracy. We used the 2002 database as a learning dataset and the 2007 database as a test dataset. Th e best results were found in the class RDCP (profi t growth in total income) where prediction was more than 85% accurate; however, in some other cases we found less than 30% ac- curacy. RECEIVED: NOVEMBER 2008 ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009282 REFERENCES AJPES (2009), Financial Data on Slovenian Companies and Sole Proprietors for 2007, Ljubljana: Agency of the Republic of Slovenia for Public Legal Records and Related Services. Bajt, R. (2008), Growth factors in Slovenian Dynamic Companies. Master Degree Th esis, University of Ljublja- na, Faculty of Economics. Birch, D. (1987), Job Creation in America, New York: Free Press Macmillan. Bosma N. et al. (2008): Global Entrepreneurship Monitor, 2008 Executive Report. Boston, USA; Santiago, Chile; London, UK. EFER (1996): Europe’s 500. Dynamic Entrepreneurs Th e Job Creators, Brussels: European Foundation for En- trepreneurship Research. Filipič, B., Pšeničny, V. (2003), Data mining models of dynamic enterprise growth, Ljubljana: IEEE Region 8, Slovenian section IEEE, proceedings 12th International conference ERK. Glas, M. (2000), Th e Entrepreneurial Society – Keynote Presentation, Round Table, Ljubljana: ICPE. GrowthPlus (2001), Europe’s Most Dynamic Entrepreneurs. Th e Real Job Creators, Brussels: GrowthPlus/Eu- rope’s 500. Makridakis, S. G. (1990), Forecasting, Planning, and Strategy for the 21st Century, New York: Th e Free Press. Mei-Pochtler, A. (1999), Growth Factors and Innovation Potential. Growth Plus: Strategies for Growth; 1999 Research Report, Brussels: GrowthPlus/Europe’s 500, pp. 97–104. Mitchell, T. M. (1997), Machine Learning, New York: McGraw-Hill. Penrose, E. (1995), Th e theory of the growth of the fi rm, 3rd ed., Oxford: Oxford University Press. Pšeničny, V. (2001), Financing of Dynamic Enterprises in Slovenia and the EU. Th ird Annual Report on Re- search Project. Portorož: Visoka strokovna šola za podjetništvo. Psenicny, V., Filipic, B. (2002), A Data Mining Approach to the Modelling of Dynamic Enterprises Growth. Entrepreneurship research conference 2002, Babson: Babson College Kauff man Foundation, 2002. Pšeničny, V. (2003), Conditions and Opportunities for Dynamic Entrepreneurship in Slovenia. Doctoral Th esis, University of Ljubljana. Pšeničny, V. (2008), 5000 Fast-growing Small Businesses Has Created a Large Part of Growth. Ljubljana: Dnevnik, Special Edition, Th e Gazelle. Quinlan, J. R. (1986), ”Induction of Decision Trees”, Machine Learning 1, 81–106. Rebernik, M. et al. (2008), Lack of Growth-oriented Entrepreneurship. GEM Slovenia 2007, Maribor: Ekonom- sko-poslovna fakulteta. Reynolds, P.D. et al. (2001), Global Entrepreneurship Monitor. 2001 Executive Report, Boston–London: Bab- son College in London Business School. Roure, J. et al. (1999), Europe’s Most Dynamic Entrepreneurs Th e 1998 Job Creators, Brussels: GrowthPlus/ Europe’s 500. V. PŠENIČNY | A LONGITUDINAL COMPARISON OF THE GROWTH FACTORS OF SLOVENIAN FAST ... 283 Roure, J. (2000), Towards a European Model of Entrepreneurial Growth. Dynamic Entrepreneurship for New Economy, Conference Proceedings, ed. Vadnjal J., Portorož: GEA College of Entrepreneurship, 16–33. Roure, J. (2001), E-500 Aggregate Data: Wealth Creation. Employment Creation. 1999 Sample Study, Barce- lona: IESE. Solymossy, E. (1998), Entrepreneurial Dimensions: Th e Relationship of Individual, Venture, and Environmen- tal Factors to Success. Doctoral thesis, Cleveland: Weatherhead School of Management, Case Western Reserve University. Stevenson, H.H. (1998), Do Lunch or Be Lunch: Th e Power of Predictability in Creating Your Future, Boston: Harvard Business School Press. Tajnikar, M. (2000), Risk Management, Book on Gazelles and Growth Business, Second edition. Portorož: GEA College - Visoka šola za podjetništvo. Vahčič, A. (1995), ”Th e Role of Entrepreneurship in Overcoming the Economic Developmental Gaps in De- veloping Countries”, Slovenian Economic Periodical 46 (4), 295–312. Wiklund, J. (1998), Small Firm Growth and Performance. Entrepreneurship and Beyond, Joenkoeping: JIBS. Joenkoeping University. Witten Ian H., Frank Eibe (2005), Data Mining: Practical Machine Learning Tools and Techniques with Java Implementations, San Francisco: Morgan Kaufmann. Zanasi, A. et al. (ed.) (2007), ”Data Mining VIII: data, text and web mining and their business applications”, WIT transactions on information and communication technologies 38. Žižek, J., von Liechtenstein H. (1994), Venture Capital and Entrepreneurship in Central and Eastern Europe, 750 CEE Dynamic Entrepreneurs Database Survey, Amsterdam: EFER. ECONOMIC AND BUSINESS REVIEW | VOL. 11 | No. 4 | 2009284