Volume 12 Issue 2 Article 3 12-31-2010 Determinants of firm entries: empirical evidence for Slovenia Dijana Močnik Follow this and additional works at: https://www.ebrjournal.net/home Recommended Citation Močnik, D. (2010). Determinants of firm entries: empirical evidence for Slovenia. Economic and Business Review, 12(2). https://doi.org/10.15458/2335-4216.1246 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. 129 ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 2 | 2010 | 129–145 DETERMINANTS OF FIRM ENTRIES: EMPIRICAL EVIDENCE FOR SLOVENIA DIJANA MOČNIK* ABSTRACT: We empirically investigate the determinants of new fi rm formations on Slovenian data set for the 6-year period across statistical regions. Analyzed are the re- lationships of the determinants classifi ed into fi ve groups: demand, unemployment, in- dustrial restructuring, local fi nancial capital, and knowledge concentration. We fi nd a positive and signifi cant impact of GDP p.c., unemployment rate, productivity growth and a negative relationship for employment density. Results show that some regions have signifi cantly worse conditions for start-up fi rms than others. Practical implications of this study would allow policy makers to better understand the dynamics in new fi rm formations. Keywords: fi rm entry, small business, entrepreneurship, regression, Slovenia JEL classification: L26 1. INTRODUCTION Since the 1970s, researchers have sought to empirically demonstrate that small and medium-sized enterprises increasingly contribute to development. Th e entrepreneur serves as a catalyst for economic growth and development (Braunerhjelm 2007). Entries of enterprises are in fact related to the processes of innovation and change in the in- dustry (Dosi et al. 1995, Callejón & Segara 1999). New businesses are, compared to old companies, more readily able to develop, use, and introduce radical innovations and changes, as refl ected in the rising revenues and productivity rates (Casson 2002a, 2002b, Baumol 2007, in Braunerhjelm 2007). Th e more small fi rms (which start-ups typically are) that exist, the more impact they have on deregulation, increased competition, and better exploitation of new technologies and knowledge (Jovanovic & Rousseau 2005, in Braunerhjelm 2007, Močnik 2009). According to Audretsch (1995a, 1995b, 1997) and Callejón and Segara (1999), an individual who wishes to realize innovation may do so by establishing a company. * University of Maribor, Faculty of Electrical Engineering and Computer Science, Smetanova 17, 2000 Mari- bor; Slovenia, e-mail: dijana.mocnik@uni-mb.si ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 2 | 2010130 Empirical research has shown that promoting entries can create long-term benefi t to society as small, innovative fi rms start their businesses in the relatively uninvestigat- ed areas of technology (Almeida & Kogut 1997, Almeida 1999, in Braunerhjelm 2007). Moreover, they are oft en willing to introduce radical innovations (Rothwell & Zegveld 1982, Baumol 2004, in Braunerhjelm 2007), resulting in greater effi ciency, productivity, and growth (Durnev et al. 2004, Aghion et al. 2004, Aghion & Griffi th 2005, Acemoglu et al. 2006, Chun et al. 2007, in Braunerhjelm 2007). Because of just described perceived benefi ts that new fi rms may bring to the society, we aim to expose the determinants that have been found as important for the fi rm entries. We study the relationship between the gross rate of entry and ten determinants across twelve Slovenian statistical regions. Th e gross rate of entry represents the ratio between the number of new established compa- nies and existing companies. Th us, it represents the percentage share of new established companies in existing companies. As many authors suggest that locations with more knowledge become attractive to en- trepreneurs (Waltz 1996, Baldwin & Johnson 1999, Black & Henderson 1999, Fujita et al. 1999, Martin & Ottaviano 1999, 2001, Baldwin & Forslid 2000, Fujita & Th isse 2002, Hendersson & Th isse 2004, in Braunerhjelm 2007), our estimation is made in a way that we are able to assess the impact of various determinants controlled for the region eff ect. Th e study is organized as follows. In the Section 2 we give an overview of the previous research and present the study’s hypotheses. Section 3 presents the model and estimation technique. Section 4 deals with the results. In the fi nal section we present the conclu- sions. 2. PREVIOUS RESEARCH AND HYPOTHESES 2.1 Th eoretical Background Many researchers have already studied the eff ects of entries. For example, Sutaria and Hicks (2004) conducted the research of diff erent factors on the creation of new busi- nesses in manufacturing regionally for all Texas metropolitan statistical areas. Brixy and Grotz (2007) studied the correlation between the intensity of new-fi rm formation and the survival rates of young businesses in West Germany. Th ere are the determi- nants of the spatial diff erences in the rates of fi rm formation that have already been the subject of many studies (e.g. Audretsch & Fritsch 1994; Sutaria/Hicks 2004). In our study, we selected the independent variables that largely following the studies cited above. Firstly this guarantees the comparability of the results obtained, and second- ly the choice of new or alternative characteristics is considerably restricted due to the availability of data. Despite the eff orts of various researchers to explain the diff erent eff ects of entries and exits through empirical research, conclusions are not uniform and many questions re- main unanswered. It is not possible to establish uniform and clear tools, assumptions, and fi ndings of this vital economic process. Naturally, research results also diff er be- D. MOČNIK | DETERMINANTS OF FIRM ENTRIES: EMPIRICAL EVIDENCE OF SLOVENIA 131 cause they are derived from diff erent conceptual models and data that cannot be directly compared; these models and data are used in the analysis of the diff erent industries for diff erent countries and periods of time or rely on diff erent methodologies for data collec- tion and processing, etc. Th us, it is not surprising that studies bring confl icting results. For example, Highfi eld and Smiley (1987, in Sutaria & Hicks 2004) and Audretsch and Fritsch (1994, in Sutaria & Hicks 2004) note that the unemployment rate is positively related to entries (i.e., the increased number of unemployed impacts the increase in en- tries), whereas Guesnier (1994, in Sutaria & Hicks 2004) and Garofoli (1994, in Sutaria & Hicks 2004) note that this link is very negative (i.e., the less the unemployment, the more the entries). Th ese results not only created confusion among scholars about the true nature of impacts of contextual factors on new fi rm formations, but also made it more diffi cult for policy makers to implement them. Confl icting research results may be evidence that establishing new businesses is un- doubtedly a very complex process that depends on various unrelated as well as correlated factors and specifi c characteristics of local conditions. To capture as many factors, our estimation takes into account various determinants of the newly established fi rms over the period 2000-2005 in Slovenia. Th e determinants are selected according to previous research (e.g. Audretsch & Fritsch 1994; Sutaria/Hicks 2004). Such selection enables the comparability of the results ob- tained, and secondly the choice of new or alternative characteristics is considerably re- stricted due to the availability of data. Hypotheses are represented in Section 2.3. 2.2 Firm Entry mechanisms Th e conceptual framework within which the hypotheses about regional factors infl u- encing new fi rm formation will be derived and tested for fi ve fi rm entry mechanisms: demand, unemployment, industrial restructuring, local fi nancial capital, and knowledge concentration. Below is the discussion of each group. Group 1 Demand: Expanding demand for goods and services increases entries. It is rea- sonable to hypothesize that new fi rms emerge to satisfy rising new demands for goods and services. We have developed two indicators representing change in local demand: the annual rate of revenue growth change and the GDP p.c. An increase in both can be expected to drive rising demand for goods and services (Reynolds 1994, Sutaria & Hicks 2004), which in turn can led to rising rates of entries. Group 2 Unemployment: When a person loses his/her job and fails to fi nd another one that is comparable, he/she may well seek to choose to create a new one for himself/ herself by starting his/her own business. Th e formation of new fi rms, in turn, may re- duce unemployment rate as the person starting a new fi rm employees not only himself/ herself but also others. At the same time, a higher level of unemployment may reduce aggregate disposable income, eff ectively reducing local demand for goods and services, ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 2 | 2010132 thereby putting downward pressure on its rate of new fi rm formation (Reynolds et al. 1994). Th ese two opposite infl uences combined with a reverse causation eff ect – new fi rms reducing unemployment rate – create uncertainty about the net impact of unem- ployment on entries (Sutaria & Hicks 2004). Ultimately, the net impact of unemploy- ment depends on which of the two infl uences, unemployment push or demand pull, dominates for a region, as well as the way in which the essential relationship is specifi ed by other factors. One indicator of unemployment was developed for this study: the unemployment rate as the share of the number of unemployed persons to a region’s total labor force. Th e indi- cator refl ects the existing status of an economy at a particular point in time in terms of number of people unemployed. Group 3 Industrial restructuring: In this paper, we have assumed that fi ve predictors can be used as measures of industrial restructuring. A fi rst predictor is mean establishment size (MES), defi ned as the mean number of em- ployees per business. Empirically has been found that new businesses emerge on a larger scale in industries, which are characterized by smaller fi rms (Audretsch 1995a, in Braun- erhjelm 2007). It is hypothesized that the smaller a region’s MES the greater the number of newly established companies (Sutaria & Hicks 2004). Th us, we expect a negative as- sociation between MES and the gross rate of entry. A second predictor is productivity growth, defi ned as the change of the annual rate of revenues per employee. We hypothesize, as Nivin (1998) does, that the growth of the pro- ductivity creates new demand for the development and manufacture of new products, which means that there will be more newly established companies if productivity growth rises. Th us, we expect a positive relationship between the productivity growth and the gross rate of entry. A third predictor is index of diversifi cation, by which we measure in how many diff er- ent industries a region creates its revenues. It was suggested that new companies more likely occur in more diversifi ed regions (Glaeser et al. 1992, Feldman & Audretsch 1999, Henderson & Th isse 2004, in Braunerhjelm 2007). Across regions we calculate the shares of created revenues across 13 industries. Th en we square these shares and calculate the sums. Finally, we calculate a diversifi cation index by dividing one with the sums of the squares across the regions. Th e diversifi cation index may lie between one and 13. Th e greater the index, the more diversifi ed is a region and consequently the greater the chance of establishing new fi rms as smaller index might imply more concentrated markets in a region. We apply the method used by Albarran (2002) who calculate the diversifi cation index for a fi rm’s revenues created in diff erent markets. Th us, a positive relationship is expected between this predictor and the gross rate of entry. A fourth predictor is investment in fi xed assets represented by the logarithm of annual gross investment in fi xed assets. With this variable we indirectly assess the optimistic D. MOČNIK | DETERMINANTS OF FIRM ENTRIES: EMPIRICAL EVIDENCE OF SLOVENIA 133 expectations of the future (e.i. favorable taxation policy) (Murphy et al. 1991, in Braun- erhjelm 2007). Each entrance leads to a certain fi xed costs, so each start-up company depends on the investment (Braunerhjelm 2007). Th e bigger the investment, the more start-ups may occur. Th e larger an investment in fi xed assets, the greater the gross rate of entry, which means that we expect a positive association. A fi ft h predictor is a gross rate of exit, calculated as a ratio of the absolute number of a region’s companies that end their activity to the region’s total number of companies. Th e greater/smaller the number of businesses that ceased to operate (exit), the greater/ smaller is the chance for starting new businesses. Th e expected connection between entries and exits is positive when the impact of competition is considered, while a negative when a multiplier eff ect is implied. Th e ambiguity in sign may be because of two opposite eff ects, competition and multiplier eff ects, which seem to work at the same time when an entry or exit occurs (Sutaria & Hicks 2004). For example, more entries may cause more exits in subsequent periods due to enhanced competition (a competition eff ect), or may cause fewer exits because the demand for all businesses’ products has increased (a multiplier eff ect). So, the expected relationship between en- tries and exits is indeterminate. We emphasize that despite the fact that there might be strong barriers to exit, companies are sometimes forced to exit markets (Karakaya 2000). Th us, our proposition is that fi rms exit when current losses exceed the present value of expected profi ts.1 Group 4 Local Financial Capital: Regions endowed with relatively high levels of per capita fi nancial assets such as local bank deposits are more likely to be areas where ac- cess to capital is comparatively easy (Garofoli 1994, Sutaria & Hicks 2004). Such pools of capital are available not only for new startups but also for the expansion of existing busi- nesses (Sutaria & Hicks 2004) and represent business expansion capital, which usually represents an amount larger than what is likely to be fi nanced through borrowing from friends or by using personal credit. Th is supposition has its roots in the resource based theory, which argues that the entrepreneur will start a business when he has suffi cient resources for doing this (Cooper et al. 1994, Cooper 1995, Penrose 1959, in Braunerhjelm 2007). Financial capital is one of the fi ve most important sources of companies (in addi- tion to human, management, sector-specifi c and access to markets and resources). In the paper, logarithm of a region’s per capita bank deposits, calculated by dividing the total bank deposits by the total population of a region, is introduced as an indicator of the availability of local fi nancial capital. Group 5 Knowledge Concentration: We have developed one indicator called employ- ment density, which is used as an indirect measure of knowledge concentration. We em- phasized the impact of networks and social capital found within a geographic region. Relational networks exist at multiple levels of analysis because they can link together individuals, groups, fi rms, industries, geographic regions, and nation-states (Audretsch 1 Interested readers are invited to identify six major exit barriers (cost of divestment, operating fi t, marketing fi t, forward vertical integration, backward vertical integration, and number of years’ association of the busi- ness unit with the fi rm) in Nargundkar, Karakaya, and Stahl 1996 (in Karakaya 2000). ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 2 | 2010134 & Feldman 2003). Th e density of employment is calculated as the percentage of a region’s employed persons in the number of the region’s inhabitants. We base our expectation on the importance of geographical proximity for knowledge spillovers in innovative- ness. Th is means that the transfer of knowledge requires proximity. Technological and entrepreneurial skills and innovation do not occur in a particular region simply because someone has the necessary skills and initial production resources; rather, this region has available all the necessary resources, which are developed in its local environment. Innovation processes are, in the opinion of many authors, subject to the processes in the local environment because innovation requires a complex exchange of knowledge, which can be obtained only in a specifi c regional environment (Lundvall 1992, Antonelli 1995, 1997, in Braunerhjelm 2007). Th ere are numerous studies that examine the determinants and extent of spatially concentrated production (Krugman 1991a, 1991b, Glaeser et al. 1992, Ellison & Glaeser 1997, Feldman & Audretsch 1999, Maurel & Sedillot 1999, Acs et al. 2002, Braunerhjelm & Johansson 2003, Braunerhjelm & Borgman 2004). We hypoth- esize a positive relationship between the employment density and the gross rate of entry. In this way we inter-relate knowledge and entrepreneurship: a decision to start a new business is a result of created and diff used knowledge. We argue that Slovenian entries are related to entrepreneurs’ innovative skills that are the result of knowledge spillover. 2.3 Hypotheses Hypotheses arising out of the group factors discussed in the previous section are the following: (1) A region’s rate of revenue growth change is positively related to its gross rate of en- try. (2) A region’s rate of per capita GDP is positively related to its gross rate of entry. (3) A region’s rate of unemployment and its gross rate of entry are related, although the direction of this relationship is indeterminate. (4) A region’s mean establishment size is negatively related to its gross rate of entry. (5) Th e rate of a region’s change in the productivity growth is positively related to a region’s gross rate of entry. (6) An index of region’s diversifi cation is positively related to a region’s gross rate of entry. (7) A region’s investment in fi xed assets is positively related to a region’s gross rate of entry. (8) A region’s gross rate of exit is related to its gross rate of entry; but the direction of this relationship is indeterminate. (9) Th e level of the per capita bank deposits in a region is positively related to its gross rate of entry. (10) A region’s employment density is positively related to a region’s gross rate of entry. Ten predictors judged to exert independent infl uences on the gross rate of entry are pre- sented in Table 1. D. MOČNIK | DETERMINANTS OF FIRM ENTRIES: EMPIRICAL EVIDENCE OF SLOVENIA 135 TABLE 1: Determinants of the gross rate of entry Panel data Variable name Code Operational defi nition Dependent variable Gross rate of entry Y The ratio of the absolute number of a region’s companies that begin their activity to the region’s total number of companies.1 Independent variables Variable name Expected eff ect Code Operational defi nition Group Independent variables Demand 1. Rate of revenue growth + X1 Annual rate of revenues change 2. GDP p.c. + X2 LOG of the GDP p.c. Unemploy- ment 3. Unemployment rate +/- X3 The number of a region’s unemployed to the region’s total labor force Industrial restructuring 4. Mean establishment size - X4 Average number of employees per company 5. Productivity growth change + X5 Annual rate change of a region’s revenues per employee 6. Diversifi cation index + X6 One divided by the region’s sum of squares of the shares of created revenues across industries 7. Investment in fi xed assets + X7 LOG of annual region’s gross investment in fi xed assets 8. Gross rate of exit +/- X8 The rate of the absolute number of a region’s companies that end their activity to the region’s total number of companies Local fi nancial capital 9. Bank deposits per capita + X9 LOG of a region’s per capita bank deposits Knowledge concentration 10. Employment density + X10 The percentage of a region’s employed persons in the number of the region’s inhabitants Source: SURS 3. MODEL AND ESTIMATION Data for the estimation were obtained from the Statistical Offi ce of the Republic of Slovenia (SURS). Most data were available on the Internet. Th e main source of informa- tion for SURS is the Statistical Business Register (SPR), maintained by the Agency of the Republic of Slovenia for Public Legal Records and Related Services (AJPES). All the data were acquired across twelve statistical regions for the period from 2000 to 2005 2 Th e analysis covers companies included in the Standard Classifi cation of Activities (SKD) in C - K activities: C - Mining and quarrying, D - Manufacturing, E - Electricity, gas and water; F - Construction; G - Trade, repair of motor vehicles and household goods, H - Hotels and restaurants; I - Transport, storage and commu- nication; J - Financial intermediation; and K - Real estate, renting and business activities. Th ese are the SKD before January 1 2008, the initiation of new regulations on the standard classifi cation of economic activities. ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 2 | 2010136 and refers to the fi rms of all NACE (SKD) activities (census data). Th us, the calculations are made with the use of the panel data (12 regions multiplied by 6 years = 72 observa- tions). Given the cross-sectional and time-series nature of the data developed for this study, the stepwise least square dummy variable (LSDV) (also named fi xed-eff ects) regression model is used (Gujarati 2004). Th e space (regional) dimension of the data is incorporated into the model through the use of eleven dummy variables for twelve regions. Region dummies are used to control for unmeasured region-specifi c infl uences on the depend- ent variable, which may be related to the primary predictors in the model. To our knowl- edge, only the study of Sutaria and Hicks (2004) used the LSDV regression for studying the phenomenon of new fi rm formations regionally in manufacturing. Th e LSDV regres- sion is able to specify relationships between dependent and independent variables in a more precise manner, and therefore it should be considered as a signifi cant improvement over the techniques used by previous empirical studies (Sutaria & Hicks 2004). Th e basic regression model analyzed is given as follows. Yi = a + bj Xji + cki Dki +ei (1) i = 1, 2, …, 72, j = 1, 2, …, 10; k = 1, 2, …, 11 where Yi is gross rate of entry of the i-th observation; i index of observations; a is model constant; bj are regression coeffi cients of the Xj variables (see Table 1); j index of inde- pendent variables; ck is diff erential coeffi cient of the model constant a for the k-th region; k index of regions; Dk is the k-th dummy variable for the k-th region (k=1, Podravska; k=2, Koroška; k=3, Savinjska; k=4, Zasavska; k=5, Spodnjeposavska; k=6, Jugovzhodna; k=7, Osrednjeslovenska; k=8, Gorenjska; k=9, Notranjsko-Kraška; k=10, Goriška; k=11, Obalno-Kraška region). Each dummy variable for a particular region has a value of 1 for the observations (cases) that refer to that region and 0 otherwise. Th e base regression refers to the Pomurska re- gion. Th us, the constant a is the average gross rate of entry of the Pomurska region when all the model’s predictors would be zero. Th e model constant changes for the signifi cant ck values. Th e ck values take into account the specifi c characteristics of the k-th region. 4. RESULTS 4.1 Th e Analysis of Zero-Order Correlations We begin the quest for evidence of possible relationships between the gross rate of entry and key independent variables by examining the degree to which correlations among the variables marked for inclusion in the model (1) to be tested actually covary with one another. D. MOČNIK | DETERMINANTS OF FIRM ENTRIES: EMPIRICAL EVIDENCE OF SLOVENIA 137 TABLE 2: Zero-order Correlation Matrix for Pooled Model Inde- pendent variables1 Y X 1 X 2 X 3 X 4 X 5 X 6 X 7 X 8 X 9 X 10 X1 -,062 1 X2 ,381** -,043 1 X3 ,009 -,140 -,606** 1 X4 -,127 ,060 -,196 ,415** 1 X5 ,042 ,833** -,079 -,106 ,003 1 X6 ,347** -,061 ,397** ,020 -,578** -,038 1 X7 ,288* -,040 ,818** -,392** -,187 -,065 ,406** 1 X8 -,071 ,002 -,320** ,348** ,129 ,021 ,076 -,294* 1 X9 ,231 -,005 ,630** -,088 ,199 -,131 ,308** ,599** -,142 1 X10 ,133 ,175 ,816** -,501** ,160 ,067 ,216 ,618** -,167 ,736** 1 Mean 6,98 9,88 4,03 11,13 6,45 10,04 6,70 5,53 6,39 5,76 27,31 Standard deviation 1,30 5,82 0,09 3,54 0,97 5,19 1,35 0,20 1,16 0,31 4,97 N = 72; 1 see X variables in Table 1, **signifi cant at the 0.01 level (2-tailed), *signifi cant at the 0.05 level (2-tailed) Table 2 presents zero-order correlations, means, and standard deviations of the variables (except dummy variables) included in the pooled model (the overall sample - regions all together). Correlation coeffi cients for three variables, GDP p.c., diversifi cation index, and investment in fi xed assets are found statistically signifi cant and have their directions consistent with the relevant hypotheses in this study. 4.2 Th e Analysis of Regression Coeffi cients Th e results of the model (1) are presented in Table 3. Th e GDP p.c., unemployment rate, employment density, productivity growth, and 6 dummy variables: Jugovzhodna, Spod- njeposavska, Gorenjska, Zasavska, Podravska, and Goriška regions explain 72.3 percent of the variability of the gross rate of entry. Th ree hypotheses out of ten (the second, third, and fi ft h) are confi rmed, whereas the tenth hypothesis has the opposite direction than it was expected. Six hypotheses are not confi rmed, which means that they are irrelevant in explaining the process of new fi rm formations. Th ese unimportant variables seem to be the following: the revenues growth change (the fi rst hypothesis), mean establishment size (the fourth hypothesis), diversifi cation index (the sixth hypothesis), investment in fi xed assets (the seventh hypothesis), gross rate of exit (the eighth hypothesis), and the per capita bank deposits (the ninth hypothesis). Th e regression coeffi cient b2 of the GDP p.c. (X2) is positive and statistically signifi cant (19.266, t = 9.15, sig. 0.000) which is in accordance with the expectation. Th is means that an increase of the GDP p.c. by 0.01 (or 1 percent), changes Y (the gross rate of entry) by ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 2 | 2010138 0.19 (0.19 = 0.01 × 19.266) units.3 As the average gross rate of entry of the pooled model is 7 (see Table 2, second column), this represents the 3-percentage increase of the average gross rate of entry. Th is variable can be regarded as the demand pull determinant of fi rm entries. Th is means that people with the higher standard of living demand more goods, which in turn enables more fi rm formations, as there is enough demand for already es- tablished companies and the new ones. Sutaria and Hicks (2004) have found no evidence of an impact on fi rm formations of per capita personal income growth. Th e rate of unemployment (X3) has a positive and signifi cant regression coeffi cient (b3 = 0.093, t = 2.065, sig. 0.043). Th us, in Slovenia an increase of unemployment rate for a unit (that is 1 percent) increases the rate of entry by 0.093 units, which means that unemployed persons in Slovenia (in a period 2000 to 2005) start their own businesses from the necessity to secure their jobs and income necessary to survive. Th us, we can say that in Slovenia, in the observed period, unemployment push was a determinant of new fi rm formations. Similar results for change in unemployment rate are calculated in studies by Highfi eld and Smiley (1987) and Audretsch and Fritsch (1994) who found signifi cant and positive impact on new fi rm formations. However, Guesnier (1994) and Garofoli (1994) found that this relationship is signifi cant and negative. Th e study of Sutaria and Hicks (2004) found no relation between unemployment and new fi rm for- mations. Th e regression coeffi cient b5 is positive and signifi cant (0.039, t = 2.312, sig. 0.024), which means that an increase of the productivity growth (X5) for 1 unit (this is annual rate change of revenues per employee for 1 percent) increases the gross rate of entry by 0.039 units. Th e result confi rms the fi ft h hypothesis stating that growth in productivity creates new demand for the development and manufacture of new products. Th is eventually impacts the decisions to start new businesses. Th e regression coeffi cient b10 is negative and signifi cant (-0.233, t = -7.099, sig. 0.000), which means that an increase of employment density (X10) by a unit (this is a one per- centage increase of a region’s employed persons in the region’s population) decreases the gross rate of entry by 0.233 units. Th is means that concentration of knowledge inactivates the formation of new companies, or it can be said that in Slovenia, there are a lot of me-too entrepreneurs. We propose a positive coeffi cient of the employment density. Th e results confi rm that Slovenian entries do not represent innovative but me-too start-up companies. Such a result for this predictor was expected when we get a positive and signifi cant regression coeffi cient for unemployment rate, which implies that new entrepreneurs are not innovators but rather people who seek to secure their jobs. Th is is additionally confi rmed also by diminishing contribution of entries to the GDP growth that was estimated in the study by Močnik (2009). All the other six de- terminants, for which we expected that may impact entries, seem not to be related to 3 GDP p.c. is expressed in a logarithm form. Th us the change of the predictor X2 for a unit changes the de- pendent variable by the regression coeffi cient b2 multiplied by the change of X2. Th e algebra is as follows: b2 = change in Y/change in ln X2 = change in Y/relative change in X2 = ΔY/(ΔX/X); ΔY = b2×(ΔX/X) (Gujarati 2004). D. MOČNIK | DETERMINANTS OF FIRM ENTRIES: EMPIRICAL EVIDENCE OF SLOVENIA 139 the gross rate of entry. Th us, rate of revenue growth, mean establishment size, diver- sifi cation index, investment in fi xed assets, gross rate of exit, and bank deposits have no signifi cant impact on the decision to start a business in Slovenia in the observed period. As estimation was done by the stepwise regression, we are able to assess the portion of the explained variance by each variable, included in the model. Th e most variability, 12.5 percent is explained by the unemployment rate (X3). Th is is followed by the GDP p.c. (X2) with 11.7 percent. Th e third variable is the employment density (X10) with 7.1 percent, while productivity growth (X5) explains 4.9 percent of the variability of the gross rate of entry (Y). All together this accounts to 36.2 percent. Another 36 percent is explained by dummy variables for regions. Th e most variability of the dependent variable is ex- plained by the dummy variable of Spodnjeposavska region (D5), namely 11.9 percent (see R2 change in Table 3, step 5). Th is is followed by Gorenjska (D8) (8.7 percent, see step 6 in Table 3), Jugovzhodna (D6) (6.8 percent, see step 3), Zasavska (D4) (3.3 percent, see step 8), Goriška (D10) (2.7 percent, see step 10), and fi nally Podravska (D1) (2.6 percent, see step 9 in Table 3) regions. Th e model constant that represents the average gross rate of entry that would be achieved when all independent or predictor variables would be zero has no meaning as its value is negative. Would the constant be positive, its value should be decreased for negative and signifi cant coeffi cients of the dummy variables of Jugovzhodna, Spodnjeposavska, and Goriška regions. Th e biggest decline of -1.428 has Spodnjeposavska region, followed by Jugovzhodna region with -1.149 and -0.862 of the Goriška region. However, positive and signifi cant coeffi cients are for Gorenjska (1.192), Zasavska (1.039), and Podravska re- gions (0.976) (see Table 3, the last column). Th us, in Jugovzhodna, Spodnjeposavska and Goriška regions have worse conditions in comparison to the Pomurska region, whereas Gorenjska, Zasavska and Podravska regions have better conditions than Pomurska re- gion. In other regions, the average gross rate of entry is the same as the model’s constant, i.e. they have pretty the same conditions as Pomurska region. Th e health of the model is tested by the variance infl ation factors (VIFs) for multi- collinearity, Koenker-Bassett (KB) test for heteroscedasticity, and autocorrelation by Durbin-Watson (DW) statistic (Gujarati 2004) (see Table 3). Th e highest VIF amounts to 5.465 for GDP p.c., otherwise these values are less than 10, which means that in the model multicollinearity is not a problem (variables are not correlated so much that this will cause a problem of getting best, linear, unbiased estimates - BLUE) (Gujarati 2004). In the KB test, squared residuals of the model are regressed on the squared predicted values of the regressand. Th e regression coeffi cient is not signifi cant, which shows that there is no heteroscedasticity. DW amounts to 2.195, which is greater than the critical value of 1.792 (that is DWU) and smaller than 2.838, which represents 4 - DWL (DWL is 1.162), so that we can accept the hypothesis of no autocorrelation. Th e inspection of residuals suggests normal distributions. Graphs and some output are not included in the paper due to space limitations but can be accessed with the author. ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 2 | 2010140 TABLE 3: Regression results of the stepwise LSDV regression Dependent variable: Gross rate of entry Stepwise LSDV regression (10 steps) Independent variables 1 2 3 4 5 6 7 8 9 10 a -11,383 -27,068 -25,752** -43,300** -54,322** -64,479** -73,436** -74,327** -70,576** -65,844** Constant (-1,884) (-3,691) (-3,645) (-4,674) (-6,144) (-7,534) (-8,486) (-8,924) (-8,577) (-8,068) b2 4,536** 8,028** 7,776** 12,873** 16,042** 18,610** 20,819** 21,041** 20,205** 19,266** X2 1 Variance infl ation factor (VIF) (3,028) 1,000 (4,601) 1,557 (4,630) 1,562 (5,266) 3,637 (6,812) 4,118 (8,207) 4,551 (9,157) 5,116 (9,616) 5,124 (9,399) 5,279 (9,150) 5,465 b3 0,143** 0,125** 0,122** 0,143** 0,189** 0,214** 0,200** 0,154** 0,093** X3 1 VIF (3,346) 1,557 ((2,989) 1,604 (3,054) 1,605 (3,911) 1,640 (5,287) 1,871 (6,141) 1,993 (5,878) 2,051 (3,998) 2,815 (2,065) 4,133 c6 -1,093* -1,156** -1,294** -1,114** -1,002** -0,949** -0,932** -1,149** D6 2 VIF (-2,579) 1,034 (-2,856) 1,037 (-3,516) 1,046 (-3,268) 1,069 (-3,088) 1,084 (-3,034) 1,089 (-3,077) 1,089 (-3,766) 1,194 b10 -0,109** -0,176** -0,207** -0,235** -0,233** -0,230** -0,233** X10 1 VIF (-2,761) 3,131 (-4,454) 3,853 (-5,563) 4,058 (-6,449) 4,368 (-6,620) 4,373 (-6,775) 4,376 (-7,099) 4,379 c5 -1,605** -1,669** -1,733** -1,594** -1,413** -1,428** D5 2 VIF (-3,939) 1,281 (-4,467 1,284 (-4,896) 1,289 (-4,619) 1,324 (-4,116) 1,399 (-4,321) 1,400 c8 1,332** 1,465** 1,506** 1,464** 1,192** D8 2 VIF (3,671) 1,211 (4,232) 1,232 (4,519) 1,235 (4,532) 1,239 (3,606) 1,401 b5 0,053** 0,051** 0,049** 0,039** X5 1 VIF (2,926) 1,152 (2,976) 1,153 (2,937) 1,156 (2,312) 1,242 c4 0,796* 0,997** 1,039** D4 2 VIF (2,475) 1,150 (3,082) 1,242 (3,330) 1,246 c1 0,801* 0,976** D1 2 VIF (2,276) 1,472 (2,816) 1,539 c10 -0,862* D10 2 VIF (-2,413) 1,636 R2 0,117 0,242 0,311 0,382 0,501 0,588 0,637 0,670 0,696 0,723 R2 Change 0,117 0,125 0,068 0,071 0,119 0,087 0,049 0,033 0,026 0,027 R2 Adjusted 0,104 0,220 0,280 0,344 0,463 0,549 0,597 0,627 0,651 0,676 F statistic 9,168** 10,860** 10,059** 10,196** 13,053** 15,211** 15,801** 15,717** 15,488** 15,623** F Change 9,168** 11,197** 6,652* 7,623** 15,514** 13,475** 8,560** 6,126* 5,180* 5,820* N3 70 70 70 70 70 70 70 70 70 70 Durbin-Watson statistic (DW) 2,195 Notes: 1 X2 – GDP p.c., X3 – unemployment rate, X5 – Productivity growth, X10 – employment density (see Table 1); 2dummy variables: D1 – Podravska, D4 – Zasavska, D5 – Spodnjeposavska, D6 – Jugovzhodna, D8 – Gorenjska, D10 – Goriška regions; Notes: t-statistics are given in parentheses; 3 two observations were omitted because of outliers; ** Signifi cant at the 0.01 level, * at the 0.05 level (both 2-tailed). D. MOČNIK | DETERMINANTS OF FIRM ENTRIES: EMPIRICAL EVIDENCE OF SLOVENIA 141 5. CONCLUSIONS We investigated the relationship between the new fi rm formations and some certain factors, which are considered as relevant in previous research. With the stepwise least square dummy variable regression we estimated the links between the gross rate of en- try as the dependent variable and ten independent variables, classifi ed into fi ve groups: demand, unemployment, industrial restructuring, local fi nancial capital, and knowledge concentration. According to the new geographical theory that argues that location is an important fac- tor of fi rm entries, we decided to study the relationships between the dependent and independent variables across Slovenian twelve statistical regions. Th e gross rate of entry is calculated as the ratio of newly established companies to the region’s total number of companies. Such gross rate of entry standardizes the number of new companies according to the existing number of companies and measures the abil- ity of the region’s population of enterprises to adapt to changing environmental condi- tions. Out of the ten hypotheses only three are confi rmed. We confi rm the positive association between the GDP p.c. and the gross rate of entry (second hypothesis). We estimate that an increase of the GDP p.c. for one percent increases the gross rate of entry by 3 percents. Th e hypothesis of a relationship between the rate of unemployment and the gross rate of entry (third hypothesis) is negative, which means that the increase of the unemployment rate for one percent increases the gross rate of entry by 1.3 percent. Th is result suggests that in Slovenia, unemployed people seek to fi nd a new job by starting a new business. Th is can also be interpreted that in Slovenia some unemployed persons become entre- preneurs from necessity. Such an observation has already ascertained the study by Duh et al. (2009), as well as the GEM study by Rebernik et al. (2009). Confi rmed is also the fi ft h hypothesis of the positive association between the productiv- ity growth and the gross rate of entry. Th e one percentage increase of the productivity growth increases the gross rate of entry by 0.6 percent, which indicates the favorable market conditions for new fi rm formations. Th e proposed positive relationship between the employment density of the region and its gross rate of entry (hypothesis ten) is partially confi rmed. Th e coeffi cient is signifi cant, but has an opposite direction than was expected. Th is means that an increase of the share of a region’s employed people in total population of the region decreases the gross rate of entry by 3.3 percents. Th is may indicate that new Slovenian fi rms are not established by entrepreneurs who start businesses to realize their innovation, but rather that new companies in general imitate existing businesses. Th e result suggests that new fi rms may represent speculative businesses, with their shorter time horizons. However, to support such an argument, further research is needed that will analyze the operations of new ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 2 | 2010142 fi rms. In fact, such a result of this hypothesis complements the third hypothesis. Namely, the more the jobs are available (which means the higher employment density), the less start-up fi rms there are. As we did not confi rm the eighth hypothesis which predicted the association (positive or negative) between the region’s gross rate of exit and its gross rate of entry, we can con- clude that the increasing number of entries did not lead to increased exits, which means that despite the increased entries incumbents are not forced to close down the business because there is enough work for new and old businesses (Sutaria & Hicks 2004, Močnik 2009). Th us, for the Slovenian government it is important to preserve a healthy competitive market structure with institutional arrangements of equal conditions for both new and old businesses. When new companies start their operations to just take a current favora- ble opportunity, soon aft er the start of operation the fi rm exit is expected, which creates unnecessary social costs (Braunerhjelm 2007). Regarding the apparently imitative nature of new businesses, it is therefore hardly ex- pected that new companies will survive, let alone grow. Th us, it is very important for new businesses to survive the fi rst few years, since the likelihood of their growth depends on the age of the company and its initial size (Brixy & Grotz 2007). However, key decisions about the future of the businesses depend on entrepreneurs’ expectations of the future success. Th erefore, the more the future is predictable, the more reliable and realistic rev- enue/cost estimates can be, and on this basis the better entrepreneurs’ aspirations on the enlargement of the business operations and employment. According to our previous work that has shown that a marginal contribution of Slovenian new fi rms to the Slovenian GDP growth has been declining since 2000 at an increased rate (Močnik 2009), the message to the Slovenian government is to provide a favorable business climate that will not only en- courage the establishment of a signifi cant number of new businesses, but also secure the long-term growth of companies with higher added value of their products and services. REFERENCES Acemoglu, D., Aghion, P. & Zilibotti, F. (2006). Distance to Frontier, Selection, and Economic Growth. Jour- nal of the European Economic Association, 4 (1), 37-74. Acs, Z. J. & Armington, C. (2002). Economic Growth and Entrepreneurial Activity. Washington D.C.: Center for Economic Studies, U.S. Bureau of the Census. Aghion, P., et al. (2004). Entry and Productivity Growth: Evidence from Microlevel Panel Data. Journal of the European Economic Association, 2, 265-276. Aghion, P. & Griffi th, R. (2005). Competition and Growth: Reconciling Th eory and Evidence. Cambridge, Mass.: MIT Press. Albarran, A.B. (2002). Media Economics Understanding Markets, Industries and Concepts (2nd ed.). Ames: Iowa State Press. D. MOČNIK | DETERMINANTS OF FIRM ENTRIES: EMPIRICAL EVIDENCE OF SLOVENIA 143 Almeida, P. (1999). Small Firms and Economic Growth. In Acs, Z., Carlsson, B. & Th urik, R. (Eds.), Small Business in the Modern Economy. Oxford: Blackwell Publishers. Almeida, P. & Kogut, B. (1997). Th e Exploration of Technological Diversity and the Geographic Localization of Innovation. Small Business Economics, 9, 21-31. Antonelli, C. (1995). Th e Economic of Localized Technological Change and Industrial Dynamics. Boston and Dordrecht: Kluwer Academic Publishers. Antonelli, C. (1997). Th e Economics of Path-Dependence Industrial Organization. International Journal of Industrial Organization, 16, 643-675. Audretsch, D. (1995a). Innovation, Growth and Survival. International Journal of Industrial Organizations, 13, 441-455. Audretsch, D. (1995b), Innovation and Industry Evolution. Boston, Mass.: Th e MIT Press. Audretsch, D. & Feldman, M. P. (2003). Knowledge Spillovers and the Geography of Innovation. Handbook of Urban and Regional Economics, Volume 4, Revised May 9. http://spol.unica.it/didattica/paci/Economia%20 applicata%20avanzata%20(1mod)/audretsch-feldman.pdf. Audretsch, D. & Fritsch, M (1994). Th e Geography of Firm Births in Germany. Regional Studies, 28 (4), 359- 366. Baldwin, J. R. & Johnson, J. (1999). Entry, Innovation and Firm Growth. In Acs, Z. (Ed.), Are Small Firms Important? Th eir Role and Impact. Dordrecht: Kluwer Academic Publishers. Baldwin, R. E. & Forslid, R. (2000). Th e Core.Peripgery Model and Endogenous Growth: Stabilizing and Destabilizing Integration. Economica, 67, 307-324. Baumol, W. (2004). Entrepreneurial Enterprises, Large Established Firms and Other Components of Th e Free-Market Growth Machine. Small Business Economics, 23, 9-21. Black, D. & Henderson, J. V. (1999). A Th eory of Urban Growth. Journal of Political Economy, 107, 252-284. Braunerhjelm, P. (2007). Entrepreneurship, Knowledge and Economic Growth. Working Paper Series in Eco- nomics and Institutions of Innovation 102. Royal Institute of Technology, CESIS - Centre of Excellence for Science and Innovation Studies. Braunerhjelm, P. & Borgman, B. (2004). Geographical Concentration, Entrepreneurship and Regional Growth: Evidence From Regional Data in Sweden, 1975-99. Regional Studies, 38, 929-948. Brixy, U. & Grotz, R. (2007). Regional patterns and determinants of birth and survival of new fi rms in West- ern Germany. Entrepreneurship & Regional Development, 19, (4), 293-312. Callejón, M. & Segarra, A. (1999). Business Dynamics and Effi ciency in Industries and Regions: Th e Case of Spain. Small Business Economics, 13, 253-271. Chun, H., et al. (2007). Creative Destruction and Firm-Specifi c Performance Heterogeneity. NBER WP 13011, Cambridge, Mass. Cooper, A. C. (1995). Challenges in Predicting New Venture Performance. In Bull, I., Th omas, H. & Willard, G (Eds.), Entrepreneurship: Perspectives on Th eory Building. London: Elsevier Science Ltd. Cooper, A. C., Gimeno-Gascon, F. J. & Woo, C. K. (1994). Initial Human and Financial Capital as Predictors of New Venture Performance. Journal of Business Venturing, 9, 371-395. ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 2 | 2010144 Dosi, G. et al. (1995). Learning, Market Selection and the Evolution of Industrial Structures. Small Business Economics, 7 (6), 411-436. Duh, M., Tominc, P. & Rebernik, M. (2009). Growth Ambitions and Succession Solutions in Family Business. Journal of Small Business and Enterprise Development, 16 (2), 256-269. Durnev, A., Morck, R. & Yeung, B. (2004). Value-Enhancing Capital Budgeting and Firm-Specifi c Stock Re- turn Variation. Journal of Finance, 59, 65-105. Ellison, G. & Glaeser, E. (1997). Geographic Concentration in U.S. Manufacturing Industries: A Dartboard Approach. Journal of Political Economy, 105, 889-927. Feldman, M.P. & Audretsch, D. B. (1999). Innovation in Cities: Science-Based Diversity, Specialization and Localized Competition. European Economic Review, 43, 409-429. Fujita, M. & Th isse, J.-F. (2002). Economics of Agglomeration. Cambridge, Mass.: MIT Press. Fujita, M., Krugman, P. & Venables, A. J. (1999). Th e Spatial Economy: Cities, Regions and International Trade. Cambridge, Mass.: MIT Press. Garofoli, G. (1994). New Firm Formation and Regional Development. Regional Studies, 28 (4), 381-394. Glaeser, E. et al. (1992). Growth of Cities. Journal of Political Economy, 100, 1126-1152. Guesnier, B. (1994). Regional Variation in New Firm Formation. Regional Studies, 28 (4), 347-358. Gujarati, D. (2004). Basic Econometrics (4th ed.). Boston, Mass.: McGraw-Hill. Henderson, V. J. & Th isse, J.-F. (Eds.) (2004). Handbooks of Regional and Urban Economics. Amsterdam: Elsevier. Highfi eld, R. & Smiley, R. (1987). New Business Starts and Economic Activity: An Empirical Investigation. International Journal of Industrial Organization, 5, 51-66. Karakaya, F. (2000). Market Exit and Barriers to Exit: Th eory and Practice. Dartmouth: University of Mas- sachusetts. http://74.125.155.132/scholar?q=cache:cntUsSf7NoUJ:scholar.google.com/&hl=sl&as_sdt=2000 &as_vis=1. Krugman, P. (1991a). Geography and Trade. Cambridge, Mass.: MIT Press. Krugman, P. (1991b). Increasing Returns and Economic Geography. Journal of Political Economy, 99, 483- 499. Lundvall, B.-A. (1992). National Systems of Innovation: Towards a Th eory of Innovation and Interactive Learn- ing. London: Pinter Publishers. Martin, P. & Ottaviano, G. I. P. (1999). Growing Locations: Industry Location in a Model of Endogenous Growth. European Economic Review, 43, 281-303. Martin, P. & Ottaviano, G. I. P. (2001). Growth and Agglomeration. International Economic Review, 42, 947-968. Maurel, F. & Sedillot, B. (1999). A Measure of the Geographic Concentration in French Manufacturing Indus- tries. Regional Science and Urban Economics, 5, 575-604. Močnik, D. (2009). Th e Impact of Slovenian Entries on Economic Performance. In Širec, K. & Rebernik, M. (Eds.), Dynamics of Slovenian Entrepreneurship: Slovenian Entrepreneurship Observatory 2008. Maribor: Faculty of Economics and Business, 47-62. 145D. MOČNIK | DETERMINANTS OF FIRM ENTRIES: EMPIRICAL EVIDENCE OF SLOVENIA Murphy, K. M., Schleifer, A. & Vishny, R. W. (1991). Th e Allocation of Talent: Th e Implications for Growth. Quarterly Journal of Economics, 56, 503-530. Nargundkar, S. V., Karakaya, F. & Stahl, M. J. (1996). Barriers to Market Exit. Journal of Managerial Issues, 8, 239-259. Nivin, S. R. (1998). Regional Innovation Potential: Th e Case of the U.S. Machine Tool Industry. Ph. D. Disserta- tion, University of Texas at Dallas. Penrose, E. T. (1959). Th e Th eory of the Growth of the Firm. Oxford: Oxford University Press. Rebernik, M., Tominc, P. & Pušnik, K. (2009). Rast podjetniške aktivnosti v Sloveniji, GEM Slovenija 2009. Maribor: Ekonomsko-poslovna fakulteta. Reynolds, P., Storey, D. & Westhead, P. (1994). Cross-National Comparisons of the Variation in New Firm Formation rates. Regional Studies, 28, 443-456. Rothwell, R. & Zegveld, W. (1982). Innovation and the Small and Medium Sized Firm. London: Printer Pub- lishers. Sutaria, V. & Hicks, D. A. (2004). New Firm Formation: Dynamics and Determinants. Th e Annals of Regional Science, 38, 241-262. Waltz, U. (1996). Transport Costs, Intermediate Goods, and Localized Growth. Regional Science and Urban Economics, 26, 671-695.