Important factors which predict entrepreneur ’ s perception in business risk

This paper seeks to examine the role of factors originated from outside (economic, political, competitive environment and relationships) and within (entrepreneur’s attitude) the organization on the business risk perceived by entrepreneurs. To test the hypothetical relationships, an ordinal regression with two link functions was applied on an original dataset of 641 small and medium-sized enterprises (SMEs) operating in Slovakia and Czech Republic. The analysis revealed that not only economic factors can predict business risk, but along with them are political and competitive environments, relationship with supply chain actors and entrepreneur’s attitude. Consistent with prior research, it is found that an unstable economic environment leads the business to expose themselves to business risk. Also, a friendly regulation framework and quality education contribute significantly to reducing the level of risk. The research triggers the interest of policymakers who design policies aimed at improving the business environment by reducing the level of risk that firms face in doing business. Also, this paper is useful for managerial perspective, since entrepreneur attitude was found to be a predictor of business risk. Gentjan Çera (Czech Republic), Jaroslav Belás (Czech Republic), Zdeněk Strnad (Czech Republic)


INTRODUCTION
The institutional environment within which firms operate determines the state of activity in the economy (Douhan & Henrekson, 2010). This implies that firm behavior is shaped by the changes in the institutional environment (Chowdhury, Audretsch, & Belitski, 2018;Stenholm, Acs, & Wuebker, 2013). Facing changes in the institutional environment, firms are exposed to a wide range of risks (García-Ramos, Gonzalez-Alvarez, & Nieto, 2017). Therefore, finding factors that predict the level of risks that firms face in doing business is vital, because exposing to risks may lead to a business termination. Managers try to adopt or apply different risk mitigation strategies in a way to avoid the failure of their business (Kim & Vonortas, 2014). On the other hand, the linkage between business risks and business failure should be a permanent concern for policymakers who tend to enhance entrepreneurship (Fereidouni & Masron, 2012). The latter can be achieved by designing policies aimed at reducing the level of business risks originated from the institutional environment (Welter & Smallbone, 2011). Thus, having a better picture over the factors that determine the level of business risk may be useful for reducing business risk. To fill this gap in the literature, the current study aims to examine the influence of institutional and business environments on business risk in the context of the Czech Republic and Slovakia, which are putting efforts to narrow the gap with advanced economies.
In the course of the two last decades, institutional theory (North, 1990) is used by scholars in studying the role of the institutional environment on entrepreneurship. The literature on entrepreneurship stresses the role of the institutional environment on business activity. Moreover, the institutional context can as well influence both the creation and termination of businesses, therefore, it is of interest to explore how the institutional environment affects business risk.
Shepherd, Wennberg, Suddaby, and Wiklund (2019) pointed out that there are research opportunities to examine further the effect of the institutional environment on the entrepreneurial endeavor. Despite the fact that research into the context and management of SMEs is consolidated and has matured with several sub-genres of focus (Herbane, 2010), studies which focus on business risk and its determinants are rare. Stimulated by Jenkins and McKelvie (2016) work which claims that business failure should not be considered as a single all-encompassing phenomenon, but rather one which includes a broader range of situations at different levels of analysis, we assumed that business termination is preceded by business risk. Thus, studying business risk and its determinants may complete a part of the entrepreneurship puzzle.
From pragmatist perspective, an enhancement of the SME sector is seen as vital for competitiveness and economic development at local, regional and national levels (Gao, Sung, & Zhang, 2011). In both countries, the Czech Republic and Slovakia, SMEs contribute significantly to 'non-financial business economy'. Compared to the European Union (EU) average, SMEs contribution to value added in both countries is slightly lower, while their contribution to employment is higher. According to the projections of European Commission (2018), during the years 2017-2019, SME value added is predicted to increase by 17.2% and SME employment by 4.4% in Slovakia, whereas in the Czech Republic these figures are expected to be 15.7% and 0.4%.
As Kelly and Ashwin (2013) have defined, the business environment consists of internal and external environments. The external business environment can be divided into macro and micro environmental factors, which affect the organization, but are beyond its control. These two environments do not act separately, but they affect each other. Taking together, it can be said that these interactions of business environments may expose the business to different risks. Therefore, in the current study, it is proposed that business risk is driven by economic, political and competitive environments along with business relationships with suppliers, competitors, employees and consumers, and entrepreneur's attitude.
The main contribution of this paper is to fill the gap found in the literature by providing new empirical evidence on the relations between business risk and business environment indicators, both from outside (economic, political, legislative, competitive environments, business relationships) and within the organization (entrepreneur's attitude). Furthermore, in the methodological aspect, the application of two link functions in an ordinal regression aims at giving more robust results.

LITERATURE REVIEW
In recent decades, scholars have studied the factors which drive business failure mainly from two perspectives: deterministic and voluntaristic (Amankwah-Amoah & Wang, 2019). Academics who used the deterministic perspective have accumulated facts and evidence, which claim that failure of business might be determined by external organizational factors over which firm managers do not have control. These external organizational factors consist of technological change, deregula-tion and competition. On the other hand, the voluntaristic perspective advocates that business failure is driven by factors in firm level such as poor leadership and not right management. Since these factors from both abovementioned perspectives may cause the failure of a business, it is reasonable to consider them as potential factors that may expose the business to different risks. In the next subsection, it is justify the inclusion of economic, political, competitive environments, business relationships (from suppliers to consumers), and entrepreneur's attitude. Boden and Nucci (2000) claimed that issues concerning business creation and survival or failure must be analyzed within the context of prevailing macroeconomic conditions. According to Boudreaux, Nikolaev, and Klein (2019), economic institutions moderate the socio-cognitive traits on opportunity entrepreneurship. If economic institutions contribute to the explanation of opportunity entrepreneurship, then it can be assumed that they may predict business risk. Thus, doing business in an unstable economic environment may lead the business to expose themselves to business risk ( (Karabag, 2019). Consequently, an instability originated from the macroeconomic environment increases the possibilities of a business to failure (Martinez et al., 2019). The growth of the Gross Domestic Product in a country is found to be a predictor of business failure (Eling & Jia, 2018). Hence, it can be said that the unstable macro environment leads to higher risks that business should face.

Economic environment
In the current study, access to finance is considered as a special dimension of the economic environment that influences business risk (Sauka & Welter, 2014). In the literature, it widely recognizes the linkages between access to finance (seen as financial difficulties) and failure of a business (Fang He, Sirén, Singh, Solomon, & von Krogh, 2018). In this line, it is found that ability to finance the venture is among the important source of the fear of failure (Cacciotti, Hayton, Mitchell, & Giazitzoglu, 2016;Kljucnikov, Belas, Kozubikova, & Pasekova, 2016). A logical extension in this line is that SMEs are seen as being dependent on external financial sources, and, as a result, a bank loan is the most frequent source available for financing business activity (Ardic, Mylenko, & Saltane, 2012;Falkner & Hiebl, 2015). Based on these considerations, businesses are not facing high risk levels when they can get easy access to finance.
Population consumption, changes in income and the structure of consumer expenditure can be a sub-dimension of economic environment. Autio and Fu (2015) found a positive association between population growth and firm entry. Moreover, having good relationship with the customers is considered by entrepreneurs as the most powerful component affecting their operations (Kadocsa & Francsovics, 2011). In addition, a study conducted in Poland concluded that as the population density increases, the firms' distress increases (Iwanicz-Drozdowska, Jackowicz, & Kozłowski, 2018). The latter is consistent with Thun et al.'s (2011) research, showing that, to meet their customers' needs, firms should offer a wide variety of goods and services. Nevertheless, this leads to higher dependence of firms on their supply chains because of high level of complexity (Falkner & Hiebl, 2015). Considering the above evidence and discussion, we may hypothesize an association between consumers and business risk.

Political environment
Beside economic environment, factors originated from the political environments such as legislation environment, regulation framework and educational system are seen critical for business risk. The domestic political condition as seen as a factor, which affects the business environment by shaping business behavior ( Karabag (2019), political risk consists of several risks that influence business performance, which might cause the failure of the business. Gao et al. (2011), by applying grounded theory, analyzed risk management capability building in SMEs, provided a useful perspective concerning the sub-genres of risk management, among them were regulation and procedures, education, networking and coordination.
Literature in the field of entrepreneurship suggests that an improvement in the quality of the legal environment lead to an improvement in the efficiency of the economy (Aristovnik & Obadic, 2015). This can be considered as an aftermath of entrepreneurship contribution. Facing with legal problems is an entrepreneur who involuntarily terminates (Fang He et al., 2018) and they affect business activity (Civelek, Ključnikov, Dobrovič, & Hudáková, 2016). The key business stakeholders can influence the legislative procedure either very limited (legitimate way) or effectively, but un-der the risk of losing their reputation. The cause is generally the deficit of legal engagement of business in related legislation, the dynamics and unpredictability of changes in legislation and the absolute absence of legitimate expectations on the part of the business.
Another strand of the literature sheds light on the relationship between the regulation framework and entrepreneurship in general. After a critical review on regulation, Mallett, Wapshott, and Vorley (2018, p. 19) claimed "the effects of regulation are not always direct, predictable or constraining. Instead, studies show how regulations vary in terms of relevance and significance for businesses, can offer competitive opportunities or protections and, ultimately, carry variable consequences". Similarly, prior research showed that the greater the regulatory complexity, the higher the business failure rate (García-Ramos et al., 2017). And this problem does not concern only the international or national level, but also the regional or local level.
Beside legislation environment and regulatory framework, the educational system may be seen as a dimension of political environment. Gao et al. (2011) identified education as a sub-genre of risk management. Quality education gained by society may equip even the firm's employees with skills and ability to deal with business risk. Higher level of education and management education is associated with business failure rate (Sauka & Welter, 2014). There is also evidence of the long-term effect on cognitive learning through simulations in the field of law (Krasnicka, Vojtko, Strnad, & Hruby, 2017). Based on the above discussion, we may assume that the political environment (represented by legislation environment, regulation framework and educational system) may expose businesses to extra risks while they perform their activity.

Competitive environment
Operating in an intensively competitive environment might expose the business to risks that may cause their termination. According to a study conducted by Delerue and Perez (2009), SME managers have a tendency to involve into a cooperative relationship when they perceive "a high level of relational risk or when they hope to gain a future competitive advantage over the competitor" (Falkner & Hiebl, 2015, p. 135). Another study (Cai & Yang, 2014) analyzed the relationship between competition and business environment indicators. The latter paper found that business environments characterized by intensive competition usually prompt customers to demand quality goods and services at low prices, and firms need effective operations processes to meet this increased demand. Facing with this type of situation is translated into additional risks for businesses. Hence, there is an association between competitive environment and business risk.

Business relationships (network)
Having problems or performing not good relationships with suppliers, competitors, employees and consumers can be a real problem for entrepreneurs to do business. Fang He et al. (2018) argued that partnership disputes lead to business failure. So, for SMEs it can be vital to have a network with other firms and interested parties (Gabbianelli, 2018). Moreover, cooperation with vertical partners reduces the effect of business failure in manufacturing sectors (Martinez et al., 2019). As mentioned earlier, according to Gao et al. (2011), networking and coordinating was identified as a sub-category of relationship that influences risk management in SMEs. Kim and Vonortas (2014) also support that networking is a risk mitigation strategy in SMEs.
Another way to manage the risk is to have close relationships with suppliers and customers. Gilmore et al. (2004) stated that continuing relations with actual customers increases the chance of having future business transactions. The latter studies show also that these relationships are created by using SME networks. Also, having good relationships with competitors can contribute to reducing risky transactions. Concerning their relationship with employees, SMEs rarely offer employee programs aimed at enhancing their know-how, skills and abilities (Sukumar, Edgar, & Grant, 2011). Almost all type of businesses may lose workers having vital know-how and being well-skilled (Rauch & Rijsdijk, 2013). Therefore, Gilmore et al. (2004) argued that losing both managers and staffs can cause extra risk for SMEs, because usually only a few individuals in the firm have the appropriate knowledge to deal with something.

Entrepreneur's attitude
From the voluntaristic perspective (Amankwah-Amoah & Wang, 2019), poor leadership and mismanagement may cause business failure. Based on this school of thought, it can be concluded that the entrepreneur's attitude is important in managing the business. Therefore, the entrepreneur's attitude may expose the business to different business risk. A logical extension in this line is that personal ability is among the most important source of the fear of failure (Cacciotti et al., 2016). However, according to Kibler, Mandl, Kautonen, and Berger (2017), entrepreneurs tend to distance themselves from the failure by claiming that external forces, which are not under their control, cause it.

Aim and unit of analysis
The aim of this paper is to examine the effect of factors originated from outside and within the organization on the business risk perceived by entrepreneurs. The analyses are conducted on a firm-level data collection through a survey in 2018. Thus, the unit of the analysis in the current research are firms operating in the Czech Republic and Slovakia. 641 firms were randomly selected from two public databases ("Albertina" in the Czech Republic and "Cribis" in Slovakia). As with Jolley, Lancaster, and Gao (2015), the respondent to our survey was either the owner or a representative person from firm's top management.
They were considered to be the appropriate individuals to represent the firm's viewpoints. Doing so, similar to Cepel, Stasiukynas, Kotaskova, and Dvorsky (2018), we have captured the entrepreneur's perception.

Dependent variable
Similar to Acar and Göç (2011), entrepreneur's perception in business risk is the dependent variable in the current research. As Mallett et al. (2018, p. 16) discussed in their paper, "studies that survey owner-manager perceptions will capture some insights into identification-interpretation processes". It was measured as a single item, which was formulated: "The business environment of my country bears the adequate risk and enables to start a business" (Cepel et al., 2018, p. 29). Respondents were asked to choose one of the five listed options: 1 -'totally disagree' to 5 -'totally agree'. For our further analysis, we transform this variable into a new one with three categories: 1 -'totally disagree' or 'disagree', 2 -'neither disagree, nor agree', 3 -'agree' or 'totally agree'.
In Table 1, some descriptive statistics of our dependent variable are summarized. As can be seen, its mean did not differ a lot between two countries (2.28 and 2.21 for Czech and Slovakian firms, respectively). Acar and Göç (2011) argued that entrepreneurs' risk appetite might change across countries. We applied the Mann-Whitney test to check for any potential difference in business risk between Czech and Slovakian entrepreneurs. The results of this test showed that there is no statistical significance in this regard. This means that, somehow, business risk is perceived similarly Note: a -assuming a scale variable; b -almost the same results were found even when business risk was measured as a fivepoint Likert type scale. SD stands for standard deviation.
among entrepreneurs operating in both countries, due to the fact that both countries share a similar culture. This is consistent with another study, which found no differences between these countries on how optimistic entrepreneurs are for the future business activity (Belas, Machacek, Bartos, Hlawiczka, & Hudakova, 2014). Therefore, the country will not be considered as an independent variable. In the second part of the H3: The more normal the competitive environment, the lower the business risk.
H4: The higher the relationships with supplier chain actors, the more adequate business risk.
H5: Positive entrepreneur's attitude toward entrepreneurship reduces business risk.

Reliability test
Before constituting composite variables by calculating the mean of the items or indicators, the reliability of the scales was checked. Reliability test is a statistical tool scholar use to checks whether the measure reflects the construct that it is measuring or not. Table 2 shows the mean, standard deviation, Cronbach's alpha for the composed variable and the correlation coefficient. Based on DeVellis's (2017) benchmarks, the construct reliability of access to finance (finance) was very good (.814), legal (.765) and education (.706) constructs were respectable (between .70 and .80), whereas the rest of constructs were minimally acceptable (between .65 and .70). In addition, regarding multicollinearity issue, all correlation coefficients were no bigger than .70 threshold, which mean that there will not be any multicollinearity problem if all the listed variables are included in a regression leading to the fact that there is no redundant variable (Dormann et al., 2013).  (Harrell, 2015). They predict the probability of a certain level or category of the dependent variable (γ) occurring with respect to the known values of the independent variables (X i ), and their formulae are presented in Table 3.
Constrained by the distribution of the categories of the outcome variable (see the second part of

EMPIRICAL RESULTS
The mean, standard deviation and number of observations by entrepreneur's perception in business risk categories for each composed variable are demonstrated in Table 4. All means of these variables had a positive trend across the business risk levels. Based on these trends in the data, it was expected that ordinal regression would find a positive relationship between business risk and composed variables covering the business environment. However, taking into account the differences between levels per each pair, it can be judged that those variables that have a low magnitude between business risk levels are expected to be not significant. Possible insignificant effect on business risk could be noticed at these variables: legal (min(diff) = 0.09), regulation (min(diff) = 0.106), education (min(diff) = 0.139), competition (min(diff) = 0.093), and network (min(diff) = 0.066). Ordinal regression will test and provide statistical evidence of whether or not these variables predict the entrepreneur's perception of business risk.
An ordinal regression analysis was performed to assess the prediction of affiliation with one of three outcome levels on the basis of nine covariates. Our outcome variable was business risk (1 -'totally disagree' or 'disagree', 2 -'neither disagree, nor agree', 3 -'agree' or 'totally agree') and the covariates used were three variables covering economic dimension of business environment, three other covering political environment, competition, relationship with suppliers, competitors, employees and consumers, and entrepreneur's attitude. After checking for cases with missing values on our covariates, data of 461 firms remained suitable for analysis (the minimum required sample size is 184 subjects (Harrell, 2015, p. 363)). In respect to the distribution of the dependent variable and to get more robust results, two types  In Table 5, the output of both conducted models is summarized. Regarding the case when cauchit was applied as link function, the results indicate that the overall model was statistically significant, χ 2 (9, n = 461) = 204.341, p < .001. Moreover, there was a good model fit (discrimination among levels) on the basis of our nine covariates, χ 2 (1267, n = 461) = 1217.863, p = .835, using a Pearson criterion. Additionally, the slope coefficients in the model were the same across dependent variable categories (and lines of the same slope were parallel), as the parallel lines assumption was not violated, χ 2 (9, n = 461) = 9.519, p = .391 (see first part of Table 4). If this assumption is violated, then a less restricted model should be used, i.e. the multinomial logit model (Harrell, 2015). Pseudo R-square of the model was reported to be .273, according to Cox and Snell statistic, or .317 based at Nagelkerke statistic.
As cauchit, similar results were found even in the case when complementary log-log was applied as a link function (see the second part of Table 5). The fitting of the model was significant, χ 2 (9, n = 461) = 208.669, p < .001, and Pearson criterion indicated a good model fit, χ 2 (1267, n = 461) = 1254.192, p = .596. In addition, its test of parallel lines was not violated demonstrating that slope coefficients are the same across the business risk categories, χ 2 (9, n = 461) = 6.921, p = .645. Pseudo R-square of the model was reported to be .278, according to Cox and Snell statistic, or .323 based at Nagelkerke statistic. As a result, taking all together, both link functions are not misleading.
In Table 6, the summary of parameter estimates of the ordinal regression when cauchit was applied as a link function is given. To distinguish the outcome variable categories, the ordinal regression calculates a continuous latent variable (Harrell, 2015). The thresholds [risk = 1] and [risk = 2] stand for the dependent variable in the ordi- Note: The numbers in parentheses are standard deviations. Disagree (1) and agree (3) categories correspond to strongly disagree or disagree and agree or strongly agree, respectively.  Table 4). On the other hand, differently from what it was expected, regulation, education and competition statistically predicted the entrepreneur's affiliation with one of the three levels of business risk. Taking all together, business risk is perceived to be more adequate when an improvement at economic environment occurs concerning macroeconomic dimension, access to finance and consumers; by doing a friendlier business environment by reducing business regulations; when efforts to get better education toward entrepreneurship are increased; when the completive environment is not very intensive.

DISCUSSION AND CONCLUSION
Our findings emphasized that a stable and progressing economic environment can be translated into fewer risk firms face during their activity, which goes in line with previous studies ( Turning back to the political environment, as it is mentioned earlier, the legal environment resulted to be not significant in predicting business risk, but regulation was. It was expected for these two factors to have similar results. We believe these findings are due to the fact that businesses face with more challenges and problems related to regulatory framework dealing with their operational activities rather than legislation, which changes not that often, compared to the regulatory framework. This is consistent with what Xheneti and Bartlett (2012) found, which was that tax regulation was the second most significant constraint identified by entrepreneurs, and surprisingly, legislation was not among them. In addition to this, Peck et al. (2018) argue that small and micro businesses are affected by regulation in a way that they are aware of the costs related to compliance with regulation framework regarding the additional money needed to be allocated to purchase specialist external advice and managing time. Our finding regarding regulation framework is in line with García-Ramos et al.'s (2017) conclusion that a complex regulation framework increases the possibility of business termination due to facing extra risks.
The competitive environment is found to be a predictor of the level of business risk entrepreneurs' perception. Our findings concerning intensive competition as a factor that impacts business risk are consistent with previous studies (Cai & Yang, 2014;Delerue & Perez, 2009). Operating in a not intensive competition may not expose firms to extra risks. The other factor which resulted to be insignificant was the one related to business relationships (network) with suppliers, competitors, employers and customers. Even though our initial intention was to capture all spectrums of business relations into a single composite, methodologically it might be fragile and not rigour. Capturing several relations (from suppliers to customers) into the same construct might lead to a scale that is not valid. With this in mind, we suggest that scholars put some extra efforts on developing more valid scale concerning business relationships. This could be considered as an issue for further research.
Both types of ordinal regressions revealed a statistical significance of entrepreneur attitude for business risk. As Cacciotti et al. (2016) claimed, personal abilities and attitude of the owner-manager can influence business risk. Having a positive attitude toward entrepreneurial activity reduces the risk business might face in daily activity.
The current research took a combined deterministic and voluntaristic perspective (Amankwah-Amoah & Wang, 2019) by using external and internal factors to predict the entrepreneur's perception of business risk. The deterministic perspective assumes that business failure is determined by external factors from the organization, which are not easy to be controlled by the owner-managers. Such factors might be macroeconomic, indicators, access to finance, changes in population and consumers income, political environment including legislation, regulation framework, educational system, competitive environment, and relationship with other actors in the supply chain. On the other hand, voluntaristic perspective claims that business failure is driven by internal forces within the organization. Since these factors originate from inside the organization, then a certain level of control over them can be assumed. Entrepreneur's attitude could be a representative factor for this perspective.
This paper contributes to the stock of knowledge in two ways: firstly, by testing the impact of different business environmental indicators on business risk perceived by owner-managers of firms operating in Czech Republic and Slovakia, countries trying to get closer or narrow the gap with the advanced economies, and, secondly, by using two type of ordinal regressions to get more robust results. Our findings emphasized that economic factors are not the only ones, which predict the entrepreneur's perception of business risk. Our study found that other factors originated from political and competitive environments and the entrepreneur's attitude determine owner-managers' perception of risk. Nevertheless, legislation environment and business relationship were not revealed as predictors of business risk.
Even though our paper has reached its aims, there are limitations in research. Firstly, study's findings are limited to two countries, which might share a similar regional, economic, institutional, historical and political environments and entrepreneurial mindset, as well with only a limited number of countries. Therefore, our findings can be generalized only for countries having an almost similar business environment with the Czech Republic and Slovakia. Secondly, although Mallett et al. (2018) stressed out that studies, which survey owner-manager perceptions, are useful for identification-interpretation processes, we believe that some external factors could be measured differently. In this regard, using secondary data could be a good approach even to validate our results.