Does informality help entrepreneurs achieve firm growth? evidence from a post-conflict economy

Abstract This article evaluates the impact of starting-up unregistered and operating informally on small firm growth in Kosovo, the country with the highest levels of informality in Western Balkans. The study uses mixed research methods as an inquiry to combining both qualitative and quantitative forms. Reporting data from 487 business owners extracted from a 2017 nationally representative survey of 8,533 households in Kosovo, 47.7 percent had started-up unregistered and were operating informally at the time of the survey. Using an instrumental variable (IV) estimation with a binary endogenous regressor to estimate the impact of informality (a treatment group) on small firm growth (control group), while controlling for other entrepreneur and firm determinants, a strong positive effect of the informality on firm growth is identified. Entrepreneurs operating informally had an 11.6 percentage points higher probability of achieving their firm growth objectives compared with entrepreneurs operating formally. Yet, the qualitative interviews revealed that growth objectives of small firms were limited, so the unregistered firms outperformed registered firms only under the modest growth objectives. Taking into account the limitations of the study, the implications for theory and policy are proposed to tackle the informality such as incentives of government small business support programs.


Introduction
Over the past decade or so, the discipline of entrepreneurship has witnessed the emergence of a burgeoning sub-field that examines informal sector entrepreneurship (see Pejic Bach et al., 2018;Sauka et al., 2016;Webb et al., 2014).The present study contributes to this strand of literature by evaluating whether firms that started-up their operations without registration and currently continue to operate informally have higher firm growth relative to firms that registered from the very beginning of their operations.
Until recently, informal sector entrepreneurship has been widely portrayed in a negative manner as poorly performing unproductive endeavour, which is harmful to economic development and growth (Baumol 1990;La Porta & Shleifer, 2008Sobel, 2008). The argument has been that informal entrepreneurship is associated with weaker firm performance (e.g., La Porta & Schleifer, 2014). In recent years, however, a small emergent literature has begun to challenge this view by showing that non-registration has a positive impact on subsequent growth. The aim of this paper is to contribute and advance this literature by disentangle the relationship between informal entrepreneurship and firm growth using fresh data from a 2017 survey conducted in Kosovo, and explaining this in terms of the institutional environment. Study also uses qualitative interviews to conceptualise growth objectives of small firms used in quantitative study.
This paper therefore advances scholarship on informal entrepreneurship in three ways. First, and theoretically, it contributes to a better evidence-based understanding of informal entrepreneurship. The paper provides both quantitative and qualitative empirical evidence to resolve the dilemma about the specified relationship, especially given the conflicting results of previous studies. Second, the original empirical contribution of this paper, using data on Kosovo, is to reveal how unregistered enterprises outperform those registering from the outset using an instrumental variable (IV) estimation with a binary endogenous regressor to show the impact of informality (a treatment group) on small firm growth (control group), while controlling for other entrepreneur and firm determinants. Finally, and from a policy perspective, it contributes to the redesign of informal sector policy and entrepreneurship support strategy by identifying specific initiatives to tackle the weak formal institutional environment and institutional asymmetry that prevent the transition to formal sector entrepreneurship.
To achieve this, Section 2 reviews the competing theoretical perspectives on the relationship between registration at start-up and future firm performance. Section 3 then introduces the mixed research methods. Section 4 reports the data and modelling framework, namely data from a study of 487 business owners in Kosovo, analysed using an instrumental variable (IV) estimation with a binary endogenous regressor to estimate the impact of informality (a treatment group) on small firm growth (control group), while controlling for other entrepreneur and firm determinants. Section 5 reports the findings from the qualitative analysis. Section 6 then discusses the implications for theory and policy. Taking into account the limitations, the study proposes some policy proposals and future research direction.

Non-registration and firm performance: literature review and hypothesis
Conventionally, entrepreneurs operating in the formal economy were deemed as the "mainstream" and appropriate focus of enquiry. In recent decades, however, there has been recognition that two-thirds of all enterprises are unregistered at startup (Autio & Fu, 2015) and that half of enterprises operate on an unregistered basis (Stenholm et al., 2013). This recognition of the prevalence of informal entrepreneurship has resulted in the rise of a burgeoning literature and new theorizations of such entrepreneurship (Aidis & Mickiewicz, 2006;Morris & Polese, 2014;Schneider & Williams, 2013).
Recently, significant advances have been made in understanding informal entrepreneurship by scholars adopting an institutional perspective. Entrepreneurship is viewed as socially constructed behaviour (Sine & David, 2010;Webb & Ireland, 2015), and institutions as "the rules of the game" which prescribe, monitor, enforce and support what is socially acceptable (Baumol & Blinder, 2009;Chaudhry et al., 2018;Denzau & North, 1994;Mathias et al., 2015;North, 1990;Shan et al., 2018;Webb et al., 2009). All societies have both formal institutions (i.e., codified laws and regulations) that set out the legal rules of the game, as well as informal institutions which are the "socially shared rules, usually unwritten, that are created, communicated and enforced outside of officially sanctioned channels" (Helmke & Levitsky, 2004: 727). Informal entrepreneurship is thus endeavor that takes place outside of formal institutional prescriptions but within the norms, values and beliefs of informal institutions (Godfrey, 2011;Kistruck et al., 2015;Siqueira et al., 2016;Webb et al., 2009;Welter et al., 2015).
Examining the scholarship on the impacts of entrepreneurs deciding to operate informally, the widespread argument has been that enterprises operating under the guiding framework of the informal institutional environment are less efficient and poorer performing than those operating in formal institutional environments (Benjamin & Mbaye, 2012;La Porta & Schleifer, 2008. Firms operating legitimately are argued to have higher levels of revenue and profits, employ more workers, and to be more capital intensive than their informal counterparts (Fajnzylber et al., 2011;McKenzie & Sakho, 2010). It is also asserted that that registration by informal firms leads to higher firm performance than if they remained unregistered (Demenet et al., 2016;Fajnzylber et al., 2011;Rand & Torm, 2012).
One prominent assumption driving much of this literature has been that firm registration is a beneficial endeavour. Grounded in the long-standing concept of liabilities of newness (Stinchcombe, 1965), start-ups are argued to lack a track record of past performance on which they can claim legitimacy, reliability and accountability (Delmar & Shane, 2004) and be deemed competent, effective, and worthy (Zimmerman & Zeitz, 2002). Registration, therefore, is deemed to be a means of enhancing firm legitimacy (Kistruck et al., 2015) and signaling to clients stability, quality and prestige (Bitektine, 2011). Conversely, non-registration is viewed to result in a lack of legitimacy and as negatively influencing firm performance (ILO, 2007;La Porta & Schleifer, 2008). Accordingly, unregistered firms will therefore suffer from worse firm performance than ventures registered from the outset, all other things being equal. However, this has been seldom evaluated. One exception is Perry et al. (2007) who compare the firm performance of formal enterprises starting-up unregistered with enterprises registered from the outset of operations. Analyzing data on just 355 unregistered start-ups in seven Latin American countries, they find that formal enterprises delaying registration suffer poorer firm performance than those registered from the outset.
Besides the weak evidence, there are also theoretical rationales for revisiting the relationship between registration and firm performance. For entrepreneurs to survive and grow, they must gain legitimacy (Scott, 2008). From an institutionalist theoretical perspective, however, the view that nonregistered firms lack legitimacy fails to recognize that they might lack legality in relation to the laws and regulations of formal institutions. However, they are often deemed socially legitimate in relation to the norms, values and beliefs that constitute the informal institutions (Siqueira et al., 2016;Webb et al., 2009) and consequently have legitimacy in the eyes of their stakeholders (e.g., consumers, suppliers and employees). This is especially so in contexts where formal institutional imperfections are greater than in the western countries (Kistruck et al., 2015;Webb et al., 2014).
Unregistered enterprises can thus achieve legitimacy with their stakeholders whilst avoiding the costs of registration (Tonoyan et al., 2010). This is particularly important in institutional contexts where the benefits of registration (e.g., access to credit, training and participation in business associations) are insufficient to outweigh the benefits of non-registration (e.g. avoiding aggressive rent-seeking by officials). Indeed, McCulloch et al. (2010) in Indonesia and McKenzie and Sakho (2010) in Bolivia reveal how registration only increases firm performance in middle-sized firms, intimating that non-registration until a certain size is reached may be an appropriate strategy. In Mexico, McKenzie and Woodruff (2006) find that smaller firms view the benefits of registration as limited and as just another significant cost imposed by the formal institutional environment. Unregistered firms may therefore outperform registered firms since they have the pre-requisite of legitimacy but without paying the costs of registration (La Porta & Schleifer, 2014).
By not registering, moreover, they can use their resources to overcome other liabilities of newness. In order to survive, firms internally need to build operational routines, trust and cohesion, and learn unfamiliar new roles, and externally establish market acceptance and stable links with partners, suppliers, customers and investors. Addressing these liabilities of newness is costly. By evading the costs of registration, resources can be devoted to these other liabilities of newness, which might result in them outperforming those who devote their resources to formal registration, which may result in relatively few benefits in developing and transition country contexts. Given these arguments regarding the relationship between registration and firm performance, the following hypothesis can be tested: Hypothesis: unregistered enterprises are more likely to achieve the growth they desire than enterprises registering from the outset, after controlling for other key determinants of firm performance.

Mixed research design
The research design used in this paper is based on mixed research methods. Mixed methods research is an approach to inquiry that combines both qualitative and quantitative forms (Aramo-Immonen, 2011). The study uses large scale representative sample survey and qualitative interviews with entrepreneurs. First, the large sample survey of entrepreneurs is used to test whether the firms operating informally outperform those operating formally. We use qualitative interviews to get better insights on why firms choose to operate formally and what are their desired growth ambitions.
The qualitative interviews are used to help better understand the findings from the econometric section. We use qualitative method for the collection of qualitative datathat is the way 'why' and 'how' a phenomenon is constructed from the perspective of the respondent (Saunders, 2011;Bell et al., 2018). According to them, this interviewing is adequate method when the concept is inappropriate or difficult to measure, such as the case of "desired firm growth" which is our dependent variable in quantitative analysis. Thus, paper seeks to provide answers on how companies define their 'desired growth' and 'why' small firms choose to operate unregistered.

Data and sample
To assess the impact of nonregistration on firm growth, data for 487 businesses owners is extracted from the labour force survey commissioned by the Millennium Challenge Corporation (MCC) conducted in 2017 with 8,533 households. The survey used multi-stage stratified cluster-randomized sampling to arrive at a representative sample for Kosovo. 1 The research design and dataset is comparable with other datasets aligned to the Eurostat approach. The unit of analysis in the survey is persons aged 15 and over living in private households and in employment who did at least one hour of work for pay or profit during the week preceding the interview. The survey also included an entrepreneurship component in line with the Global Entrepreneurship Monitor (GEM) to collect data on individuals' perceptions of entrepreneurship and their involvement in entrepreneurial activity (both actual and aspirational). Therefore, the sub-sample of individuals analysed here are entrepreneurs (8.5% of total sample) who completed an extended interview module to identify their entrepreneurial activities and aspirations. Respondents were asked not only about registration at start-up but also if they still operate informally at the time of the survey. In addition, respondents were asked whether they own the business, their firm performance in terms of growth, and about the firm characteristics, such as its size, number of employees, sector of activity, and their perceptions of the business environment.

Variables
Dependent variable. Several firm growth indicators are used as dependent variables in the entrepreneurship and small firm growth literature, the most common of which are employment growth, sales growth, and productivity growth (Lajqi & Krasniqi, 2017). For transition country contexts, employment growth is most widely used in empirical studies of small firm growth (Bartlett & Bukvi c, 2001;Capelleras & Rabetino, 2008;Nason & Wiklund, 2018). Some scholars, however, argue that subjective growth measures can be even better and more successfully employed as estimators of organizational performance and growth (Singh et. al., 2016), especially in transition countries. In the transition economy context of Kosova, there are additional reasons to opt for such a subjective growth indicator. As in other transition economies where firms commonly underreport their activities, self-reported satisfaction with whether they have achieved their growth objective is more reliable than reported indicators of firm growth such as sales. Similarly, employment figures are likely to be underreported, especially where there is unregistered employment or under-declared employment in terms of either time worked or wages paid (Aidis & Mickiewicz, 2006;Brown et al., 2005;Galloway & Mochrie, 2006). In this paper, therefore, a subjective growth measure reported by entrepreneur is the dependent variable. Respondents who declared "yes" to the following question, "Have you been able to grow your business as much as you want to?" were recorded 1, and 0 otherwise. Independent variable. To evaluate the impact of nonregistration on firm growth, a binary variable is used coded as 1 if respondent declared "yes" to the question: "Are you formally registered at the Agency for Business Registration of Kosovo", and 0 otherwise.
Control variables. To measure the impact of nonregistration on firm growth, it is necessary to control for other key determinants of firm growth. In line with previous studies on firm growth, several control variables are added. Given that firm size is considered to be a key variable in any firm growth model which improves model fit, in line with other studies, "firm size" is included, measuring the number of employees in the firm. Sector of activity also has an important effect on firm growth. Evidence shows that performance significantly varies across sectors (Nabar & Yan, 2013;Siqueira et al. 2016). Perry et al. (2007) suggest that given that unregistered firms may be heavily concentrated in labour-intensive sectors with fewer returns to scale, controlling for sector is important. Sector is here a categorical variable indicating the sector of the firm.
There is in addition ample evidence of the impact of different types of entrepreneurship on firm growth. The literature often refers to entrepreneurs as either necessity-driven or opportunity-driven. The clear view is that necessity-driven entrepreneurs are poorer performing compared with opportunity-driven entrepreneurs (Benjamin & Mbaye, 2012;La Porta & Schleifer, 2008. Therefore, we included a dummy variable "necessity-driven entrepreneur" coded as 1 for those who stated "I would prefer to do something else, but there are no other opportunities available to me" in response to the question "Why are you involved in this business?", and 0 otherwise. To take into account the importance of urban location we control for location for firms in cities. The literature argues that the firms in cities may benefit from agglomeration economies and networking which translates into firm growth (see Krugman, 1991;Thisse, 2002). This line of literature argues that positive externalities in developed and dynamic urban areas have highlighted the important role of location for economic growth in general and for the growth of firms in particular (Alc acer & Chung, 2007;Black & Henderson, 1999). These benefits include knowledge spill-over and the role of urban institutions that lead to efficient growth, access to specialized resources that may not be able to be developed internally but are found in urban areas (Maine et al., 2010).The variable "Urban location" takes a value of 1 if located in urban areas (cities) and 0 otherwise.
An entrepreneur-level factor important in explaining firm growth is human capital, such as the educational level, especially university degree, and the skills and experience of the owners, which are found to have a significant impact on firm performance (Belas et al., 2019;Black & Lynch, 1996;Camis on-Haba et al., 2019;Gennaioli et al., 2013) and also contribute to social and technological change (Dalmarco et al., 2018;Sansone et al., 2019). However in some previous studies in transition, university education is found to have negative impact on firm growth signalling the low quality of the education system in such countries (Bartlett & Bukvi c, 2001;Xheneti & Bartlett, 2012). We therefore introduce the variable 'Education" coded as 1 if the entrepreneur holds a university or postgraduate degree and 0 otherwise. We also include the variable "entrepreneur's age", which is a continuous variable indicating the number of years engaged in such endeavour which is a proxy for skills and experience. Two other individual variables "gender" coded as 1 for male and 0 otherwise and "marital status" were introduced as control variables too, reflecting literature which asserts that this influences firm growth.

Empirical model
To develop an empirical model to estimate the impact of nonregistration on the likelihood of an entrepreneur achieving their desired growth, it is necessary to recognize that the registration of an enterprise is an endogenous choice, suggesting that the sample is not random, raising the issue of the sample section bias. Therefore, taking into account the binary nature of both the variables nonregistration and the variable firm growth, and controlling for potential endogeneity problem, we adopt the instrumental variable (IV) estimation with a binary endogenous regressor (similar to Heckman twostep). Following Cameron and Trivedi (2010, p.186), the estimation takes into account the latent binary variable of the endogenous regressor by changing the nature of the first -stage model to be a latent variable model similar to the probit. Let y 1 (firm growth) depend in part on y 2 (nonregistration) as an endogenous binary variable. We introduce an unobserved latent variable, y Ã 2 that determines whether the y 2 ¼ 1 or 0 or whether the firm operates unregistered or not. Formally, we represent our model as follows: The errors ðu i , v i Þ are assumed to be correlated bivariate normal with Var u i ð Þ ¼ r 2 , Var v i ð Þ ¼ 1, and Covðu 1 , v i Þ ¼ qr 2 : The binary endogenous regressor y 2 (nonregistration), can be viewed as a treatment indicator. If y 2 ¼ 1, the entrepreneur received treatment (i.e. nonregistration) and if y 2 ¼ 0, otherwise. We apply the variable perception of the entrepreneurship environment as the single instrument to identify the equation. This instrument is expected to have influence on the likelihood of being unregistered but not on the firm growth.
The proposed model helps address the endogeneity issue (i.e., unobserved heterogeneity between registered and unregistered firms). By estimating a first-stage model that predicts the likelihood of registration and including the correction factor (i.e. unregistered) in the second-stage model that predicts firm performance. The registration is an endogenous choice of firms, and as such, the sample of firms available for analysis cannot be taken as random.
To proceed with estimation of the model, we have included an instrumental variable, to identify the equation in the first stage. For the purposes of the identification strategy, we argue that the general perception of the entrepreneurs about starting-up and entrepreneurship as a career choice are critical in affecting the endogenous decision of firms to register or not. The instrumental variable in this case is the perception of the entrepreneur about starting an entrepreneurial career and the status of the entrepreneur in society. This variable is defined as binary which assumes value of 1 if entrepreneurs responded the general environment for pursuing an entrepreneurial career as positive and 0 otherwise. It was included in the first stage equation (the probability of being unregistered) and excluded in the second-stage equation (i.e. the primary equation estimating firm growth). The variable indicating the entrepreneur's perception about the desirability of the entrepreneurship career choice is supposed to directly influence the registration decision, but is unlikely to affect the firm growth directly.

Econometric findings
The first finding is that of all entrepreneurs interviewed, 52.3% had formally registered with the Kosovo Business Registration Agency (KBRA), while 47.7% had not and were therefore unregistered enterprises. Table 1 reports why the entrepreneurs had not registered their enterprise. The most common reason, cited by 54% of entrepreneurs who had not registered their enterprise, was because did not see the benefit in doing so. This mirrors in Kosovo the finding in many developing world contexts, discussed above, that entrepreneurs do not see the benefits of registration. The second most important reason, again reflecting the earlier discussion, is that entrepreneurs cannot afford to do so. The costs of registration prevent them from doing so.
What, therefore, is the relationship between nonregistration and firm performance? Table 2 reports a set of regressions and if the error correlation q ¼ 0, then the errors u i and v i are independent and there is no endogeneity problem. The last line in  Table 2 cleraly rejects H 0 : q ¼ 0, suggesting that the nonregistration variable is indeed an endogenous regressor (v 2 (1) ¼ 75.87; p¼ Prob >0.00). In line with the econometric procedure to deal with endogeneity issues in our model, we have included our variable 'Perception of entrepreneurship environment' which enters highly statistically significant.
The key finding is that nonregistration is positively and significantly associated with a greater likelihood of achieving the entrepreneurs' desired firm growth. To elaborate more, Lokshin and Sajaia (2011), based on Aakvik et al. (2005), suggest that after estimating the model's parameters, the effect of the treatment on the treated, or the expected effect of the treatment on individuals with observed characteristics x who participated in the program (in our case the firms who started up unregistered). Accordingly, while controlling for other entrepreneur and firm determinants of growth, the finding is that entrepreneurs operating unregistered had an 11.6 percentage points higher probability of having achieved their growth objectives compared with entrepreneurs who had registered their business from the outset of operations. This finding support the hypotheses that nonregistration has a positive significant influence on firm growth.

Selection of cases
We combine the firm's survey data with information taken from qualitative surveys of sample of 8 small firms operating formally and informally (see Table 3). We conducted in-depth interviews with a small number of firms to better understand the choices they make and the institutional environment in which they operate (e.g. Rothenberg et al., 2016). We considered also, that the construct of entrepreneurial growth-orientation and growth objectives would be a useful indicator to get a clearer picture for the conceptualization of dependent variable ("desired firm growth') from quantitative analysis. We engaged with ideas of motivation to go formal and informal and to compare our cases in a thematic, inductive analysis. We tried to learn more about small firm's owners that operate in formal and informal sector, and investigate the difference in their expectations in the growth of their firms. This will help to cross-validate the results from the quantitative analysis used in previous section. For this purposes we used purposeful sample, on the characteristics of SMEs that interested us (see Anderson et al., 2013). The sample of 8 cases is extracted from the sample used in quantitative analysis. Following Eisenhardt's (1989) we ensured diversity of cases in order to reflect extreme situational and polar types. Therefore, the entrepreneurs selected for this study varied in size, experience, sector, and other characteristics (Domi & Krasniqi, 2019). The selected companies (cases) should be categorized as small firms (less than 250 employees) 4 of which operate formally and 4 of which informally. In this sense, the 4 selected informal companies have similar size characteristics and belong to the same sector of operation to 4 formal businesses. This approach was used to make comparisons across the same sector because as literature argues some sectors are more biased towards informal business activities the study used open-ended interviews to find patterns that made sense of the problem (see Thornhill et al., 2009).
Data were collected through personal in-depth interviews conducted in Kosovo during the December 2019-January 2019. Interviews lasted about 1 hour. In all cases, the owner/founder was interviewed. To ensure the anonymity of reported responses, each entrepreneur was allocated a case code descriptor. Following protocol developed by research team interviews consisted of 2 questions: How you explain your "desired firm growth" and "what are the reasons for operating in formal or informal sector".

Findings from interviews
This section provides qualitative analysis from interviews with entrepreneurs. The key issues in the literature and theory of informality remains whether the formality vs informality decision depends on the growth objectives of entrepreneurs. The widespread argument of the literature is that that enterprises operating informally are less efficient and poorer performing than those operating in formal institutional  environments (Benjamin & Mbaye, 2012;La Porta & Schleifer, 2014). Firms operating legitimately are argued to have higher levels of growth and profits, employ more workers, and to be more capital intensive than their informal counterparts (Fajnzylber et al., 2011;McKenzie & Sakho, 2010). However, in specific post-conflict and transition context is especially crucial for small firms which largely have modest growth objectives motivated by push factors (Krasniqi, 2014;Krasniqi & Lajqi, 2018).
To better understand the concept of desired growth as defined in our quantitative analysis, we have asked entrepreneurs about their future growth objectives in 5 years period. In line with discussion on entrepreneurial orientation by Ling et al. (2019) we controlled for growth orientation of the small firms which is crucial for explaining quantitative results in the previous section. Our interviewees found that majority of cases (6 from 8) explained their desired growth objective to be very modest citing "remaining stable in business" so they can secure a stable stream of income. They consider growth of 5-15 percent in their sales to keep up with competition. Only two cases (C1 and D1) from businesses that operate in the formal sector have higher growth objectives and both of these two companies export their products/services to European markets. We have asked all companies (formal or informal) to explain why they have chosen to operate in formal or informal sector. This was illustrated by the interviewee A2 who says: "I started my own online business at age 19. I have learned through internet how this business works, so I have opened a Facebook account and started to promote products for sale. I bought these products in the open market near capital city in Prishtina. I usually bought one product so it enabled me to make picture and put in the internet. So, initially I didn't know how this business will go. Now, my business has grown and I employ 7 employees and contract services via private mailing service company for delivery of my products online. I did not register still my company because it is complicated to adhere to all procedures. I plan to register my company if I keep growing this business. I import products from China and sell online, so I started to think to register because now I am in a better position to earn sufficient profit and pay taxes." (Interviewee A2) In contrast to A2, which operates in the same business but formally, the A1 company states that "the main reason for operating formally was that the company since the beginning had plans for growth into international market to sell". He explained his "desired growth" as a growth in international markets. This two cases show that differences in the concept of "desired growth".
Indeed, all the respondents from the group of firms operating informally stated that they had operated formally because they wanted to test their business ideas if first can succeed then they can move to informal sector. Only one company -B2 in the construction was an exception who stated: "I have been in construction business for 15 years. We are 5 brothers and work together. We did construction work entirely informally. All my employees which I engage seasonally are not registered in employment office. I don't have plan to grow my business. If I do register my company then I will be left with small profit which is not sufficient for my family." (Interviewee B2).
Slightly different view and strategy was used by interview B1 who stated: "I did register my company after 5 years in operation in the construction business. When I have started to get contracts from government I had to register in order to get them. Now my business is stable". (Interviewee B1) As can be illustrated from these two cases the motivation to register formally often was because of increased opportunities to work with government or as interviewee D1 explains, to get financing to support growth in their business. In terms of growth he explains that "I had growth ambitions to increase my business in an international business" Overwhelming majority of cases we found that they have very modest desired growth objectives suggesting that small firms in post-conflict economy are not growth-oriented. Their desired growth is to remain stable in business, which according to C2 and D2 means to experience small growth considered sufficient to keep same number of employees. This finding is in line with La Porta and Shleifer (2008), who argue that informal firms surveyed by the World Bank-are much smaller than formal firms. However, the qualitative analysis revealed that small firm largely are not growth-oriented and their informality is linked with their "desired growth" to remain in the business and keep up with competition.
In addition, our interviews also show that some interviewees did not register companies because they think that other companies in their sector do operate informally. This is illustrated by the B2 and C2: "I didn't register my company because other firms in my sector operate informally. If I get registered then I will pay taxes and need to charge higher prices for my services" (Interviewee C2) " … other business don't pay any tax, why should I pay taxes and earn less when no one from competitors pays taxes" (Interviewee B2) Going back to our quantitative analysis, this finings helps to better understand the econometric results. Quantitative analysis suggest that firms operating informally can use informality to better achieve their 'desired growth" compared to those informally. Some authors (Krasniqi, 2007) argue that operating unregistered can decrease the cost of operations and at the same time can enable firms to become more competitive. Williams and Krasniqi (2018) in the study for Kosovo found a statistically significant correlation between sales under-reporting and the level of horizontal trust, that is, firms that consider that other competitors do not pay taxes. However, this finding applies only to small start-up firms which usually use this phase to test their initial business ideas in informal sector first. If they succeed this phase they consider then expansion which requires more formal type of arrangements and contracts as well as financing, therefore may choose to enter formal sector. Therefore, the findings from econometric section should be analyzed in the light of the argumentation that small firms usually are not growth-oriented and can only be interpreted in the context of modest growth intentions (see Krasniqi & Mustafa, 2016).

Conclusion and policy implications
This analysis of entrepreneurs in Kosovo reveals that nearly half (47.7%) operate unregistered and that operating unregistered is positively and significantly associated with a higher likelihood of achieving their desired growth objectives. However, the results of this study should be interpreted with cautions in the context of small firms which are not growth-oriented and can only be interpreted in the context of modest growth intentions Theoretically, therefore, this study contributes to the small emergent body of literature that is contesting the thesis that informal entrepreneurship negatively influences firm performance. The strong intimation, therefore, is that the deficiencies of the formal institutional environment result in the benefits of registration being low and the costs of registration high, resulting in nonregistration being a rational economic decision. Indeed, entrepreneurs who do not register outperform those that do but only for small firms with modest growth intentions as informed by our qualitative analysis. The costs of registering therefore outweigh the benefits of formal registration, and the benefits of nonregistration outweigh the costs.
This has important policy implications. Conventionally, and in line with Allingham and Sandmo (1972), there is a need to adopt a rational economic actor approach and seek to change the costs and benefits of operating unregistered and registered. Allingham and Sandmo (1972) and many governments, however, focus near enough entirely upon increasing the costs of operating unregistered by increasing the penalties for doing so and the risks of detection. This study intimates that this is only a partial solution to the problem. There is also a need to reduce the costs and improve the benefits of registration since the benefits currently appear insufficient to outweigh the benefits of nonregistration in Kosovo. It is important that government policy should target the education and training of small business owners on the benefits of registration. The survey findings show that in question "why you have not registered your company", 54 percent of respondents stated that: "don't see benefit of doing that", reinforcing the need for expanding program for start-ups, but at the same time making information available for access to all potential entrepreneurs. Despite the increasing the business start-up programs (see Lajqi et al., 2019), those programs need to have as a prequalifying criteria certificate of registration at business registry and tax administration authorities. These criteria will encourage start-ups and small firms to operate formally. In Kosovo, there are available programs to support start-ups, but there is need to shift the start-up support towards more growthoriented small firms in order to incentivize firms to enter in formal sector. Put it differently, there is a need to move away from 'deterring informal entrepreneurship and towards the use of incentives and indirect controls if a formalization of informal entrepreneurship is to occur' (Williams, 2014).
Nevertheless, there are limitations to this study. First, it is only a single country study, and the average size of the firm in the sample is very small (up to 12 employees) suggesting that we cannot argue that this holds for sample of larger firms. Future studies could use more differentiated approach to evaluate the effect of informality at different stages of firm growth. An additional limitation of the study is that it cannot evaluate the reasons for entrepreneurs operating unregistered (high costs and limited benefits of registration) and benefits of registration (better access to finance, attracting new investment for growth, more access to government and donor support programs). Future studies should take into account these and use multi-country studies to design tailored made policy measures. From the theoretical perspective, the findings suggest that rigorous research is needed to evaluate other negative and potentially positive assumptions of informal entrepreneurship, such as whether nonregistration can support start-ups as a platform to test their viability and then put in place steps to facilitate the transition to formality. Future studies should take into account the growth ambitious of entrepreneurs because the relationship between informality and firm growth could be limited only to small firms with modest growth ambitious.
In sum, this paper has revealed in one country context the positive influence of nonregistration on achieving firm growth under modest growth ambitions of small firms. If this stimulates further research on this association between the registration and firm performance, then it will have achieved one of its intentions. The results could be used to explain behaviour of the entrepreneurs operating informally in similar institutional contexts in transition and emerging economies. If this then leads to governments considering how to improve the benefits and reduce the costs of registration, to encourage the transition to formalization, then it will have achieved its fuller intention.

For details see Millennium Challenge Corporation (2018) Kosovo Labor Force and Time
Use Study, Research Report, Prishtina: Millennium Challenge Corporation.

Disclosure statement
No potential conflict of interest was reported by the author(s).