Small- and medium-size enterprises’ access to finance: The role of institutions

(1)

1

Small- and medium-size enterprises’ access

to finance: The role of institutions

Eizo Pieter ter Veer S1907719*

Master’s Thesis

International Business & Management Small Business & Entrepreneurship

Supervisors: Dr. A.A.J. van Hoorn

Dr. Ir. H. Zhou

Januari 20 2015

Word count: 18.646 Abstract

This study addresses the problem that many small and medium firms face concerning access to finance. It is tried to find the institutions that have an influence on access to finance. It is argued that small and medium firms are more positively affected by institutional development in countries with institutional voids than large firms. In this research, financial institutions, rule of law, accounting standards, regulatory environment and corruption are addressed. There is no evidence found that size plays a role when considering the effect that institutions have on access to finance. The effects are equal or even bigger for large firms.

Keywords: Small and medium sized enterprises; access to finance; institutions

*_{Master student at the Faculty of Economics and Business, University of Groningen, The Netherlands. E-mail:}

(2)

2

1. Introduction

Small and medium sized enterprises (SMEs) form a large part of the private sector in developed and developing countries. For instance, in 2013 99.8% of the companies were SMEs in the Netherlands. These companies employed 66.3% of the working population and added 63.8% of value (European Commission enterprise and industry,2013). According to Raynard and Forstater (2002), SMEs are especially important in the development process in a country. They argue that SMEs tend to employ more labor intensive production processes and therefore contribute significantly to the employment process. Furthermore, countries with a high share of SMEs have succeeded in making the income distribution more equitable and that SMEs are an important source of innovation. Therefore, it could be argued that SMEs play a vital role in the development of the economic market.

It is argued that there is substantial evidence that small firms face larger growth constraints and have less access to formal sources of external finance, potentially explaining the lack of SMEs’ contribution to growth. Furthermore, the development of financial and institutional organizations can help to take away these growth constraints for SMEs and increase their access to external finance (Beck and Demirgüç‐Kunt, 2006). Kerr and Nanda (2009) argue that financing constraints are one of the biggest concerns impacting potential entrepreneurs around the world. And given the important role that entrepreneurship is believed to play in economic growth, taking away financial constraint for would-be entrepreneurs and startups is an important goal for policymakers worldwide.

This means that SMEs are expected to have a problem when it comes to access to finance. Access to finance is commonly referred to as the availability of supply of quality financial services at reasonable costs. There are several internal and external reasons for the finance problems of SMEs. Some of the internal reasons are: lack of reputation, small firms are more opaque and small firms have less collateral (Biggs and Shah, 2006, Beck and Demirgüç‐Kunt, 2006). These internal reasons can be seen as the characteristics of the firm itself and these are hard to change by policymakers. Therefore, it is more interesting to look at the external reasons for the problem with access to finance.

(4)

4

aimed at SMEs are based on the premises that market and institutional failure impede their growth which justifies government intervention. Therefore, it is interesting to look at these failures and research how to overcome them. To do this, one can look at institutions. The most common

definition of institutions is the one from North (1990 p.3); “Institutions are the rules of the game of a society”. This means that institutions include written and unwritten rules where culture and norms are governed by the latter, while laws, political systems taxes etc. are examples of the former. The last important element is the enforcement mechanisms, what are rules if they cannot be enforced. Being that the institution encompass so many different aspects and that rules for a society also create the rules of the marketplace, I think it is more interesting to look at the institutions than at the market failure.

One thing that is often associated with institutional failure, are the so-called ‘institutional voids’. Institutional voids are defined by Mair and Marti (2009) as; “situations where institutional arrangements that support markets are absent, weak, or fail to accomplish the role expected of them.” (p. 1) Meaning that there is a ‘gap’ between what is expected from the institutions and what they do. These voids emerge when economic growth advances faster than social and institutional structures. This means that these gaps are more apparent in countries that are less sophisticated than the western world. Khanna, Palepu & Sinha (2005) argue that countries with institutional voids, lack specialized intermediaries, regulatory systems, and contract-enforcing methods. Lacking these institutions can make it harder for companies, for instance, to invest. Since they cannot be sure if the data they have about an investment is right and, moreover, if the contract they signed is enforced. Furthermore, Beck and Demirgüç‐Kunt (2006) argue that corporations, often large firms, report fewer financing, legal and regulatory obstacles than unincorporated firms, often smaller firms, and that this advantage is greater in countries with more developed institutions. Emphasizing the importance of institutional voids for SMEs. To understand the institutional voids, Khanna, Palepu & Sinha (2005) argue that companies need to understand the institutional variation in countries in five different contexts. One of these is the capital market, since the development of this market varies a lot between countries. They state that in a lot of less developed countries, apart from a few stock exchanges and government-appointed regulators, there are not many reliable intermediaries. Which leads to that investors and creditors do not have access to accurate information on companies and Businesses cannot easily assess the creditworthiness of other firms or collect receivables after they have extended credit to customers.

(5)

5

defined as the result of institutional failure. Therefore, it can be interesting to research which institutions have an influence on access to finance for firms, but especially if this differs a lot between small, medium and large firms. This way, it can be seen whether size can overcome institutional failure or that the lack of institutions has the same effect on all firm sizes. All of this leads to the following research question;

How do national institutions influence the access to finance for Small and Medium sized companies and does this differ compared to large firms?

In this paper I will look at the following institutions: financial institutions, the rule of law, accounting standards, regulatory environment and corruption. To address this problem the Business Environment and Enterprise Performance Survey (BEEPS) 2008-2009 is used. This is a database that consists of data from 29 countries in eastern Europe and central Asia. In this database there are several countries that are in different stages of development and that have institutional voids. The added value of this research lies in the fact that researching this research question, leads to a

research in which not only a too little researched group, SMEs, are dealt with, but it also gives a clear picture of the differences in access to finance between small, medium and large firms. Also, in many on the literature, the relation between SMEs growth and institutions have been discussed, only a few papers look directly at the relation between one of the main problems for SME growth, access to finance, and the institutions. Furthermore, it combines several unique databases to one dataset combining firm level data with country level data on institutions.

(6)

6

2. Theoretical Background

In this chapter the important topics for this research will be discussed. The institutional theory, and with that the institutional voids theory, are the overarching theory and will be explained first. After this, access to finance for SMEs will be explained on the basis of equity and debt finance, followed by the institutions and access to finance.

2.1 Institutions

From an institutional economics perspective, there are two main views on institutions; the Original Institutional Economics (OIC) and the New Institutional Economics (NIE). The New institutional economics are called ‘new’ since the economics are embedded in a framework of institutions, formal and informal, in which they encompass the old institutionalism school and the neo-classical

economics (Williamson, 2000). This means that it can be described as an analytical system born within the frame of neoclassical economics, but it offers answers to several issues by modifying and extending it (Boliari, & Topyan, 2011). In the NIE, there are several definitions about institutions. According to Scott (2001), institutions can be seen as multifaceted, durable social structures, made up of symbolic elements, social activities and material resources resistant to change and transmitted across generations. Polanyi (1944) argues that market access and market activity are affected by the social, cultural, and political institutions in which they are embedded. For these markets, institutions are especially important as they need specific institutions and rules in order to come into existence and function (Fligstein, 2001). However, the most commonly agreed upon definition of institutions is the one from North (1990, p.3): “Institutions are the rules of the game of a society, or, more

(7)

7

Furthermore, Williamson (2000) argues that there are four main levels of social analysis, in which the first three are consistent with the three important elements suggested by North (1990). He argues that at the first level are the informal or unwritten rules. This influences all the levels below and is the most difficult level to change (approximately 100 to 1000 years). The level below that is the formal/written rules or the institutional environment. Here, the executive, legislative, judicial and bureaucratic functions of government as well as the distribution of power across the different levels of government are established. Important features are the definition and enforcement of property rights and contract laws. Changes on this level take approximately 10 to 100 years. The third level is concerned with aligning the governance structure with contracts. Thus, the

enforcement of contracts or the enforcement mechanism. This is based on the premise that enforcing contracts is not free of charge and that there are transaction costs. Changes on this level take one to ten years. In general, there are three types of transaction costs: 1. search and

information cost, finding a suitable buyer, 2. Bargaining and decision cost, time and advice used to negotiate an agreement, 3. Supervision and enforcement cost, the cost of supervising and enforcing the deal.

This means that, when looking at the influence of institutions, we can do this on three different levels; the informal or unwritten rules, the institutional environment and the enforcement mechanisms.

2.1.1 Institutional voids

In many developing countries institutions that support markets are either absent or weak (Easterly, 2006). The literature calls this institutional voids. These institutional voids are defined by Mair and Marti (2009) as; situations where institutional arrangements that support markets are absent, weak, or fail to accomplish the role expected of them. These voids emerge when economic growth

advances faster than social and institutional structures.

There is a lot of debate concerning markets in developing countries and who should create and maintain the necessary institutions. It is widely accepted that the state, and therefore the government, is responsible for the building of these institutions (Mair, Marti, 2009). Discussions subsequently center on the degree, as well as the manner of state intervention, and range from demands to set property rights and guarantee their enforcement (De Soto, 1989).

(8)

8

The capital and financial markets in many less developed countries are remarkable for their lack of sophistication. Apart from a few stock exchanges and government appointed regulators there are no reliable intermediaries such as credit-rating agencies, investment analysts, merchant bankers, or venture capital firms. Companies cannot count on raising debt or equity locally and do not have accurate information on companies. Corporate governance is weak as well. Companies cannot trust their partners to adhere to local law. They cannot even assume that the partner is driven by profit alone since there is still a lot of corruption. The authors argue that companies should always analyze, before investing in a country, industry structure, such as entry barriers, scale economies etc., to understand a countries institutional context. This since the industry profitability varied widely across countries. (Khanna, Palepu, 2009)

Companies may have to adapt to the voids in a country’s product markets, its input markets, or both. But companies must retain their core business propositions even as they adapt their

business models. Three strategies; adapt your strategy to the institutional voids, change the context, ergo change the institutional voids (the big four accounting firms setting up branches in Brazil which raised financial reporting and auditing standards in Brazil) or stay away. (Khanna, Palepu 2009)

2.2 SME Access to finance

Generally, access to finance refers to the availability of quality financial services at reasonable costs. However, depending on what one considers ‘quality’ services and ‘reasonable’ costs, the

measurement of access to finance needs to be altered accordingly. Measurement of access to finance is also influenced by the definition and priority of its various dimensions about access to finance for SMEs, one can talk about two different sources of finance; equity finance and debt finance. These two sources of finance will be discussed below, followed by the institutions that are important for access to finance.

2.2.1 Equity finance

(9)

9

develop, they soon are in need of other capital than internal equity and turn to external capital, including venture capital, corporate investment and bank debt (Falkena et al., 2002). In South Africa, as in many other countries, it is falsely believed that when there is enough debt financing for SMEs, they will overcome most of their development problems. In practice it is often the case that SMEs with a proven track record hardly ever experience a shortage of debt financing. In the UK, the Cruickshank Commission emphasized the implied consequence thereof, being that Government interventions in the past to stimulate debt finance to SMEs had been misdirected. According to Falkena (2002), governmental interventions should take a different path, away from making it easier for SMEs to get a loan, and focus more on policies that facilitate the provision of equity to SMEs. Governments should therefore support the expansion and/or establishment of venture capital funds. Following from this, one can say that for SMEs to grow, a better allocation of public and private money is needed. The government has a key role in this process.

2.2.2 Debt Finance

With debt finance, the problem is not so much the availability of debt finance as such, but

inefficiencies in terms of product range, the cost of debt finance and the services provided to SMEs. The reasons for these inefficiencies relate mainly to competitive factors, to barriers for the entry of potential new providers of financial services and to SMEs’ need for non-financial services. A crucial subject of this chapter is that, as stated by Falkena (2002), the recognition that, for SMEs, access to debt cannot be separated from the broader topic of access to banking services.

According to Berger and Udell (2006), there are two important overarching lending technologies. These are; transactions lending that are based primarily on ‘hard’ quantitative data and, relationship lending, which is based on ‘soft’ qualitative information. Berger and Udell (2006) argue that in previous research the SMEs opacity is the most important determinant in determining whether SMEs use transaction lending or relationship lending. They argue that SMEs with the highest financial strengths and the highest transparency have easy access to finance mostly through

traditional financial statement lending. On the other side of the continuum are the SMEs with the lowest financial strength and high opacity, which are unqualified borrowers that should not receive finance. In the middle is the so-called funding gap. The SMEs in this gap can be firms with a good business model, skilled entrepreneurs and high growth potential but they are also; undercapitalized, smaller which makes them more opaque and, more vulnerable to financial shocks.

(10)

10

of the whole company meaning that there is collateral for the loan. This could be helpful in overcoming the funding gap for SMEs. However, for this to work, a certain degree of development from financial institutions is required since it is important that both parties get the right information (Berger & Udell, 2006). As can be concluded from the above, for SMEs, especially debt finance and external equity finance are important in improving access to finance. This is all external finance, and therefore this term will be used throughout the paper. Furthermore, it can be concluded that especially for debt finance the banking sector is of high importance while for external equity finance the government has a big influence.

2.2.3 Institutions and access to Finance

As argued in the introduction of this paper, Small and Medium sized enterprises play an important role in the economic development of a country. However, one of the main problems for SMEs is access to finance. Beck et al. (2006) show that institutional development is the most significant country characteristic that can explain cross-country variation in firms’ financing obstacles. For example, Sleuwaegen and Goedhuys (2002), show in their study that smaller firms grow relatively faster in Germany than in the Ivory Coast, while the opposite is true for large firms. According to Beck and Demirgüç‐Kunt (2006), these differences in growth rate between small and large firms exist primarily because small firms suffer more from market frictions such as transaction costs and

information asymmetries than large firms. Beck et al (2004) argue that a positive effect of financial and institutional development can mostly be seen in the use of external finance. Especially the better protection of property rights increases external financing.

(11)

11

idiosyncratic risk proprietors face. This shows again that, if in the absence of well-developed institutions, it is optimal for firms to stay small.

Concluding from the above, institutions can be seen on three levels, the informal or

(12)

12

3. Hypotheses development

In this chapter, five hypotheses will be developed. These five hypotheses concern, 1. The financial institutions, 2. The rule of law, 3. The access to information, 4. The tax and regulatory environment and 5. The social environment.

3.1 Financial institutions

According to Siklos (2001), there are three types of major financial institutions. These are: Depositary Institutions: institutions that take deposits and manage them. Contractual institutions of which examples are insurance companies and pension funds and investment institutions, for instance, investment banks. Siklos (2001) states that the function of financial institutions is to provide services as intermediaries of financial markets and that they facilitate the flow of money through the

economy. Therefore, they are an intermediary in getting the funds from investors to firms in need of these investments. From this one can conclude that financial institutions are of a vital importance in getting access to finance.

Demirgüç‐Kunt and Maksimovic (1998), Shows in their paper that in countries with better developed financial and legal institutions, firms have better access to external finance than in countries with less developed financial and legal institutions. This paper shows an especially strong relation between the development of financial institutions and the access to finance for firms. However, the sample in this paper includes mostly large firms. Beck, Demirgüç‐Kunt, and Maksimovic (2005) argue in their paper that large firms internalize many of the capital allocation functions carried out by financial markets and financial intermediaries. Therefore, when the financial market develops, it should

disproportionately benefit SMEs. Moreover, the main conclusion of their article is that the smallest firms are the most constrained and that firm size has a constraining effect. These constraining effects become less when financial institutions develop. This leads to believe that when the financial

institutions develop, SMEs are less constraint to grow and, more importantly, have better access to external finance.

(13)

13

(depth), the degree to which individuals can and do use financial services (access), the efficiency of financial intermediaries (efficiency), and the stability of financial institutions (stability).

Concluding from the above, it can be stated that , following Demirgüç‐Kunt and Maksimovic, (2005) and Siklos (2001), that the development of financial institutions play a key role when

considering access to finance. Moreover, it is argued that this development of financial institutions has a disproportional effect on the growth rate of SMEs. Therefore, one could argue that the development of financial institutions has a disproportional effect on the access to finance for SMEs as well. This leads to the following hypothesis:

Hypothesis 1: In a country with good financial institutions, there is better access to finance for firms with a bigger effect on SMEs than on large firms.

In which financial institutions shall be measured with variables for depth, access, efficiency and stability since it can be interesting to see which of these four has the biggest influence.

3.2 Rule of law

Berger and Udell (2006), argue that the rule of law is one of the most important dimensions in the extension of credit to SMEs. They state that the Rule of law consists of a legal, judicial and

bankruptcy environment. The legal environment that affects business lending consists of the

commercial laws that specify property right associated with the commercial transaction. The judicial and bankruptcy environment show how well those laws can be enforced in disputes. This determines the confidence that contracting parties have in financial contracts. Thus, both the legal environment and the enforcement of the laws have to be high for a country to have a good rule of law. If there are good commercial laws but they cannot be enforced, contracting parties still will lack confidence in that countries rule of law. If the enforcement of the commercial laws is high, but there are no laws or ambiguous laws, there still would not be high confidence by contracting partiers. Countries differ significantly on the rule of law dimension. For some countries the commercial laws are incomplete and the enforcement is weak, while for other countries the laws are unambiguous and enforcement is predictable (EBRD, 2003).

(14)

14

They stated that small firms may be more affected by this than larger firms. In another study Beck and Demirgüç‐Kunt (2006) found that the effect of financial, legal, and corruption problems consistently constrained the growth of smaller firms more than larger firms in a cross-country analysis.

Jappelli, Pagano, and Bianco (2005) state that the efficiency of the judicial and bankruptcy systems are important as well for credit availability. Especially high-cost judicial procedures, which they call judicial inefficiency, are associated with decreased access to credit. An important dimension of the efficiency is the length of time in bankruptcy.

To measure the strength of a countries lending system one could look at the different kind of loans that are mostly used in that country. An important aspect of countries that do have a strong lending infrastructure is asset-based lending. This is any kind of lending secured by an asset. This means, if the loan is not repaid, the asset is taken. Berger and Udell (2006), argue that since asset based lending only has a significant presence in four countries there must be significant hurdles in the legal, judicial and bankruptcy environment to use these kinds of loans. Therefore it is only used in countries with strong lending structures. In countries with a weak lending infrastructure the use of factoring and leasing are probably more popular. This while in these cases the assets for which they take a loan is still property of the lender and not from the borrower. This gives the lender more power since they are still the owner of the asset and can claim it back in the case of not paying. This means that in countries with strong lending systems, assets can be used as collateral in loans but in countries with weak lending systems loans such as leasing, where the borrower is not the owner of the asset, are more important. Overall it can be concluded that a good law system improves the use of external finance and improves the efficiency of financial institutions. Furthermore, a weak rule of law has a bigger effect on the growth of SMEs than on large firms. Therefore, it could be argued that improvements in these institutions have a disproportionally bigger effect on SMEs than on large firms. Therefore it is argued;

Hypothesis 2: In countries with a good legal system, there is better access to finance for firms with a bigger effect on SMEs than on large firms.

3.3 Access to information

(15)

15

conditions for good financial statements. These financial statements are key in the financial statement lending technology (Berger & Udell, 2006). According to La Porta, Lopez-de-Silanes, Shleifer, & Vishny (2001), there is considerable variation between accounting standards. As well between developing and developed countries as between developed countries.

To have a good information environment, sharing of information is important. Commercial and consumer credit bureaus can help this sharing. They provide mechanisms for the exchange of payment performance data (Berger & Udell, 2006). Furthermore they have shown to have predicting power in firm failure beyond financial ratio (Kallberg & Udell, 2003). These bureaus are important in combination with all the lending technologies since they have the necessary data on SME loan performances. In general, better accounting infrastructure has a positive effect on the available information which is important for several financial institutions. Therefore it is argued:

Hypothesis 3: In a country with a good accounting infrastructure, SMEs have better access to finance.

3.4 Tax and regulatory environment

The tax environment can have a large effect on the availability of credit for SMEs. Different taxes can influence lending technologies in different ways. Governments can try to stimulate lending to SMEs by, for instance, giving several tax advantages. The same holds for the regulatory environment. Government policies often affect the entry of different types of financial institutions, their market shares, their abilities to compete, and their corporate governance structures (Berger & Udell, 2006). Levine (2003) and Beck, Demirgüç‐Kunt and Maksimovic (2004b) found that larger market shares for foreign-owned banks are often associated with greater SME credit availability in developing nations. Therefore, government policies that restrict foreign entry may have a large effect on access to finance for SMEs. These restriction policies may especially effect the transaction lending technologies in developing economies. Since the foreign institutions probably have an advantage in collecting and processing these hard data (Buch, 2003). Thus, as well tax rules that promote small business lending as government policies that do not restrict foreign entry are important for access to finance for SMEs.

(16)

16

that this would improve the access to finance for SMEs however, it could have the opposite effect due to the lack of market discipline for the institution. Also, the funding could go to SMEs that are not creditworthy, the loans are not paid back at market rates and it could go to the wrong companies for political rather than economic motives (Cole, 2009). Stated-owned institutions are often

associated with unfavorable macroeconomic consequences and less developed financial and economic systems (Barth, Caprio, Levin, 2004, La Porta, Lopez‐de‐Silanes,& Shleifer, 2002). Furthermore, some research also suggest that a greater percentage of state-owned banks has an negative influence on the access to finance for SMEs (Berger et al, 2004). Furthermore, Falkena Abedian, Von Blottnitz, Coovadia, Davel, Magungandaba and Rees,(2002), argue that for a better access to finance for SMEs, governments should support the establishment and expansion of venture capital funds since as successful SMEs develop they outgrow internal equity and the next step is external equity-based finance. Thus, one could argue that a good regulatory environment in which there is room for foreign-owned firms and little stated-owned institutions, lowers the cost of business and has a positive effect on access to finance for SMEs.

Hypothesis 4: A restrictive regulatory environment has an negative effect on access to finance for SMEs.

3.5 The social environment

(17)

17

Hypothesis 5: In countries with less corruption, there is better access to finance for firms with a bigger effect on SMEs than on large firms.

(18)

18

4. Data and method

4.1. Data

The main data source for this paper is the database of the Business Environment and Enterprise Performance Survey (BEEPS). This is a joint initiative of the European bank for reconstruction and development and the World Bank Group. In this survey there are businesses of different sizes and questions about several subjects, including access to finance. In the BEEPS, small firms are firms with 5 to 19 employees, medium sized firms have 20 to 99 employees and large firms have more than 99 employees. From the latest version of the BEEPS, 2012, only the data about Russia is available at this moment. Therefore, I choose to use the dataset from 2008-2009. This dataset covered companies in 29 different countries in Eastern Europe and Central Asia. These countries are: Albania, Armenia, Azerbaijan, Belarus, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Estonia, FYR Macedonia, Georgia, Hungary, Kazakhstan, Kyrgyz Republic, Latvia, Lithuania, Moldova, Mongolia, Montenegro, Poland, Romania, Russia, Serbia (including Kosovo), Slovak Republic, Slovenia, Tajikistan, Turkey, Ukraine, and Uzbekistan. This is an interesting area since it encompasses countries in different stages of development. The BEEPS survey does not include information about countries institutions. Therefore, I added data on country-level variables from several other databases. These are; the Global financial development database, the World Bank Worldwide Governance Indicators, the global competiveness report and the Doing Business indicators. From all of these databases the data about 2009 is used. The complete sample used consists of 11.668 firms in 29 countries. Unfortunately, not every company gives information about access to finance. Therefore, I choose to filter out the companies that responded to the question: “How much of an obstacle is access to finance?” with “Do not know” or “Does not apply”. This leaves 10.078 companies in the sample.

4.1.1 Dependent variable

In the BEEPS survey, several questions are asked concerning access to finance. However, I think that there is one question that covers the topic of access to finance best; “Is access to finance, which includes availability and cost, interest rates, fees and collateral requirements, No Obstacle, a Minor Obstacle, a Moderate Obstacle, a Major Obstacle, or a Very Severe Obstacle to the current

(19)

19 How much of an obstacle is access to finance Size less than 5 small >=5 and <=19 medium >=20 and <=99 large >=100 Total don’t know % of total 8 0.1% 146 1.3% 80 0.7% 52 0.4% 286 2.5% does not apply

% of total 5 0.0% 85 0.7% 44 0.4% 29 0.2% 163 1.4% no obstacle % of total 85 0.7% 1248 10.7% 1266 10.9% 985 8.4% 3584 30.7% minor obstacle % of total 53 0.5% 715 6.1% 715 6.1% 501 4.3% 1984 17.0% moderate obstacle % of total 62 0.5% 939 8.0% 954 8.2% 690 5.4% 2645 22.7% major obstacle % of total 34 0.3% 717 6.1% 635 5.4% 472 4.0% 1858 15.9% very severe obstacle

% of total 20 0.2% 407 3.5% 413 3.5% 308 2.6% 1148 9.8% Total % of total 267 2.3% 4257 36.5% 4107 35.2% 3037 26.0% 11668 100%

Table 1 present the descriptive statistics of the question: how much of an obstacle is access to finance among the firms.

As can been seen from the table displayed above, almost half of the firms (48.4%) considers access to finance as at least a moderate obstacle. It is interesting to see that the differences between Small and Medium sized firms are very small with every answer. 17.6% of the small firms see access to finance as a moderate or worse obstacle while, 17.1% of the medium sized firms and 12% of the large firms see it as a moderate or worse obstacle. For the remainder of this paper, I chose to incorporate the micro firms with the small firms.

4.1.2. Key independent variables

(20)

20

literature as private credit deposit money bank credit to the private sector as a percentage of GDP (Čihák et al., 2012).

For the access to financial institutions Čihák et al. (2012) propose several measures. The one most commonly used is the number of bank accounts per 1000 adults. However, for the countries used in this sample, this information is not available for almost half of the countries. Therefore, I chose to use another measure, the percentage of firms with a line of credit. This information is both available in the Global Financial development database as in the BEEPS questionnaire. Since this variable measures the access to financial institutions one could expect that the correlation with the dependent variable is high. After testing this correlation, this is not the case and therefore this variable is included. The efficiency of the financial system can be measured using the banks return of equity or assets. For the financial stability it is commonly accepted to use the z-score of banks (Čihák et al., 2012).

For the second hypothesis, the rule of law measure of the World Governance Indicators is used. They define Rule of Law as: “It Reflects perceptions of the extent to which agents have

confidence in and abide by the rules of society, and in particular the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence.” From this indicator the estimate will be used; “Estimate of governance (ranges from approximately -2.5 (weak) to 2.5 (strong) governance performance)”. (Worldwide governance indicators 2008-2009)

With the third hypothesis the accounting infrastructure among different countries will be compared. This can be done with the global competiveness indicators 2008-2009 since one of these indicators is the Strength of auditing and reporting standard.

For the fourth hypothesis, about the restrictive regulatory environment, several variables from the Doing business indicators 2008-2009 are used. The Doing business indicators is a yearly report from the World Bank Group in which is determined how easy it is to conduct a business in several countries. This is measured with several indicators. These are; starting a business, dealing with construction permits, registering property, getting credit, protecting minority investors, paying taxes, trading across borders, enforcing contracts and resolving insolvency. Since getting credit is highly correlated with access to finance, this variable will not be used. The same holds for enforcing

contracts since this is covered in the Rule of law variable. For the last hypothesis about corruption, the World Governance Indicators 2008-2009

(21)

21

as well as ‘capture’ of the state by elites and private interests.” And the estimate as: “Estimate of governance (ranges from approximately -2.5 (weak) to 2.5 (strong) governance performance)” (Worldwide governance indicators 2008-2009). Table 2 presents the descriptive statistics for these variables.

Variable Obs. Mean Std. Dev. Min Max

Bank Z-score 29 12.47972 11.73738 -4.545584 58.40426 Firms with a line of credit 29 34.69615 24.42051 0 71.2 Deposit money bank asset to GDP 29 41.02488 34.77029 0 112.5934 Bank return on assets 29 -1.953526 10.35922 -51.41264 3.061513

Rule of law 29 -.0531712 .6921099 -1.316931 1.090953

Starting a Business 29 81.87308 10.6617 48.89 95.92 Dealing with construction permits 29 50.32385 22.5906 7.73 85.12 Registering property 29 71.57885 15.6408 47.5 96.86 Protecting minority investors 29 53.58962 10.62241 33.33 73.33

Paying taxes 29 59.83346 17.14965 17.98 85.51

Trading across borders 29 59.16423 27.15569 0 92.08 Resolving insolvency 29 35.66 10.9548 9.85 51.67

Corruption 29 -.2506724 .6094917 -1.230568 1.023537

Accounting standards 29 4.365988 .6147263 3.326661 5.690338 Rule of law factor 29 .1217704 1.085886 -2.009103 1.682086

Table 2 presents the descriptive statistics of the independent variables

4.1.3. Control variables

To address potentially spuriousness, I add various independent variables to my analysis as control variables. Beck, Demirgüç-Kunt, and Martinez Peria, (2008) state in their paper that macroeconomic instability is the main obstacle for lending money to SMEs. Therefore, I will control the results for the macroeconomic stability in the country. Furthermore, Beck, Laeven et al. (2006) find that when everything else is equal, older, larger and foreign owned SMEs face less financing obstacles. Therefore, I will control for this variables as well. For the macroeconomic stability a measure from the global competiveness report is used. For the older, larger and foreign owned dimensions, the BEEPS questionnaire is used. In the BEEPS questionnaire, size is a categorical variable in which 1 is for small firms, 2 for medium sized firms and 3 for large firms. For the older firms, the year of

establishment is used. Table 3 shows the descriptive statistics for these variables.

Table 3 presents the descriptive statistics of the control variables.

Variable Obs. Mean Std. Dev. Min Max

Macroeconomic environment 10078 4.725769 1.06095 0 5.72

Year of establishment 10078 1958.612 257.4751 -9 2008

(22)

22

4.2. Methodology

My sample consists of firms nested in different countries. To deal with this hierarchical structure of the data I apply multilevel or hierarchical linear modeling. A multilevel model is used since all the observations in the BEEPS questionnaire are on the firm level (level 1). However, all the observations for the institutions are on a country level (level 2). The general model is the following:

Access to finance = β0 + β1 FI + β2 ROL + β3 COR + β4 AcS + β5 RE + ε

In this model, FI are the financial institutions, ROL is the rule of law, COR is corruption, Acs is accounting standards, and RE is the regulatory environment. Subsequently, β0 is the constant, the other β’s are the coefficients of the direction and ε is the residue.

When using the multilevel model, for both levels, an regression equitation is used. For level 1 the following regression equitation is used:

Yij = β0j + β1j(Xij) + eij

In this equitation, Yij refers to the score on the dependent variable for a firm observation. Xij refers to

the predictor of level 1. β0j is the intercept of the dependent variable of level 2 which is calculated in

the level 2 equitation.β1j is calculated on level 2 as well and represent the relationship between the

predictor of level 1 and the dependent variable. eij refers to the random errors of prediction for the

Level 1 equation. At Level 1, it can be either fixed or random. In this case it is Fixed, meaning that all groups have the same values.

For the second level, the following regression equitation is used:

β0j = γ00 + γ01Wj +u0j

β1j = γ10 + u1j

The dependent variables are the intercepts and the slopes for the independent variables at Level 1. Meaning that on this level, β0j and β1j are calculated and then substituted in the level 1 equitation. In

general, γ00 refers to the overall intercept. This is the mean of the scores on the dependent variable

across all the groups. Wj refers to the Level 2 predictor. γ01 refers to the slope, between the

dependent variable and the Level 2 predictor. u0j refers to the random error component for the

(23)

23

Level 1 predictor. u1j refers to the error component for the slope, meaning the deviation of the group

slopes from the overall slope.

To get the results for the different sizes of firms, interaction terms are created. These are made by multiplying the different independent variables with a dummy variable that is dependent on the size of the firm. Due to the fact that all the variance in the model is explained by summing up the

variance in small + medium + large firms, one of the three has to serve as the base category. The variance in this base category, in this case the large firms, can be explained by the estimates of the independent variables since these are the results for the total minus the small and medium results and thus, the results of the large firms.

Unfortunately, there is a high correlation between these last three factors. The Standardized score of corruption correlates 0,946 with the Rule of law estimate and with 0,849 with the accounting standards while accounting standards and Rule of law correlate 0,911 with each other. This

correlation is not a big surprise, since it is to be expected that with a better rule of law, there is less corruption and better defined rules regarding the accounting practices. Since the correlation is this high, it seems not appropriate to use all three of these measures in the model. Therefore I decided to do a factor analysis. This confirmed that these three variables can be used as one factor. To

determine this one has to look at the communalities that have to be higher than 0.5, the eigenvalues that have to be higher than one for the factor created and the ‘Cronbach Alpha’. The first two are created by SPSS. The ‘Cronbach Alpha’ determines the internal consistency or average correlation of items in a survey instrument to gauge its reliability (Santos, 1999). In most of the literature a

Cronbach Alpha of 0.7 is regarded as satisfactory. All three variables had higher communalities than 0.5 and a ‘Cronbach Alpha’ of 0.964. For all the other factors that SPSS created in this analysis, one of these criteria was not met. Table 4 shows the results of these factor analysis.

Construct Consist of Rotated factor score Cronbach’s Alpha

Abiding the law Rule of law Corruption Accounting system 0.874 0.838 0.740 0.964

Table 4. Factor analysis result for rule of law, corruption and accounting system

Substituting this factor analysis variable in to the general model leaves; Access to finance = β0 + β1 FI + β2 ROLFA + β3 RE + ε

(24)

24

Chapter 5. Empirical Results

In this chapter the results of the analysis are discussed. First the baseline results of the model given above are presented. For this a step-by-step method is used. The model will start off with the variables representing the financial institutions and the control variables. In the second model the factor analysis variable for the rule of law, accounting systems and corruption will be added. In the third model the variables representing the regulatory environment will be added. In the last model two of these variables will be left out to improve the model. In the second part of this chapter several robustness checks are done. First other dependent variables are used to test the robustness of the dependent variable. After this, two different measures of size are used and the control variables will be checked for robustness.

5.1. Baseline Results

As stated above, the results are computed trough a multilevel analysis in SPSS. There are four different models that are tested. The first model consists of the four different variables to measure the financial institutions and the control variables. This model is tested first since it can be expected that financial institutions have the biggest influence on access to finance. Table 5 shows the results of this analysis. As can be seen in the results, none of the variables measuring financial institutions seem to be significant and three out of the four variables have a negative influence on the

dependent variable. This is not surprising since with the dependent variable, the higher the score, the higher the problems with access to finance. Therefore, a negative impact means that it lowers the problems with access to finance. When looking at the control variables included, the

(25)

25

finance. For the control variables, almost the same results are found as in the first model.

Percentage of foreign owned is significant (p<0.05) and the year of establishment is almost significant (0,64).

In the third model, the ease of doing business indicators, minus getting credit and enforcing contracts, are included. Although none of these variables seem to be significant, they have a big influence on several other variables. The variables firms with a line of credit and deposit money bank asset to GDP, are almost significant with scores of 0.086 and 0.092. Another noteworthy change is in the Rule of law Factor variable which goes from being highly significant to not significant on the five percent level (0.088). Again, the effects for the control variables are minimal. Following these results, I decided to leave out two ease of doing business factors that had the highest correlation; Dealing with construction permits and Trading across borders. This led to the fourth model.

Without the two variables stated above, the fourth model is quite similar to the second model. Again, only the Rule of law factor analysis variable is significant and the control variable percentage of foreign owned. Following these results several other analysis where done. The variables that where added every model where also solely tested. This did not result in other noteworthy results. Furthermore, reducing the variables more by , for instance, creating one variable for financial institutions and one for the regulatory environment (ease of doing business indicators), did not have any noteworthy effects as well. Overall, it seems that this last model is the best model. Therefore, in the remainder of this paper this model will be used unless stated otherwise.

In general it can be concluded that the factor variable for Rule of law has a significant influence on the access to finance for firms. For the control variables the percentage of being foreign owned has a highly significant effect for all models. This means that a higher percentage of being foreign owned has a positive effect on the access to finance for firms. The year of establishment variable is almost the same in every model. However not significant, the difference is very small. The estimate of this control variable is negative which means that a higher year of establishment lowers the problems with access to finance. This is a surprising result as it was expected that older firms, thus with a lower year of establishment, would have less problems with getting access to finance. To check the results, the last model is done once more with the year of establishment variable standardized. This gives an estimate of -0.0241 with a Std. Error of 0.013. Which leads to believe that older firms do not have less problems with getting access to finance.

(26)

26

different countries is 0.070987/(0.070987/1.506552) = 0.0449. Thus there is not much difference between the different countries. In the other models the variance gets even smaller.

Following from these results, hypothesis one: in a country with good financial institutions, there is better access to finance for firms with a bigger effect on SMEs than on large firms, can be rejected. Since we used a factor analysis to combine hypotheses 2,3 and 5 it is hard to say something for these specific hypothesis. When substituting the rule of law factor analysis variable with one of these three variables, all three of them are significant as well. Therefore, it can be concluded that rule of law, accounting infrastructure and corruption do have a significant influence on the access to finance for firms and thus can be accepted. For hypothesis 4, restrictive regulatory environment has an negative effect on access to finance for firms, it can be concluded that this is not the case and thus should be rejected.

(27)

27 Parameter Model 1 Estimate Model 1 Std. Error Model 1 Sig. Model 2 Estimate Model 2 Std. Error Model 2 Sig. Model 3 Estimate Model 3 Std. Error Model 3 Sig. Model 4 Estimate Model 4 Std. Error Model 4 Sig. Intercept 2.076675 .309849 .000 1.795405 .310656 .000 1.895282 .353913 .000 1.625847 .321352 .000 Standardized: Bank Z-score .030236 .056666 .598 .004083 .053023 .939 -.011238 .057288 .846 -.008677 .058068 .882 Standardized: Firms with a

line of credit

-.013722 .068137 .842 .059038 .069644 .404 .118573 .066494 .086 .108721 .068519 .124 Standardized: Deposit

money bank asset to GDP

-.070636 .059352 .244 -.065165 .054157 .239 -.094767 .054204 .092 -.061997 .052255 .246 Standardized: Bank return

on assets

-.103861 .080242 .207 -.030938 .079508 .700 -.099793 .095765 .307 -.031350 .089207 .728

Rule of law Factor -.170528 .073975 .029 -.314015 .177423 .088 -.189122 .078628 .024

Standardized: Starting a Business

.070228 .093609 .460 .040154 .085382 .642 Standardized: Dealing

with construction permits

-.104430 .132021 .436 Standardized: Registering property .024258 .090411 .790 -.015566 .070404 .827 Standardized: Protecting minority investors -.024796 .082632 .766 -.052151 .070710 .467

Standardized: Paying taxes -.088036 .091925 .347 -.068286 .093880 .474

Standardized: Trading across borders .236327 .180900 .203 Standardized: Resolving insolvency -.114418 .075208 .140 -.107382 .077715 .179 Macroeconomic stability -.075940 .059690 .214 -.015515 .060320 .799 -.031882 .070741 .656 .024367 .063706 .705 Year of establishment -9.44E-9 5.07E-9 .063 -9.39E-9 5.07E-9 .064 -9.39E-9 5.07E-9 .064 -9.37E-9 5.07E-9 .065 % of Foreign owned -.003444 .000552 .000 -.003434 .000552 .000 -.003428 .000552 .000 -.003434 .000552 .000 Size .000044 .000058 .452 .000042 .000058 .456 .000044 .000058 .448 .000043 .000058 .456

(28)

28

To get an impression about how good the model is, a linear regression is performed to see the R². Since it is expected that the last model is the best model, this one is used. The entire model seems to be significant (p<0.05) with an adjusted R² of just 0.04. Meaning that only 4% of the problem of getting access to finance is explained. From table 6 one can conclude that the coefficients for some individual variables may be insignificant while the regression as a whole is significant. A possible reason can be correlation of the independent variables, a condition known as multicollinearity. There is multicollinearity when highly correlated independent variables are explaining the same part of the variation in the dependent variable, so their explanatory power and the significance of their

coefficients are ‘divided up’ between them. It can be concluded that there is

no multicollinearity problem between the different independent variables (1.314 <VIF

<2.943). When the value of the ‘Variance Inflation Factor' or ‘VIF’ is higher than a certain critical limit one can state that there is a problem of multicollinearity. The critical VIF value is very subjective according to Haan (2002). He states that some researchers use a VIF 10 or 5 as the critical threshold. Pen & Jackson (2008) argue that this critical value should be even 4. In this research the highest VIF value (2.943) is lower than these numbers.

Variable Beta t-Value Significance VIF-value

Constant 1.457 17.243 .000

Direct effects

Standardized: Bank Z-score -.033 -1.789 .074 1.891

Standardized: Firms with a line of credit .108 5.540 .000 2.188 Standardized: Deposit money bank asset to GDP -.056 -3.708 .000 1.316

Standardized: Bank return on assets .007 .349 .727 2.503

Rule of law Factor -.241 -11.254 .000 2.614

Standardized: Starting a Business .107 4.814 .000 2.805

Standardized: Registering property -.061 -3.027 .002 2.333

Standardized: Protecting minority investors -.082 -4.245 .000 2.135

Standardized: Paying taxes -.137 -6.082 .000 2.913

Standardized: Resolving insolvency -.019 -.915 .360 2.432

Macroeconomic stability .023 1.330 .184 2.003

Year of establishment -.026 -1.938 .053 1.006

% of Foreign owned -.004 -6.809 .000 1.017

Table 6 shows the results of the linear regression.

In general it can be stated that, using multiple linear regression leads to severe more significant variables. However, since this is not the complete correct way of analyzing this model, the hypothesis cannot be accepted on these premises.

(29)

29

interesting to look whether size has an effect on the independent variables effect on access to finance, a so-called interaction effect. Table 7 shows the results of this multilevel analysis.

Table 7 presents the results of the multilevel analysis with the interaction terms

All the variables in this analysis have been standardized. As can be seen in the table, there are interaction terms for small and medium firms but not for large firms. For the large firms holds that since the model explains the estimates for small and medium sized firms, what is left for the general terms to explain is the estimates of the large firms. Since, small medium and large firms together represents the whole sample. As in the first analysis, the variables for small and medium have a far

Parameter Estimate Std. Error Sig.

Small .0285083 .0343238 0.406

Medium -.0014356 .0341074 0.966

Bank Z-score -.0045514 .0589331 0.938

Firms with a bank loan or line of credit .1443008 .0761576 0.058

Deposit money banks assets to GDP -.0328228 .0545476 0.547

Bank return on assets -.0834922 .0940368 0.375

Rule of law Factor -.2085056 .0809645 0.010

Starting a business -.0449622 .0849713 0.597

Registering Property .005267 .0698452 0.351

Protecting minority investors .004863 .0784231 0.951

Paying taxes -.1050233 .1008121 0.298 Resolving insolvency -.113024 .081148 0.164 Macroeconomic stability .0237401 .0620731 0.702 Year of establishment .0000951 .0000508 0.061 % of foreign owned -.0033594 .0005551 0.000 Small*Bank Z-score .0035438 .0401711 0.930

Small*Firms with a bank loan or line of credit -.0266611 .0482024 0.580

Small*Deposit money banks assets to GDP -.0487339 .034675 0.160

Small*Bank return on assets .0691376 .0529006 0.191

Small*Rule of law factor .0228177 .0474656 0.631

Small*Starting a business .1219164 .0510057 0.017

Small*Registering property .0984531 .0094855 0.111

Small*Resolving insolvency -.0024068 .0438245 0.956

Small*Protecting minority investors -.0784286 .052094 0.132

Small*Paying taxes .0425611 .0577338 0.461

Medium*Bank Z-score .0113763 .0454367 0.802

Medium*Firms with a bank loan or line of credit -.0936626 .0499377 0.061

Medium*Deposit money banks assets to GDP -.0181978 .0356334 0.610

Medium*Bank return on assets .0592764 .0521182 0.255

Medium*Rule of law factor .0665992 .0493302 0.177

Medium*Starting a business .0682855 .0529272 0.197

Medium*Registering property .0589423 .0046854 0.267

Medium*Resolving insolvency .0238261 .0442644 0.590

Medium*Protecting minority investors -.0607255 .0529459 0.251

(30)

30

from significant effect. The first ten independent variables after the size variables, represent the base category and thus the estimates of the effect for the large variable. From these variables, only the rule of law factor is significant (P<0.05). It has an negative effect on the dependent variable of -0.209. Since it is expected that when the rule of law gets better, the problems with access to finance go down. Furthermore, the firms with a bank loan or line of credit variable, representing the access to financial institutions, is almost significant. However, the estimate is positive, meaning that when the access to financial institutions goes up, the problems with access to finance go up as well which is a strange result. From the control variables, the percentage of being foreign owned is highly

significant (P<0.05) while the year of establishment variable is almost significant. The latter has an positive very small estimate because when the year of establishment goes up with one, means that a firm is younger and thus that the problems with access to finance are bigger. This is, in contrast to the results in table 5, what was expected. For the percentage of being foreign owned the opposite is true, a higher percentage means less problems with access to finance and thus a negative estimate.

When looking at the interaction terms for the small firms, only the variable for starting a business is significant (P<0.05). This result is expected since small firms are more likely to be in a start-up process where the ease of starting a business is important. The estimate of 0.122 is related to the base category. This means that the effect of the starting a business variable is 0.122 times bigger for small firms then for large firms. Furthermore, the variables small*firms with a bank loan or line of credit and the small*rule of law factor are far from significant while they were significant for the large firms.

Considering the interaction terms with the medium firms, none of the variables are significant, although firms with a bank loan or line of credit is close. It is interesting to see that the rule of law factor variable is again not significant, although it is closer to being significant than for the small firms, it is still far off. While this variable was two of the three times significant in the original models, based on this results, it can be concluded that it only influences large firms on a significant level. The opposite can be said for firms with a bank loan or line of credit variable. While only significant one out of four times in the original models, it is almost significant for both the large and medium firms.

Considering the hypotheses, the following can be said:

Hypothesis 1: In a country with good financial institutions, there is better access to finance for firms with a bigger effect on SMEs than on large firms.

(31)

31

Hypothesis 2: In countries with a good legal system, there is better access to finance for firms with a bigger effect on SMEs than on large firms.

Hypothesis 3: In a country with a good accounting infrastructure, SMEs have better access to finance. Hypothesis 5: In countries with less corruption, there is better access to finance for firms with a bigger effect on SMEs than on large firms.

Hypotheses 2,3 and 5 represent the rule of law factor variable. As stated above, this variable is significant or almost significant in the original model. But when checked for size, only significant for large firms. Since the hypotheses all consider SMEs, these can all be rejected. It would appear that it is only significant for large firms.

Hypothesis 4: A restrictive regulatory environment has an negative effect on access to finance for SMEs.

This hypothesis can be rejected as well. Only once one of the variables is significant, the interaction term for Small*starting a business. This is far too little ground and thus should be rejected.

The only real conclusion that can be made from the above is that size does not seem to have an influence on access to finance. With the variables for Small and Medium being far from significant and only one of the interaction terms being significant, size seems to have a very small influence.

5.2. Robustness Checks

To check the robustness of the results, some variables will be replaced with other appropriate variables. This will be done with the size variable, the dependent and the independent variables.

5.2.1 Independent variable

To check the robustness of the independent variable, the independent variable is replaced with two different question from the BEEPS questionnaire. These are; “To what degree are crime, theft and disorder an obstacle for this establishment”, and, “At this time, does this establishment have a line of credit or a loan from a financial institution?” The results can be found in Appendix B. The results show that the independent variables clearly give a different result and thus that the dependent variable is robust.

5.2.2 Continuous size variables

(32)

32

firms have less than 50 employees and medium sized firms less than 250 (What is an SME?, 2013). This shows that the definition used in the BEEPS questionnaire is a very conservative one. To get a better view of the impact that size has, it can be interesting to look at a continuous variable for size so that one can better observe how the variables change with every extra employee or euro of sale added.

5.2.2.1 Employees for determining size

First, I will look at the effect every extra employee added has. To do this, another question from the BEEPS questionnaire is used; “At the end of fiscal year 2007, how many permanent, full-time

employees did this establishment employ?” The results of this analysis can be seen in table 8. Again, all the variables are mean centered.

(33)

33

Table 8 shows the result of the analysis with the continuous size variable.

5.2.2.2 Sales for determining size

Besides looking at a continuous size variable, it is also important to consider other definitions of size. Up to now, all the analysis are done with number of employees as the definition of size. However, a labor intensive company can have more than 100 employees but still can be seen smaller, when looking at sales, than a company that is very IT-intensive and has less employees. Therefore, I will look at the effect size has on the effect independent variables have on the access to finance variable. For this, a different question from the BEEPS questionnaire is used: “In fiscal year 2007, what were this establishment’s total annual sales?” Table 9 shows the results of this analysis.

Constant 1.413928 .2946691 0.000

Number of fulltime employees .0000152 .0000427 0.721

Bank Z-score .0002898 .0553475 0.996

Starting a business .0438551 .0774204 0.571

Protecting minority investors -.0431325 .0680531 0.526

Paying taxes -.0770273 .0925549 0.405

Resolving insolvency -.1114019 .0751419 0.138

Macroeconomic stability .0245311 .062167 0.693

Year of establishment .0001188 .0000541 0.028

% of foreign owned -.0035925 .000552 0.000

Size* Bank Z-score -.000054 .0000846 0.523

Size* Firms with a bank loan or line of credit -.0000113 .0000734 0.878

Size* Deposit money banks assets to GDP .0000475 .0000512 0.354

Size* Bank return on assets -1.84e-06 .000083 0.982

Size* Rule of law factor .0001349 .0000749 0.072

Size* Starting a business -.0001869 .0000761 0.014

Size* Registering property .0000785 .0000684 0.223

Size* Resolving insolvency .0000628 .0000538 0.242

Size* Protecting minority investors 2.58e-06 .0000709 0.971

(34)

34

Table 9. Shows the results with the Sales as size variable

The results in table 9 are similar to those in the analysis with the number of employees as a measure of size, in that sales again is a continuous variable. Meaning that the estimate of the interaction terms with sales represent the change in the independent variables effect if one extra euro of sales is made. Furthermore, the estimates are even smaller in this analysis while the effect of one extra euro is even smaller than the effect of one extra employee since one added employee costs more than one added euro of sales.

When looking at the independent variables, only the rule of law factor variable is significant and the firms with a bank loan or line of credit is almost significant while the control variable for percentage of being foreign owned is significant as well. While being (almost) significant in almost all of the earlier tests, year of establishment does not have a significant effect on access to finance when sales are included. For the interaction terms with sales and the independent variables, none of the results are significant. This leads to believe that is does not matter if size is expressed in sales or employees.

Constant 1.42685 .2922312 0.000

Sales -2.36e-11 2.64e-11 0.371

Bank Z-score -.0193904 .0544345 0.722

Starting a business .02366 .0767917 0.758

Protecting minority investors -.0458973 .0681519 0.501

Paying taxes -.0878304 .0920399 0.340

Resolving insolvency -.1065272 .0749017 0.155

Macroeconomic stability .0289461 .0616281 0.639

Year of establishment .000076 .0000679 0.263

% of foreign owned -.0035962 .0005993 0.000

Sales* Bank Z-score 9.44e-11 7.32e-11 0.197

Sales * Firms with a bank loan or line of credit -3.56e-11 4.64e-11 0.443

Sales * Deposit money banks assets to GDP -1.39e-11 4.44e-11 0.754

Sales * Bank return on assets -5.03e-11 4.43e-11 0.257

Sales * Starting a business -5.32e-11 6.81e-11 0.435

Sales * Registering property 4.37e-11 4.46e-11 0.219

Sales * Resolving insolvency 5.47e-11 5.48e-11 0.319

Sales * Protecting minority investors 2.41e-11 6.46e-11 0.710

Sales * Paying taxes 2.61e-11 2.47e-11 0.291

(35)

35

5.2.3 Institutions and control variables

To further check the robustness of the model, two variables from the Global competiveness

Indicators report have been added (World economic Forum, 2012). These are the public and private institutions which together are all the institutions. Replacing these variables for all the other

variables except the control variables, gives almost the same result. Both are not significant and only percentage of being foreign owned is significant while year of establishment is almost significant. The results of this test can be seen in appendix C.

Since two of the three control variables do seem to be significant, it is important to also check for other control variables. Since the Macroeconomic stability is far from significant it may be good to look at other macroeconomic variables. Therefore, I choose to change the macroeconomic stability for the different countries inflation Gross Domestic Product per capita (GDP). When using these two control variables in model 4 of the original model, not much changes. The Rule of law factor is still significant and no other independent variables. The GDP and inflation do not have a significant influence. The inflation however is significant on the 10% level (0.094). These results can be seen in appendix C.

Small- and medium-size enterprises’ access to finance: The role of institutions