THE USE OF REGRESSION ANALYSIS IN THE EVALUATION OF THE IMPACT OF DIGITALISATION AND TECHNOLOGICAL INOVATIONS IN THREE MEASURES OF THE DEVELOPMENT OF ECONOMY AND SOCIETY

Cilj rada je da se ispita korelacija između dostignutog nivoa tehnologije i inovacija i razvoja finansijskog tržišta, indeksa humanog razvoja i bruto domaćeg proizvoda po glavi stanovnika. Istraživačko pitanje glasi: „Da li su, i u kakvoj vezi, nivo tehnologije i inovacija sa indeksom razvijenosti tržišta, indeksom ljudskog razvoja i BDP-om po glavi stanovnika?” Cilj istraživanja je da utvrdimo da li ima osnova za verovanje da je „kopiranjem” određene zemlje po nivou digitalizacije i inovacija moguće dodatno razviti finansijsko tržište, uticati na nivo ljudskog razvoja ili na povećanje BDP-a po glavi stanovnika. Metode korištene u radu su regresiona analiza, odnosno prosta linearna regresija, te analiza i sinteza prethodnih istraživanja i teoretskih nalaza, da bi zaključci bili izvučeni metodom indukcije. Rad predstavlja doprinos autora ekonomskoj teoriji i praksi, te široj javnosti. Rezultati istraživanja još ukazuju i da bi bilo racionalno da finansijski posrednici u zemljama u razvoju razmotre izmjenu svojih poslovnih modela i mogućnosti za njihovo prilagođavanje ubrzanim tehnološkim promjenama.


Introduction
In the past three centuries, civilization went through three industrial revolutions, while according to the opinion of the world's leading economists, the fourth industrial revolution is ongoing, the one popularly called the Industry 4.0. Each of these industrial revolutions was characterised by technological innovations that had a key impact on the development of the entire mankind. What is characteristic of the Industry 4.0 is that it already affects all business activities, in different ways, while simultaneously developing digital and other technologies, but also affecting the entire global lifestyle. After the process of globalization and intertwining of the world into one global market, which resulted in an unobstructed expansion of business, a new era began, which may be called the age of digital transformation. The basic characteristic of the new, digital age is that it takes on new dimensions and new forms, with each passing day. Accelerated technological development, the expansion of smart devices, and the mass production of mobile devices, as some of the elements of the fourth industrial revolution, have challenged financial institutions, which have responded with digitalisation, the creation of new communicative channels towards clients, as well as with a variety of other innovative services (Šehović, 2017, 136).
Although banks are less likely to accept changes due to their structure, line of business and other characteristic, they have largely adjusted their business to the changes in the business environment and, consequently, adopted and implemented certain processes imposed by the digitalisation process. The continuous process of creating new banking products and services which are directly linked to the digitalisation process, is a clear sign that the banking sector has seriously acknowledged the upcoming changes, which certainly result in the creation of a competitive advantage and a better position in the market. According to experts, technological innovations and clients will "set new rules of the game" in the banking sector (Ćukić, 2013), which will significantly affect banks, especially the ones in which traditional banking is prominent (Tornjanski, V., Petrović, D., & Milanović, M., 2016). In order for banks to retain their competitiveness, growth and development, and to continually create valuable products for banking service users, as well as shareholders, bank management should without any delays acknowledge these trends and redefine existing business strategies (Fasnacht, 2009;Huo & Hong, 2013;Tornjanski et.al., 2014) and develop new models for the expansion of knowledge.
Research so far indicates that financial institutions and the population largely benefit from the process of digitalisation and innovations (Laursen & Salter, 2006;Fasnacht, 2009). On the other hand, despite the digitalisation representing an important step in the application of innovations, there is not enough research on these processes in the existing body of scientific research, i.e. on how digitalisation, innovations and information and communication technologies contribute to the performances of open innovations in banking (Tornjanski et al., 2016). Bearing that in mind, this paper aims at deepening and expanding the body of research on the effects of digitalisation and innovations in the financial market and the population, along with indicating the significance of these processes.
The first research hypothesis reads: "Technology and innovations do not affect the FD index." The second research hypothesis reads: "Technology and innovations do not affect the HDI." The third research hypothesis reads: "Technology and innovations do not affect GDP per capita."

Previous Research
Industry 4.0 or the so-called fourth industrial revolution, i.e. industry digitalisation, has for a few years been the key topic explored in order to find answers to how economies can be more competitive in global markets (Mekinjić, 2019).
The essence of industry 4.0 is in a new approach, i.e. intertwining smart digital devices and products, tools, robots and people, while its basic objective are smart factories able to adjust and efficiently integrate clients and business partners into a unique process (Mekinjić, 2019).
Upravo odatle proizilazi cilj ovog istraživanja. Cilj je da se utvrdi da li je "kopiranjem" određene zemlje po nivou digitalizacije i inovacija moguće dodatno razviti finansijsko tržište, uticati na nivo ljudskog razvoja ili na povećanje BDP-a po glavi stanovnika. S tim u vezi, pod terminom "tržište u razvoju" podrazumijevamo zemlјe koje odlikuju institucionalne turbulencije, nizak nivo korporativnog upravljanja i ekonomskog razvoja u odnosu na razvijene zemlјe. Hoskisson i saradnici kao zemlje u tranziciji izdvajaju sve zemlje Zapadnog Balkana (Hoskisson et al. 2000, 249-267 In this way, productivity and efficiency would increase and secure competitiveness in the global market.One of the interesting aspects of this revolution is that it was scheduled in advance, i.e. out of necessity due to the crisis, recession and slowdown of economic activities, which led leading countries in the European Union to search for a solution to strengthen their economies and global competitiveness. According to research conducted by the European Banking Federation (EBF, 2018), it is believed that a unique digital market will improve the development of companies performing business in this system and that it will be beneficial for all clients, further economic growth and further employment. Therefore, the fourth industrial revolution has brought thorough and essential changes and it has resulted in a completely new economicsdigital economics (Lazarević, Đuričković 2018, 27).
There are three main components of digital economics (Lazarević, Đuričković 2018, 27): • e-business infrastructure (hardware, software, telecoms, networks, human capital, etc.); • e-business (a focus on how business is realised, i.e. any business performed by the organisation on a computer through a network); • e-trade (transfer of goods, i.e. a book being sold online). At the same time, OECD considers digital economics "an umbrella term for the description of markets focused on digital technology". It includes trade in information goods or services through electronic trade. It functions through a layered foundation with separate segments for the transport of data and applications.
The results of existing research indicate the need for introducing the concept of openended innovations in the banking sector, i.e. adequate incorporation of external knowledge into innovation processes through suitable technologies (Tornjanski, Petrović, & Milanović, 2016). In accordance with that, the role of banks in this transformation is not only to be innovative partners investing in innovative financial technologies, but also to contribute to the economic growth and development in the overall financial market (Mekinjić, 2019). . Therefore, the diversity of financial systems in different countries means that it is necessary to look at multiple indicators in order to measure financial development. For example, Grujić (2019) explored the impact of pension fund structure on the development of financial markets across countries and showed that different pension fund structures can affect the development of financial markets.
Indeks humanog razvoja predstavlja kompromis između sveobuhvatnosti i mjerljivosti ( characterised by institutional turbulence, low level of corporative management and economic development in comparison with developed countries. Hoskisson and associates consider all the countries in Western Balkans as countries in transition (Hoskisson et al. 2000, 249-267). For illustration purposes, the institutional legacy of communism in those markets is reflected in large, undisciplined and inefficient administration, bureaucratic approach by institutions, as well as in corruption (Haramija & Njavro, 2016). Namely, "bureaucratic and restrictive governance enabled corruption and bribery of public officials because most citizens saw that as the only way of achieving the desired goal" (Dimitrova-Grajzl & Simon 2010, 206). Even a brief insight by bureaus of statistics in countries in Western Balkans confirms the continuation of such practice, i.e. an increase in the number of employees in budget financed fields: administration, public administration, education and art. On the other hand, it is obvious that there are fewer employees in the processing industry. Besides, relevant research also confirms a high level of corruption, as a consequence of the communist system, in new member states of the EU in comparison with the "old" member states (Transparency International 2016). In addition, when one observes research on the trust in institutions, countries in transition are at the bottom of such lists (Bjørnskov 2007). Besides, in all small and open-ended economies, such as in states constituting the Western Balkans, the capacity of monetary policy is limited by numerous factors (Benazić & Rami, 2016, 1039. Hence, the criticism of transition is based on significant increase in poverty and deterioration of mostly middle class (Cifrić 1996, 137).
By observing developing markets it may be noted that, in both countries of the Western Balkans and Bosnia and Herzegovina, noneconomic factors in the region play the most important role in determining the amount of trade between countries (Trivić & Klimczak 2015, 57). Economic instability is a consequence of "frequent reforms which completely disregard economic growth and the social impact of changes, low rates of domestic and foreign investments, foreign-trade deficit and low rate of GDP" (Duvnjak, 2018, 198). In BiH there is a number of social and economic issues which have still not been resolved, the improvement of which entails complex and demanding solutions (Amidžić et al. 2016, 57).
Da bi se prevazišli nedostaci pojedinih indikatora kao zamjena za finansijski razvoj, kreirano je mnoštvo indeksa koji pokazuju kako su razvijene finansijske institucije i finansijska tržišta u smislu njihove dubine, pristupa i efikasnosti, što kulminira konačnim indeksom finansijskog razvoja (ilustracija 1). Ovaj indeks je prvobitno razvijen u kontekstu napomene MMF-a za diskusiju o osoblju "Preispitivanje finansijskog produbljivanja: stabilnost i rast na tržištima u razvoju" (Sahay et al. Each of these financial functions can affect decisions on savings and investments and the efficiency of allocation of assets. As a result, finances affect the accumulation of capital and the total factor productivity, i.e. the three factors determining economic growth. To the extent that financial development diminishes the asymmetries of information and financial limitation and promotes the distribution of risk, development may increase the capacity of financial systems to absorb shock and diminish the amplification of cycles through a financial accelerator (Bernanke, Gertler & Gilchrist 1999), while diminishing macroeconomic volatility and inequality (Svirydzenka, 2016, 4).
Most empirical literature from the seventies relates financial development with the two measures of financial depth -ratio of private credit to GDP and, to a lesser extent, the ratio of the capitalisation of the shares market, also to GDP. For instance, in an influential study at an industrial level Rajan & Zingales (1998) use both measures to show that a higher degree of financial development facilitates economic growth. As for macroeconomic volatility, Dabla-Norris & Srivisal (2013) consider that financial development, measured through private credit against GDP of banks and other financial institutions, plays a significant role in mitigating the instability of production, consumption and investment growth, but only to a certain extent. Most researchers in this field use variations of these measures to examine the role of the financial system in economic development. široki multidimenzionalni pristup definisanju finansijskog razvoja prati matricu karakteristika finansijskog sistema koju je razvio Čihák sa saradnicima. (2012). Imajući u vidu bogatstvo informacija o svojstvima finansijskog sistema -postoji 105 različitih indikatora u GFDD i 46 indikatora u FinStats -nije moguće pratiti sve ove različite pokazatelje pojedinačno, posebno u empirijskom radu. Čak i da je to bilo moguće, nijedan pojedinačni indikator, kada bi se koristio sam, ne bi pružio sveobuhvatno razumijevanje nivoa finansijskog razvoja. Podindeksi i konačni indeks povezuju ove različite pokazatelje i omogućavaju sveobuhvatnu procjenu pojedinih karakteristika finansijskih sistema i ukupnog nivoa finansijskog razvoja. Kao rezultat toga, indeksi dozvoljavaju da se utvrdi gdje su nedostaci u finansijskom razvoju nedoslijedni ili koji aspekti finansijskog razvoja utiču na makroekonomske rezultate, koji bi zatim mogli biti detaljnije istraženi koristeći raščlanjene podatke od FinStats ili GFDD.
Bearing in mind the wealth of information on the features of the financial system -there are 105 different indicators in GFDD and 46 indicators in FinStats -it is not possible to follow all these different indicators individually, especially in empirical work. Even if that were possible, no individual indicator, if used on its own, would provide a thorough understanding of the level of financial development. Sub-indices and the final index connect these different indicators and enable a thorough estimate of specific features of the financial system and the overall level of financial development. As a result, indices enable the determination of aspects in which flaws in financial development are inconsistent or aspects of financial development which affect macroeconomic results, which could in turn be explored in more detail by using the divided data from FinStats or GFDD.
In the remainder of the paper the described methodology is used for the construction of indices, including data sources, the treatment of missing values, functional form and weight used in aggregation. It indicates how the new indices are compared to traditional measures and key stylised facts on the financial development across the world. In the discussion we will look at some limitations and shortcomings of indices with the objective of demonstrating the extent to which the structure and size of pension funds affects the result of indices. The goal is to determine whether it is possible to further develop the financial market by imitating a certain country in terms of the structure and size of indices. Therefore, under the term "developing market" we mean countries characterised by institutional turbulence, low level of corporative management and economic development in relation to developed countries.  the other hand, it is obvious that there are fewer employees in the processing industry. Besides, relevant research also confirms a high level of corruption as a consequence of the communist system in new member states of the EU in comparison with the "old" member states. In addition, when one observes research on the trust in institutions, countries in transition are at the bottom of such lists (Bjørnskov 2007

Data
Data on the level of technological development and innovations were taken from the publication Readiness for the Future of Production Report 2018, data on the market development were expressed through FD index, and the level of human development index was found in data from

Methodology
In the paper we have observed the data given by the National Bureau of Economic Research The relationship between the realised rates of non-payment of liabilities and macroeconomic indicators may be checked in several ways. We used linear regression because we supposed that there is a linear relationship between the independent variable (X) and dependent variable (Y).
Hypotheses were constructed in the following manner: H0 -null-hypothesis 0 = negative H1 -alternative = affirmative In relation to that, the research question was formulated so as to ask whether variable X affects the variable Y. Therefore, the hypotheses are: H0 1 : Technologies and innovations do not affect the FD index.

Results
By analysing the relation between digitalisation and innovations across countries of OECD and FD index, we have obtained the table 4.
R squared (R 2 ) equals 0.137910. That means that the independent variable (variable X -the level of digitalisation and innovations) explains 13.8% of variable Y -FD index. In other words, digitalisation and innovations affect around 13.9% of the financial market development. The multiple correlation coefficient (R) equals 0.371362188 which means that there is a weak direct relation between the independent and dependent variable.
Considering that the p value of 0.02575015 is higher than 0.01, with a 99% certainty we cannot dismiss the hypothesis that "technology and innovations do not affect the FD index" and we can conclude that the "levels of digitalisation and innovations is not in a statistically significant relation with the FD index".
However, when we look at the relation between digitalisation and innovations by countries of OECD and the HD index, we obtain the table 5.
R squared (R 2 ) equals 0.5119. That means that the independent variable (variable X -level of digitalisation and innovations) accounts for 51.2% of variable Y -HD index. In other words, digitalisation and innovations affect the level of human development index in a country, in the degree of about 52.2%. The multiple correlation coefficient (R) equals 0.7155027 which means that there is a strong direct relationship between the independent and the dependent variable. At this moment we can offer arguments for the cause-and-effect relationship in both directions, for all variables and digitalisation. For example, GDP creates conditions for the development of digitalisation, while the digitalisation of economy facilitates the growth of GDP, enabling more assets  S ciljem da još iste hipoteze provjerimo na drugom uzorku izabrali smo manji uzorak slučajno izabranih zemalja za koje postoje parametri koje posmatramo (Prilog 2).
Considering that the p value, 0.0000009365 is significantly lower than 0.01, we can dismiss with a 99% certainty the hypothesis that "technology and innovations do not affect the HD index" and we can conclude that the level of digitalisation and innovations is related to the FD index, assuming that other variables remain unchanged.
In addition, when we look at the relation between digitalisation and innovations by countries of OECD and GDP per capita, we obtain the data from table 6.
After repeating the examination of hypotheses, we reach these conclusions: we cannot dismiss the first one and we will dismiss the second and third hypotheses with a 99% certainty. Therefore, technology and innovations do not affect the FD index but they do affect HDI and GDP per capita. We have shown that there is a weak relationship, i.e. a low correlation (0.49%) between the level of digitalisation and innovations and FDI and that there is a 24.65% determination. However, we have shown that there is a high correlation between the level of digitalisation and innovations and HDI (a 68.88% correlation and about 47.45% determination), as well as the level of digitalisation and innovations and GDP per capita (a 67.34% correlation and about 45.34% determination).
U radu smo pokazali da veza inovacija i indeksa finansijskog razvoja jeste nešto slabija. about 19.5%. The multiple correlation coefficient (R) equals 0.44167693, which means that there is an average direct relationship between the independent and dependent variable.
With the aim of examining the same hypotheses in a different sample, we have chosen a smaller sample of randomly chosen countries for which parameters observed exist (Supplement 2).

Concluding Remarks
We have proved that the level of digitalisation and innovations does not affect the level of the financial market development. This hypothesis was confirmed in a sample of countries outside OECD and in a sample of less developed countries. On the other hand, we have shown that the level of digitalisation and innovations affects the level of human development and the level of GDP per capita.
We have shown that digitalisation and innovations affect the financial market development in countries of OECD to a degree of only 13.9%. The multiple correlation coefficient equals about 37.13%, which means that there is a low direct correlation between the independent and the dependent variable. Considering that the p value of 0.02575015 is higher than 0.01, with a 99% certainty we cannot dismiss the hypothesis that "technology and innovations do not affect the FD index" and conclude that the level of digitalisation and innovations are in a statistically significant correlation with the FD index.
On the other hand, we have shown that digitalisation and innovations affect the level of human development index in a country to a degree of about 52.2%. The multiple correlation coefficient (R) equals 0.7155027, which means that there is high direct correlation between the independent and dependent variable. Considering that the p value of 0.0000009365 is significantly lower than 0.01, with a 99% certainty we can dismiss the hypothesis that "technology and innovations do not affect the HD index" and conclude that the level of digitalisation and innovations significantly affects the FD index, provided that other variables remain unchanged.
In the end, digitalisation and innovations affect the level of GDP per capita in a country, to a degree of about 19.5%. The multiple correlation coefficient equals 0.44167693, which means that there is a moderate direct correlation between the independent and dependent variable.
We reached the same conclusions when we examined the hypotheses in a randomly chosen sample of countries outside the OECD.
After repeating the examination of the hypothesis, we reach the same conclusions: we cannot dismiss the first one and we dismiss the second and third hypotheses with a 99% certainty. Therefore, technology and innovations do not affect the FD index but they do affect HDI and GDP per capita. We have shown that there is a weak relationship, i.e. a low correlation (0.49%) between the level of digitalisation and innovations and FDI and there is a 24.65% determination. However, we have shown that there is a high correlation between levels of digitalisation and innovations and HDI (a 68.88% correlation and about 47.45% determination), as well as between the level of digitalisation and innovations and GDP per capita (a 67.34% correlation and about 45.34% determination).
By analysing the existing trends and indicators of the level of development of financial services digitalisation, it is obvious that financial intermediaries in developing countries will be obliged to change their business models and adjust them to accelerated changes in the market, or to form alliances with large technological companies, as well as with smaller companies which have complementary solutions in the same manner as banks. Besides, the financial sector will develop towards openended financial services, which will lead banks and other participants in the financial industry to further adjust their business and services. In accordance with that, the regulations need to be changed in the direction of facilitating digitalisation and innovations in the financial sector, as well. The results of the paper indicate that financial institutions in developing countries, primarily banks, since they are the closest to the populations, should react as soon as possible in order to prepare for the future, when innovations and new technologies will have the principal role, by constructing adaptable and digitally prepared business models which will help them answer all the upcoming challenges. The paper also points to Međutim, to još uvijek ne znači da digitalizacija i inovacije nisu prodrle u finansijski sektor onoliko koliko u ostatak privrede. Prvo, to bi moglo da znači i da digitalizacija i inovacije ne mijenjaju dubinu, pristup i efikasnost finansijskog sistema, iako su u njemu zastupljene. Drugo, to bi moglo da znači da digitalizacija jeste važna za finansijski razvoj, ali da stari indeks to ne mjeri dobro, dakle da bi možda trebalo menjati indeks koji se koristi. Treće, to bi moglo i da znači ono što se u radu naglašava, a to je da digitalizacija finansijskog sektora kaska za digitalizacijom ostatka privrede. Međutim i dve interpretacije su moguće. S tim u vezi ostavljen je i prostor budućim istraživanjima da razvrstaju koji je od ovih faktora odgovaran za manju vezu između inovacija i finansijskog razvoja u poređenju sa ljudskim razvojem ili BDP.
Mekinjić B., Grujić M., Vujičić-Stefanović D. Primjena regresione analize u procjeni uticaja digitalizacije i tehnoloških inovacija na tri mjere razvoja privrede i društva recommendations for further research in this field, because it is necessary to explore which innovations contribute the most to specific segments of the financial system and society.
A superficial interpretation of the results would lead to the conclusion that a lower R 2 of regressions having the FDI as a dependent variable points to a weaker cause-andeffect relationship between innovations and digitalisation on the one hand and financial development on the other hand, disregarding the direction of the relationship (i.e. from innovation toward financial development, or vice versa). However, it is true that both these variables are affected by a variety of factors and a higher R 2 of regressions between, for example, innovations and human development does not necessarily mean that these two variables significantly affect each other, but it might mean that they are determined by a third factor, like for example GDP. For example, the system of values followed by two brothers may be almost the same, but not because they influenced each other, but because they were raised by the same parents. Therefore, the significance assigned to R 2 should have less emphasis than in this research, but it is still worth to emphasise that it is lower in regressions with FDI. Hence, among the interrelationships of all four variables, the one with the FDI is the weakest -without considering if that is the result of a mutual weaker influence of these factors, or of other different factors not considered in this paper and potentially influential.
Secondly, the issue remains of the interpretation of the statistical significance of the regression coefficient. The paper insists on the relation between innovations and financial development is not significant, while the relation to other variables is substantial. Such interpretation is based on using a 1% threshold, while it is not sustainable when using, for example, a 5% threshold, which would also be quite acceptable in this context. This fact needs to be emphasised so as not to exaggerate the significance of the distinction found through data analysis.
In the paper we have shown that the relation of innovations and the FDI is somewhat weaker. However, that still does not mean that digitalisation and innovations have not entered the financial sector to the extent that they already permeate the rest of the economy. First, that might also mean that digitalisation and innovations do not change the depth, access and efficiency of the financial system, although they are present in it. Secondly, it might mean that digitalisation is important for financial development, but that the old index does not measure that properly, i.e. that potentially the index being used should be changed. Thirdly, it could also mean what is emphasised in the paper, i.e. that the digitalisation of the financial sector is falling behind in comparison with the digitalisation of the rest of the economy. However, two interpretations are also possible. In relation to that, there is some space for future research to determine which of these factors facilitates a weaker relationship between innovations and financial development, in comparison with the human development or GDP.
Undoubtedly, the financial sector's development is falling behind when it comes to changes in digitalisation. This idea fits in with the higher p-value of the coefficient, which digitalisation uses to explain the FDI over the other two variables. This claim becomes significant if we know how the FDI is constructed. Namely, the factors affecting it are not significantly linked to digitalisation by their very definition, hence the weaker relation.
The paper indicates correlations between digitalisation and technological innovations on the one hand and FD index, HD index and GDP per capita on the other hand. Determining what and to what extent determines the other is left for further research.
After repeating the examination of the hypothesis, we reached the same conclusions: we cannot dismiss the first one, and we dismiss the second and third hypotheses with a 99% certainty. Therefore, technology and innovations do not affect the FD index but they do affect HDI and GDP per capita. We have shown that there is a weak relationship, i.e. a low correlation (0.49%) between the level of digitalisation and innovations and FDI, and there is 24.65% determination. However, we have shown that there is a high correlation between the levels of digitalisation and innovations and the HDI (a 68.88% correlation and about 47.45% determination), as well as between the levels of digitalisation and innovations and GDP per capita (a 67.34% correlation and about 45.34% determination).