European School which is primary based on the works of Freeman and Lundvall understands the term of innovation system in the broad sense. It is considered not only as a production, but also as a distribution, acquisition and use of knowledge through learning processes.

B.Lundvall and C.Freeman are among the leaders of the development of the national innovation system’s (NIS) theory. Lundvall defines two general methodological principles. The first is that knowledge plays special role in the economic development [7]. The second principle is refers to  the main factor of economic dynamics which  is the competition among entrepreneurs based on innovation. Similarly, Freeman maintains these main principles [3].

Representatives of considered approaches are focused primarily on learning and accumulation of knowledge, They also highlight institutional aspects of knowledge  and various forms of interaction between innovators. The main idea of this approach is that economic activity and dynamics caused by different types of innovation and learning processes, which create and use new technologies. In this idea is considered that the learning process is collective, universal and is influenced both by formal institutions and social norms and values. The first group of institutions are universities, corporations and different regulatory systems. The second one is refers to creating and spreading of knowledge. These two groups are basic premises of the economic activity. Education is understood in the broadest sense. It is considered as the knowledge accumulation and using it in future activities.

Lundvall suggests that the innovation system is formed from the elements and relationships which interact in the production, distribution and use of new and economically useful knowledge [7]. Also he claims that  national system includes elements and relationships ruled by the government . Thus, the definition is focused on the intra characteristics such as elements of the system and the interactions between them without limiting the scope of these specific elements.

Freeman defines innovation system as a network of institutions in the public and private sectors whose activities and interactions initiate, import, modify and diffuse new technologies [7].  Also in his work special attention is paid to the historical development of national institutions which are understood as the result of historical socio-economic trends. An example is the study of Japanese economy, in which C. Freeman introduces the concept of NIS. He  analyze the rapid development of Japan in the second half of the twentieth century.

Thus, in these definitions similarities of NIS are observed:

• the systemic nature of the NIS or an aggregate consideration of its special way of interacting elements;
• the institutional aspect or the impact of society’s formal and informal institutions on the pace and scale of development of innovation;
• the functional feature which considers that the main function of NIS is spreading of new knowledge and technologies.

The concept of innovation systems take into account the entire complex of determinants of knowledge’s production and especially its spread. Therefore, now this concept is widely used by scientists, analysts and international organizations in order to analyze the problems of innovative systems and to develop proposals for eliminating these troubles. Nevertheless, in the world there is no unified understanding of innovative system’s specific structure and it differs significantly in different researches.

In the European Union there is a two-tier measurement system of innovation development: level EIS (European Innovation System) and at the level of EU regions (RIS). The evaluation system of innovative development of European countries has been used since  2000 year. Also the system of evaluation innovative development of the regions of the EU is based on EIS . RIS has become part of the performance of the country survey. Thereby now innovative activity of European Union countries is measured on the basis of 29 indicators, and in order to assess the innovative development in regions 16 indicators are used.

Scientific figures such as the Fritsch, Slavtchev, Andre’s Rodri’guez-Pose, Riccardo Crescenzi,  Jan Fagerberg, Martin Srholec, Bart Verspagen, Laura de Dominicis, Raymond Florax,  Henri de Groot and others are contributed to the study of innovation development. Each author has his own opinion on the factors that affect the development process.

Michael Fritsch and Viktor Slavtchev in their article cover the topic of differences in the efficiency of different innovation systems in regions [5]. The main purpose of the study is to answer the question of what indicators are able to explain differences in the effectiveness of innovation systems in different regions. The basis of their work is the concept of the knowledge production function. As a result, they have identified a list of factors that have positive and negative effects on the efficiency of the system. These researchers find that knowledge spillovers sector within the private and public sector have a positive impact on private innovation. For instance, universities and similar institutions interacted with private-sector firms have a positive effect on the innovative sphere. The authors emphasize that the magnitude of this effect increases when the innovation of the private sector coincides with research topics in scientific universities. According to the article another factor that has a positive effect on innovation performance is population density. In their study  they claim that activities of research and development (R&D) are more productive in agglomerations then in the countryside. They prove their claim and come to conclusion that the West German innovation system is more efficient than the East one. They points out that a close interaction between elements of RIS and the division of labor greatly strengthen the effectiveness of innovation system.

The theme of innovation efficiency is also discussed in the article of Andre’s Rodri’guez-Pose and Riccardo Crescenzi [7]. Unlike other listed authors they try to combine three different approaches in their paper. According to the work, they analyze the interaction of investment with R&D, features of economic growth, the existence of effective regional systems and inhomogeneous distribution of knowledge. In order to analyze these three components the authors construct a regression model for 25 countries which are part of the European Union. In the empirical analysis they find that a complex and close relationship of local and external researches enhances the innovative activity in all regions.

Another work connected with the topic is written by Riccardo Crescenzi, Andreґs Rodrıґguez-Pose and Michael Storper [1]. Scientists also create a regression model and point out that, on the one hand, knowledge, people and capital mobility have a positive effect on innovation. On the other hand, the authors claim that institutional and cultural barriers prevent innovative sector from developing.

Jan Fagerberg, Martin Srholec and Bart Verspagen also consider the topic of innovative activity. In their study they consider the impact of innovation in the broad sense [4]. According to this work innovation is refers to trying of doing something new or improving something old. As a result, scientists have concluded that the country should have special opportunities to improve the efficiency of innovation.

Almost all the authors have different methods of innovative development’s measuring. However, they all come to the conclusion that the innovative sector of the economy develops unevenly in the regions. Listed scientists identify one major problem – spatial inequality of regions. They agree that  inequality inhibits the development of innovation and reduces its efficiency. However, scientists comments this situation in different ways.

Michael Fritsch and Viktor Slavtchev say that the location of innovation is not accidental and add that innovative structures are concentrated in certain areas [5]. Scientists also suggest that one reason for this phenomenon is the variation in access and quality of the material. The second reason is geographically restricted distribution of knowledge. These reasons lead to different levels of release. Authors do not provide any piece of empirical evidence and conclude that the problem is rather sparse.

The study of Andre´s Rodrı´guez-Pose and Riccardo Crescenzi also touch the problem of spatial inequality of region’s innovative development [7]. In their paper they estimate what role the geographical distance plays in the process of innovation. Also authors propose a comparative analysis of innovative systems in two regions: United States and  Europe. Using the linear model of innovation, scholars maintain that local innovative activities are crucial for the production of new knowledge and for the economic exploitation of existing knowledge. Moreover they points out that it is an important factor in European regions. Economic geography also contributes to the development of socio-economic differences. The authors conduct several empirical analysis and conclude that the uneven development of the innovation system appears due to three factors : some regions have low local capacity; information is not equivalent to economically useful knowledge; technological improvements in communication infrastructures do not affected all kinds of information in the same way. Moreover,  the scientists  declare that knowledge production in both continents is governed by different geographical processes. In the paper it is said that the United States should be considered as leader in innovation area. There are several reasons why Europe cannot be identified as a leader. Firstly, in Europe regions level of factors mobility is lower than in the United States. Secondly, despite rapid economic integration, distinct national and regional systems of innovation persist in Europe. A third reason is refers to American’s unity in understanding what innovative system is and for what purposes it is needed. In conclusion the scientists say that geographical processes is very important factor of innovative activity.

Schartinger, Schibany and  Gassler have written one more work which touches the topic of innovative systems [8]. Considering the Austrian regions in particular, they come to conclusion that one another reason of spatial inequality of region’s innovative development is connected with some barriers between universities and business sector. This problem makes it difficult to transfer knowledge from one area to another. According to results of logistic regressions this trouble is caused by differences in cultures of areas and by the lack of information.

To sum up all the studies it can be inferred that regional environment have a rather strong impact on region’s development and business activities itself. The availability of the region to use education, organization, technological innovation  products, learning networks and entrepreneurship influence in the right way improves innovative and economic development. Only complex of factor can be considered in the research. However, most of authors mention following crucial factors: useful knowledge and effective collaboration of different elements of innovative system.

In result of analysis of different articles several general factors of innovation efficiency have been found. These variables touch such aspects as: Institutes which generate knowledge and encourage its dissemination; innovative business activity; innovation infrastructure; integration elements of RIS.

Table 1 Variables

According to the information included in the first table nine variables are picked up for all four issues which are proposed for further research.

Secondly, it is necessary to conduct a descriptive analysis of the sample for a preliminary assessment of its quality. For analyzing quantitative data of samples I calculate such statistical values as the mean, median, minimum, maximum, lower quartile, upper quartile, and the coefficient of variation. Than after generating of a box-plot and a histogram for GDP It is possible to identify whether there are any ejection or not. In the first case ejection should be eliminated for approximation to the normal distribution. In the second case this procedure is ought to be missed. For the next move it is reasonable to use several formal tests on normal distribution: Shapiro-Wilk W test for normal data; Skewness/Kurtosis tests for Normality. The results of three tests for normality allow us to accept the null hypothesis of normality GDP, or reject it. Now the relationship between quantitative variables should be analyzed. After conductions I construct scatter diagrams and the correlation matrix. Based on these results it is possible to make an assumption about the most important factors and the type of the relationships.

The third step is to build a basic linear model and diagnose it. Diagnosis involves finding significant and insignificant factors and the analysis of the coefficient of determination. Insignificant factors should be eliminated and the analysis is to be made again. Also there is a need to conduct formal tests for functional form and heteroscedasticity: Breusch-Pagan-Godfrey; White; Glejser. The next stage is to assess the basic model’s multicollinearity and to create the correlation matrix. If this matrix includes factors which have a strong linear relationship, I get rid of them and repeat this step again.

For the fourth step I evaluate alternative model specifications and to diagnose them according to the previous step. Than receiving new model specifications should be compared with each other. For this move I use comparative analysis of the next three values: Adjusted R-squared; kurtosis; asymmetry. The final stage of the fourth step is selecting the best model. This decision is based on picking up the pattern with the highest R-squared and without heteroscedasticity.

As a result we have a certain method, which is described above. The mythology of the research includes next six steps:

1. to collect data from the OECD Regional Database;
2. to conduct a descriptive analysis of the sample for a preliminary assessment of its quality;
3. to build a basic linear model and diagnose it;
4. to evaluate alternative model specifications and to diagnose them;
5. to assess the impact of indicators on GDP.

Interpreting multiple logistic regression coefficients: the regression coefficient at regressor expresses the elasticity of the dependent variable on this factor at a constant other variables. As it has been mentioned before model proposes that the regressor is regional per capita gross domestic product in country’s regions. On the on hand, if some factors have a weak degree of influence on the GDP, I make the assumption that there are problems in some of considered areas. On the other hand, in the case where factors have a strong influence relevant areas have a favorable environment in the region. Thus we can identify weaknesses and strengths of country’s Innovative system.