Innovative Development of Kazakhstan’s Raw Material (Oil and Gas) Regions: Multifactorial Model for Empirical Analysis

The following paper reveals the applied aspects of multi-factor analysis allowing identifying patterns of innovative development of Kazakhstan’s raw material (oil and gas) regions in connection with modernization transformations. Initial signs of indicators for the period between 2008 and 2020 are currently under investigation using factor analysis. All data have been analyzed with IBM SPSS 23. In considerable detail, authors describe the methodology of this study and provide the results of statistical analysis. The analysis has revealed factors determining implementation of modernization transformations and taking effect on innovative development of Kazakhstan’s raw material regions, which are as follows: Regional Economic Development and Agglomeration Effects, Market Potential and Infrastructure, Structural Factor of Innovative Development, Human Factor of Innovative Development, and Investment Factor of Innovative Development. It is concluded that stimulation of innovation activity can be based on the following public policy measures: increasing investment in fixed assets; growth of gross regional product and product and process innovation costs; regional development of information and communication technologies; expansion of lifelong learning programs; poverty reduction; increase in the share of R&D employees; development of small businesses; increasing investment in education and in the number of technical and STEM students. The obtained results also allowed us to conclude about the completeness of identified factors of innovative development of Kazakhstan’s raw material (oil and gas) regions and the need for further research in the context of studying the stated issue.


INTRODUCTION
Back in 2020 and 2021, the share of mineral raw materials and products would account for 66% in the structure of Kazakhstan's exports.This shows that Kazakhstan's economic power directly depends on sales of natural resources, i.e. revenues from oil and gas exports form a significant part of the republican budget of the country.
The raw material factor interpreted in the scientific literature (Wang et al., 2021;Rahim et al., 2021;Khan et al., 2020;Li et al., 2020) as a "resource curse" for the countries with economies in transition and rich in mineral resources, is determining the country's socio-economic development.The processes of institutional transformations are being hindered in the raw materials economy (Aljarallah, 2021;Haque, 2020;Pelzman et al., 2018;Vakulchuk and Overland, 2018), since raw material factor in the economy contributes to the obscurity of social distribution of natural resource rent.An inefficient institutional environment hinders economic growth by reducing quality indicators, which in turn has a negative impact on the quality of life, which manifests itself in an increase in social stratification and in socio-economic inequality.
Raw material specialization determines regional imbalances in the level of economic development.Raw material (oil and gas) regions This Journal is licensed under a Creative Commons Attribution 4.0 International License B u k e t o v U n i v e r s i t y attract mobile and skilled labor resources; attract investments thereby turning into leading centers on certain socio-economic indicators.Concurrently, adverse weather conditions and the high cost of infrastructure maintenance form an obstacle to their sustained socio-economic development.
In our previous studies (Kurmanov et al., 2020), results of the analysis of Kazakhstan's raw material regions indicate a low level of innovation activity, instability of regional development, which predetermines the strengthening of the search for factors and new tools and measures to ensure the boost of the existing potential for the creation and implementation of regional innovation with oil and gas production's predominance in the economy.

LITERATURE REVIEW
Various studies (Zemtsov et al., 2017;Crescenzi and Jaax, 2017;Ó hUallacháin and Leslie, 2007;Bottazzi and Peri, 2003;Feldman and Florida, 1994;Jaffe, 1989) show that R&D costs, investment climate, availability and quality of human capital, the level of economic diversification, the flow of knowledge positively influence the effective regional innovation activity.
In addition, scientific literature reveals a link between the level of economic development and the level of innovative activity of regional enterprises; however, its direction is definitely impossible to assume something of.Development of regional innovative entrepreneurship is facilitated by the growth of GRP and, particularly, GRP per capita as an indicator of the volume of consumer markets, the solvency of the population and the quality of life (Reynolds et al., 1994).A number of studies have revealed that startup activity take an effect on GRP per capita (Fritsch and Storey, 2014;Audretsch and Keilbach, 2004).
Research in Kazakhstan conditions requires taking into consideration the characteristics of the economic structure of a raw material (oil and gas) region.On the one hand, extractive industry's dominance in the structure of the regional economy can cause a "Dutch disease:" a decrease in the economic activity of enterprises and monospecialization, ultimately leading to a decrease in the level of entrepreneurial and innovative activity (Egert and Leonard, 2008).On the other hand, regions with a raw-material economy enjoy higher incomes of the population; accordingly, the purchasing power is growing as well, ultimately contributing to the growth of mass entrepreneurship in the service sector.
In their empirical study, Reynolds et al. (1994) conclude that investments take a positive effect on innovative activity of regional enterprises.Global technology giants (Samsung, HP, Apple, Huawei, Google, etc.) invest heavily in R&D, support startups, maintain research units, and implement joint innovative projects.Educational level of the population serves as an indicator of concentration and quality of human capital.This rate also indicates informal rules and norms in society.Through the education system, the government can influence the development of creative entrepreneurship and innovation (Abad-Segura and -Zamar, 2019).Training and introduction of advanced training courses for the population contribute to the acquisition by individuals of necessary competencies to engage in innovative entrepreneurship.Therefore, the study should consider this indicator.

González
The scientific literature (Zemtsov et al., 2021;Fritsch and Wyrwich, 2018;Lee et al., 2004;Audretsch and Fritsch, 1994) demonstrates that regions with major markets, agglomerations and adjacent territories with high incomes (and therefore, high purchasing power) show increased demand for new products and services.This opens up market niches for creating and implementing innovations in them.
The smaller the average size of one regional entity, the higher the barriers to entry to the local market and the lower the density of innovative enterprises in it (Plummer, 2010;Lee et al., 2004).Chepurenko et al. (2017), Audretsch and Belitski (2017) note that the regional development of innovation activities requires highquality information, communication, and innovation infrastructure including access to digital resources and online markets.Modern digital platforms provide access to global consumers, technologies, and the labor market.

DATA SOURCES AND METHODS
Singling out "raw regions" is not customary in the practice of classification of Kazakhstan regions.As a key criterion in the study, a comprehensive analysis of socio-economic development, development of mechanisms for managing the innovative development of Kazakhstan's raw material (oil and gas) regions use the share of gross value added from oil and gas production in the structure of gross regional product for the period between 2008 and 2020 (Figure 1).
To assess the level of organization of innovation activity in selected regions, we shall use the following research methods: panel data (Appendix A), and a factor analysis.Table 1 shows the main variables for factor analysis.
After collecting statistical data, we initiated the next stage of the study, which was to evaluate the data with the statistical analysis software, SPSS 23.Then we applied the method of reducing the amount of data.This method of factor analysis is used primarily to compress information and reduce the number of variables based on their classification.Concurrently, variables that strongly correlate with each other are grouped.For factor analysis, we used eighteen variables influencing innovation activity in the raw material region.
According to the analysis conditions, all signs of variables were expressed quantitatively.
The factor analysis was performed on the basis of its main stages, which were the following: 1. Assessment of model quality and verification of data suitability for analysis using indicators of the Kaiser-Meyer-Olkin sample adequacy measure and the Bartlett criterion, 2. Calculation of initial factor loadings using the principal component method, 3. Varimax factor selection and rotation.Coefficients are rotated to find factors facilitating interpretation, and 4. Data interpretation.As a hypothesis of the study, we have identified the following confirmatory (confirming) provisions, which factors are currently important in the innovation activity of the raw material region and how complete they are in determining them.

Indicators of Development of Kazakhstan's Oil and Gas Regions
The indicator of the share of gross value added from oil and gas production in the GRP identifies four raw material regions of Kazakhstan: Atyrau, West Kazakhstan, Mangystau, and Kyzylorda regions.Oil and gas production dominates over the extraction and export of other types of natural resources in the socio-economic development of these regions (Figure 1).
Let us take a closer look at the economic indicators of the development of Kazakhstan's oil and gas regions (Appendix A).
Over  The level of poverty of the population in the regions in question has been significantly reducing over a thirteen-year period.Accordingly, Atyrau region saw the reduction in the poverty level by 9.9%, West Kazakhstan region by 6.3%, Kyzylorda region by 18.5%, and Mangystau region by 26.7%.However, we feel important to note that the average poverty level in Kazakhstan for 2020 was 5.3%.Kyzylorda and Mangystau regions exceed this level with indicators of 5.8 and 5.7, respectively.
Accordingly, we can conclude that among Kazakhstan's oil and gas regions, Atyrau and West Kazakhstan are developing most rapidly.It should be noted that from oil and gas companies pay the profits from oil and gas sector exports to state budget in two ways.First, tax payments including special subsoil user payments (redirected to the National Fund of the Republic of Kazakhstan).
Second is export customs duties on crude oil and oil products (to the republican budget).Specifics of Kazakhstan's financial system consist in the following: regional budgets are formed from certain types of taxes and fees (IIT, social tax, environmental charges, etc.).Insufficient funds to finance the expenses of regional budgets Accordingly, selected regions show the following common features: • High endowment of natural resources in demand on the world market, • Primary natural resource allocation in areas with adverse weather conditions, • Poor regional infrastructure (social, industrial, transportation, innovation), • Region's landlocked location increasing transportation and logistical costs, • Low population density and underdevelopment of the settlement system, and • Regional technological backwardness.

Factor Analysis Results
Following statistical data collection, first thing to do is to check their suitability for factor analysis.The results have shown the following.The first indicator is a Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy, a value that characterizes the degree of applicability of factor analysis to this sample.High values (0.5 to 1.0) usually indicate that factor analysis is applicable to these data (IBM Knowledge Center).The value below 0.50 shows the impropriety of the factor analysis.In our case, KMO is 0.577 > 0.5, which is a good result.The second indicator, the Bartlett's test, is used to verify sufficiency of correlation of initial variables.This test should be significant (P < 0.05), otherwise factor analysis will be inappropriate.In the model we built, this indicator is 0.000, which also indicates reliability of the model.Table 2 shows the results of both KMO measure of sampling adequacy and Bartlett's criterion.
For the next stage, we calculated initial factor loadings using the method of principal components to obtain certain data: • Initial communalities are estimates of each variable's variance considered by all components or factors.In correlation analysis for extraction of the main components, their values of 1.0 are always the same, • Extraction communalities are estimates of each variable's variance considered by the components (IBM Knowledge Center).Low values indicate variables not suitable for a factor solution and may need to be excluded from the analysis.
The communality value of 0 tells that the factor does not affect the variable.Value of 1 implies the variance of the variable is determined by the selected factor in its entirety.
The analysis showed that generalities in this table are high.This indicates that the extracted components represent the variables .000 Compiled by the authors based on IBM SPSS 23 data B u k e t o v U n i v e r s i t y well.Table 3 shows variable names and their communalities (Table 3).
The next step of the study was the Varimax factor selection and rotation.The purpose of factor extraction is to reduce a larger set of variables to a smaller set of "artificial" variables called principal components, which account for most of the variance of the original variables.To decide which factors to keep for further analysis, we use formal criteria.These are all factors whose individual values are greater than one.
The leftmost section of the Table 4 shows the variance explained by the initial solution of factor extraction.Only five factors at the initial stage of the solution have eigenvalues exceeding 1.These factors will serve as the basis further.Together, they account for almost 83% of the variability of the baseline variables.This suggests that innovation activity in the raw material regions of Kazakhstan is influenced by five hidden factors; however, there is also room for many unexplained variations.The second section of this table shows the variance explained by the extracted factors before rotation.
The rightmost section of this table shows the deviations explained by the extracted factors after rotation.The rotated factor model introduces changes to all factors.
To confirm the factors found, we used the method of factor extraction -the Rocky Scree criterion by R. Kettell.It consists in finding the point where the decrease in eigenvalues slows down the most.Figure 2 shows five main factors that have eigenvalues greater than one.We can also see the importance of each factor by comparing them with each other.
After extracting the factors, for a more apparent interpretation of the solution, we used the Varimax rotation method of the initials, which allowed us to trace a clear factor structure and to identify variables marked by high values of correlation coefficients with a given factor.Correlation is considered strong if the correlation coefficient value exceeds 0.7.
The rotated component matrix helps to determine what the components are.5).
As a result of the analysis, we have identified five main factors affecting innovation activity in the raw material regions of Kazakhstan.In general, these factors explain 83% of the total variance.
Next, let us give an interpretation of these factors according to the results of which officials and interested persons can make appropriate tactical and strategic decisions (Table 6).
The above matrix shows that the "strength" of factors presented, or the weight of all identified factors is 83%, which confirms the possibility and necessity of considering them as priorities for the innovation activities of oil and gas regions of Kazakhstan.We could not identify the remaining 17% of factors; this is the scope for future research.
The most important value is an assessment of interrelationships of the initial indicators with the obtained factors.The conducted assessment allows us to establish an economic rationale to the factors identified as a result of the analysis.
Accordingly, such indicators as Fixed Asset Investments, Gross Regional Product per Capita, Gross Regional Product, Product and Process Innovation Costs, and Region's Central City Residents that formed Factor 1 are advised to interpret as regional economic development and agglomeration effects.The main focus is on the economic indicators of regional development:

Indicators forming
• On investments to a greater extent, • On GRP growth, and

B u k e t o v U n i v e r s i t y
Public Administration Bodies) depend on market potential and infrastructure development.
Indicators forming Factor 3 are mainly presented in the form of structure and shares and include such indicators as R&D Employment, Innovation Activity Level, Population at the End of the Year, Share of Crude Oil and Natural Gas Production in GRP, and Average Organization Size.Based on the content of these indicators, we define Factor 3 as a structural component of innovative development.
The set of indicators that determined the economic content of Factor 4 characterize the human factor of innovative development of the oil and gas region.Such indicators as Average Organization Size and Students per 1.000 Population characterize the quality of human capital.Therefore, this group is interpreted as a human factor of innovative development of the oil and gas region.
Indicators forming Factor 5 reflect investment support for innovation.In particular, Education Investments and Innovative Production Volume significantly depend on large-scale investment support.This explains the investment factor of innovative development.

CONCLUSIONS AND RECOMMENDATIONS
The raw material structure of the economy of Kazakhstan's oil and gas producing regions hinders innovative development due to low demand for new technologies and other aspects of the "resource curse."The existing literature has yet to analyze this dependence further.
Indicatively, raw material (oil and gas) regions establish relatively weak demand for new technologies and, accordingly, for innovations.Their number, as Fritsch and Wyrwich (2018)'s example of Germany shows, is historically lower in the lands adjacent to the coal mining regions due to the "resource curse" squeezing local capital and personnel out to a more profitable raw materials sector.Dependence on natural rents leads to destruction of local institutions and corruption depriving technological entrepreneurs of incentives to initiate new projects.Unlike large commodity companies, which, as a rule, are not happy with the emergence of competitors, large diversified agglomerations see higher innovation activity due to player concentration and competition, market scale and diversity, etc. (Zemtsov et al., 2021;Beaudry and Schiffauerova, 2009;Audretsch and Fritsch, 1994).
Regional high-tech clusters (Belitski and Desai, 2015) provide entrepreneurs with access to the appropriate infrastructure and knowledge and generate the effect of knowledge flow from large companies and universities to innovative startups.
This study was conducted with a goal to determine the main factors affecting the ability of raw material (oil and gas) regions of Kazakhstan to encourage innovation processes.In the course of the work, we used a statistical research method and factor analysis.Factor analysis allowed us to identify five main factors taking effect on innovation activity in the raw material regions of Kazakhstan:

Figure 1 :
Figure 1: Share of gross value added from oil and gas production in the regions's Gross Regional Product for the period between 2008 and 2020 Changes for 2008-2020 in terms of regions' central city population are as follows: • The center of Atyrau region (the city of Atyrau): population increased by eleven pp; • The center of West Kazakhstan (the city of Uralsk): population increased by nine pp; • The center of Kyzylorda region (the city of Kyzylorda): population increased by six pp; • The center of Mangystau region (the city of Aktau): population decreased by nine pp.

Table 1 : Variables selected for factor analysis
with their own revenues push them to use transfers from the state budget as a source of missing funds.Consequently, if budget revenue redistribution is inefficient, there may be a situation where regions with a great potential receive less subsidies than regions with less capacity for economic development, which in the long term may slow down the economic growth of the state as a whole.
BNSCity_ residents Region's Central City Residents, % Calculations Internet1 Internet User Organizations (Incl.Public Administration Bodies), units BNS Internet2 Share of Internet Users Aged 16-74, % BNS 1) Compiled by the authors.2) BNS is Bureau of National statistics, Agency for Strategic planning and reforms of the Republic of Kazakhstan.3) HDI is the United Nations' Human Development Index In 2020 compared to 2008, Fixed Asset Investments increased by 285% in Atyrau region, by 113% in West Kazakhstan region, by 70% in Kyzylorda region, and by only 52% in Mangystau region.Over the same period, Education Investments in Atyrau, West Kazakhstan and Kyzylorda regions increased by 152%, 245%, and 28%, respectively, and decreased by 36% in Mangystau region.In 2008-2020, Product and Process Innovation Costs increased in Atyrau, West Kazakhstan and Kyzylorda regions, and decreased in Mangystau region.The Average Organization Size by region for 2020 was as follows: 32 in Atyrau region, 38 in West Kazakhstan region, 40 in Kyzylorda region, and 27 in Mangystau region.In Kazakhstan, the Average Organization Size is more associated with the number and role of budgetary institutions, while correlation with the number of innovative enterprises in the raw material region requires verification.The Average Expected Education during the Coming Life in all regions has increased by 6%.Number of Students per 1.000 Population in all studied regions decreased over the analyzed period by 47% in Atyrau region, by 2% in West Kazakhstan region, by 44% in Kyzylorda region, and by 50% in Mangystau region.Indicators characterizing accessibility of Internet in all oil and gas regions of Kazakhstan have increased significantly: on average, by 220% for Internet User Organizations, and by 330% for the Share of Internet Users Aged 16-74.

Table 4 : Total variance explained Component Initial eigenvalues Extraction sums of squared loadings Rotation sums of squared loadings Total % of variance Cumulative % Total % of variance Cumulative % Total % of variance Cumulative
Table 5 shows that the first, i.e., the general component correlates most strongly with Fixed Asset Investments, Gross Regional Product per Capita, Population Engaged in Professional, Scientific and Technical Activities, Gross Regional Product, Product and Process Innovation Costs, and Region's Central City Residents.The second component correlates most strongly with Share Internet Users Aged 16-74, Average Expected Education during the Coming Life, and Proportion of the Population with Incomes below the Subsistence Minimum (Poverty Level).It also has an average connection with Internet User Organizations (Incl.Public Administration Bodies).Extraction method: Principal component analysis.

Table 3 : Communalities
Innovation Activity Level, and Population at the End of the Year.It also has an average relationship with Share of Crude Oil and Natural Gas Production in GRP, and Average Organization Size.The fourth component has a relationship with Average Organization Size, and Students per 1.000 Population.The last fifth component is closely related to Education Investment, and Innovative Production Volume (Table

Table 5 : Rotated component matrix a
Extraction method: Principal component analysis.Rotation method: Varimax with kaiser normalization.a Rotation converged in 9 iterations Regional Economic Development and Agglomeration Effects, Market Potential and Infrastructure, Structural Factor of Innovative Development, Human Factor of Innovative Development, and Investment Factor of Innovative Development.Accordingly, for the innovative development of raw material (oil and gas) regions of Kazakhstan, state bodies need to focus mainly on the following measures: Second, on increasing the variables for analysis, since the factors we have obtained explain only 83% of the total variance.The remaining 17% of the variance are factors yet to be found, and • Third, confirmation of the reliability of the results obtained requires different statistical analysis methods.