CC BY
20УДК: 303.725.33
ЭКОНОМЕТРИЧЕСКОЕ ИССЛЕДОВАНИЕ ФАКТОРОВ ИННОВАЦИОННОГО РАЗВИТИЯ СТРАН МИРА
Жигляева А.В., студент Финансовый университет при Правительстве Российской Федерации, Москва, Россия
E-mail: [email protected] Невежин В.П., профессор Финансовый университет при Правительстве Российской Федерации, Москва, Россия
E-mail: [email protected]
Аннотация. Целью настоящего исследования является выявление факторов, оказывающих наиболее существенное влияние на инновационное развитие национальных экономик стран мира. В качестве основы для проведения исследования приняты основные компоненты, учитываемые при расчете Глобального инновационного индекса (ГИИ). Эти ежегодно публикуемые статистические данные позволяют оценить место страны в международном инновационном развитии. Необходимо специфицировать эконометрическую модель, в которой в качестве объясняемой переменной выбрана величина Глобального инновационного индекса, а в качестве объясняющих - показатели по основным группам в соответствии со структурой ГИИ. Для достижения поставленной цели были решены следующие основные задачи: 1) проведен эконометрический анализ данных по выборке из 30 стран на основе доклада «Глобальный инновационный индекс» за 2018 год; 2) построены множественные регрессионные модели - линейная, полиномиальная, гиперболическая и степенная; 3) с помощью специальных тестов реализована проверка моделей на гетероскедастичность и автокорреляцию случайных остатков; 4) проведена проверка параметров и регрессионных уравнений в целом на значимость и адекватность. По результатам исследования выбрана модель, которая наилучшим образом аппроксимирует исходные данные. Сформулированы рекомендации по применению модели для прогнозирования уровня инновационного развития в Российской Федерации.
Ключевые слова: эконометрическая модель, Глобальный инновационный индекс, инновационное развитие, факторы, прогнозирование
ECONOMETRIC STUDY OF THE FACTORS OF INNOVATIVE DEVELOPMENT OF THE WORLD COUNTRIES
A.V. Zhiglyaeva, student Financial University under the Government of the Russian Federation, Moscow, Russia E-mail: [email protected] V.P. Nevezhin, professor Financial University under the Government of the Russian Federation, Moscow, Russia
E-mail: [email protected]
Abstract. The purpose of this study is to identify the factors that have the most significant impact on the innovative development of national economies of the world. The main components taken into account when calculating the Global Innovation Index (GII) are taken as the basis for the study. These annually published statistics allow us to estimate the country's place in international innovation development. It is necessary to specify an econometric model, in which the Global Innovation Index value is chosen as the explained variable, and the indicators for the main groups in accordance with the GII structure as the explanatory variables. To achieve this goal, the following main tasks were solved: 1) an econometric analysis of data from a sample of 30 countries was carried out based on the Global Innovation Index report for 2018; 2) built multiple regression models - linear, polynomial, hyperbolic and power; 3) with the help of special tests, model checking for heteroscedasticity and autocorrelation of random residues was implemented; 4) the parameters and regression
equations in general were tested for significance and adequacy. According to the research results, a model was chosen that best approximates the original data. Recommendations for using the model to predict the level of innovation development in the Russian Federation are formulated.
Key words: econometric model, Global Innovation Index, innovative development, factors, forecasting
1. INTRODUCTION
The study of innovative economies acquires special relevance today. This is due, firstly, to the close relationship between technological innovation and economic growth. Investing in innovation generates significant profits [10, P. 55]. The rate of turnover of capital and the life cycle of innovation are in direct correlation. Countries with a high level of innovation development have relatively stable economic growth, which directly affects the social well-being of the population. Therefore, innovative development is a priority of economic policy [1, P. 38]. The importance of innovation in resource-intensive countries is also increasing. The irreplaceable nature of the reserves of fuel and raw materials in these countries, fluctuations in world prices for raw materials make the economies of these countries extremely unstable and vulnerable from the point of view of future development. This is especially true for Russia. Learning theoretical concepts and practical experience of other countries in the field of innovative development is necessary for further socio-economic development. The role of the innovation factor with its interrelation with other growth factors in country development is considered by many authors as a determinant of transition to a new social formation. A number of authors speak about the formation of a postindustrial society, where knowledge and innovation will play the main role (N. Shter, A. Gorts, H. Wilke, M. Castells). It is the production innovation forms the basis of a postindustrial knowledge society. Nevertheless, for strategic planning of innovation development, enhancing the country's competitiveness, ensuring a higher position in the international ranking of innovation development, it is necessary to clearly define the key drivers of this development, the main factors that directly affect the resulting innovation indicators. With the help of econometric studies, it is possible to build models with the inclusion of the most important factors. These models, if they are significant, adequate, can be used to predict
indicators of the country's innovative development, as well as, for example, be used by state authorities to determine the directions of budget investments and priority directions for state budget expenditures [4, P. 35].
The main aggregate indicator, giving a comprehensive assessment of the level of innovative development of the country, is the Global Innovation Index, calculated by the International Business School INSEAD (in points, separately for each country). When calculating the GII, 2 subindexes are taken into account: 1. The subindex of innovation resources (institutions, human capital and science, infrastructure, development of the internal market, business development); 2. Subindex of innovation results (development of technologies and knowledge economy, development of creative activity) [8, P. 367]. For the construction of multiple regression models in the present study, the variable is chosen as the explanation: Y - Global Innovation Index (in points, max = 100). The following variables were selected as explanatory variables: X1 - institutions (institutional environment, regulation), X2 - human capital and research, X3 - infrastructure, X4 -domestic market development, X5 - business development (innovative entrepreneurship), X6 -results in science and technology, X7 - results in the field of creating intangible assets and the development of creative activities. In order to achieve homogeneity and comparability of baseline data, only countries with the highest Global Innovation Index (top - 30 international GII rankings based on 2018) were included in the sample of 30 countries. The leader of the GII-2018 rating is Switzerland (the index value is 68,40 out of 100). Slovenia closes the group of 30 selected countries (46,87 out of 100 points). The selected data was used to build 4 different regression models with their subsequent comparison, assessment of the influence of factors and the choice of the model that best fit the original data.
2. RESEARCH METHODS
For the econometric analysis, the Excel tabular processor was used, its main tools — Correlation and Regression — were used (to estimate the parameters of the regression models and to obtain calculated values for further testing of the models). In the process of building a multiple regression linear model, at the stage of analyzing the vector of correlation coefficients and matrix of correlation coefficients, 2 factors (X5 and X7) were excluded from further consideration, however, the test results on the choice of specification (choice of "long" or "short" regression), as well as the results of the comparison of models according to the Akaike criterion and the Bayesian Schwarz information criterion, showed that the specification of the "long" econometric model is better than the "short" one. Consequently, a baseline regression with seven factors was chosen for further research. For this model, the VIF test (analysis of the inflation factor of variation) showed that there is no multicollinearity of the explanatory factors (since the value VIFmax = 3,28 does not exceed the critical value, which is VIF = 10) [9, P. 131]. The verification of the residues of the obtained linear regression for heteroscedasticity using the Goldfeld-Quandt test and the Spearman's rank correlation test gives grounds for concluding that random perturbations are homoscedastic. For each value of the factor, random residues have the same variance. Therefore, it can be assumed that the application of the least squares method will result in unbiased and effective parameters. The results of the verification of the obtained linear regression for autocorrelation using the Durbin - Watson test, the Sved - Eisenhart series method and the Breusch -Godfrey test suggest that there is no autocorrelation of residuals (random residues (ui) and (uj) are independent of each other). Consequently, there will be no deterioration in the quality of estimates of regression parameters by the method of least squares, as well as an overestimation of the test statistics, by which the quality of the model is checked. According to the results of the Helvig agreement test and the Shapiro - Wilk test, random residues of the obtained linear regression are distributed according to the normal law. Therefore, regression parameters are also normally distributed. This condition is also
necessary to test other conditions of the Gauss-Markov theorem in order to use the least squares method to estimate model parameters [7, P. 135].
A similar set and order of methods and procedures was applied in the process of building three multiple non-linear regression models. According to the results of checking the parameters of the models for significance (t - test) and for adequacy, as well as checking the obtained regression models for significance (F - test) and adequacy, it was found that all parameters are significant and adequate, and all regression models are also significant and adequate. In particular, when comparing the values of ai with the intervals calculated for them, it can be seen that the intervals for the parameters do not pass through the value zero, and therefore all parameters should be considered adequate. Similar results are obtained if for the considered parameters to compare their values presented in the section "Variance analysis" in the columns "Lower 95%" and "Top 95%". 3. RESULTS
According to the results of the study, four multiple regression models were obtained (one linear, three non-linear). The multiple regression linear model is (1):
y = 0,1145 + 0,0996x1 + 0,1003x2 + 0,0986x3 + 0,1001x4 + 0,0990x5 + 0,2502x6 + 0,2505 x7 + u (1)
The determination coefficient for this model is 0,999995289 (the variation of the explained variable is more than 99,99% due to the variation of the explanatory variables), and the value of the average approximation error (0,02%) does not exceed 1215%, which indicates good quality of the obtained regression.
The constructed multiple polynomial econometric model has the following form (2): y = 27,75 + 0,0005x2 + 0,00096x2 + 0,0009x| + 0,0008x4 + 0,0011x| + 0,0024x6 + 0,0024x2 + £ (2)
The determination coefficient for this model is 0,994610202 (the variation of the explained variable is more than 99,46% due to the variation of the explanatory variables), and the value of the average approximation error (0,51%) does not exceed 12-
15%, which indicates good quality of the obtained regression.
The type of the resulting hyperbolic econometric
model is presented as follows (3):
1 i
y = 108,43 - 746,49 —- 320,75 —-
Xi #2
1111 371,63 —- 304,05 —- 509,02 —- 702,10 —+
X4 X5 Xg X7
e (3)
The obtained value of the average approximation error (1,05%) does not exceed 12-15%, which indicates the good quality of the obtained regression. Since the coefficient of determination is 0,9774 (97,74%), this also indicates the good quality of the resulting regression.
And finally, the multiple power econometric model is as follows (4):
y =
D qqqar0,138 0,093 0,115 0,117 0,112 0,197 0,225 e \jf j j j ^ 2 3 5 6 7
(4)
The obtained value of the average approximation error (0,08%) does not exceed 12-15%, which indicates the good quality of the obtained regression. Since the coefficient of determination is 0,9982 (99,82%), this also indicates the good quality of the resulting regression. The variation of the explained variable by 99,82% is due to the variation of the explanatory variables.
So, as a result of the research, linear and nonlinear regressions were obtained to approximate the initial data. Based on the presented data, linear regression is selected for subsequent use, which has the highest value of the coefficient of determination (0,999995) and the lowest average approximation error (0,02%). 4. CONCLUSION
The resulting multiple linear regression equation, chosen as the best fit to the original data, is meaningful and adequate, has good quality and can be used for further research. The most important factors are X7 - results in the field of creative activities and X6 - results in science and technology. These are the results that are located on the "output" in the national innovation system, which serve as a certain reflection of the results already achieved, the effectiveness of the innovation, scientific and technical policy pursued in the state. At the same time, the largest contribution among the indicators at
the "entrance" (innovative resources) is made by X2
- human capital and research and X4 - domestic market development. When implementing the state strategy of innovation development (innovation strategy), in particular, in the modern Russian realities, it is especially important to pay attention to the development of precisely these highlighted components of the country's integrated innovation potential. The obtained results are confirmed in the works of many authors - for example, the domestic researcher in the field of innovation and innovative development Zueva O.A. believes that it is human, scientific and technical elements-potentials that are central to the structure of the innovative potential of the country [2, P. 27]. According to I. V. Shevchenko, E. N. Alexandrova, I. V. Shlyakhto, S. G. Alekseeva, V. S. Vasiltsova, the most significant elements of the potential of a national innovation system are social and institutional components [5, P. 3]. Analytically and empirically confirmed the allocation of the most significant factors of innovative development - X2, X4 (innovation resources), X6 and X7 (achieved innovation results)
- will allow more rational and efficient allocation of public funding, support for innovation activity in the country. These factors are the "points of growth" that should be actively influenced in order to increase the country's innovation rating in the international community, as well as to ensure an increase in the contribution of innovation to GDP growth. According to Russian researchers A.N. Kozitsina, I.V. Filimonenko, on the availability and status of resources (personnel, finance, research base) and the possibilities of their use - in essence, disposable potential - directly depends on the choice of innovation development strategy, and, consequently, the effectiveness of innovation activity [3, P. 170].
The obtained and tested multiple linear regression model can be used to predict the values of the Global Innovation Index for the Russian Federation, in particular, to more effectively achieve a number of national innovation strategy goals. In addition, using this model it is possible to build scenario forecasts of the country's innovative development, for example, by predicting the values of individual factors using various modern methods of macroeconomic planning
and forecasting. It should be noted that for Russia innovation development in the regional, spatial and territorial aspects is also very important. In order to predict at the regional level, taking into account the high interregional differentiation of socio-economic and innovative development, it is recommended to use the Regional Innovation Index, use its main components for calculations [6, P. 115]. Finally, along with forecasting innovation development with the help of the constructed econometric model, it is recommended to analyze changes in the relationship between indicators of resources and results (as components of the GII), in order to determine growth trends or reduce innovation efficiency, effectiveness, degree of results with initial resources.
СПИСОК ИСПОЛЬЗОВАННЫХ ИСТОЧНИКОВ
[1] Жигляева, А.В. Инновационный потенциал Российской Федерации: роль в обеспечении экономической безопасности // Скиф. Вопросы студенческой науки. - 2018. - № 4 (20). - С. 38-41.
[2] Зуева, О.А. Методологические подходы к исследованию сущности и измерения инновационного потенциала // Научный журнал НИУ ИТМО. - 2016. - № 3. - С. 20-31.
[3] Козицина, А.Н., Филимоненко, И.В. Стратегическая взаимосвязь инновационного развития и управления кадровым потенциалом региона // Экономические и социальные перемены: факты, тенденции, прогноз. - 2015. - № 1 (37). - С. 167-179.
[4] Новиков, А.Г. Зарубежный опыт инновационной инфраструктуры региона // Стратегии бизнеса: анализ, прогноз, управление. - 2017. - № 9 (41). - С. 31-43.
[5] Реймер, В. Инновационная система России: проблемы управления и перспективы / В. Реймер, А. Бреусов // Международный сельскохозяйственный журнал. - 2013. - № 2. - С. 3-6.
[6] Рейтинг инновационного развития субъектов Российской Федерации. Выпуск 5 / Г.И. Абдрахманова, П.Д. Бахтин, Л.М. Гохберг и др.; под ред. Л.М. Гохберга; Нац. исслед. ун-т «Высшая школа экономики». - М.: НИУ
ВШЭ, 2017. - 260 с. [Электронный ресурс]. URL:
https://issek.hse.ru/data/2017/06/09/1170533818 /RIR2017.pdf (дата обращения: 20.03.2019).
[7] Эконометрические исследования: учеб. пособие /В.П. Невежин, А.В. Богомолов, Л.А. Чаговец. - М.: ИНФРА-М, 2018. - 337 с.
[8] Cornell University, INSEAD, and WIPO (2018): The Global Innovation Index 2018: Energizing the World with Innovation. Ithaca, Fontainebleau, and Geneva. - 386 p. [Электронный ресурс]. URL:
https://www.wipo.int/edocs/pubdocs/en/wipo_pu b_gii_2018.pdf (дата обращения: 15.03.2019).
[9] Dougherty, С. Introduction to Econometrics. (4th ed.). 2011, Oxford University Press. - 573 p.
[10] Drucker, P. Innovation and entrepreneurship. Practice and principles. 1993, Collins. - 293 p.
[11] Невежин В.П. Практическая эконометрика в кейсах: учеб. пособие /В.П. Невежин, Ю.В. Невежин. - М.: ИД «ФОРУМ»: ИНФРА-М, 2017. - 317 с.
REFERENCES
[1] ZHiglyaeva, A.V. Innovacionnyj potencial Rossijskoj Federacii: rol' v obespechenii ehkonomicheskoj bezopasnosti // Skif. Voprosy studencheskoj nauki. - 2018. - № 4 (20). - P. 3841.
[2] Zueva, O.A. Metodologicheskie podhody k issledovaniyu sushchnosti i izmereniya innovacionnogo potenciala // Nauchnyj zhurnal NIU ITMO. - 2016. - № 3. - P. 20-31.
[3] Kozicina, A.N., Filimonenko, I.V. Strategicheskaya vzaimosvyaz' innovacionnogo razvitiya i upravleniya kadrovym potencialom regiona // EHkonomicheskie i social'nye peremeny: fakty, tendencii, prognoz. - 2015. - № 1 (37). - P. 167-179.
[4] Novikov, A.G. Zarubezhnyj opyt innovacionnoj infrastruktury regiona // Strategii biznesa: analiz, prognoz, upravlenie. - 2017. - № 9 (41). - P. 3143.
[5] Rejmer, V. Innovacionnaya sistema Rossii: problemy upravleniya i perspektivy / V. Rejmer, A. Breusov // Mezhdunarodnyj
sel'skohozyajstvennyj zhurnal. - 2013. - № 2. -P. 3-6.
[6] Rejting innovacionnogo razvitiya sub"ektov Rossijskoj Federacii. Vypusk 5 / G.I. Abdrahmanova, P.D. Bahtin, L.M. Gohberg i dr.; pod red. L.M. Gohberga; Nac. issled. un-t «Vysshaya shkola ehkonomiki». - M.:NIU VSHEH, 2017. - 260p. [EHlektronnyj resurs]. URL: https://issek.hse.ru/data/2017/06/09/ 1170533818/RIR2017.pdf (data obrashcheniya: 20.03.2019).
[7] EHkonometricheskie issledovaniya: ucheb. posobie /V.P. Nevezhin, A.V. Bogomolov, L.A. CHagovec. - M.: INFRA-M, 2018. - 337 p.
[8] Cornell University, INSEAD, and WIPO (2018): The Global Innovation Index 2018: Energizing
the World with Innovation. Ithaca, Fontainebleau, and Geneva. - 386 p. [Электронный ресурс]. URL:
https://www.wipo.int/edocs/pubdocs/en/wipo_pu b_gii_2018.pdf (дата обращения: 15.03.2019).
[9] Dougherty, С. Introduction to Econometrics. (4th ed.). 2011, Oxford University Press. - 573 p.
[10] Drucker, P. Innovation and entrepreneurship. Practice and principles. 1993, Collins. - 293 p.
[11] Nevezhin V.P. Practical econometrics in the case studies: proc. manual /V.P. Nevezhin, Yu.V. Nevezhin. - Moscow: publishing house "FORUM": INFRA-M, 2017. - 317 p.
V V
