Algorithm for Defining Clusters based on Input–Output Tables: Case of Construction Cluster of Russia

This research presents an algorithm for cluster identification based on input–output matrixes. Authors present an algorithm for downstream and upstream analysis of the symmetrical input–output matrix, which allows definition of the top input and output suppliers and consumers for each industry. As a result of the algorithm, related industries and clusters can be defined. The program, which implements the proposed algorithm, was written using Python. In this paper, the algorithm is applied to the analysis of «Construction» industry of Russia. We used the latest input–output matrix available for Russia for 2016, which contained information on 98 industries. We defined clusters and industries that are the top suppliers and consumers of the «Construction» industry. Among the top suppliers for the «Construction» industry are «Metal manufacturing», «Automotive cluster», and «Chemical products cluster», which account for 15.01%, 9.63%, and 5.95% of overall consumption, respectively. Top consumers of the «Construction» industry are «Public administration and defense; compulsory social security», «Real estate activities», and «Human health and social work activities», which account for 23.3%, 12.19%, and 6.26%, respectively, in the volume of output. The proposed algorithm can be used for analyzing input–output matrixes and cluster identification. Using the results of its application, the decision-makers can elaborate on policy for supporting the cluster-based development of the regions.

Construction cluster; Input–output matrixes; Regional specialization

    Currently, in the context of forming Industry 4.0 and digitalizing industrial enterprises, innovatively active industrial clusters start playing a key role (Tashenova et al., 2020). Cluster development positively impacts the regional economy and employment levels (Moeis et al., 2020). Identification of industrial clusters, analysis of relationships between cluster presence, and economic performance are widely explored topics (Ketels and Protsiv, 2020). One of the main ideas behind the cluster is maximization of agglomeration effects, which can arise from localization, competition, and knowledge exchange. Researchers use two main approaches in order to identify clusters (Skhvediani and Sosnovskih, 2020). The first approach is based on analysis of localization quotients of the related industries. This approach allows to identify industries that have relatively high concentration at the regional level compared to the country average (Slaper et al., 2018; Kudryavtseva et al., 2020). The data on the values of regional localization coefficients make it possible to assess their competitive specialization and allow them to identify efficient regional clusters (Pavlov et al., 2015). The localization coefficient characterizes the concentration of enterprises belonging to a certain industry, but clusters can consist of enterprises of various industries connected by buyer-supplier relations. Another approach for cluster identification is based on the downstream and upstream analysis of symmetrical input–output matrix (Titze et al., 2011; Morrissey and Cummins, 2016). Input-Output Analysis shows that, in the overall economy, there are interrelationships and interdependencies between sectors. The tables present data on sales or shipments between companies in different industries, which allow to calculate what portion of its resources a company in one industry purchased from enterprises of other industries. The idea of cluster analysis lies in identifying strong patterns of cross-industry interaction. Groups of industries with strong connections are called value-added production chains or clusters. Despite the lack of practice in compiling cross-industry balance sheets at the regional level, such tables provide an overview of the possible relationship between enterprises and industries. Also, a combination of these methods can be used, as was done by Delgado et al. (2015).

    Research on defining clusters based on input–output analysis is quite popular in scientific literature. For example, a worldwide, input-output network was built based on the global, multi-regional, input-output tables (Cerina et al., 2015). Also, input–output analysis was used to identify clusters of industries with high carbon emissions in Japan (Kanemoto et al., 2019), to identify clusters in German industries with intensive research and development (Kosfeld and Titze, 2017), and to find industrial clusters in the Beijing-Tianjin-Hebei region in China (Guo et al., 2019). Analysis of input-output tables was carried out in a study by Thai scientists to identify Thai rubber cluster (Tengsuwan et al., 2019). Indonesian researchers worked to identify the role of agricultural sectors in Jambi economy (Fitri et al., 2019). For Russian cases, this method was used to analyze 40 industries’ input-output flows for 2007 (Markov and Markova, 2012), and the composition of the textile cluster of the Ivanovo region (Valitova et al., 2021). Therefore, there is quite a limited amount of research dedicated to cluster identification, based on input–output tables.

    In this paper, we present a program algorithm, which allows to identify top suppliers and consumers. This algorithm was used in the «Construction» industry using the 2016 input–output table (Federal State Statistics Service, 2020). 

This paper contributes to the topic of cluster identification based on the input–output tables. We developed the program in Python and presented the programming algorithm for downstream and upstream analysis of the symmetric input and output tables. We used this algorithm at example of Russian data on 98 industries for 2016. In this paper, we presented the results for the «Construction» industry. In particular, we defined top suppliers and consumers of this industry and identified clusters that are related to it. To the best of our knowledge, this is one of the first works that used data on Russian input-output linkages for the last years in order to define related industries and form clusters. Previous works about Russia have mainly used localization quotients for cluster identification and international data on linkages between industries. There are several limitations to our research. The first limitation is that we determined the connectivity at the country level; we did not consider the presence of these clusters in the regions. The second is that we did not use coefficient on the localization employment variable in order to check whether related industries located at the same region. The third is that we extrapolate data from 2016 to the present date. Future research should consider the development of the programming algorithm to receive the map of interconnected industries automatically. In addition, it should allow to identify changes in structure of interconnected industries in time. Input–output table for 2021 is expected to be published by a Russian statistical service in 2022-2023. In the future, it will identify the spatial localization of clusters at the regional level, considering intersectoral relations.

    This research was funded by the Russian Science Foundation. Project No. 20-78-10123.

Badusheva, V., Palagin, A., 2020. Development of the Construction Industry under the Influence of COVID-19. Vestnik Akademii Znanii, Volume 39(4), pp. 81–85

Cerina, F., Zhu, Z., Chessa, A., Riccaboni, M., 2015. World Input-Output Network. PloS One, Volume 10(7), pp. 1–21

Delgado, M., Porter, M., Stern, S., 2015. Defining Clusters of Related Industries. Journal of Economic Geography, Volume 16(1), pp. 1–38

Fitri, Y., Amir, A., Murdi, S., Syafaruddin., 2019. Using Input-Output Analysis Approach to Identify the Role of Agricultural Sectors in Jambi Economy. Russian Journal of Agricultural and Socio-Economic Sciences, Volume1(8), pp. 552–562

Federal State Statistics Service, 2020. Official Site Federal State Statistics Service. Available Online at https://rosstat.gov.ru/accounts, Accessed on October 21, 2021

Guo, J., Lao, X., Shen, T., 2019. Location-Based Method to Identify Industrial Clusters in Beijing-Tianjin-Hebei Area in China. Journal of Urban Planning and Development, Volume 145(2), June 2019

Kanemoto, K., Hanaka, T., Kagawa, S., Nansai, K., 2019. Industrial Clusters with Substantial Carbon-Reduction Potential. Economic Systems Research, Volume 31(2), pp. 248–266

Ketels, C., Protsiv, S., 2020. Cluster Presence and Economic Performance: A New Look Based on European Data. Regional Studies, Volume 55(2), pp. 208–220

Kosfeld, R., Titze, M., 2017. Benchmark Value-added Chains and Regional Clusters in R&D-Intensive Industries. International Regional Science Review, Volume 40(5), pp. 530–558

Kudryavtseva, T., Skhvediani, A., Berawi, M.A., 2020. Modeling Cluster Development using Programming Methods: Case of Russian Arctic Regions. Entrepreneurship and Sustainability Issues, Volume 8(1), pp. 150–176

Luo, S., Yan, J., 2009. Analysis of Regional Industrial Clusters’ Competitiveness based on Identification. In: International Conference on Electronic Commerce and Business Intelligence, Beijing, China , pp. 471–474

Markov, L.S., Markova, V.M., 2012. Revealing Reference Clusters: Methodical Questions and the Practical Application to the Domestic Industry. Bulletin of Nsu Social Economic Sciences, Volume 12(1), pp. 95–108

Moeis, A.O., Desriani, F., Destyanto, A.R., Zagloel, T.Y., Hidayatno, A., Sutrisno, A., 2020. Sustainability Assessment of the Tanjung Priok Port Cluster. International Journal of Technology, Volume 11(2), pp. 353–363

Morrissey, K., Cummins, V., 2016. Measuring Relatedness in a Multisectoral Cluster: An Input–Output Approach. European Planning Studies, Volume 24(4), pp. 629–644

Pavlov, K., Rastvortseva, S., Cherepovskaya, N., 2015. A Methodological Approach to Identifying Potential Clusters in Regional Economy. Regional Economics: Theory and Practice, Volume 10(385), pp. 15–26

Peeters, L., Tiri, M., Berwert, A., 2001. Identification of Techno-Economic Clusters using Input-Output Data: Application to Flanders and Switzerland. Innovative Clusters: Drivers of national innovation Systems. In: OECD proceedings, 251-272

Semenenko, V.Y., 2020. Global Construction Market in the Context of the Covid-19 Pandemic: Trends and Factors of Influence. ASR: Economics and Management, Volume 3(32), pp. 313–316

Skhvediani, A., Sosnovskikh, S., 2020. What Agglomeration Externalities Impact the Development of the Hi-tech Industry Sector? Evidence from the Russian Regions. International Journal of Technology, Volume 11(6), pp. 1091–1102

Slaper, T.F., Harmon, K.M., Rubin, B.M., 2018. Industry Clusters and Regional Economic Performance: A Study Across U.S. Metropolitan Statistical Areas. Economic Development Quarterly, Volume 32(1), pp. 44–59

Tashenova, L., Babkin, A., Mamrayeva, D., Babkin, I., 2020. Method for Evaluating the Digital Potential of a Backbone Innovative Active Industrial Cluster. International Journal of Technology, Volume 11(8), pp. 1499–1508

Tengsuwan, P., Kidsom, A., Dheera?Aumpon, S., 2019. Economic Linkage in the Thai Rubber Industry and Cluster Identification: Input-Output Approach. Asian Administration & Management Review, Volume 2(2), pp. 147–159

Titze, M., Brachert, M., Kubis, A., 2011. The Identification of Regional Industrial Clusters using Qualitative Input–Output Analysis (QIOA). Regional Studies, Volume 45(1), pp. 89–102