Cluster Modelling of Labour Resources Employment in the Context of Globalisation

Abstract

The article examines the essence of the phenomenon, the state and dynamics of unemployment in the structure of the modern economy, analyses the literature on forecasting the employment of labour resources. A new approach to solving the problem of analysing and forecasting the development of the labour market and indicators of labour force employment using the Kohonen self-organising maps is proposed. The basis of this approach is the limited data series for individual countries to obtain meaningful conclusions or forecasts. Therefore, to improve the accuracy of the modelling, it is advisable to expand the information base with relevant data for other countries. However, given the significant differences between different countries, there is a need to identify groups of countries that are similar in terms of the state and development of the labour market. This is where clustering methods come in handy. The study selected more than 40 primary indicators that determine the level of unemployment, employment, labour market conditions, demographic and macroeconomic characteristics of 203 countries over the 12-year period from 2010 to 2021. As a result of the data filtering, 173 countries remained, on the basis of which further analysis and clustering are carried out. When filling in the gaps for these countries, the average values for the corresponding indicator for groups of countries with the same level of human development were taken. The authors also argued for the expediency of using relative indicators in clustering to enable comparison of countries of different sizes. Accordingly, a number of relative indicators from the original list were selected for the final list of factors, and a number of new relative predictors were constructed on the basis of others. A total of 30 indicators were used to build the Kohonen self-organising map, which allowed segmenting countries by their level of socio-economic development and labour force potential. As a result of numerous experiments, it was found that the most effective distribution, in which the indicators of countries retain the greatest similarity in groups, is observed when dividing the worldʼs countries into 12 clusters. In this case, Ukraine falls into a cluster with the following countries: Croatia, Czech Republic, Greece, Hungary, Poland, Slovenia, etc. Ukraineʼs position on the self-organising map indicates a high level of labour market development. Moreover, in 2018, Ukraine changed its position within the same cluster, moving closer to the group of more developed countries.