OECD ÜLKELERİNDE PANDEMİ SONRASI İŞSİZLİĞİN K-MEANS YÖNTEMİYLE KÜMELENMESİ (2021–2023)

Abstract

Bu çalışma, k-means kümeleme yöntemini uygulayarak 2021, 2022 ve 2023 yıllarında OECD ülkelerinde işsizlik dinamiklerini incelemektedir. Yıllık işsizlik oranı verileri, Dünya Bankası'nın Dünya Kalkınma Göstergeleri veritabanından alınmış ve analizden önce standartlaştırılmıştır. Kümelerin sayısı, siluet puanı, dirsek yöntemi ve boşluk istatistiği gibi çeşitli teşhis araçları kullanılarak belirlenmiştir. Bu sonuçlara dayanarak, ülkeler işsizlik oranlarındaki benzerliklere göre gruplandırılmış ve böylece hem kesitsel hem de zamansal karşılaştırmalar yapılabilmiştir. Analiz, her yıl ekonomik gelişmişlik düzeyleri, işgücü piyasası yapıları ve kurumsal kapasitelerdeki farklılıkları genel olarak yansıtan üç farklı küme belirlemektedir. İsviçre, Japonya, Almanya ve Norveç gibi gelişmiş ekonomiler, çeşitlendirilmiş ekonomik temeller, güçlü mesleki eğitim sistemleri ve etkili aktif işgücü piyasası politikaları sayesinde sürekli olarak düşük işsizlik grubunda yer almaktadır. Buna karşılık, Yunanistan, İspanya ve Türkiye, köklü yapısal katılıklar, beceri uyumsuzlukları ve turizm ve diğer düşük katma değerli hizmetlere aşırı bağımlılık nedeniyle, dönem boyunca yüksek işsizlik grubunda kalmaktadır. Bazı ülkeler zaman içinde kümeler arasında geçiş yapmıştır, bu da pandemi sonrası toparlanma dinamiklerinin, enflasyonist baskıların ve sektöre özgü büyüme modellerinin rolünü vurgulamaktadır. Bulgular, işsizlik eşitsizliklerinin, hem uzun vadeli yapısal faktörlerin hem de kısa vadeli ekonomik dalgalanmaların etkisiyle OECD ülkeleri genelinde kalıcı bir özellik olmaya devam ettiğini göstermektedir. Bu çalışma, güncellenmiş bir pandemi sonrası perspektif sunarak mevcut literatüre katkıda bulunmakta ve politika yapımına ilişkin önemli bilgiler sağlamaktadır. İşsizliğin yüksek olduğu ülkelerde, çabalar ekonomik sektörlerin çeşitlendirilmesi, mesleki ve teknik eğitimin geliştirilmesi ve istihdam yaratılmasını kolaylaştırmak için işgücü piyasası kurumlarının güçlendirilmesine odaklanmalıdır. Tersine, işsizliğin sürekli düşük olduğu ülkeler, teknolojik değişim ve demografik dönüşümler karşısında işgücü piyasasının dayanıklılığını korumaya öncelik vermelidir. Bu sonuçlar, hızla gelişen küresel ekonomide işsizliği etkili bir şekilde ele alan, hedefli ve kanıta dayalı stratejiler uygulamayı amaçlayan politika yapıcılar için pratik bir rehber niteliğindedir.

Keywords

• OECD
• Pandemi Sonrası İşsizlik
• Kümelenme Analizi
• İşgücü Piyasası

CLUSTERING OF POST-PANDEMIC UNEMPLOYMENT IN OECD COUNTRIES USING THE K-MEANS METHOD (2021–2023) ABSTRACT

Abstract

This study examines unemployment dynamics across OECD countries during 2021, 2022, and 2023 by applying the k-means clustering method. Annual unemployment rate data were drawn from the World Bank’s World Development Indicators database and standardized before analysis. The number of clusters was identified using several diagnostic tools, including the silhouette score, the elbow method, and the gap statistic. Based on these results, countries were grouped according to similarities in unemployment rates, allowing for both cross-sectional and temporal comparisons. The analysis identifies three distinct clusters each year, which broadly mirror differences in economic development levels, labour market structures, and institutional capacities. Advanced economies such as Switzerland, Japan, Germany, and Norway consistently fall into the low-unemployment group, supported by diversified economic bases, strong vocational training systems, and effective active labour market policies. By contrast, Greece, Spain, and Türkiye remain in the high-unemployment group throughout the period, reflecting entrenched structural rigidities, skill mismatches, and a heavy reliance on tourism and other low value-added services. A subset of countries shifted between clusters over time, highlighting the role of post-pandemic recovery dynamics, inflationary pressures, and sector-specific growth patterns. The findings indicate that unemployment disparities remain a persistent feature across OECD countries, influenced by both long-term structural factors and short-term economic fluctuations. By offering an updated post-pandemic perspective, this study contributes to the existing literature and provides insights relevant for policymaking. In countries with high unemployment, efforts should focus on diversifying economic sectors, enhancing vocational and technical education, and strengthening labour market institutions to facilitate job creation. Conversely, countries with consistently low unemployment should prioritize maintaining labour market resilience in the face of technological change and demographic shifts. Taken together, these results offer practical guidance for policymakers aiming to implement targeted, evidence-based strategies that address unemployment effectively within a rapidly evolving global economy.

Keywords

• OECD
• Post-Pandemic Unemployment
• Clustering Analysis
• K-Means
• Labour Market

References

1. Ardiansyah, M.F.H., Amany, N., Anugrah, C.I., & Syafitri, U.D. (2024). K-Means Clustering Application of Open Unemployment in 2020 Caused by COVID-19 in West Java Province. ENTHUSIASTIC, International Journal of Applied Statistics and Data Science, 4(1), 1- 12.
2. Aretz, B., Frey, S., & Weltermann, B. (2024). Regional socioeconomic characteristics and density of general practitioners in Germany: A nationwide cross-sectional and longitudinal spatial analysis. Public Health, 236, 338–346. https://doi.org/10.1016/j.puhe.2024.09.010
3. Arulampalam, W. (2001). Is unemployment really scarring? Effects of unemployment experiences on wages. The Economic Journal, 111(475), F585–F606. https://doi.org/10.1111/1468-0297.00664
4. Bell, D. N. F., & Blanchflower, D. G. (2011). Young people and the Great Recession. Oxford Review of Economic Policy, 27(2), 241–267.
5. Blanchard, O., & Johnson, D. R. (2013). Macroeconomics (6th ed.). Pearson.
6. Blasques, F., Hoogerkamp, M.H., Koopman, S.J., & Werve, I. van de. (2021). Dynamic factor models with clustered loadings: Forecasting education flows using unemployment data. International Journal of Forecasting, 37, 1426–1441. https://doi.org/10.1016/j.ijforecast.2021.01.026
7. Demir, Y. (2021). Balkan Ülkelerine Ait İşsizliğin Mekânsal Panel Ekonometri Yaklaşımı İle Analizi. Balkan Sosyal Bilimler Dergisi, 10(19) 26–35.
8. Engeloğlu, Ö., & Yurdakul, F. (2025). The Factors Causing Consumer Behavior: Asymmetric Causality And Cluster Analysis For EU Countries Through Consumer Confidence Index. EGE AKADEMİK BAKIŞ/EGE ACADEMIC REVIEW, 25(1), 21-42. https://doi.org/10.21121/eab.20250102

9. Eygü, H. (2018). Enflasyon, İşsizlik Ve Dış Ticaret Arasındaki İlişkinin İncelenmesi: Türkiye Örneği (1990-2017). Kastamonu Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi, 20(2), 96-112. https://doi.org/iibfdkastamonu.408823
10. Hartigan, J. A., & Wong, M. A. (1979). A K-means clustering algorithm. Journal of the Royal Statistical Society: Series C (Applied Statistics), 28(1), 100–108. https://doi.org/10.2307/2346830
11. İşleyen, Ş. (2021). Clustering Analysis of Employment Sectors According To OECD Countries Using The K-Average Method. International Journal of Contemporary Economics and Administrative Sciences, 11(1), 093–105. https://doi.org/10.5281/zenodo.5136506
12. International Labour Organization (ILO). (2022). World employment and social outlook: Trends 2022. https://www.ilo.org/sites/default/files/wcmsp5/groups/public/@dgreports/@dcomm/@publ/documents/publication/wcms_834081.pdf
13. Kanberoglu, Z., Görür, Ç., Yıldırımçakar, İ., & Türkmenoğlu, M. (2021). Analysis of Covid-19 Pandemic Data According to the Countries Including in the Human Development Index by Using Clustering Analysis K-Average Method. International Journal of Contemporary Economics and Administrative Sciences, 11(2), 454–468. https://doi.org/10.5281/zenodo.5831774
14. Lietzmann, T., & Hohmeyer, K. (2024). Unemployed and then? The role of non-standard employment in labour market trajectories after unemployment. Int J Soc Welf, 34, e12698. https://doi.org/10.1111/ijsw.12698
15. MacQueen, J. (1967). Some methods for classification and analysis of multivariate observations. In Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, 1, 281–297. University of California Press.
16. Manea, L.D., & Sorcaru, I.A. (2021). Cluster Analysis on the Unemployment Rate during Lockdown Period. Annals of “Dunarea de Jos” University of Galati Fascicle I. Economics and Applied Informatics, 31-36. https://doi.org/10.35219/eai15840409220
17. Masserini , L., Bini, M., Zeli A., & Forciniti, A. (2024). Measuring the impact of the 2008 and 2011 financial crises and the 2015 recovery on the unemployment rate in Italy. Socio-Economic Planning Sciences, 95, 102032. https://doi.org/10.1016/j.seps.2024.102032
18. Mogos, I.R., Dinu, M., Constantinescu, V.G., & Istrate, B. (2022). Unemployment in European Union During the COVID-19 Pandemic. A Cluster Analysis. In: R. Pamfilie, V. Dinu, C. Vasiliu, D. Pleșea, L. Tăchiciu eds. 8th BASIQ International Conference on New Trends in Sustainable Business and Consumption. Graz, Austria, 25-27 May 2022. Bucharest: ASE, 117-124. https://doi.org/10.24818/BASIQ/2022/08/014
19. Monfort, M., Cuestas, J. C., & Ordóñez, J. (2018). Real convergence in Europe: A cluster analysis. Economic Modelling, 689-694. https://doi.org/10.1016/j.econmod.2013.05.015
20. Monfort, M., Ordóñez, J., & Sala, H. (2018). Inequality and Unemployment Patterns in Europe: Does Integration Lead to (Real) Convergence? Open Econ Rev, 29, 703–724. https://doi.org/10.1007/s11079-018-9488-x
21. Nichols, A., Mitchell, J., & Lindner, S. (2013). Consequences of long-term unemployment. Urban Institute. https://www.urban.org/sites/default/files/publication/23921/412887-Consequences-of-Long-Term-Unemployment.PDF
22. OECD (2023). OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market. OECD Publishing, Paris, https://doi.org/10.1787/08785bba-en
23. Ogryzek, M., & Jaskulski, M. (2025). Applying Methods of Exploratory Data Analysis and Methods of Modeling the Unemployment Rate in Spatial Terms in Poland. Appl. Sci., 15, 4136. https://doi.org/10.3390/app15084136
24. Ok, Y. (2022). Bulanık C - Ortalamalar İle Ülkelerin İşsizlik Göstergeleri Temelinde Kümelenmesi. Avrupa Bilim ve Teknoloji Dergisi, (34), 507-512. https://doi.org/10.31590/ejosat.1083246
25. Paul, K. I., & Moser, K. (2009). Unemployment impairs mental health: Meta-analyses. Journal of Vocational Behavior, 74(3), 264–282. https://doi.org/10.1016/j.jvb.2009.01.001
26. Poutanen, J., Gluschkoff , K., Kausto, J., & Joensuu, M. (2024). Main activity trajectory clusters of unemployed people with partial work ability and cluster features. Scandinavian Journal of Public Health, 52, 918–926. https://doi.org/10.1177/14034948231210347
27. Salimova, G., Ableeva, A., Gusmanov, R., Sharafutdinov, A., & Nigmatullina, G. (2024). Employment in the Digital Economy Development: Regional Clustering. Public Organization Review, 24, 141–160. https://doi.org/10.1007/s11115-023-00746-w
28. Seppälä, P., Zhu, N., Hietamäki, J., Häkkilä, L., Gawel, A., & Toikko, T. (2024). The threshold of child protection notifications is higher in municipalities with a high level of risk factors – Is this evidence of the inverse intervention law? Child Abuse & Neglect, 155,106963. https://doi.org/10.1016/j.chiabu.2024.106963
29. Tatarczak, A., & Boichuk, O. (2018). The multivariate techniques in evaluation of unemployment analysis of Polish regions. Oeconomia Copernicana, 9(3), 361–380. https://doi.org/10.24136/oc.2018.018 Tibshirani, R., Walther, G., & Hastie, T. (2001). Estimating the number of clusters in a data set via the gap statistic. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 63(2), 411–423. https://doi.org/10.1111/1467-9868.00293
30. Trentini, M. (2024). Labour market trajectories and unemployment of older workers in Europe after the Great Recession. Sociology Compass, 18(5), 1-16. https://doi.org/10.1111/soc4.13215
31. World Bank. (2024). World Development Indicators. Retrieved (Access Date:12/06/2025), from https://data.worldbank.org/indicator/SL.UEM.TOTL.ZS
32. Yılmaz, T. (2022) . OECD Ülkelerinin İş Gücü Endeksi Açısından Kümeleme Analizi ile İncelenmesi. Journal of International Management, Educational and Economics Perspectives, 10(1), 84–101.
33. Zaharia, M. (2024). The influence of pre-university education resources on school dropout and the unemployment rate in Romania is significant? Journal of Research and Innovation for Sustainable Society (JRISS), 6(2). https://doi.org/10.33727/JRISS.2024.2.22:202-214