CHAPTER


DOI :10.26650/B/SS10.2021.013.32   IUP :10.26650/B/SS10.2021.013.32    Full Text (PDF)

Investigation of Unemployment with Nonlinear Time Series in Turkey

Fatma İdil Baktemur

1. Introduction

It has long been debated that the majority of time series are nonlinear. Attempts have been made to explain cyclical fluctuations in economics using the “sunspots” approach. Neftçi (1984) and Hamilton (1989) mentioned an asymmetric structure in a time series, stating that linear models cannot be sufficient to define the processes of macroeconomic series such as real production and unemployment rate. In his study, Neftçi (1984) examined the unemployment rate in the US using the Markov approach. Asymmetric movements in the business cycle form the basis of nonlinear time series studies. Unemployment is an important economic problem, and the nonlinear nature of unemployment has been the subject of many studies. For example, Van Dijk, Teräsvirta, and Franses (2002) used a logistic smooth transition autoregressive (LSTAR) model to examine US unemployment.

2. Methodology

For the smooth transition autoregressive (STAR) model estimation, the appropriate delay length for the autoregressive model is determined according to the Akaike information criteria (AIC) and Schwarz information criteria (SIC). Subsequently, transition variables are formed, and the linearity of the model against its nonlinearity is tested by the Lagrange Multiplier (LM) tests. If the linearity hypothesis is rejected, the appropriate model (LSTAR or exponential smooth transition autoregressive [ESTAR]) is determined. Finally, the LSTAR/ ESTAR model is estimated, and the diagnostics are checked.

3. Results

In this study, the nonlinear structure of the unemployment rate in Turkey between 2005:01 and 2018:12 was analyzed using STAR models. First, the graph of the series was examined, and the asymmetry and seasonal effects of the unemployment series were mentioned. Further, the stationarity was tested with a nonlinear unit root test (Kapetanios, Shin, and Snell). Thus, the natural unemployment rate hypothesis and hysteresis effect were emphasized. In the natural unemployment rate hypothesis, temporary shocks have a temporary effect on the natural unemployment rate, whereas in the hysteresis effect, they have permanent effect. The series is nonstationary, and it emphasizes hysteresis effect. After determining the appropriate delay length for the autoregressive model according to the AIC and SIC, a transition variable was formed, and the linearity of the model against its nonlinearity was tested using the LM tests. In the formation of transition variables, a 12-month difference (Δ12 yt-d) was applied based on Skalin and Teräsvirta’s (2001) study. After rejecting the linearity hypothesis, the choice between LSTAR and ESTAR models was made. The results show that the ESTAR model type is appropriate. In the estimation results of the ESTAR model, the threshold parameter and the parameter indicating the speed of transition are statistically significant. Transition from one regime to another is smooth. Autocorrelation and heteroscedasticity, normality of residuals, parameter constancy, and remaining nonlinearity tests have also been reported in the model.

4. Conclusion

In this study, the nonlinearity of unemployment for Turkey was analyzed using the 2005:01–2018:12 period. The series is nonstationary according to nonlinear unit root test. Thus, hysteresis effect is valid. Therefore, temporary shocks have permanent effects on unemployment. Linearity hypothesis was rejected, and the appropriate model was ESTAR. The transition from one regime to another is smooth.


DOI :10.26650/B/SS10.2021.013.32   IUP :10.26650/B/SS10.2021.013.32    Full Text (PDF)

Türkiye İçin İşsizliğin Doğrusal Olmayan Zaman Serileri ile İncelenmesi

Fatma İdil Baktemur

Zaman serilerinin büyük çoğunluğunun doğrusal olmadığı uzun süreden beri tartışılmaktadır. İktisatta konjonktürel dalgalanmalar “güneş lekeleri” olarak bilinen yaklaşım ile açıklanmaya çalışılmıştır. Zaman serisinde asimetrik yapıdan Neftçi (1984) ve Hamilton (1989) çalışmalarında bahsetmişlerdir. Reel üretim ve işsizlik oranı gibi makroekonomik serilerin süreçlerini tanımlamada doğrusal modellerin yeterli olamayacaklarını ifade etmişlerdir. Neftçi (1984) çalışmasında ABD için işsizlik oranını Markov yaklaşımı ile incelemiştir. Görüldüğü gibi iş çevrimi döngülerindeki asimetrik hareketler doğrusal olmayan zaman serileri çalışmalarının temelini oluşturmaktadır. İşsizlik önemli bir ekonomik problemdir ve işsizliğin doğrusal olmayan yapısı da literatürde pek çok çalışmaya konu olmuştur. STAR modelleri LSTAR ve ESTAR model tipleri ile asimetrik davranışları açıklayabilecek esnek bir yapıdadır. Bu çalışmada da 2005:01-2018:12 dönemi Türkiye için işsizlik oranlarının doğrusal olmayan yapısı STAR modelleri ile incelenmiştir. Serinin ilk olarak grafiği incelenerek işsizlik serisindeki asimetri ve mevsimsel etkilere değinilmiş, daha sonra doğrusal olmayan birim kök testi (KSS) ile durağanlığı sınanmıştır. Böylece doğal işsizlik oranı hipotezi ve histeri etkisi üzerinde de durulmuştur. Doğal işsizlik oranı hipotezinde geçici şokların doğal işsizlik oranında geçici etkisi varken; histeri etkisinde kalıcı etkisi vardır. Otoregresif model için uygun gecikme uzunluğunun AIC ve SIC bilgi kriterlerine göre belirlenmesinden sonra geçiş değişkeni oluşturulmuş ve modelin doğrusallığı doğrusal olmamasına karşı LM testleri ile test edilmiştir. Geçiş değişkeninin oluşumunda Skalin ve Teräsvirta (2001) çalışmasından hareket ederek on iki aylık fark (Δ12 yt-d) uygulanmıştır. Doğrusallık hipotezi reddedildikten sonra LSTAR ve ESTAR modelleri arasında seçim yapılmıştır. Uygun model tipi ESTAR olarak bulunmuştur ve bir rejimden diğer bir rejime geçiş yumuşak bir şekilde gerçekleşmektedir. Modelde otokorelasyon ve değişen varyans, parametre sabitliği ve kalan doğrusal olmama testleri de ayrıca rapor edilmiştir.



References

  • Acemoglu, D., & Scott, A. (1994). Asymmetries in the cyclical behaviour of UK labour markets. Economic Journal, 104(427), 1303-1323. google scholar
  • Ağazade, S. (2014). Türkiye’de İşsizlik ve İşgücüne Katılım İlişkisinin Doğrusal ve Doğrusal Olmayan Koentegrasyon Yöntemleri ile Analizi. Eskişehir Osmangazi Üniversitesi İİBF Dergisi, 9(3), 145-161. google scholar
  • Akram, Q. F. (2005). Multiple unemployment equilibria and asymmetric dynamics—Norwegian evidence, Structural Change and Economic Dynamics, 16(2), 263-283. google scholar
  • Bayat, T., Kayhan, S., & Koçyiğit, A. (2013). Türkiye’de İşsizliğin asimetrik davranişinin rejim değişim modeliyle incelenmesi. Business and Economics Research Journal, 4(2), 1-15. google scholar
  • Bozkurt, E., & Altıner, A. (2018). Doğrusal ve Doğrusal Olmayan Birim Kök Testleriyle Türkiye’de İşsizlik Histerisinin Tespiti. Uluslararası İktisadi ve İdari İncelemeler Dergisi, UİİİD-IJEAS, 2018 (Prof. Dr. Harun Terzi Özel Sayısı):167-180. google scholar
  • Camarero, M., & Ordonez, J. (2012). Nonlinear adjustment in the real dollar-euro exchange rate: The role of the productivity differential as a fundamental. Economic Modelling, 29, 444-449. google scholar
  • Cengiz, S., & Şahin, A. (2014). Modelling Nonlinear Behavior of Labor Force Participation Rate by STAR: An Application for Turkey. International Journal of Economic Sciences and Applied Research, 7(1), 113-127. google scholar
  • Deschamps, P. J. (2008). Comparing smooth transition and Markov switching autoregressive models of US unemployment. Journal of Applied Econometrics, 23, 435-462. google scholar
  • Dikmen, F. C., & Dursun, G. (2018). Unemployment Hysteresis and Convergence: Evidence from Latin American Countries. EconWorld2018@Lisbon 23-25 January, 2018; Lisbon, Portugal. google scholar
  • Dornbusch, R., & Fischer, S. (1994). Macro economics (Sixth Ed.). McGraw-Hill, Inc. google scholar
  • Eitrheim, 0., & Terasvirta, T. (1996). Testing the adequacy of smooth transition autoregressive models. Journal of Econometrics, 74, 59-75. google scholar
  • Granger, C. W. J., & Terasvirta, T. (1993). Modelling nonlinear economic relationships. Oxford: Oxford University Press. google scholar
  • Güriş, B., Tiftikçigil, B. Y., & Tıraşoğlu, M. (2017). Testing for unemployment hysteresis in Turkey: evidence from nonlinear unit root tests. Quality & Quantity, 51(1), 35-46. google scholar
  • Hamilton, J. D. (1987). A new approach to the economic analysis of nonstationary time series and the business cycles. Econometrica, 57(2), 357-384. google scholar
  • Kapetanios, G., Shin, Y., & Snell, A. (2003). Testing for a Unit Root in the Nonlinear STAR Framework. Journal of Econometrics, 112, 359-379. google scholar
  • Luukkonen, R., Saikkonen, P. & Terasvirta, T. (1988). Testing linearity against smooth transition autoregressive models. Biometrika, 75, 491-499. google scholar
  • Mishkin, F. (2000). The Economics of Money banking and financial markets. google scholar
  • Neftçi, S. N. (1984). Are economic time series asymmetric over the business cycles. Journal of Political Economy, 92, 307-328. google scholar
  • Papell, D. H., Murray, C. J., & Ghiblawi, H. (2000). The structure of unemployment. The Review of Economics and Statistics, 82, 309-315. google scholar
  • Peel, D. A., & Speight, A. E. H. (1998). The nonlinear time series properties of unemployment rates: some further evidence. Applied Economics, 30, 287-294. google scholar
  • Saraç, T. B. (2014). İşsizlikte Histeri Etkisi: Türkiye Örneği, Ege Akademik Bakış, 14(3), 335-344. google scholar
  • Skalin, J., & Terasvirta, T., (1999). Another look at Swedish business cycles, 1861-1988. Journal of Applied Econometrics, 14, 359-378. google scholar
  • Skalin, J., & Terasvirta, T. (2002). Modeling asymmetries and moving equilibria in unemployment rates. Macroeconomic Dynamics, 6, 202-241. google scholar
  • Terasvirta, T. (1994). Specification, estimation and evaluation of smooth transition autoregressive models. Journal of the American Statistical Association, 89, 208-218. google scholar
  • Van Dijk, D., Franses, P. H., & Paap, R. (2002). A non-linear long memory model for the US unemployment. Journal of Econometrics, 102, 135-165. google scholar
  • Van Dijk, D., Terasvirta, T., & Franses, P. H. (2002). Smooth transition autoregressive models—a survey of recent developments. Econometric Reviews, 21, 1-47. google scholar
  • www.tcmb.gov.tr google scholar


SHARE




Istanbul University Press aims to contribute to the dissemination of ever growing scientific knowledge through publication of high quality scientific journals and books in accordance with the international publishing standards and ethics. Istanbul University Press follows an open access, non-commercial, scholarly publishing.