Türkiye'deki Enflasyonun Bayesci Vektör Otoregresyon Modeller ile İncelenmesi

Bu çalışmada, zaman serileri analizinde yaygın biçimde kullanılan Bayesci Vektör Oto-regresyon (BVAR) modeller ayrıntılı bir şekilde açıklanmıştır. Bu çerçeve içinde Vektör Oto-regresyon (VAR) modeller de incelenmiştir. Bilindiği gibi Türkiye, çok uzun zamandan beri yüksek enflasyon oranlarının yol açtığı sıkıntılarla uğraşmaktadır. Öyle ki yüksek enflasyon oranları Türkiye’nin en önemli sorunlarından biri haline gelmiştir. Bu faktörün önemi, bizi onu modellemeye yönlendirmiştir. Modelleme için iki ayrı dönem, Ocak 1986 – Aralık 2001 ve Ocak 1986 – Aralık 2000 seçilmiştir. Bu dönemler için yedi farklı BVAR modeli oluşturulmuş ve daha sonra 2002 ve 2001 yılları için bu modellerin öngörü performansları VAR modeller ile karşılaştırılmıştır. Sonuçlar incelendiğinde, BVAR modellerinin Ocak 2002 – Aralık 2002 döneminin öngörüsünde VAR modeline oranla iyi bir performans sergileyemediği anlaşılmıştır. Bu duruma, 2001 yılında yaşanmış olan ekonomik krizin yol açtığı düşünüldüğünden, Ocak 1986 – Aralık 2000 dönemi için ayrı bir modellemeye gidilmiş ve Ocak 2001 – Aralık 2001 öngörülerine bakılmıştır. Çıkan sonuçlar BVAR modellerinin, VAR modeline göre 2001 yılı gerçek değerlerin tahmininde çok daha başarılı olduğunu kanıtlamıştır.

Analysing the Inflation in Turkey with Bayesian Vector Autoregression Models

In this study, Bayesian Vector Auto-regression (BVAR) models, which are widely used in time series analysis, are explained in details. In this perspective, Vector Auto-regression (VAR) models are also examined. As is well-known, Turkey suffers from high inflation rates for a long time. In other words, high inflation rates have become one of the most important problems of Turkey. The significance of this factor has led us to model of itself. To model, two different periods, January 1986 – December 2001 and January 1986 – December 2000 are selected For these periods different seven BVAR models are constructed, and then the forecast performances of these, for years 2002 and 2001 are compared with VAR model. When the results are examined, it is found that the forecast performances of BVAR models aren’t better than VAR models for January 2002 – December 2002 period. The reason for this can be explained by the economic crisis happened at 2001. With this respect, another model is constructed for January 1986 - December 2000 period and the January 2001 - December 2001 forecasts are examined. Finally the results showed that, BVAR models are much better than VAR models for estimating the real values of 2001.

Kaynakça

[1]Todd, R., (1984). Improving Economic Forecasting with Bayesian Vector Autoregression. Quarterly Review, Federal Reserve Bank of Minneapolis, 8(Fall), 18-29.

[2]Litterman, R. (1986). A Statistical Approach to Economic Forecasting. Journal of Business and Economic Statistics, 4(1), 1-4.

[3]Litterman, R. (1986). Forecasting with Bayesian Vector Autoregression – Five Years of Experience. Journal of Business and Economic Statistics, 4(1), 25-38.

[4]Spencer, D. (1993). Developing a Bayesian Vector Autoregression Forecasting Model. International Journal of Forecasting, 9(3), 407-421.

[5]Bessler, D.A. & Kling, J.L. (1986). Forecasting Vector Autoregressions with Bayesian Priors. American Journal of Agricultural Economics, 68(1), 144-151.

[6]Nelson, C.R. & Plosser, C. (1982). Trends and Random Walks in Macroeconomic Time Series. Journal of Monetary Economics, 10(2), 139-162.

[7]Tiao, G.C. & Box, G.E.P. (1981). Modelling Multiple Time Series with Application. Journal of The American Statistical Association, 76(376), 802-816.

[8]Tiao, G. & Tsay, R. (1983). Multiple Time Series Modeling and Extended Sample Cross-Correlations. Journal of Business and Economic Statistics, 1(1), 43-56.

[9]Engle, R.F. & Granger, C.W.J. (1987). Cointegration and Error Correction: Representation, Estimation and Testing. Econometrica, 55(2), 251-276.

[10]Engle, R.F. & You, B.S. (1987). Forecasting and Testing in Cointegrated Systems. Journal of Econometrics, 35(1), 143-159.

[11]Fırat, B.E. (1994). Estimating and Forecasting Exchange Rate: Comparison of Structural and VAR Models. A Master’s Thesis, Middle East Technical University. Master of Science in Economics, Ankara.

[12]Doan, T.; Litterman, R.B. & Sims, C. (1983). Forecasting and Conditional projection Using Realistic Prior Disributions. NBER Working Papers Series. No:1202, Cambridge.

[13]Larson, H.J. (1982). Introduction to Probability Theory and Statistics Inference. Third Edition. New York: John Wiley and Sons.

[14]Hsu, P.; Wang, C.; Shyu, J.Z. & Yu, H. (2003). A Litterman BVAR Approach for Production Forecasting of Technology Industries. Technological Forecasting & Social Change, 70(1), 67-82.

[15]Ramos, F. (2003). Forecasts of Market Shares from VAR and BVAR Models: A Comparison of Their Accuracy. International Journal of Forecasting, 19(1), 95-110.

[16]Doan, T. (1992). RATS. Users Manual Version 4. Illinois: Estima.

[17]Bischoff, C.W.; Belay, H. & Kang, I. (2000). Bayesian VAR Forecasts Fail To Live Up to Their Promise. Business Economics, 35(July), 19-29.

[18]Wi, S. (1999). An Evaluation Of Combining Forecasts and a Strategy For Searching For An Optimum Bayesian Var Prior To Forecast Business Cycle Turning Points. Dissertation of Doctor of Philosophy in Economics. State University of New York, Graduate School of Binghamton University, New York.

[19](www.tcmb.gov.tr). [10.05.2001].

[20]Joutz, F.L.; Maddala, G.S. & Trost, R.P. (1995). An Integrated Bayesian Vector Autoregression and Error Correction Model for Forecasting Electricity Consumption and Prices. Journal of Forecasting, 14(3), 287-310.

[21]Lutkepohl, H. (1993). Introduction to Multiple Time Series Analysis. Second Edition. Berlin: Springer-Verlag.

Kaynak Göster