Türkiye’de enflasyon direngenliğinin bai-perron yöntemi ile incelenmesi

Bu çalışma son 30 yılda para politikası rejimlerinde önemli değişimler geçiren Türkiye'de fiyatlar genel seviyesinde meydana gelen yapısal değişimleri göz önüne alarak enflasyon direngenliği incelemek ve para politikası rejimlerindeki değişikliklerin bu süreçte etkili olup olmadığını ortaya koymak amacıyla gerçekleştirilmiştir. Bunun için enflasyon serisinde oluşan yapısal kırılmaların rast geldiği tarihleri tarafsız bir gözle ve bilimsel gerekçelerle saptamak amacıyla değişim tarihlerinin içsel olarak belirlenmesinin, bu tarihlerdeki iktisadi olayları tarafsız bir gözle analiz edebilmek imkanı vereceği düşünülmektedir. Yabancı literatürde "Persistency" karşılığı olarak kullanılan enflasyon direngenliği uygulamada çeşitli ekonometrik yöntemlerle tahmin edilebilmektedir. Çalışmada enflasyon serisine gelen şokların serinin ortalamasında yarattığı etkilerin kalıcı olup olmadığı, bu seriyi oluşturan oranların tarihi davranış kalıplarından çıkarılması amaçlandığı için tek değişkenli zaman serisi yöntemi kullanılmıştır. Bu amaçla, 1982–2010 yıllarına ait üçer aylık TÜFE endekslerinden hesaplanan enflasyon oranı serisinde meydana gelen yapısal değişim dönemleri ve bu süreç içerisinde enflasyon direncindeki değişmeler Bai ve Perron (1998, 2003a) tarafından önerilen çoklu yapısal değişim modeli, ardışık seçim yöntemi çerçevesinde analiz edilmiştir. Analiz sonuçlarında elde edilen katsayılar topluca değerlendirildiğinde örtük enflasyon hedeflenmesine geçilen 2003 yılından itibaren, dış konjonktürün de etkisi ile enflasyon kontrol altına alınmaya başlandığı, ayrıca bu dönemde enflasyon direngenliğinde de değişme olduğu saptanmıştır. Ancak enflasyonun istenilen düzeylere indirilebilmesi ve bunda kalıcılığın sağlanabilmesi için enflasyon direngenliğindeki azalmanın bir süre daha devam etmesi gerektiği sonucuna varılmıştır.

Çoklu Yapısal Değişim Modelleri ,

Enflasyon Oranı ,

Direngenlik ,

Multiple Structural Change Models ,

Inflation Rate ,

In this study, we aim to investigate inflation persistency in Turkey, which has experienced significant changes in monetary policy regime in the last 30 years, and to find out if the policies that were implemented have been effective in this process or not. Inflation persistency has aroused as an issue due to the difficulties in reducing the inflation rate, which seems to be taken under control since 2003, to a desired level such as below % 4-5, To reduce the inflation rate to a desired level and to keep it stable there by applying appropriate economic policies depends on how low the inflation persistency is. Therefore, it is important to know the structure and the level of the persistency of inflation rate and also it necessitates investigating inflation persistency according to econometrics. With this aim I conducted a study which has four parts. First part has the summary of economic policies and price movements in the time period that was investigated to have a better understanding of the structure of the inflation persistency. Second part is the summary of literature review. Third part shows the method, data collection and analysis. Conclusion section has the data evaluation. We can define inflation persistence as inflation's tendency to turn its average long term level slowly. When inflation has a high persistency, central bank needs to apply more effective policies than the time periods when it had low persistency in order to get inflation closer to the targeted level. Inflation persistency can be estimated by using a variety of econometric methods. If these estimates are done by considering the structural breaks in the sequence, then we could get more reliable results. Therefore, our study has been conducted by considering structural breaks. However, the dates of structural breaks can be both taken exogenously and endogenously from the data. In this study, in order to determine structural breaks points more objectively, they were taken as endogenously. In this study, we used univariate time series method because we aimed to understand if the shocks that were effecting inflation series were permanent or not by examining the form behavior of the series. Therefore, we adopted Bai-Perron method which is used when there are structural breaks in the inflation series because the test makes it possible to have more than one break and it accepts break points endogenously so it is different from other methods. In our study, we tried to determine the structural change periods of the inflation rate in Turkey by using quarterly data from 1982 to 2010. According to the suggested method, during the analysis phase we firstly need to determine the structural change number in the series. Bai-Perron suggest to use SupFT(l) and SupFT(l|l+1) test statistics to determine structural change number during the analysis, minimum observation number between the estimated structural change points was chosen as h=11 when maximum breaking number is 8 and ε=0.10. All methods are implemented in the GAUSS program. As a result, sequential test results have shown that there are 4 structural change period in inflation rate series in Turkey. These periods were estimated by minimizing the sum of squared residuals of general model. Also, confidence interval is calculated for these estimated periods and they are all presented in tables. Considering the time periods that was analyzed, it has been seen that there were structural changes in the average of inflation rate series in Turkey during 1987:Q3, 1993:Q4, 1998:Q1 and 2001:Q4. According to these structural change dates, the time between the 1982 and third quarter of 1987 is called as first period; second period is the time between the last quarters of 1987 and 1993; third period is the time from the first quarter of 1994 to the first quarter of 1998; fourth period is the time from the second quarter of 1998 till the fourth quarter of 2001; and finally the time after 2002 is called as fifth period. The data that have been found is parallel with the internal and external cyclical economic occasions. When analysis results were evaluated according to this, it was seen that the economic policies that were applied to reduce the inflation until 2004 have not been permanently effective. 1, j ˆ and 2, j ˆ (j=1,…,5) coefficients show average of inflation rate and the value of AR coefficient in every sub period respectively. On the other hand, these 2. j ˆ coefficients which show the value of AR coefficients in the estimated model are actually the coefficients which show the change in the inflation persistence in terms of periods. These coefficients were evaluated in terms of periods and as a result of structural tests, they were found statistically meaningful only after the fourth structural change period. This proves that compared to the previous period, there was a considerable change in the inflation persistency in the last period. When the coefficients that were obtained from the results of the analysis were considered as a whole, it was seen that with the effect from external conjecture inflation has started to be under control in this period when we transferred to implicit inflation targeting and also, in this period inflation persistency has started to break. However, in order to reduce the inflation to a desired level and to keep it stable there, inflation persistency should continue to decrease for a while.

Persistency ,