of  

or
Sign in to continue reading...

The Role of Expectations in the Inflation Process in Turkey: Have the Dynamics Changed Recently?

Umit Koc
By Umit Koc
3 years ago
The Role of Expectations in the Inflation Process in Turkey: Have the Dynamics Changed Recently?


Create FREE account or Login to add your comment
Comments (0)


Transcription

  1. The Role of Expectations in the Inflation Process in Turkey : Have the Dynamics Changed Recently? Ümit Koç Fethi Öğünç Mustafa Utku Özmen February 2021 Working Paper No: 21/02
  2. © Central Bank of the Republic of Turkey 2021 Address: Central Bank of the Republic of Turkey Head Office Structural Economic Research Department Hacı Bayram Mah. İstiklal Caddesi No: 10 Ulus, 06050 Ankara, Turkey Phone: +90 312 507 80 04 Facsimile: +90 312 507 78 96 The views expressed in this working paper are those of the author(s) and do not necessarily represent the official views of the Central Bank of the Republic of Turkey.
  3. The role of expectations in the inflation process in Turkey : Have the dynamics changed recently? Ümit Koç a, Fethi Öğünç b, and Mustafa Utku Özmen c ǂ Abstract In this paper, we analyze the role of inflation expectations in inflation dynamics. The hike in inflation following the exchange rate shock in 2018 provides an interesting period to analyze whether the sensitivity of inflation to its main determinants, including expectations, has changed. To this end, we estimate a time-varying parameter Phillips curve model to focus on the changes in inflation dynamics. We also jointly study the formation of inflation expectations to further investigate how the setting of inflation expectations evolved over the course of the rapid rise and the following gradual decline in inflation observed since the second half of 2018. Our results reveal that inflation expectations play an important role in inflation dynamics; and that the sensitivity of inflation to expectations did not change much recently. Meanwhile, the sensitivity of inflation to the exchange rate has sharply risen and corrected only partially afterwards. However, the most notable change has been witnessed in the weight attached to the past inflation in forming expectations; agents pay higher attention to inflation realizations. Overall, our results reveal that inflation expectations and the exchange rate movements are the leading driving forces of inflation in Turkey, in which the interaction between them further amplifies the impact on inflation. Keywords: Inflation, survey-based inflation expectations, state-space model, Turkey JEL codes: E31, C32, C36 a Umit.Koc@tcmb.gov.tr, Central Bank of the Republic of Turkey, Information Technologies Department; Fethi.Ogunc@tcmb.gov.tr, Central Bank of the Republic of Turkey, Structural Economic Research Department; c Utku.Ozmen@tcmb.gov.tr, Central Bank of the Republic of Turkey, Research and Monetary Policy Department. ǂ Authors thank Çağrı Sarıkaya, Eda Gülşen and the anonymous referee for constructive comments and recommendations which helped improve the final version of the paper. The views and opinions presented in this paper are those of the authors and do not necessarily represent those of the Central Bank of the Republic of Turkey or its staff. b 1
  4. Non-technical summary In this study we analyze the role of inflation expectations in the inflation process . Although the medium-term inflation expectations hover above the medium-term inflation target in Turkey, changes in inflation expectations transmit into changes in inflation. The changes in inflation dynamics can be better understood if the role of inflation expectations and the formation of these expectations are well comprehended. Following the exchange rate shock in 2018, inflation and inflation expectations increased noticeably. Analyzing the dynamics of inflation and inflation expectations in Turkey during a period covering the exchange rate shock, our results suggest that indeed inflation expectations have a sizeable and robust association with inflation, and that the sensitivity of inflation to expectations did not change much recently. We also find that the exchange rate pass-through has sharply risen and, despite gradually decelerating afterwards, is still above the historical averages. On the other hand, we document a slowdown in the extent of the import price pass-through. Meanwhile, the sensitivity of inflation to the output gap and real unit labor costs are found to be relatively stable. Regarding the formation of inflation expectations, the most notable change has been witnessed in the weight attached to the past inflation as agents pay higher attention to inflation realizations. Overall, our results reveal that inflation expectations and the exchange rate movements are the leading driving forces of inflation in Turkey, in which the interaction between them further amplifies the impact on inflation. 2
  5. 1 . Introduction Following the exchange rate shock in 2018, inflation recorded a substantial increase in Turkey. Meanwhile, inflation expectations not only increased but also the dispersion of expectations around the mean expectation deteriorated considerably. Accordingly, 12-month ahead inflation expectations stood at 16.5 percent at the end of 2018 with a sizeable variation around the mean expectation (Figure 1).1 Given the interplay between inflation and inflation expectations, a detailed analysis of the role of expectations in inflation dynamics is deemed necessary. Questions regarding the drivers of the rise in inflation expectations and the possibility of a change in the behavioral role of expectations in inflation dynamics, i.e. whether the economic agents put more weight on expectations, stand out. Figure 1: Inflation, Expectations and Dispersion of Expectations a) Inflation, Expectations and the Inflation Target b) Dispersion of Inflation Expectations (12-month ahead) 30 Inflation Expectation (12-month ahead) Consumer Inflation Target 25 20 15 10 5 12.19 03.19 06.18 09.17 12.16 03.16 06.15 09.14 12.13 03.13 06.12 09.11 12.10 03.10 06.09 09.08 12.07 0 Source: CBRT. Note: In the right panel, each blue dot represents an individual survey participant’s inflation expectation, whereas the orange dot refers to the mean expectation at that period. The workhorse theoretical framework for investigating above mentioned questions is the New Keynesian Phillips Curve.2 The New Keynesian economists reconstructed the Phillips curve relation based on micro-foundations and forward-looking expectations since the expectations of economic agents are deemed to have an important effect on price setting. Fischer (1977) and Taylor (1979) were the first to work on the New Keynesian Phillips Curve (NKPC), while Gali and Gertler (1999) developed the hybrid version of the NKPC that relates inflation to future expected inflation, lagged inflation and the real marginal costs. In this setup, the monetary policy affects inflation also by 1 Note that actual inflation was 11.8 percent at the end of 2019. Phillips curve has been on the agenda of the economy for more than 60 years. Literature started with Phillips (1958), who investigated the relationship between unemployment and nominal wages; followed by Samuelson and Solow (1960), who characterized the negative relationship between inflation and unemployment. During this period, the Phillips curve has become one of the main arguments for conducting monetary policy. Phelps (1967) and Friedman (1968) also included expectations into the Phillips curve, where the expectations of economic agents were shaped according to the adaptive expectations hypothesis. 2 3
  6. managing inflation expectations . Mishkin (2010) and Woodford (2003) suggest that the management of expectations and efforts of policymakers in this endeavor have become one of the pillars of the monetary policy setting, which has a catalyst role in achieving both economic growth and macroeconomic stability. Mishkin (2007) further argues that to evaluate changing inflation dynamics, it is vital to examine the possible changes in the formation of expectations as well. Recent studies enhance the open economy NKPC framework for investigating the inflation process with factors such as globalization, openness, exchange rates, import price index and oil prices (i.e. Auer, Borio and Filardo, 2017; Bianchi and Civelli, 2015; Blanchard, Cerutti and Summers 2015; Borio and Filardo, 2007; Ciccarelli and Mojon, 2010; Jorda and Nechio, 2018; Kamber, Mohanty and Morley, 2020). It is widely argued that in the last 20 years, the Phillips curve has become flatter (Blanchard, 2016; Mavroeidis, Plagborg-Moller and Stock, 2014; Mishkin, 2007; Pfajfar and Roberts, 2018) especially in developed countries, where low inflation have been observed and recently inflation mostly remained below the inflation targets.3 The response of wages and prices to economic activity continued to weaken. Despite sustainable economic growth, low unemployment, and historically low policy rates, inflation is persistently below the targets. The flattening of the Phillips curve is addressed to two issues: to the success of central banks in controlling inflation, and to structural changes in the economy.4 The success of central banks has led to the anchoring of inflation expectations to the inflation targets. Meanwhile, the level of anchoring of inflation expectations has also changed after the global financial crisis. Mazumder (2018) argues that while short and medium-term inflation expectations were tightly anchored until the end of 2008 in Euro area, this anchor has started to weaken since the global crisis, where the decline of inflation in the Euro-zone has been very sharp and inflation is persistently below the ECB’s targets. Mazumder (2018) shows that short-term (one-year ahead) inflation expectations have become more significant than longer-term expectations in Europe. Similarly, Fuhrer (2012) finds that the impacts of short term expectations (four-quarter ahead) have become more relevant and more visible for US inflation. Guinigundo (2016) and Galati, Poelhekke 3 Pfajfar and Roberts (2018) suggest two reasons for flattening. The first one is that the prices are stickier at the microlevel, and thus they are adjusted less often. Second, the expectations of households and firms regarding future inflation do not include the information set about macroeconomic conditions. As expectations are important to determine current inflation, the sensitivity of inflation to macroeconomic conditions has decreased. 4 Despite many studies disputing the validity of the Phillips curve, Blanchard (2016) argues that the relationship is still working. It is found out that inflation expectations have become more anchored, which leads to a relationship between unemployment and level inflation rather than the change in inflation. As a general view, the Phillips curve has become closer to its shape in the 1960s. Another finding is about the slope of the Phillips curve, which declined significantly. But the decline dates back to the 1980s, not the recent global crisis. 4
  7. and Zhou (2011) also point to less anchoring of inflation expectations during and after the global financial crisis. Expectations play a very important role in the inflation process. If well-anchored, inflation expectations can help stabilize inflation faster in case of supply and demand shocks. In Turkey, the medium-term inflation expectations hover above the medium-term inflation target. Nonetheless, changes in inflation expectations transmit into changes in inflation. The changes in inflation dynamics can be better understood if the role of inflation expectations and the formation of these expectations are well comprehended. In this perspective, we analyze the dynamics of inflation expectations as well as their impact on inflation in Turkey. Accordingly, in this paper, we estimate reduced-form Phillips curve models for consumer inflation based on the study of Kara, Öğünç and Sarıkaya (2017) to understand the role of expectations in inflation dynamics. We analyze both the fixed-parameter and time-varying parameter (TVP) models and extend their model in two dimensions. First, we utilize the direct survey measure of inflation expectations, similar to Roberts (1995), Adam and Padula (2011), Fuhrer (2012), Blanchard, Cerutti and Summers (2015) and Mazumder (2018), using the CBRT’s Survey of Expectations data. In this regard, we allow for nonrational expectations. Second, we model the formation of survey expectations at a macro level based on the micro-level findings of Gülşen and Kara (2020) in a time-varying parameter setting.5 Accordingly, the novelty of our approach is that we estimate the evolution of inflation and inflation expectations simultaneously in a time-varying framework for the Turkish case. Our results suggest that indeed inflation expectations have a sizeable and robust relationship with inflation with a fairly stable coefficient; meanwhile some of the coefficients of other determinants of inflation are subject to considerable fluctuation. Our results also reveal that the sensitivity of inflation expectations to the exchange rate and inflation surprise grew stronger in 2018. More importantly, the sensitivity of expectations to past inflation substantially increased. In this respect, as the decline in headline inflation became evident throughout 2019, the improvement in inflation expectations in this period also accelerated. The remainder of the paper is organized as follows. Section 2 examines the role of expectations in inflation dynamics with fixed-parameter Phillips curve models and attempts to model the survey expectation formation as well. Section 3 extends the models in the previous section to the timevarying parameter case and discusses main empirical findings. Section 4 concludes the study. 5 Note also that one of the strengths of our study is using data from an inflationary environment, which allows us to identify the changing relationships even with relatively small sample size. 5
  8. 2. The Role of Expectations in the Inflation Dynamics In this section, we investigate the role of inflation expectations in explaining inflation dynamics through the lens of fixed-parameter Phillips curve models. To this end, we extend the previously estimated Phillips curve models for Turkey (such as Kara, Öğünç and Sarıkaya (2017), Koca and Yılmaz (2018)) by including the survey inflation expectations into the model. In the first part, we estimate the following specification for inflation: