• Sonuç bulunamadı

An open economy new Keynesian macroeconomic model: The case of Turkey

N/A
N/A
Protected

Academic year: 2021

Share "An open economy new Keynesian macroeconomic model: The case of Turkey"

Copied!
20
0
0

Yükleniyor.... (view fulltext now)

Tam metin

(1)İktisat İşletme ve Finans 26 (305) 2011 : 37-56 www.iif.com.tr doi: 10.3848/iif.2011.305.2978. l. An open economy new Keynesian macroeconomic model: The case of Turkey†. 12. ih:. ar. ], T. .44. .54. ge. /06. /20. 20. 09 :4. 7:0 5. Abstract . A new consensus in macroeconomics called the New Neo-Classical Synthesis or New Keynesian Macroeconomic Model emerged at the end of the 1990s. The main characteristics of this consensus are formed by the synthesis of the New Classical, Real Business Cycle and New Keynesian approaches. Although The New Keynesian Macroeconomic Model is based on a general equilibrium model, it can typically be reduced to a three-equation system, consisting of an aggregate supply equation (Phillips curve), an aggregate demand equation (IS equation) and a monetary policy rule. The basic model assumes a closed economy. However, for a small open economy, such as Turkey, whose growth is largely affected by international capital flows, the ability of the standard New Neo-Classical Synthesis model to fully take into account economic dynamics is limited. In this paper, we attempt to extend the New Neo-Classical Synthesis model to a small open economy case by adding equations that would capture the exchange rate movements and the current account dynamics. Results indicate that the open economy New Neo-Classical Synthesis model for Turkey can provide a framework to capture the dynamics of the Turkish economy for the post financial liberalization era and explain the behavior of the Central Bank of the Republic of Turkey. Keywords: Open economy, New neo-classical synthesis, Monetary policy JEL Classification: F41, E52, C32. si] , IP. ite. rs. ive. [Ç u. ku. ro. va. Ün. il. :[. 19. 3.1. 40. Özet. Açık ekonomi yeni Keynesyen bir makroekonomik model: Türkiye uygulaması Makro iktisatta 1990’ların sonunda Yeni Neo-Klasik Sentez ya da Yeni Keynesyen Makroekonomik Model olarak adlandırılan yeni uzlaşı oluşmuştur. Bu uzlaşının temel özelliği Yeni Klasik, Reel İş Çevrimleri ve Yeni Keynesyen yaklaşımların sentezinden meydana gelmiş olmasıdır. Yeni Keynesyen Makroekonomik Model bir genel denge modeline dayanmakla beraber, toplam arz denklemi (Phillips Eğrisi), toplam talep denklemi (IS denklemi) ve bir para politikası kuralından oluşan üç denklemli bir sistem olarak da ifade edilebilmektedir. Temel model kapalı bir ekonomi varsaymaktadır. Bununla birlikte, Türkiye gibi büyümesi uluslararası sermaye hareketlerine bağlı küçük açık ekonomiler için, standart Yeni NeoKlasik Sentez modelinin ekonomik dinamikleri açıklama yeteneği sınırlıdır. Bu çalışmada, standart bir Yeni Neo-Klasik Sentez modeli döviz kuru hareketlerini ve cari açık dinamiklerini dikkate alan denklemler eklenerek küçük bir açık ekonomi için genişletilmiştir. Elde edilen sonuçlar açık ekonomi Yeni Neo-Klasik Sentez modelinin finansal serbestleştirme sonrası Türkiye ekonomisinin dinamiklerini ve Türkiye Cumhuriyet Merkez Bankasının davranışlarını açıklamak için bir çerçeve olarak kullanılabileceğini göstermektedir. Anahtar Kelimeler: Açık ekonomi, Yeni neo-klasik sentez, Para politikası JEL Sınıflaması: F41, E52, C32. İnd ire. n:. † Earlier versions of this paper were presented at the 9th International Conference of the Middle East Economic Association, Istanbul Technical University, Maçka-Istanbul, June 24-26, 2010, and the 2nd International Conference on Economics, Turkish Economic Association, Girne-Turkish Republic of Northern Cyprus, September 1-3, 2010. We are grateful to both conference participants and an anonymous referee for helpful comments. Special thanks to Pier Roberts for proofreading the entire manuscript and helping to season it. * Çukurova University, Department of Economics, erhanyil@cu.edu.tr ** Çukurova University, Department of Econometrics, klopcu@cu.edu.tr *** Çukurova University, Department of Economics, selimc@cu.edu.tr 2011© Her hakkı saklıdır. All rights reserved.. +0 30 0. se. Received 01 November, 2010, received in revised form 15 February 2011; Accepted 21 March, 2011. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. Erhan Yıldırım*, Kenan Lopcu**, Selim Çakmaklı***.

(2) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. +0 30 0. 7:0 5. 09 :4. ar. ih:. 12. /06. /20. 20. se ], T. .44. .54. rs. ite. si] , IP. :[. 19. 3.1. 40. ge ive. [Ç u. ku. ro. va. Ün. il n:. İnd ire. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. 1-Introduction A new consensus in macroeconomics called the New Neo-Classical Synthesis (NNS) or New Keynesian Macroeconomic Model (NKMM) emerged at the end of the 1990s. The main characteristics of this consensus are formed by the synthesis of the New Classical, Real Business Cycle and New Keynesian approaches. The new consensus models offer a combination of various elements, such as intertemporal optimization, rational expectations, imperfect competition and staggered price adjustment. These elements are brought together under a dynamic stochastic general equilibrium (DSGE) framework. Typically DSGE models are represented in three-equation systems, consisting of an aggregate supply equation (Phillips curve), an aggregate demand equation (IS equation) and a monetary policy rule. The new consensus provides a new framework for implementations of economic policies and outcomes. In particular, in terms of economic stability, it emphasizes the role of monetary policy in the short run. Monetary policy is implemented by central banks, using the short-run interest rate as a monetary policy instrument in order to ensure price stability. The NNS model has generally been analyzed in a closed economy context. However, for a small open economy, such as Turkey, whose growth is largely affected by international capital flows, the standard NNS model’s ability to fully account for economic dynamics is limited. In this paper we estimate an extended version of the NNS model for Turkey. Our model is largely based on Arestis (2007) and Buncic and Melecky (2008). As shown in Figure 1, the current account deficit (as a percent of GDP) and the GDP gap have similar dynamics in Turkey. One of the principal rationales for our estimating the open economy version of the NNS model for Turkey is this similarity of fluctuation in current account and GDP. The number of studies dealing with an open economy NNS model for Turkey is extremely limited. Two noteworthy recent studies are Huseynov (2010) and Ünalmış, Ünalmış and Ünsal (2010) which rely on Bayesian methodologies. The main purpose here is to estimate the structural parameters of the open economy NNS model and investigate whether it can be used to explain observed fluctuations of the Turkish economy, rather than performing a model calibration. Structural parameters of the model are estimated by the Full Information Maximum Likelihood (FIML) method. Needless to say the estimates of structural parameters obtained here can be used for calibration purposes.  We prefer to use the term New Neo-Classical Synthesis (NNS) rather than New Keynesian Macroeconomic Model (NKMM) to avoid confusion with the New Keynesian Economics.  We are thankful to the anonymous referee for bringing this study to our attention.. 38.

(3) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011 . l. . . . . . . . . . . +0 30 0 . . /20. . . 7:0 5. . 20. . 09 :4. se. .          . ih:. ar. ], T. .44. .54. 40. 3.1. 19. :[. si] , IP. ite. rs. ive. . Ün. . va. . ro. . 2-Brief History of the New Consensus Macroeconomic Model As stated in Blanchard (2000, p.1376), the history of macroeconomics in the 20th century can be analyzed in three different eras. According to this classification, the period before the 1940s was a period of discovery, and the theory at that time did not have a unified framework. Macroeconomic issues then consisted of monetary policy and business cycle theory. After 1940, macroeconomics was subject to extensive discussion. From the 1940s to 1980, the process of macroeconomic analysis evolved towards an integrated approach. Beginning with the IS-LM model, the dynamic general equilibrium model was developed, and the roles of shocks and transmission mechanisms were clarified. During this period, Neo-Classical Synthesis emerged and became the dominant theory in macroeconomic policy discussions. The period from the 1980s to date, on the other hand, can be considered a new era of discovery and exploration. In this new era, macroeconomic studies have focused on macroeconomic imperfections, such as the relevance of wage and price settings, incompleteness of markets, asymmetric information, search and bargaining in decentralized markets and increasing returns in production.. ku. . [Ç u. . il. . n:. . This study consists of 5 sections, including the introduction. A brief historical development of the new consensus in macroeconomics is provided in section 2. Data and methodology are explained in section 3. Results and discussion are presented in section 4. And finally, some concluding remarks are offered in section 5.. İnd ire. . ge. . 12. /06. Figure 1. Current account and GDP gap for Turkey, 1990-2009 . B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. . 39.

(4) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. ar. ih:. 12. /06. +0 30 0. 7:0 5. 09 :4. /20. 20. se ], T. .44. .54. rs. ite. si] , IP. :[. 19. 3.1. 40. ge ive. [Ç u. ku. ro. va. Ün. il n:. İnd ire. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. The Neo-Classical Synthesis shaped after the post-war period was based on two fundamental principles. According to the first principle, the rational behavior of individuals and companies can be studied using the standard methods of microeconomics. The second principle holds that market prices and wages cannot adjust immediately to the full employment equilibrium level. However, full employment can be achieved with the use of monetary and fiscal policies (Blanchard, 2008). The consensus which was built around the Neo-Classical Synthesis in macroeconomics until the mid-1970s was unraveled because of the empirical failure of the Phillips curve and the lack of microeconomic principles of macroeconomic inferences. After the demolition of the Neo-Classical Synthesis, economic researchers in general divided into two camps: on the one hand, there was the New Classical Economics and the Real Business Cycle Theory with rational expectations and continuous market clearing assumptions; on the other hand, there was the New Keynesian Economics with its attempt to find micro foundations which are consistent with rational expectations for the Keynesian wage and price rigidities. Although the New Keynesian models accept the assumption of rational expectations, they differ from the New Classical Economics in terms of price setting. In New Classical Economic models, firms are price takers in competitive markets, but supply decisions are taken under incomplete information conditions. In contrast, in the New Keynesian Economics, firms are price setters in imperfectly competitive markets (monopolistic competition) (Snowdon, Vane & Wynarczyk, 1995, p.291; Gordon, 1990, p.1136). New Keynesian models that have been developed since the 1980s take into account both nominal and real rigidities and have demonstrated that economic policy can be used to ensure economic stability in the short run. Although some disagreements continue about various theoretical points, there is a new consensus in macroeconomics wherein a satisfactory macroeconomic model should include a rational expectations hypothesis, wage and price rigidities and optimization behavior of individuals, firms and policy makers (Carlin & Soskice, 2006, p. 563). The main feature of this consensus is that it brings together the intertemporal optimization and rational expectations hypothesis of the New Classical Economics and the Real Business Cycle Theory, imperfect competition and the costly price adjustment elements of the New Keynesian Economics (Goodfriend & King, 1997, p.25). The key idea in the new consensus is that the firm’s price setting behavior causes some temporary nominal rigidities, thus enabling monetary policy to have real effects in the short run while maintaining the neutrality of  See Mankiw (1990) for a detailed discussion.  See Gordon (1990) for a comprehensive survey of New Keynesian Economics.. 40.

(5) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. +0 30 0. 7:0 5. 09 :4. ar. ih:. 12. /06. /20. 20. se ], T. .44. .54. rs. ite. si] , IP. :[. 19. 3.1. 40. ge ive. [Ç u. ku. ro. va. Ün. il n:. İnd ire. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. money in the long run (Clarida, Gali & Gertler, 1999). Nevertheless, the New Neo-Classical Synthesis differs from the old consensus in terms of its reliance on dynamic micro foundations and the imperfect competition equilibrium models (Dixon, 2008, p.3). Another important difference is that the money supply is endogenous in the new synthesis as opposed to being exogenous in the old one. The NNS approach to monetary policy analysis includes the systematic use of the dynamic stochastic general equilibrium framework. A typical general equilibrium model can be formulated in the form of three blocks: aggregate demand, aggregate supply, and a policy rule. Aggregate demand and aggregate supply are derived from the maximization problem of households and firms, respectively, and the policy rule is derived from the minimization of the social loss function. Firms are modeled in the context of monopolistic competitive markets. Each firm has a well-defined demand curve for the goods it produces, and it sets prices to maximize its discounted profits. In the model, the main source of the non-neutrality of monetary policy is nominal rigidities. Generally, nominal rigidities are introduced in the form of restrictions on the frequency of the price (wage) setting process of firms (employees). These limitations on price (wage) decisions imply the forward-looking component of the price (wage) decision, because each economic agent recognizes that prices (wages) will remain constant for a certain time. In this case, it will be optimal to take into account expectations about the future demand and cost conditions. In this way the new consensus models bring a new perspective to the nature of inflation dynamics. One source of observed fluctuations in inflation in the NNS model is the output gap, which is defined as deviations of the current output level from its equilibrium level in the absence of nominal rigidities. The central bank responds to deviations of inflation and output from their equilibrium levels by adjusting the short run interbank interest rate. The transmission of monetary shocks to real variables works through a conventional interest rate channel. Thus, dynamic stochastic general equilibrium models with nominal rigidities provide a utility based welfare criterion to evaluate the desirability of alternative monetary policies. This policy rule associated with small welfare losses (derived from the minimization of welfare loss function) provides a good approximation to the optimal rule and is not subject to the Lucas critique (Gali, 2002, pp. 2-3 and Gali, 2009, pp. 1-2). 3-Data and Methodology As stated above, a typical closed economy version of the NNS model is represented with a three-equation system, consisting of an aggregate supply equation, aggregate demand equation and a monetary policy rule. 41.

(6) . İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. .54. .44. ], T. ar. ih:. 12. /06. +0 30 0. 7:0 5. 09 :4. /20. 20. se. ive. rs. ite. si] , IP. :[. 19. 3.1. 40. ge [Ç u. ku. ro. va. Ün. il n:. İnd ire. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l.  These equations can be estimated separately or simultaneously as a system.  However, the closed economy version of the NNS model could be inadequate for a small open economy, such as Turkey, whose growth is largely affected  π = δ E t πcapital ( −has   π t − +and δ  −lasting δ  )(er ) + λyt +deficit ε AS,t  problems. t + + δ flows t + π t account by international current  t Yıldırım, Lopcu, Çakmaklı and Özkan (2010) conclude that the standard   NNSmodel’s ability to fully account for economic dynamics is limited for Turkey.  π = δ E π + δ π + ( − δ − δ )(er + π  ) + λy + ε     t t +  t −   t t AS,t We t employ the following model which ist a modified version of Arestis’s Eλy   π t interpretation ( −Open )(ert + rπt t−) +New    δ  − δ  Economy εAS,t   (2007) of t πt t+ +Consensus  E t π t + + δ  π t −  +an  yt==δα  +ψ qt +θyMacroeconomic   y + Et yt+ + ( −  ) yt − + ϕ  t + ε IS,t  Model.  + Et πt+     (1)  π t = δ E t π t+ + δ  π t − + ( − δ  − δ  )(ert + π t ) + λy t + ε AS,t      rt − Et πt+   y = α + E y + ( − ) y  r +− ϕ Et πt+   +ψ qt + θyt + ε IS,t     t y t t+ t−t    +ψ q + θ  t = α y + Et yt+ + ( −  ) yt − + ϕ   y + Etπtt+ yt + ε IS,t     + E π  t t+    rt − Et πt+  (2)  yr ==αα + E   ]ψ ( ) [ y + −  y + + θt yt + εIS,t    ϕ + ρr + (  − ρ) βE +εqtMP,   y mp t t+t −  t −   t π t + + γyt+  tt  + Et πt+     r − Et πt+     +ψ q + θyt + ε IS,t   yt = α y + Et yt+ + ( − ) yt − + ϕ  t   rt = α mp + ρr t − + (  − ρ) [ βE t π t + +  γy+t ]E+ πε MP, t   t      (3) rt = α mp + ρr t −  + (  − ρ) [ βE t π t +t +t+γy t ]+ ε MP, t        rt − E t π t +    − (rt  − E t π t+  ) + φ ca t + ψ  E t  q t +   + ε  q t = α q + ϕ       + E t π t +   rt − E t π t +      (4) q = α q + ϕ   − (rt − E t π t +  ) + φ ca t + ψ  E t  q t +   + ε q  t    rt t = α mp + ρr t−+ +E (t πt +− ρ) [ βE π + γy t ]+ ε MP, t       rt − E t π t +  t t +    − (rt − E t π t +  ) + φ ca t + ψ  E t  q t +   + ε  q = α q + ϕ   t  + E t π t +          (5)  ca t = α ca + ψ  q t + ω  y t + ω  y t + ε ca t  ca = α + ψ q + ω y + ω y  + ε         t ca  t  t  t ca  t  r − E t π t +    q t = α q + ϕ   t − (rt  − E t π t+  ) + φ ca t + ψ  E t  q t +   + ε q  er ==αq −+ ψ pt* q+t ++ptωE t yπ tt ++ ω  y t + ε ca t     (6)  ca    t tt ca   q − pt +p tpt + p t   erer t = t =t q t −               Equation (1) is the aggregate supply equation with inflation determined    by past  and expected inflations (πt-1 and Et πt+1), current output gap (yt), the    * worldca ), the exchange +ψ q +change ω  y t +inωthe   rate (Δer  t), and  the   t = α ca (π  y nominal t + ε ca  t   inflation t  t er = q − p + p  t t t t  stochastic aggregate supply   shock  ε AS,t  . In the NNS model,   aggregate   the     supply  equation or Phillips curve is derived from the aggregation of the price   of firms, based on Calvo (1983). In Calvo (1983), inflation setting  decision    was expressed as a function of expected inflation and output gap. In the Calvo   er = q − p + p  setting,t each t t t    to change   price whenever  a random    his   price setter is allowed   42              .

(7) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. +0 30 0. 7:0 5. 09 :4. ar. ih:. 12. /06. /20. 20. se ], T. .44. .54. rs. ite. si] , IP. :[. 19. 3.1. 40. ge ive. [Ç u. ku. ro. va. Ün. il n:. İnd ire. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. signal is received. However, all price setters do not receive a signal at the same time, so price revisions are not simultaneous. In any period, only a fraction of firms will receive the signal and revise their price decisions. Each firm has the same probability of being one of the firms receiving the signal. Importantly, it is optimal for an individual firm to take into account the expected future price level and demand conditions because its price will remain unchanged  for some time in the future.  Equation (2) is the aggregate demand or IS equation with current output  gap defined as the function of past and future output gaps, real interest r −Eπ  rate  1t + E tπ t+1 , real exchange rate (qt), the world output gap (yt*), and the t t +1 aggregate demand shock  ε IS,t  The aggregate demand equation represents  the current output gap with   backward and forward dynamics. Such hybrid  identification allows us to capture the persistence that can be seen in a time  series. In the aggregate demand equation, backward dynamics are introduced  formation. However, there is no consensus on  through some form of habit  whether the consumption habits of households should be internal or external.  “With internal habits a household’s marginal utility of consumption depends  on the history of its own consumption, whereas with external habit formation  it depends on the history of other households’ consumption” (Dennis, 2005,    pp. 1-2). In equation (2) external habit persistence is considered. This external habit  formation gives a  micro foundation of backward dynamics in the   The real exchange rate affects the demand for aggregate demand equation.   imports and exports, and thereby the level of the current output gap.    Equation (3) is the central bank reaction function or monetary policy    rule with the past nominal interbank interest rate, expected inflation, current  output gap and the monetary policy shock term  ε MP,t  The coefficient ρ   prepresents the degree of smoothness in the interest  rate change. As mentioned   in Clarida, Gali and Gertler (2000) any weight to past values of the interest    implies the gradual adjustment of the interest rate rate in the reaction function   to its equilibrium value.   Equation (4) gives the real exchange rate as a function of the real interest  rate differentials, current account (cat), the expected   value of the real exchange   rate itself, and a random shock term  ε q,t  Equation (5) determines the   current account as a function of the current real exchange rate, domestic and   . . . . . . . . . 43.

(8) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. . +0 30 0. 7:0 5. 09 :4. 12. /06. /20. 20. se. rs. ite. si] , IP. :[. 19. 3.1. 40. .54. .44. ], T. ar. ih:. ge ive. [Ç u. ku. ro. va. Ün. il n:. İnd ire. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l.  world output gaps, and the random shock term  ε ca,t  Finally, equation (6) is an identity equation which describes the nominal exchange rate in terms of  the real exchange rate and domestic (p) and foreign (p*) price level differences,  all in natural logarithms. The change in the nominal exchange rate appearing as     − π  + π   There  in equation (1) can be derived from equation (6) are six equations and six unknowns: output, interest rate, inflation, real  exchange rate, current account and nominal exchange rate.   As an alternative to the Arestis model, we also estimate a model based on  Buncic and Melecky (2008) for the open economy version of the NNS model.   Buncic and Melecky (2008) use a two block model in which the first block  represents the domestic economy, akin to equations (1)-(3), and the rest of the   world is modeled as the closed economy following Cho and Moreno (2003,  affects, but is not affected 2006). In this specification the rest of the world  by, the domestic economy. That is, the rest of the  world is exogenous to the  domestic economy. We estimate this model and compare the results with the   Arestis version of the open economy NNS model.  The NNS model is a system and there are twoways for estimating structural parameters of the system. First, each equation in the system can be estimated separately. In general, a single equation instrumental variables (IV) method   does not provide efficient estimators. System IV estimators, such as three  provided that there exists stage least squares (3SLS), will be more efficient  a sufficient number of good instruments. An alternative system method, the  Full Information Maximum Likelihood (FIML), on the other hand, is efficient  among all estimators. Linde (2005) illustrates via Monte Carlo simulations that single equation methods, e.g., GMM estimators, are likely to be biased   and argues that employing the FIML is a way of obtaining better estimates. We, therefore, determine FIML as the method of estimation in this study, as   in many earlier studies.   data for the 1990:Q1-2009: The NNS model is estimated using quarterly Q4 period. The interest rate is obtained from the IFS (International Financial Statistics-IMF) database. The GDP, the consumer price index (CPI, 2005=100)  OECD. The output gap is and the current account series are obtained from the calculated by applying the Hodrick-Prescott filter to the seasonally-adjusted,  logarithmic GDP series. The world output gap is calculated similarly. The. .  We should note that all contemporaneous shock terms are assumed to follow a multivariate normal distribution since we estimate the model using the Full Information Maximum Likelihood (FIML) method, as indicated below.   See Henry and Pagan (2004) for a discussion of the econometrics of the NNS models.   GDP: Millions of US dollars, volume estimates, fixed PPPs, OECD reference year, annual levels, seasonally adjusted.. 44.    .

(9) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. +0 30 0. 7:0 5. 09 :4. ar. ih:. 12. /06. /20. 20. se ], T. .44. .54. rs. ite. si] , IP. :[. 19. 3.1. 40. ge ive. Ün. il İnd ire. n:. [Ç u. ku. ro. va.  We convert the current account series from the OECD into constant US dollars using US CPI (2005=100). Then, we use the IFS GDP series in national currency along with the IFS market exchange rate and convert the series into current US dollars. We then calculate the real GDP series in 2005 dollars. Next, we compare the real GDP series we obtained with the OECD-based PPP-adjusted real GDP series and adjust the real current account. Finally, we divide the adjusted real current account by the OECD-based PPP-adjusted GDP series to obtain the current account as a percentage of GDP (CA).  We use the CBRT buying rates for the original series. 10 Tests are performed using a standard z-test since Maximum Likelihood estimators are asymptotically normal estimators. 11 The value of the Likelihood Ratio test is 6.66, which, with 3 degrees of freedom, corresponds to a p-value of 0.0835. 12 The only exception to this is the effect of the real interest rate on the GDP gap becoming significant, which is the result we obtain below when we drop the insignificant variables from the model.. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. interbank interest rate and the CPI-based series are used for the monetary policy instrument and inflation, respectively. The rest of the world is defined by G7 data. We compute the weights using each country’s share in Turkey’s trade, and then normalize and use these weights to calculate the interest rate, the GDP, and the inflation rate for the rest of the world. The current account series from the OECD is in current US dollars and is seasonally adjusted. It is converted into real dollar terms and adjusted in order to obtain the current account as a percentage of the OECD-based PPP-adjusted real GDP (CA). The exchange rate is defined as the amount of domestic currency per world currency and calculated using the weights described above, after which the natural logarithm is taken. The original series are from the Central Bank of the Republic of Turkey (CBRT). The real exchange rate series is calculated using the exchange rate and the domestic and world price series and in logarithm as well. 4-Results and Discussion Estimation results of the Arestis model of NNS are given in Table 1. Firstly, all of the variables with the exception of coefficients on current account deficit and the real exchange rate in equations (4) and (5) are associated with the expected coefficient signs, though some are not statistically significant. Tests for each of the first three coefficients being equal to 1/3 in our aggregate supply equation indicate that the nulls of the true parameters being equal to 1/3 cannot be rejected at any conventional significance level.10 This finding is further strengthened by a Likelihood Ratio test of the joint null hypothesis that the true values of all first three parameters are equal to 1/3.11 Further, re-estimating the model with the above restrictions on the first three parameters leaves all the other estimated coefficient values and significance levels practically unaffected.12 Based on these findings, we conclude that the hypothesis of economic agents giving the same weight to the first three variables in Turkey cannot be rejected. The output gap coefficient in the aggregate supply equation, on the other hand, is insignificant.. 45.

(10) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. Table 1. FIML Estimation of Modified Version of Arestis (2007). l. . Arestis 2007 without insignificant variables Coefficient. δ1. 0,3592. 0,1334. 0,0071. δ2. 0,4491. 0,1589. 0,0047. λ. 0,0121. 0,4014. 0,9758. αy. 0,0142. 0,0135. 0,2921. . 0,5030. 0,0860. 0,0000. φ1. 0,4263. 0,2830. 0,1320. ψ1. 0,0011. 0,0304. 0,9708. 0. θ. 0,5179. 0,2964. 0,0806. 0. αmp. 0,0026. 0,0233. 0,9097. 0. ρ. 0,4531. 0,1710. 0,0081. 0,4607. β. 1,3169. 0,3672. 0,0003. 1,3130. γ. 0,9429. 0,6975. 0,1765. 0. αq. 0,0252. 0,0563. 0,6542. 0. φ2. 0,3368. 1,3163. 0,7980. φ. 1,2002. 1,1432. 0,2938. 1,0108. 0,1415. 0,0000. 0,0023. 0,0076. 0,7556. 0. 0,0454. 0,0210. 0,0308. 0,7117. 0,2020. 0,2175. 0,4753. B. ω2. 0,1450. 0,0193. 0,4711. 0,1142. 0,0000. 0,0103. 0,0055. 0,0654. 0,5667. 0,0643. 0,3150. 0,1796. 7:0 5. 0,0083. 0,0000. 0,0004. 0,6101. 0,0907. 0,0000. 0,6473. 0. .54. 3.1. 19. :[ si] , IP. ite. rs. ive. Ün. 0,9955. n:. [Ç u. 20. 0,0500. .44. 0,0000. 0. 0,0279. 0,0249. 0,0512. 0,0526. 0,0943. 0,0967. 0,0223. 0,0214. İnd ire. /20. /06 12 ih: ar ], T. ge. 0,0342. 0. σIS,t. 46. 0,0795. 0,0000. 0,0481. σCA,t. 0,0000. 0,1658. 0,0484. σq,t. 09 :4. 0. σAS,t. σMP,t. +0 30 0. 0,3394. 0,0000. va. ω1. ro. ψ3. ku. αca. P Value. 0,0720. il. ψ2. Standard Error. se. Coefficient Standard Error P Value. 40. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. Arestis 2007.

(11) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. +0 30 0. 7:0 5. 09 :4. ar. ih:. 12. /06. /20. 20. se ], T. .44. .54. rs. ite. si] , IP. :[. 19. 3.1. 40. ge ive. [Ç u. ku. ro. va. Ün. il n:. İnd ire. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. According to the aggregate demand equation, the past and future output gaps have the same weights. The real interest rate and real exchange rate coefficients have the right signs but are insignificant. Lastly, the world output gap has the expected sign and is marginally significant. Equation (3) shows that the Central Bank responds to expected inflation by increasing the interest rate around 0.72 percent. The output gap coefficient in equation (3) is insignificant, implying that the Central Bank does not respond to fluctuations in the output gap. In equation (4) the real exchange rate is not influenced by the current account and the difference between the real domestic and real world interest rates, but only by the expected real exchange rate. The last equation shows that the depreciation of the real exchange rate increases the current account deficit, contrary to the theoretical expectations. Increases in the domestic gap raise the current account deficit, a fact consistent with observations for the Turkish economy. The second column of Table 1 presents results when insignificant coefficients are eliminated from the model one by one, starting with the coefficient that has the highest p-value.13 In this version of the model, the effect of the real interest rate on the GDP gap becomes significant and has the right sign, consistent with theoretical expectations. The effect of the world gap on the domestic gap, on the other hand, becomes insignificant. Overall our results are consistent with the other empirical studies, such as Giordani (2004), Buncic and Melecky (2008), and Yıldırım, et al. (2010). Figure 2 presents the actual and predicted values of the model’s endogenous variables. Actual and predicted values coincide well for output gap but differ for inflation and interest rate variables, especially in times of economic crises. For the current account deficit, on the other hand, the actual and predicted values coincide better during times of economic crises and differ at other times. The model performs the worst in terms of the actual and predicted values of the real exchange rate. Given the poor performance of the model, especially in the real exchange rate equation, we perform a Zivot and Andrews (1992) unit root (ZA) test, allowing a structural break in the mean and the trend of the real exchange rate and reject the null of the unit root. The ZA test picks the break point of the real exchange rate consistently as the 2000:Q4 period. We apply the Quandt-Andrews type of tests, called SupF, for structural breaks in the mean. 13 Here we eliminate the coefficient with the highest p-value and re-estimate the model, then eliminate the coefficient in the re-estimated model with the highest p-value and continue this process until all the insignificant coefficients at the 10% level are eliminated. It should be noted that the p-values for each re-estimated model are different, and some of the coefficients with a lower p-value in the initial model become insignificant in the final model.. 47.

(12) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. +0 30 0. 7:0 5. se. . (. 20. ). ). /06. /20. (.    . 09 :4.  r − Et πt+  r − Et πt+  qt = α q + α dm + ϕ  t − rt − Etπ t+  + τD t − rt − Etπ t+  + Et πt+  + Et πt+  + φcat + τ  Dcat  +ψ qt + + τ  Dqt +  + τ trend + τ  Dtrend + ε qt. 12. . ih:. ge. . si] , IP. ite. rs. ive. . il. :[. 19. 3.1. 40. .54. .44. ], T. ar.  results presented in Table 2 reveal that the signs and significance The of the original coefficients are essentially the same, except the world gap coefficient in equation (2) turning out to be insignificant and the intercept and  the current account coefficient in equation (4) becoming significant. When  the actual and the predicted values of the model’s endogenous variables are  compared (Figure 3) with the benchmark model, the model’s performance  improves significantly, especially in terms of the exchange rate variable. Even if the insignificant coefficients are excluded, the model performs well, as the  differences between the actual and predicted values are essentially unaffected.  The above result illustrates that it is essential to test and account for structural   for meaningful comparisons. changes. . va. Ün. π t = δ Et π t+1 + δ  π t −1 + ( − δ  − δ  )qt + λy t + ε AS,t  . . ku.  . . [Ç u.  . ro. (7)  π t = δ Et π t+1 + δ  π t −1 + ( − δ  − δ  )qt + λy t + ε AS,t . B 48. n:.  r − Et π t+1   * (8) yt = α y + Et yt+1 + (1 −  ) yt −1 + ϕ  t  1 + E π  +ψ qt − + θ yt + ε IS,t   r − Et πt+1   t t+1  +ψ qt − + θ y*t + ε IS,t  yt = α y + Et yt+1 + (1 −  ) yt −1 + ϕ  t  1 + E π  t t+1    14 See Andrews (1993) and Hansen (1997&2000) for a detailed discussion.   İnd ire. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. and mean and trend of the real exchange rate (by calculating a Chow type test statistic for every point from the 10th percentile of the sample to the 90th percentile of the sample and picking the largest F-value).14 Additionally, we test the real exchange rate model of equation (4) and a modified version of equation (4) by adding a time trend. In all the cases, again the period 2000: Q4 gives the largest F-value with p-value near zero. Thus, we conclude that the parameters of the exchange rate equation are different before and after the 2001 crisis. We then re-estimate the model with the new exchange rate equation (4`) given below, where D is the dummy variable and takes the value of 1 starting with 2001:Q1.. . r t = α mp + ρr t − 1 + (1 − ρ )[ βE t π t + 1 + γy t ] + ε MP, t  .  r t = α mp + ρr t − 1 + (1 − ρ )[ βE t π t +1 + γy t ]+ ε MP, t  . .  .  . .

(13)          . +0 30 0.   . /20. 20. . /06. 12. ih:. si] , IP ite rs ive. İnd ire. n:. [Ç u. ku. ro. va. Ün. il. :[. 19. 3.1. 40. .54. .44. ], T. ar. ] ] ] ]. 7:0 5. ) ) ) ). 09 :4. ( ( ( (. se. ) ) ) ) [ [ [ [. ge. ( ( ( (. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. r1t + − Ett π tt++11    +ψ qt − + θ y*t + ε IS,t  y t = α y + Et y t+1 + (1 −  ) y t −1 + ϕ  rt1−+EEttππtt++11    +ψ qt − + θ y*t + ε IS,t  y t = α y + Et y t+1 + (1 −  ) y t −1 + ϕ   1 + E π t t+1     rt = α mp + ρr t − 1 + (1 − ρ )[βE t π t +1 + γy t ]+ ε MP, t      İktisat İşletme ve Finans 26 (305) Ağustos / August 2011 r = α mp + ρr t − 1 + (1 − ρ )[βE t π t +1 + γy t ]+ ε MP, t    t  rt = α mp + ρr t − 1 + (1 − ρ )[βE t π t +1 + γy t ]+ ε MP, t     (9) rt = α mp + ρr t − 1 + (1 − ρ )[ βE t π t +1 + γy t ]+ ε MP, t     * * * * * * * * π = δ 1 E t π t + 1 + (1  δ 1 )π t  1 + λ y t + ε AS, t       *t π t = δ 1* E t π *t + 1 + (1  δ 1* )π t* 1 + λ * y *t + ε *AS, t     * * * * * * * * (10) π t = δ 1 E t π t + 1 + (1  δ 1 )π t  1 + λ y t + ε AS, t      π *t = δ 1* E t π *t + 1 + (1  δ 1* )π t* 1 + λ * y *t + ε *AS, t      * y t = α *y +  * E t y *t + 1 + 1   * y t*1 − ϕ * rt* − E t π *t + 1 + ε *IS, t    * * * * * * * * * * (11) y t = α y +  E t y t + 1 + 1   y t 1 − ϕ rt − E t π t + 1 + ε IS, t    y * = α * +  * E y * + 1   * y * − ϕ * r * − E π * + ε * t y t t+1 t 1 t t t+1 IS, t    * * * * * * * * * * y t = α y +  E t y t + 1 + 1   y t 1 − ϕ rt − E t π t + 1 + ε IS, t    * * * * * * * * * * (12) rt = α mp + ρ rt  1 + (1  ρ ) β E t π t + 1 + γ y t + ε MP, t     * * * * * * * * * * rt = α mp + ρ rt  1 + (1  ρ ) β E t π t + 1 + γ y t + ε MP, t     r * = α * + ρ * r * + (1  ρ * ) β * E π * + γ * y * + ε *    t 1 t mp t t+1 t MP, t * * * * * * * * * * rt = α mp + ρ rt  1 + (1  ρ ) β E t π t + 1 + γ y t + ε MP, t      For comparison purposes we estimate a modified version of the Buncic   and Melecky (2008) open economy NNS model described by equations (712). Here, the difference from the Arestis model is the addition of the real  exchange rate, qt in equation (7) and the replacement of qt by qt −1 in equation  (8). Equation (9) is the same as in the Arestis model, and the equations (10-12) describe the standard closed economy NNS model for the rest of the world.. 49. .   .

(14) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. .  .  . l. . se.  . /20. 20. .          . 12. /06. . ih:. ge. . ], T .44. . .54. 3.1. . 19. .   . . .          . . ro. va. Ün.          . ite. . . rs. . ive. il. . si] , IP. :[. . .           . 40. . . ar. . . 09 :4.           . 7:0 5. .  . +0 30 0.  . B. [Ç u n:.       Figure   50. ku. . İnd ire. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. . 2. Actual and fitted values of modified version of Arestis (2007).

(15) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. Table 2. FIML Estimation of the Modified Version of Arestis (2007) with              Structural Break Standard Error. P Value. δ1. 0,3126. 0,1882. 0,0967. δ2. 0,4878. 0,1787. λ. 0,0565. 0,3459. 0,8702. αy. 0,0204. 0,0166. 0,2202. . 0,4451. 0,1156. 0,0001. φ1. 0,4629. 0,3088. 0,1339. ψ1. 0,0117. 0,0361. 0,7442. θ. 0,5022. 0,3408. 0,1406. αmp. 0,0001. 0,0309. 0,9965. ρ. 0,4377. l. Coefficient. 0,0576. /06. 0,2305. 0,9051. 0,1281. 0,0428. 1,1593. 0,5571. 0,1695. 0,9616. 0,7312. 1,1256. 0,5159. 3,0514. 1,1021. 0,0056. 1,7640. 0,4398. 0,1700. 0,0033. 0,3314. 0,2917. 0,0014. 0,7098. 0,0069. 0,0803. 0,0093. 0,9695. 1,3626. ih: ar ], T. .44. .54. 0,0042 0,2513 0,0028 0,0374 0,8600. τ3. 0,3494. τ4. 0,0005. τ5. 0,0120. αca. 0,0003. ψ3. 0,0525. 0,0257. 0,0414. ω1. 0,7195. 0,2071. 0,0005. 0,0906. 0,5210. 0,8619. ite rs ive. Ün. va. ku [Ç u n:. σAS,t σIS,t σMP,t σq,t σCA,t. İnd ire. ω2. si] , IP. 0,4988. il. ψ2. ro. τ2. 40. φ. 3.1. τ1. 19. φ2. :[. αdm01. 12. 0,4780. 1,0384. ge. 1,3681. γ αq. +0 30 0. 7:0 5. 09 :4. /20. 20. se. 0,0064. β. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. . 0,0484 0,0286 0,0521 0,0730 0,0224. 51.

(16) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. . . . . . l. . .          . /20. 20. . /06. . ih: ar. ge. . 19. .  . si] , IP. :[. . .   .          . . ro. va. Ün.          . ite. . rs. . ive. il. . . .44. 3.1. 40. . .           . .54. . . ], T. . . 7:0 5. . .           . 09 :4. . +0 30 0. se . . 12. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300.  .  . ku. [Ç u n:. . İnd ire. B. . Figure 3. Actual and fitted values of Arestis (2007) with structural break in equation (4)  52 . .

(17) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. 0,0722 0,0176. 0,0203. 0,3582. 0,0326. 0,4373. 0,5369. 0,8929. 0,0697. 0,0195. 0,3668. 0,0134. 0,1854. 0,0184. 0,4226. 0,8690. 0,0120. 0,2660. 0,1724. 0,0199. . 0,3900. 0,1680. φ. 0,4718. 0,4875. ψ. 0,0033. 0,0240. 0,8905. 0,0241. 0,0606. 0,6903. θ. 0,0019. 0,7028. 0,9977. 0,0210. 0,6820. 0,9754. αmp. 0,0121. 0,0358. 0,7353. 0,0084. 0,0327. 0,7976 0,0331. ge. 0,4014. 0,4389. 0,2799. ], T. 0,4742. .44. 0,3331. 0,2135. 0,0901. 0,4281. 0,2009. 1,1877. 0,3700. 0,0013. 1,2169. 0,3438. γ. 0,4504. 0,8225. 0,5839. 0,5002. 1,0494. δ1*. 0,5498. 0,1236. 0,0000. 0,5474. 0,1173. 0,0000. *. 0,0133. 0,0648. 0,8363. 0,0155. 0,0691. 0,8224. αy*. 0,0003. 0,0009. 0,7121. 0,0003. 0,0009. 0,6940. *. . 0,4821. 0,0796. 0,0000. 0,4803. 0,0870. 0,0000. φ*. 0,0479. 0,1612. 0,7663. 0,0494. 0,1696. 0,7706. 0,6336. 0,0003. 0,8338. 0,0000. 0,0004. 0,8626. 0,9530. 0,0324. 0,0000. 0,9536. 0,0335. 0,0000. *. 0,6558. 0,8930. 0,4627. 0,6761. 0,9865. 0,4931. γ*. 0,9690. 0,9549. 0,3102. 0,9686. 0,9950. 0,3303. ku. ro. va. Ün. 0,0000. ρ* β. 0,0541. σIS,t. 0,0281. σMP,t. 0,0547. σ* AS,t. 0,0033. 0,0034. 0,0023. 0,0023. 0,0008. 0,0008. IS,t. σ* MP,t. İnd ire. n:. [Ç u. σAS,t. *. 0,0004. 40. 3.1. 19. :[. si] , IP ite. rs. ive. αmp. *. .54. 0,3619. il. ρ β. λ. σ. 0,0874. 09 :4. λ αy. 0,0343. 20. 0,2234. /20. 0,4773. P Value. /06. 0,5205. δ2. Standard Error 0,2095. 12. P Value Coefficient. ih:. δ1. Standard Error 0,2458. 7:0 5. With ( ert + π t* ). ar. Coefficient. +0 30 0. With Real Exchange Rate. se.  Table 3. FIML Estimation of Buncic and Melecky (2008) Version of NNS. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. The estimation results of the Buncic and Melecky (2008) model are given in Table 3. The coefficients generally have the expected signs, and overall the results are consistent with the results of the Arestis model. What is different from the Arestis model is that the point estimates imply noticeably stronger forward dynamics in the aggregate supply equation and stronger backward dynamics in the aggregate demand equation.. 0,0483 0,0284 0,0542. . 53.

(18) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. +0 30 0. se. 54. İnd ire. n:. [Ç u. ku. ro. va. si] , IP. ite. rs. ive. Ün. il. :[. 19. 12. ih:. ar. ], T. .44. .54. 3.1. 40. ge. /06. /20. 20. 09 :4. 7:0 5.  5-Conclusion The results of the analysis for the Turkish economy for the 1990:Q1 2009:Q4 period are summarized as follows:  estimation results of the Arestis (2007) and the i. According to the Buncic and Melecky (2008) versions of NNS models, the hypothesis that the  economic agents in Turkey give similar weight to past and future inflations in  their decision making process cannot be rejected. ii. The output gap does not have any statistically significant impact on inflation.  iii. Backward and forward dynamics have similar weight in the aggregate demand equation.  iv. The real interest rate and the real exchange rate do not have a significant  impact on the output gap for all but one model. For the Arestis model  without insignificant variables, the real exchange rate becomes marginally significant.  v. In all the specifications, the Central Bank gives significant weight to inflation expectations in the monetary reaction function but does not respond to fluctuations in the output gap.  vi. The actual and fitted values are better matched with the structural  equation. break in the exchange rate These results clearly  indicate that the NNS model can provide a useful framework to explain the fluctuations in the Turkish economy. It is equally  clear that further investigations regarding the potentials of alternative NNS models to better mimic the Turkish economy are needed. In particular, the impact of the real exchange rate on current account and conversely current  are controversial. Future research could focus on account on exchange rate exchange rate and current  account dynamics to better capture the fluctuations of the Turkish economy. . B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l.    When we replace the real exchange rate, qt in equation (7) and qt −1 in equation (8) with    + π   and its first lag, respectively, all point estimates are essentially the same as their Arestis model counterparts. The  only noteworthy difference is the insignificance of the world gap coefficient and again relatively stronger-backward dynamics in the IS equation. Thus,  overall it is reasonable to say that estimates of the main parameters of the NNS model are robust across the specifications of both the Arestis and the Buncic and Melecky models and their several variations. .   .

(19) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011 References. +0 30 0. 7:0 5. 09 :4. ar. ih:. 12. /06. /20. 20. se ], T. .44. .54. rs. ite. si] , IP. :[. 19. 3.1. 40. ge ive. [Ç u. ku. ro. va. Ün. il n:. İnd ire. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. Andrews, D.W.K. (1993). Tests for parameter instability and structural change with unknown change point. Econometrica, 61 (4), 821-856. doi:10.2307/2951764. Arestis, P. (2007). What is the new consensus in macroeconomics? In P. Arestis (Ed.), Is there a new consensus in macroeconomics? (pp. 22-41), Basingstoke: Palgrave Macmillan. Blanchard, O. (2000). What do we know about macroeconomics that Fisher and Wicksell did not? The Quarterly Journal of Economics, 115 (4), 1375-1409. doi:10.1162/00 3355300554999. Blanchard, O. (2008). Neoclassical synthesis. In S. N. Durlauf & L. E. Blume (Eds.), The new Palgrave dictionary of economics online, (2nd ed.). Palgrave Macmillan. <http://www.dictionaryofeconomics.com/article?id=pde2008_N000041> doi:10.1057/9780 230226203.1172. Buncic, D. & Melecky, M. (2008). An estimated new Keynesian policy model for Australia. The Economic Record, The Economic Society of Australia, 84 (264), 1-16. doi: 10.1111/j.1475-4932.2008.00443.x doi:10.1111/j.1475-4932.2008.00443.x. Calvo, G. A. (1983). Staggered prices in a utility-maximizing framework. Journal of Monetary Economics, 12 (3), 383-398. doi:10.1016/0304-3932(83)90060-0 doi:10.1016/0304-3932(83)90060-0. Carlin, W. & Soskice, D. (2006). Macroeconomics: Imperfections, institutions and policies. Oxford, UK: Oxford University Press.. Cho, S. & Moreno, A. (2003). A structural estimation and interpretation of the new Keynesian macro model. Facultad de Ciencias Económicas y Empresariales, Universidad de Navarra, Faculty Working Papers, No: 14/03.. Cho, S. & Moreno, A. (2006). A small-sample study of the new-Keynesian macro model. Journal of Money, Credit, and Banking, 38 (6), 1461-1481. doi: 10.1353/ mcb.2006.0078. doi:10.1353/mcb.2006.0078. Clarida, R., Gali, J. & Gertler, M. (1999). The science of monetary policy: A new Keynesian perspective. Journal of Economic Literature, 37 (4), 1661-1707. doi:10.1257/ jel.37.4.1661. Clarida, R., Gali, J. & Gertler, M. (2000). Monetary policy rules and macroeconomic stability: Evidence and some theory. The Quarterly Journal of Economics, 115 (1), 147-180. doi:10.1162/003355300554692. Dennis, R. (2005). Specifying and estimating new Keynesian models with instrument rules and optimal monetary policies. Federal Reserve Bank of San Francisco Working Papers, No: 2004-17.. Dixon, H.D. (2008). New Keynesian macroeconomics. In S. N. Durlauf & L. E. Blume (Eds.), The new Palgrave dictionary of economics online (2nd ed.). Palgrave Macmillan. <http://www.dictionaryofeconomics.com/article?id=pde2008_N000166> doi:1 0.1057/9780230226203.1184. Gali, J. (2002). New perspectives on monetary policy, inflation, and the business cycle. NBER Working Paper, No. 8767.. Gali, J. (2009). The new Keynesian approach to monetary policy analysis: Lessons and new directions. http://www.crei.cat/people/gali/jg09cfs.pdf (25.02.2010).. Giordani, P. (2004). Evaluating new Keynesian models of a small open economy. Oxford Bulletin of Economics and Statistics, 66 (Supplement), 713-733. doi: 10.1111/ j.1468-0084.2004.099_1.x doi:10.1111/j.1468-0084.2004.099_1.x. Goodfriend, M. & King, R. G. (1997). The new neoclassical synthesis and the. 55.

(20) İktisat İşletme ve Finans 26 (305) Ağustos / August 2011. .54. .44. ], T. ar. ih:. 12. /06. +0 30 0. 7:0 5. 09 :4. /20. 20. se. ive. rs. ite. si] , IP. :[. 19. 3.1. 40. ge n:. [Ç u. ku. ro. va. Ün. il 56. İnd ire. B. İndiren: [Çukurova Üniversitesi], IP: [193.140.54.44], Tarih: 12/06/2020 09:47:05 +0300. l. role of monetary policy. <http://www.richmondfed.org/publications/research/working_ papers/1998/pdf/wp98-5.pdf > (15.01.2009).. Gordon, R.J. (1990). What is new Keynesian economics? Journal of Economic Literature, 38 (3), 115-1171.. Henry, S. G. B. & Pagan, A. R. (2004). The econometrics of the new Keynesian policy model: Introduction. Oxford Bulletin of Economics and Statistics. 66 (Supplement), 581-607. doi: 10.1111/j.1468-0084.2004.00094.x doi:10.1111/j.1468-0084.2004.00094.x. Hansen, B.E. (1997). Approximate asymptotic P values for structural-change tests. Journal of Business and Economic Statistics, 15(1), 60-67. doi:10.2307/1392074. Hansen, B.E. (2000). Testing for structural change in conditional models. Journal of Econometrics, 97(1), 93-115. doi:10.1016/S0304-4076(99)00068-8. doi:10.1016/S03044076(99)00068-8. Huseynov, S. (2010). An estimated DSGE model for Turkey with a monetary regime change. (Master thesis). Central European University, Budapest/Hungary. <http:// www.etd.ceu.hu/2010/huseynov_samir.pdf> (25.01.2011).. Linde, J. (2005). Estimating new Keynesian Phillips curves: A full information maximum likelihood approach. Sveriges Riskbank Working Paper Series, No: 129.. Mankiw, G. N. (1990). A quick refresher course in macroeconomics. Journal of Economic Literature, 27 (4), 1645-1660.. Snowdon, B., Vane, H. & Wynarczyk, P. (1995). A modern guide to macroeconomics: An introduction to competing schools of thought. Cheltenham, UK: Edward Elgar Press.. Ünalmış, D., Ünalmış, İ. & Ünsal, D.F. (2010). On the sources of oil price fluctuations. TCMB Working Paper Series, No:10/05.. Yıldırım, E., Lopcu K., Çakmaklı, S. & Özkan, F.Ö. (2010). Yeni Keynesyen makro ekonomik bir model: Türkiye uygulaması. Ege Academic Review, 10 (4), 1269-1277.. Zivot, E. & Andrews, D. W. K. (1992). Further evidence on the great crash, the oil price shock and the unit root hypothesis. Journal of Business and Economic Statistic, 10 (1), 251-270. doi:10.2307/1391541.

(21)

Referanslar

Benzer Belgeler

In a magnetic particle imaging (MPI) scanner, utilizing a tunable gradiometer receive coil can aid in achieving greater degree of decoupling of direct feedthrough signal.. However,

with significant changes in the direction of trade; the significant and negative sign of the interactive dummy variable cudeu in the export equation and its significant and

Although military statistics and records were highly developed in Russia compared with the Ottoman Empire, the number of Russian troops during the Crimean War

特別企劃 文◎胸腔內科 劉文德醫師 睡眠障礙影響健康,整合團隊提供個別化服務

When patients with allergic rhinitis stop using their drugs to control their symptoms, especially during the pollen period, their complaint of sneezing will be increased, and if

Hastalar 4 farklı yaş grubuna göre sınıflandırıldığında; en sık başvuru yapan hasta grubunun adölesan dönemde, en az başvuru yapan hasta grubunun ise okul ön- cesi dönemde

Kalça ekleminin harabiyeti ile sonuçlanan ve hastalarda günlük hayatlarını olumsuz etkileyen şiddetli ağrıya neden olan hastalıkların konservatif tedavilerden yarar

Ayrıca, siyasi partilerin seçim müziklerinde seçim dönemi boyunca liderlerinin dile getirdiği konular (tek bayrak, tek vatan, tek millet, tek devlet, Türkiye vakti, artık tamam,