2021 Msü Ye Kaç Kişi Girdi

1 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES

2 УДК ББК I 73 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Editors: Selahattin Sarı Jusup Primbaev Alp H. Gencer Ainura Turdalieva Batuhan Tufaner Copyright Eurasian Economists Association & Kyrgyzstan-Turkey Manas University ALL RIGHTS RESERVED. This book contains material protected under International and Federal Copyright Laws and Treaties. Any unauthorized reprint or use of this material is prohibited. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system without express written permission from the publisher. ISBN: B p. Typesetting: Alp H. Gencer & Hülya Üzümcü Print: Turar I УДК ISBN ББК

3 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Bishkek - Kyrgyzstan October

4 iv INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES FOREWORD We are proud to have completed the nith International Conference on Eurasian Economies, held in Bishkek, Kyrgyzstan with the cooperation of Kyrgyzstan-Turkey Manas University and Eurasian Economists Association on October abstracts were submitted to the conference from 5 countries, and in the end 57 accepted papers from 4 countries were presented at the conference. Many academicians attended the conference even though they had no papers to present, so the total attendance was around academicians. The fact that the rectors Manas University, Vice Minister of Department of Agriculture, Vice Rectors, as well as deans of many faculties, professors and associate professors, people from the financial sector and executive boards attended the conference show that interest in the conference is high and continual. We would like to extend our deepest appreciation to all participants and sponsors who have trusted and supported the conference s organizing committee. As in previous years, presented papers are collected into this proceedings book. A considerable amount of time and effort is put into evaluating and formatting the papers for the book. Each paper is evaluated by at least two judges before being accepted. We would like to thank our judges, scientific and organization committee members who contributed their valuable time and support to evaluating every paper. Additionally, we would like to extend our gratitude to Dr. Alp Gencer, who helped to design our web platform with great patience despite all of the difficulties and prepared the book of proceedings for publication. Our appreciation is also extended all members of Kyrgyzstan-Turkey Manas University who put every effort in the perfect organization, starting with the rector Prof. Dr. Sebahattin Balcı, dean of Faculty of Economic and Administrative Sciences, seafoodplus.info Hakan Çetintaş, the faculty and students of the department of economics. They have done a wonderful job in making sure that the entire organization runs smoothly, press participance was ensured and organized a cultural feast for all attendees of the conference dinner. We would like to extend our greatest appreciation for the continuous support of International Conference on Eurasian Economies by Beykent University s Board of Trustees Chairman and members, Rector Prof. Dr. Mehmet Emin Karahan, the Secretary General Murat Süslü and Public Relations Department, as well as the administrative staff. Finally, our greatest appreciation goes to all of the academicians, researchers and participants whose contributions have made this conference a success. We hope and believe that the International Eurasian Conference has made important contributions to the Eurasian region and scientific community at large. We hope to see you at the 10 th International Conference on Eurasian Economies. Prof. Dr. Selahattin Sarı and Prof. Dr. Cusup Primbayev Conference Co-Chairs

5 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES v ÖNSÖZ Uluslararası Avrasya Ekonomileri Konferansı nın dokuzuncusu; Beykent Üniversitesi, Kırgızistan- Türkiye Manas Üniversitesi ve Avrasya Ekonomistler Derneği nin ortak organizasyonuyla Ekim tarihlerinde Bişkek-Kırgızistan da gerçekleştirilmiştir. Konferansa özet gönderim aşamasında 5 ülkeden 74 başvuru olmuştur. Özet başvurularının tamamına yakını bildiriye dönüşmüş 57 bildiri konferansta sunulmuştur. Diğer taraftan konferansta bildirisi olmayan çok sayıda öğretim üyesi dileyici olarak iştirak etmiş ve konferans yaklaşık civarında katılımcının iştiraki ile gerçekleşmiştir. Katılımın yüksekliği konferansa olan ilginin büyüklüğünü göstermektedir. Konferansa Manas Üniversitesi Rektörü Prof. Dr. Sebahattin Balcı, Tarım Bakan Yardımcısı, Manas Üniversitesi Rektör Vekili Prof. Dr. Asılbek Kulmirzayev, Rektör Yardımcıları ve gerek Türkiye den gerekse Kırgızistan dan çok sayıda dekanın katılması, profesör ve doçentlerin ağırlıklı olması finans sektörü ve üst kurullardan katılımlar konferansın bilim dünyasındaki yerini ve önemini bir kez daha öne çıkarmıştır. Konferans düzenleme komitesine olan güven ve desteklerini konferansa iştirakleriyle şereflendiren tüm katılımcılara öncelikle teşekkürlerimizi arz ediyoruz. Konferansta sunulan bildiriler elinizde bulunan bu kitapta toplanmıştır. Gerek bildirilerin değerlendirilmesi aşamasında, gerekse kitap haline getirilmesi aşamasında ciddi emek sarf edilmektedir. Her bildiri en az iki hakem tarafından değerlendirilmektedir. Bu değerlendirmelere zaman ayıran ve destek veren hakemlerimize, bilim kurulu üyelerimize, konferans hazırlıklarının başlamasından itibaren çok yoğun mesai harcayan organizasyon komitesindeki arkadaşlarımıza, web sayfası dizaynında her türlü müdahaleye rağmen sabırla işlemleri yerine getiren Dr. Alp Gencer e teşekkürlerimizi sunarız. Uluslararası Avrasya Ekonomileri Konferansı nın dokuzuncusu eksiksiz ve mükemmel olarak gerçekleştirilmesinde, planlama aşamasından konferansın sonlandırılmasına kadar her türlü desteği esirgemeden veren Manas Üniversitesi Rektörü Prof. Dr. Sebahattin Balcı, İktisadi ve İdari Bilimler Fakültesi Dekanı Prof. Dr. Hakan Çetintaş, İktisadi ve İdari Bilimler Fakültesinin İktisat Bölümünün çok değerli öğretim elemanları ve destek veren diğer bölümlerden öğretim elemanları, öğrencilerine sonsuz teşekkürler. Basın koordinasyonu üst düzeyde gerçekleştirilmiş, gala yemeğinde yer alan sanatçılar ve folklor gösterisi tam bir ziyafet olmuştur. Misafirlere Türk Dünyasının esintilerini taşımış ve yaşatmışlardır. Emeği geçen tüm yönetici ve personele teşekkürler. Bu konferansta da Beykent Üniversitesi Mütevelli Heyet Başkan ve üyeleri, Üniversite Rektörü Prof. Dr. Mehmet Emin Karahan, Genel Sekreter Murat Süslü ve başta halkla ilişkiler olmak üzere tüm idari birimler personeli destek vermişlerdir; kendilerine sonsuz teşekkürler. Konferansın organizasyonunda birlikte hareket ettiğimiz Doğuş Üniversitesi ve Avrasya Ekonomistler Derneği nin tüm yöneticilerine teşekkürler. En büyük teşekkürümüz elbette ki konferansın oluşmasını sağlayan akademisyen, araştırmacı ve katılımcılaradır. Uluslararası Avrasya Ekonomileri Konferansının dokuzuncusunun başta Avrasya ülkeleri olmak üzere bilim dünyasına katkı sağlayacağı inanç ve dileğiyle, Uluslararası Avrasya Ekonomileri Konferansı nda buluşmak üzere. Prof. Dr. Selahattin Sarı ve Prof. Dr. Cusup Primbayev Organizasyon Kurulu Başkanları

6 vi INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES AÇILIŞ KONUŞMASI Prof. Dr. Sebahattin Balcı Marifet iltifata tâbidir. Ne kadar çok çalışırsanız çalışın, yaptığınız işin bir yansıması olmazsa, bunlar çok fazla değer ifade etmiyor demektir. Katılımcı sayısına bakıldığında, konferansın çok ilgi çektiği görülüyor. Avrasya ekonomileri dendiği zaman hem çok değişken hem de çok dinamik bir bölgeden söz ediyoruz. Bir taraftan değişim ve dönüşüm sürerken, bir taraftan da küresel ekonomiyle uyum ve aynı zamanda rekabet içerisinde olmak mecburiyeti var. Hepsini bir arada götürüp ayakta kalmak ve ileri gitmek zorunluluğu var. Sonuç itibariyle biz şunu yapmak istiyoruz: Avrasya da bir ekonomi bölgesi oluşturup bu ekonominin küresel ekonomideki yerini almasını istiyoruz. Küresel rekabette bölgesel ekonomiler çok büyük bir önem ifade etmeye başladı; çünkü eğer sadece küresel ekonomide meseleye bakarsanız o zaman tek kutuplu bir ekonomi ortaya çıkıyor. Başını ABD nin çektiği ve her şeyi onun düzenlediği bir ekonomi ortaya çıkıyor. Diğerleri de ona hizmet etmek ve onun maksatlarına yardım etmek durumunda kalıyor. İşte bunu biraz olsun dengelemek adına bölgesel ekonomiler büyük önem ifade ediyor. Biz de geniş bir coğrafya olan Avrasya da bulunanlar olarak bunu bilmeliyiz. Burada özellikle Türkler bulunuyor. Buna çok dikkat etmek lazım. yüzyılda hak ettiği değeri alabilmek, zenginliğe ulaşabilmek, dünyadaki karar organlarında söz sahibi olabilmek için Türk Dünyası olarak ekonomimizin güçlü olması gerekiyor. civarında ülkenin olduğu Birleşmiş Milletler de kararları 5 ülke alıyor; çünkü bu 5 ülkenin gücü ekonomisinden geliyor. Biz Türkiye olarak Türk dünyasının hem büyük bir ekonomisi hem de büyük bir sorumluluk taşıyan ülkesiyiz; çünkü bütün Avrasya da eğer bir gelecek olacaksa bu Türkiye nin taşıyacağı omurgada yürümek durumunda olan bir gelecek olacak. Hem ekonomisi hem nüfusu hem insan kaynakları hem bulunduğu coğrafya hem de dünyayla entegrasyonu bunu bize gösteriyor. Ancak tek başına da olmuyor. O halde ekonomik hareketi, bütün Avrasya daki kardeş topluluk ve ülkelerin birlikte kalkınacağı bir ekonomi olarak algılamak gerekiyor. Bilime baktığımız zaman ilk 10 ülke şu anda üretilen bilimin yüzde 90 ını üretiyor. Gerçek bu. Geriye kalan ülke de yüzde 10 u üretiyor. Pozisyonlarımızı buna göre belirlememiz gerekiyor. Buna göre hareket etmemiz gerekiyor. O bakımdan güç birliği şart. Tek başımıza böyle bir rekabet ortamında varlığımızı ifade etmemiz çok zor. Ekonomi de bunun tam omurgasını oluşturuyor. Bu bakımdan bütün tecrübelerimizi birleştirmemiz gerekiyor. Bu konferanslarda genel konuşmalar yerine konjonktürün ön gördüğü, vizyoner, temalı konuşmalara, temalı çalışmalara doğru geçmek gerekiyor. Bir de bu salonlara iş dünyasını, iş adamlarını getirmek gerekiyor. O zaman gerçekten ses, tam adresini bulmuş olacak. Birbirimizden gerçek anlamda öğrenmeye başlayacağız. Birlikte daha iyisini yapmaya gayret edeceğiz. Bizim bölgemizde hangi sektörlerde fırsatlar var bunu bilmek gerek. Sınırlı olan enerjimizi ve gücümüzü her tarafa yayarsak hiçbir yerde etki oluşturamıyoruz. Oysa odaklandığımız bir yerden ses getirmemiz mümkün. Bölgemizde ekonomi açısından baktığımız zaman fırsatları ve tehditleri görüp buna göre hareket etmeli ve yoğunlaşacağımız konuları belirlemeliyiz. Böyle konferansların da buna ışık tutması gerekiyor. Eğitim programlarımızı böyle konferanslar sonucunda gerekirse yeniden gözden geçirmeliyiz. Gerçekleştirdiğimiz konferanslarda elbette dünya ekonomisini de ele almalıyız; ama kendi bölgemiz ve ekonomilerimizle ilgili olarak çok iyi çalışmalar ortaya çıkarmamız gerekiyor. Gençleri o şekilde donatarak yetiştirmemiz gerekiyor. Bölgesini çok iyi anlamadan dünya ekonomisini anlatan bir mezun yetiştirmenin bölgeye çok fayda olmaz. Kırgızistan da 5 yıl öğrenim gördükten sonra Kırgızistan ekonomisini ve Orta Asya ekonomilerini anlamadan mezun olan bir mezunun başarılı bir mezun olduğu söylenemez. Onlar çevrelerini tanımalı. Biz bu şekilde bir eğitimle öğrenci yetiştirirsek onlar sahanın adamı olurlar. Gittikleri uluslararası alanda da o sahayı en iyi şekilde sunabilir, savunabilir, fırsatları değerlendirebilir ve kendilerini öne çıkarabilir. Bu konferanslardan böyle sonuçlar çıkaralım. Kırgızistan nüfusunun yüzde 65 inin kırsal alanda yaşaması ve tarımın yıllık ekonomik hasılaya olan katkısının yüzde 54 olması verileri üzerinden hareketle bir değerlendirme yapmak önemli. Kırgızistan ın gerçek anlamda tarım üzerinden bir ekonomi ile kalkınmasını öngörmek gerekiyor. Biz Ziraat Fakültemizi kurarken bunları dikkate aldık. Üstelik beslenmenin, gıdanın stratejik bir alan olduğunu görmek gerek. Yeraltı kaynakları bir süre sonra tükenecek olsa da insanların temel ihtiyaçları olan beslenme ve gıda ihtiyacı devam edecektir. Gıdanın da geldiği yer topraktır. Bunu bilen ve iyi değerlendirenler gelecekte çok stratejik bir önem ifade edeceklerdir. Kırgızistan hayvancılıkta son derece önemli bir potansiyele sahip. Sovyetler Birliği döneminde 20 milyon baş hayvanın yaşadığı Kırgızistan da bugün 6,5 milyon hayvan yaşıyor. Demek ki daha çok kazanılacak zenginlik var.

7 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES vii Kaybedilenlerin kazanılması gerek. Bu noktada fırsatlar ekonomisini yakalamamız gerek. Enerji, su, bunların koridorları. Bütün bunları tartışıp hem iş dünyasına hem Devlet yöneticilerine bir taban bilgi ve yol gösterici sonuçlar çıkarabiliyorsak düzenlenen konferanslar amacına ulaşmış demektir. Yapılmasına da her zaman devam etmek gerekmektedir. İşte şimdi geldiğimiz noktada, inşallah bundan sonra da bu çizgide devam ederek, bütün kesimlerin faydasına olan, ihtiyaç duyulan ve istenilen çalışmalar yaparız. Bu duygu ve düşüncelerle bu tebliğ kitabında basılan makalelerin bilim dünyasına, kamu ve özel sektör mensuplarına ve okuyuculara yararlı olmasını diliyor, konferansı düzenleyen ve emeği geçenlere içten teşekkürlerimi sunuyorum. Prof. Dr. Sebahattin Balcı Kırgızistan-Türkiye Manas Üniversitesi Rektörü

8 viii INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES ADVISORY BOARD Prof. Dr. Sebahattin Balcı (Rector, Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Prof. Dr. Mehmet Emin Karahan (Rector, Beykent University, Turkey) Prof. Dr. Hakan Çetintaş (Dean, Faculty of Economic and Administrative Sciences, Kyrgyzstan- Turkey Manas University, Kyrgyzstan) Prof. Dr. Turgut Özkan (Dean, Faculty of Economic and Administrative Sciences, Beykent University, Turkey) ORGANIZING COMMITTEE Prof. Dr. Selahattin Sarı (Beykent University, Turkey) Prof. Dr. Jusup Pirimbaev (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Asst. Prof. Dr. Aynura Turdalieva (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Ph.D. Candidate Mustafa Batuhan Tufaner (Beykent University, Turkey) Asst. Prof. Dr. Sıtkı Sönmezer (Beykent University, Turkey) Assoc. Prof. Dr. İlyas Sözen (İstanbul Kavram MYO, Turkey) Assoc. Prof. Dr. Ayşen Hiç Gencer (Istanbul Aydın University, Turkey) Assoc. Prof. Dr. Damira Japarova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Assoc. Prof. Dr. Metin Bayrak (Kyrgyzstan-Turkey Manas University, Turkey) Asst. Prof. Dr. Junus Ganiev (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Asst. Prof. Dr. Raziya Abdiyeva (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Asst. Prof. Dr. Damira Baigonushova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Asst. Prof. Dr. Zamira Oskonbaeva (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Ph.D. Candidate Galip Afsin Ravanoglu (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Dr. Burulcha Sulaimanova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Dr. Alp Gencer (Germany) Mr. Alpin Gencer (Germany) Mrs. Hülya Üzümcü (Turkey) WEB & PRINT DESIGN

9 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES ix SCIENTIFIC BOARD Prof. Dr. Ahmet Ay (Selçuk University, Turkey) Prof. Dr. Harun Bal (Çukurova University, Turkey) Prof. Dr. Mehmet Balcılar (Eastern Mediterranean University, Cyprus) Prof. Dr. Viktor Barhatov (Chelyabinsk State University, Russia) Prof. Dr. Tansu Barker (Brock University, Canada) Prof. Dr. Ersan Bocutoğlu (Avrasya University, Turkey) Prof. Dr. Cihan Çobanoğlu (University of South Florida, United States of America) Prof. Dr. Mustafa Kemal Değer (Karadeniz Technical University, Turkey) Prof. Dr. Ljubomir Drakulevski (Ss. Cyril and Methodius University, Macedonia) Prof. Dr. Hüseyin Avni Egeli (Dokuz Eylül University, Turkey) Prof. Dr. Ömer Selçuk Emsen (Atatürk University, Turkey) Prof. Dr. Cevat Gerni (Doğuş University, Turkey) Prof. Dr. Ahmet İncekara (İstanbul University, Turkey) Prof. Dr. Danijela Jacimovic (University of Montenegro, Montenegro) Prof. Dr. Gani Kaliev (Academy of Agricultural Sciences, Kazakhstan) Prof. Dr. S. Rıdvan Karluk (Antalya AKEV University, Turkey) Prof. Dr. Nuriddin Kayumov (Tajikistan Academy of Sciences, Tajikistan) Prof. Dr. Merim Koichueva (Kyrgyz-Russian Slavic University, Kyrgyzstan) Prof. Dr. Recep Kök (Dokuz Eylül University, Turkey) Assoc. Prof. Dr. Igor Kokh (Kazan Federal University, Russia) Prof. Dr. Turar Koychuev (National Academy of Sciences, Kyrgyzstan) Prof. Dr. Fevzi Okumuş (University of Central Florida, United States of America) Prof. Dr. Zekai Özdemir (Istanbul University, Turkey) Prof. Dr. Jusup Pirimbaev (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Prof. Dr. Tatiana Pyshkina (Academy of Economic Studies of Moldova (ASEM), Moldova) Prof. Dr. Orazaly Sabden (Institute of Economics, Kazakhstan) Prof. Dr. Mehmet Sarıışık (Sakarya University, Turkey) Prof. Dr. Euphrasia Susy Suhendra (Gunadarma University, Indonesia) Prof. Dr. Ferenec Szavai (Kaposvar University, Hungary) Prof. Dr. Muammer Tekeoğlu (Çukurova University, Turkey) Prof. Dr. Mustafa Topaloğlu (Özyeğin University, Turkey) Prof. Dr. Khodjamakhmad Umarov (Institute of Economics, Tajikistan) Prof. Dr. Bahadir Umurzakov (G.V. Plekhanov Russian University of Economics in Tashkent, Uzbekistan)

10 x INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Contents OPENING SESSION Session Chair: Prof. Dr. Selahattin Sarı # Bir Ekonomik Avrasya Öyküsü: Arakan Eyaleti Prof. Dr. Ersan Bocutoğlu (Avrasya University, Turkey) * SESSION 1A: Microeconomics Session Chair: Prof. Dr. Erdoğan Taşkın # Effect of Price and Brand on Purchase Decision An Application on Turkish Smart Phone Consumers 2 - Ph.D. Candidate Barış Batuhan Geçit (Beykent University, Turkey) * - Asst. Prof. Dr. Murad Kayacan (Beykent University, Turkey) # Gender Differences, Risk Attitude and Entrepreneurship in Kyrgyzstan Asst. Prof. Dr. Raziya Abdiyeva (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Dr. Burulcha Sulaimanova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * - Asst. Prof. Dr. Kamalbek Karymshakov (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) # Matrix Management Approach and a Survey Research in the Industry Asst. Prof. Dr. Ceren Gül Artuner (Beykent University, Turkey) - Prof. Dr. Erdoğan Taşkın (Beykent University, Turkey) * SESSION 2A: Growth & Development Session Chair: Prof. Dr. Ersan Bocutoğlu # Seasonal Cointegration Approach on Expenditure Based Gross Domestic Product and Its Some Sub- Components for Turkey Prof. Dr. Mehmet Özmen (Çukurova University, Turkey) * - Ph.D. Candidate Sera Şanlı (Çukurova University, Turkey) # On The Purchasing Power Parity in Turkey: The Role of Structural Changes Assoc. Prof. Dr. Şaban Nazlıoğlu (Pamukkale University, Turkey) - Ph.D. Candidate Çağın Karul (Pamukkale University, Turkey) - Ph.D. Candidate Ahmet Koncak (Abant İzzet Baysal University, Turkey) - Assoc. Prof. Dr. İlhan Küçükkaplan (Pamukkale University, Turkey) * # Export-led Growth Hypothesis in MINT Countries: A Panel Cointegration Analysis Prof. Dr. Harun Bal (Çukurova University, Turkey) * - Ph.D. Candidate Shahanara Basher (Çukurova University, Turkey) - Ph.D. Candidate Abdulla Hil Mamun (Çukurova University, Turkey) - Ph.D. Candidate Emrah Akça (Çukurova University, Turkey) # Relationship Between Trade Openness, Capital Formation and Economic Growth: A Panel Data Analysis for African Countries Prof. Dr. Ahmet Ay (Selçuk University, Turkey) - Asst. Prof. Dr. Fahri Kurşunel (Selçuk University, Turkey) * - Ph.D. Candidate Mahamane Moutari Abdou Baoua (Selçuk University, Turkey)

11 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES xi SESSION 3A: Sectoral Analysis Session Chair: Prof. Dr. Cevat Gerni # The Impact of Tax Apportionment on Fiscal Equalization: A Case Study of Turkish Metropolitan Municipalities Prof. Dr. Naci Tolga Saruç (İstanbul University, Turkey) * - Prof. Dr. İsa Sağbaş (Afyon Kocatepe University, Turkey) - Dr. Recep Yorulmaz (Afyon Kocatepe University, Turkey) # Natural Resources Revenue, Fiscal Policy and Economic Growth: Panel Data Analysis for Sub-Saharan Africa Countries Asst. Prof. Dr. Savaş Çevik (Selçuk University, Turkey) * - Prof. Dr. Ahmet Ay (Selçuk University, Turkey) - Ph.D. Candidate Mahamane Moutari Abdou Baoua (Selçuk University, Turkey) # The Education-Job Mismatch Determinants Among Youth of Kyrgyzstan Asst. Prof. Dr. Kamalbek Karymshakov (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Dr. Burulcha Sulaimanova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * # The Public Policy in Agricultural Product Markets and Effectiveness of Regulations Prof. Dr. M. Okan Taşar (Selçuk University, Turkey) * SESSION 4A: Развитие I Session Chair: Prof. Dr. Jusup Pirimbaev # Найти Своё Лицо Prof. Dr. Turar Koychuev (National Academy of Sciences, Kyrgyzstan) * # Глобальные Тренды Современного Мирового Экономического Развития Prof. Dr. Kaisyn Khubiev (Lomonosov Moscow State University (MSU), Russia) * - Assoc. Prof. Dr. Turusbek Asanov (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Dr. Marat Kudaikulov (Kyrgyz-Russian Slavic University, Kyrgyzstan) # Образование и экономический рост: на примере Кыргызстана Asst. Prof. Dr. Raziya Abdiyeva (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * - Asst. Prof. Dr. Damira Baigonushova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) # Управление Региональным Развитием Кыргызской Республики В Условиях Интеграции Asel Dzhailova (Kyrgyz Economic University, Kyrgyzstan) * SESSION 5A: Развитие II Session Chair: Prof. Dr. Turar Koychuev # Источники Экономического Роста В Кыргызстане Asst. Prof. Dr. Aynura Turdalieva (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * # Кыргызстанда Акчалай Которуулардын Экономикалык Өсүүгө Таасири Dr. Burulcha Sulaimanova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Daniyar Jasoolov (Public Fund AVEP, Kyrgyzstan) *

12 xii INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES SESSION 1B: Büyüme I Session Chair: Prof. Dr. İsmail Bircan # Petrol Fiyatlarının Düşüşünde Shale (Şeyl) Gazın Etkisi, Geleceği ve Türkiye nin Shale Gaz Potansiyeli Assoc. Prof. Dr. Kamil Uslu (Kavram MYO, Turkey) * # İhracata Dayalı Büyümenin Geçerliliği Üzerine Ampirik Bir Araştırma: Türkiye den Bulgular Prof. Dr. Harun Bal (Çukurova University, Turkey) * - Ph.D. Candidate Emrah Akça (Çukurova University, Turkey) # Kuznets Eğrisi ile Ekonomik ve Siyasal Liberalizasyon Uygulamaları Arasındaki İlişkiler Prof. Dr. Murat Nişancı (Erzincan University, Turkey) * - Asst. Prof. Dr. Ahmet Fatih Aydemir (Atatürk University, Turkey) - Ph.D. Candidate Bengü Tosun (Atatürk University, Turkey) - Prof. Dr. Ömer Selçuk Emsen (Atatürk University, Turkey) # Doğrudan Yabancı Sermaye Yatırımları ile Sanayileşme İlişkisi: Türkiye Üzerine Bir Uygulama Prof. Dr. Harun Bal (Çukurova University, Turkey) * - Asst. Prof. Dr. Erhan İşcan (Çukurova University, Turkey) - Ph.D. Candidate Ahmet Kardaşlar (Çukurova University, Turkey) SESSION 2B: Büyüme II Session Chair: Prof. Dr. M. Okan Taşar # İhracat ve İthalatın Ekonomik Büyüme Üzerine Etkileri: Merkezi Asya Ekonomileri Örneği Prof. Dr. Selahattin Sarı (Beykent University, Turkey) - Prof. Dr. Cevat Gerni (Doğuş University, Turkey) - Asst. Prof. Dr. Çağatay Karaköy (Cumhuriyet University, Turkey) - Prof. Dr. Ömer Selçuk Emsen (Atatürk University, Turkey) * # Kırgızistan da İşçi Gelirleri ile Büyümedeki Dalgalanmalar Arasındaki İlişki Prof. Dr. Hakan Çetintaş (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Asst. Prof. Dr. Damira Baigonushova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * # Ev Sahipliği ve İşsizlik: Oswald Hipotezinin Türkiye için Test Edilmesi Asst. Prof. Dr. Mahmut Erdoğan (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * - Dr. Dastan Aseinov (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) # Kentsel Dönüşüm İçin Bir Finansman Modeli: Gayrimenkul Sertifikası Dr. Ayça Doğaner (İTO Meslek Komiteleri Koordinatörlüğü, Turkey) * - Prof. Dr. Zekai Özdemir (Istanbul University, Turkey) SESSION 3B: Sektörel Analizler Session Chair: Prof. Dr. Şaban Kayıhan # Öğrencilerin Otel İşletmelerindeki İş Etiğine Yönelik Algıları Muharrem Kaya (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Dr. Azamat Maksudunov (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) *

13 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES xiii # İklim Değişikliğinin Gıda Fiyatları Üzerine Etkisi: OECD Örneği Prof. Dr. Muammer Tekeoğlu (Çukurova University, Turkey) * - Asst. Prof. Dr. Hakkı Çiftçi (Çukurova University, Turkey) - Asst. Prof. Dr. Erhan İşcan (Çukurova University, Turkey) - Ph.D. Candidate Duygu Serin (Çukurova University, Turkey) # Karbon Emisyonu ve Uluslararası Ticaret İlişkisi Prof. Dr. Harun Bal (Çukurova University, Turkey) - Asst. Prof. Dr. Erhan İşcan (Çukurova University, Turkey) * - Ph.D. Candidate Duygu Serin (Çukurova University, Turkey) - Selçuk Loğoğlu (Çukurova University, Turkey) # Banka ve Müşteri Sırlarının Ceza Hukuku Tarafından Korunması Prof. Dr. Çetin Arslan (Hacettepe University, Turkey) * - Ph.D. Candidate Didar Özdemir (Hacettepe University, Turkey) SESSION 4B: Региональные исследования Session Chair: Prof. Dr. Dzhumabek Dzhailov # Учет Потерь При Попроцессном Калькулировании Производственных Затрат Prof. Dr. Obolbek Kurmanbekov (Kyrgyz Russian Slavic University, Kyrgyzstan) * # Роль И Значение Диагностики И Оценки Состояния Автомобильных Дорог Для Их Воспроизводства Jannat Ysyrailova (Kyrgyz State University of Construction, Transport and Architecture, Kyrgyzstan) * - Prof. Dr. Anara Sherbekova (Kyrgyz State University of Construction, Transport and Architecture, Kyrgyzstan) # Рейтинг Социально-Экономического Положения Регионов Как Основа Их Инвестиционной Привлекательности Prof. Dr. Damira Bektenova (Kyrgyz-Russian Slavic University, Kyrgyzstan) * # Региональные Аспекты Инвестиционных Процессов В Кыргызской Республике Prof. Dr. Anara Sherbekova (Kyrgyz State University of Construction, Transport and Architecture, Kyrgyzstan) - Dr. Sabina Esenbekova (Kyrgyz State University of Construction, Transport and Architecture, Kyrgyzstan) * SESSION 5B: Секторальный анализ Session Chair: Prof. Dr. Kaisyn Khubiev # Использование Финансовых Ресурсов В Государственных Организациях Здравоохранения В Кыргызской Республике Assoc. Prof. Dr. Damira Japarova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * - Prof. Dr. Anara Kamalova (Kyrgyz National University, Kyrgyzstan) # Өткөөл Экономикада Социалдык Нормалардын Салык Маданиятына Тийгизген Таасири: Кыргызтан Мисалында Asst. Prof. Dr. Raziya Abdiyeva (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * # Проблемы И Перспективы Интеграции Аграрного Сектора Кыргызстана В ЕАЭС Prof. Dr. Dzhumabek Dzhailov (Ministry of Agriculture, Kyrgyzstan) * - Mardalieva Leila (Kyrgyzstan)

14 xiv INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES # Кыргызстандын Айыл-Чарба Тармагын Өнүктүрүү Маселелери Asst. Prof. Dr. Damira Baigonushova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * - Saikal Otorova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Asst. Prof. Dr. Junus Ganiev (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Prof. Dr. Jusup Pirimbaev (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) SESSION 1C: Bölgesel Çalışmalar Session Chair: Prof. Dr. Muammer Tekeoğlu # Avrasya Ekonomik Birliği nin Kırgızistan ile Türkiye İlişkilerine Etkileri Ph.D. Candidate Ömer Faruk Karaman (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * # Çin in Bir Kuşak-Bir Yol Projesinin Ekonomik ve Jeopolitik Sonuçları Üzerine Düşünceler Prof. Dr. Ersan Bocutoğlu (Avrasya University, Turkey) * # Ekonomik Entegrasyon Teorisi Çerçevesinde Türkiye İçin Avrupa Birliği&#;ne Bir Alternatif: Avrasya Gümrük Birliği Ph.D. Candidate Mustafa Batuhan Tufaner (Beykent University, Turkey) * - Ph.D. Candidate Hasan Boztoprak (Beykent University, Turkey) - Assoc. Prof. Dr. İlyas Sözen (İstanbul Kavram MYO, Turkey) # Körfez İşbirliği Konseyi Ülkelerinde Doğal Kaynak Laneti Hipotezi Geçerliliğinin İncelenmesi Ph.D. Candidate Mehmet Akif Destek (Gaziantep University, Turkey) - Ph.D. Candidate Müge Manga (Çukurova University, Turkey) - Assoc. Prof. Dr. Neşe Algan (Çukurova University, Turkey) * SESSION 2C: Mikroekonomi Session Chair: Prof. Dr. Ahmet Ay # Avrasya Ülkeleri Üniversiteleri Arasında Ar-Ge ve İnovasyon İşbirlikleri ile Sürdürülebilir Kalkınmanın Desteklenmesi Prof. Dr. İsmail Bircan (Atılım University, Turkey) * # Örgüt Kültürü, Örgütsel Öğrenme ve Yenilikçilik İlişkisi Üzerine Bir Araştırma: Bir Üniversite Hastanesinde Anket Çalışması Vasfi Kendir (İstanbul University, Turkey) - Assoc. Prof. Dr. Bahaddin Sinsoysal (Beykent University, Turkey) - Ph.D. Candidate Hasan Boztoprak (Beykent University, Turkey) * # Lojistik Performans Endeksi Kapsamında Orta Asya Türk Cumhuriyetleri nin Lojistik Performanslarının Analizi Ph.D. Candidate Bayram Şirin (Beykent University, Turkey) * - Asst. Prof. Dr. Hakan Emanet (Beykent University, Turkey) # Otomotiv Endüstrisindeki Muhtemel Dönüşümün Ekonomik Boyutu: Türkiye Örneği Asst. Prof. Dr. Osman Simav (Beykent University, Turkey) * - Asst. Prof. Dr. Ayfer Ustabaş (Beykent University, Turkey)

15 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES xv SESSION 3C: Finans I Session Chair: Prof. Dr. Ömer Selçuk Emsen # Konut Kredi Hacminin Gelişimi ile Makroekonomik Değerler ve Finansal Getirilerin İlişkisi Asst. Prof. Dr. Sıtkı Sönmezer (Beykent University, Turkey) * - Yusuf Pala (Kheiron Reklam Organizasyon Mimarlık Tic. Ltd. Şti., Turkey) # Monte Carlo Metodu ve Yeni Zıplama ve Daralma Sürecinin BIST (Borsa İstanbul) Getirileri Uygulaması Assoc. Prof. Dr. Kutluk Kağan Sümer (İstanbul University, Turkey) * # Konut Sektöründe Finansman Modellerinin Analizi: Türkiye Uygulaması Asst. Prof. Dr. Gülay Selvi Hanişoğlu (Beykent University, Turkey) - Fidan Güler (Okan University, Turkey) * SESSION 4C: Finans II Session Chair: Assoc. Prof. Dr. Celaleddin Serinkan # Kırgızistan da Faiz ve Kur Politikalarinin İhracat Üzerine Etkisi Prof. Dr. Fuat Sekmen (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Ph.D. Candidate Galip Afsin Ravanoglu (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * # Kırgızistan da Bankacılık Sektörünün Yönetim ve Organizasyon Sorunları Assoc. Prof. Dr. Celaleddin Serinkan (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * # İşçi Dövizleri ve Reel Efektif Döviz Kuru: Kırgızistan Üzerinde Bir Uygulama Dr. Burulcha Sulaimanova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Daniyar Jasoolov (Public Fund AVEP, Kyrgyzstan) * SESSION 5C: Kalkınma Session Chair: Prof. Dr. Hakan Çetintaş # Dış Borç ve Yabancı Yatırımların Toplam Çıktıya Etkisi: Kırgızistan Örneği Asst. Prof. Dr. Junus Ganiev (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * - Asst. Prof. Dr. Damira Baigonushova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Ph.D. Candidate Mairam Baigonusheva (Marmara University, Turkey) # Kırgızistan da Dış Borçların Ekonomik Büyüme Üzerindeki Etkisi Tolkun Zhumakunova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) - Zhainagul Kydyralieva (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * # Kırgız Cumhuriyeti Bölgelerinde Anne Ölüm Oranları Belirleyicileri Ph.D. Candidate Nurbek Madmarov (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * - Assoc. Prof. Dr. Metin Bayrak (Kyrgyzstan-Turkey Manas University, Turkey) # Orta Asya da Bölge Merkezli Bütünleşme: Çabaları ve Sonuçları Nargiza Sakmurzaeva (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) *

16 xvi INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES # Kazakistan ve Kırgızistan da Kamu Harcamalarının Ekonomik Büyümeye Etkisi Serhat Atmaca (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) * - Assoc. Prof. Dr. Metin Bayrak (Kyrgyzstan-Turkey Manas University, Turkey) * Presenting author

17 OPENING SESSION 1 An Economic Eurasian Tale: Rakhine State Bir Ekonomik Avrasya Öyküsü: Arakan Eyaleti Prof. Dr. Ersan Bocutoğlu (Avrasya University, Turkey) Abstract Genocide or, to say the least, ethnic cleansing towards Muslims in Rakhine State, Myanmar, conducted by fanatic Buddhist monks and the military has been on the agenda in recent months. This opening speech aims at finding out real causes of this inhumane incident and investigating whether or not it is solely a result of some kind interreligious conflict in Myanmar. My research has convinced me that Rakhine inter-ethnic question has international economic and security related roots that deserve close and detailed investigation. In my point of view, the Rakhine Question depends heavily on economic security considerations such as: a) security of natural gas and petrol reservoirs in Rakhine State and pipelines connecting Rakhine State to China, b) security of railway link connecting Kyaukpyu Deep Water Sea Port in Kyaukpyu Special Economic Zone in Rakhine State to China which has developed by China to bypass Malacca Strait. The potential that Muslim dominated Rakhine State may cause security threats to some of the foreign Chinese investments in Myanmar in medium term should be taken as a main cause that gives way to ethnic cleansing against Muslim Rakhine population. Sayın Bakanım, Rektörüm, Rektör Vekilim, Dekanlarım, öğretim üyesi meslektaşlarım, sevgili öğrencilerimiz ve kıymetli misafirler, konuşmama başlarken hepinizi saygı ve sevgilerimle selamlarım. Myanmar (Burma) da iki asırdan beri Budistler ile Müslümanlar arasında çatışmalar yaşanmakla birlikte, son aylarda ortaya çıkan etnik temizlik ve soykırım, Budizm gibi barışçı yönüyle temayüz etmiş bir inancın sorgulanmasına yol açtı. Sürgündeki Budist lider Dalay Lama bile, Budistlerin bu tür eylemlerini anlaşılmaz bulduğunu ifade etti. Bu durum beni, Myanmar ın Arakan Eyaletindeki etik temizlik ve soykırımın arkasında başka motivasyonların yatabileceği düşüncesine götürdü. Konu ile ilgili olarak yaptığım araştırmalar sonucunda, Arakan Eyaletinin karasal hudutları içinde zengin doğalgaz, deniz hudutları içinde ise zengin petrol ve doğalgaz yataklarının bulunduğunu; bu enerji kaynaklarının, boru hatları aracılığı ile Çin e nakledildiğini gördüm. Araştırmalarımda; Çin ihracatının ve enerji ithalatının % 80 inin Çin bakımından güvenli kabul edilmeyen Malaka Boğazından geçmesinin yarattığı sıkıntı anlamına gelen Malaka Sendromu nu çözmek amacıyla, bu boğazı bay-pas eden yeni bir alternatif olarak Çin in Arakan Eyaletindeki Kyaukpyu şehrini Çin e bağlayan bir demiryolu hattını geliştirdiğini, Kyaukpyu şehrinde Özel Ekonomi Bölgesi kurduğunu ve Kyaukpyu Derin Deniz Limanını inşa ettiğini tespit ettim. Arakan Eyaletinin Müslüman ağırlıklı bir nüfus yapısına sahip olduğu dikkate alındığında, bu bölgenin orta vadede Çin in petrol, doğalgaz üretimi, boru hatları, özel ekonomi bölgesi, demiryolu taşımacılığı ve derin deniz limanı alanlarındaki yatırımlarının güvenliği bakımından sorun yaratma potansiyeline sahip olduğu dikkatlerden kaçmamaktadır. Müslümanlar aleyhindeki uluslararası konjonktürün de uygun olması nedeniyle, güvenliği sağlamanın en kestirme yolunun Müslümanları bu bölgeden bir şekilde sürüp çıkarmak olduğu açıktır. Bu nedenle Myanmar da Müslüman ağırlıklı Arakan Eyaletinde ortaya çıkan Müslüman soykırımı ve etnik temizliğin sadece dini nedenlere dayandığını varsaymak ve meselenin ekonomi ve güvenlikle ilgili yönlerini ihmal etmek kanaatimce meseleyi yüzeysel değerlendirmek anlamına gelecektir. Elimde kesin veriler bulunmadığı için, olayın sorumluluğunu hiç bir ülkeye yıkamıyorum. Ancak sonuç yerine şu soruyu sormak istiyorum: Acaba Arakanda olan bitenlerle Hint Okyanusundaki Çin-Hindistan-ABD rekabetinin bir ilişkisi var mıdır? Eğer varsa ve bu Çin in Bir Kuşak-Bir Yol Projesi ile ilgiliyse, Çin e komşu olan birinci kuşak devletlerin orta vadede benzer sorunlarla karşılaşmasının kaçınılmaz olduğunu ifade ederek konuşmamı bitiriyorum. Sabrınız ve dikkatiniz için teşekkür ederim.

18 2 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Effect of Price and Brand on Purchase Decision An Application on Turkish Smart Phone Consumers Ph.D. Candidate Barış Batuhan Geçit (Beykent University, Turkey) Asst. Prof. Dr. Murad Kayacan (Beykent University, Turkey) Abstract Purpose of this study is to discuss whether brand and price has a significant effect on consumers smart phone purchase decision in Turkey. For this purpose; a survey has been conducted to the Turkish smart phone consumers about the factors affecting their smart phone purchase decision. The results obtained from the survey has been analyzed by using IBM SPSS program. According to the results obtained from the survey the model is accepted as p value of the model is smaller than 0, According to the analyze both brand and price has a significant effect on Turkish consumers smart phone purchase decision as both variables&#; p values has been found to be smaller than 0, Brand has a positive effect on consumers&#; smart phone purchase decision whereas price has a negative effect on consumer&#;s smart phone purchase decision. 1 Introduction Until the end of the World War II, there was no problem in selling the produced goods, especially because of the lack of serial production and the fact that demand was not more than production (Durmaz, ). With the end of the war and the increase in the production capacities of companies, a market structure has emerged in which every manufactured good could be sold. However, the period called "sales approach" has emerged with the increase of production capacities and increase of demand, as a result, forced sales with psychological pressure has increased. The approach was "Sell whatever you produce, as long as know how to sell it" and the selling strategy was completely misleading, which in fact, did not fully reflect the wishes and needs of consumers (Taşkın, ). In the advanced times, the mass production of businesses meant that unit prices for consumers were getting cheaper. As a result, the pace of demand started to increase and a new approach called "production approach" became valid (Koç, ). However, over time, inability to respond to the needs and developments of the market in terms of production (sales) and insufficient supply-demand balance has led to the emergence of "modern marketing approach". In modern marketing approach, which is accepted today, first step is to determine requirements and desires of consumers then the next is step is production (Taşkın, ). As consumer requirements became the focal point; consumer requirements, consumer behavior became an important concept that has to be analyzed by marketers. Many firms have begun to focus on customers for their business activities and as a quality investment tool (Kotler and Armstrong, ). One of the key concepts that creates huge difference in the perception of consumers for the product, which has become one of the focal points of marketing, is brand (Blythe, ). Brand is a concept that has an important position in marketing strategies that creates awareness among consumers and differentiate them for companies. Besides the brand concept, another important concept for marketers is the price. Price is the mechanism for obtaining value for firms (Iacobucci, ). For consumers, the price is the amount needed to get a product (Hawkins and Mothersbaugh, ). One of the most important markets where brand perception is most common and the effects on consumers are required to be examined the most is smartphone market. In the smartphone market, giant companies who constantly rise prices of their products also continue to sell at higher numbers and this fact makes smartphone market an extraordinary market to be analyzed as increasing prices also increases the sell numbers. The fact that brands&#; market positions and values are changing and evolving with the developing technologies in the smartphone market is the reason that Apple has been a leader in determining the rules of the market right now, even they were an obscure brand couple of years ago. Apple went from being an obscure brand to being number one brand of the World (Interbrand, ), Samsung utilizing the brand image they had obtained from various product groups for years in the smart phone market and even they suffered Note 7 Explosion Scandal, they can still sell at high rates by courtesy of their high brand image, brands like Huawei and Xiaomi went from being completely unknown brands to preferred brands, Sony went from being one of the biggest technology brands of the world to a poor smart phone brand, Blackberry had a %20 market in and currently has a market share of % and is in the process of disappearing (Business Insider, ). All these rapid changes in smart phone market shows how dynamic this market is and it is vital to analyze the facts that effect consumer preference in smart phone market.

19 SESSION 1A: Microeconomics 3 Alongside with its dynamic nature, smart phone market it creates very high amount of profit and has a very big transaction mass. Smart phones are one of the most common used products in the World. According to Statista () there are Million smart phone users in the World and by the Statista is also forecasting that there will be Million smart phone users in the World. While smart phone is that common in the whole World, Turkey is one of the top countries in the World in terms of smart phone usage. According to Newzoo () Turkey is the 12 th country in terms of smart phone users with 40 Million Smart phone users. According to Poushter () Turkey is once again 12 th in terms of smart phone usage percent/total population ratio. So according to these numbers Turkish consumers can be counted as top 15 for the smart phone brands. All these facts make Turkey, which according to The World Bank () is the 17 th biggest economy of the world in terms of GDP, an ideal and an appropriate market to be analyzed in terms of smart phone consumption trends and the main reasons of smart phone preference. To analyze the facts that effect consumer preference in smart phone market; price and brand has been selected as the most appropriate variables after reviewing the academic literature. 2 Literature In terms of marketing, it is undoubtedly very important to examine the factors that affect the purchase decision of consumers. In the group of products that is analyzed, being able to know which reasons lead consumers to buy certain brands or products; may be important in terms of businesses that want to become big brands. That is because they can determine how to act in the future and what strategies they should implement, and how they can eventually achieve a superior position in the competition. The type of consumer analyzed in this study is not the industrial consumer, but the final consumer. Because the smartphone is a final consumer. Consumer behavior can be defined as the processes directly related to the acquisition and use of goods and services, which are the economic values of individuals, and the decisions that determine and determine these activities (Tek, ). As mentioned above smart phone is a final consumer product and consumer buying behavior is purchasing behavior for personal or family use (Dibb, Simkin and Pride, , p). Price is the amount of money charged for a product or a service. In other words, price is the sum of all the values that customers give up to gain the benefits of having or using a product or service. Historically, price has been the major factor affecting buyer choice. In recent decades, however, nonprice factors have gained increasing importance. Even so, price remains one of the most important elements that determines a firm s market share and profitability (Kotler and Armstrong, , p). Price is the most visible element of the marketing mix, and pricing policies are being questioned by consumers frequently. If consumers think that prices are unfair, they can leave the firm or spread negative information to other consumers. Low price can have a positive effect as consumers would demand to pay lower prices but it can also have a negative effect as low price usually decreases the brand image of products and brands. High price can vice versa have a negative effect on consumers as they don t want to pay higher prices but it can increase the brand image. If product development, promotion and distribution are the seeds of business success, price is the harvest (Kotler and Armstrong, , p). While price is seen as an important factor to affect the dynamics of a market, brand is also regarded as an important variable to affect the dynamics of competition and perception of consumers. Brand is largely associated with the attributes of the product, benefits, use values, user culture, and personality and user behavior. Every brand has a unique chronology and experience. These are built over time through strong and consistent communication wherever the brand manifests itself from business cards to electronic communication such as television commercials, internet and print media, and finally to billboards (Rajagopal, , p.1). Brand generates image for product which creates awareness for consumers. Brand helps companies to position themselves in the market. Positioning the product or the brand helps company to decide the desired consumer group, potential market rivals and competition and as a result the long term plans of the company. Brand basically helps firms to differentiate and position themselves within the market competition and for the consumers. If we give an example for the importance of awareness on consumers; Apple&#;s easy-to-use, aesthetic and technological music player series, launched in , has been successful in a short time. On top of that, the ipod shuffle, IPod nano and IPod touch, which are sold in different prices and quality varieties of IPod, have been introduced into the market. IPod, which sold over million in eight years, was also the pioneer of IPhone &#;s launch of Apple&#; s actual explosion (Aaker, , p. ). Thanks to the brand image and the awareness on consumers obtained by ipod s, iphone product made an explosive departure in the smart phone market.

20 4 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Marketing literature mainly supports the idea that brand and price have a significant effect on consumers purchase decision but a statistical analysis should be done to prove these ideas and especially for a certain product group like smart phone, which has a very different market dynamic compared to other products, it is very important to do such an analysis. 3 Model Model of this study is as following (Table 1): Table 1: Price and Brand&#;s Effect on Consumer Decision H1: Price has a significant effect on consumers smart phone purchase decision. H2: Brand has a significant effect on consumers smart phone purchase decision. 4 Methodology A survey has been conducted to the Turkish smart phone consumers. The results have been conducted via survey has been analyzed with using IBM SPSS 22 programme and regression analysis has been made for the significance and the positivity of the analysis. 5 Analysis Firstly, normality analysis has been made to find out the generalizability of the dataset. Normality analysis is as following (Table 2): Variable and Question Number of the Variable Skewness Kurtosis Purchase Decision -1, 3, Brand ,, Brand ,, Brand ,, Price ,, Price 2 1, 0, Price ,, Table 2: Normality Analysis When the normality distributions of the data are examined; The p values in Kolmogorov-Smirnov and Shapiro- Wilk must be greater than in order to be able to accept the normal distribution of data. The p values in the analysis results were mainly lower than On the other hand, when we look at the values of Skewness and Kurtosis, another value that is considered valid for normality distribution of the data, it is seen that a large part of the data is within the acceptable values of -3 and +3. In this direction, the data can be interpreted as normally distributed. Regression analysis has been made to expresses the significance between the variables and analyze its power. Regression analysis is as following (Table 3):

21 SESSION 1A: Microeconomics 5 Model Sum of Squares df Mean Square F Sig. 1 Regression 12, , 16,, b Residual , , Total , Table 3: Regression Analysis I According to the regression analysis; the model is valid (accepted) as the Sig. (p) value is lower than 0, Standardized Unstandardized Coefficients Coefficients Model B Std. Error Beta t Sig. 1 (Constant) 4,, 35,, Brand,,, 2,, Price -,, -,,, Table 4: Regression Analysis II After obtaining the result of the model as valid, the second table shows the significance of the effect of two independent variables to dependent variable. Again Sig. (p) value should be lower than 0,05 for independent variables to have a significant effect on consumer preference dependent variable. As seen from the table, both brand and price variables have a significant effect on consumers purchase preference for smart phone. Beta value shows the positivity and negativity of these relations. According to the Beta value; brand has a positive significant effect on smart phone purchase decision. Which means the consumers tend to purchase smart phones who has a better brand image. According to International Data Corporation () Samsung has a % market share and Apple has a % market share. These two brands have a total of %38 market share in a market that has numbers of companies. Interbrand () lists Apple as the number one most valuable brand of the World, whereas they list Samsung as the number seven most valuable brand of the World. So as a result proven by these lists, even Apple switched from number one position in most valuable brand list to number two position in market share; two most valuable brands by has the biggest market share by far. These numbers can be seen in table below. Switch of Apple and Samsung can be explained by two factors. Firstly, Samsung produces various number of different smart phones like, Samsung S8, Samsung Note, Samsung A series, Samsung J series, etc. However, Apple only produces iphone. One other fact that may have an effect on consumers preferring Samsung more than Apple is, Apple has been selling smart phones at a very higher price compared to Samsung throughout the years. However, with the Samsung S8 model, Samsung has reached to the price levels of Apple and even higher prices. Samsung S8 is being sold at Turkish Liras whereas Apple iphone 7 is being sold at Turkish liras. But while segmenting a very high price level of a product, as Samsung still produce lower priced and lower segmented products, they also target the lower and middle incomed consumers and as a result they can sell more. MARKET SHARE BRAND VALUE BRAND VALUE FORTUNE GLOBAL BRAND (IDC, ) (Interbrand, ) (Brandz, ) LIST () Apple %14,7 M$ (1st) M$ (2nd) 15th Samsung %23,3 51M$ (7th) 24M$ (37th) 13th Table 5: Samsung and Apple Data Comparison Price has been found to have a significant effect on consumers smart phone purchase decision but Beta value has been found as negative so which means price has a negative effect on consumers smart phone purchase decision. That means as prices increase people don t tend to buy these products. It can be interpreted as no consumer would prefer to buy a product that increases its prices but that kind of analysis cannot be ensured without such an analysis because it both submits a statistical result and even though Apple increases its price range in every single product As mentioned above, Samsung also target lower and middle incomed consumers, they sell at high numbers compared to Apple with a %23,3 market share. It can be easily interpreted that these two firms apply different company strategies. Apple tends to only sell at high price tag and retain high brand image whereas Samsung tries to apply a more balanced strategy where they both try to sell at every price segments and at the same time retaining a high brand image.

22 6 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES To conclude and summarize; brand and price has been found to have a significant effect on consumers smart phone purchase decision. Brand has a positive effect whereas price has a negative effect. Which means as brand attributes increase purchase decision increases and when price increases purchase decision decreases. Examining the sale numbers, these hypotheses are supported by two giant companies sales numbers and apart from this statistical study, brand and price seem to have an effect on each other which can be studied in further studies. References Aaker, Brand Relevance - Making Competitors Irrelevant. Jossey Bass. Blythe, Marketing. Sage Publications. Brown, Top Global Brands Brandz. Business Insider, BlackBerry&#;s share of the global smartphone market is now officially 0%. Dibb, Simkin, Pride and Ferrell, Marketing. London: Cengage Learning. Fortune, Global Iacobucci, Marketing Management. Cengage Learning. Interbrand, Best Global Brands Ranking. International Data Corporation Research, Smartphone Vendor Market Share, Q1 Hawkins and Mothersbaugh, Consumer Behavior - Building Marketing Strategy. McGraw Hil. Koç, Tüketici davranışı ve pazarlama stratejileri: global ve yerel yaklaşım. Seçkin, 7 th edition. Kotler and Armstrong, Principles of Marketing. Pearson. Newzoo, Top 50 Countries by Smartphone Users and Penetration. Poushter, Smartphone Ownership and Internet Usage Continues to Climb in Emerging Economies. Rajagopal, Brand Management Strategy. Nova Science Publishers. Statista, Number of smartphone users worldwide from to (in billions). Taşkın, Pazarlama esasları. Türkmen Kitabevi. Tek, Pazarlama İlkeleri Global yönetimsel yaklaşım türkiye uygulamaları. Beta Yayınevi. 8 th edition. World Bank, World Bank Databank Database.

23 SESSION 1A: Microeconomics 7 Gender Differences, Risk Attitude and Entrepreneurship in Kyrgyzstan Asst. Prof. Dr. Raziya Abdiyeva (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Dr. Burulcha Sulaimanova (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Asst. Prof. Dr. Kamalbek Karymshakov (Kyrgyzstan-Turkey Manas University, Kyrgyzstan) Abstract This study analyses the role of risk attitude for entrepreneurship by gender differences in Kyrgyzstan. Multinomial probit analysis is applied to the data set drawn from the nationally representative survey Life in Kyrgyzstan for Entrepreneurship is measured through the agricultural and non-agricultural sample. Results of the study show that more risk-taking preferences are associated with higher entrepreneurship probability. However, this effect is not persistent for women in further estimations for non-agricultural entrepreneurship sample, while for men higher positive effect of risk loving behavior remains in off-farm self-employment too. 1 Introduction Economic development requires development of entrepreneurship activity in economy. Along with such measures as strengthening legislature for private property rights or development of financial and other infrastructure, individual behavioral characteristics and perceptions are important for self-employment propensity of individuals. Among these behavioral features empirical studies indicate that risk attitude of individuals is important factor for entrepreneurship choice of individuals (Hvide and Panos, ; Ekelund et al., ; Wagner, ). Moreover, majority of the papers revealed gender differentials in risk aversion behavior of individuals. In particularly, it is asserted that women are more risk averse than men (Bruce and Johnson, ; Eckel and Grossman, ; Al- Ajmi, ). Following these empirical findings, it is of particular interest to study the risk tolerance effect on the of entrepreneurship of women. The aim of this paper is to analyze the effect of risk tolerance by gender on the choice of entrepreneurship in the specific case of Kyrgyzstan. To our best knowledge this study is the first that analyze the relation between risk attitude and entrepreneurship in Kyrgyzstan. According to the World Bank () 59 percent of women and 75 percent of the male population between 15 and 64 years of age in were in the labor force. Female labor force participation rates in emerging Europe and Central Asia are near the global average, but the male participation rates are well below the global average. According to the National Statistical Committee () of Kyrgyz Republic in , % of women were employed. Most of them are employed in services sectors, such as education, healthcare, real estate etc. In Europe and Central Asia region Kyrgyzstan are among the countries with higher rates of female participation in ownership (World Bank, ). This study draws on the nationally representative cross-sectional data, where responses of respondents are used as subjective measure of the risk tolerance of individuals. Binary response logit models are applied for empirical estimation. The paper is structured as follows: In the Section 2 literature review on risk tolerance and entrepreneurship, taking into account gender differentials, is given. Section 3 describes data and provides with summary statistics. Section 4 includes empirical methodology and Section 5 discusses results of estimation. Section 6 concludes. 2 Literature Review Risk attitude of individual can affect his/her choices and decisions. Bonin et al. () find that the more risk lover individuals select occupations with higher wages and higher variation in wages. Dohmen and Falk () with complementary laboratory experimental study showed that relatively risk averse workers prefer fixed payments than variable payment schemes such as piece-rate or tournament contracts. More risk lovers are more likely to choose entrepreneurship than wage employment. According to some theoretical and empirical studies (Wagner, ; Ekelund et al., ) entrepreneurship requires making risky decisions in an uncertain environment. So only those persons who are able to bear higher risks may start as an entrepreneur. From this point of view, the risk attitude of a person is one of the crucial variables in a person s choice between entrepreneurship and a salaried job (Caliendo et al, ). Guiso and Paiella () find that less risk averse individuals are more likely to be self-employed. Also several other empirical studies revealed that more risk lover men more inclined to be engaged in entrepreneurship as selfemployed (Hartog et al., ; Guiso and Paiella, ; Ekelund et al., ; Dohmenet al.,; Kan and Tsai, ; In Soo Cho, ). On the other hand, some studies found mixed results related the effect of risk attitude of

24 8 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES individuals on the choice of entrepreneurship. Rosen and Willen () came to conclusion that risk attitude is not a dominant factor in his/her decision to start an own business. Cramer et al. () too could not be confident enough to conclude that there is causality link between risk aversion and entrepreneurial selection of individual. Also, Blanchflower and Oswald () found positive correlation between the wealth status of a person and his/her risk attitude. Generally empirical studies showed that women are more risk averse than men (e.g., Bruce and Johnson, ; Hartog et al., ; Eckel and Grossman, ; Croson and Gneezy, ; Al-Ajmi, ; Çağlayan and Abdieva, ). Charness and Gneezy () found a strong difference in investment behavior across gender and results showed that women make smaller investments in the risky asset than do men, and so appear to be financially more risk averse. Soo Cho () using National Longitudinal Surveys of Youth data and multinominal logit model find that more risk tolerant men are more likely to enter self-employment but they didn t find any significant effect of risk tolerance on women s entrepreneurship. Brody () affirms that in anticipation of negative outcomes, women report more nervousness and fear than men, thus it is logical that women are more risk averse when facing a risky situation. Niederle and Vesterlund () and Soll and Klayman () have found that men are more overconfident than women. According to several studies gender differences in risk attitude is eliminated with experience and profession. Atkinson, Baird and Frye () analyzed mutual fund managers and didn t found gender differences in the way funds are managed in terms of performance and risk. Johnson and Powell () found differences among the non-managerial population, but did not observe such differences in the managerial population (managers and potential managers with management education). There are several studies focusing risk attitude and gender relationship in Kyrgyzstan case. Çağlayan and Abdieva () investigated risk tolerance of individual investors in Kyrgyzstan case. Using multinomial logit model they found that men are more risk lovers than women in Kyrgyzstan. When the age increases people become less risk lover. Having non-wage income increases taking risk and increases in the rate of investment. The findings also indicate that income has a positive effect on the risk tolerance. Abdieva et al. () analyzed the determinants of the risk tolerance in Kyrgyzstan. The results of multinomial probit models showed that men more willing to take risk and that as the age increases individuals become risk-averse. Increasing income and education level has a positive effect on the risk taking decision of the individual. The regional distribution of risk tolerance of individuals shows that individuals living in rural areas and in south region are more likely to be in risk-averse category. Results of study by Esenaliev and Anderson () on gender wage gap in Kyrgyzstan showed that the level of gender wage gap is %, and risk attitude is used as one of the possible factors for explaining this. Although these studies explore empirical evidence on risk attitude by gender and raise its importance as potential factor for wage earnings differentials, to our best knowledge studies did not focus on the risk attitude and entrepreneurship activities by gender in case of Kyrgyzstan. This paper aims to fill this gap. 3 Methodology The decision of individual to be entrepreneur has binary response character, which calls for using binary response Logit models, which are conditional on individual and household characteristics. Formally, model is given below (Wooldridge, ): P(y i = 1 x i ) = G(β 0 + β 1 x β k x k ) = G(β 0 + xβ) (1) G(z) = exp(z) = Λ(z) (2) [1 + exp(z)] where G is the logistic function. y i is the discrete dependent variable, taking values of zero or one, showing the probability of individual to be entrepreneur; x i is the vector of variables at individual level, which includes individual s age, marital status and education level and household levels variables, which includes household composition, expenditure level, regional characteristics (for detailed description of variables see Table A in appendix). Occupational choices of individuals are based on their employment status given in the questionnaire. It should be noted that in this paper entrepreneurship is measured in a broad sense and those who indicate their wage employment as the own-account worker are grouped as the entrepreneurs. However, this broad definition of entrepreneurship may give biased results, since in most of the developing countries not all own-account working activities can be considered as entrepreneurship. As Karymshakov et al. () note in Kyrgyzstan members of household who own land are considered as own-account workers. In this case it appears that even if individuals just work as family worker in agriculture with low productivity, they may be considered as own-account workers. Therefore, working in agriculture as own-account worker may not have characteristics of entrepreneurs. This fact raises the necessity for measurement of entrepreneurship in in-farm and off-farm sector. Following this issue we estimate the model by agriculture and non-agriculture sector.

25 SESSION 1A: Microeconomics 9 In the dataset Life in Kyrgyzstan for used for this study, there is a special section on subjective well-being of individual, where individual is asked particular question on how they asses their risk tolerance, as person who is fully willing to take risks or a person avoiding taking risks, which is measured from 0 to The answer for this question is taken as the main explanatory variable. From these answers three categories of risk tolerance are formed: risk-averse individuals (from 0 to 3), risk-neutral (from 4 to 6) and risk-lovers (from 7 to 10). 4 Data and Descriptive Statistics In this study the risk attitude and gender behavior impact on entrepreneurial decision is investigated based on the second wave of "Life in Kyrgyz Republic" survey data, which was conducted by DIW Berlin in collaboration of Humboldt University of Berlin, the Center for Social and Economic Research (CASE-Kyrgyz Republic), and the American University of Central Asia (AUCA) in This survey includes wide range information both on individual and household level and representative at the national level. The table 1 describes the main individual and household characteristics both for men and women. The total amount of observation is equal to individuals between ages, where % are women. The mean age of sample is years, and there is no significant difference in ages between genders. While marital status of women is higher than men, showing that women have more propensities to be married rather than men. The education attainment shows that both men and women more likely to have basic or secondary education, while men more likely to have technical education and women more inclined to have tertiary education. Total sample Men Women Obs. % Obs. % Obs. % Individual characteristics: Age (mean) Marital status (1=married) Education level: - Basic or secondary Technical Tertiary Risk attitude: - Risk lover Risk neutral Risk averse Household characteristics: Household size (mean) Children ratio ( years) (mean) Expenditure per capita (mean) Residence (1=rural) Regions: - North South Central Table 1. Descriptive Statistics Source: LIK data The risk attitude indicators show that % and % of men and women count their self as risk lovers respectively, while % and % of men and women report their self as risk-averse person. Precisely, men more likely to take risk rather than women do. The household characteristics do not display gender differences, and in general the average household size consist of 5 people, the ratio of children in the household with respect to household size is around and total expenditure per capita in household is around 38 thousand soms. More than half of individuals reside in rural areas and most of observation comes from south regions, which are highly populated. 5 Estimation Results The estimation results for entrepreneurial decision of the individual are given in the Table 2. Almost all estimated parameters have expected signs. The primary variable of interest, the impact of risk attitude of individual on decision to be entrepreneur shows that risk lovers more likely to be self-employed, and that there is significant gender difference on impact of risk attitude on labor supply decision. Thus, risk-lover women more likely to be entrepreneur, but this effect is not significant in further estimations by agricultural and non-agricultural samples. Women with risk averse preferences have less probability to be entrepreneur in non-agricultural sector. Interestingly, risk effect for men is more evident in non-agricultural sector. Risk lover men show higher probability

26 10 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES to be entrepreneur in non-agricultural sector, while in agricultural sector it shows negative effect. Moreover, the risk averse preferences have less negative effect for off-farm entrepreneurship for men. There is statistically significant impact of age on the individuals decision to engage in entrepreneurial activities, thus with increase of age individual more likely to be entrepreneur, however after definite ages this probability decreases. Generally marital status of individual is positively correlated with decision to be entrepreneur, while there is no significance of marital status on entrepreneurs working in agricultural sector, especially if they are women. The educational level of individual shows that, with higher education individuals more likely to be working in other occupation types. While having technical or tertiary education do not present any impact on decision of individual to be self-employed in agricultural sector. The household composition has twofold impact on entrepreneurship. While the size of household does not influence the entrepreneurship, the ratio of children ( years) with respect to the total size of household size impacts the decision of individuals of men and women differently. Hence with increase of children ratio the probability of men to be self-employed is increasing, while for women is decreasing. The per capita expenditure level of household has significant influence on individual decision to be entrepreneur in non-agricultural sector and general positive correlation with women decision to engage in self-employment both in agricultural and non-agricultural sectors. The residence of household in rural area is increasing the probability of individual to be self-employed in agricultural sector rather than in non-agricultural sector. Moreover, the regional characteristics of household show that individuals from north area more likely to be working in own account in agricultural sector, while in south region individuals more likely to conduct theirselves to selfemployment in non-agricultural sector. Total sample Agricultural sector Non-agricultural sector Total Men Women Total Men Women Total Men Women Individual characteristics Age *** *** *** *** *** * *** *** *** Age squared *** *** *** *** *** *** *** *** Marital status (1=married) *** *** ** ** *** *** * Education level: - Technical *** * Tertiary *** *** *** *** *** * Household characteristics Household size * *** * ** Children ratio ( years) ** ** *** ** * * Expenditure per capita ** ** *** *** ** Residence (1=rural) *** *** *** * *** ** *** Regions: - North *** *** *** *** South *** *** ** *** *** *** * Risk attitude: - Risk lover *** ** *** *** ** Risk averse *** *** *** ** ** *** ** *** No. of Obs LR chi *** *** *** *** *** *** *** *** *** Log likelihood Pseudo R *,** and *** show statistical significance at the 10, 5 and 1% level, respectively. For coefficient estimates of the models see the table in Appendix Table A.2 6 Conclusion Table 2. Logit Estimation Results for Entrepreneurship (Marginal Effect Estimates) Development of entrepreneurship activities for developing countries is important long-term objectives, which may depend on many factors ranging from government policies to other formal measures. Along with these factors, individual risk-tolerance characteristics are important for carrying out entrepreneurial activities. The objective of this paper was to study the impact of risk tolerance over the entrepreneurship by focusing on the gender and non-agricultural sector of employment. This study applied binary response techniques on the large nationally representative cross-sectional data. In general, estimation results showed that individual preferences for taking more risk have positive effect over the entrepreneurship. However, further estimations by gender and non-

27 SESSION 1A: Microeconomics 11 agricultural sector revealed that women entrepreneurship in off-farm sectors is not associated with risk loving behavior. However, the effect of risk loving over the entrepreneurship remains in non-agricultural sector for men. Thus, although risk tolerance important for entrepreneurship, this effect may not be strong through all sector of employment and among men and women. Especially these results suggest that agricultural and non-agricultural measurement of entrepreneurial activities should be taken into account in order to have the true picture of entrepreneurship. For more comprehensive exploration of the relationship between risk tolerance and entrepreneurship further studies may test the reverse causality changes in risk preferences after being entrepreneur. Findings of this paper to some extent underline the fact that women entrepreneurship in off-farm sector of employment in Kyrgyzstan case is limited. Therefore, development of entrepreneurial activities among women in off-farm sector requires in priority government policies oriented supporting women self-employment. References Abdieva, R., Karymshakov, K. and B. Sulaimanova, Determinants of Risk Tolerance: Empirical Evidence from Kyrgyzstan, Annual Life in Kyrgyzstan Conference, October , Bishkek Al-Ajmi, J. Y Risk Tolerance of Individual Investors in an Emerging Market. Journal of Risk and Diversification. ISSN Issue 2. Atkinson, S. M., Baird, S. B. and Frye, M. B Do female mutual fund managers manage differently? Journal of Financial Research, 26(1), Blanchflower, D., and A. Oswald, What makes an Entrepreneur? Journal of Labor Economics, 16, Bonin, H., Dohmen, T., Falk, A., Huffman, D., and Sunde, U. () Cross-sectional earnings risk and occupational sorting: The role of risk attitudes, Labour Economics, 14, Brody, L. R. (). On understanding gender differences in the expression of emotions. In Human Feeling: Explorations in Affect Development and Meaning, ed. S. L. Ablon, D. Brown, E. J. Khantzian and J. E. Mack, (pp ). Hillsdale, N.J.: Analytic Press. Bruce, A. and Johnson, J., , "Male and female betting behavior: new perspectives, Journal of Gambling Studies 10, Çağlayan, E. and R. Abdieva How s the Risk Tolerance of Individual Investors in Kyrgyzstan?, Asian Economic and Financial Review, Volume 4, Number 3, p Caliendo, M., Fossen F. and Kritikos, A Risk Attitudes of Nascent Entrepreneurs: New Evidence from an Experimentally-Validated Survey, IZA DP No. , DISCUSSION PAPER SERIES, Forschungsinstitutzur Zukunft der Arbeit, Institute for the Study of Labor, June. Charness, G. and U., Gneezy, Strong Evidence for Gender Differences in Risk Taking, Journal of Economic Behavior & Organization, Cramer, J., J. Hartog, N. Jonker, and C. Van Praag Low Risk Aversion encourages the choice for entrepreneurchip: an empirical test of a truism, Journal of Economic Behavior and Organization, 48, Croson, R., Gneezy, U Gender Diferences in Preferences, Journal of Economic Literature, 47, Dohmen, T., A. Falk, D. Huffman, U. Sunde, J. Schupp, and W. Wagner (): Individual Risk Attitudes: New Evidence from a Large, Representative, Experimentally-Validated Survey, IZA Discussion Paper, Dohmen, T., and A. Falk Performance Pay and Sorting Productivity, Preferences and Gender, IZA Working Paper. Eckel, C.C., Grossman, P.J Forecasting Risk Attitudes: An Experimental Study Using Actual and Forecast Gamble Choices, Journal of Economic Behavior and Organization, Ekelund, J., E. Johansson, M.-R.J arvelin, and D. Lichtermann, Self-Employment and Risk Aversion - Evidence from Psychological Test Data, Labour Economics, 12(5), Gary Charness and Uri Gneezy, Strong Evidence for Gender Differences in Risk Taking, Journal of Economic Behavior & Organization, Guiso, L., Paiella, M The Role of Risk Aversion in Predicting Individual Behavior, CEPR Discussion Paper, Hans K. Hvide and Georgios A. Panos, Risk Tolerance and Entrepreneurship, IZA Discussion Paper No February

28 12 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Hartog, J., Ferrer-i-Carbonell, A., Jonker, N Linking Measured Risk Aversionto Individual Characteristics, Kyklos, 55, In Soo Cho, Do Gender Differences in Risk Preferences Explain Gender Differences in Labor Supply, Earnings or Occupational Choice? Working Paper No , December, IOWA STATE UNIVERSITY, Department of Economics. Johnson, J. E. and Powell, P. L., Decision making, risk and gender: are managers different?, British Journal of Management, 5(2), Kan, K. and Tsai, W., Entrepreneurship and risk aversion, Small business economics, 26, Karymshakov, K., Abdieva, R., Sulaimanova, B., & Sultakeev, K., Remittances impact on youth labour supply: evidence from Kyrgyzstan ( No. PEP) National Statistical Committee of Kyrgyz Republic, Women and Men in Kyrgyz Republic, years (in Russian). Niederle, M. and Vesterlund, L., Do women shy away from competition? Do men compete too much?, Quarterly Journal of Economics, (3), Rosen, H., and P. Willen, Risk, Return and Self-employment, Discussion Paper, University of Princeton. Soll, J. B. and Klayman, J., Overconfidence in interval estimates, Journal of Experimental Psychology: Learning, Memory and Cognition, 30(2), Wagner, J The Impact of Personal Characteristics and the Regional Milieu on the Transition from Unemployment to Self-Employment: Empirical Evidence for Germany, Jahrb ucherf urnational okonomie und Statistik, (2), Wooldridge, J. M., Introductory econometrics: a modern approach. World Bank, Sarosh Sattar. Opportunities for Men and Women: Emerging Europe and Central Asia Women in Entrepreneurship, Chapter three Women Entrepreneurship

29 SESSION 1A: Microeconomics 13 Matrix Management Approach and a Survey Research in the Industry Asst. Prof. Dr. Ceren Gül Artuner (Beykent University, Turkey) Prof. Dr. Erdoğan Taşkın (Beykent University, Turkey) Abstract This academic research aims to determine the approach of the employees in different organizations to the matrix management style. Although the practice of matrix management style is quite difficult and its implementation costs are high, the main reasons of its adoption by the managers are discussed. In this research, the facts of establishing effective and reliable communication with all of the employees in an enterprise, working as efficient teams, the interrelation among the teams, solving problems as a team as well as receiving education for conflict resolution and ensuring cooperation among people while maintaining the values of the enterprises are investigated. In addition, the attitude of the managers in the performance of important tasks towards prioritizing knowledge and competence rather than authority and line relationships, the establishment of an integral and equitable performance evaluation system, the existence of a consistent information flow and a multidimensional accounting are inquired. Besides, by the integration of the conflict management process into the planning process, the use of the appropriate planning software, and clearly defining the tasks with all the details and responsibilities involved, the building-up of needed connections to do the work and at the same time, which add value to the work, are evaluated. 1 Introduction While the external environment of an enterprise is changing especially in the terms of competition and technology, its internal structure has to adapt to these changes. A huge, complex and international enterprise with a functional structure where the line relationships are defined traditionally and where the departments are organized in an interlinked manner as production management, finance, human relations, management information systems, research and development has experiences more difficulties to adapt to the external environment. Particularly, in the case of enterprises which have numerous projects and a wide range of products, the matrix organization structure is outstanding as the most suitable option for many years among the long existing organizational structures. Matrix structure is a form of departmentalization or organizational design in an enterprise. The recent evolutions in the means of communication, the introduction of sophisticated technologies and an increase in the practice of project management has reinforced the frequency of use and applications of matrix organization type (Dylon et al., ) However, the most important challenges experienced in applying the matrix structure are the fact of reporting of an employee to more than one manager and the practice of cross working between different teams. As products are getting more complex and customers increase in number, most of the companies tend to refer to teamwork (Johnson and Geal, ; Kotler and Armstrong, ). Contemporary communication means has transformed the companies in a drastic manner. The managers are taking advantage of the communication means more intensely during the planning, implementation and control activities. Matrix structure is providing efficiency during the exercise of management functions (Joyce, ; Gunn, ). Acquiring knowledge of developing technologies is facilitating the economic struggle of a country. All of the large companies operating with restrained abilities and limited financial resources must ultimately benefit from the matrix organization structure in order to carry out production efficiently, to take advantage from specialization, to be distinguished from its competitors in the market, to augment its sales, to utilize efficient technologies, to adapt to the environment and to realize its goals and objectives (Kingdon, ; Erdoğan, ). 2 The Importance of the Matrix Structure During the 2nd World War, the matrix structure was used for the first time in order to benefit from the experiences of talented engineers and to produce complex guns and war tools, but it cannot be a remedy for resolving all the problems concerning the contemporary organizations structure (Numerof and Abrams, ). In the post war period, project management has been widely used in different industries like pharmaceutics, construction, naval construction and aeronautics all along with defense and spatial industries. The introduction of new production techniques and the adoption of new practices in various fields required the careful implementation of matrix structure in global enterprises in today s civilized world. In addition, matrix structure is commonly used in defense industry in Turkey (Pretorius and Taylor, ; Gunn, ; Baysak, ).

30 14 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES In times of economic fluctuations where societal transformations were occurring and customers expectations were shifting, the managers used the matrix organization structure as a flexible and dynamic mean to adapt to the ongoing changes (Shreiber and Rosenberg, ). Matrix structure is not a classical organization type or old fashioned departmentalization style where the job description of all of the employees are strictly and clearly determined in advance (Peters, ). This contemporary form could be adopted by the management with a fresh and different look to the departments, employees and the ongoing works in the company. The employees do not have to consider their tasks in a restricted manner squeezed in a predetermined organizational structure. Indeed, matrix structure requires every manager and employee to see his / her task from a broader point of view and to attach importance to the ultimate details while performing it (Kolodny, ). In fact, the matrix structure is not only an organizational structure but also a managerial thinking system and consists of building shared values. It also means creating a common point of view among the employees, augmenting the workforces efficiency and ensuring its efficient functioning, together with integrating the management s efforts and succeeding the needed change and transformation (Barlett and Ghoshal, ). In the fully developed large-scale global company, product or business, function, area and customer know-how are simultaneously focused on the organization s world-wide marketing objectives. This type of total competence is a matrix organization. Management s task in the matrix organization is to achieve an organizational balance that brings together different perspectives and skills to accomplish the organization s objectives (Keegan and Green, ). In mighty organizations with a considerable number of employees, producing on a big scale and targeting a variety of markets, rethinking to organize different departments like marketing communication, product management, production, data processing, research and development in an integrated matrix organizational structure would facilitate the accomplishment of goals of the company as a whole (Burton et al., ). A company has to have some apparent skills in order to adopt and implement the matrix organization structure as a management concept in a successful manner. In the context of this research, the opinions of employees in different organizations concerning the level of the apparent institutional skills required for a successful implementation of matrix structure is asked. Undoubtedly, other institutional properties are also needed for the success of a matrix structure application in the practice. However, this research is analyzing the traits or particularities of companies who adopted a matrix structure with reference to sixteen questions. Matrix structure does not mean solely a change in the organizational structure, but it also involves a change of paradigm at every managerial level. In order to ensure this change, favorable conditions for learning and training opportunities must be created for the employees. People who would carry these efforts in the company, especially those who are in key positions and the managers supposed to bring about this change must have excellent communication skills. The thoughts and the actions as well as the performance of this workforce must traverse the company s actual boundaries (Galbraith, ; Galbraith, ; Gunn, ). 3 The Problems Encountered and Their Solutions Contemporary organizations have to cope first with a lot of problems in order to accomplish their goals and to survive. The nature of the products and services introduced to the market are getting more and more complex. In addition to that, customers are asking for more integrated services. There is a considerable constraint on minimizing the costs. However, the talented and experienced employees are hard to find. Most of the time, the management reveals to be inadequate in satisfying the employees expectations, in establishing a flawless communication, in giving an account of its actions and in making powerful decisions. In such a case, a matrix structure may offer new solutions to the above mentioned problems by handling them more carefully (Numerof and Abrams, ). If there is growing conflict between the teams and the team leaders, a need for a more flexible teamwork, the work performed by the employees doesn t serve to the realization of the main objectives of the organization, the managers can t manage their subordinates without referring to their authorities, there exists general optimization problems throughout the organization, the resources are not exploited at the optimum level, the organization becomes more complex and gets hard to manage, the goals must be adjusted more quickly and more simply, the planned projects take more time to be implemented strategically and the production time gets longer, rumors and blaming others becomes familiar among the employees, some employees are fired but it doesn t contribute to the solving of the problems, there s only a few persons who are doing their best while the others are reluctant for working, there s no innovation, the organization may need to be restructured in an effective manner. Briefly, if the contemporary organizations managers are willing to cope with these problems in an efficient and effective way, they have to learn about matrix structure and implement it. In an integrated manner with the global economic system, the companies whose targeted customers are not loyal to any product nor service are trying to adapt themselves to a continuously changing environment. They may reshape their organizations according to the matrix

31 SESSION 1A: Microeconomics 15 structure in order to realize their strategies, to ensure a powerful leadership, to take profit from teamwork, to reach the top, to change the game s rules and to achieve their competitive transformation (Martin, ). Needless to say that the matrix organization type is not a remedy for all the diseases. There is a need in the organizations for a more creative managerial view, which detects and identifies the problems from a broader perspective instead of criticizing all. The hiring of talented workforce to the company, their training, leading, assigning them in appropriate teams and projects could grow the matrix organization s success. In the enterprise, all the managers perform their most key tasks based on knowledge and competence, not according to their line authority. The power rests on knowledge and competence in the company. The high-ranking persons in the company are accustomed to impress the employees by practice without referring to their line authorities. A complete and equitable performance appraisal is put into practice trough the organization. The company spares time to measure all the different performance types, both quantitative and qualitative. The workforce and the management resolves effectively the arising problems together. The companies are developing the principles and values, which would encourage the cooperative problem solving. The workforce reacts to the arising problems as: How can I contribute to the solution of this problem? Instead of That doesn t concern me! In an effective matrix structure, it is not acceptable to blame anyone for the coming out problems. The management follows the company s conduct by consulting the employees in the whole organization regularly. Like every organizational structure, matrix management style has also its weaknesses and its strengths. Planning management and human resources department have to assume this responsibility. Especially, people who are reporting to more than one superior must be followed closely. People who has something to say must be heard in private sessions or they must have to be part of focus groups. It is also proven by a research that, for the sake of the company s future, the following of such precautions is extremely important (Galbraith, ). 4 Materials and Methods The aim of this academic study is to determine the matrix organization type management perception of the managers. Throughout the research, we have referred to the capabilities that a company needs to successfully execute a matrix organization. The discussion of these capabilities are spread throughout the survey. The survey can test how effective the leaders believe the company is at executing the capability. The following scale are the competencies and capabilities necessary for executing a matrix, and these are that we research for in effective matrix organization under twelve headings: 1. The company possesses high social capital or no silos. As the elements of a reliable workforce, a high level of communication is established with all the employees. As the most important feature of the matrix organization, formal or informal communication with every employee working for the company must be established in a reliable and straightforward manner (Burns and Wholey, ). 2. All the workforce work within the boundaries of the enterprise as effective teams. Matrix structure is naturally based on teamwork. The team players and the leaders identify the problems, discuss on them and work together to solve them. That is why the teams are working in a cross-borders, cross-companies and cross-functions manner (Davis and Lawrence, ). 3. Every team works in an interlinked manner with the others in the company. The teams operating in a matrix structure are linked to the others in different ways and in various domains. The planning process is driven with the team and the estimations or the prospective problems are communicated to the management (Kinor and Francis, ). 4. The workforce had a training as a team on problem solving and conflict management skills. All the teams within the matrix structure face conflicts and disparities. In order to resolve those conflicts and to perform better, the workforce needs similar skills and common processes (Spencer and Cox, ). 5. The most important task of the managers is to ensure the habitual cooperation and to protect the values. The company chooses the employees who would succeed in a matrix structure, hires them, train them, coaches them and promotes them (Brown and Agnew, ; Galbraith, ). 6. Continuous information flow and multidimensional accounting practices exist in order to measure the performance in the organization. The company can measure all of the dimensions of the matrix functioning with profit-loss charts and can use them to take decisions. All those data are made accessible to the teams planning and implementing the work to be done (Galbraith, ). 7. The planning process is at the same time a conflict resolution process. Planning is also an organizational process aiming to resolve the existing conflicts (Galbraith, ). 8. The company s management uses the appropriate planning software to effect the necessary alignments for doing the work. The teams using multidimensional accounting system do the planning which ends by the definition of goals and objectives (Galbraith, ). 9. Leaders build relationships required to do the work and which add value to the workforce. Leaders organize a variety of activities, work the employees alternately and arrange training activities, which

32 16 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES create value. Developing the cross-relationships effectively in its own unit is evaluated positively in the performance appraisal of a manager (Galbraith, ). Leadership means communicating entirely to the employees their tasks, responsibilities and the required details to do the work. In order to preserve the matrix structure, the manager should give an account of important details concerning the responsibilities and tasks designated to the workforce. Within the context of a team, people to do the work, to implement the decisions, to give an account of the results of a decision or of an activity as well as people to consult or to provide with knowledge must be predefined elaborately (Galbraith, ). The actual work processes and the problems are defined altogether in the enterprise. The problems are defined and solved cooperatively. The different work processes or prospective problems are brought to the leadership team. Most of the time, a planning process gather them and fixes the priorities (Larson and Gobeli, ). A strong leadership team sets a good example for all of the employees. The leadership in the matrix organization, particularly the top management looks like a team sport. Substantial difficulties encountered during the performance of a work is transmitted to the top management. The top management must set a good example for the employees in dealing with tricky situations and resolving the embarrassing problems (Bazigos and Harter, ; Sy and d Annunzio, ). Orientation Main Category: (1) The company possesses high social capital or no silos. (2) People work effectively in teams across organizational boundaries. (3) People are trained in the skills of team problem solving and conflict management. Communication Main Category: (4) The teams are interconnected. (5) Managers in the key roles are naturally collaborative, and they live in values. (6) The leadership builds networks and values networkers. (7) The leadership ensures that roles, responsibilities, and interfaces are clearly defined. (8) Joint escalation processes are known and used intelligently. (9) A strong leadership team sets the example. Coordination Main Category: (10) There are multidimensional accounting systems to track multidimensional performance. (11) The planning process is also the conflict resolution process. (12) There is a spreadsheet planning process that aligns the goals of the dimensions of the matrix. These properties are further classified under three categories, which will form our dependent variables in the context of this research: The communication, orientation and coordination levels perceived by the managers. The independent variables are classified under four headings: The managerial position, the age of the manager, the size of the organization, and the educational level of the managers. Briefly, the variability in the perception of the managers towards the matrix organization structure regarding twelve competency and capability features organized under three main groups required to form the matrix organization structure are measured using one-way variance analysis ANOVA. The one-way ANOVA is considered a robust test against the normality assumption. This means that it tolerates violations to its normality assumption rather well (McClave, Sincich, ). Within the context of this research, first, one sample T-Test is applied. The test value is obtained as 1. The P (significance) value is 0. < is the rejection area of the H0 hypothesis. So, the H0 hypothesis is rejected and H1 hypothesis is accepted. H0 : The managers cannot perceive matrix organization type management properties in the enterprises they are working in. H1 : The managers can perceive matrix organization type management properties in the enterprises they are working in. In all the companies we have examined, the perception level regarding the matrix organization type management of the managers are similar. Sum of Squares Grding df Mean Square F Sig. Between Groups 1, 4,,, Within Groups , , Total , Table 1. Relationship between Orientation and Managerial Position The relationship between orientation variable and managerial position is insignificant ( > ). The managerial position of the manager does not affect the orientation in the enterprise.

33 SESSION 1A: Microeconomics 17 Sum of Squares df Mean Square F Sig. Between Groups 5, , 1,, Within Groups , , Total , Table 2. Relationship between Orientation and Educational Level The relationship between orientation variable and educational level is insignificant ( > ). The educational level of the manager does not affect the orientation in the enterprise. Sum of Squares df Mean Square F Sig. Between Groups, 2,,, Within Groups , , Total , Table 3. Relationship between Orientation and Age The relationship between orientation and the age of the manager is insignificant ( > ). The age of the manager does not affect the orientation in the enterprise. Sum of Squares df Mean Square F Sig. Between Groups, 3,,, Within Groups , , Total , Table 4. Relationship between Orientation and the Size of the Organization The relationship between orientation and the size of the organization is insignificant ( > ). The size of the organization does not affect the orientation in the enterprise. Sum of Squares df Mean Square F Sig. Between Groups 9, , 3,, Within Groups , , Total , Table 5. Relationship between Coordination and the Size of the Organization The relationship between coordination and the size of the organization is significant (< ). When the organization s size is larger, there is more tendency towards a positive perception of coordination function in the enterprise. Sum of Squares df Mean Square F Sig. Between Groups 1, 3,,, Within Groups , , Total , Table 6. Relationship between Coordination and the Educational Level of the Employee The relationship between coordination and the educational level of the manager is insignificant ( > ). The educational level of the manager does not affect the coordination in the enterprise. Sum of Squares df Mean Square F Sig. Between Groups 1, 4,,, Within Groups , , Total , Table 7. Relationship between Coordination and Managerial Position The relationship between coordination and managerial position is insignificant ( > ). The managerial position of the manager does not affect the coordination in the enterprise.

34 18 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Sum of Squares df Mean Square F Sig. Between Groups 2, , 1,, Within Groups , , Total , Table 8. Relationship between Coordination and Age The relationship between coordination and the age of the manager is insignificant ( > ). The age of the manager does not affect the coordination in the enterprise. Sum of Squares df Mean Square F Sig. Between Groups 1, 3,,, Within Groups , , Total , Table 9. Relationship between Communication and the Size of the Organization The relationship between communication and the size of the organization is insignificant ( > ). The size of the organization does not affect the communication in the enterprise. Sum of Squares df Mean Square F Sig. Between Groups 1, 3,,, Within Groups , , Total , Table Relationship between Communication and the Educational Level of the Employee The relationship between communication and the educational level of the manager is insignificant ( > ). The educational level of the manager does not affect the communication in the enterprise. Sum of Squares df Mean Square F Sig. Between Groups 5, , 1,, Within Groups , , Total , Table Relationship between Communication and the Managerial Position of the Employee The relationship between communication and the managerial position of the manager is insignificant ( > ). The managerial position of the manager does not affect the communication in the organization. Sum of Squares df Mean Square F Sig. Between Groups 5, , 2,, Within Groups , , Total , Table Relationship between Communication and Age The relationship between communication and the age of the manager is significant ( = ). As the manager grows older and gets more experienced, he / she perceives more positively the communication in the enterprise. Managerial Size of The Age Educational Level Position Organization Communication Positive Correlation Irrelevant Irrelevant Irrelevant Coordination Irrelevant Irrelevant Irrelevant Positive Correlation Orientation Irrelevant Irrelevant Irrelevant Irrelevant Table One-Way ANOVA Comparisons

35 SESSION 1A: Microeconomics 19 5 Discussion This survey is effectuated with the managers of companies operating in different sectors. In sum, managers have participated to this research. Therefore, in the future, this research has to be reconducted with the managers of companies operating in a specific industry s context. Consequently, the data obtained from this new study could be compared with the results of the present research. 6 Conclusion As the organization gets larger, the age of the manager is affecting more the inclination towards the matrix organization management type. There are three sub-dimensions identified. (1) The communication function. (2) The coordination function. (3) The orientation function. According to the factor analysis, we defined the Cronbach Alpha Coefficient for each of these sub-dimensions as follows: (1) Communication: (2) Coordination: (3) Orientation: Tukey HSD and Bonferroni post-hoc tests are also applied to clarify the relationship between the size of the organization and the perceived coordination function. According to the Bonferroni test, it can be inferred that in local organizations, the perceived coordination is weaker than the regional organizations (the p value (significance) is in local organizations, while it is 1, in regional ones). It has been recorded a 10 % difference between national and regional organizations: It is estimated that the perceived coordination is 10 % weaker in national organizations than in the regional ones. Besides, the perceived coordination in regional organizations is 5% weaker than in local organizations. So we can state that the perceived coordination level is as follows according to the size of the organizations: Local > Regional > National. All along with that, according to the Bonferroni test applied to identify the age of the manager and perceived communication level variables, we have assessed that, between the managers who are years old and the managers who are years old, there is a 10 % increase in the perception level, while between the managers who are years old and the managers who are years old, there is an increase of 5 % in the perception level. Thus, we can assume that, as the managerial experience increases, an increase in the tendency towards the matrix organization type management is recorded. Furthermore, the tendency towards simultaneous treatment of work processes to be defined and problems to be solved, the cooperation degree of the management and the workforce in solving the problems all along with the follow-up of the company s actual position by getting continuous feedback from the employees contacting them regularly and leading them by giving examples are observed and measured. Additionally, the opinion of the employees concerning the change in the management view, their assessments and evaluations are taken into consideration. Acknowledgements The authors would like to thank Hakan Bal, Bahattin Sinsoysal, and Barış Batuhan Geçit for help with the tables. References Amy Kates and Greg Kesler, Activate the Matrix for Integration, People and Strategy, Volume 36 Issue 1, pp Barry Johnson and Mandy Geal, January Matrix Management, Training Journal, pp Brent Durbin, 23 Sep Matrix Organization, Britannica Online, Christopher A. Barlett and Sumantra Ghoshal, July-August Matrix Management: Not a Structure, a Frame of Mind, Harvard Business Review, Volume 68 Issue 4, pp Donald Ralph Kingdon, Matrix Organization Managing Information Technologies; Organizations, People, Society, Tavistock Publications, London. Dylon Moodley and Margie Sutherland, Pieter Pretorius, Comparing the Power and Influence of Functional Managers with that of Project Managers in Matrıx Organisations: the Challenge in Duality of Command, South African Journal of Economic and Management Sciences, Volume 19 Issue 1, pp Erik W. Larson and David H. Gobeli, Summer, Matrix Management: Contradictions and Insights, California Management Review, Volume 29 Issue 4, pp

36 20 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES F. I. H. Pretorius, and R. G. Taylor, September Conflict and Individual Coping Behaviour in Informal Matrix Organizations within the Construction Industry, Construction Management and Economics, Volume 4 Issue 2, pp Feridun Baysak, SASAD, 29 September Interview with General Secretary of Defense and Aviation Industry Manufacturers Association, Ankara. Harvey F. Kolodny, October Evolution to a Matrix Organization, Academy of Management Review. Volume 4 Issue 4, pp James T. McClave and Terry Sincich, Statistics, Tenth International Edition, Pearson Prentice Hall, England. Jay R. Galbraith, Designing Matrix Organizations That Actually Work: How IBM, Procter & Gamble, and Others Design for Success, The Jossey-Bass Business and Management Series, San Francisco. Jay R. Galbraith, February Matrix Organization Designs, Business Horizons, Volume 14 Issue 1, pp John L. Brown and Neil McK. Agnew, November/December The Balance of Power in a Matrix Structure, Business Horizons, Volume 25 Issue 6, pp Lawton Burns and Douglas Wholey, February Adoption and Abandonment of Matrix Management Programs: Effects of Organizational Characteristics and Interorganizational Networks, Academy of Management Journal, Volume 36 Issue 1, pp Len Kinor and Ed Francis, February , Navigating Matrix Management, Leadership Excellence Essentials, Volume 33 Issue 2, pp M. S. Spencer and J. F. Cox, May An Analysis of the Product-Process Matrix and Repetitive Manufacturing, International Journal of Production Research, Volume 33 Issue 5, pp Michael Bazigos, and James Harter, Revisiting the Matrix Organization, McKinsey Quarterly, Issue 4, pp Nick Shreiber and Mike Rosenberg, How to Make the Matrix Work, IESE Insight, 3rd Quarter, Issue 26, pp Paula K. Martin, Matrix Management Reinvented- The New Game in Town, International Matrix Management Publishing, USA. Peter F. Drucker, Management: Tasks, Responsibilities, Practices, Harper and Row Publishing, New York. Richard Burton and M.Børge Obel, Dorthe Døjbak Håkonsson, , How to Get the Matrix Organization to Work, Journal of Organization Design, Volume 4 Issue 3, pp Richard Lepsinger, Influencing In A Team, February Leadership Excellence Essentials, Volume 33 Issue 2, pp Rita E. Numerof and Michael N. Abrams, Summer Matrix Management: Recipe for Chaos?, Directors and Boards, Volume 26 Issue 4, pp Ronald A. Gunn, Matrix Management Success-Method Not Magic, Infinity Publishing, USA. Philip Kotler and Gary Armstrong, Principles of Marketing, Sixteenth Global Edition, Pearson Education, England, Stanley M. Davis and Paul R. Lawrence, May/June, Problems of Matrix Organizations, Harvard Business Review, Volume 56 Issue 3, pp Thomas Sy and Laura Sue D&#;Annunzio, Challenges and Strategies of Matrix Organizations: Top- Level and Mid-Level Managers&#; Perspectives, Human Resource Planning, Volume 28 Issue 1, pp Tom Peters, 20 September Beyond the Matrix Organization, McKinsey Quarterly, Septemper / Warren J. Keegan and Mark C. Green, Global Marketing, Seventh Global Edition, Pearson, England. William F. Joyce, September Matrix Organization: A Social Experiment, Academy of Management Journal, Volume 29 Issue 3, pp

37 SESSION 2A: Growth & Development 21 Seasonal Cointegration Approach on Expenditure Based Gross Domestic Product and Its Some Sub-Components for Turkey Prof. Dr. Mehmet Özmen (Çukurova University, Turkey) Ph.D. Candidate Sera Şanlı (Çukurova University, Turkey) Abstract In this study, it has been aimed to investigate the existence of co-integration relationship between quarterly gross domestic product (GDP), final consumption expenditures of resident households (CONS), exports of goods and services (EXP), government final consumption expenditures (GOV) and private sector machinery-equipment (PRIEQ) series for the period QQ4 for Turkey. Since, Engle and Granger () cointegration test does not take unit roots at seasonal frequencies into account; seasonal cointegration approach proposed by Engle, Granger, Hylleberg and Lee (EGHL) () has been conducted in order to be able to detect the long-run equilibrium relationship among variables which are integrated at the same seasonal frequency. With the aim of determining the stationarity order of series, HEGY seasonal unit root test has been applied. Consequently, there has been found a cointegrating relationship only between GDP and GOV series at quarterly frequencies for only the auxiliary regression including constant term and seasonal dummies. 1 Introduction All the studies regarding time series methods are useful only in case the series in interest do not display seasonal patterns. That is why it is of great importance to take the time series properties of the series like seasonal patterns or trends into account while dealing with economic time series data. The analyses of seasonal unit roots are generally conducted with the most popular approach developed by Hylleberg, Engle, Granger and Yoo () called HEGY by working with different models that include trends, constants and seasonal dummy variables. Cointegrated series are series which are non-stationary alone but stationary in their linear combinations. The concept of seasonal cointegration is valid for models including stochastic seasonals just as the concept of cointegration showing itself in models including stochastic trends (Maddala and Kim, ). One advantage of HEGY test procedure is that it enables to test for unit roots at each frequency separately. So, concerning quarterly data including the four roots which are 1, -1, i ; Engle, Granger, Hylleberg and Lee () propose different levels of seasonal cointegration. In conclusion, cointegration relationship will be analyzed at frequencies in which series are all integrated at the same order. For each frequency, separate cointegration tests are applied. Through seasonal cointegration analysis, whether the variables that are integrated at the same seasonal frequency have a stationary relationship in time is investigated. However, in cointegration test proposed by Engle and Granger (), unit roots at seasonal frequencies are not taken into consideration. In this case, if the presence of seasonal unit roots in series is ignored; the parameter in interest will not be estimated in a consistent way. For this reason, it is more appropriate to conduct seasonal cointegration analysis (Caglayan, ) (Ayvaz Kızılgol, ). As associated with the unit root concept, the relationship between cointegration and error correction models was first suggested by Granger () and then it was also introduced by Granger and Weiss (). Engle and Granger () also offer a theorem based on Granger () which associates the moving average (MA), autoregressive (AR) and error correction representations for cointegrated systems and estimation methods. Engle, Granger and Hallman () and Hylleberg, et al. () introduce the concept of seasonal cointegration in their papers. Kunst () tries to evaluate the effects of modelling seasonal cointegration on predictive accuracy for German and United Kingdom (U.K.) macroeconomic series. HEGY seasonal unit root test has originally been derived for quarterly seasonality and extended to data with different frequencies. Contrary to the DHF test proposed by Dickey, Hasza and Fuller (), HEGY procedure enables to test for unit roots at each seasonal frequency as well as the zero frequency separately and the techniques are applied to quarterly U.K. data for the period to in order to examine the cointegration relationship between consumption and income variables at different frequencies. As a result of application, Hylleberg et al. () find the unit elasticity error correction model to be invalid at any frequency. The asymptotic distributions of the t-statistics from their testable model have been analyzed by Chan and Wei (). In their paper, Chan and Wei () characterize the limiting distributions of the least square estimates as a functional of stochastic integrals. In their paper, Kunst and Franses () deal with the impact of deleting, restricting or not restricting seasonal constants on forecasting seasonally cointegrated time series for Austria, Germany and the U.K. In their paper, Hamori and Tokihisa () analyze the stability of Japanese money demand function using seasonal integration and seasonal cointegration and they find that there exist unit roots in money balances, interest rates and real gross domestic product (GDP) series in different cycles. Because of the rejection of seasonal

38 22 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES cointegration in every case, it is also expressed that there is no stable relationship between money supply and the real economy for the period under consideration. In the study by Lof and Lyhagen (), the comparison of the forecasting performance of the seasonally cointegrated model of Johansen and Schaumburg () and of the specification proposed by Lee () with a parameter restriction included at the annual frequency has been covered. For three data sets from Austria, Germany and U.K., each including six variables: GDP, private consumption, gross fixed investment, goods exports, real wages and the real interest rate; it is also dealt with how the inclusion of restricted or unrestricted seasonal dummies may have an influence in the seasonal cointegration models. Since the semi-annual frequency for Austria appears to have full rank and the U.K. data set shows a rather weak cointegration evidence at the seasonal frequencies, only the German data are used in the forecasting example. Through Monte Carlo study, Lof and Lyhagen () have found some evidence that for the smaller sample sizes the specification of Johansen and Schaumburg () may result in worse forecasts in the case of the inclusion of more cointegrating relations and for larger sample sizes the study results have been found to favour of this specification. In her paper, Caglayan () investigates the presence of seasonal unit root for the monthly series of personal consumption expenditures made to non-durable and semi-durable goods and services, per capita disposable income and stock market returns that are concerned with the life-long permanent income hypothesis over the period and examines if cointegration exists among given variables by using HEGY procedure. In her study, the presence of seasonal unit root has been found in consumption expenditures and disposable income series for both 0 and ¼ frequencies and in stock market returns series for ¼ frequency. Also, it is concluded that consumption expenditures and disposable income variables are cointegrated at zero frequency. In her study, Ayvaz Kızılgöl () has examined if GDP, export, consumption and investment series have seasonal unit roots and display a seasonal cointegration relationship by using quarterly series for the period QQ3 and through Engle, Granger, Hylleberg and Lee () tests, she concluded that there is no seasonal cointegration relationship between series at zero and biannual frequencies. However, a seasonal cointegration relationship has been detected between gross domestic product and consumption series at ¼ (and ¾ frequency) for the model with intercept and seasonal dummy variables. Mert and Demir () have aimed to examine the seasonal patterns to detect if seasonal cointegration relationship exists between export and import series over the quarterly periods. Two series have been found to be cointegrated at ¼ and ¾ frequencies with one cointegrating vector and not cointegrated at zero (long-run) frequency. The results have shown that error correction mechanism works at ¼ frequency. However, at ¾ frequency, because of the error correction term is positive signed contrast to the expectations, the error correction mechanism has been determined not to operate. In this study, it has been aimed to investigate the existence of co-integration relationship between quarterly gross domestic product (GDP), final consumption expenditures of resident households (CONS), exports of goods and services (EXP), government final consumption expenditures (GOV) and private sector machinery-equipment (PRIEQ) series for the period QQ4. The rest of this paper has been organized as follows: Section 2 considers the theoretical approach for seasonal cointegration with single and multiple equations; Section 3 provides the information about the data set and application. Finally, Section 4 presents the brief conclusions. 2 Theoretical Approach for Seasonal Cointegration The concept of seasonal cointegration is valid for models including stochastic seasonals just as the concept of cointegration showing itself in models including stochastic trends (Maddala and Kim, ). One advantage of HEGY test procedure is that it enables to test for unit roots at each frequency separately. So, concerning quarterly data including the four roots which are 1, -1, i ; Engle et al. () propose different levels of seasonal cointegration. Assume that y t and z t series are seasonally cointegrated so that 4 yt and z 4 t are stationary. When these two series have a common non-seasonal unit root (that is, they are cointegrated at long-run zero frequency at root 1), we have the error term u (1 L L L ) y (1 L L L ) z (1) t t 1 which is stationary. If seasonal cointegration exists at frequency ½ corresponding to unit root 1, we have v (1 L L L ) y (1 L L L ) z (2) t t 2 which is stationary (so, it does not require ( 1 L) filter to be stationary) and finally if seasonal cointegration exists at frequency ¼ corresponding to unit roots w t 2 i and (1 L ) filter we have ( 1 L ) yt 3(1 L ) zt 4(1 L ) yt 1 5(1 L ) zt 1 (3) t t

39 SESSION 2A: Growth & Development 23 is stationary. In case all three series in a simple form as u t, v t and w t are stationary, the seasonal cointegration model is represented 4 yt 11ut 1 21vt 1 31wt 2 41wt 3 1 t 4zt 21ut 1 22vt 1 32wt 2 42wt 3 2t (4) (5) where s represent the error correction terms. In addition; constant, seasonal dummies and trend variables can be incorporated into these equations. This method with two-step proposed by Engle et al. () is similar to the Engle-Granger approach applied for nonseasonal time series: in the first step, equations (1) to (3) are estimated by ordinary least squares (OLS) procedure and in the second step Augmented Dickey Fuller (ADF) unit root tests are applied to û, vˆ and ŵ (in other words, this transaction allows us to check if estimated residuals û to ŵ t t t are stationary). The tests for û t and vˆ t have the same critical values as those in Engle and Granger (). However, critical values for testing ŵ are different. For this case, the critical values are tabulated in Engle et al. () (Maddala and Kim, ). t As mentioned above, subsequent to estimating 1 to 5 by OLS for bivariate time series involving z, the stationarity condition is checked for estimated residuals t auxiliary regressions: (Lof, ). (1 L) uˆ (1 L)ˆ v 2 (1 L ) wˆ t t t l1 1uˆ 1 (1 ) ˆ t i L ut i i 1 û t to t t y t and ŵ t. This is executed by using the following (6) l2 2 ( vˆ 1 ) (1 ) ˆ t i L vt i i 1 t (7) t 2 t 1 3 i 1 t 2 ( wˆ ) ( wˆ ) (1 L ) wˆ (8) 3 4 As seen here, the lagged dependent variables may be added to these auxiliary regressions given above. To detect the cointegration at the zero and semi-annual frequencies, t-statistic values of 1 and 2 should be compared to the critical values in the paper of Engle and Yoo () and the null hypotheses of no cointegration at zero frequency and no cointegration at ½ frequency should be tested for two auxiliary regressions in (6) and (7). On the other hand, for ¼ (and ¾ frequencies), F( 0 ) test statistic value should be compared to the critical 3 4 values which take place in the paper of Engle et al. () and here the null hypothesis should be constructed as H : No cointegration at ¼ (and ¾) frequencies for the third auxiliary regression given in (8) (Mert and Demir, ). 3 Data Set and Application In this application, it has been aimed to investigate the existence of co-integration relationship between quarterly gross domestic product (GDP), final consumption expenditures of resident households (CONS), exports of goods and services (EXP), government final consumption expenditures (GOV) and private sector machinery-equipment (PRIEQ) series for the period QQ4. Data that are based on expenditure based GDP time series at fixed prices have been obtained from Central Bank of the Republic of Turkey (CBRT). First, in order to linearize the exponential growth in these series, their logarithms have been taken. Since by taking logarithm, variance is stabilized and the effects of outliers are reduced (Ture and Akdi, ). In order to determine which series have a cointegrating relationship, it is necessary to find out at which frequencies series are integrated of the same order (or at which frequencies unit roots are present). For each series, three different models including constant (C), constant+dummies (C,D) and constant+dummies+trend (C, D, T) have been constructed. Also, the lagged values of the dependent variable have been added into these models. Since the series discussed are at quarterly frequency, seasonal unit root test results of the series at frequencies have been presented in Table 1. l i t i t 0, 1 1 3,, 4 2 4

40 24 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Variables Auxiliary Regressions t 1 t t 2 3 t 4 Lags ( ) ( ) ( ) ( ) LNGDP Intercept * * * * * Intercept + Dummies * * * * Intercept + Dummies * * * * Trend LNCONS Intercept * Intercept + Dummies * * * Intercept + Dummies * * * Trend LNPRIEQ Intercept * Intercept + Dummies * * Intercept+ Dummies * * Trend LNGOV Intercept * * * * * Intercept + Dummies * * * * Intercept + Dummies * * * * * Trend LNEXP Intercept * * * * Intercept+ Dummies * * Intercept + Dummies+ Trend * * * Notes. 1 * denotes insignificant values at 5% level. 2 t-statistics for t ) shows whether there is a unit root or not at zero frequency H : 0). t-statistics 1 ( 1 ( 0 1 ( H ( 4 for t ) tests the presence of the semi-annual unit root : 0). F- statistic for F( 3, )) tests 2 ( 2 whether there is a unit root at quarterly frequency or not. 3 Critical values have been taken from Hylleberg et al. () for N= observations and 5% level. For zero frequency, critical values are , , and for semi-annual frequency, they are , , respectively for intercept, intercept+dummies, intercept+dummies+trend models. Table 1. HEGY Seasonal Unit Root Test Results for Quarterly Macroeconomic Series F 3, ) ( 4 In the application for seasonal unit root test, the appropriate lag length has been chosen in that way: Regression equation has been estimated first with Lag 1 and it has been investigated if first order and fourth order autocorrelations exist between residuals. For this investigation, it has been utilized from Lagrange Multiplier (LM) test statistics (thus for first order: LM(1) and for fourth order: LM(4)). If any one of the null hypotheses of H 0 : There is no 1st order autocorrelation and There is no 4th order autocorrelation is rejected, lag length has been increased by one and LM test has been applied again. This process has been continued until the null hypothesis cannot be rejected for each order and homoscedastic residuals are obtained. LM(1) and LM(4) statistics results have been presented in Table 2: H 0 :

41 SESSION 2A: Growth & Development 25 VARIABLES AUXILIARY REGRESSIONS LAGS LM(1) LM(4) LNGDP Intercept () () Intercept + Dummies () () Intercept + Dummies+ Trend () () LNCONS Intercept () () Intercept + Dummies () () Intercept + Dummies+ Trend () () LNPRIEQ Intercept () () Intercept + Dummies () () Intercept + Dummies+ Trend () () LNGOV Intercept () () Intercept + Dummies () () Intercept + Dummies+ Trend () () LNEXP Intercept () () Intercept+Dummies () () Intercept + Dummies+ Trend () () Note. LM(1) and LM(4) represent LM test statistics investigating the presence of 1 st and 4 th order autocorrelations and the values given in parentheses indicate the significance levels. Table 2. LM(1) and LM(4) Statistics for Quarterly Macroeconomic Series As it is clear from Table 2 that for selected lags, there are no first order and fourth order autocorrelation problems for all macroeconomic series. If looked at the Table 1 results, it is seen that the presence of a unit root at zero frequency has been accepted for all variables in all three auxiliary regression models. When t ( 1 ), t ( ( 2 ) and F 3, 4 ) columns are examined, it is concluded that the null hypotheses that there is a (non-seasonal) unit root at zero frequency and there are seasonal unit roots at other seasonal frequencies cannot be rejected for three auxiliary regression models of LNGDP series at 5% significance level. Thus, LNGDP series has a non-seasonal unit root at zero frequency and seasonal 1 unit roots at semi-annual ( 2 frequency) and quarterly frequencies. While both LNCONS and LNPRIEQ series have the zero frequency unit root for three models with deterministic components given in Table 1, according to the results they both do not include any annual unit root (at quarterly frequency). For LNCONS series, the presence of semi-annual unit root has been accepted for two models except the intercept model. However, no semi-annual unit root has been detected in any model for LNPRIEQ series. When looked at the LNGOV and LNEXP series, both series are seen to include the zero frequency unit root. However, while LNGOV series has a seasonal unit root at semi-annual frequency for two models except the intercept+dummies model, LNEXP series rejects the presence of the semi-annual unit root for all three models. Finally, while LNGOV series has annual unit roots at 1 3 quarterly 4 4 frequencies for all deterministic models, LNEXP series has seasonal unit roots at quarterly frequencies for two models except only intercept+dummies model. In conclusion, cointegration relationship will be analyzed at frequencies in which these series are both integrated at the same order. In this case, it is necessary to determine which series are integrated of the same order at which frequencies. In all series, the presence of the zero frequency unit root has been detected in common. LNGDP, LNCONS and LNGOV series have been found to include seasonal unit root at semi-annual frequency. On the other hand, it has been determined that LNGDP, LNGOV and LNEXP series include seasonal unit roots at the quarterly frequencies 1 3. The results of seasonal 4 4 cointegration analyses of the series at 0, ½ and ¼ (and ¾) frequencies have been presented in Table 3, Table 4 and Table 5 respectively.

42 26 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES In this application, regression models obtained from the linear components of the variables that are integrated at the same frequency have been estimated through OLS procedure. Before applying to cointegration analysis, it is necessary to give the transformations of variables that will be used in cointegration models. As a matter of example, it will be sufficient to present only LNGDP series (the other series will be transformed in the same way with LNGDP): Now let us mention about the cointegration models to be used. Seasonal cointegration has been mentioned in section In addition, as also summarized by Ayvaz Kızılgöl (), in cointegration analysis the regression model to be estimated for all variables that are integrated of the same order at the zero frequency is Y Z u. The residuals ( u ) obtained from this cointegration model will be used in order to estimate 1t 1 1t t t auxiliary regression model at the zero frequency. Regressand LCONS1t LCONS1t LCONS1t Regressand LPRIEQ1t LPRIEQ1t LPRIEQ1t Coefficient Regressor LGDP1t ( ) ( ) ( ) Coefficient Regressor LGDP1t ( ) ( ) ( ) 2 3 LNGDP t (1 L L L ) LNGDP LNGDP t (1 L L L ) LNGDP 2 2 LNGDP t (1 L ) LNGDP 3 4 LNGDP t (1 L ) LNGDP 4 Cointegration Analysis: LGDP and LCONS Deterministic Components Included Auxiliary Regression Tests for Unit Roots in Residuals (9) (10) (11) (12) t statistic 2 R Augmentation DW t ( 1 ) C , 4, C, D C, D, T Cointegration Analysis: LGDP and LPRIEQ Deterministic Components Included Auxiliary Regression Tests for Unit Roots in Residuals t statistic 2 R Augmentation DW t ( 1 ) C , 4, C, D C, D, T Cointegration Analysis: LGDP and LGOV , 2, , 4, 5, Auxiliary Regression Tests for Unit Roots in Residuals Regressand LGOV1t LGOV1t LGOV1t Coefficient Regressor LGDP1t ( ) ( ) ( ) Deterministic t statistic Components 2 R Augmentation DW t ( 1 ) Included C , 3, C, D , C, D, T Table 3. Cointegration Test Results at Zero (Long Run) Frequency

43 SESSION 2A: Growth & Development 27 Cointegration Analysis: LGDP and LEXPORT Auxiliary Regression Tests for Unit Roots in Residuals Regressand LEXPORT1t LEXPORT1t LEXPORT1t Coefficient Regressor LGDP1t ( ) ( ) ( ) Deterministic Components Included t statistic 2 R DW Augmentation t ( 1 ) C , 2, C, D C, D, T 1, 2, , Notes. 1 These tests at zero frequency are based on the (ADF) auxiliary regression model u t u 1 t 1 k j 1 b u t j e (without deterministic components) where represents the residuals obtained from cointegration model that are used to estimate this auxiliary regression model. The distribution of t statistic is as characterized in Engel and Granger () and Engle and Yoo () (Engle, et al., ). As it is clear, the necessary significant lagged values of the dependent variable j have been added into auxiliary regression model in order to whiten the residuals (the lagged variables with insignificant coefficients at 5% significance level have been removed from the model). 2 The values in parentheses are t-statistics. t 3 C, D and T denote constant, seasonal dummies and trend terms respectively. 4 The basic hypothesis to be tested is H : There is no cointegration at zero frequency ( 1 0). 5 Critical values have been obtained from Engle and Yoo (). See Appendix. Table 3 (Continued) For semi-annual (½) frequency, the cointegration model to be used is 0 u t and for quarterly frequencies, it is Y Z Z w. Also, the residuals obtained from these models ( and ) will 3t 1 3t 2 3, t 1 t be used for estimating auxiliary regressions at specified frequencies respectively. As mentioned before, in order to detect the long-run equilibrium relationship between the series, first of all it is necessary to determine the stationarity order of the series. In this application, for investigating the presence of seasonal cointegration relationship between the series, firstly seasonal unit root test has been applied in order to make inference about at which frequencies there are unit roots if they exist. The series discussed here have quarterly frequencies. Therefore, HEGY seasonal unit root test which is developed by Hylleberg et al. () has been applied in order to detect seasonal unit roots and general results have been presented in Table 1 for three models with deterministic components that are C, C,D, C,D,T. Now, Table 3 presents the cointegration test results at zero frequency. As a result, when cointegration test results are evaluated at the zero frequency, although the explanatory variables that take place in the cointegrating regression have been found to be statistically significant, no cointegrating relationship has been found between LNGDP and LNCONS, LNGDP and LNPRIEQ, LNGDP and LNGOV, LNGDP and LNEXP at 5% significance level in the long-run. ut Y 2t 2 Z 2t v t v t w t

44 28 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Cointegration Analysis: LGDP and LCONS Auxiliary Regression Analysis of the residuals Regressand Coefficient Regressor LGDP2t Deterministic Components Included 2 R Augmentation DW t statistic t 2 LCONS2t C, D , ( ) LCONS2t C, D, T , ( ) Cointegration Analysis: Auxiliary LGDP and LGOV Regression Analysis of the residuals ( ) Regressand Coefficient Regressor LGDP2t Deterministic Components Included 2 R Augmentation DW t statistic LGOV2t C , 2, ( ) LGOV2t ( ) C, D, T Notes. 1 In lag augmentations, only significant lags have been added into the auxiliary regressions (insignificant lags have been removed) in order to get white noise residuals. 2 These tests at semi-annual frequency are based on the auxiliary regression k t 1 ) 2 ( t 1 ) b j ( t j t j 1 ) j 1 ( t e t (here without deterministic components) where ( ) represents the residuals obtained from cointegration model that are used to estimate the auxiliary regression models. The distribution of t statistic is as characterized in Engel and Granger () and Engle and Yoo () (Engle, et al., ). For critical values see Appendix. 3 The basic hypothesis to be tested is H : There is no cointegration at semi-annual frequency ( 2 0) 0 Table 4. Seasonal Cointegration Test Results at Semi-Annual (½) Frequency t 2 LNGDP and LNCONS series have been found to be integrated of the same order for C,D and C,D,T models at ½ frequency. Also, LNGDP and LNGOV series have been found to be integrated of the same order for C and C,D,T models at ½ frequency. Therefore, cointegration analysis results at ½ frequency have been shown in Table 4 for LNGDP, LNCONS and LNGOV series. When Table 4 results are compared to the Engle and Yoo () critical values for 5% significance level, no cointegration relationship has been found between LNGDP & LNCONS series and LNGDP & LNGOV series at ½ frequency. Thus, these series in interest do not seem to be cointegrated at the semi-annual frequency. t Cointegration Analysis: LGDP and LGOV Auxiliary Regression Analysis of the residuals Regressand LGOV 3t ( ) LGOV 3t ( ) LGOV 3t ( ) Coefficient Regressor LGDP 3t LGDP 3t ( ) ( ) ( ) Deterministic Components Included Augmen- t statistic 2 tation R t ( 3 ) t statistic t 4 ( ) F statistic C C, D , * * , Table 5. Seasonal Cointegration Test Results at ¼ (¾) Frequencies 3 4

45 SESSION 2A: Growth & Development 29 Cointegration Analysis: LGDP and LEXPORT Auxiliary Regression Analysis of the Residuals HEGY test Regressand Coefficient Regressor LGDP 3t LGDP 3t-1 Deterministic Components Included Augmentation t statistic t ( 3 ) t statistic ( ) F statistic LEXPORT 3t C , 4, 6, ( ) ( ) LEXPORT 3t C, D , * ( ) ( ) LEXPORT 3t ( ) ( ) , 4, 5, Notes. 1 These tests at ¼ (and ¾) frequencies are based on the auxiliary regression k ( w (here without deterministic components) where t wt 2 ) 3 ( wt 2 ) 4 ( wt 1 ) b j ( wt j wt j 2 ) e w t t j 1 represents the residuals obtained from cointegration model that are used to estimate the auxiliary regression models (Engle, et al., ). 2 C denotes constant, D denotes seasonal dummies and denotes no deterministic component. 3 * denotes significant values at 5% significance level. 4 Critical values have been obtained from Engle et al. (). See Appendix for critical values. 5 The basic hypothesis to be tested is H : There is no cointegration at ¼ (and ¾) frequencies 0). Table 5 (Continued) Table 5 presents seasonal cointegration test results at quarterly ¼ (¾) frequencies. According to the Table 5 results, it can be said that there has been found a cointegration relationship between LNGDP and LNGOV series at quarterly frequencies ¼ (and ¾) for only the model with constant and seasonal dummies ( C,D ). In other saying, the null hypothesis saying that there is no cointegration at quarterly frequencies has been rejected with a significant joint F statistic of On the other hand, no cointegration relationship has been detected for no models between LNGDP and LNEXP series at ¼ (and ¾) frequencies. 4 Conclusion 0 In this paper, whether a cointegration relationship exists or not between quarterly GDP, CONS, EXP, GOV and PRIEQ series has been investigated. As a result of HEGY application, the presence of a zero frequency (nonseasonal) unit root has been detected for all series for the three models with constant, constant+dummies and constant+dummies+trend. LNGDP, LNCONS and LNGOV series have been found to include a seasonal unit root at semi-annual frequency. In addition, LNGDP, LNGOV and LNEXP series have been detected to have 1 3 seasonal unit roots at quarterly 4 4 frequencies. It should be noted that cointegration analysis should be evaluated among the series having unit roots at the same frequency. When cointegration test results are evaluated thoroughly at the zero (long-run) frequency, there has been found no cointegrating relationship between LNGDP & LNCONS, LNGDP & LNPRIEQ, LNGDP & LNGOV, LNGDP & LNEXP at 5% significance level. Similarly, no cointegrating relationship has been detected between LNGDP&LNCONS series and LNGDP&LNGOV series at semi-annual (½) frequency. However, there has been found a cointegrating relationship between LNGDP & 1 3 LNGOV series at quarterly 4 4 frequencies for only the model with constant+dummies. On the other hand, no cointegrating relationship has been found between LNGDP & LNEXP series for no models at these quarterly frequencies. References Ayvaz Kızılgol, Mevsimsel Eşbütünleşme Testi: Türkiye&#;nin Makroekonomik Verileriyle Bir Uygulama, Ataturk University - Journal of the Faculty of Economic and Administrative Sciences, 25, pp Caglayan, Yaşam Boyu Sürekli Gelir Hipotezinde Mevsimsellik, Marmara University, Journal of the Faculty of Economic and Administrative Sciences, 18, pp Chan and Wei, Limiting Distributions of Least Squares Estimates of Unstable Autoegressive Processes, Annals of Statistics, 16, pp Cubadda, Complex Reduced Rank Models for Seasonally Cointegrated Time Series, Oxford Bulletin of Economics and Statistics, 63, pp R t ( 3 4

46 30 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Dickey, Hasza and Fuller; Testing for Unit Roots in Seasonal Time Series, Journal of the American Statistical Association, 79, pp Engle and Granger, Co-integration and Error Correction: Representation, Estimation and Testing, Econometrica, 55, pp Engle and Yoo, Forecasting and Testing in Co-Integrated Systems, Journal of Econometrics, 35, pp Engle, Granger and Hallman; Merging Short and Long Run Forecasts: An Application of Seasonal Cointegration to Monthly Electricity Sales Forecasting, Journal of Econometrics, 40, pp Engle, Granger, Hylleberg and Lee; Seasonal Cointegration: The Japanese Consumption Function, Journal of Econometrics, 55, pp Granger, Some Properties of Time Series Data and Their Use in Econometric Model Specification, Journal of Econometrics, 16, pp Granger, Cointegrated Variables and Error Correction Models (UCSD Discussion Paper a), University of California, San Diego. Granger and Weiss, Time Series Analysis of Error-Correction Models, in Samuel Karlin, Takeshi Amemiya and Leo A. Goodman (Eds.), Studies in Econometrics, Time Series and Multivariate Statistics, Academic Press, New York, pp Hamori and Tokihisa, Seasonal Cointegration and the Money Demand Function: Some Evidence from Japan, Applied Economics Letters, 8, pp Hylleberg, Engle, Granger and Yoo; Seasonal Integration and Cointegration, Journal of Econometrics, 44, pp Johansen and Schaumburg, Likelihood Analysis of Seasonal Cointegration, Journal of Econometrics, 88, pp Kunst, Seasonal Cointegration, Common Seasonals and Forecasting Seasonal Series, Empirical Economics, 18, pp Kunst and Franses, The Impact of Seasonal Constants on Forecasting Seasonally Cointegrated Time Series, Journal of Forecasting, 17, pp Lee, Maximum Likelihood Inference on Cointegration and Seasonal Cointegration, Journal of Econometrics, 54, pp Lof, On Seasonality and Cointegration, Ph. D. Dissertation, The Economic Research Institute - Stockholm School of Economics, Elanders, Gotab. Lof and Lyhagen, Forecasting Performance of Seasonal Cointegration Models, International Journal of Forecasting, 18, pp Maddala and Kim, Unit Roots, Cointegration and Structural Change, Cambridge University Press, Cambridge. Mert and Demir, Mevsimsel Eşbütünleşme ve Mevsimsel Hata Düzeltme Modeli: İthalat - İhracat Verileri Üzerine Bir Uygulama, Suleyman Demirel University Journal of Faculty of Economics & Administrative Sciences, 19, pp Sanli, The Econometric Analysis of Seasonal Time Series: Applications on Some Macroeconomic Variables, Master s Thesis, Cukurova University, Adana. Ture and Akdi, Mevsimsel Kointegrasyon: Türkiye Verilerine Bir Uygulama, Paper presented at the VII. National Econometrics and Statistics Symposium, May, Istanbul University. Information Notes This study has been derived from the Master Thesis that has been prepared in consultancy of Assoc. Prof. Mehmet Ozmen called The Econometric Analysis of Seasonal Time Series: Applications on Some Macroeconomic Variables (Sanli, ). This study has been supported by TUBITAK (The Scientific and Technological Research Council of Turkey) BIDEB (Scientist Support Department) within the scope of E Direct National Scholarship Programme for PhD Students. This study has been supported by Cukurova University Rectorate Department of Scientific Research Projects (BAP).

47 SESSION 2A: Growth & Development 31 Appendix: Critical Values for Seasonal Cointegration (for observations) Number of Variables (k=5, N=) ve 2 Significance Level 1% 5% 10% Critical Value Table 6. Critical Values for Seasonal Cointegration at Zero and Semiannual Frequencies Source. Engle and Yoo, N= Deterministic Component 1% 5% 10% 1% 5% 10% 99% 95% 90% C C, D Table 7. Critical Values for Seasonal Cointegration at ¼ (and ¾ ) Quarterly Frequencies Source. Engle, et al.,

48 32 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES On the Purchasing Power Parity in Turkey: The Role of Structural Changes Assoc. Prof. Dr. Şaban Nazlıoğlu (Pamukkale University, Turkey) Ph.D. Candidate Çağın Karul (Pamukkale University, Turkey) Ph.D. Candidate Ahmet Koncak (Abant İzzet Baysal University, Turkey) Assoc. Prof. Dr. İlhan Küçükkaplan (Pamukkale University, Turkey) Abstract Turkey as an emerging country and one of the fastest growing economies during the last decade has been implementing the trade-oriented growth model since The exchange rate policy in that respect is at the center of trade and monetary policies. Given the importance of constructing fundamental equilibrium exchange rates, the long-run PPP hypothesis has been empirically investigated during the last decade. We re-examine the purchasing power parity (PPP) hypothesis for Turkey with her ten major trading partners and find out that when the structural shifts are taken into account, there is a strong evidence in favor of the validity of PPP hypothesis. An interesting finding also is that the PPP hypothesis seems to hold for the European Union countries. 1 Introduction Given the importance of determining equilibrium exchange rates, a special attention has been given to testing for the long-run PPP hypothesis during the last decade. The PPP hypothesis implies a stationary real exchange rate series and therefore researchers benefit from the unit root analysis. Turkey which is an emerging market and a fastly growing economy (during the last decade) has put into effect the trade-oriented growth model in that leads the exchange rate policy to be at the core of trade and monetary policies. Turkey decided to shift from pegged to flexible exchange rate system because of the economic crisis and hence the dynamics of Turkish exchange rates play a crucial role for sound monetary policy. The literature on the persistence of shock to Turkish real exchange rates shows that there is no consensus whether the shocks are permanent. The lack of consensus on PPP hypothesis holds for Turkey motivate us to re-examine the PPP hypothesis within the context of new testing procedures. This study re-analyzes the validity of PPP hypothesis in Turkey by employing a battery of unit root and stationarity tests. This study contributes to the literature by looking at the validity of PPP in Turkey by focusing on modelling the structural changes. Empirical analysis indicates that there is a strong evidence in favor of the validity of PPP hypothesis when the structural shifts are taken into account in testing procedures. It is important to emphasize that even though this finding is not new for international trade literature, it provides a new empirical evidence with respect to the dynamics of exchange rates in Turkey. 2 Testing Procedures Unit Root Tests To test for unit root of Turkish exchange rates, we first start with the Dickey-Fuller (DF) test developed by Dickey and Fuller () and estimate the no-shift regression model as y t = Z t δ + αy t 1 + ε t (1) where y t is the first difference of y t, Z t includes the deterministic terms defined by [1, t], and ε t is the error term. The null hypothesis of unit root (H 0 : α = 0) is tested against the alternative hypothesis of stationarity (H 1 : α < 0). Accordingly, if the null hypothesis is rejected, then an evidence is supported in favor of PPP hypothesis. The test statistic denoted by τ is defined by the t-ratio of α. Under the null hypothesis, the t-ratio corresponds to y t 1 does not follow the asymptotic t-distribution and therefore one needs to use the critical values provided by Dickey and Fuller (). In the no-shift model, Z t is assumed not to have any structural changes. However, ignoring structural shifts leads to misleading inferences because of incorrectly retaining a false unit root null hypothesis (Perron, ). In order to handle this problem, Zivot and Andrews () allow a sudden structural break in the DF test. To formulize the sharp-shift model with one break, Z t is described as [1, t, DU 1t, DT 1t ] where DU 1t = 0 for t T B and 1 otherwise and DT 1t = 0 for t T B and t T B otherwise that T B denotes the break date. Narayan and Popp () extend Zivot and Andrews () s sharp shift model for two sudden breaks that Z t becomes [1, t, DU 1t, DT 1t, DU 2t, DT 2t ], where DU it = 0 for t T Bi and 1 otherwise and DT it = 0 for t T Bi and t T Bi otherwise and T Bi (i = 1,2) shows the break dates. The statistic for testing the null hypothesis of unit root with structural shifts is described as in the DF test. Both Zivot and Adrews () and Narayan and Popp () approaches use the dummy variables to capture sudden structural changes and require estimating the break dates. The location of break (T Bi ) is endogenously determined to be where the test statistic is minimized (i.e., the most negative) by a grid search

49 SESSION 2A: Growth & Development 33 procedure by considering all possible break points as δ = infτ(δ) where δ = T Bi /T and δ [0,1]. Finally, τ statistic with structural shifts does not follow the asymptotic t-distribution and hence the simulated critical values are used. It is worthwhile noting that the sharp-shift models entail knowing a priori the number, dates and form of breaks. In practice, it is however difficult to have such a priori knowledge and moreover economic series may contain multiple smooth breaks at unknown dates. More recently, Enders and Lee () propose the smooth-shift model type of the DF unit root test by using a Fourier approximation for Z t which does not require selecting the dates, number, and form of the breaks. The Fourier expansion for Z t is described as [1, t, sin( 2πkt ), cos(2πkt)] where k T T represents an integer frequency. The test statistic is again described as in the DF test, but its distribution now depends on k that requires using the critical values for different values of the Fourier frequency (see, Enders and Lee, ). It is finally should be emphasized that the smooth-shift model testing procedure now requires determining the Fourier frequency and the number of lags. Following Enders and Lee (), we apply the generalto-specific approach. The maximum number of Fourier frequency is set to 3 and the maximum number of lags is set to 4. We first determine the optimal lag for each of frequency with the significance of the last lagged term by looking at its t-statistic at the 10 percent level. The optimal number of Fourier frequency component is then selected by the minimization of sum of squared residuals of the regression model. Stationarity Tests Testing the null hypothesis of stationarity against the alternative of unit root is one of ongoing research areas in the time series analysis. The stationarity test of Kwiatkowski et al. () (the KPSS test) is widely utilized which is based on the data generating process (DGP) y t = γ Z t + r t + ε t (2) r t = r t 1 + u t (3) where t=1,,t time dimension, r t is random walk process with initial values r 0 = 0. The null hypothesis of stationarity H 0 : σ 2 u = 0 is tested against the alternative hypothesis of unit root. H 1 : σ 2 u > 0. In the conventional KPSS test, Z t is assumed not to have any structural changes and is defined as Z t = [1, t] for a trend stationary process. Lee et al. () investigates the distribution of the KPSS test under a structural break and find out that the test diverges from the distribution under the null hypothesis when the structural break is ignored. The simulations moreover indicate that the KPSS statistics is over-sized in the case of a structural shift. Kurozumi () allows a break in the KPSS test by defining Z t as [1, t, DU 1t, DT 1t ] where DU 1t = 0 for t T B and 1 otherwise and DT 1t = 0 for t T B and t T B otherwise that T B denotes the break date. Carrion-i-Silvestre and Sansó () (hereafter CS) extends the testing framework for two breaks that Z t becomes [1, t, DU 1t, DT 1t, DU 2t, DT 2t ], where DU it = 0 for t T Bi and 1 otherwise and DT it = 0 for t T Bi and t T Bi otherwise and T Bi (i = 1,2) show the break dates. Both Kurozumi and CS approaches use the dummy variables to capture structural shifts and require estimating the location of break (T Bi ) which is endogenously determined by minimizing the test statistic by considering all possible breaks points as δ = infτ(δ) where δ = T Bi /T and δ [0,1]. As discussed earlier, the dummy variable approach assumes sharp shifts and also entails to know a priori the number and dates of breaks. In practice, it is difficult to have a such a prior information and moreover economic series may contain an unknown form breaks at unknown dates. Becker et al. () (hereafter BEL) use a Fourier approximation for Z t which does not require selecting the dates, number, and form of the breaks. The Fourier expansion for Z t is described as [1, t, sin( 2πkt ), cos(2πkt)] where k represents an integer frequency. The testing T T procedure now requires determining the Fourier frequency. Following Becker et al. (), we apply the generalto-specific approach. The maximum number of Fourier frequency is set to 3 and the optimal number of Fourier frequency component is then selected by the minimization of sum of squared residuals of the regression model. T t=1 S 2 t The stationarity statistic is defined as LM = t where S t = j=1 ε j is the partial sum process by using the σ ε2 OLS residuals from equation (1) and σ ε2 is an estimate of the long-run variance of ε t that is defined as σ 2 ε = lim T T 1 E(S 2 T ) (Note 1). Since LM statistics does not follow the asymptotic t-distribution and hence the simulated critical values are used. 3 Empirical Findings We collect the real exchange rates data between Turkey and Canada, Denmark, Eurozone, Japan, Norway, Saudi Arabia, Sweden, Switzerland, United Kingdom, and USA for the period January December (Note 2). The period is started from January because the euro banknotes and coins were put into circulation from that time. We describe the real exchange rate by y it = e it + p it p t where e is the natural logarithm of nominal

50 34 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES exchange rates, p is the natural logarithm of foreign CPI and p is the natural logarithm of domestic CPI which are based the indexes (=). The data is retrieved from the International Financial Statistics (IFS). Table 1 reports the results from the unit root and stationarity tests. The Panel A is for the results from the unit root tests. The no-shift model indicates that the PPP hypothesis seems to be valid only for Norway. When we account for the one-sharp structural change, the PPP is supported in six trading partners (Denmark, Japan, Norway, Sweden, Switzerland, and UK). Moreover, if the number of structural changes raises to two-sharp shifts, the PPP appears to be valid for all countries. If the structural changes are modelled as a gradual/smooth process, the PPP is supported for six countries (Denmark, Eurozone, Norway, Sweden, Switzerland, and UK) which are all the European Union countries during the sample period. Testing null hypothesis of stationarity is useful to confirm results those from the tests with the null hypothesis of unit root. The conventional KPSS test shows that the null hypothesis of stationarity cannot be rejected in for countries (Canada, Denmark, Eurozone, and Norway), implying that the PPP holds. If we account for sharp structural change, while the PPP is supported in nine countries partners (except UK) for the one-shift model, it is supported in all the countries for the two-shift model. If the structural changes are modelled as a gradual/smooth process, the PPP is rejected only for two countries (Japan and Sweden). The model with sharp shifts assume that the trend is not strictly linear and hence is composed to a pre-specified number of discrete linear segments. Moreover, there may be a discontinuity at the breakpoint because the end and start of linear segments do not need to be concur. Last but not least, the sharp shift assumption requires an immediate transition from one segment to another. An arguably more realistic approach is to accommodate structural changes as a gradual process (King and Dobson, ). From this point of view, the results taken together from both the unit root and stationarity tests based on the smooth-shift model provide an interesting finding is that the PPP hypothesis seems to hold for Denmark, Eurozone, Norway, Switzerland, and UK which are the European Union countries and is not valid for Canada, Japan, Saudi Arabia, Sweden, and USA. No-shift Sharp-shift Smooth-shift Panel A: Unit root tests Stat. Stat. Break Stat. First Break Second Break Stat. k Canada M M M Denmark M M M Eurozone M M M Japan M M M Norway M M M Saudi Arabia M M M Sweden M M M Switzerland M M M UK M M M USA M M M Panel B: Stationarity tests Canada M M M Denmark M M M Eurozone M M M Japan M M M Norway M M M Saudi Arabia M M M Sweden M M M Switzerland M M M UK M M M USA M M M Bold numbers denote that PPP hypothesis is valid at least at 10% level of significance. 4 Conclusion Table 1: Empirical Results We examine the PPP hypothesis in Turkey based on the unit root framework and benefits from the recent developments in time series testing procedures by paying attention to controlling for structural shifts. We first estimate the no-shift model and then employ its extensions to which accounting for structural breaks, namely the sharp-shift model and the smooth-shift model. The results from the point of modelling strategy imply that controlling for structural shifts plays an important role in order to determine the behavior of Turkish exchange rates. The different testing strategies result in the different inferences and policy implications. The conventional noshift model unit root and stationarity models imply the PPP is valid only for Norway. In contrast to the no-break test, the sharp shift tests based two-break model description show the strong evidence on the PPP. Finally, the testing procedure -based on the assumption that any structural shifts are gradual in nature- supports that the PPP hypothesis seems to hold for Denmark, Eurozone, Norway, Switzerland, and UK which are the European Union countries and is not valid for Canada, Japan, Saudi Arabia, Sweden, and USA.

51 SESSION 2A: Growth & Development 35 Acknowledgements Saban Nazlioglu gratefully acknowledges that this study is carried out under the Outstanding Young Scientists Award Program of the Turkish Academy of Sciences (TÜBA-GEBİP ). References Adiguzel, U., Sahbaz, A., Ozcan, C.C., Nazlioglu, S. (). The Behavior of Turkish Exchange Rates: A Panel Data Perspective. Economic Modelling, 42, Becker, R., Enders, W., Lee, J. (). A Stationarity Test in the Presence of an Unknown Number of Smooth Breaks. Journal of Time Series Analysis 27, Carrion-i-Silvestre, J.L., Del Barrio-Castro, T., Lopez-Bazo, E., (). Breaking the panels: An application to GDP per capita, Econometrics Journal, 8, Carrion-i-Silvestre, J. Ll., Sansó, A., (). A guide to the computation of stationarity tests. Empirical Economics, 31, Carrion-i-Silvestre, J.L., Sansó-i-Rosselló, A.S. (). The KPSS Test with Two Structural Breaks. Spanish Economic Review, 9(2), Dickey, D.A. and Fuller, W.A., (). Distribution of the Estimators for Autoregressive Time Series with a Unit Root. Journal of the American Statistical Association, 74(), Enders, W. and Lee, J. (). The flexible Fourier form and Dickey-Fuller type unit root tests. Economics Letters, , King, A. and Dobson, R.C. (). Are income differences within the OECD diminishing? Evidence from fourier unit root tests. Studies in nonlinear dynamics and econometrics, 18, Koedijk, K.G., Tims, B., Dijk, M.A. (). Purchasing Power Parity and the Euro Area, Journal of International Money and Finance. 23, Kurozumi, E. (). Testing for Stationarity with a Break. Journal of Econometrics. (1), Kwiatkowski, D., Phillips, P.C.B., Schmidt, P., Shin, Y. (). Testing the Null Hypothesis of Stationary against the Alternative of a Unit Root. Journal of Econometrics 54, Lee, J., Huang, C.J., Shin, Y. (). On Stationary Tests in the Presence of Structural Breaks. Economics Letters, 55, Narayan, P.K. and Popp, S. (). A new unit root test with two structural breaks in level and slope at unknown time. Journal of Applied Statistics, 37, Perron, P. (). The Great Crash, the Oil Price Shock and the Unit Root Hypothesis, Econometrica, 57, Zivot, E., and Andrews, D.W.K. (). Further evidence of the great crash, the oil-price shock, and the unit root hypothesis. Journal of Business & Economic Statistics, 10, Footnotes: (Note 1) See Carrion-i-Silvestre and Sansó () for the methods on estimating the consistent long-run variance and their small sample comparisons. (Note 2) The selected trading partners account for about 97% and 96% of Turkish total exports and imports which is generally quoted by euro and US dollar. Following Koedijk et al. (), we treat the Eurozone countries as a single currency. The euro explains about 48 percent of Turkey s exports and 30 percent of Turkish imports (see, Adiguzel et al., ).

52 36 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES Export-led Growth Hypothesis in MINT Countries: A Panel Cointegration Analysis Prof. Dr. Harun Bal (Çukurova University, Turkey) Ph.D. Candidate Shahanara Basher (Çukurova University, Turkey) Ph.D. Candidate Abdulla Hil Mamun (Çukurova University, Turkey) Ph.D. Candidate Emrah Akça (Çukurova University, Turkey) Abstract The contribution of exports to GDP in MINT countries that improve substantially just after their implantation of export promotion strategy in the late s raises the issue of whether the growth in these countries is led by export or not. While a good number of studies have been found investigating whether economic growth is promoted by exports for developing countries having an outstanding share of export in GDP, no study investigating the export-led growth hypothesis for MINT countries has been found until recent times. The main purpose of this study is to fill up the void. The study employs panel cointegration technique with an aim to examine whether the export is the key factor of economic growth for MINT countries employing yearly secondary data that covers the period. Results of the study imply that economic growth of these countries is considerably exports driven. Moreover, there is an indication of improvement of efficiency as exports work along with the rise capital formation. As the employment opportunity of an economy is expanded through capital formation, the emerging MINT countries endowed with large population and favorable demographics are expected to become the major exporters with strong GDP growth by being able to attract adequate foreign investment. 1 Introduction ELG hypothesis that studies whether exports are a fundamental element to stimulate growth in developing countries has been the topic of interest of plenty of studies over the last few decades. Export can foster economic growth by confirming better resource allocation, cost advantages and productive efficiency (Kruger, ; Williamson, ; Balassa, ; Bhagawati, ; Srinivasan, ; Awokuse, ). Therefore, despite the cut in investment in other sectors, greater investment in export sector is expected to have a substantial positive impact on the economy. Moreover, exports help increase imports of equipment, raw materials and technologies necessary for an economy, which allows an economy to increase its investment that results in higher output (Rana & Dowling, ). The macroeconomic policies of most of the developing economies were characterized as closed until the late s. Even the East Asian countries were not the exception. All they were following import substitution industrialization policy that was successful in achieving the socio-economic objectives and economic growth. By the early s, there was considerable doubt about the acceptability of import substitution industrialization as a measure to realize the development objectives due to its certain limitations in promoting the growth of industrial sector of domestic economy. It is inevitable to import equipment, raw materials and technologies to enhance the production capacity of an economy, and hence import substitution industrialization cannot completely eliminate the need for import. Moreover, as import substitution industries were not exposed to international competition, it resulted in a production process which is incompetent. MINT is a group of countries tipped as the next economic powerhouse because of their bright economic prospects for the future. Common features that all MINT countries (Mexico, Indonesia, Nigeria and Turkey) share are their growing population, which is a potential source of young labor in coming decades for growing faster, and geographical position suitable for taking advantage of large markets nearby. Mexico is sharing its border with US and also neighbors to Latin America, Indonesia is located at the center of South-east Asia profound link with China, Turkey is a bridge between West and East. Though not yet obvious, but with the expansion of the economies in this continent, Nigeria, the largest economy in Africa, has the potential of becoming the hub of Africa&#;s economy owing to the endowment of natural resources, massive population, well-controlled and well-capitalized banks, and scope to spread retail credit. Apart from these, MINTs are fortunate enough to have systems supportive to industrial development and business growth and they are not overly dependent on a single industry. MINTs have become the major exporters of raw and finished goods since after their inauguration of outwardoriented or export promotion strategy in the late s. Presently, the export sector of these countries have substantial contribution to their GDP. In the s, Mexico&#;s export is percent of GDP, Indonesia&#;s percent, Nigeria&#;s percent and Turkey&#;s percent. Therefore, it is commonly expected that exports are the main driving force of economic development of these countries. While a good number of studies have been found investigating whether exports promote the economic growth of developing countries having an outstanding share of export in GDP, no study investigating the ELG hypothesis for MINT countries has been found until recent times. Hence, the study mainly aims to examine whether exports are the core element of growth of MINT countries.

53 SESSION 2A: Growth & Development 37 2 Export Growth and Overall Macroeconomic Performance of MINT Countries Prior to study the starring role of export in growth of MINT countries, a precise evaluation of the performance of growth and exports of these countries over the past two decades will be appropriate as all of these economies have undertaken structural reform programs as a key for growth emphasizing liberalization and promotion of exports since the s (Thornton, ; Nguyen et al., ; Alimi & Muse, ; Rahmaddi & Ichihashi, ). Mexico: Trade liberalization in Mexico began in the early s through a range of economic reform initiatives undertaken by Mexican government following the dramatic balance of payment crisis. Country s decision to join the GATT in together with the enactment of NAFTA in reduced the tariff and non-tariff barriers considerably. Besides, since , initiatives like facilitating imports of equipment, raw materials and technologies, ensuring exporters easy and better access to credit and removal of barriers to the use of export earnings promote non-oil exports. Consequently, during the period , exports grew at percent per year, as percent of GDP, it was percent in but at the end of it climbed to percent. Imports grew at over 8 percent, climbing from % to percent of GDP. On the export side, manufactures were the leading sector. Its manufactures mounted to around 80 percent in the early s from 38% of total exports at the beginning of liberalization in (Ibarra, ). Nigeria: The failure of the policies to promote Import Substitution Industrialization prior to the s led Nigeria to adopt export promotion industrialization strategies as part of its Structural Adjustment Programme (SAP) in This policy is nowadays followed with the objective of translating it into economic growth and considerable efforts have been given and are still being continued to boost production of domestic exports particularly in sectors other than oil sector with regard to increase the variety of products in the country export structure. Exports of agricultural products were the main source of foreign exchange during the s and s. But crude oil turned out to be the core export item from the mids which now founds nearly 96 percent of total export. However, there is a declining share of non-oil exports, from about 48% in to about 7% in Concerning the contribution of non-oil export to GDP, the downgoing trend reached from 7% in to % in (CBN Statistical Bulletin, ). Hence the performance of the non-oil export sector in the recent decades is not impressive (Ojide et al., ). Consequently, in spite of the development in Nigeria s total exports earnings, the country has been encountering a significant amount of deficit in the balance of payment over the years (Alimi & Muse, ). Main problems that are hindering the non-oil exports of Nigeria are poor funding for export promotion schemes, inefficient implementation, and corruption. Besides, lack of government support and subsidies for non-oil export sector are also mentionable. Exports & Imports s s s s Exports of goods and services % of GDP Turkey Mexico Indonesia Nigeria Imports of goods and services % of GDP Turkey Mexico Indonesia Nigeria Table 1: Performance of Export and Import in MINT Countries. Sources: Authors&#; Calculation from World Development Indicators, World Bank, Turkey: From the beginning of the s, Turkey adopted policies in order to achieve an economic system which is more liberal and market-oriented. As part of the policies, government undertook a growth strategy conducive to promote export and the external competitiveness of the economy was preserved through management of exchange rate and export subsidies. These primary initiatives undoubtedly helped regain the confidence of international creditors. The stand-by loans provided by IMF along with the adjustment loans of World Bank were promptly arranged and disbursed together with supplementary debt relief operations (Taş and Kar, ). On the other hand, the gradual but deliberate reduction in real wages in s targeting to generating an exportable surplus enhance the competitiveness of export with lower labor costs (Taban S. & Aktar, ). These export-promoting strategies were successful in enhancing exports substantially. Its exports reach to percent of GDP in the s from percent in s at a percent annual growth rate. Indonesia: Export promotion in Indonesia was embarked on in response to the breakdown of import substitution industrialization due to the fall in oil prices throughout the period of first half of the s that exhausted the gains that Indonesia achieved due to the mids oil boom. The economy experienced a quick surge in foreign direct investment owing to the intrepid and influential series of outward-oriented reform programs

54 38 INTERNATIONAL CONFERENCE ON EURASIAN ECONOMIES undertaken from the mids ahead. Owing to the reduction and removal of tariff as well as non-tariff barriers, manufacturing exports raised five times in 9 years from that of (Rahmaddi & Ichihashi, ). However, regardless of the fall of manufacturing exports during Asian economic crisis in, it maintained a modest real GDP growth rate over 5 percent with more that 22 percent share of exports in GDP through to Literature Review There has been a good number of empirical studies on the export-growth relationship based on both time series as well as panel data. The results of the studies backed by time series data are inconclusive as some of them suggest that ELG hypothesis is valid for certain countries while it is not the case for others. While Medina-Smith (), Hossain & Karunaratne (), Siliverstovs & Herzer (), Baharumshah and Almasaied () and Balcilar & Ozdemir () find evidence to support ELG hypothesis, studies include Chang et al. (), Awokuse (), Ahmed and Uddin (), Mishra () and Malhotra & Kumari () oppose it. In addition, Shan and Sun (), Awokuse () and Paul () find that the causality between exports and GDP growth is bidirectional. Among the panel cointegration analysis, Sharma and Smyth (), Seabra and Galimberti (), Chandra Parida and Sahoo (), Biyase and Zwane () and Olson () find exports as a driving force of economic growth, but Bahmani-Oskooee et al. () and Pazim et al. (), did not find sufficient evident to support evidence of the earlier studies. Results obtained from the studies on MINT countries employing time series data are also inconclusive. Studies mainly rely on Granger causality test with the aim of justifying the ELG for MINT countries. Studies that support ELG of MINT countries include McCarville & Nnadozie (), Thornton (), Lorde (), Ozturk and Acaravci (), but the hypothesis is opposed by Nguyen et al. (), Taban and Aktar (), Thangavelu and Rajaguru (), Alimi (). In addition, Lorde () finds an inverse association between exports and GDP for Mexico. Granger causality from real GDP to exports is found to be unidirectional in the study of Nguyen et al. () for Turkey and Alimi & Muse () for Nigeria. Taban and Aktar () identify a bidirectional causality relationship between real GDP growth and growth of exports in Turkey not only in the short run but also in the long run. The same result is found by Alimi () for Nigeria, while Ojide et al. () finds sufficient evidence in support of non-oil ELG hypothesis beyond the ELG hypothesis in Nigeria based on autoregressive distributed lag (ARDL) model and co-integration analysis. However, there is no study found addressing whether the growth of MINT countries are essentially driven by export making use of panel cointegration analysis and the study is an attempt to fill up the void. 4 Data Sources and Methodology The study considers yearly data accumulated from secondary sources to examine the ELG hypothesis for MINT countries. The balanced panel consists of data for export, import, gross capital formation and GDP for MINT countries for the period of and data of each variable is measured in US dollars. Data are in real form considering the base year The data are collected from World Development Indicators, World Bank. A substantial amount of equipment, raw materials and technologies are imported by all the MINT countries to support their local production and also exports. Therefore, it will be appropriate to include import as a component in the estimation of growth. To assess the change in the efficiency of production, the study also considers the stock of physical capital as dependent variable. But the stock of physical capital is proxied by gross fixed capital formation mainly due to the limitation of data and complications in measuring the stock of physical capital. Hence the theoretical model for the study to examine the export-growth relationship can be given by- ln y t ln EXP ln IMP ln CAP (1) it i i 1i it 2i it 3i it it Where y is gross domestic product, EXP is export, IMP is import and CAP is gross fixed capital formation. All the variables are in real form. The probable sign of 1 i, 3i is positive and 2i is negative. The limitation of pure time series and pure cross section analysis is that they are not able to detect all the information available in data while estimating the equations, which is not the case for panel data techniques. That is why the study estimates the export-growth relationship employing the panel cointegration technique. Moreover, coefficient estimates in panel data estimation are superior to time series estimation due to the increase in the power of the tests for small data span, provided that the sample covers only 30 observations for each of the countries. The study begins with panel unit root test for each series, applying Im et al. () (IPS) techniques. Different methods are used for the test of panel cointegration. The null hypothesis of no-cointegration is assumed in the Engle & Granger () tests and stochastic disturbances produced from the panel regression are used. This method is employed in the panel cointegration tests offered by Pedroni (), McCoskey & Kao (). Kwiatowski et al. (), Harris & Inder () and Shin () proposed another method that considers the null hypothesis of cointegration. Heterogeneity in the cointegrating coefficients is permitted by all these tests of panel cointegration. Primarily, the two-step methodology offered by Engle Granger is performed for panel cointegration

55 SESSION 2A: Growth & Development 39 tests where unit root tests are directly applied to the residuals. However, this procedure yields biased test statistics towards accepting stationarity (Pedroni, ). Pedroni () challenges the suitability of applying panel unit root tests directly to the regression residuals for several reasons that include lack of exogeneity of the regressors and the residuals are dependent on the distribution of the estimated coefficients. As a result, it is necessary to have a robust test procedure for cointegration in the presence of heterogeneity as an alternative. The study prefers to employ the cointegration test offered by Pedroni which permits considerable heterogeneity. After testing for cointegration, analysis of equation (1) will be the next issue to be addressed. Fully-Modified OLS (FMOLS) approaches can be applied to attain the panel data estimates for β i. For heterogeneous panel with I(1) variables, application of this methodology is able to produce results superior to OLS. Its distribution is standard and follows the asymptotic properties of unbiasedness. Moreover, FMOLS performs better in producing consistent standard errors and therefore consistent t-statistics that the OLS estimator could not yield. 5 Empirical Results and Discussion Table 2 summarizes the IPS panel unit root test results both at levels and at first difference. The results of the panel unit root tests imply that there is no way of rejecting the null hypothesis of non-stationarity at log level for all variables. Therefore, all the variables considered for the study are stationary at their first difference level. Variable GDP Export Import Capital Level Intercept () () () () Intercept and Trend () () () () First Difference Intercept Intercept and Trend * * () () * * () () * * () () * * () () Table 2: Im, Pesaran, Shin (IPS) Unit Root Test Decision Since all the variables are stationary at first difference level, panel cointegration test can be conducted. As all the variables are integrated at same order, i.e. I(1), Pedroni s panel cointegration tests can be applied in order to examine the long-run equilibrium relationship among the variables. The findings of cointegration analysis based on equation (1) are reported in Table 3. The null hypothesis of non-cointegration against the alternative of cointegration is rejected in the case of equation (1) as both panel-adf and group-adf statistics are significant at 10% and 5% levels respectively. Apart from the adf-statistics, the results also imply significant panel-v and panel-pp statistics. The Kao test also rejects the null hypothesis- no cointegration as shown in table 4. Test Intercept p-value Intercept and Trend p-value Panel v-statistic * * Panel ρ-statistic * Panel pp-statistic * * Panel adf-statistic * * Group ρ Statistic Group pp-statistic * Group adf-statistic * Table 3: The Pedroni Panel Cointegration Test t-statistic Prob. ADF Residual variance HAC variance Table 4: Kao Residual Cointegration Test Overall cointegration results suggest that there exists a long run cointegrating relationship among the variables. Therefore, long run cointegrating relationship can be estimated. Table 5 summarizes the findings of group FMOLS of equations (1). Results lead to the inference that the coefficients hold the signs theoretically expected that are also statistically significant. Hence, the growth of MINT countries as a whole is export driven, but whether it differs from individual country experiences requires investigation. I(1) I(1) I(1) I(1)

56