Analysis of Fractal Market Hypothesis in Latin American Countries Index

M. E. Soykan

Makale detayları

Analysis of Fractal Market Hypothesis in Latin American Countries Index

Yazarlar:

M. E. Soykan

Yayıncı:

Ankara Hacı Bayram Veli University

Ana Başlıklar:

Uluslararası Risk Yönetimi

Dergi:

Akademi Sosyal Bilimler Dergisi Vol.12 Issue 34, pp.17-30

Belge Tipi:

Makale

Yayın Tarihi:

1/2025

Belge Dili:

Türkçe

Document ID (Org No.):

2001340008

pISSN:

2667-4866

Anahtar kelimeler:

Özet:

 The Efficient Market Hypothesis is of great importance for policy makers, portfolio managers, institutional and individual investors. In the Efficient Market Hypothesis, it is argued that all information available in the market is reflected in the stock price and abnormal returns cannot be obtained. On the contrary, the Fractal Market Hypothesis defends the existence of long memory in the index examined and suggests that future prices or returns can be predicted by analyzing past prices, in a sense, technical analysis etc. and that doing so will be advantageous . The target of this research is to analyze if the Fractal Market Hypothesis is valid in the MSCI Emerging Markets Latin America Index-USD in the context of long memory availability. The countries by this index are Brazil, Chile, Colombia, Mexico and Peru. Daily closing prices between 05111997 and 16112023 are used in the study. The data is obtained from Refinitiv-Eikon database. The most suitable model for the relevant data is the ARMA4,6-FIEGARCH1,d,1 model with GED distribution. Since the d coefficient in the model is greater than 0.5 and significant, long memory does not exist in the series. In other words according to the applied methods, the Fractal Market Hypothesis is not valid for this index. Since the d coefficient is greater than 0.5, even if the process tends to return to the mean, stationarity in terms of covariance will not be ensured and it will have infinite variance. Therefore, it will take a very long time for the process to return to the mean. Since the coefficient showing leverage is negative and significant, it can be mentioned that there is leverage effect in the series and negative news increase volatility more than positive news.

 The Efficient Market Hypothesis is of great importance for policy makers, portfolio managers, institutional and individual investors. In the Efficient Market Hypothesis, it is argued that all information available in the market is reflected in the stock price and abnormal returns cannot be obtained. On the contrary, the Fractal Market Hypothesis defends the existence of long memory in the index examined and suggests that future prices or returns can be predicted by analyzing past prices, in a sense, technical analysis etc. and that doing so will be advantageous . The target of this research is to analyze if the Fractal Market Hypothesis is valid in the MSCI Emerging Markets Latin America Index-USD in the context of long memory availability. The countries by this index are Brazil, Chile, Colombia, Mexico and Peru. Daily closing prices between 05111997 and 16112023 are used in the study. The data is obtained from Refinitiv-Eikon database. The most suitable model for the relevant data is the ARMA4,6-FIEGARCH1,d,1 model with GED distribution. Since the d coefficient in the model is greater than 0.5 and significant, long memory does not exist in the series. In other words according to the applied methods, the Fractal Market Hypothesis is not valid for this index. Since the d coefficient is greater than 0.5, even if the process tends to return to the mean, stationarity in terms of covariance will not be ensured and it will have infinite variance. Therefore, it will take a very long time for the process to return to the mean. Since the coefficient showing leverage is negative and significant, it can be mentioned that there is leverage effect in the series and negative news increase volatility more than positive news.

 The Efficient Market Hypothesis is of great importance for policy makers, portfolio managers, institutional and individual investors. In the Efficient Market Hypothesis, it is argued that all information available in the market is reflected in the stock price and abnormal returns cannot be obtained. On the contrary, the Fractal Market Hypothesis defends the existence of long memory in the index examined and suggests that future prices or returns can be predicted by analyzing past prices, in a sense, technical analysis etc. and that doing so will be advantageous . The target of this research is to analyze if the Fractal Market Hypothesis is valid in the MSCI Emerging Markets Latin America Index-USD in the context of long memory availability. The countries by this index are Brazil, Chile, Colombia, Mexico and Peru. Daily closing prices between 05111997 and 16112023 are used in the study. The data is obtained from Refinitiv-Eikon database. The most suitable model for the relevant data is the ARMA4,6-FIEGARCH1,d,1 model with GED distribution. Since the d coefficient in the model is greater than 0.5 and significant, long memory does not exist in the series. In other words according to the applied methods, the Fractal Market Hypothesis is not valid for this index. Since the d coefficient is greater than 0.5, even if the process tends to return to the mean, stationarity in terms of covariance will not be ensured and it will have infinite variance. Therefore, it will take a very long time for the process to return to the mean. Since the coefficient showing leverage is negative and significant, it can be mentioned that there is leverage effect in the series and negative news increase volatility more than positive news.

 Etkin Piyasa Hipotezi, politika oluşturucular, portföy yöneticileri, kurumsal ve bireysel yatırımcılar bakımından büyük önem taşımaktadır. Etkin Piyasa Hipotezinde, piyasada bulunan bütün bilginin hisse senedi fiyatına yansıdığını ve anormal getiri elde edilemeyeceği savunulmaktadır. Fraktal Piyasa Hipotezi ise bunun tam tersine incelenen endekste uzun hafızanın varlığını savunmakta, geçmiş fiyatlar analiz edilerek ilerideki fiyatların veya getirilerin tahmin edilebileceğini öne sürmekte yani bir bakıma teknik analiz vs. yapmanın avantaj sağlayacağını ifade etmektedir. Bu çalışmanın amacı MSCI Latin Amerika gelişmekte olan ülkelerinin oluşturduğu endekste Fraktal Piyasa Hipotezinin geçerli olup olmadığını uzun hafıza varlığı bağlamında analiz etmektir. Bu endeksin kapsadığı ülkeler Brezilya, Şili, Kolombiya, Meksika ve Perudur. Çalışmada 05111997 ile 16112023 tarihleri arasında günlük kapanış fiyatlarından faydalanılmaktadır. Veriler Refinitiv-Eikon veri tabanından sağlanmıştır. İlgili veri için en uygun model GED dağılımlı ARMA4,6-FIEGARCH1,d,1 modelidir. Modelde d katsayısı 0.5den büyük ve anlamlı olduğundan seride uzun hafıza yoktur. Yani uygulanan yöntemlere göre Fraktal Piyasa Hipotezi bu endeks için geçerli değildir. d katsayısı 0.5den büyük olduğu için sürecin her ne kadar ortalamaya dönme eğilimi sergilense bile kovaryans açısından durağanlığı sağlanamayacak ve sonsuz varyanslı olacaktır. Bu yüzden sürecin ortalamaya dönüşü çok uzun bir zaman geçmesini gerektirecektir. Kaldıracı gösteren katsayı negatif ve anlamlı çıktığından seride kaldıraç etkisi bulunduğu olumsuz haberlerin olumlu haberlere kıyasla volatiliteyi daha fazla artırdığı ifade edilebilir.

Makale detayları

Analysis of Fractal Market Hypothesis in Latin American Countries Index

Yazarlar:

M. E. Soykan

Yayıncı:

Ankara Hacı Bayram Veli University

Ana Başlıklar:

Uluslararası Risk Yönetimi

Dergi:

Akademi Sosyal Bilimler Dergisi Vol.12 Issue 34, pp.17-30

Belge Tipi:

Makale

Yayın Tarihi:

1/2025

Belge Dili:

Türkçe

Document ID (Org No.):

2001340008

pISSN:

2667-4866

Fractal Market Hypothesis

FIEGARCH

long memory

volatility

efficient market hypothesisFraktal piyasa hipotezi

FIEGARCH

uzun hafıza

volatilite

etkin piyasa hipotezi.

Sayı

Makale detayları

Yazarlar: M. E. Soykan

Yayıncı: Ankara Hacı Bayram Veli University

Ana Başlıklar: Uluslararası Risk Yönetimi

Dergi: Akademi Sosyal Bilimler Dergisi Vol.12 Issue 34, pp.17-30

Belge Tipi: Makale

Yayın Tarihi: 1/2025

Belge Dili: Türkçe

Document ID (Org No.): 2001340008

pISSN: 2667-4866

Sayı

Yazarlar:

M. E. Soykan

Yayıncı:

Ankara Hacı Bayram Veli University

Ana Başlıklar:

Uluslararası Risk Yönetimi

Dergi:

Akademi Sosyal Bilimler Dergisi Vol.12 Issue 34, pp.17-30

Belge Tipi:

Makale

Yayın Tarihi:

1/2025

Belge Dili:

Türkçe

Document ID (Org No.):

2001340008

pISSN:

2667-4866