Journal of Economic Policy Researches
Testing the Efficient Market Hypothesis in G8 Countries: New evidence from Unit Root Tests with Fourier Shiftsİlhan Küçükkaplan, Emre Kılıç, Şevket Pazarcı, Asım Kar
The aim of this study is to test the validity of the efficient market hypothesis for stock market indices in G-8 countries (USA, Germany, France, England, Italy, Japan, Canada, and Russia). The augmented Dickey-Fuller (ADF), residual augmented least squares (RALS)-ADF, Fourier-ADF, and Fourier-Kapetanios-Snell-Shin (Fourier-KSS) unit root tests are used to do this. The longest possible period was used as the analysis period by considering the data available for each index. Unlike the literature, the validity of the efficient market hypothesis for stock market indices in the G8 countries has been comprehensively and comparatively examined by simultaneously considering Fourier breaks, non-normal distribution, and non-linearity. According to the empirical findings, the null hypothesis was not be rejectable for all the unit root tests that were applied to the stock market indices from Germany, France, and Japan. In other words, strong evidence has been obtained for the validity of the efficient market hypothesis for the stock market indices from these three countries. On the other hand, the null hypothesis was rejected for Russia’s stock market index as a result of the unit root tests applied apart from ADF, with the results that emerged indicating the efficient market hypothesis to be invalid for that case. Different results were observed to have been caused by taking into account Fourier breaks and nonlinearity in other indices. The Fourier-KSS test considers the nonlinearity in a dataset using Fourier breaks and was found to provide more evidence for the invalidity of the efficient market hypothesis compared to the other types of tests. This reveals the importance of choosing the appropriate test for the data.
G-8 Ülkelerinde Etkin Piyasa Hipotezinin Test Edilmesi: Fourier Kırılmalı Birim Kök Testlerinden Yeni Kanıtlarİlhan Küçükkaplan, Emre Kılıç, Şevket Pazarcı, Asım Kar
Bu çalışmanın amacı G-8 ülkelerinde yer alan borsa endeksleri için (ABD, Almanya, Fransa, İngiltere, İtalya, Japonya, Kanada ve Rusya) etkin piyasa hipotezinin geçerliliğini test etmektir. Bunun için ADF, RALS-ADF, FourierADF ve Fourier-KSS birim kök testlerinden yararlanılmıştır. Analiz dönemi olarak her bir endeks için veri bulunabilirliği göz önüne alınarak en uzun dönem kullanılmıştır. Literatürden farklı olarak G-8 ülkelerinde yer alan borsa endeksleri için etkin piyasa hipotezinin geçerliliği aynı anda hem fourier kırılmalar hem normal dağılmama durumu hem de doğrusal olmama durumu dikkate alınarak kapsamlı ve karşılaştırmalı bir şekilde incelenmiştir. Elde edilen ampirik bulgulara göre Almanya, Fransa ve Japonya’nın borsa endekslerinde uygulanan tüm birim kök testlerinde boş hipotez reddedilememiştir. Yani bu üç ülkenin borsa endeksleri için etkin piyasa hipotezinin geçerliliği için güçlü kanıtlar elde edilmiştir. Aksine Rusya’nın borsa endeksinde ise ADF dışında uygulanan birim kök testleri sonucunda boş hipotez reddedilerek etkin piyasa hipotezinin geçersiz olduğuna yönelik sonuçlar ortaya koyulmuştur. Diğer endekslerde de fourier kırılma ve doğrusal olmama durumunun dikkate alınmasına göre farklı sonuçların neden olduğu bilgisine ulaşılmıştır. Fourier kırılmalarla birlikte veri setindeki doğrusal olmama durumunun dikkate alındığı Fourier-KSS testi etkin piyasa hipotezinin geçersizliği yönünde diğer tip testlere göre daha fazla kanıt sunduğu gözlemlenmiştir. Bu durum, veri setine uygun test seçiminin önemini ortaya koymaktadır.
Studies on market efficiency are often used to examine the behavior of stock markets. The basic assumption of the efficient market hypothesis (EMH) is that the prices of financial assets reflect all available information. In an efficient market, prices reflect all information and investors cannot make abnormal profits, nor can predictions be made about the future using past stock prices, with this being characterized by the random walk process. Stock prices that have a random walk process where the unit root process is namely integrated as I(1) show the efficient market hypothesis to be valid. The study uses the unit root approach in this context to determine the validity of the efficient market hypothesis. In this case, the effect from shocks will be permanent and therefore, predicting the future based on past price movements will not be possible. If stock prices follow a static process where the unit root process integrates as I(0), then the EMH is invalid. In this case, shocks will have a temporary effect, and predictions about the future price can therefore be obtained by using the past price movements of a stock.
Testing stock prices with traditional unit root tests will result in low power characteristics. When using tests with low power characteristics, false inferences can be made about market efficiency. In this context, the study takes nonlinearity, non-normal distribution, and Fourier breaks in the data into account by analyzing the behavior of the series for G8 countries, which constitute the sample group of the study. G8 countries constitute 48% of gross world product, 35% of total exports, 37% of imports, 49% of foreign direct investments, and approximately 13% of the total population of the world. In this context, G8 countries have an important place in the functioning of the economy.
The study examines the validity of EMH for G8 countries using unit root tests with different specifications and reveals new evidence from the tests using Fourier breaks. According to the augmented Dickey-Fuller (ADF) test, which is the classical unit root test, EMH was concluded to be valid for seven countries, while the Fourier-Kapetanios-SnellShin (Fourier-KSS) test, which most comprehensively takes the structure in the data into consideration, showed EMH t be valid in only three countries. All the unit root tests examined in Germany, France and Japan revealed the markets to be effective (i.e., EMH to be valid); this shows that investors cannot make above-normal profits in these markets. Therefore, investors who will invest in the stock market indices of this country will be able to earn the best returns in the market by following the buy-and-hold strategy.
Obtaining different results from tests that take into account different features of the data structure in empirical analysis reveals the importance of the test used.. In this regard, the study suggests that researchers that first determine the method to use in their analysis by examining the structure of the data. The study also draws attention to the effect of using the method that will strongly explain the data in terms of the results.