PETROL FİYATLARI İLE BORSA İSTANBUL’UN KAPANIŞ FİYATLARI ARASINDAKİ SAKLI İLİŞKİNİN ANALİZİ
ANALYZING THE HIDDEN COINTEGRATION BETWEEN OIL PRICES AND STOCK PRICES
Sefer ŞENER,Veli YILANCI,Muhammed TIRAŞOĞLU
Bu çalışmada Borsa İstanbul’un kapanış fiyatları ile petrol fiyatları arasındaki ilişki 2002-2012 dönemi için günlük veri kullanılarak, Granger ve Yoon (2003) ile Hatemi-J ve Irandoust (2012) tarafından literatüre kazandırılan saklı eşbütünleşme testleri ile incelenmiştir. Granger ve Yoon (2003) testi iki serinin hem pozitif hem de negatif bileşenleri arasında uzun dönemli bir ilişki olmadığını gösterirken, Hatemi-J ve Irandoust (2012) testi ise her iki serinin hem iki bileşeni arasında uzun dönemli bir ilişki olduğuna işaret etmektedir. Elde edilen bu sonuçlar, petrol fiyatlarında meydana gelecek artış veya azalışların hisse senetleri fiyatlarının oluşmasında etkili olacağını göstermektedir.
The long term relationship between oil prices and stock prices has been a
popular subject for both politicians and academicians for a long time since the
determinants of the stock prices is an important issue for investors and the bulk of
literature find the long run relationship between oil prices and stock prices.
Determining real economic activities that determines stock prices is very
important for investors to determine their investment instrument. The main variables that
have an impact on stock prices; market interest rate, industrial production index money
supply, exchange rate, inflation, goldpriceandoilpricedrawattention ( Fama (1981),
MookerjeeandYu (1997), KwonandShin (1999), Özer, (2011). Besides these variables, it
is possible to state that there are the psychological effects on investigators which effect
their investment plans.
Oil prices is one of the most important variables that impacts on stock prices.
Unexpected changes that may occur in the factors that impact on oil prices may lead to
fluctuations in oil prices, therefore also to the risk. Fluctuations in oil prices increase the
uncertainty that has a negative impact on wealth and investment. Several other macroeconomic
variables with this uncertainty can be affected by changes in oil prices.
Rising oil prices are usually indication of pressure related to inflation that
central banks bring under control by raising interest rates. High interest rates make the
bills more attractive as opposed to stocks. The overall effect of rising oil prices depends
on whether companies are consumers or manufacturers of oil and oil products. Since the
oil consuming companies in the world are more than oil producing companies, the overall
impact of increase in oil prices on stocks is expected to be negative.
Changes in oil prices create different effects according to being oil importer or
oil exporter of a country. Oil pricesin oil-exporting countries increases export revenues
and leads to a significant increase in national income. However, the increase in oil prices
in oil-importing countries will make a reducing effect on national income. These effects
vary according to countries' oil expenditure share of the national income.
In the literature, the relationship between the stock and real economic activity is
examined in many studies, some of these studies are between oil prices and stock returns.
However, the number of researchs examining the relationship between the price of stocks
and oil prices in Turkey is not much.
The main studies in the literature on this subject; Sadorsky (1999), Maghyereh
and Al-Kandari (2007), Henriques and Sadorsky (2008), Chiou and Lee (2009), Fayyad
and Daly (2011). For Turkey, Güler et al. (2010), İşcan (2010), Kapusuzoglu (2011),
andÜnlü and Topcu (2012) investigate the relationship between stock price and oil prices.
If there is a stationary linear combination of two integrated variables, we
conclude that there is a cointegration relationship between two variables. Engle and
Granger (1987) and Johansen (1988, 1991) cointegration tests are among the most
employed cointegration tests in the empirical literature. Although popularity of these
cointegration tests, since there are some deficiencies of the tests, several new techniques
have been introduced to the literature. For example Gregory and Hansen (1996) and
Hatemi-J (2008) developed new cointegration tests which consider structural breaks
andBalke and Fomby (1997) and Kapetanios, et al. (2006) introduced new
cointegrationtests which allow nonlinearity in the long run relationship. In this study we employ recently introduced hidden cointegration techniques.
Granger and Yoon (2003) state that even if economic data do not respond to same shocks
together, they could response a certain kind of shocks and define this new type of
cointegration as hidden cointegration. The hidden cointegration test which introduced by
Granger and Yoon (2003) based on the simple cointegration test of Engle and Granger
(1987) while the cointegration test of Hatemi-J and Irandoust (2012) based on the
cointegration test of Johansen (1988). Hidden cointegration is a special case of standard
cointegration and indeed a type of nonlinear cointegration test.
To implement hidden cointegration techniques, we first decompose the data into
the negative and positive shocks and obtain the hiddencointegration test results, we obtain
theEngle and Granger (1987) and Johansen (1988 cointegration tests to these
componentsif they are integrated at the same levels.
The data, used in this study covers the period 2002-2012 and obtained in daily
frequency. Closing prices of Istanbul Stock Exchange (BIST) were obtained from the
Electronic Data Dissemination System of the Central Bank of the Republic of Turkey and
Brent oil prices (oil) were obtained from the U.S. Energy Information Agency. In order to
reduce the variability of the data used in this study, the logarithmic form of them was
We first test the stationary characteristics of the related variables by using
Augmented Dickey Fuller unit root test. The test results show that all the components of
the variables are integrated at the first level, so we can advance to the stage where we test
the existence of the long run relationship between the components of the oil prices and
closing prices of Borsa Istanbul.
Test results of Granger and Yoon (2003) show that there is no long run
relationship between the components of two variables whereas the results of the Hatemi-J
and Irandoust (2012) cointegration test show the existence of cointegration relationship
between the components of two variables.
The reason for the difference of the results obtained is the superiorities that
Johansen cointegration test has over Engle and Granger cointegration test. In Johansen
cointegration test, the estimation of the cointegration vectors are based on the Maximum
Likelihood estimation method and as the short-term economic changes are made
concurrently, the effectiveness of the estimate increases. Therefore, relying on the
Hatemi-J and Irandoust (2012) that were based on the Johansen cointegration test, will
provide more accurate decisions.
There arelots of studies in the literature investigating the existence of a
relationship between the prices of stocks and oil. While most of the results of the studies
support this idea, some studies have not reached the existence of a relationship between
these two variables. In this study, by using Hatemi-J and Irandoust (2012) cointegration
test, it was concluded that Istanbul Stock Exchange closing prices are cointegrated with
the price of oil. The findings are consistent withthe those of Maghyereh and Al-Kandari
(2007), Henriques and Sadorsky (2008) and Fayyad and Daly (2011). As for Turkey, the
results are consistent to those of Güler, et.al. (2010), Kapusuzoglu (2011) and Unlu and
Topcu (2012) (the second sub-period, 2001:03-2011:12).
According to the results of the study, it is apparent that oil price is among the
determinants of stock prices in Istanbul Stock Exchange. The results shows that rising oil
prices, can lead to an increase in production costs in the absence of a full substitution
among the factors of production, the increase in production costs can lead to the decrease in the cash flow and reduce stock prices. In other words, future increases or decreases in
oil prices will affect the formation of stock prices.