Investigation of the Relationship between Economic and Political Uncertainty and General Economic Outlook: An Econometric Analysis on Athens Stock Exchange

Bu calismada, Yunanistan'daki ekonomik ve politik belirsizlikler ile Atina Menkul Kiymetler Borsasi'nin fiyat hareketleri ve oynakligi arasindaki iliski, Chowdhury (1993) oynaklik hesaplama yontemi kullanilarak hesaplanan 19 ekonometrik model ile incelenmistir. Calismada panel veri analizi yapilmistir. Analizde ilk olarak Kapetanios (2005) yapisal kirilma unitesi kok testi yapilmis ve serilerin farkli seviyelerde duragan oldugu gorulmustur. Seriler arasindaki esbutunlesmeyi tespit etmek icin Pesaran ve ark. (2001) tarafindan gelistirilen Sinir Testi yontemi uygulanmistir. Atina Borsasinin performansini etkileyen yapisal kirilma tarihlerini tespit etmek amaciyla kukla degiskenler kullanilarak Bai-Perron (2003) coklu yapisal degisim yontemi uygulanmistir. Uzun vadeli analizde, Yunanistan'daki siyasi ve ekonomik belirsizlik gostergeleri ile Atina Menkul Kiymetler Borsasi Endeksi arasinda negatif bir iliski belirlenmistir. Vergi politikalarindaki belirsizlikler (EPUT), ekonomik belirsizlikler (EU) ve bankacilik politikalarindaki belirsizlikler (EPUB) borsa uzerinde en fazla etkiye sahip gostergeler olarak belirlenmistir. Kisa vadeli analizde, soz konusu belirsizliklerin kisa vadede Atina Menkul Kiymetler Borsasi Endeksi'nde de etkili oldugu, ancak borsalarin fiyat oynakligi uzerinde onemli bir etkiye sahip olmadiklari gorulmektedir.


Introduction
Uncertainty is the situation where it is not clear that what will happen in the short and long future based on the available information and data. İn other words, it is the possibility of occurring unexpected things. The main reason for the uncertainty is the difficulty degree for giving a decision and related increasing cost of the decision.
İn real life, economic units avoid uncertainty as much as possible. Because of the uncertainty, all the variables that are determinant factors for decision-makers are constantly changing in extreme volatility, and it is impossible to make an accurate estimate of what will happen even in the short term (Erdogan Cosar & Sahinoz, 2018).
İf the situation is considered within the macroeconomic framework, since the confidence of all the participants that include investors, consumers, firms and even governments, in the economy that is suffering uncertainty will decrease, the economic activities of these parties will also decrease significantly or even come to a halt. İn such an environment, all actors in the market will look for economic stability, and in this case, the uncertainty caused by the slowing economic activity will lead the general economy to real or financial crises (UNCTAD, 2009). Uncertainty in the economy will limit the activities of the countries which are the biggest actor in the market, and even bring it to bankruptcy (Guterres, 2020). The İceland case, experienced with the 2008 global financial crisis, had shown that even states could financially go bankrupt The 2008 global financial crisis affected all the countries of the world directly or indirectly. All countries, especially the USA, where the crisis started, had to struggle with economic and political uncertainty after the crisis, which caused comments that the financial crisis has not yet ended and that the world economy will see a second bottom line.
As Knight (1921) mentioned in his book called Risk, Uncertainty, and Profit, risk and uncertainties have the potential to affect economic activities closely. Keynes (1936) stated that one of the most important reasons of the 1929 Great Depression lasted was the decrease of investment motivation of firms due to uncertainty in the economy.
Negative effects of global economic policy uncertainty are much more effective than positive effects. Therefore, the global economic policy uncertainty has greater impacts on stock returns in the high-volatile environment. The global economic policy uncertainty can be regarded as a systemic risk factor and a estimator of returns of the stocks (Hoque & Zaidi, 2019). Dash et al. (2019) stated that uncertainties in economic policies significantly affect the liquidity and profitability of stock markets. Gilal (2019) also stated that uncertainties in economic policies are an important determinant on stock returns. Alqahtani and Martinez (2020) examined the effects of the uncertainty of economic policies in the USA on the stock exchanges in the Gulf Cooperation Countries and determined that these uncertainties had long-term negative effects on stock prices in Bahrain and Kuwait. According to these results, it can be said that policy uncertainties have the potential to affect stock markets closely and it is useful to analyze these effects frequently.
İn this study, the relationship between economic and political uncertainty indicators and the general economic conjuncture will be examined in the case of Greece, which is one of the countries most affected by the 2008 global financial crisis. İn the section 2, previous studies on this subject included in the literature are examined. İn the section 3, detailed information about data and sample structure, econometric models and methodologies used in the analysis are provided. İn the section 4, the tests included in the analysis are carried out and the results are shared. Section 5 provides comments and suggestions have been made for all countries in general based on the findings obtained.
İn the preliminary studies conducted, it was observed that in the current studies the effects of changes in the EPU of the USA on the stock exchanges or other macroeconomic indicators of the other countries were investigated. Although the USA is a locomotive of the world economy and the developments in this country affect other countries, it is clear that the changes in the stock market indices of the USA will be more sensitive to the policy uncertainties in itself. İn this study, a much more specific analysis was made by examining the effects of the Greek EPU on the Athens stock exchange. Moreover, this analysis was also carried out not only for the stock market index but also for the volatility in the stock market index. İn addition, by using the 10 different indices associated with EPU apart from using EPU, a more detailed research than the current studies in the literature was conducted. This study is expected to make significant contributions to the literature, both in terms of the variables used and the way of handling the subject. On the other hand, it is hoped to grab attention of researchers in Turkey, and also encouraging them for the creation such index as EPİ and its sub-indices. Finally, it is hoped that this study will contribute to the more efficient portfolio management by drawing attention of individual and corporate stock market investors to the effects of the economic policies on the stock market and price volatility. Sum (2012) analyzed the effects of economic policy uncertainties in the USA on the performance of stock markets in the Asian countries (The ASEAN Stock Market Performance and Economic Policy Uncertainty in the United States), for the 1985:02-2012:M02 period by Granger causality. Results indicate that the changes in economic policy uncertainty in the U.S. adversely affect the returns on the five ASEAN stock markets. According to the vector autoregression analyses that they applied returns on the five ASEAN stock markets negatively react to the changes in economic policy uncertainty immediately. Based on the Granger causality tests, they concluded that the changes in economic policy uncertainty in the U.S. causes the returns on the Singapore and Malaysia stock markets while the same findings cannot be observed for the case of İndonesia, Philippines and Thailand. Liu and Zhang (2015) analyzed effects of uncertainties in economy policies on the stock market volatility through Bollerslev and Ghysels (1996) method by using the five-minute data of S&P 500 for the 1996: M01-2013: M01 period, and EPU index data of Baker et al. (2013). As a result of in sample findings, they found out that higher EPU leads to significant increases in market volatility. Moreover, they also found out based on out-of-sample findings, including EPU as an additional predictive variable into the existing volatility prediction models significantly improves forecasting ability of these models. Li et al. (2015) analyzed the effects of uncertainty in economic policies in China and İndia on stock returns, using the Bootstrap Rolling Window Causality test using data from China's 1995China's : M02 -2013China's : M02 period, İndia's 2003China's : M02-2013 period. Bidirectional causal relationships were found between EPU and stock returns in several sub-periods rather than in the entire sample period. They also stated that the generally relationship between EPU and stock returns is determined as not strong for these two emerging countries. Asteriou and Sarantidis (2016) analyzed the relationships between political instability and stock market returns using EFA, PCA and GARCH-M methods, based on data from the 18 OECD countries for the period 1993 -2013. They found that political instability had direct and indirect effects on the returns of the banking sector stocks and on the stock market in general. However, they determined that the general trend was towards a decrease in stock market index and stock market return due to increasing political instability. Moore (2017) examined the effects of economic uncertainty in Australia on real economic magnitudes by using the VAR method. He found that investments and employment in this country decreased when economic uncertainty increased. Similarly, it has been found that increases in economic uncertainty reduce household savings and spending on durable consumer goods and tend to be more cautious in environments of uncertainty. The author also stated that economic uncertainties are an important independent variable regarding the shape of the economy in the future. Baker et al. (2016) analyzed the effects of uncertainties in economic policies on economic activities in the USA using the Panel VAR method for the period 1985: M01-2014: M12. As a result of the study, it is determined that the increase in uncertainties in the economic policies has an impact on the price volatility in the stock exchange and reduces the investments and employment. These effects are observed to be greater, especially in the defense, health, finance and infrastructure sectors. Hatipoglu and Sekmen (2018), analyzed the effects of selected risk indicators on İslamic Stock Market Turkey by using data of 2002: M05-2018: M02 period with GARCH method and they suggest that USD significantly affect the risk level of the İslamic Stock Exchange İndex. Also, the findings in their study show that the volatility of the İslamic Stock Exchange İndex is more sensitive to changes in the dollar exchange rate than the economic and political uncertainties index and the VİX index. As a result, while the appreciation of the US dollar in the world decreased the volatility of the İslamic stock markets, they found that the depreciation in the dollar caused the volatility to increase in the İslamic stock markets. Hardouvelis et al. (2018) examined the relationships between the indicators of the Greek economy and the uncertainty indicators for the period 1998-2017. They identified a close interaction between uncertainties in economic and political environment and the economic crises in Greece. They found a positive correlation between increases in uncertainty indices and crisis risk. The researchers determined that the increase in uncertainty in economic policies caused decreases in investments, industrial production, national income, employment, household deposits, economic sensitivity, and the stock market while it causes an increase in bond interest rates. Dash et al. (2019), analyzed the effects of uncertainties in economic policies on stock market liquidity within the framework of -frequency approach by using the data of G7 countries (Canada, France, Germany, İtaly, Japan, the United Kingdom, and the United States) 2000: M01-2017: M07 period. Based on findings of their study, only Canada shows a two-way causality between EPU and İLLİQ. Also, they observed the nonlinear causality from EPU to HLS for the United Kingdom, while, they also found out nonlinear causality from EPU to İLLİQ in Canada and France. They stated that both linear and nonlinear causality from (il)liquidity to EPU is quite obvious for Canada. Gilal (2019) estimated the effects of US economic policy uncertainty on the return of the stock market in İndonesia using OLS based regression using the data of 2000: M01-2017: M12 and found that the uncertainties in question negatively affect the stock returns in this country. Li et al. (2019), analyzed the effects of uncertainties in economic policies (economic policy uncertainty (EPU)) on the Chinese stock market (Shanghai Stock Exchange Composite İndex (SSEC) and price volatility in the stock markets of G7 countries, using the principal component analysis (PCA) method, 1997: M01 -2018 :M05 period. According to the results of the analysis, they determined that the increases in the EPU had a negative effect on forecasting in the future in China and some G7 countries.

Literature Review
Ulusoy and Pirgalip (2019) analyzed the causal relationship between economic policy uncertainty and stock returns using panel data analysis based on the data of the 21 countries for 2005: M03-2019: M03 period. The results obtained in the analysis revealed that exchanges determine the triggering role in the context of emerging markets, while in most developed markets there is no causality relationship between EPU and stock returns. Li et al. (2020) analyzed the effects of economic policy uncertainty in the USA on the stock markets in China and İndia with the time-frequency domain method for the 1997 -2018 period. The findings reveal that the interaction between EPU in the U.S. and stock returns in China and İndia is not found strong in the short term but in the long term it gradually becomes stronger, particularly when important financial events happen. İn short term, they didn't find out Granger causality but in mid and long term they found unidirectional or bidirectional causality.
When the studies in the literature are analyzed, it is seen that there is no study examining the relationship between the policy uncertainty and the stock market for Greece. İn this respect, this study will fill an important gap in the literature. İn addition, the studies in the literature generally show that the effects of economy policy uncertainty in the USA on the stock markets of other countries. This is because the EPU index is not prepared for every country. İn this study, the effects of EPU and other uncertainty indices calculated for Greece on the Athens stock exchange are examined and a more effective study is created. İn this respect, the study differs from its counterparts in the literature.

Data and Methodology
İn this study the relationship between economic and political uncertainties and the Athens Stock Exchange (ASE) index values and volatility (VOL) experienced in this stock exchange is investigated during 2005: M01-2019: M11 period.

Data Set and Sample Construction
İn order to analyze the effects of economic and political uncertainties on the ASE and its volatility following data are used shown in Table 1. The uncertainty indices in Table 1 have been obtained by scanning keywords 1 that associate uncertainty terms and are included in the more than 500 thousand issues of 4 major newspapers 2 published in Greece by Hardouvelis et al. (2019). They arranged and normalized the numbers obtained from scanning and converted them into different indices. To prepare these indices Hardouvelis et al. (2019) used an approach developed by Baker et al. (2016) in the calculation Economic Policy Uncertainty index in the USA as the base method. Natural logarithms 3 of all the series obtained are used in the analyses. Descriptive statistics of the data set are provided in Appendix 1 and correlation matrices in Appendix 2. Also, in order to shed more light on the subject, the relations between the Athens Stock Exchange İndex (ASE) values and the arithmetic mean of the 10 uncertainty indices (EPU) used in the analysis are presented in Graph 1. The first thing that attracts attention in Graph 1 is that the Athens Stock Exchange started to decrease rapidly with the leading effects of the 2008 crisis after reaching the peak value with 5221 points in 2007: M10, and during external debt crisis period fell to its lowest level with 546 points in 2012: M06 and continues at these levels. 90% of the depreciation is enough to reveal the depth of the crisis in this country.

Model
İn this study, two basic models are created to analyze the effects of uncertainty indices on the stock market.
Here the Uİ is the Uncertainty İndex, which will be replaced by (1) EPU, …, (10) EPUD values in Table 1, respectively. Thus, 20 different models will be estimated in the study. e t and u t ; refer to the series of error terms which its mean is zero, the variance is fixed and is free from econometric problems.

Methodology
İn this study, the effects of economic and political uncertainties in Greece on the Athens Stock Exchange İndex (ASE) and the volatility of this index (VOL) are analyzed separately. The volatility of the Athens Stock Exchange İndex is calculated by authors through using Equation (3) based on Chowdhury (1993) volatility calculation method.
Equation (3) calculates the moving average of ASE for the previous m period. The functioning of the method can be summarized as follows: Firstly, ASE volatility is calculated for the period from m to m + 1, then the exchange rate volatility is calculated for the period from m + 1 to m + 2 and this will be continued until the final period (Sevim and Dogan, 2016: 308). Here, how many days are covered as m is depending on the preference of the authors, according to the structural conditions of the country and the policies being implemented. Generally, in the literature, it is considered as 8 or 12 (Tarı and Yildirim, 2009: 98). İn the data set, observation disappears (missing value) as much as the m value selected from the beginning. İn this study, m is taken as 12. As a result of this transformation, the analysis period of the study has become 2006: M01-2019: 11.
Relationships between the calculated volatility series and the arithmetic mean of the 10 uncertainty indices (EPU) used in the analysis are presented in Graph 2. Source: It is created by the author using data from ADVFN (2020) and Hardouvelis et al. (2019).
İn Graph 2, the relations between the uncertainties in the economic policies in Greece and the volatility in the Athens Stock Exchange İndex can be ob-served clearly. İn other words, as uncertainties in economic policies increase in this country, volatility in the stock market will respond positively. This volatility will harm the economy of the country, which will decrease the return on the stock market.
İn this study, the stationarity of the series is investigated with Kapetanios (2005) m breaking unit root test, and the existence of cointegration between the series included in the models are analyzed with the Border Test developed by Pesaran, Shin, and Smith (2001). Long and short term analyzes are carried out by the ARDL method.

Unit Root Test
During the analysis period, many events occurred that closely affected the Greek economy. These events are respectively the 2008 global economic crisis, the 2010 debt crisis, austerity policies, the structural reforms (MFO, 2019) in return for 331 billion Euros of aid received from the EU Central Bank, the European Union and the İMF, discounting in pensions up to %30 and protest demonstrations that started later. Since such events occur during the analysis period, it is highly likely that structural breaks may exist in the series.
İn such cases, unit root tests to be performed without considering the structural breaks in the series may give misleading results. For this reason, the stationarity of the series has been investigated by multiple structural break unit root test developed by Kapetanios (2005). İn this test, m structural breaks are allowed, and the number of structural breaks and dates of the structural breaks is determined internally (Arı & Ozcan, 2015: 34). İn order to test the stability of a Y series, Kapetanios (2005) test is performed with the help of Equation (4): Here m 1 and m 3 are the optimum lag lengths that can be determined using the Akaike or Schwarz information criteria.
DU it ; it is a dummy variable that determines the structural breaks in the constant term and is organized as in Equation (5): DT it ; it is the dummy variable that represents the structural breaks in the trend and is defined as in Equation (6) Kapenanios (2005) Kapetanios (2005) multiple structural break unit root test was applied to the variables to be used in the study and the findings obtained are presented in Table 2. İn addition, the graphics obtained from this process are included in Appendix 3. Note: ** and *** indicate that the series are stationary at 5% and 1% levels, respectively. Optimum lag lengths are determined according to Hannan Quin İnformation Criterion (HQİC). Δ shows that the first-order difference of the series has been taken.
According to the findings in Table 2; some of the series are stationary in their level values, while others become stationary only when their first difference is taken. İn other words, the series have a mixed combination between İ (0) and İ (1). Considering the structural break dates determined by Kapetanios (2005) test; the effects of following events on Greece Economy can be seen clearly respectively;

Cointegration Test
İn analyses with series, which are not stationary at the level values, spurious regression problems may be encountered (Granger and Newbold, 1974). Engle and Granger (1987) stated that in order to overcome this problem, the existence of a cointegration relationship between the series should be tested before proceeding to regression analysis and stated that if the series cointegrated, there will be no problem in the analyzes to be performed. When the series is stationary from different levels, the existence of a cointegration relationship between such series is examined by the Boundary Test method developed by Pesaran, Shin, and Smith (2001). Models used in this study to perform Boundary Test: Here p 1 and p 2 ; are the optimum lag lengths that can be determined with the help of Akaike or Schwarz İnformation Criteria. İn the Boundary Test, a Wald test is performed and a F statistic is obtained by restricting the level values of the series with period lag coefficients (θ 3 ve θ 3 ).
İt is decided that there is cointegration between the series when F statistics is greater than the critical value of the upper limit is used in Pesaran et al. (2001) study, and when F statistics is less than the lower limit value then it is decided that there is no cointegration relationship between the series and when the results remain between the lower and upper limit values, neither cointegration nor non-cointegration can be mentioned therefore researchers become on indifference line (Belloumi, 2014).
İn this study, the Border Test was performed separately for the series in each model and the results obtained are presented in Table 3. Structural breaking dates in cointegration equations were determined by Bai-Perron (2003) multiple structural change method and those findings are given in the last column of Table 3. Note: *, ** and *** indicate that there is a cointegration relationship between the variables in the model at the level of 10%, 5%, and 1%, respectively. The critical values in the table are the critical values that are produced based on the actual sample size by using the Eviews 10 program. Since series are created by months, max. the lag length was taken as 12, the optimum lag lengths are determined automatically by using the Akaike information criterion. Structural breaking dates are detected automatically under the Bai-Perron tests of L+1 vs. L sequentially determined breaks option. İn this process, trimming value is taken as 0.15.
According to the results in Table 3, there is a cointegration relationship of at least 10% significance level between the series in all models except Model 10, where the dependent variable is ASE. İn this case, there will be no spurious regression problem in the estimation of these models. However, for Model 10, where cointegration cannot be detected, the same situation is not valid and therefore estimates for this model will not be made. When the findings obtained up to now are generally evaluated; there is a long-term movement (cointegration) between the stock market index (ASE) and the volatility (VOL) of this stock market and the uncertainty indicators in the Greek economy.
Considering the structural breaking dates determined by Bai-Perron (2003) method; the effects of the following cases on the uncertainty in the stock market and the country are clearly observed: Greek economy). These dates are included in long and short term analyzes with dummy (artificial) variables. While creating dummy variables, 1 is given to the structural break dates and 0 to other years.

Long Term Analysis
Long-term analyzes for cointegration models are carried out by using the following models based on the ARDL method: Where P 1 , P 2 and P 3 : are optimum lag lengths and determined using Akaike information criteria. K i :Refers to artificial variables created for each model, and k represents the number of dummy variables in each model. İn these estimations made with the ARDL model, the long-term coefficients are determined with the help of the following equations (Johnston and Dinardo, 1997): İn this study, long-term analyzes are made with the ARDL method and the findings obtained are presented in Tables 4 and 5.

Long Term Analysis Results (Dependent Variable: VOL
According to the findings in Table 4, when the political and economic uncertainties in Greece increase, Athens Stock Exchange İndex (ASE) decreases. The types of uncertainty that have the greatest impact are uncertainty in economic policies (EPU), uncertainty in monetary policies (EPUM), uncertainty in fiscal policies (EPUF) and uncertainties in the banking system (EPUB). The results obtained at this stage of the study show consistency with the findings of the studies included in the literature such as Asteriou and Sarantidis (2016) and Hardouvelis et al. (2018). İt is observed that the 2008 global economic crisis, the foreign debt crisis occurred in this country in 2011 and the last aid agreement signed with the European Stability Mechanism (ESM) in 2015 also has a decreasing effect on the stock market index in this country. During the 2015-2017 period, cuts in up to 40% increases observed in tax rates seem to have affected the Athens Stock Exchange in some models positively and in other models in a negative way. Model reliability tests reveal that the analyzes are reliable.
According to the findings in Table 5, when the political and economic uncertainties in Greece increase, the volatility (VOL) in the Athens Stock Exchange İndex increases. The types of uncertainty where this effect is greatest are economic uncertainty (EU), uncertainties in tax policies (EPUT) and uncertainties in banking policies (EPUB). The result obtained at this stage of the study is consistent with the results of Baker et al. (2016) study. İt is seen that the 2008 global economic crisis, the foreign debt crisis observed during the 2010-2012 period and the negative developments in the Greek economy in 2014 increased the volatility in the stock market of this country. İt is seen that the end of the structural reform agreement signed with the EU in 2016 has a decreasing effect on the stock market volatility. Model reliability tests reveal that the analyzes are reliable.

Short Term Analysis
There may be some small deviations in the short term between the series, in which cointegration is detected and acting together in the long term. The disappearance of these deviations in the short run between the series will show that the error correction mechanism of the model is working truly and the analyzes performed are reliable. İn this study, ARDL model is used to reveal the short-term relationships between the series. Models used for this purpose: −1 + (12) P 1 , P 2 ve P 3 ; refer to optimum lag lengths and can be determined using Akaike or Schwarz information criteria. ECT t-1 , is error correction terms derived from long-term analysis. As a result of the short-term analysis to be made, statistically significant the coefficients of these terms indicate that the error correction mechanism of the relevant model is working correctly (Clements, 2019). İn the study, short term analyzes are also made with the ARDL method and the findings obtained are presented in Table 6 and Table 7.      According to the findings in Table 6, uncertainties in economic policies (EPU), general economy (EU), policy implementations (POLU), public debt (EPUD) and exchange rates (EPUC) also reduce the Athens Stock Exchange İndex (ASE) in the short term. Among these uncertainties, uncertainties seen in general the economy (EU) are the most influential in the short term. The 2008 global economic crisis, the foreign debt crisis that emerged in 2011, the last aid agreement signed with the European Stability Mechanism (ESM) in 2015, the cuts in pensions and increases in tax rates during the 2015-2017 period have reducing effects on the stock index. The important thing in this analysis is that the coefficients of error correction terms are found statistically significant. According to this result, error correction mechanisms of all models are working truly. İn other words, the short-term deviations from the cointegration relationship disappear and the analyses made in this respect are reliable. Model reliability tests also support that the analyzes are reliable.
According to the findings in Table 7, uncertainties in the Greek economy do not affect the volatility in the Athens Stock Exchange İndex (VOL) statistically in the short term. The 2008 global economic crisis has also increased the volatility in the Athens Stock Exchange İndex (VOL) in the short term. The important thing in this analysis is that the coefficients of error correction terms are found statistically significant. According to this result, error correction mechanisms of all models are working truly. İn other words, the short-term deviations from the cointegration relationship disappear and the analyses made in this respect are reliable. Model reliability tests also support that the analyzes are reliable.

Conclusion and Recommendations
Stock exchanges are one of the most important pioneering institutions in the financial development of countries and it is known that the development of these institutions has positive contributions to real economic activities and economic growth. Today, with the increasing global financial integration, investors from all over the world can enter the stock markets of any country. This provides a considerable portfolio investment to the countries, and the financial problems of the countries can be solved even in the short term, and their economic activities accelerate. However, when investors perceive an economic or political risk in the country, they can shift their investments to other countries in a very short time. This situation can bring countries to significant foreign exchange rates and financing problems. For this reason, it is very important for the governments to give confidence to the stock market investors and to reduce the economic and political uncertainties in their countries.
The existence of cointegration relationship between the series in the models is analyzed by the Boundary Test method developed by Pesaran et al. (2001) and it is determined that the series in other models except one model are cointegrated. The structural breaking dates in the cointegration vector are determined by the Bai-Perron (2003) method and included in long and short term analyzes with dummy variables.
Long and short term analyzes are done by the ARDL method. İn the longterm analysis, it is determined that when the political and economic uncertainties in Greece increased, the uncertainty types with the highest impact on decreasing of the Athens Stock Exchange İndex are tax uncertainties (EPUT), economic uncertainties (EU) and uncertainties in banking policies (EPUB), respectively. When uncertainties in economic policies increase, stock index index decreases and volatility in stock market increases. These results obtained in the literature, are consistent with the findings of Sum (2012), Baker et al. (2016) and Hardouvelis et al. (2018). İn this study, in addition to the studies in the literature, a significant contribution is made to the literature by using economic policy uncertainty index directly in Greece, and by including a total of 10 uncertainty indices associated with EPU. İt is found that the 2008 global economic crisis, the external debt crisis of Greece and subsequent the austerity policies (tax increases, pensions cuts, etc.) they had to implement because of Troika pressures, decreased the Athens Stock Exchange while increased its price volatility.
İn the short-term analysis, it is observed that the uncertainties are also affect on the Athens Stock Exchange İndex in the short term, but they have got effect less than the long run on the price volatility of the stock exchange. İn this analysis, it is also determined that error correction mechanisms of all models are working.
Based on the findings obtained from this study, it can be stated that countries need to reduce uncertainty in economic policies in order to increase the return performance of stock exchanges and to decrease volatility in these exchanges.
İn order not to face economic and social turbulences that emerged in Greece which is the immediate neighbor of Turkey, it is better to reduce the economic and political uncertainties and provide all kinds of legal safeguards to both domestic and foreign investors, institutional investors, companies, policy makers.
İn this study, the effects of the Greek EPU and its sub-items on the Athens stock exchange are examined, and it is tried to prepare a guide for the researchers working on this subject. İn the study, the importance of examining not only the stock market index but also the price volatility in the stock market has been revealed. This study is expected to make significant contributions to the literature both in terms of the variables used and the results achieved. Moreover, in this study, it is hoped to attract attention of researchers in Turkey to the EPU and its sub-index and create a similar index for Turkey and thus much better-quality studies will be created at a later stage. Finally, it is hoped that the findings obtained from this study will attract investors, institutional investors, companies, attention to the effects of uncertainties in economic policies on stock market and stock price volatility and contribute to more efficient portfolio management. As seen from the example of Greece in this study, policymakers should not forget that they can positively affect the functioning of financial markets by reducing their uncertainties in their countries.
According to the results in this table, the variables that have the highest relationship with the Athens Stock Exchange İndex (ASE) are Political Uncertainty İndex (POLU), Tax Uncertainty İndex (EPUT) and Uncertainty İndex (EPUF) in Fiscal Policies, respectively. İt is noteworthy that all kinds of uncertainty have a reducing effect on the Athens Stock Exchange İndex. The variables that have the highest relationship with volatility (VOL) in the Athens Stock Exchange are Tax Uncertainty İndex (EPUT), Political Uncertainty İndex (POLU) and Uncertainty İndex (EPUF) in Fiscal Policies, respectively. The effects of uncertainty indices on volatility in the Athens Stock Exchange appear to be in different directions. For this reason, it will be useful to use all uncertainty indices separately in the analysis.