A Study Regarding the Advances of Political Stability and Economic Development Experienced in Turkey during the Periods of 1980-2015

In this study, the causality relationship between the political stability and gross domestic product (GDP) in Turkey with the annual data for the period 1980-2015 has been analyzed. The cointegration and Engle-Granger causality tests have been applied in the study. As the Series were cointegrated, the Error Correction Method has been used in order to carry out the causality test. According to the analysis results, the cointegration test suggests the presence of a long-term relationship between political stability and gross domestic product (GDP). As for the Error Correction Model, it reveals the presence of one-way causal relationship from political stability to GDP.


Introduction
In recent times, the needs regarding the impact of political stability on economic development and the studies related to this issue in the economic development literature have started to increase.The importance of political stability in view of the economic development is better understood when we look at the economic indicators of the countries in this process.For this purpose, the relationship between political stability and economic development has been researched and tested for Turkey's economy in the study.The economic development indicators are an issue which the researchers are trying to find the root causes and consider them as the most important indicator of a country's development level.The "political stability" within the framework of relations of "politics and economics" is evaluated within the scope of govermental periods in general (Aksu, Kanca, & Başar, 2015, p. 77).In this regard, the terms of office of the governments emerge as an important variable for the purpose of building the political stability in Turkey on solid foundations.Suggesting the truth that the political stability had positive impact on the development of the countries (Akıncı, 2015, p. 50) in general sense, necessitated the suggestion of importance of political stability in view of Turkey's growth process.In this regard, the objective of this study has been defined as investigating the political stability and the likely impact of political stability on the economic development, and evaluating the period of 1980-2015 for Turkey within this scope.
The incidents such as coups, crisis, takeover of power by non-democratic means experienced in Turkey's recent political history have revealed an ambiguity regarding the country's economic policies.That's why, it has been observed that the governments had important roles in economic issues and held responsible for the results of economic indicators in the countries, in general.This situation makes necessary to analyze the interaction between the political stability and the economic advances in Turkey.The variables used in this study consist of the governmental periods between 1980-2015, and GDP annual data.The data used in the study has been made more convenient for the stability test by taking their logarithms.In this study, first of all, the Augmented Dickey-Fuller (ADF) unit root test has been performed with aim of determining the the stability of the variables, then Engle-Granger cointegration test has been carried out with the aim of testing the presence of cointegration between the variables.Finally, the causality relationship between the co-integrated series has been tested and evaluated by using error correction model.This study consists of four sections together with the introduction and conclusion.The second section involves the summary information regarding the literature at a conceptual level outside the introduction and conclusion sections.In the third section, the information about the method and data set has been provided and the empirical findings related to study has been put forward and the assessments has been made upon them.

Literature Search
The development and growth concepts emerges among the major issues discussed in each period over the economy.We may say that the main reason of overemphasizing these concepts is that the economic development has a multifaceted structure concerns both developed and developing countries in view of social and economic aspects.Accordingly, economic growth, the increase in real gross domestic product per capita and the country's economic activities, refers to a scale of increase (Kibritçioğlu, 1998, p. 208).In other words, the economic growth is the increase occurred in the capacity of goods and services produced and a country's economic growth means continuous increase of GDP expressed as the equivalent in market price for all the measurable values produced in that economy (Gündüz, 2008, p. 39).The economic development consists of changes upon the economic and socio-cultural structures due to the increase in production and per capita national income (Tolunay & Akyol, 2006, p. 118).The economic development is a more complex concept as it includes the changes in both economic and social structures.In this regard, the main purpose of economic development is to use the increasing economic opportunities with the aim of solving the social problems in order for people to live in prosperity (Aktuğ, 1998, p. 8).
Since the concept of economic development handled the prosperity from a broad perspective, this ensures the social, cultural and political variables included in the concept (Küçükkalay, 2001, p. 61).The development can be talked about in the full sense when the increase in production volume is stable and real (Aktuğ, 1998, p. 8).When we consider all above then we can say that the essential condition of ensuring the economic growth is to achieve the economic growth.Because, as a result of ensuring the economic growth, the structural improvements in resolving the social issues can be achieved.On the other hand, the concept of political stability is expressed as the creation of efficient and effective political structure for the country under the rule of law in a democratic process (Ozturk, 2004, p. 9).The political stability, in other words, represents stability and continuity in the management of a country.However, it is very important to examine the impact of political stability on the economic performance of a country and to put forth the existence of relations between political stability and economic development.The major argument among those suggesting a country's political stability has impact on the economic development is that political stability affects economic decision-making processes, such as production, employment and investment.However, the frequency of changes experienced in governments in office has positively or adversely affect on GDP (Alesina, Ö zler, Roubini, & Swagel, 1992, pp. 3-4).In particular, the foreign investors prefer countries that has a political environment experienced less uncertainty and chaos.However, it is possible to sort the various obstacles in front of ensuring the political stability in a country depending on their degrees.Accordingly, the indicators for the political instability in a country are listed as wars, revolutions, coups, fall of governments, frequent changes in goverments and assassinations (Carmignani, 2003, p. 1).When viewed from the theoretical point of view, it may be argued that the political instability is one of important mechanisms that affect economic performance in a country.Because the political instability is responsible for the emergence of several weak macroeconomic indicators such as high inflation, unemployment, low growth rate and high debt level (Miller, 1997, p. 12).To sum up, that the investments showing improvement in direct proportion with the stability and the national income affected by this are one of major factors underlying in the relation with the politically experienced uncertainties and the economic performance in theoretical aspect (Karahan & Karagül, 2014, p. 2).In general, it is accepted that the governments of the countries have important roles in economic matters.The governments are held responsible for the consequences of the economic indicators.This situation makes it necessary to analyze the relationships between the political stability and economic developments.
The empirical studies carried out in the economics literature points out to various aspects of the relations between political instability and economic growth.In some of these studies done, it is stated that there was a direct relationship among the variables and in some others stated that the relation existing is indirect.The empirical studies to investigate the effect of these arguments studied from a theoretical point of view, have tried to put forward the influences of the political stability environment of the countries upon economic growth.According to this, in a study covering the process between periods of 1960 and 1985 in 70 countries, the relationship between the political instability and economic growth was correlated with the physical capital accumulation and investment.It has been put forth that the physical capital accumulation and investment could be realized only depending the political stability.In addition to this, that the political instability was a variable leads to both income inequality and decrease in investments has been highlighted in this study (Alesina & Perotti, 1993, pp. 1-2).In another study it has been examined the relationship between political stability and economic growth between the periods of 1950-1982 in 113 countries.The government changes have been assessed as a political instability criterion.According to this, the government changes (the transfer or fall of government and so on.)have been found to have negative impacts on economic growth performance (Alesina, Ö zler, Roubini, & Swagel, 1992, p. 1).Another study to test the validity of arguments based on the relationships that have been established theoretically between the economic growth and political stability, has been carried out for the process covering 100 countries and 1960-1985 periods.In this study, it has stated that the current governmental decisions determined the public expenditures but its long terms decisions was bound to the political stability.However, it was stressed that the political instability posed an obstacle in taking long-term decisions for the governments by providing a balance in public expenditures between the consumption and investment (Persson & Svensson, 1989, pp. 325-326).Based on all this, we could say that the uncertainty resulting in a country where the political stability is not achieved in the political system, could reduce the private investments and affect economic growth in a negative direction.The greater the ratio of changes of Governments, increases the potential uncertainities about the new policies of a new Government.This situation causes the economic entities avoiding the risk, give up vital economic decisions or move these investment decisions out of the country (Alesina etc., 1992, p. 4).At the present time where the globalization and financial liberalization are prevailing, to establish the political stability and make it sustainable has great imprtance in securing the financial capital and not to cause the investments move out of country.Otherwise, political instability and the resulting political uncertainties may have negative impact on the economic growth by changing the kind of investments or by affecting the structure of public expenditures (Demirgil, 2011, p. 124).In this context, the results obtained from empirical studies conducted in 169 different countries from the year 1960 to 2004 revealed the fact that the political instability and in addition, the Cabinet changes have significantly decreased and had a downward effect upon GDP per capita ratio (Aisen & Veiga, 2010, p. 4).In the studies examining the impact of political instability on economic growth, as a model, it has been demonstrated that it increased the uncertainty of the political instability, negatively affected the investment and capital accumulation and had negatively influences on economic growth (Alesina et al., 1992, p. 4).However, the political instability environment as seen in the general of coalition governments formed, leads to a weak structure emerging on both social lifes and macroeconomic indicators of the countries (Sobacı, 2015, p. 8).A large number of empirical studies has been made since the 1960s with the aim of examining and identifying the relationships between political instability and economic growth.(Alesina etc., 1992), (Persson & Svensson, 1989), (Alesina & Perotti, 1993), (Alesina & Rodrik, 1994) and (Asteriou & Price, 2001) can be noted as examples to such studies.However, in these studies, that the relationship between economic growth with political instability has been reviewed just in one country in particular and that the country's political structure and properties have not been revealed in a specific manner, emerges as a shortcoming.In addition, many different variables in the studies examining the political stability (ability to implement government programs, if they come to office with the elections, wars, coups, changes of government, coalition, etc.) have been used.In this study, the durations of government in office have been used as an indicator of political stability.On the other hand, to investigate the impact on GDP have been found to be more meaningful due to the variability of growth differences between countries, both in order to obtain current data and to determine the effect of the relationship.

Data and Method
The relationship between political stability and economic growth which is fundamental condition for economic development have been analyzed in the study.The variables used for econometric analysis consists of annual data for the period of 1980-2015 in the case of Turkey.In this study, the cointegration test has been conducted according to the Engle-Granger (1987) method.The Cointegration is a technique developed that was developed to examine the correlation between the two non-stationary time series (Demirhan, 2005, p. 80).However, the causal relationship between the data in the short and long term for co-integrated series has been tested with error correction model.The cointegration analysis is a process examined if there is a long-term equilibrium relation between series.If the variables are cointegrated, the causal relationship between short and long-term data is created by adding the error correction term to the causal model (Sandalcılar, 2012, p. 8).That the cointegration has been found at the result of cointegration test conducted in order to determine if there are long-term relationship between the variables, means that there is a long-term relationship between these two variables (Demirhan, 2005, p. 80).The stationary test used in the study has been conducted with Augmented Dickey-Fuller (ADF) unit root test.In case that the variables are stationary in their first differences, then it should be looked at if the variables are cointegrated.If the variables are cointegrated that it means there is a long-term relationship between variables (Demirhan, 2005, p. 80).For this reason, the cointegration test has been conducted in the study after determining that the variables were cointegrated in the first degree.
In the analysis section of this study; real gross domestic product (Y), political stability (SX) variables have been used.The variable used as an indicator of the political stability is the date created in the scope of durations of governments in office.The statistics from Turkish Republic Prime Ministry site have been used in compiling data of durations for the goverments in office.The Economic and Social Indicators (2015) statistics from TR Ministry of Development have been used in compiling data for GDP variable.However, the impacts on GDP have been included as the economic growth is the essential condition of economic development and express the market price equivalent of the all measurable values produced in a economy and the growth rates vary among countries.

Unit Root Test
The Augmented Dickey Fuller-ADF (DICKEY/FULLER,1981) unit root test should be conducted in order to test if the variables are stationary (Demirhan, 2005, p. 79).Accordingly, the stationarity properties of the time series used in the study has been applied by using the Augmented Dickey-Fuller (ADF) unit root test developed by Dickey Fuller (1981).This test includes three different regression relationship shown in equations revealed by Dickey Fuller.
(3) Accordingly, the simple model created from Yt series with the aim of explaning the Dickey-Fuller test applied to determine if the series contain unit root in the conducted application and consisting first-order Autoregressive (AR1) process may be shown as below.
∆  = ∆ −1 +   −1 ≤  ≤ 1 (4) In the equation above (4), x refers to the Autoregressive coefficient of the model.If it is x the IγI < 1 Y t series does not contain unit root or stationary.
IγI = 1 If it is x then Y t series contains unit root or not stationary (Sandalcılar, 2012, p. 7).
Many different unit root tests have been developed by making different assumptions from the model expressed with the equation in (4).ADF unit root test refers to three different regression relationships shown in the equations (1), ( 2) and (3) above.In these equations ∆  x refers to first difference ∆ − of variable stationarity analysis made, t = time trend, k delay length, delayed   difference terms, x error term with zero average, constant variance, probability non consecutive-dependent (Demirhan, 2005, p. 80), (Sandalcılar, 2012, p. 7) and (Doru, 2013, pp. 48-50).Here, ADF test  tests if the coefficient is statistically equal to zero.If zero hypothesis is an  = 0, alternative  < 0 hypothesis then it is x.As a result of the test, if the null hypothesis is not rejected than it is concluded that the series contains unit root (Demirhan, 2005, p. 80 The results of the ADF test of the variables used in the study are shown in Table 1.The variables belong to the ADF unit root test results are shown as with constant, with constant-trend, with no constant-trend.The results for unit root tests conducted for variable in level indicate that the variables are not constant.On the determination that the variables are not constant in their level values, they are made constant by taking the differences of non-constant series.In this regard, the stationarity test has been reapplied by taking the differences of the series. According to the results shown in table 1 above, the variables are not stationary in level but they become stationary when their first differences are taken.This shows that the series I(1) are integrated in level, in other words, the variables are integrated in first I(1) degree.It is possible to apply the co-integration test between variables.The optimal delay length in ADF test has been determined according to Schwarz Information Criterion (SC).

Engle Granger Cointegration Test
Linear combination of two non-stationary series is not stationary.The variable should be provided as stationary in same degree.In this respect, if the residual series are stationary then it may be stated that these two series will go in long term and they may have a new long-run equilibrium relationship (Ata and Yücel).For this purpose, if these two series have long-term relationship has been analyzed by Engle Granger method.In the study, firstly the regression model was established between the two series and then tried to guess residuals of the model.The presence of cointegration between GDP and political stability variables has been estimated by applying Engle-Granger cointegration test.The results obtained are shown in Table 2.
According to the results, shown in Table 2 above, the coefficient showing long-term relationship between two variables were positive and significant.However, since the variables were stationary in same degree, the co-integration situations between variables should be identified.For this, the regression model error values shown above should be obtained and stationarity test should be conducted for the error values.The stationarity of this series is tested by applying Augmented Dickey-Fuller test.Accordingly, the error series of the model obtained was called as error (residue) and ADF test was applied to this error series.The stability results regarding the error values of the model were shown in table 3 below.H 0 :  1 = 0 Series is not stable, unit root is present.
H 1 :  1 <0 Series is stable, unit root is not present.
According to the test results shown in Table 3, Model 3 is constant and trend in line with the ADF ADF test statistic value compared to the critical value of a unit root test results in a 5% significance level that the calculated values of the small table value H 0 hypothesis is rejected.Therefore, there is no unit root and it is stationary.In other words, null hypothesis where the error term has unit root is rejected and stated that it is stationary in level.Based on all of this, we can say that two series examined are cointegrated.Series are cointegrated.Therefore, a long-term relationship between the series are present.We can state that there are at least one cointegration relationship between two variables.That there are a long-term relationship between the variables justifies the arguments based on that the political stability affects the country's macroeconomic indicators.The most important argument among them is that the political stability environment is decisive in decision-making processes, such as investment and production.Because, the political stability environment is determining factor for the investors determine their direction of decisions and in the roles where the economic actors play in market and their decisions.Based on all this, once the there is a cointegration between GDP and political stability, that is, there is a long-term relationship among the variables, degree and direction of this cointegration should be identified.For this reason, error correction model is used in determining the degree of relationship between cointegrated variables.

Error Correction Model
The error correction model is a stage followed after the determination of the long-term relationship between series and is used to determine the causal relationship and its direction between the examined series (Aktas, 2009, p. 40).In other words, the error correction model is used to determine the causal relationship between the cointegrated series in short term.According to Granger (1988), if there is a cointegrated relationship between the studied variables then it must be at least one-way causality between the variables.In this regard, it is more appropriate that the causality analysis is tested with error correction model (Kıran, 2007, p. 273).The error correction model is a model that is implemented using the values of variables in level where they are constant and one-term delayed value of error term in order to predict the relationships occured between the variable in short term (Yıldız ve Aksoy, 2014, p. 17).In this context, we may say that the error correction model is created with by adding the error term delayed value to the model and in addition it is used in determining the direction of causality relationship.In addition, Granger (1988) stated that when the variables discussed are cointegrated the the standart Granger casuality will not be valid and it is more appropriate that the casuality analysis of the series in this state is conduected with Error Corection Model (ECM) (Aktaş, 2009, pp. 40-41).
The error correction models are used when a long-term relationship between the variables are present.Because, error correction model indicates the deviation in long term (balance) relationship.The long-term relationship between them can be considered that the variables are cointegrated.The series should be stationary in order for to mention about the formation of cointegration concept.The difference process is applied to the series to provide stationarity.However, losses in long term information occur during the implementation of difference process.Therefore, using the error correction model, it is tried to eliminate these imbalances.In the error correction models, a delayed state of error terms is added to the model established with the stationary variables.
The error terms should be stationary in the level value.Accordingly, the error correction is shown as in the following equations in order to test the casuality between the GDP and political stability variables discussed in the study.
∆ − +  3  −1 +   (7) In the equations 6 and 7, the delayed value of error terms obtained as a result of ECT cointegration relationship is shown.The x parameters in the α 3i ve  3 model are the error correction parameters enforcing the variables come close to their long-term balance value.If these parameters are statistically significant, then there is a deviation from the balance.Also the parameter is expected to be negative and significant for coming close to value of the balance in the long term (Kıran, 2007, p. 273).The causality relationship between GDP and political stability variables will be tested in the model established by expressing the equations 6 and 7 above.On the other hand, it is tested that the causality relation to GDP from the political stability in the equation 6 above established in order to test the causality between the variables and if there is a causality relation from GDP to political stability in the equation 7 above.The hypotheses regarding the equations done accordingly are established as follows.
(6) H 0 : Political stability is not the cause of GDP.
H 1 : Political stability is the result of GDP.
(7) H 0 : GDP is not the cause of political stability.
H 1 : GDP is the cause of political stability.
Accordingly, the regression results obtained based on the equation ( 7) and equation ( 8) are shown in Table 4 below.The casuality test results based on error correction model in order to test causality from political stability to GDP in the equation ( 6), causality from GDP to political stability in the equation ( 7) are shown In Table 4.According to the error correction model test results, H0 null hypothesis established in the way "political stability is not a cause of GDP" in the equation ( 6) established is rejected.The alternative hypothesis is accepted.In response to this, H0 null hypothesis established in the way "GDP is not the cause of political stability" in the equation ( 7) establied is accepted.According to the regression results shown in Table 4, error correction parameter was found to be statistically significant and negative in the equation ( 6).No casuality was found from GDP to political stability in the equation ( 7).The coefficient of error correction term for the equation ( 7) is not statistically significant.According to these results, the error correction model indicates that there is a unidirectional causality between variables.The error correction parameter that was found to be statistically significant for the equation ( 6) forces the variables come close their long term equilibrium value.However, that the error correction parameter is statistically significant, shows the deviation from long-term equilibrium and indicates that when any deviation occurs then they can come to their equilibrium state again.That the coefficient for the error terms was found to be negative here shows us that some portion of imbalance was removed for every term and they came close to equilibrium level in long term, that is, they have a cointegration relationship.In this study, the coefficient for error correction term obtained for the equation ( 6) is 0.35.The error term coefficient shows us that how much the deterioration occurred in the balance last year improved for this period.According to these results, approximately 35% of 1 unit deviation improves in the next period.In other words, 35% portion of imbalance that may occur in political stability, recovers in the next period.
It reveals that 35% portion of imbalance is removed in each period.According to this, the deviations from the equilibrium/balance value will be disappeared after approximately (1/ 0.35≅ 2.85) periods.The imbalances in the short term is isolated in this way.As a result, in the short term, the casuality from political stability to GDP is true but there are no causality findings from GDP to political stability observed.

Conclusion
The main objective of this study was to analyze and evaluate the relationship between the variables with the annual data for the period of 1980-2015 in Turkey and econometric analysis using the political stability and GDP series in the scope of cointegration and causality tests.As a result of analysis done, long-term equilibrium relationship between the political stability and GDP has been identified.Depending on this, the error correction model has been established to determine the direction of causality.As a result of the error-correction model, it has been found that there was a casuality relationship from political stability to GDP.However, no finding supporting the casualty from GDP to political stability was obtained.In another words, the results of the analysis proves that there is a one-way causal relationship between GDP and political stability in Turkey.Based on the findings obtained, we can say that the political stability is one of the major causes of economic development for the periods of 1980-2015 in Turkey.Therefore, the durations of governments in office in Turkey are considered to be essential cause influencing the economic growth that is one of major factors of economic development.
One of the fundamental requirements to ensure the development is the economic growth.As the economy grows, GDP grows in other words, the level of the country's income and prosperity increases and standard of living rises.As a result, rate of economic development captures an acceleration in a positive direction.It is assessed that the political stability environment has been established discussed for the economic development defined according to the symptoms in economy is the major criteria.In a matter supporting this, in the periods of 1980-2015 discussed in this study for determing the importance of political stability, the economic growth has been influenced in different rates according to the duty durations of the governments as one-party or changing coalitions in office.In this context, the situation becomes clear when the impacts upon the economic growth of political stability demonstrated by the short-term coalition governments and long-term one-party government periods are concerned.Accordingly, in the period between 1980-2015 in Turkey, when we compare the rates of economic growth according to the periods of government for one-party and coalition then single-party goverments seem to be more successful.
The growth rates were respectively 5% and 4.8% during the periods of single party governments for the years of 1983-1991 and 2002-2015 while the average growth rates were respectively around 4.5%, 2.4% and 1.4% for the periods when different coalition governments were in office formed during the years of 1992-1996, 1997-1999 and 1999-2002 (TR Ministry of Development Economic and Social indicators, GDP growth rates through harmonized expenditures in prices of 1998 (1981-2014), 2015, p. 47), (http://www.tuik.gov.tr).This situations shows us that the political stability environment where one-party government is in power has more positive impact upon the economy in comparison to the coalition governments formed for short durations.Consequently, it was analyzed and proved the fact that the political stability was the cause of economic growth in Turkey for the period discussed and studied.The size of this relationship has been discussed in the scope of impact of political stability upon a macroeconomic indicator GDP and a casuality relationship between them has been found.Therefore, it was demonstrated that the political environment depending on the durations of governments in office in Turkey, is a cause for economic growth.In light of the results obtained from this study, the relevant period series is related with in long-term and have causality between them.Thus, under the constraints of the data set used, depending on the durations of governments remain in office, the predition saying that the political stability environment may have positive impact upon the economic development becomes acceptable.When we compare the political instability environment experienced during the short-term coalition governments are in office which demonstrates different scenes from each other in terms of political stability with the political stability environment experienced especially during single party government periods in 2000s, the macroeconomic advances experienced in Turkish economy gives a view supporting our study suggesting that the political stability is a cause of economic growth.However, this revealed result shows that the economic advances in a country are sensitive to the political developments.Based on all this, it is of great importance that Turkey maintains political stability environment experienced starting from the 2000s up to date in terms of achieving the targeted macroeconomic indicators.In the global new world finance order where the fact that each politics have economics and each economy have politics was known, it is of great importance to establish and maintain the political stability in order to reach to a economically developed level.In this empirical study performed, the result obtained for Turkey that the political stability is a cause of economic growth is in a way supporting the empirical studies conducted for the different countries worlwide. ).

Table 4 .
Causality test results based on error correction model