The Impact Of Financial Liberalization On Economic Risk In The Asia Pacific Countries

In this study, we analyze the non-linier effect of financial liberalization policy toward the economic risk of the Asia Pacific countries with trade openness as a threshold variable. To do that, we apply panel regression threshold proposed by Hansen (1999) as our method of analysis. Based on yearly data from 1975 – 2015, we find that there is a non-linier effect of financial liberalization on economic risk depending on the certain level of trade openness. Regarding to this finding, we find that when the trade openness is below the threshold value, financial liberalization policy can reduce the economic risk of Asia Pacific countries. However, when the trade openness exceeds the threshold value, financial liberalization will increase the economic risk. So that, we conclude that an open domestic financial market that is followed by high degree of trade openness will tend to create an economic instability.


INTRODUCTION
Globalization, which is marked by the strengthening of economic relations between countries, has increased the economic risk of countries in the world. In practice, the strengthening of economic relations between countries makes economic shocks that happen in one country can be quickly transmitted to other countries whether through financial or trade channels. As a result, there is no country in the world that can be fully protected from the cycle occured in the world economy (Aizenman et al. 2013).
Financial liberalization and the discussion of its effects on global economic risk has become one of the important issues studied in the economic literature. In relation to economic risk, Ozcan et al. (2003) found that financial liberalization policies can reduce a country's economic risk. This is because financial openness can help a country in obtaining additional capital which allows the country to improve its production quality. So that, financial liberalization can maintain economic stability.
However, it is still unclear whether financial liberalization really can provide positive benefit for the economy (Kose et al. 2003). This is because several other studies have shown different results compare to Ozcan (2003) and Mirdala et al. (2015). Easterly et al. (2001) which examines factors that determine economic risk, have found that a liberal financial system can actually increase the volatility risk of output growth. In addition to this, Kose et al. (2003) also found that open financial sector has led to an increase in the risk of consumption and income in a country. Neaime (2005), who analyzed the Middle East and North Africa (MENA) countries, found that financial openness has increased the economic risk of countries located in this region.
However, it is still unclear whether financial liberalization really can provide positive benefits for the economy (Kose et al. 2003). This is because several other studies have shown different results compare to Ozcan (2003) and Mirdala et al. (2015). Easterly et al. (2001), for example, which examines factors determining economic risk, have found that a liberal financial system can actually increase the volatility risk of output growth. In addition to this, Kose et al. (2003) also found that open financial sector has led to an increase in the consumption risk and income risk in a country. Neaime (2005), who analyzed the Middle East and North Africa (MENA) countries, found that financial openness has increased the economic risk of countries located in this region.
The difference in results that is found in the aforementioned studies indicate that there is still no clear conclusion regarding the impact of financial liberalization on economic risk in a country. One of the reasons why this difference could happen is that the effect of financial liberalization toward the economic risk might not be linear as always assumed by aforementioned studies. In response to this gap, the purpose of this study is to analyze the effect of financial liberalization toward economic risks in a non-linear manner. This non-linear aspect is still not widely discussed in previous studies.
To achieve the research objective, this study adapted the panel regression threshold approach developed by Hansen (1999). One of the main advantages of this approach is that the threshold value is not determined exogenously, but endogenously determined by the data. In its operations, this study suspects that the non-linear effect Source: External Wealth of Nations, dan Indeks Chinn-Ito (diolah) Picture 1. The Condition of Financial Liberalization in the Asia Pacific Region of financial liberalization on economic risks is due to differences in the characteristics of trade in each country. Financial liberalization followed by trade openness can affect the economic conditions of a country/region (Mirdala et al. 2015). Therefore, in relation to the financial liberalization and economic risk, the nonlinear effect of financial liberalization This study uses countries in the Asia Pacific region as its sample. These Asia Pacific countries were chosen because from 1970, the openness of the financial sector in this region was very accelerated compared to other regions (Chinn et al. 2007). The increasing movement of international capital flows occurring in the Asia Pacific, according to Borensztein (2011), is characterized by the presence of free and open capital mobility. This increasing capital mobility, as a consequence of the implemen-tation of financial liberalization policy, may effect the economic risk of the countries. toward economic risk is observed through trade openness that are treated as threshold variables. This study also includes several variables that are considered important in influencing a country's economic risk as a control variable, which includes inflation volatility, fiscal policy, institutional quality and volatility of terms of trade. Positive impact: if the countries specialized in a

LITERATURE REVIEW
Studies discussing macroeconomic uncertainty began to develop around 1990. At that time, Ramey and Ramey (1995) had discovered a detrimental effect of the output volatility on economic growth which ended in a decline of economic prosperity. Even in the short term, the economic volatility has the opposite effect and is detrimental to the poor people. Macroeconomic volatility affects expectations of economic actors related to production, consumption, price uncertainty, and unemployment. The negative relationship between volatility and growth that ultimately affects the level of well-being becomes one of the important problems that raises further research on what factors and sources that cause economic volatility to occur. Internally, in the modern view of macroeconomics, according to Easterly et al. (2000), uncertainty in the economy is determined by rational actions between companies and households, policy intervention by the government and even the complexity of collective behavior that brings the economy to quickly return to full employment.
In relation to the context of globalization, research on the economic crisis, which is an extreme manifestation of high economic volatility, has begun to be a concern since the 2000s. In this period, Kose et al. (2005) reveal that many crises are caused by external factors. This is because many developing countries open their economies faster than they should for trade and global finance. Theoretically, the effect of increasing trade openness and financial flows on output volatility, as a proxy for economic risk, depends on various factors, including the composition of trade and financial openness, patterns of specialization and other sources of shocks. Kose et al. (2003) revealed that in the relationship between trade openness and risk, increasing specialization of the production structure affects output volatility in the business cycle scheme. This is because the more specialized a country in certain industries, the effect of international shocks will tend to provide negative transmission and increase the economic risk.
Related to the financial openness, Kalemli-Ozcan, Sorensen and Yosha (2003) revealed that financial openness has strengths and weaknesses. The advantage of financial openness is that it can help reduce economic risk in developing countries by providing access to capital, so that it can help them diversify the production base. On the contrary, the weakness of financial openness is because it allows the creation of specialization of production so as to make the economy more vulnerable to a specific industrial shock. As important findings from Kose et al. (2003) who found that trade openness increases output volatility in developing economies because trade flows can increase the possibility of risk sharing. Important findings by Kose et al. (2003) regarding how external shocks exposed to the domestic economy are relevant to study. To date, the results of various studies on the effects of financial openness and trade are still debatable. This is because previous studies find both positive relationship and a negative relationship between financial openness and trade toward economic risks. From the point of view of financial openness, the debate arises because there are no clear conclusions and there are still many debates about the relationship between financial liberalization and macroeconomic volatility. The lack of clarity about the relationship is expected to occur because of the two major forces in financial liberalization. These two forces, on the one hand, may reduce economic risks, but on the other hand, they can also increase the economic risk of a country. In this case, international financial openness can reduce the economic risk because of diversification in sharing risks. But at the other side, financial openness can lead to greater specialization and thus increase the risk of the domestic economy. Neaime (2005) found that financial and trade openness in poor regions such as the JBPE Journal of Business and Political Economy, Vol 2 (1), 2020 Middle East and North Africa (MENA) have a positive relationship to economic volatility. That is, the more open and integrated financial and trade systems in MENA to the world economy, the more it will have a detrimental effect on economic risk in terms of output and consumption. Later, Ahmed and Suardi (2009) have examined the effect of trade and financial liberalization on macroeconomic volatility in the Sub-Saharan African region. The results show that there are differences in the effect between financial openness and trade on macroeconomic volatility. Using the Panel-GMM system method, the results show that financial openness has a positive and significant influence on macroeconomic volatility in the Sub-Saharan African region. It means that the more financial openness increases, the more it will increase the macroeconomic uncertainty. On the contrary, trade openness policy has a negative impact on macroeconomic risks. This means that the more liberal and integrated the trade systems in Sub-Saharan Africa, the more it will reduce the macroeconomic volatility. Similarly, Mujahida and Alam (2013) have examined the effects of trade and financial openness toward economic risks in Pakistan. Using the Autoregressive Distributed Lag (ARDL) method, the results show that in the long run trade openness has a negative and significant effect on the economic risks. Meanwhile, financial openness has a negative impact on investment risks, which reflects that an increase in financial openness causes a decrease in investment risk in the short term.
Related to the above, Yang (2011) analyzed the impact of political democracy and economic liberalization toward economic risks using the difference in difference method for 158 countries from 1970 -2005. The results have showed that economic liberalization %which are financial and trade openness% can reduce economic risks, but the same thing is not found in political democracy. However, the implementation of democracy after economic openness is a good choice because it gives a positive effect in reducing the economic risk. This result serves to provide additional support for policy recommendations that developing countries must liberalize their economies first and then consider to liberalize their political system by implementing democracy.
Another interesting finding, that gave rise to a new theoretical perspective is the study done by Barrot et al. (2018). Barot et al. (2018) has identified four structural shocks -demand, supply, monetary and commodity shocks and linked their impact to state policies and structural frameworks. Shocks originating from external factors, such as trade, in the past few decades have had a greater impact on increasing the economic risk of a country compare to the shocks originating from domestic factors. Global monetary policy shocks are one of the main external sources of economic risk in developing countries. An increase in openness will increase external disruption to the economy. Another interesting fact is that in the case of trade openness, the contri-bution of variance to global disturbances is non-linear, following a Ushaped pattern. In this case, external disturbances will have a greater effect on low and high openness than intermediate levels. In addition to this, an important finding is that commodity intensive countries show are riskier for all types of external shocks.

1Panel Threshold Model
The main purpose of this study is to identify whether there is a non-linear effect from financial liberalization policy toward economic risk in the Asia Pacific countries. Thus, to accommodate this purpose, this reseach adopt an approach called panel threshold model introduced by Hansen (1999). The main advantage of this approach is that the threshold value is not arbitrarily determined, so that it allows us to get the confidence interval for the chosen interval. Beside that, the other advantage of the endogenous threshold regression technique include (1) it does not require any specified functional form of non-linearity, and the number and location of thresholds are endogenously determined by the data, and (2) asymptotic theory applies, which can be used to construct appropriate confidence intervals. A bootsrap method to asses the statistical significance of the threshold effect is also available in order to test the null hyphothesis of a linear formulation against a threshold alternative (Chang et al. 2010). Hansen (1999) developed an econometric model that is suitable for threshold regression with panel data. The panel threshold model divides the observation into two or more regimes, depending on whether each observations is above or below the threshold level. The subscript i indexes the individual and t indexes time. The depen-dent variable, y t , and the threshold variable, q it , is scalar, and the regressor x it is a k vector. The structural equation of interest is Where I(.) is an indicator function.
The observed data will be divided into two regimes, depending on whether the threshold variable q it is smaller or larger than the threshold ã. The regimes are distinguished by different regression slope ² 1 and ² 2 . In order to be able to identify ² 1 and ² 2 , regressor variable x it and threshold variable q it are not time-invariant; u it is the fixed individual effect, and error e it is assumed to be independently and identically distributed (iid) with zero mean and finite variance.
Threshold level ã is estimated using least square method introduced by Hansen (2000). A bootstrap procedure is adapted to get approximate critical values of the test statistics which allows us to perform the hyphothesis test for the threshold effect. In this case, if the asymptotic p value is smaller than the desire critical value, we conclude that the null hyphothesis of no threshold is rejected. After a threshold value is found, the confidence intervals for the threshold value and slope coefficients are then estimated (Chang et al. 2010). The same procedure can also be applicated in the case of multiple thresholds. Related to this case, the potentiality of presence of more than one threshold represent another advantage of this method over the traditio-nal approach.
Our purpose of this study is to analyze the effect of financial liberalization on economic risks with trade openness as the threshold variable. This research includes some variables that are important in effecting economic risk in a country as control variables, including (1) volatility inflation, (2) discretionary fiscal policy, (3) institu-tional quality, and (4)  finlib is a financial liberalization policy which is proxied by two measures, namely (1) financial integration, and (2) financial deepening. These approaches have been adapted in several previous studies, including Kose et al. (2003), Kose et al. (2006), Bekaert, Harvey, & Lundblad (2006), Ahmed & Suardi (2009), and Nicolò & Juvenal (2012). Financial integration, which is used as a proxy for external financial liberalization in this study, was built using Lane & Milesi-Ferretti (2007) calculation methods. The measure of financial integration is the sum of international financial gross assets and international financial liabilities relative to GDP. Beside financial integration, this study also adapt financial deepening as a proxy for internal financial liberalization as measured by the value of M2 relative to GDP.
Trade openness, which is treated as threshold variable, is a measure of the sum total exports and imports relative to total GDP. This measure has been widely used by various literatures as a proxy of economic openness such as Kose et al. (2003), Neaime (2005), Dupasquier & Osakwe (2006), Ahmed & Suardi (2009), Pancaro (2010) and Balavac & Pugh (2016). To address the country specific effects, we also consider some control variables to be included in empirical model which are terms of trade volatility, inflation rate volatility, institutional volatility and descretio-nary fiscal policy. Specifically, the discretionary fiscal policy is constructed by using the method proposed by Fatas & Mihov (2003), Where G is the logarithm of real government spending and Y is the logarithm of real GDP. Deterministic time trends are used to capture the observed trends in govern-ment spending at all times. Discretionary fiscal policy's data is represented by the value of å t . To calculate this data, we use annual data for 20 Asia-Pacific countries from the period 1975-2015 and estimate the following regression for each country.
All data are collected at an annual frequency. The data used is collected through various sources, and it covers 20 countries in the Asia Pacific for period 1975 -2015.
Operationally, as we mentioned before, GDP Growth Volatility is used as a proxy for economic risk; meanwhile, the financial integration and financial deepening are used as a proxy for financial liberalization. These proxies are used because it represents the impact of financial liberalization on financial sector both externally (financial integration %capital inflows and outflows)

Table 1. Data and Data Source
and internally (financial deep-ening). Trade openness, which is defined as the sum of import and export per GDP, is used as a proxy for trade openness condition in each country and is also treated as the threshold variable. Other variables, including terms of trade volatility, inflation rate volatility, discretionary fiscal policy, dan institutional quality are used as control variables.

Tabel 2. Descriptive Statistics
Before discussing the main result of this study about the effect of financial liberaliza-tion on economic risk, the descriptive statistics of each variable is represented in the study for financial liberalization, the score range between 13.43 and 344.21. For the trade openness, the score lies between 0.13 and 4.39. The somewhat high score range for volatility infla-tion indicate that there is a high variation of inflation in some of the sample countries.

RESULTS
As we mentioned earlier, this study suspects that there is a non-linear impact of financial liberalization policy on economic risk in the Asia Pacific, depending on the degree of trade openness in each country. In order to be able to detect this condi-tion, in terms of whether an economy with a high level of trade openness can show a different trend compare to the economy with a relatively low level of trade openness, this study accommodate the posibility of nonlinear effect (or threshold effect) of trade openness related to the impact of financial liberalization on economic risks. As we discussed earlier, it is still unclear based on empirical results whether financial liberalization can give a positive impact by reducing the economic risk in a country (Kose et al. 2003).
Related to the above, to test our hyphothesis regarding the non-linear effect between financial liberalization and economic risk with trade openness as a threshold variable, we test for the existence of a threshold effect. This paper uses the bootstrap method to calculate the F statistics and the p value. The results are estimated with two different financial liberalization measures, which are financial integration (model 1) and financial deepening (model 2). To be noticed, the test statistic for a single thres-hold is significant for both financial integration and financial deepening, but the test for double and triple thresholds are insignificant. Based on this findings, we may conclude that there is a strong evidence that our hyphotesis stating that there is a non-linear effect of financial liberalization toward economic risks can be accepted. The threshold regression using Hansen (2000) method are shown in Table  3 below. Table 3 reports the results for models 1 and 2. Based on the threshold results, the bootstrap p value shows significant results for both models. These results confirm the existence of a threshold effect between Source: Data processed (2018) financial liberalization and economic risk. So that, rather linear, we can conclude that the effect is non-linear. The estimated threshold of trade openness is 143.8 percent for both financial integration (model 1) and given is calculated, the LR2 value at the estimated threshold value will always be equal to zero. The dashed line represents a critical value of 5%.

Table 3. Threshold Effect Estimation
After confirming the presence of nonlinear effects, we now try to divide the sample countries according to the level of In this study, financial liberalization is proxied by two variables which include (1) financial integration, and (2) financial deepening. Financial integration, which is measured by the sum of capital inflow and outflow divided by GDP, is used as a proxy its degree of trade openness threshold value which receive a score of 143.8 percent. Table  4. shows that the majority of countries, that are almost 80 percent from 1980 -2015, have a relatively lower trade openness value compare to the threshold level. In contrast, it is only 20 percent of the sample countries have a higher trade openness value compare to its threshold level. Based on this findings, we may say that only a small portion of our sample countries that are vulnerable to external shocks because of their trade openness condition higher than the threshold level.

Table 4. Percentage of Countries in Each Regime by Year
for financial liberalization in terms of external. Financial integration has been used as a proxy for financial liberalization in many researches including ), Neaime (2005), Bakaert et al. (2006, Ahmed et al. (2009) andMirdala et al. (2015). 1980 1985 1990 1995 2000 2005 2010 2015 Trade Moreover, the financial deepening is used as a proxy for financial liberalization in terms of internal according to Kose (2003), Neaime (2005) and Ito (2006). Table 4 reporting the regression results regarding the effect of financial integration toward economic risk of Asia Pacific Countries. The results show that financial liberalization as proxied by financial integration increases the economic risk when the countries liberalize their trade beyond the predetermined threshold (which receive a value of 1.438). This shows that the higher level of economic liberalization as indicated by the high financial liberalization and trade liberalization can simultaneously increase the risk of the economy by increasing the volatility of GDP growth. The similar result were also obtained when analyzing the effect of financial liberalization on economic risk using financial deepening as a proxy. Regarding to this concern, we found that financial deepening increase the GDP growth volatility when the countries liberalize their trade over the trade threshold. It means that an open domestic financiall market that is followed by high degree of trade opennes will tend to create an economic instability. It is because that the potential loss of foreign exchange reserves that are needed to finance the country's international trade are greater in the financial liberalization regime. According to Cardarelli, Elekdag and Kose (2009), capital inflows create important challenges for policymakers because excessive capital inflows, that can be triggered by financial liberalization policy, may give the potential excessive pressure, lose competitiveness due to the appreciation of the exchange rate and increase the vulnerability to the economic crisis. Related to this concern, Stiglitz (2002) revealed the negative effect of financial liberalization is that it is able to create an instability in the financial market if the economy is still not well developed. In the case of this research finding, the financial liberalization policy will tend to increase the country's economic risk that has a high degree of trade openness.  Table 4 and 5 also explain other factors that may affect the economic risk of Asia Pacific countries. It is found in the estimation result that terms of trade volati-lity show a positive but not significant effect toward economic risk. The estimation results indicate that the inflation rate volatility significantly effect the economic risk in a positive manner. It means that inflation rate volatility increase the economic risk. This finding is in line with Ahmed andSuardi (2009) andNeaime (2005). Adverse effect of inflation volatility toward economic risk is because increasing uncertainty in inflation, which is represented by increasing inflation rate volatility, can distort price mechanism effectivity in allocating resources, thus it may increase the economic risk.
Related to the above, it also found that the discretionary fiscal policy has a positive effect to the economic risk in both model. This is due to undisciplined fiscal policy can lead to output fluctuation giving pressure to the economic risk of a country through the decreasing governments credibility. In terms of the institutional quality, the result show that this variable can reduce economic risk of a country. It means that the existence of a good institutional quality can help a country to lower its economic risk. This risk reducing role played by institutional quality that is found in this study in line with Ahmad and Suadi (2009) which showed that good financial market insti-tution can help to reduce capital flight so that it will maintain the economic stability.

CONCLUSION
The purpose of this study is to analyze the effect of financial liberalization toward economic risks in a non-linear manner. In analyzing the non-linier effect of financial liberalization, we use trade openness as threshold variable. By using panel regression threshold introduced by Hansen (1999), we confirm that there is a non-linier effect of financial liberalization toward economic risk in Asia Pacific countries depending on the value of trade openness in each country. Regarding to this finding, when the trade openness is below the threshold value, financial liberalization policy can reduce the economic risk of Asia Pacific countries. However, when the trade openness exceeds the threshold value, the financial liberalization will increase the economic risk. So that, we conclude that an open domestic financial market that is followed by high degree of trade openness will tend to create an economic risk in the Asia Pacific countries.