Good Coups, Bad Coups: Evidence from THAILAND’S Financial MARKETS

This study investigates the short-run and long-run impact of coups on Thailand’s financial markets. Using daily data from the stock and foreign exchange markets during the period 2005–2017, the study shows (1) both coups in 2006 and in 2014 exert short-run impact on Thailand’s stock and foreign exchange markets; (2) however, the direction and magnitude of impact are different and opposite in the two coups; and (3) in the long run, the coups exhibit minimal impact on the currency market, but induce better market performance (positive return and decrease in the return volatility) despite an increase in liquidity risk of the stock market. Against common beliefs about negative consequences of the coup d’états, this study suggests that the uncertainty surrounding coups can bring good investment opportunities for investors to earn abnormal profits. Moreover, in the long term, the coup can drive the country to better stability and development. Sutsarun Lumjiak (Thailand), Nguyen Thi Thieu Quang (New Zealand), Christopher Gan (New Zealand), Sirimon Treepongkaruna (Australia) BUSINESS PERSPECTIVES LLC “СPС “Business Perspectives” Hryhorii Skovoroda lane, 10, Sumy, 40022, Ukraine www.businessperspectives.org Good Coups, Bad Coups: Evidence from THAILAND’S Financial MARKETS Received on: 11th of January, 2018 Accepted on: 16th of April, 2018


INTRODUCTION
Thailand is known as a country with high political instability in Asia among other countries such as the Philippines, Bangladesh, and North Korea. As of 2015, Thailand was ranked 163 rd out of 194 countries in the world in terms of political stability (TheGlobalEconomy.com, 2015). Within eight years since 2006, Thailand has seen two coups and the most recent coup in 2014 is marked as the 12 th coup in the country since 1932. The coups and associated political instability were blamed for the country's slow economic development (International Monetary Fund, 2015). However, Thailand's financial market has grown rapidly since the recovery from the 1997 Asian financial crisis. In 2016, Thailand was considered as the second best-performing stock market in Asia with more than 19% year-to-date return of Thai SET50 index (Chandran, 2017). In addition, the Thai Baht appreciated 7.8% against the US dollar to become Asia's best performing currency in 2017 (Reuters, 2017). It is, therefore, interesting to determine if there is any relationship between the coups and financial market quality. In other words, whether the coups affect financial market quality in Thailand.
This study extends the study of Lumiajiak, Treepongkaruna, Wee, and Brooks (2014) by taking into account the effect of the most recent 2014 coup in addition to the 2006 coup. As suggested by Duggan (2004), investors may behave differently to different coups due to their systematic differences. We also investigate the short-run and long-run impact of coups on Thailand's financial markets, specifically the Stock

LITERATURE REVIEW
Coup is a political event related to the use of military power and the change of government leaders, and is a source of political risk (Powell & Thyne, 2011). Previous studies have shown that political risk has great influence on stock market development (Perotti & van Oijen, 2001;Yartey, 2008). Particularly, the effect is stronger in emerging than developed markets due to more barriers, such as low transparency, less reliable data, and higher level of information asymmetry (Bilson, Brailsford, & Hooper, 2002;Brooks & Mosley, 2007;Erb, Harvey, & Viskanta, 1996). It has been argued that higher risk associated with political risk can bring abnormal stock returns (Amihud & Wohl, 2004 However, political risk also affects market volatility. For example, Chan and Wei (1996) showed that political shocks such as the news about Sino-British collaboration on Hong Kong affairs significantly affected the volatility of the stock mar-ket due to the impact of both market-wide and substitution effects. Białkowski, Gottschalk, and Wisniewski (2008) when studying the behavior of stock market during national elections in 27 OECD countries also showed that the stock market participants acted more aggressively when there were increasing election shocks caused by narrow margin of victory or changes in the political inclination of the government. We hypothesize the following relationships: H2: The coup brings more volatility to Thai stock market.
H3: The coup increases the liquidity risk of Thai stock market.
However, not many studies investigate how political risk affects the stock trading volume, except for a few studies, such as Chan, Chui, and Kwok (2001) and Leblang and Mukherjee (2005). Chan et al.'s (2001) study showed that political news negatively impacted the trading volume on the Hong Kong Stock Exchange due to investors' perceptual biases and the less quality of information from political news. On the other hand, Leblang and Mukherjee (2005) found that the political effect depended on the anticipation of the winning party. We hypothesize the following relationships: H4a: The coup has a negative effect on the liquidity of Thai stock market.
H4b: The coup has a positive effect on the liquidity of Thai stock market.
The effect of political risk is also prominent in the foreign exchange market. Bachman (1992) argued that political risk played an important role in explaining the significant changes in forward exchange bias. Blomberg and Hess (1997) also criticized the poor performance of standard exchange rate forecasting models and posited that this was due to the omission of political factors. Freeman, Hays, and Stix (2000) added that in democratic-politic countries, the currency traders were forced to continuously revise their expectations about election results and government survival. Therefore, exchange rate would be affected in a way that exhibited different market regimes. In addition, this effect depended on the degree of democracy and the transparency of the policymaking (Hays, Freeman, & Nesseth, 2003). The following hypothesis is offered: H5: The coup has a positive effect on abnormal return of Thai foreign exchange market.
In terms of market volatility, Lobo and Tufte (1998) found that the exchange rate volatility of the JPY, GBP, DEM and CAD against the USD was affected by political factors, such as electoral cycle and/ or the political party. Similar findings were also found for the Greek foreign exchange rate (Siokis & Kapopoulos, 2003). In Latin America, Cermeño, Grier, and Grier (2010) proved that the foreign exchange rate was more volatile in post-election period. Liu and Pauwels (2012) also showed that external political pressures from other countries such as the U.S., the EU and Japan caused the volatility of Renminbi. The following hypothesis is offered: H6: The coup has a positive effect on return volatility of Thai foreign exchange market.
Given the substantial effort in investigating the effect of political risk on the currency market, studies on liquidity effects have not received adequate attention. Lumiajiak

Data and key variables of interest
where ,, idt r is the d th five-minute return on instrument i during day t and D is the total number of all five-minute return intervals during a trading day. The ,, idt r on the SET Index is calculated from the price level of the SET Index at a five-minute interval. The ,, idt r for the USD/THB is computed from the mid-price of the USD/THB at a five-minute interval.
Each instrument volatility is measured using intra-day data following Andersen et al. (2003). Thus, the daily realized volatility it RV is calculated as follows: For liquidity measure, we use different proxy for the USD/THB and the SET Index, since bidask prices are not available for the SET Index. Specifically, we use the daily average of the bidask spread to measure the liquidity of USD/THB as follows: where , dt s is the d th five-minute bid-ask spread in the USD/THB during day t and D is the total number of all five-minute intervals during day . t A high bid-ask spread indicates the less liquidity of the foreign exchange market. Data on the bid-ask spread are not available for the SET Index, thus the natural logarithm of the daily trading volume is employed as our proxy for liquidity of the SET Index (LNV). The liquidity proxy for the SET Index exhibits opposite direction of the USD/THB such that the greater trading volume in the SET Index represents the better liquidity of the Thai stock market.
Finally, to measure variability of liquidity, we use the daily standard deviation of the liquidity. The liquidity volatility of the USD/THB is defined as follows: Similarly, the variability of liquidity on the SET Index is defined as follows: idt v is the natural logarithm of the d th five-minute trading volume in day ; t t s and t v are the daily average of the five-minute bid-ask spread in the USD/THB and the daily average of the natural logarithm of the trading volume at five-minute intervals in day , t respectively.

Method
We conduct both short-run and long-run analysis as follows.
where it R is the daily return for instrument i on day . t where it u is the normally distributed error term with zero mean and variance of one.
Our key independent variables include SR5_C1, SR10_C1, SR20_C1, SR40_C1 and SR60_C1 which capture the short-term effect of the coup d'états of 2006; SR5_C2, SR10_C2, SR20_C2, SR40_C2 and SR60_C2 to capture the shortterm effect of the coup d'états of 2014. These are dummy variables and set to 1 for the windows [0, 5], [6,10], [11,20], [21,40] and [41,60], respectively, with date 0 being the coup date (i.e. September 19, 2006 for first coup and May 22, 2014 for second coup). We also include the variable "Military_led" to capture the long-term effect of the coup. This is also a dummy variable equal 1 for the period the military takes control, and 0 otherwise. In addition, the financial market is also affected by the trading activities of foreign investors. To capture this effect, we include "FC_Buy" variable, which is the percentage of net purchases in Thai Baht on the Thai stock market by foreign investors. This variable is also interacted with coup-related variables to examine the modera-tion effect of the foreign investors' net purchase on the relationship between the coup and the financial markets.
The control variables include lags of dependent variables to capture the AR structure in return, return volatility, liquidity and liquidity volatility for both stock and currency markets. The choice of lag lengths to be included closely follows the study of Lumiajiak et al. (2014). Acknowledging that the performance of the financial market of a country can be affected by the financial market in other countries, we include daily return on S&P 500 (S&P) to control for the world stock market cycle. In addition, since the study covers the global financial crisis, a dummy variable "GFC" is included to capture this effect. GFC is set to 1 for the period from July 26, 2007 to July 30, 2010, and 0 otherwise. For the stock market, we also include "Quote" to control for the trading volume and "AvgTBT" to control for the frequency of trading activities. For the currency market, we include "Quoter" to control for the competition in the market.
Definition of the regression variables (equation 7) and their measurements are provided in the Appendix.

Summary statistics
Together with changes in the government, the Thai stock market and foreign exchange market exhibit different movement patterns. The foreign exchange market shows the most volatility during the Independent led period, which might be attributed to the high uncertainty of the new election result (see Figure 1). Similar trend is also found for the Thai stock market. However, the market fluctuates greatly during Shin led period, particularly during the protests (see Figure 2). The returns of both the stock and foreign exchange markets are also the highest during these uncertain periods.
The ratio of net stock market purchase by foreign investors (FC_Buy) experiences an increase trend before June 2007 and then decreases sharply during the second half of the Independent led and Shin led periods (Sundaravej as PM). After a relatively stable period during the Democracy and Shin led government (Yingluck as PM), the foreign stock purchase reduces before and after the coup in 2014, but rebounced from the beginning of 2015 (see Figure 3). The summary statistics in Table 2 further shows that when the Military led government is in power, the USD/THB return is less volatile, has higher liquidity and lower liquidity risk. However, for the Thai stock market, although the SET return is less volatile, its liquidity risk during this period is higher than that in non-Military led period. In both periods, there was no difference in the return of stock and foreign exchange markets.  Notes: Table 2 reports the summary statistics of the variables used in the regression analysis for the USD/THB (Panel A) and SET Index (Panel B). Samples are partitioned into two sub-periods based on whether the government is led by military or civilian. T-statistics for mean differences between two sub-samples are displayed in two last columns of the This "against-expectation-effect" can be justified by the uncertain information hypothesis suggested by Brown, Harlow, and Tinic (1988). Accordingly, investors react to bad news more than to good news. Since the Thai Baht depreciated and the stock index declined during the political turmoil in 2013 (Lee, 2013), there would be upward adjustments after the coup. Therefore, the coup was considered as necessary for the longterm competitiveness and stability of Thai's economy (Schmidt, 2007).       in the first 5 days after the coup. The less impact on the currency market during 2014 coup can be explained by its expectation beforehand, when the martial law was declared 2 days before. Table 4 show the differences in the average daily abnormal return, volatility, liquidity and liquidity risk of the SET Index between pre and post windows for the 2006 and 2014 military coup d'états, respectively. Interestingly, we find no difference on the Thai stock market performance and quality before and after the 2006 coup. The little damage in the stock market in the 2006 coup was partly because the country's stock market was already lagged for several years and the coup took place peacefully without bloodshed (Cheng, 2006 (Shaffer, 2014). The coup helped to put an end on that political uncertainty which blocked the economic development for a long term. The approvals of many billion-dollar-worthinvestment projects right after the coup raised the investors' confidence and encouraged them to trade more. We can conclude that the coup and preceding protests, on the one hand, created a risky investment environment which was highly volatile, but on the other hand, brought investment opportunities for investors to earn higher returns favored by investors.   , we find investors react differently in the currency and stock markets both in the short run and long run. We find the stock market reacts stronger than the currency market. Table 5 shows when the net flow of trading by foreigners in the Thai stock market (FC_Buy) increases, the Thai Baht appreciates by 0.003 point against the US dollar, the currency market exhibits more volatility (coefficient of FC_Buy for volatility regression is 0.0112), less liquid (bid-ask spread increases by 0.011) and more liquidity risk (liquidity increases by 0.00974). These findings exhibit foreigners need to convert their currency in-to the Thai baht before investing in the Thai stock markets. This effect is supported by Gyntelberg, Loretan, Subhanij, and Chan (2009) study, which suggests that foreign investors' net purchases of domestic equities would create an appreciation of its currency. Lumiajiak et al. (2014) also found similar effect when studying the effect of the 2006 coup on the financial markets.

Short-run and long-run market reactions
Based on the short-run effect on the currency market (see Table 5 However, these effects reversed when the foreign money flows in during these short-term windows (interaction variables between SR5_C1 to SR60_ C1, and SR5_C2 to SR60_C2 with FC_Buy). These results indicate that the short-term effect of coup d'état on Thai currency market is largely affected by the net stock purchase of foreign investors. During the days following the coups, the higher purchase on the stock market by foreign investors will reduce the volatility of both market return and liquidity. This effect is against the expectation of more market volatility as hypothesized (H6 and H8). In terms of the market return and liquidity, the short-term effect of coups is less clear. The coefficients of interaction terms between short-term windows and net foreign investors' stock purchase are contrary in the two coups.
For the long-run effect of the military intervention (see Table 5), we find minimal impact in the currency market such that the currency market becomes slightly more volatile during the Militaryled government (support of H6). The coefficient of "Military_led" government is positive and significant in the regression of volatility. This finding is consistent with the financial behaviour theory that investors tend to overact or underact to the news in the short run, but in the long run, when they realize their errors, a return reversal will be detected (Daniel, Hirshleifer, & Subrahmanyam, 1998). Furthermore, the coefficients on the interaction term between "Military_led" and ratio of foreign buy in the stock market (FC_Buy) indicate that foreign investors place more confidence in the military-led government as the currency market is less volatile during the military-led government. The higher purchase by foreign investors during this time also increases the liquidity of the currency market and thus, supports the positive expectation about the effect of coup on the currency market liquidity (H7). It was reported that after the 2014 coup, the Thailand's Board of Investment was appointed on June 8 and approved investment incentives for 603 projects worth more than 400 billion Thai Baht. Budget spending also increased, together with massive investment plans in infrastructure (Saiyasombut, 2014). Therefore, there is a strong belief that the coup promotes Thailand's stability and economy competency in the future.
Turning to the short-term effect on the Thai stock market (see Table 6), we find investors react differently to the two military coup d'états. The 2006 military coup d'états results in immediate increase in return (the coefficient of SR5_C1 is positive and significant), drop in volatility, but lower liquidity (negative coefficients of SR5_C1 to SR60_C1) and higher liquidity risk (positive coefficients of SR5_ C1 to SR60_C1). It also appears that foreign investors do not time market well, as they earn negative return 5 to 10 days after the 2006 military coup d'états. These coefficients are -0.079 and -0.062, respectively. On contrary, market drops slightly 5 and 60 days after the 2014 military coup d'états.
Coefficients of SR5_C2 and SR60_C2 in the regression of stock return are -0.0439 and -0.0179, respectively. However, the foreign investors earn positive return during these two short-term windows (positive and significant of variables SR5_ C2*FC_Buy and SR6-_C2*FC_Buy).
The results from the short-term effects of the coup provide inconclusive result about its effect on the stock market return given the different market reactions in the two coups. However, there are supportive evidences about the positive effect of coup on liquidity risk (H3) and the negative effect on liquidity of the Thai stock market (H4a). In addition, against the negative expectation about the coup's effect on market volatility, in the short run, it shows that the coup reduces the market risk (rejection of H2). Nevertheless, these effects are lessened by the increase in foreign investors' net stock purchase.
For the long-run effect in the stock market (see Table 6), when the net flow of trading by foreigners in the Thai stock market (FC_Buy) increases, we find higher return and lower risk. The coefficient of FC_Buy in Return regression is 0.0317 and significant at 0.01 level. In the volatility regression, this coefficient is -0.000597 and also significant at 0.01 level. Similarly, higher return and lower risk is detected during the Military-led govern-ment (coefficients of Military_led in Return and Volatility regressions are 0.0046 and -0.000153, respectively) with the exception for an increase in liquidity risk (coefficient = 0.0351).
While the result is inconclusive for short-term effect of the coup on the stock market return, the long results show that the coup has a positive effect on the abnormal return of Thai stock market (support of H1). The results also provide support for the positive effect of the coup on the Thai stock market's liquidity risk (H3). However, similar to short-term results, the hypothesis on the positive relationship between coup and market volatility (H2) is rejected. These effects are also lessened by the increase in foreign investors' net stock purchase. The coefficients on the interaction term between the Military-led and ratio of foreign buy in the stock market (FC_Buy) indicate that foreigners do not time market well during the Militaryled government. It appears foreign investors enter the market when liquidity risk is relatively For the effect of coup on the liquidity of Thai stock market, we do not have enough evidence to support either hypothesis H4a or H4b and conclude that there is no effect of coup on the stock market liquidity in the long run.

CONCLUSION
This study investigates the effects of coups on Thai financial markets. Using daily data from the stock and foreign exchange markets during the period 2005-2017, the study shows the coups exert both shortrun and long-run impact on the stock and foreign exchange markets. However, the effect is different between the coups and markets.
In the short run, the 2006 coup induces Thai Bath to depreciate, more volatile, more liquid with higher liquidity risk. However, the 2014 coup strengthens the Thai Bath, reduces its liquidity but increases the liquidity risk. The effect lasts longer for the 2006 coup compared to the 2014 coup. For the stock market, both coups reduce market volatility and liquidity, but increase the liquidity risk. The 2006 coup increases stock returns, while the 2014 coup leads stock returns to decrease. The persistence effects is high for both coups, but more for the 2006 coup.
In the long run, both coups show their positive impact on the stock market by increasing the SET index return and reducing its risk, despite affects the USD/THB negatively by making it more volatile. The study also shows that the effects of the coups on Thai financial markets are largely dependent on the net stock purchase of foreign investors. The moderation effect takes place in the way that it reverses the "pure" effect of the coup (when there is no purchase of foreign investors).
Against common beliefs about negative consequences of the coup d'états, this study suggests that the uncertainty surrounding the coups can bring good investment opportunities for investors to earn abnormal profits. Moreover, in the long term, the coups can drive the country to better stability and development.