Were Foreign Exchange Markets Reacting Negatively to Political Events? The Case of Malaysia

Abstract

This article explores the effects of political events on foreign exchange returns in Malaysia. We identify five political events in recent history, namely the 13th General Election (GE13), the imprisonment of a key opposition politician, the scandal from the 1MDB exposé, the appointment of a new Central Bank Governor and the 14th General Election (GE14). Using event studies, our findings show that the imprisonment of the opposition party leader triggered a favourable response from the foreign exchange market. However, market reactions to the 1MDB scandal were largely unfavourable. The GE13 triggered unfavourable market response, while the reverse is true for market reactions to GE14. Market response to the appointment of the new Central Bank Governor was rather positive. The Event Study is the first of its kind that examines the foreign exchange market implications of key political events in Malaysia. There are practical considerations that emanate from these findings.

JEL Classification: F31, D72, D73, O38

Keywords

1MDB general election market model exchange rate confidence Malaysia

Introduction

The effects of political events on financial markets have been well documented in the literature. One strand of the literature focuses on how political risks/events influence stock market returns and volatility (see, for instance, Bialkowski et al., 2008; Bowes, 2018; Chesney et al., 2011; Chia, 2018; Hou & Li, 2019; Kollias et al., 2011; Lean, 2010; Liew & Rowland, 2016; Nazir et al., 2014; Wong & Hooy, 2016; Yusoff et al., 2015). Meanwhile, a second strand in the literature examines the effects of political events on the foreign exchange markets (Lobo & Tufte, 1998, Mpofu & Peters, 2017). While the literature can be seen as related to studies on foreign exchange market efficiency (Ahmad et al., 2012, Khuntia et al., 2018) and rational expectations (Echavarria & Villamizar-Villegas 2016), it, nonetheless, serves a broader and more general purpose of understanding how investors react to changes in political environments.

In this article, we examine how key political events affect the foreign exchange market in Malaysia. We are interested to explore the implications of five rather recent events that induced political uncertainty on investors. These events cover the 13th General Election in 2013 (GE13), the imprisonment of Anwar Ibrahim (the erstwhile leader of the opposition coalition) in 2014, the 1MDB exposé in 2015, the appointment of Muhammad Ibrahim as the eighth Governor of Bank Negara Malaysia (Central Bank of Malaysia) and the 14th General Election in 2018 (GE14).

There are compelling reasons for motivating this research, besides the fact that no other author has directly addressed this topic at the time of writing. The country was affected by the Asian financial crisis that triggered volatile exchange rate fluctuations. The crisis was touted by many observers to be a result of fundamental macroeconomic weaknesses partly arising from political–economic factors like corruption and crony capitalism (IMF, 1997). In this regard, the rise of political instability had undermined the confidence of financial markets, while refusal to carry out reforms made matters worse (Corsetti et al., 1999). It has been observed by Haggard (2000) that countries which carried out much-needed institutional reforms tended to fare better. The unhealthy nexus between politics and business had been a major factor behind the weakening of fundamentals and the unravelling of the Malaysian economy (Perkins & Woo, 2000; Rasiah, 2001). Nonetheless, reforms were often hampered by substantial obstacles in many other countries, so little has changed since the crisis (Sen & Tyce, 2017). Hence, the politics–asset market nexus remains highly relevant.

Moreover, from a political–economic point of view, these events are watershed moments in history. Since gaining independence from the British Empire in 1957, the country was ruled by the Barisan Nasional (BN) coalition with no disruption in continuity for over 60 years. The absence of change in political leadership suggested that Malaysia was a pseudo-democracy (Chin, 2015). In the 12th General Election in 2008, BN suffered one of the worst election outcomes in history. The opposition coalition improved their performance in the 2013 General Election. More obstacles followed, thereafter, when the erstwhile opposition leader Anwar Ibrahim was sentenced to a 5-year prison term.¹

Not long after the GE13 concluded, investigative journalists produced reports on financial mismanagement concerning 1MDB, a sovereign wealth fund founded in 2009 (Gunasegaram, 2018). Claims arose about 1MDB being a Ponzi scheme and also a conduit for the vast network of global money laundering trail, with allegations surfacing in July 2015 that the fund’s cash flow was channelled into the personal bank account of the sixth prime minister (Brown, 2018). The local anti-corruption agency had managed to open up investigations, but political interference compromised the integrity of the investigations.² Meanwhile, there was a change of leadership in the Central Bank of Malaysia, with the appointment of the eighth Governor Muhammad Ibrahim.

While all these events were occurring, the ringgit/US dollar exchange rates displayed a large degree of volatility with some commentators linking exchange volatilities to political factors (The Economist, 2016). No research papers have confirmed nor rejected these claims about the nature of the exchange rate movements. Not long after this, BN was defeated in the GE14 in May 2018 and subsequently lost its stranglehold on political control for the first time in 60 years. Curiously, there seemed to be a rally in the currency markets as ringgit strengthened, perhaps reflecting signs of confidence among investors (see Figure 1).

Figure 1.

Source: Bank Negara Malaysia.

While there seems to be a link between key political events and news, on the one hand, and exchange rate fluctuations, on the other hand, such links are as yet unconfirmed due to the absence of research on this topic. To the extent that fluctuations in currency values are an indication of investor beliefs, these fluctuations also present a challenge to investors and speculators who constantly work out hedging strategies. Our study remedies this gap in the literature.

We deploy event studies as a tool to explore the impact of political events on foreign exchange markets. We use the spot ringgit/USD as a proxy for the foreign exchange market. In this framework, we first estimate a market model of the exchange rate returns, drawing from and improving upon the model adopted by Frankel (1981) and Adam et al. (2013). The data for this estimation cover daily observations of the spot ringgit/US dollar exchange rate time series from 13 February 2012 to 2 April 2013—this is basically our 250-day estimation window. The events being analysed encompass the GE13 in 2013, the 1MDB exposé in 2015 and the GE14 in 2018. For each event, we set 20 pre-event days, one event day and 60 post-event days, giving a total of 81 days in one event window. Next, the abnormal returns (ARs) and cumulative abnormal returns (CARs) are calculated for each of the event. Parametric statistical tests are then applied on the ARs and CARs to determine their statistical significance.

An overview of our findings is as follows. First, CARs from the scandal arising from the 1MDB exposé are significantly positive. This indicates that the ringgit had depreciated more than expectations, signifying an unfavourable market response. Second, the responses to the results of the general elections are not uniform. Particularly, subsequent CARs in reaction to the GE13 were mostly positive, again implying a larger than expected depreciation of the ringgit and an unfavourable market reaction. In contrast, the CARs for the GE14 were mostly negative, implying a favourable market response since the ringgit depreciated less than expected. Meanwhile, CARs in reaction to Anwar Ibrahim’s imprisonment had been negative, generally an unfavourable market reaction. Finally, the subsequent CARs in response to the appointment of a new Central Bank Governor was negative towards the end of the event window, implying a favourable market reaction.

The article is organised as follows. The second section presents a brief view of the extant literature on political events and reactions from the asset markets, and also the event study methodology. The third section discusses the data and methodology for our study. After that the fourth section reports the main results and findings, before suggesting practical and policy implications. The fifth and final section concludes the article.

Literature Survey: A Brief Overview

Political Events and Financial Markets: What Are the Theories, Scope of Study and Key Findings?

This article makes a contribution to the wider literature concerning political events and their implications on mean returns and volatilities in financial markets. Some of the previous studies that fall under this area include Lobo and Tufte (1998), Bialkowski et al. (2008), Lean (2010), Kollias et al. (2011), Chesney et al. (2011), Nazir et al. (2014), Yusoff et al. (2015), Bin (2015), Liew and Rowland (2016), Wong and Hooy (2016), Mpofu & Peters (2017), Bowes (2018), Chia (2018) and Hou and Li (2019). The financial markets that are frequently covered encompass the stock and foreign exchange markets, with the exception of Chesney et al. (2011) who considered commodity and bond markets as well. The literature is also a mix of country case studies and cross-country analyses. The definitions of political events are wide ranging, but most aforementioned studies cited would cover general elections, political news and events concerning power struggles in the government. However, existing studies do not focus much on transmission mechanisms—how news of some events get propagated. There is also a heavy concentration of studies on the stock market, while other assets are not so well covered.

Some interesting cases of other political events are anti-corruption campaigns (Hou & Li, 2019), terrorist attacks (Chesney et al., 2011, Kollias et al., 2011) and the style of governance in terms of the more autocratic versus more democratic type of rule (Nazir et al., 2014). Such events, when they pose uncertainty to investors, could lead to significant negative outcomes, in terms of both smaller mean returns and larger volatilities. This is largely a reflection of the desires of risk-averse investors who reject situations where risks escalate. This is the first theoretical explanation for observed asset market reactions.

As a proxy for investor sentiments, market reactions could signal the degree of approval with regard to a particular political event (Colon-De-Armas et al., 2017)—this is the second theoretical explanation. Following this point, negative reactions could imply disapproval. A contrasting theoretical argument, and also the third theoretical explanation, is that investors may display too much optimism initially but eventually make corrections in their expectations when the newly elected regime performs below expectations (Booth & Booth, 2003).

Among the studies that cover the effects of general elections and political news, political uncertainty was found to have significant negative effects, not just on asset returns (Lean, 2010; Liew & Rowland, 2016) but also on volatility (Bialkowski et al., 2008; Bowes, 2018; Lobo & Tufte 1998). There are exceptions such as the study by Wong and Hooy (2016) who in contrast found that banking stocks had larger CARs in the election period, a finding similar to Chia (2018) who observed that different sectors of industry respond differently to general elections. Moreover, the study also suggested that some stocks perform better than others. A similar observation can be found in Yusoff et al. (2015), who reported that stock returns in politically connected firms tend to underreact in response to negative news. It would appear that the degree of uncertainty induced by political events matter. The perception that things are under control is important, as demonstrated in Nazir et al. (2014), where a more autocratic style of leadership leads to less uncertainty. Anti-corruption drives, on the other hand, may not necessarily lead to positive responses by the market, as Hou and Li (2019) discovered. Particularly, in the context of China, the decline in the stock market in response to the political purges of the Communist Party is interpreted by the authors as the ‘price’ that one pays for corruption. Studies that consider the effects of terrorist attacks (e.g., Chesney et al., 2011) report negative initial responses from the stock market, but the recovery thereafter is considerably less uniform between different markets.

Methodologies Adopted by Previous Researchers

Within the corpus of the literature mentioned earlier, there are currently three widely used methodologies favoured by researchers in this field. The first concerns the specification and estimation of time series models, with rate of return on a financial asset as the dependent variable and proxies of political risks as a regressor. Studies by Lean (2010), Bin (2015), Liew and Rowlands (2016) and Bowes (2018) are some noteworthy examples. Meanwhile, the second methodology relates to event studies (Bialkowski et al., 2008; Hou & Li, 2019; Mpofu & Peters, 2017; Nazir et al., 2014; Yusoff et al., 2015). Other researchers, such as Chesney et al. (2011), Kollias et al. (2011) and Chia (2018), combine both the methodologies in a single study. In empirical finance, the deployment of event studies has been immensely popular as a research tool to quantify the magnitude and type of response that financial securities have in relation to shocks (Kliger & Gurevich, 2014).

Studies coming under the first methodology are characterized by the specification and estimation of some econometric framework over a specific period of time, which includes the political events whose effects are being examined. To capture the effects of such events on returns and volatility of returns, dummy variables are regularly deployed usually in either a linear regression or a Generalized Autoregressive Conditional Heteroskedasticity (GARCH) family type of model. Needless to say, the level of sophistication in modelling varies from one author to the other. Meanwhile, papers adopting the second methodology tend to apply the econometric models of the first methodology to estimate benchmarks for normal asset returns or volatility over a sample period prior to the event whose effects are being investigated. Parameter estimates of the empirical models are then used to ‘extrapolate’ normal asset returns or volatility to a future period in which political events are contained; actual asset returns or volatility are compared with the extrapolated values to arrive at ARs/volatility. Summing up the ARs/volatility produces CARs/volatility. The main difference between the first and second methodologies is the estimation window; the first methodology has estimation windows covering the events of interest, whereas, in the second methodology, the event window is separate from the estimation window.

Data and Methodology

Data

The present study attempts to assess the impact of selected political events on the foreign exchange market, which we proxy with the spot ringgit/US dollar (RM/USD). Before this goal can be attained, a market model of exchange rate behaviour must first be estimated, as per the event study literature in the previous section. The independent variables of the market model are interest rate differentials (between domestic and foreign interest rates) and the spot exchange rates of US dollar/Euro (USD/EUR), yen/US dollar (Y/USD) and Chinese yuan/US dollar (RMB/USD). More descriptions of the market model are found in the next section. In addition to this, we also compiled daily interest rate data for Malaysia (proxied by overnight interbank rates) and foreign interest rates (proxied by US 3-month Treasury bill rates). Daily data are appropriate because it leads to a higher degree of accuracy (Brown & Warner 1985). Our rationale for this set-up will be justified in the next section on methodology. Notably, the RMB/USD exchange rate data and Malaysia’s interest rate data are obtained from the Monthly Statistical Bulletin of Bank Negara Malaysia (Central Bank of Malaysia). Data for US interest rates come from the Federal Reserve Economic Data (FRED). The USD/EUR data come from the European Central Bank (ECB). Data for the SGD/USD and Y/USD exchange rates are from Monetary Authority of Singapore (MAS) and Bank of Japan (BOJ), respectively. The market model is estimated using daily data covering the period from 13 February 2012 to 2 April 2013.

Meanwhile, the dates of political events covered in this article have to be spelt out clearly. As suggested, we consider five political events that are monumental, namely the two general elections (GE13 and GE14, respectively), the leakage of the 1MDB scandal by the media, the appointment of Muhammad Ibrahim as the eighth Central Bank Governor and the imprisonment of erstwhile opposition leader Anwar Ibrahim. The polling day of GE13 was 5 May 2013, whereas the polling day of GE14 commenced on 9 May 2018. There is no controversy regarding the dating of these events. In contrast, there will certainly be more disputes about how to date the progression of the 1MDB scandal, given that negative news reporting had begun since March 2013 (Gunasegaram, 2018). We suggest 2 July 2015 to be the date of interest in this regard. There may have been numerous news leaks during the 2013–2017 period in the run-up to the GE14 election, but 2 July is, indeed, the most crucial date because this was when two news portals, namely Sarawak Report and the Wall Street Journal, connected the financial returns of 1MDB to the personal bank account of Najib Razak, who was the Prime Minister of Malaysia and head of the BN coalition that was running the government in the country (Gunasegaram, 2018). Needless to say, this was the mother of all leaks in the context of the 1MDB scandal as it triggered major investigations that allegedly led to the downfall of Najib Razak and the BN coalition. The announcement of the appointment of Muhammad Ibrahim as the new Central Bank Governor took place on 27 April 2016, while the news of Anwar Ibrahim’s imprisonment was released on 7 March 2014.

Methodology

In an event study, there are two important procedures, the first being the identification of the event window and its size. MacKinlay (1997) recommends the use of a 41-day event window, which covers 20 pre-event days, the event day and 20 post-event days. However, as the foreign exchange market may be inefficient, it would be prudent to use 60 post-event days instead (Wong & Hooy, 2016).

The second procedure is to set the estimation window, over which a market model will be estimated, to calculate ARs and CARs of returns on an exchange rate arising from a particular event. MacKinlay (1997) and Wong and Hooy (2016) both adopted a 250-day estimation window, placing the estimation window just before the event window. In contrast, Mpofu and Peters (2017) used a 100-day estimation window, while Adam et al. (2013) deployed an arbitrary window size for an estimation spanning 4 months. We adopt a 250-day estimation window in this study in line with the literature, with the estimation window stretching from 13 February 2012 to 2 April 2013.

As highlighted in the literature review section, the main shortcoming of event studies on exchange rate returns is the lack of consensus on what constitutes a typical market model. This problem is less pronounced in event studies on stock market returns, where modern finance theory presents a wide array of asset pricing models. In this regard, data limitations and the nature of the present study as a time series analysis necessitate the adaptation of the news model used by Frenkel (1981) (as cited by Mpofu & Peters, 2017) and Adam et al. (2013). In Frenkel (1981), the spot exchange rate was regressed on lagged forward rate and expected interest rate differential. Adapting this news model, Mpofu and Peters (2017) substituted the lagged spot exchange rate for the lagged forward rate. The market model of Adam et al. (2013) regressed the EUR/PLN exchange rate on EUR/USD (proxy for global factor), EUR/CZK and EUR/ HUF (proxies of regional factors). We initially experimented with each of these model types individually and found that the fit of the model was quite poor. Hence, we decided to combines both the elements of Frenkel (1981) and Adam et al. (2013) in Equation (1), which is expressed as:

R_{t} = f [((i - i *) - E (i i *)), r (U S D / E U R), r (R M B / U S D), r (Y / U S D)]

(1)

where R_t represents exchange rate returns at time t, modelled as a function of the variables in the parenthesis. In this article, the baseline market model uses the RMB/USD spot exchange rate returns on returns on USD/EUR (r(USD/EUR) and returns on Y/USD (r(Y/USD) (both are global factor proxies),³ returns on RMB/USD (r(RMB/USD) (a regional proxy) and expected interest rate differential ((i − i*) − E(i − i*)).⁴ To better fit the high-frequency data, the autoregressive distributed lag (ARDL) model structure is adopted here. The other two market models with returns on RMB/EUR and returns on RMB/GBP, respectively, follow a similar model set-up as the baseline market model.

In the event study methodology, the residuals from the market model regressions, that is, ε_t is also defined as ARs. When the ARs are aggregated across time, we obtain the CARs. Typically, the statistical significance of ARs and CARs need to be tested via t-tests. The null hypotheses are specified as follows:

H_{0} : A R_{t} = 0

(2)

H_{0} : C A R_{t} = 0

(3)

Meanwhile, the t-statistics are calculated as the ratio of ARs or CARs over the standard deviation:

t (A R) = \frac{A R_{t}}{σ (A R)}

(4)

t (C A R) = \frac{C A R_{t}}{\sqrt{T_{1}} σ (A R)}

(5)

In the t-statistic for ARs (Equation [4]), the denominator is the standard deviation of the regression residuals of the market model. In Equation (5), T₁ represents the size of the event window.

Results and Findings

We regressed the market model specified in the previous section, using an ARDL framework. All the variables are expressed in logarithmic first difference, as this is the convention for expressing total asset returns. Prior to running the regression, preliminary unit root tests were applied to each series to confirm that they are, indeed, stationary. Given the large number of estimates and to conserve space, the unit root tests, summary statistics and estimated models are not reported but are available upon request. We only briefly mention, here, that the estimated models are well specified,⁵ passing the standard diagnostics tests like the Lagrange Multiplier and Ljung–Box test of the existence of serial correlation in residuals (at various lags), the Breusch–Pagan–Godfrey and the Cumulative Sum (CUSUM) and Cumulative Sum of Squares (CUSUMSQ) tests for structural stability, while also demonstrating that there is no strong evidence of heteroskedasticity and ARCH effects.

Using the estimated parameters of the model, we proceed to calculate the ARs and CARs for the three events mentioned in the earlier part of the article. In the interest of preserving space, we show only the charts depicting the CARs (Figures 2–6). In an 81-day event window, Figure 2 shows that the CARs were initially negative in response to the outcome of GE13. This implies that the ringgit had been depreciating at a rate that was less than that predicted by the market model—a favourable response by the market. Almost 2 weeks after GE13, however, the CARs became positive, signifying that ringgit had depreciated more quickly than expected. The observed reactions are consistent with the third theoretical explanation for exchange rate behaviour in relation to political events mentioned in the literature review. Particularly, the newly elected government was able to emerge to move the country forward, indicating approval by the public, but the lack of subsequent political reforms dampened the enthusiasm for the new government.⁶

Figure 2.

Source: The author.

Paradoxically, the imprisonment of popular politician Anwar Ibrahim induced negative CARs throughout the event window in the returns of RMB/USD (Figure 3). Negative CARs (as explained previously) means that ringgit was depreciating much less than one would expect, and it does imply a rather positive development. This is in contrast to common-sense explanations, as the persecution of a key opposition politician would have been detrimental to political reforms. We argue that this is due to the fact that the public felt relieved that the persecution of the key opposition politician had finally concluded.

Figure 3.

Source: The author.

The CARs from the exchange rate returns were clearly positive in response to the leakage of news regarding monies from the 1MDB sovereign wealth fund being found in the private account of the sixth prime minister of Malaysia (Figure 4). Market reactions were far more adverse than the predictions of the market model. Drawing upon the first theoretical argument, it would seem that the news leakage induced ever greater political uncertainty.

Figure 4.

Source: The author.

A very different picture emerges in Figure 5. The market reacted quite negatively to news of a change in leadership in the Central Bank at the initial period of the post-event days. However, sentiments were reversed later in the post-event period. This observation is in line with the third theoretical explanation of exchange rate behaviour, namely that investors were cautious initially but corrected for their initial pessimism at a later stage. This would make a lot of sense, given that the appointment of Muhammad Ibrahim as the eighth Central Bank Governor was seen to herald a continuity of the central banking style consolidated by his predecessor, Dr Zeti Akhtar Aziz.

Figure 5.

Source: The author

Figure 6 projected a rather favourable image for the newly elected government in the aftermath of GE14. Apparently, this also fits closely with the second theoretical explanation for exchange rate responses, namely that the market is in approval of the newly formed government. Given that this was the first change in political regime in approximately 60 years, there was a chance for much-needed reforms to be pushed through.

Figure 6.

Source: The author.

The patterns of the CARs seem to be consistent with the movements of portfolio capital flows⁷ (Table 1). The GE13 took place in May 2013 (mid-2Q 2013) and had a larger-than-expected depreciation of the ringgit 60 days later (i.e., July 2013 or early-3Q 2013).⁸ This is matched by a negative capital outflow position in third quarter of 2013. Similarly, the announcement of the prison sentence of Anwar Ibrahim occurred in March 2014 (end-1Q 2014) resulting in a larger-than-expected ringgit appreciation 60 days later in May 2014 (or mid-2Q 2014)—we observe capital inflow also taking place in second quarter of 2014. Meanwhile, the negative publicity surrounding the 1MDB exposé in early-July 2015 (early-3Q 2015) led to a larger-than-expected ringgit depreciation 60 days later in September 2015 (late 3Q 2015)—in the same quarter, massive capital outflows were observed. The appointment of a new Central Bank Governor in end-April 2016 (early-2Q 2016) saw a larger-than-expected appreciation in ringgit and a matching observation of capital outflows 60 days later (end-2Q 2016). Finally, the GE14 commenced in May 2018 (mid-2Q 2018) and resulted in a larger-than-expected appreciation of the ringgit 60 days later in July 2018 or early-3Q 2018 during which capital inflows were also occurring.

Table 1.

Capital Flows After Key Political Events

	Date of Commencement	Portfolio Capital Flows (after 60-day post-event window)
Event 1: GE13	5 May 2013/mid-2Q 2013	3Q 2013: RM 9,662 million
Event 2: Imprisonment of Anwar Ibrahim	7 March 2014/end-1Q 2014	2Q 2014: RM 6,321 million
Event 3: 1MDB exposé	2 July 2015/early-3Q 2015	3Q 2015: RM 24,431 million
Event 4: appointment of new Central Bank Governor	27 April 2016/early-2Q 2016	2Q 2016: RM 70 million
Event 5: GE14	9 May 2018/mid-2Q 2018	3Q 2018: RM 864 million

Source: Author’s calculation based on Bank Negara Malaysia.

Of concern, the shocks to exchange rates and capital flows may affect the real economy. To ensure that only the short- and medium-run fluctuations were being captured, we extracted the cyclical components of the time series using the techniques discussed in Sayan (2006).⁹ Next, the time series are subject to a correlation analysis. We find that the correlations between real GDP and capital flows are statistically insignificant across three leads and three lags. Thus, the results are not reported here.

The other highlight of this article is the testing of statistical significance of the ARs and CARs, the results of which are reported in Tables A1—A5. We are unable to utilize the standard critical t-values due to the presence of non-normality in the regression residuals. To overcome this problem, we calculate bootstrap critical values in a manner similar to Bialkowski et al. (2008). With these new critical values, we are able to proceed with the tests of statistical significance. It can be seen that a number of ARs and CARs are statistically significant at the 5% level, particularly on the post-event days. A number of the statistically significant t-values appear towards the end of the event window for all the five events considered in this article.

Robustness Checks

We have conducted some additional robustness checks on our results.

First, we have used other proxies of foreign exchange returns as the dependent variable in the market model, namely the RM/EUR and RM/GBP exchange rates.¹⁰ The pattern of the CARs turned out to be very similar to those depicted in Figures 2–6. Moreover, the statistical significance of the CARs and ARs closely resemble the ones reported in Tables A1—A5.

In addition, we have also experimented with other explanatory variables, such as SGD/USD exchange rate as a proxy for regional factor (one of the independent variables of the market model), instead of the RMB/USD variable used in the initial analysis. Once again, the results of the analysis do not differ substantially from the ones reported in the previous section. In the interest of conserving space, we do not report the details of these checks, but they are available on request.

Furthermore, the results of the analysis are consistent with observations of capital flows highlighted earlier, that is a larger-than-expected appreciation (depreciation) of the ringgit in reaction to a positive (negative) political event is followed by capital inflows (outflows). Next, our results are not contradicted by other analysts’ reports in the media, in relation to the effects of GE13,¹¹ GE14,¹² appointment of a new Central Bank Governor,¹³ imprisonment of opposition leader Anwar Ibrahim¹⁴ and the largest corruption scandal in history (1MDB exposé).¹⁵

We add that our market model already controls for global and regional factors that might affect the value of the ringgit. The local/national factors are not directly controlled for but are to be captured by the event windows instead. In this regard, there are no other similarly significant and momentous local factors occurring in each of our five chosen event windows. Hence, any confounding factors are minimized. These additional checks imply that our results and findings are robust.

Conclusion and Implications

This article attempts to assess the response of the foreign exchange market in Malaysia to a number of political events. These events include two general elections (namely the GE13 and GE14), the 1MDB exposé, the imprisonment of erstwhile opposition leader Anwar Ibrahim and the appointment of Muhammad Ibrahim as the eighth Governor of the Central Bank of Malaysia. Our study is conducted within the framework of an Event Study. We adopt the spot RM/USD exchange rate as the proxy for the foreign exchange markets. Using daily data in a 250-day estimation window, we estimated market model of RM/USD exchange rate returns in an ARDL framework and calculated the resulting ARs and CARs. The statistical significance of the ARs and CARs is also determined. To our knowledge, this is the first study of its kind in Malaysia and represents a meaningful contribution to the literature.

Some of our salient findings are that the CARs from the scandal, arising from the 1MDB exposé, are significantly positive. This indicates that the ringgit had depreciated more than expectations, signifying an unfavourable market response. Second, the response to the results of the general elections are not uniform. Particularly, subsequent CARs in reaction to the GE13 were mostly positive, again implying a larger-than-expected depreciation of the ringgit and an unfavourable market reaction. In contrast, the CARs for the GE14 were mostly negative, implying a favourable market response since the ringgit depreciated less than expected. Meanwhile, CARs in reaction to Anwar Ibrahim’s imprisonment had been negative—generally an unfavourable market reaction. Finally, the subsequent CARs in response to the appointment of a new Central Bank Governor was negative towards the end of the event window, implying a favourable market reaction. We also notice that these results are robust, even if RM/USD is substituted with RM/GBP and RM/EUR, or if different independent variables are used.

There are a number of practical implications following the results of our article. First, the market’s reaction to general election outcomes could be used as input to measure how well a newly elected government is performing as a gauge of public sentiments. For instance, the loss of public enthusiasm for the new government almost 2 months after GE13 is captured very succinctly in the CAR plots in Figures 2–4. The advantage that exchange rates have over public surveys and polls is that the former can be available real-time and can be analysed more quickly. Second, the manner in which the markets responded to political events suggests evidence of market inefficiency, that is, a violation of the efficient markets hypothesis (EMH). As such, there exists arbitrage opportunities among players in the foreign exchange markets. For instance, the exposure of corruption (in the case of the 1MDB exposé) and the subsequent unfavourable market response imply arbitrage opportunities, shorting the ringgit before the market reaction is fully reflected. Finally, as the response of the three exchange rates to key political events are quite uniform, there is very little room for the practice of diversification among RM/USD, RM/EUR and RM/GBP when a key political event is triggered.

Footnotes

Acknowledgements

The author would like to thank the anonymous referee and participants of the Nottingham University Business School i-CON Conference 2020 for comments. Gratitude is also extended to HELP University for financing this research.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author received no financial support for the research, authorship and/or publication of this article.

Appendix

Table A5.

Selected t-Values of Abnormal Returns (AR) and Cumulative Abnormal Returns (CAR) in Response to the 14th General Election outcome (GE14)

Event Days	CAR	AR
d − 15	2.048*	0.000
d − 14	2.261*	0.251*
d − 10	0.795	0.336*
d − 9	0.254	0.365*
d − 8	−0.471	0.312
d + 15	1.375	−0.560*
d + 23	−4.321*	−1.244
d + 33	1.949*	−2.206
d + 37	−2.779	−2.345
d + 43	−3.328*	−3.200
d + 55	−1.796	−3.838*
d + 56	2.010*	−3.615
d + 57	−0.905	−3.716*
d + 58	−0.404	−3.760*
d + 59	−3.071*	−4.102*
d + 60	−0.286	−4.133*

Source: Author’s calculations.

Note: To conserve space, the actual AR and CAR values are not reported, but are available upon request. * Indicates a 5% level of significance based on bootstrapped critical values.

Notes

References

Ahmad

, Rhee

S. G.

, & Wong

Y. M.

(2012). Foreign exchange market efficiency under recent crises: Asia-Pacific focus. Journal of International Money and Finance, 31(6), 1574–1592

Adam

, Kozinsky

, & Zielinski

(2013). To what extent can central banks influence exchange rates with foreign exchange interventions? The case of Poland (BIS Papers No. 73). BIS.

Bowes

D. R.

(2018). Stock market volatility and presidential election uncertainty: Evidence from political futures markets. Journal of Applied Business Research, 34(1), 143–150.

Bialkowski

, Gottschalk

, & Wisniewski

T. P.

(2008). Stock market volatility around national elections. Journal of Banking and Finance, 32(9), 1941–1953.

Bin

(2015). Political leadership events and stock market reactions: Evidence from the Greater China region. Journal of Accounting and Finance, 15(8), 81–94.

BIS (2016). Foreign exchange turnover in April 2016 (Triennial Central Bank Survey). Bank for International Settlements.

BIS (2019). Global foreign exchange market turnover in 2019 (Triennial Central Bank Survey). Bank for International Settlements.

Booth

J. R.

, & Booth

L. C.

(2003). Is presidential cycle in security returns merely a reflection of business conditions? Review of Financial Economics, 12, 131–159.

Brown

C. R.

(2018). The Sarawak report: The inside story of the 1MDB exposé. Gerakbudaya Enterprise.

10.

Brown

S. J.

, & Warner

J. B.

(1985). Using daily stock returns: The case of event studies. Journal of Financial Economics, 14(1), 3–31.

11.

Chesney

, Reshetar

, & Karaman

(2011). The impact of terrorism on financial markets: An empirical study. Journal of Banking and Finance, 35(2), 253–267.

12.

Chia

R. C. C.

(2018). General election and political uncertainty in the Malaysia Stock Market: Evidence from stock market returns and their volatility (PhD dissertation). Yokohama National University.

13.

Chin

(2015). Malaysia: Pseudo-democracy and the making of a Malay-Islamic State. In Case

(Ed.), Routledge handbook of Southeast Asian democratization (pp. 399–409), Routledge.

14.

Colon-De-Armas

, Rodriguez

, & Romero

(2017). Investor sentiment and US presidential elections. Review of Behavioral Finance, 9(3), 227–241.

15.

Corsetti

, Pesenti

, & Roubini

(1999). What caused the Asian currency and financial crisis? Japan and the World Economy, 11(3), 305–373.

16.

Echavarria

J. J.

, & Villamizar-Villegas

(2016). Great expectations? Evidence from Colombia’s exchange rate survey. Latin American Economic Review, 25(3), 1–27.

17.

Frenkel

J. A.

(1981). Flexible exchange rates, prices, and the role of ‘News’: Lessons from the 1970s. Journal of Political Economy, 89(4), 665–705.

18.

Gunasegaram

(2018). 1MDB: The scandal that brought down a government. The Strategic Information and Research Development Centre (SIRD).

19.

Haggard

(2000). The political economy of the Asian Financial Crisis. Peterson Institute for International Economics.

20.

Hou

, & Li

(2019). The price of official-business collusion: Evidence from the stock market reaction to ‘Hunting the Tigers’ in China. China Finance Review International. https://mpra.ub.uni-muenchen.de/id/eprint/98149

21.

IMF 1997. World economic outlook: Crisis in Asia, regional and global implication. Interim assessment. IMF.

22.

Lean

H. H.

(2010). Political general election and stock performance: The Malaysian evidence. In Ismail

M. T.

(Ed.), Research in mathematics and economics (pp. 111–120). USM Press.

23.

Liew

V. K. S.

, & Rowland

(2016). The effect of Malaysia general election on stock market returns. SpringerPlus, 5, 1–13.

24.

Lobo

B. J.

, & Tufte

(1998). Exchange rate volatility: Does politics matter? Journal of Macroeconomics, 20(2), 351–365.

25.

Kliger

, & Gurevich

(2014). Event studies for financial research: A comprehensive guide. Palgrave Macmillan.

26.

Kollias

, Papadamou

, & Stagiannis

(2011). Terrorism and capital markets: The effects of the Madrid and London bomb attacks. International Review of Economics and Finance, 20(4), 532–541.

27.

Khuntia

, Pattanayak

J. K.

, & Hiremath

G. S.

(2018). Is the foreign exchange market efficiency adaptive? The empirical evidence from India. Journal of Asia Pacific Business, 19(4), 261–285.

28.

MacKinlay

A. C.

(1997). Event studies in economics and finance. Journal of Economic Literature, 35(1), 13–39.

29.

Mpofu

T. R.

, & Peters

(2017). The impact of monetary policy announcements and political events on the exchange rate: The case of South Africa (ERSA Working Paper 700). Economic Research Southern Africa (ERSA).

30.

Nazir

M. S.

, Younus

, Kaleem

, & Anwar

(2014). Impact of political events on stock market returns: Empirical evidence from Pakistan. Journal of Economics and Administrative Sciences, 30(1), 60–78.

31.

Perkins

D. H.

, & Woo

W. T.

(2000). Malaysia: Adjusting to deep integration with the world economy. In Woo

W. T.

, Sachs

J. D.

, & Schwab

. (Eds.), The Asian financial crisis: Lessons for a resilient Asia (pp. 227–255). MIT Press.

32.

Rasiah

(2001). Pre-crisis weaknesses and vulnerabilities. In Jomo

K. S.

(Ed.), Malaysian eclipse: Economic crisis and recovery (pp. 47–66). Zed Books.

33.

Sayan

(2006). Business cycles and workers’ remittances: How do migrant workers respond to cyclical movements of GDP at home? (IMF Working Paper WP/06/52). IMF.

34.

Sen

, & Tyce

(2017). The elusive quest for high income status—Malaysia and Thailand in the post-crisis years. Structural Change and Economic Dynamics, 48, 117–135.

35.

The Economist (2016, 24 November. Malaysia’s central bank tries to stem a slide in the ringgit. The Economist, pp. 64–65.

36.

Wong

W. Y.

, & Hooy

C. W.

, (2016). The impact of election on stock market returns of government-owned banks: The case of Indonesia, Malaysia and Thailand. Asian Journal of Business and Accounting, 9(1), 31–58.

37.

Yusoff

W. A.

, Salleh

M. F. M.

, Ahmad

, & Idris

(2015). Short-run political events and stock market reactions: Evidence from companies connected to Malaysian bi-power business-political elite. Procedia—Social and Behavioral Sciences, 211, 421–428.