Does foreign portfolio investment moderate the impact of exchange rate volatility and investor sentiment on country index crash risk?

Abstract

This study evaluates the relationship investor sentiment, exchange rate volatility, net foreign portfolio investment and the country index crash risk. The moderating variable, net foreign portfolio investment, is introduced. While previous crash risk studies typically focus on individual firms, this study takes a country-level perspective. CRASH, NCSKEW and DUVOL represent the Country Index Crash risk. The data will be analyzed using EViews software, including panel data from logistic regression and OLS regression using a two-dimensional clustered standard error method. The findings demonstrate the importance of exchange rate fluctuations and investor mood in affecting the country index crash risk. The influence of Net Foreign Portfolio Investment on the crash risk is negligible. Moreover, the study reveals that higher Net Foreign Portfolio Investment does not strengthen the impact of Investor Sentiment but weakens its influence in conjunction with Exchange Rate Volatility on the country index crash risk.

Keywords:

• Net foreign portfolio investment
• investor sentiment
• exchange rate volatility
• country index crash risk
• CRASH
• NCSKEW
• DUVOL

REVIEWING EDITOR:

• David McMillan, University of Stirling, UNITED KINGDOM

SUBJECTS:

• Economics; Finance; Business
• Management and Accounting

JEL Classification Code:

• E60
• F20
• F21
• F40
• F60
• F65
• G10
• G15
• G33

1. Introduction

Research on crash risk studying in various countries have been conducted. Chen et al. (2001) used Down to Up Volatility (DUVOL) and Negative Conditional Skewness (NCSKEW) as proxies for stock price crash risk in their investigation of crash risk. Research by Chen et al. (2001) has been followed by Jin et al. (2006) and Hutton et al. (2009) who found indications of stock prices crash risk using the NCSKEW and DUVOL methods. Then, research on crash risk began to be widely studied, including by Kim et al. (2011) conducted research utilizing a dataset comprising U.S. companies from 1993 to 2009. The findings revealed the presence of crash risk, identified through the application of NCSKEW and DUVOL methodologies. Subsequent research by Callen and Fang (2015) found that The likelihood of a crash decreases with increasing religiosity. The latest research conducted by Petri (2020) found crash risk in 73 sample countries using the NCSKEW and DUVOL methods.

Global and regional issues will affect a country’s crash risk. The Market Outlook reported that for 2021 the sharpest fall in stock prices occurred in 1998 As a consequence of the worldwide financial crisis, the collapse in subprime mortgage stock values brought on by the 2008 financial crisis. A stock price crash was noted in 2018 and ascribed to a trade spat between the US and China. Stock values declined as a consequence of the COVID-19 epidemic in 2020. The financial crises significantly influenced the volatility of regional stock prices in places like Europe, the Middle East and Asia. National and international concerns and events can strongly influence a country’s index crash risk.

Kim et al. (2011) used the American sample, while Xu et al. (2014) used the Chinese sample. Lim et al. (2016) conducted with a sample of Korean countries. Then Ni and Zhu (2016) used a sample of countries included in the Emerging Markets for 2008–2014. Furthermore, Lee researched the Asian Emerging Market from 1997 to 2003. The latest research conducted by Petri (2020) used a sample of 73 countries included in MSCI's Global Investable Market Index (GIMI). Exploring crash risk is crucial in the field of risk management, making it a particularly intriguing and essential area of investigation. A short time crash risk occurs due to negative sentiment in the capital market (Cui & Zhang, 2019). Sometimes investors overreact, causing stock prices to fall quickly. Negative sentiment can be in the form of this, causing the market conditions to be bearish. Risk is one of the most important factors in making investment decisions. The probability of negative returns from an investment is a risk that must be avoided. Cui and Zhang (2019) found that investor sentiment and the likelihood of a stock market crash are positively and significantly correlated. Furthermore, as Baker and Wurgler (2006) imply, the trade volume is a gauge of investor sentiment.

Following the 1997 Asian Financial Crisis, economists, international investors and policymakers have shown considerable interest in exploring the connection between stock prices and exchange rates (Yau & Nieh, 2006). The likelihood of foreign debt will rise along with the weakening a country’s currency. Foroni et al. (2017) examined the potential risk transmission from the sovereign debt market to the currency market by introducing a novel risk premium factor for forecasting returns on exchange rates associated with sovereign risk. Their study indicated that random and alternative models yielded less precise predictions. Additionally, short-term economic shocks, including fluctuations in interest rates, the trade balance, exchange rates, employment and inflation, can impact stock prices (Yao & Luo, 2009). Investment in dollars tends to rise in the context of prolonged depreciation in exchange rates. This shift redirects resources that could otherwise be invested in currencies toward assets denominated in dollars, as noted by Coleman and Tettey (2011).

Foreign investors’ effects on stock market crash on the Ho Chi Minh City stock exchange was examined by Vo (2018) who classified them as an independent variable. His research indicates a strong relationship between foreign investors and the chance of a future stock market crash. Halale (2014) argues that the role of foreign portfolio investment is substantial in stock price movements. The results of his research show that daily inflows and outflows of foreign investment portfolios significantly influence the daily price index. Based on the description and explanation of the background presented earlier, there appears to be a research gap in both the empirical and theoretical gaps regarding the stock price crash risk. As a result, we arrive at a tentative hypothesis that this study will further investigate the influence of foreign portfolio investments, exchange rate volatility and investor sentiment on the likelihood of a country’s index crash risk.

2. Literature review

2.1. Investor sentiment and country index crash risk

Investor sentiment refers to investors’ attitude and outlook toward the market and its future direction. Various factors, including economic data, news events and market trends, can drive it. Positive investor sentiment may lead to an increase in the purchase of stocks and other assets since investors are often upbeat about the state of the market. On the other hand, investors may be more cautious and inclined to sell their investments when the mood among investors is unfavorable. Country index crash risk refers to the likelihood that the stock price index will experience a significant and rapid decline. It can be driven by various factors, including economic shocks, political events and other unforeseen events that can impact the market. Investor sentiment is a well-established relationship with stock price crash risk. When overly optimistic, investor sentiment can create an environment where investors are more likely to take risks and engage in speculative behavior. It can lead to a buildup of market imbalances and excesses, increasing the likelihood of a market crash. Similarly, when excessive pessimism among investors, it can create an environment where investors are more likely to panic and sell their holdings, contributing to a market crash. Market fluctuations are closely related to market sentiment.

Pan (2019) found that stock price bubbles were strongly correlated with negative investor sentiment, indicating that irrational sentiment contributes significantly to excess market volatility. Moreover, the study indicates increased volatility and returns result from an inefficient market’s asymmetric structure. A significant correlation between investor sentiment and the likelihood of further stock market crashes was found by Yin and Tian (2015). They also emphasized how a need for more basic data on stock prices and optimistic market circumstances strengthens this beneficial association. However, Gong et al. (2016) concluded that stock price volatility is unaffected by investor sentiment. Considering these findings, the hypothesis is:

Hypothesis 1 (H1): Investor sentiment has an effect on country index crash risk.

2.2. Exchange rate volatility and country index crash risk

The extent of variation in a nation’s currency in value relative to other currencies is known as exchange rate volatility. Several things, including shifts in interest rates, developments in geopolitics and the publication of economic data, can cause this volatility. When there is volatility in currency rates, businesses and investors participating in international commerce and investment may face uncertainty and risk. A growing corpus of evidence suggests an exchange rate volatility and stock market crash probability are related. In particular, increased volatility in exchange rates might increase the likelihood of a stock market crashes or significant drop. This link can be explained, in part, by the fact that enterprises operating in foreign markets face greater risk and uncertainty due to heightened exchange rate volatility. This uncertainty may eventually impact stock prices if investment and economic activity decline. Extremely volatile exchange rates can also affect the profitability of multinational corporations, which in turn can affect stock prices.

According to research by Dungey and Martin (2015), stock market crashes in several nations, including the US, Canada and Japan, were strongly predicted by exchange rate volatility. Likewise, Huang et al.’s (2021) study discovered that exchange rate volatility greatly influenced the chance of a crash of the Chinese stock market. A stock market crash probability is directly linked to exchange rate volatility. Excessive fluctuations in exchange rates can put investors and companies at risk, affecting stock prices and raising the possibility of a market crash. To reduce investment risk, investors should thus keep a close eye on exchange rate volatility and other market health markers.

Hypothesis 2 (H2): Exchange rate volatility has an effect on country index crash risk.

2.3. Net foreign portfolio investment and country index crash risk

Several research papers have investigated the connection between foreign portfolio investments and the likelihood of stock market crashes. The study by Shruti and Thenmozhi (2024) investigates the impact of foreign institutional investors (FII) on stock price crash risk in India. Panel regression findings indicate that higher levels of FII ownership, signifying positional trading, exacerbate stock price crash risk. Foreign investment in the domestic financial market was studied by Hamao and Mei (2001). Their study’s conclusions show that, in contrast to the domestic market, trading by overseas investors exhibits greater market volatility. Besides, foreign investors often employ more advanced technology than the technology used by domestic investors, making foreign investors lose money on them, forcing foreign investors to prioritize short-term earnings over long-term profits. Compared to long-term gains, Çitak (2019) identified that Foreign Portfolio Investment was important variable that cause the possibility of a bubble forming in the Turkish stock market. Based on this, the hypothesis is:

Hypothesis 3 (H3): Net foreign portfolio investment has an effect on the country’s index crash risk.

2.4. Net foreign portfolio investment, investor sentiment and country index crash risk

Investor sentiment is excessively high, and foreign portfolio investment positively influences country index crash risk. An abundance of foreign investment amid elevated the sentiment of investors might raise the risk of a market crash. An investor sentiment analysis examines various financial topics, particularly bubbles, market crash and financial crisis prediction. Once economic variables are considered, investor emotion exhibits complementary information that can lead to a wide range of future financial and economic applications (Pan, 2019). Stock price crash risk contributes to the complexity and intricate facets of the stock market (Peng & Hu, 2020). Sentiment assumes a pivotal role in investor decision-making and stock price movements. Sentiment-driven investor behavior influences different financial market trading strategies. According to Bouteska (2020), investors are likelier to act on investor surveys information as opposed to sentiment indices derived from the market, underscoring the importance of sentiment in driving stock prices. Investor behavior is correlated with bullish and bearish markets. In bear markets, trading choices are more susceptible to sentiment-driven fluctuations. Based on this, the hypothesis is:

Hypothesis 4 (H4): The higher the net foreign portfolio investment, the stronger the influence of investor sentiment on crash risk.

2.5. Net foreign portfolio investment, exchange rate volatility and country index crash risk

The worse the financial crisis, the greater the need for capital and liquidity, resulting in stronger pressure on the exchange rate (Fratzscher, 2009). Sui and Sun (2016) state that the financial crisis exacerbates the spillover effect on the exchange rate on stock returns. Omrane and Savaşer (2017) find that the euro and pound markets display signs of earlier and gradual weakening compared to the yen throughout the financial crisis of 2007. The impact of macroeconomics on stock prices before and during the 2008 financial crisis was investigated by Sheikh et al. (2020). The 48 observations that comprise the pre-economic crisis era ran from January 2004 to December 2007, in contrast, the 120 observations that comprise the time following the economic crisis occurred between January 2008 and December 2018. The four tests used to determine if the data are stationary are the Philips Schmidt Shin Kwiatkowski test, the Dickey-Fuller test, the Zivot-Andrew unit root, and the Philips Perron test. The findings demonstrate that investors’ responses to the price of gold and oil vary over the long run and pre-global financial crisis. Long-term and post-crisis, investors have responded to all macroeconomic swings in varying ways. It demonstrates how investors respond to both favorable and unfavorable shocks to gold. Another intriguing feature is that, since the global financial crisis, investors have only responded positively to shocks associated with gold prices, interest rates and currency exchange rates over the long run. Coleman and Tettey (2011) a declining currency quickly impacts the volume of trade on the stock exchange and the price index because money moves overseas, allowing foreign portfolio investors who are significant players to shift their money to other, more exciting markets. Based on this, the hypothesis is:

Hypothesis 5 (H5): The higher the net foreign portfolio investment, the stronger the influence of exchange rate volatility on the Country’s Index crash risk.

3. Methodology

3.1. Calculation of dependent variable (country index crash risk)

CRASH: If the daily rate of return for an index of stock prices is 3.09 or less than the average standard deviation, there is a crash risk (Hutton et al., 2009). This study uses a standard deviation of 3.09 because the stock price index between developed and developing countries has different volatility. For developed countries, a 1000-point drop does not yet indicate a crash risk, but for developing countries, it has entered the crash category. If a crash risk occurs is 1 and 0 if there is no crash risk (Harymawan et al., 2019).

NCSKEW is the negative trend of a company’s daily returns over a monthly period. Monthly data are found by multiplying minus one by splitting the cube of the daily return’s standard deviation after obtaining the cube of the daily return. The likelihood of a stock market crash increases with the NCSKEW number. (1) $${\mathit{\text{NCSKEW}}}_{\mathit{\text{it}}}=\frac{-(\mathrm{n}{(n-1)}^{3/2}\sum R_{\mathit{\text{it}}}^3)}{\left({\left(n-1\right)\left(n-2\right)\left(\sum R_{\mathit{\text{it}}}^2\right)}^{3/2}\right)}$$(1)

Where:

$R_{\mathit{\text{it}}}:\mathit{\text{daily returns of stock}}\mathrm{ }i\mathrm{ }\mathit{\text{in period}}\mathrm{ }t$

$N:\mathit{\text{number of observations in daily period}}$

DUVOL is the abbreviation for daily return volatility. The natural logarithm was used to determine the ratio of daily return standard deviations below monthly average to daily return standard deviations above monthly average. A metric of DUVOL is the standard deviation log ratio. The likelihood of stock prices crashing increases with the DUVOL number. (2) $${\mathit{\text{DUVOL}}}_{\mathit{\text{it}}}=\mathrm{\text{Log}}\mathrm{ }\left\{\frac{\left(n_u-1\right)\sum _{\mathit{\text{DOWN}}}{R}_{\mathit{\text{it}}}^2}{\left(\left(n_d-1\right)\sum _{\mathit{\text{UP}}}{R}_{\mathit{\text{it}}}^2\right)}\right\}$$(2)

3.2. Calculation of independent variable

The exchange rate data is typical of the rate established by national authorities and was gathered from the websites of each country’s central bank and the CEIC Database. Exchange rate volatility was calculated using the exchange rate’s absolute percentage change. For the Exchange rate volatility using standard deviation of daily values aggregated to a monthly volatility data of exchange rate dollar to local currency.

An assessment of investor sentiment by Williams %R is a simple formula based on the highest high, lowest low and closing price of an asset (Zhou, 2018).

The net foreign portfolio investment variable is calculated by the nominal amount of foreign portfolio investment in the stock exchange of each country converted to US dollars.

The Table 1 is the summary of the dependent and independent variables in this study.

Table 1. Variables.

Variables	Description
CRASH	If there are one or more daily returns, the composite index will be assigned a value of 1, and if it falls by 3.09 standard deviations that are lower than the mean daily return during a given month, it will be assigned a value of 0. It is because throughout the financial year period, 3.09 will lead to a frequency of 0.1% in the standard distribution and 0 otherwise.
NCSKEW	Negative skewness of daily returns. Monthly data is found by multiplying minus one by the cube of the weekly return and then dividing by the daily return’s cube of standard deviation.
DUVOL	The natural logarithm is used to indicate the ratio of the daily return standard deviation above the monthly average to the daily return standard deviation below the monthly average.
Exchange rate volatility	standard deviation of daily values aggregated to a monthly volatility data of exchange rate dollar to local currency.
Investor sentiment	Williams %R is a simple formula based on the highest high, lowest low and closing price of an asset (Zhou, 2018) $\%R=\frac{\text{Highest High}-\text{Closing Price}}{\mathit{\text{Highest High}}-\mathit{\text{Lowest Low}}}x100\%$
Net foreign portfolio investment	Monthly data on the net amount of a country’s foreign portfolio investment

4. Empirical model

The regression equation for panel data in this study are: $$Y\mathrm{ }\left(\mathit{\text{Crash}}\right)=\alpha +{\beta }_1{\mathit{\text{IS}}}_{1\mathit{\text{it}}}+{\beta }_2{\mathit{\text{VER}}}_{2\mathit{\text{it}}}+{\beta }_3{\mathit{\text{FPI}}}_{3\mathit{\text{it}}}+{\beta }_4{\mathit{\text{IS}}.\mathit{\text{FPI}}}_{4\mathit{\text{it}}}+{\beta }_5{\mathit{\text{VER}}.\mathit{\text{FPI}}}_{5\mathit{\text{it}}}{+\mu }_{\mathit{\text{it}}}$$ $$Y\mathrm{ }\left(\mathit{\text{NCSKEW}}\right)=\alpha +{\beta }_1{\mathit{\text{IS}}}_{1\mathit{\text{it}}}+{\beta }_2{\mathit{\text{VER}}}_{2\mathit{\text{it}}}+{\beta }_3{\mathit{\text{FPI}}}_{3\mathit{\text{it}}}+{\beta }_4{\mathit{\text{IS}}.\mathit{\text{FPI}}}_{4\mathit{\text{it}}}+{\beta }_5{\mathit{\text{VER}}.\mathit{\text{FPI}}}_{5\mathit{\text{it}}}{+\mu }_{\mathit{\text{it}}}$$ $$Y\mathrm{ }\left(\mathit{\text{DUVOL}}\right)=\alpha +{\beta }_1{\mathit{\text{IS}}}_{1\mathit{\text{it}}}+{\beta }_2{\mathit{\text{VER}}}_{2\mathit{\text{it}}}+{\beta }_3{\mathit{\text{FPI}}}_{3\mathit{\text{it}}}+{\beta }_4{\mathit{\text{IS}}.\mathit{\text{FPI}}}_{4\mathit{\text{it}}}+{\beta }_5{\mathit{\text{VER}}.\mathit{\text{FPI}}}_{5\mathit{\text{it}}}{+\mu }_{\mathit{\text{it}}}$$

Where:

a: constant

β_1: coefficient of regression for Investor sentiment volatility

β_2: coefficient of regression for exchange rate volatility

β_3: coefficient of regression for net foreign portfolio investment

β_4: coefficient of regression for investor sentiment with the moderating variable of net foreign portfolio investment.

β_5: coefficient of regression for exchange rate volatility with the moderating variable of net foreign portfolio investment

4.1. Data and sample

The data is the secondary data from the websites of each country’s central banks, Investing.com, yahoo finance, Bloomberg, World Bank and the CEIC database. The list of countries will be from the World Bank website. The data for the crash risk variable from the composite stock price index data from the investing.com and yahoo finance websites. Meanwhile, data on investor sentiment, exchange rate volatility and net foreign portfolio investment are from the CEIC database and the IMF. The sample of this study is 31 samples of countries stock index (Malaysia, Indonesia, Saudi Arabia, Turkey, Switzerland, United Kingdom, United States, Canada, Thailand, India, Russian Federation, Germany, Philippines, France, Spain, Bosnia and Herzegovina, Czech Republic, Denmark, Finland, Hong Kong SAR (China), Croatia, Iceland, Italy, Japan, South Korea, Netherlands, Norway, Poland, Portugal, Slovenia, Sweden). The decision to use the sample period of 2015–2020 and the 31 countries mentioned was likely made based on the data for the selected countries and periods readily available from the IMF and World Bank websites. Additionally, this particular period (2015–2020) mark by several significant global events, such as Brexit, the US-China trade war and the COVID-19 pandemic, which substantially impacted global financial markets. The criteria in Table 2.

Table 2. Sample selection.

Sample selection from 2015 to 2020 available from the IMF website and world bank website	96
Uncompleted data	65
Total sample	31

Source: Prepared by authors, 2023.

4.2. Descriptive statistics

The independent variables used in this study are investor sentiment, exchange rate volatility and foreign portfolio investment. In contrast, the dependent variable market crashes are proxy Crashes (a dummy variable), NCSKEW and DUVOL. The Data from Table 3 show 446 samples had a risk of crashes, while those that did not crash were 1786 compared to the non-crash sample, which has a negative value (−0.22), the average positive NCSKEW crash value is 0.97. It demonstrates a positive correlation between the NCSKEW value and the likelihood of a crash. It implies that both groups’ NCSKEW variables are adversely skewed. Additionally, the NCSKEW standard deviations for Crash and Non-Crash are 0.870 and 0.850, respectively, suggesting that the data in both groups was well distributed.

Table 3. Descriptive.

Variables	Y	N	Mean	Min	Max	Standard deviation
NCSKEW	Crash	446	0.973	−1.562	4.262	0.870
	Non-crash	1786	−0.223	−1.985	2.297	0.850
DUVOL	Crash	446	0.501	−0.507	1.872	0.360
	Non-crash	1786	−0.118	−3.540	1.642	0.429
Investor sentiment	Crash	446	47.28	0.000	100	30.48
	Non-crash	1786	42.41	0.000	100	29.54
Exchange rate volatility	Crash	446	0.0062	0.000	0.063	0.008
	Non-crash	1786	0.0061	0.000	0.062	0.007
Net foreign portfolio	Crash	446	2.680	0.065	1.588	3.744
Investment	Non-crash	1786	2.918	0.066	1.698	3.848

Source: Prepared by authors, 2023.

For a sample where the crash risk is 0.5, the average value of the variable Y (Proxy DUVOL) is more significant than the sample where the crash risk is −0.12. It suggests a greater chance of a crash with a higher DUVOL rating. It implies that in both groups, DUVOL has a positive skew. Additionally, the DUVOL standard deviations for the Crash and Non-Crash groups are 0.360 and 0.429, respectively, suggesting substantial data dispersion in both categories.

The mean value for the Investor Sentiment variable is 42.41 for non-crash and 47.28 for Crash. Furthermore, the investor sentiment standard deviations for the crash and non-crash categories are 30.48 and 29.54, respectively, suggesting that the data in both categories is balanced. The average value of investor sentiment during an impact is positive. A positive outlook among investors may lead them to take greater chances and allocate more capital to riskier investments. Because exposure to hazardous assets can result in high market volatility, raising the possibility of a market crash.

The exchange rate volatility crash averages 0.0062, higher than the non-crash sample’s 0.0061. It demonstrates that the likelihood of a crash increases with exchange rate volatility. The maximum value for the Exchange Rate Volatility variable is 0.062 for a non-crash and 0.063 for a crash. Exchange Rate Volatility standard deviations for Crash and Non-Crash are 0.008 and 0.007, respectively, suggesting that the data in both groups is not too distributed.

The mean value for the Net Foreign Portfolio Investment variable is 2.918 for non-crash and 2.680 for Crash. The Net Foreign Portfolio Investment variable is marginally more significant in the non-Crash group. Moreover, the Net Foreign Portfolio Investment standard deviations for the Crash and Non-Crash groups are 3.744 and 3.848, respectively, suggesting significant data dispersion in both sets. The average value of the foreign portfolio investment crash was 2.68, more significant than the sample that did not crash, 2.92. It shows that the higher the value of the foreign portfolio investment, the less likely a crash will occur.

Overall, the descriptive statistics suggest that there are differences between the Crash and Non-Crash groups in the NCSKEW, DUVOL and Net Foreign Portfolio Investment variables, but not in the Investor Sentiment and Exchange Rate Volatility variables. The dispersion of the data is also different across the variables, with some variables having highly dispersed data (such as Net Foreign Portfolio Investment) and others having moderately dispersed data (such as DUVOL).

5. Data analysis and interpretation

The correlation matrix’s output is in Table 4. According to Table 4’s data, Investor sentiment and CRASH have a somewhat positive correlation (r = 0.005). Exchange Rate Volatility and Crash have a very slight positive link (correlation coefficient of 0.004). Foreign Portfolio Investment and CRASH have a weakly negative link, as seen by their correlation coefficient of −0.025. A modest negative association exists between Investor Sentiment and NCSKEW, as indicated by the correlation coefficient of −0.07. Exchange Rate Volatility and NCSKEW have a weakly negative link, as seen by the correlation coefficient −0.05. A weak negative link exists between Foreign Portfolio Investment and NCSKEW, as indicated by the correlation value of −0.03. Investor Sentiment and DUVOL have a −0.05-correlation coefficient, which suggests a slight negative relationship.

Table 4. Correlation matrix.

Variables	CRASH	Investor sentiment	Exchange rate volatility	Foreign portfolio investment
CRASH	1
Investor sentiment	0.004978	1
Exchange rate volatility	0.003593	0.009655	1
Foreign portfolio investment	−0.024868	−0.006621	−0.014037	1
Variables	NCSKEW	Investor sentiment	Exchange rate volatility	Foreign portfolio investment
NCSKEW	1
Investor sentiment	−0.073257	1
Exchange rate volatility	−0.054629	0.009655	1
Foreign portfolio investment	−0.033963	−0.006621	−0.014037	1
Variables	DUVOL	Investor sentiment	Exchange rate volatility	Foreign portfolio investment
DUVOL	1
Investor sentiment	−0.048802	1
Exchange rate volatility	−0.037348	0.009655	1
Foreign portfolio investment	−0.035593	−0.006621	−0.014037	1

Source: Prepared by authors, 2023.

5.1. Dependent variable crash

Table 5 shows model 1 – model 5 using the dependent variable crash. Based on these data, the significance value of model 5 is the most significant compared to the other models. However, the highest % correct level is model 5, 80.20% compared to other models. Thus, this study uses model 5. Goodness-of-Fit Evaluation for Binary that shows the probability level of 0.7639> alpha 0.05, which means that this research model is acceptable. The model accuracy test’s value is to determine how good the data is from the prediction results of the model and as a measure of model accuracy. The model accuracy test calculates the correct and incorrect estimated values in the Expectation-Prediction Evaluation table. Table 5 shows that the accuracy rate is 80.20% with the variable Foreign Portfolio Investment, Investor Sentiment and Exchange Rate Volatility.

Table 5. Logit regression.

	Model 1		Model 2		Model 3		Model 4		Model 5
Variable	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob
C	−1.394603	0.0000	−1.394722	0.000	−1.347355	0.0000	−0.818925	0.0000	−1.366583	0.0000
Investor sentiment (IS)	0.005445	0.8652	0.070891	0.8154	0.067323	0.8229	−0.056071	0.7566	−0.105014	0.7427
Exchange rate volatility (ER)			1.146669	0.8671	1.027923	0.8806	0.533479	0.8909	4.151561	0.5929
Net foreign portfolio investment (FI)					−0.000167	0.2420	−0.00958	0.2216	−0.00853	0.6409
Moderating 1 (IS*FI)							0.001094	0.1314	0.001953	0.1250
Moderating 2 (ER*FI)									−0.015138	0.4331
McFadden R-squared	0.000013		0.000037		0.001680		0.001699		0.001983
S.D. dependent var	0.399955		0.399955		0.399955		0.399955		0.399955
Akaike info criterion	1.002087		1.002959		1.003108		1.003089		1.003700
Schwarz criterion	1.007204		1.010634		1.015900		1.015881		1.019051
LR statistic	0.028622		0.083186		3.750612		3.792750		1.009305
n	2232		2232		2232		2232		2232
H/L statistic	8.1239		7.7135		7.1629		7.2771		4.9368
Prob. Chi-sq	0.4215		0.4619		0.5192		0.5071		0.7643
%Correct	80.02%		80.02%		80.05%		80.05%		80.06%
%Incorrect	19.98%		19.98%		19.95%		19.95%		19.94%
Obs with dep = 0	1786		1786		1786		1786		1786
Obs with dep = 1	446		446		446		446		446

Source: Prepared by authors, 2023.

The logistic regression equation for panel data in Model 5 can be written as: $$\mathrm{\text{CRASH}}=-1.728990+0.008770\left(\mathrm{\text{IS}}\right)+5.127870\left(\mathrm{\text{ER}}\right)+0.000364\left(\mathrm{\text{FI}}\right)‐0.000012\left(\mathrm{\text{ISFI}}\right)‐0.104766\left(\mathrm{\text{ERFI}}\right)+\mathrm{e}$$

The country index crash risk was significantly and negatively impacted by investor sentiment with a significantly value of 0.0001 < alpha 0.05. Meanwhile, Exchange Rate Volatility (0.2 > alpha 0.05) and Foreign Portfolio Investment (0.14 > alpha 0.05) have no impact on the likelihood of a country index crash risk. Net foreign portfolio investment does not influence Exchange Rate Volatility (0.1 > alpha 0.05. However, net Foreign Portfolio Investment does strengthen the effect of Investor Sentiment (0.01 > alpha 0.05) on the country index crash risk.

The Investor Sentiment variable has a positive coefficient in each model, which means that the likelihood of a crash increases with increasing investor sentiment. Additionally, the probability value (Prob) indicates this, which is very small in each model, such that it may be determined that the investor sentiment variable is very significant in influencing the occurrence of crash risk. Model 2 adds an Exchange Rate Volatility variable. However, the coefficient for this variable is not significant, with a probability of 0.87. Model 3 adds the variable Net Foreign Portfolio Investment. The coefficient for this variable is also not significant, with a significant probability of 0.24. Model 5 adds two moderator variables, namely IS*FI and ER*FI. The coefficients for IS*FI and ER*FI are negative, with a significant probability of 0.01 and 0.135, respectively. The interaction between Investor Sentiment and Foreign Portfolio and Exchange Rate Volatility and foreign portfolio investment is an essential factor in forecasting crash risk. Table 5 also shows the H/L Statistics, which tests the null hypothesis that the model fits the data. All models have relatively high significant probabilities, indicating they fit the data reasonably well. However, Model 5 has the lowest significant probability among all the models, indicating that it best fits the data.

5.2. Dependent variable NCSKEW

Table 6 Panel Regression shows model 1 – model 5 using the dependent variable NCSKEW. Based on the EViews software output. In model 4 and 5, two variables have a significant effect at the 5% significance level. Exchange rate volatility and investor sentiment in model 1 to model 5 does show a negative and significant effect on market crashes. but net foreign portfolio investment does not show significant effect in the model 3 to model 5. But Net Foreign Portfolio Investment does not strengthen the likelihood of investor sentiment and exchange rate volatility on country index crash risk.

Table 6. Panel data regression.

	Model 1		Model 2		Model 3		Model 4		Model 5
Variables	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob
C	0.016955	0.4121	0.059444	0.0250	0.085682	0.0055	0.085694	0.0055	0.089414	0.0061
Investor sentiment (IS)	−0.410839	0.0005	−0.407919	0.0006	−0.409200	0.0005	−0.427575	0.0008	−0.427358	0.0008
Exchange rate volatility (ER)			−6.895533	0.0106	−6.958449	0.0099	−6.691405	0.0099	−7.549033	0.0172
Net foreign portfolio investment (FI)					−0.00009	0.0951	−0.00005	0.0952	−0.000105	0.1272
Moderating 1 (IS*FI)							0.000199	0.6864	0.000198	0.6888
Moderating 2 (ER*FI)									0.002476	0.7231
Model	Common effect		Common effect		Common effect		Common effect		Common effect
Root MSE	0.976018		0.974591		0.973982		0.973946		0.973918
Mean dependent var	0.015964		0.015964		0.015964		0.015964		0.015964
S.D. dependent var	0.978867		0.978867		0.978867		0.978867		0.978867
Akaike info criterion	2.791122		2.789090		2.788735		2.789558		2.790398
Schwarz criterion	2.796239		2.796765		2.798970		2.802351		2.805749
Durbin-Watson stat	2.104524		2.108964		2.792472		2.111124		2.111505
R-squared	0.005367		0.008274		0.009514		0.009586		0.009642
Adjusted R-squared	0.004921		0.007385		0.008180		0.007807		0.007418
S,E, of regression	0.976456		0.975246		0.974855		0.975039		0.975230
F-statistic	12.03201		9.298785		7.133484		5.388873		4.334522
Prob(F-statistic)	0.000533		0.000095		0.00091		0.000256		0.000628

Source: Prepared by authors, 2023.

The regression equation for NCSKEW in Model 5 can be written as follows: $$\mathrm{\text{NCSKEW}}=0.089414\mathrm{ }‐\mathrm{ }0.427358\mathrm{ }\left(\mathrm{\text{IS}}\right)‐\mathrm{ }7.549033\mathrm{ }\left(\mathrm{\text{ER}}\right)-0.000105\mathrm{ }\left(\mathrm{\text{FI}}\right)+0.000198\mathrm{ }\left(\mathrm{\text{ISFI}}\right)+0.002476\mathrm{ }\left(\mathrm{\text{ERFI}}\right)+\mathrm{e}$$

5.3. Dependent variable DUVOL

Table 7 panel regression output shows Model 1 – Model 5 using the dependent variable DUVOL. The five models show that Model 5 is the best one. With a significance level of 0.0213, the constant term (C) has a coefficient of 0.037197, suggesting a strong positive connection. The probability of a country index crash and investor sentiment are highly positively correlated, as shown by the Investor Sentiment (IS) coefficient of 0.00141196 and significance level of 0.0247. The correlation of exchange rate volatility and the likelihood of a country index crash risk is non-significantly negative, as evidenced by the Exchange Rate Volatility (ER) coefficient of −2.884708 at the significance level of 0.0662. A non-significant negative correlation has been found between foreign portfolio investment and country index crash risk, as indicated by the coefficient of −0.000596 for net foreign portfolio investment (FI) at a significance level of 0.0818. The coefficient of 0.00674 for Moderating 1 (ISFI) at a significance level of 0.7831 indicates no significant impact of the interaction between Investor Sentiment and Net Foreign Portfolio Investment on the likelihood of country index crash risk. The country index crash risk is not significantly influenced by the exchange rate volatility and net foreign portfolio investment, as indicated by the coefficient of 0.002186 for moderating 2 (ERFI) at a significance level of 0.5280.

Table 7. Panel data regression.

	Model 1		Model 2		Model 3		Model 4		Model 5
Variables	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob
C	0.006103	0.5512	0.020476	0.1193	0.033908	0.0265	0.033912	0.0265	0.037197	0.0213
Investor sentiment (IS)	−0.135380	0.0211	−0.134392	0.0220	−0.135048	0.0213	−0.141388	0.0245	−0.0141196	0.0247
Exchange rate volatility (ER)			−2.332645	0.0812	−2.364853	0.0770	−2.365873	0.0770	−2.884708	0.0662
Net foreign portfolio investment (FI)					−0.00046	0.0848	−0.000461	0.0849	−0.000596	0.0818
Moderating 1 (IS*FI)							0.000688	0.7787	0.00674	0.7831
Moderating 2 (ER*FI)									0.002186	0.5280
Model	Common effect		Common effect		Common effect		Common effect		Common effect
Root MSE	0.483512		0.483182		0.48280		0.482851		0.482808
Mean dependent var	0.005776		0.005776		0.005776		0.005776		0.005776
S.D. dependent var	0.484197		0.484197		0.484197		0.484197		0.484197
Akaike info criterion	1.386309		1.385841		1.385404		1.386264		1.386981
Schwarz criterion	1.391426		1.393517		1.395638		1.399057		1.402332
Durbin-Watson stat	2.039323		2.042310		2.044917		2.044915		2.045371
R-squared	0.002382		0.003742		0.005069		0.005104		0.005282
Adjusted R-squared	0.001934		0.002848		0.003729		0.003317		0.003048
S,E, of regression	0.483728		0.483507		0.483293		0.483393		0.483458
F-statistic	5.323620		4.185658		3.783853		2.856461		2.364222
Prob (F-statistic)	0.021129		0.015332		0.010091		0.022401		0.037666

Source: Prepared by authors, 2023.

The regression equation for DUVOL in Model 5 can be written as follows: $$\mathrm{\text{DUVOL}}=0.037197\mathrm{ }‐\mathrm{ }00141196\left(\mathrm{\text{IS}}\right)‐\mathrm{ }2.884708\left(\mathrm{\text{ER}}\right)‐\mathrm{ }0.000596\left(\mathrm{\text{FI}}\right)+0.00674\left(\mathrm{\text{ISFI}}\right)+0.002186\left(\mathrm{\text{ERFI}}\right)+\mathrm{e}$$

5.4. Robustness test

According to Huang et al. (2021) state that the difference in the log of stock trading volume between the current and prior periods is used to compute the investor sentiment assessment indicator. In the previous test using the indicator of Williams %R (Zhou, 2018). Moreover, standard deviation from daily data of exchange rate dollar to local currency was used for exchange rate volatility, while another indicator, the difference in exchange rate dollar to local currency following Thursby and Thursby (1987); Bailey et al. (1986) from McKenzie (1999) article, was used in the robustness test. Using the trade volume indicator, the test findings in Tables 8, 9 and 10 demonstrate a correlation between investor sentiment and the possibility of a country index crash risk.

Table 8. CRASH (logit regression panel data).

Variable	Model 1		Model 2		Model 3		Model 4		Model 5
	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob
C	−1.387634	0.0000	−1.389170	0.0000	−1.343042	0.0000	−1.342727	0.0000	−1.340866	0.0000
Investor sentiment (IS)	0.071410	0.8141	0.067580	0.8240	0.064046	0.8315	−0.112977	0.7247	−0.110091	0.7313
Exchange rate volatility (ER)			1.154001	0.5663	1.070105	0.5941	1.252146	0.5322	−0.227010	0.9235
Net foreign portfolio investment (FI)					−0.000164	0.2486	−0.000174	0.2247	−0.000180	0.2117
Moderating 1 (IS*FI)							0.002000	0.1174	0.002170	0.0913
Moderating 2 (ER*FI)									0.007673	0.2163
McFadden R-squared	0.000025		0.000171		0.000779		0.001862		0.002554
S.D. dependent var	0.399955		0.399955		0.399955		0.399955		0.399955
Akaike info criterion	1.002075		1.002825		1.003113		1.002925		1.003129
Schwarz criterion	1.007192		1.010501		1.013347		1.015718		1.018481
LR statistic	0.055346		0.381749		1.739342		4.157734		5.702065
n	2232		2233		2234		2235		2236
H/L statistic	20.8630		8.3567		8.6983		6.3042		5.1452
Prob. chi-sq	0.0075		0.3994		0.3684		0.6132		0.7420
%Correct	80.02		80.02		80.03		80.05		80.07
%Incorrect	19.98		19.98		19.07		19.95		19.93
Obs with dep = 0	1786		1786		1786		1786		1786
Obs with dep = 1	446		446		446		446		446

Source: Prepared by authors, 2023.

Table 9. NCSKEW (panel data regression).

Variables	Model 1		Model 2		Model 3		Model 4		Model 5
	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob
C	0.016955	0.4121	0.012716	0.5368	0.035590	0.1669	0.035517	0.1678	0.035665	0.1661
Investor sentiment (IS)	−0.410839	0.0005	−0.423805	0.0003	−0.424830	0.0003	−0.457077	0.0003	−0.456557	0.0003
Exchange rate volatility (ER)			3.865814	0.0000	3.827255	0.0000	3.862904	0.0000	3.613252	0.0001
Net foreign portfolio investment (FI)					−0.0000795	0.1391	−0.0000794	0.1396	−0.0000790	0.1420
Moderating 1 (IS*FI)							0.000349	0.4794	0.000378	0.4462
Moderating 2 (ER*FI)									0.001225	0.6011
Root MSE	0.976018		0.970829		0.970352		0.970243		0.970183
Mean dependent var	0.015964		0.015964		0.015964		0.015964		0.015964
S.D. dependent var	0.978867		0.978867		0.978867		0.978867		0.978867
Akaike info criterion	2.791122		2.781354		2.781268		2.781940		2.782713
Schwarz criterion	2.796239		2.789030		2.791502		2.794732		2.798064
Durbin-Watson stat	2.104524		2.104513		2.106516		2.106037		2.104672
R-squared	0.005367		0.015916		0.016882		0.017103		0.017224
Adjusted R-squared	0.004921		0.015033		0.015559		0.015338		0.015016
S,E, of regression	0.976456		0.971482		0.971223		0.971331		0.971490
F-statistic	1.203.201		18.02560		12.75328		9.687892		7.802456
Prob(F-statistic)	0.000533		0.000000		0.000000		0.000000		0.000000

Source: Prepared by authors, 2023.

Table 10. DUVOL (panel data regression).

Variables	Model 1		Model 2		Model 3		Model 4		Model 5
	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob	Coefficient	Prob
C	0.006103	0.5512	0.005136	0.6160	0.017650	0.1682	0.017628	0.1688	0.017700	0.1671
Investor sentiment (IS)	−0.135380	0.0211	−0.138338	0.0184	−0.138899	0.0179	−0.148338	0.0184	−0.148088	0.0186
Exchange rate volatility (ER)			0.881783	0.0251	0.860689	0.0288	0.871125	0.0273	0.751004	0.1030
Net foreign portfolio investment (FI)					−0.0000435	0.1038	−0.0000435	0.1041	−0.0000433	0.1060
Moderating 1 (IS*FI)							0.000102	0.6773	0.000116	0.6378
Moderating 2 (ER*FI)									0.000589	0.6131
Root MSE	0.483512		0.482968		0.482681		0.482662		0.482634
Mean dependent var	0.005776		0.005776		0.005776		0.005776		0.005776
S.D. dependent var	0.484197		0.484197		0.484197		0.484197		0.484197
Akaike info criterion	1.386309		1.384954		1.384662		1.385480		1.386262
Schwarz criterion	1.391426		1.392630		1.394896		1.398273		1.401613
Durbin-Watson stat	2.039323		2.038383		2.040720		2.040751		2.039883
R-squared	0.002382		0.004625		0.005807		0.005884		0.005998
Adjusted R-squared	0.001934		0.003732		0.004468		0.004098		0.003765
S,E, of regression	0.483728		0.483293		0.483114		0.483204		0.483284
F-statistic	5.323620		5.178382		4.337495		3.295223		2.686441
Prob(F-statistic)	0.021129		0.005705		0.004685		0.010575		0.019898

Source: Prepared by authors, 2023.

Table 8 demonstrates how net foreign portfolio investment successfully moderates the connection between investor sentiment and country index crash risk in model 5. The findings demonstrate a negative correlation between CRASH and the independent factors in all five models. The negative the independent variables’ coefficients suggest that the likelihood of a country index crash risk decreases with increasing investor sentiment, exchange rates and net foreign portfolio investments.

Meanwhile, in Table 9, investor sentiment negatively and significantly impacts country index crash risk (Indicator: NCSKEW) in models 1 to 5. The findings demonstrate a negative correlation between investor sentiment and NCSKEW across all five models. In all models, the coefficients of net foreign portfolio investment and exchange rate are positive, indicating that the NCSKEW increases with the values of these variables.

Similarly, Table 10 demonstrates that in models 1 through 5, The likelihood of a country index crash risk is significantly correlated negatively with investor sentiment (Indicator: DUVOL). The findings indicate that investor sentiment, exchange rate and net foreign portfolio investment are significant predictors of CRASH, NCSKEW and DUVOL. Additionally, net foreign portfolio investment moderates the interaction between investor sentiment and exchange rate with country index crash risk.

H1: Investor sentiment has an effect on country index crash risk.

When investor sentiment is negative, a country’s index crash risk is more likely to happen because investor sentiment affects a country’s index crash risk negatively. This finding implies that investor sentiment has the potential to stabilize the stock market and lessen the likelihood of market crashes. Investors’ general perception of a certain securities or financial market is investor sentiment (Habibah et al., 2017). Positive investor sentiment significantly impacts the country’s index crash risk (proxied by DUVOL and NCSKEW). Stock price volatility may be accurately predicted by investor sentiment (Cevik et al., 2022). According to Rupande et al. (2019) investor sentiment and stock return volatility have a substantial association. Financial behavior can account for a considerable portion of the volatility of the Johannesburg Stock Exchange stock returns. This analysis shows a substantial negative correlation between investor sentiment and the country’s index crash risk. It demonstrates how unfavorable sentiment boosted stock market turbulence. A positive outlook among investors may lead them to take greater chances and allocate more capital to riskier investments. Because exposure to hazardous assets can result in high market volatility, raising the danger of country’s index crash risk. Alnafea and Chebbi (2022) found that the possibility of a stock market crash is increased by investor sentiment and triggers future stock crashes. This study are the same as research conducted by Anastasiou et al. (2022) that found that Investor Sentiment has a negative and significant effect on crash risk. The sentiment coefficient in sentiment-sensitive sectors is negative, indicating an average reversal of the sentiment effect (Muguto et al., 2022). Previous predictions that the relationship between Investor Sentiment is negative for the country index crash risk. However, the results of this study show the different expected sign, possibly because the sample is from 2020, where Investor sentiment can influence market volatility. If investor sentiment shows high optimism, it can lead to increased market volatility due to rapid and irrational price changes. High volatility tends to raise the possibility of market crash.

H2: Exchange rate volatility has an effect on country index crash risk

The term exchange rate volatility describes the level of ambiguity or volatility in a country’s currency’s value relative to other currencies. High exchange rate volatility might cause investors to lose in the economy, which would lower the value of the nation’s stock market index and lead to investor withdrawals. This is because high Exchange Rate Volatility often signals economic uncertainty, which can cause investors to panic and sell off their investments in the stock market. The likelihood of a stock market crash is negatively correlated with exchange rate volatility. This results align with Irani et al. (2021) The stock values of Turkish travel agencies show long-term negative impacts of exchange rates. Exchange Rate Volatility and country index crash risk have a negative and substantial link for the NCSKEW proxy. A country crash is more likely if the exchange rate declines. This study supports the findings of J. Liao et al. (2022) that the actual exchange rate has a negative and substantial coefficient. Exchange rate hazards amid more stringent global financial circumstances (Martin & Sokol, 2022). However, the results of this study show the different expected sign, possibly because by the unique characteristics of the sample or the specific context in which the research was conducted. Where the research sample consists of several countries that have price stability.

H3: Net foreign portfolio investment has an effect on the country’s index crash risk

CRASH, NCSKEW and DUVOL proxy, the Net Foreign Portfolio Investment has a negative and insignificant effect on the country index crash risk. The results of this study are the different as those of Derbali and Lamouchi (2020) in the pre-crisis period. He found that the Philippines has a negative net foreign portfolio on the average daily return on the capital market. Hong Kong, Taiwan, Singapore, Korea, Myanmar, Philippines, Thailand, India and Indonesia’s net foreign portfolios had a negative effect on the average daily return. India’s net foreign portfolio had a negative effect, It resulted in increased volatility and a decline in the domestic stock market.

H4: The higher the net foreign portfolio investment, the stronger the influence of investor sentiment on country index crash risk

When Net Foreign Portfolio Investment rises, investor sentiment has a greater influence on the likelihood of a stock market crash. The difference between the inflow and outflow of foreign investments is net foreign portfolio investment. When there is a higher Net Foreign Portfolio Investment, it indicates that there is a more significant amount of foreign capital flowing into the country’s stock market.

Positive emotion and market optimism among foreign investors might encourage more money to be invested, which would raise the market index. On the other hand, if this attitude shifts negatively, there might be an abrupt withdrawal of foreign capital and a market crash. The impact of this abrupt outflow of investment is more significant when Net Foreign Portfolio Investment is higher, which raises the possibility of a market crash.

The more the Net Foreign Portfolio Investment, as measured by CRASH, NCSKEW and DUVOL proxies, the less impact it has on investor sentiment on the likelihood of a country index crash. This research differs from Pan (2019) discovery that stock price bubbles were positively correlated with negative investor sentiment. Crash risk is caused, in part, by bubbles in stock prices. An excessive or illogical gain in the stock price does not necessarily result in a correction, which causes the stock price to crash. This study implies that foreign investment inflows cannot amplify the influence of investor attitudes on market risk. By distributing risk among different assets, diversification may reduce the market’s sensitivity to sentiment-driven swings.

H5: The higher the net foreign portfolio investment, the stronger the influence of exchange rate volatility on the country’s index crash risk

This finding emphasizes how crucial it is to consider foreign investment when analyzing the mechanics of stock market crashes. It also shows how crucial it is for policymakers to monitor and control the risks associated with significant inflows or outflows of foreign money. Increased foreign portfolio investment levels might be a factor in improving the economic climate. This stability may buffer against the adverse effects of exchange rate fluctuations on the country index crash risk. The foreign capital inflow might act as a buffer, reducing the market’s vulnerability to sudden swings brought on by exchange rate volatility.

The more the Net Foreign Portfolio Investment, as measured by CRASH, NCSKEW and DUVOL proxies, the less the impact of Exchange Rate Volatility on the country index crash risk. Meanwhile, the CRASH, NCSKEW and DUVOL proxies insignificantly affect the country crash risk index. Coleman and Tettey (2011) found that a depreciating currency rapidly affects the price index and the stock exchange’s trade volume because funds divert abroad so that foreign portfolio investors who play a significant role can also direct their investments elsewhere to other markets that are more interesting.

6. Conclusion

This study looks at the factors influencing the potential of a stock market crash in different countries. The country index crash risk is the dependent variable, and it is determined by whether the daily rate of return of the stock price index is 3.09 or less than the average standard deviation. The independent variables are the ratio of the daily return standard deviation’s natural logarithm to the negative skewness of the returns, exchange rate volatility, investor sentiment and net foreign portfolio investment. The study uses panel data regression analysis to examine the association between these characteristics and the likelihood of a country index crash risk in 31 different nations. Secondary data from a variety of sources, including the World Bank and websites of central banks.

Using three proxy indicators – CRASH, NCSKEW and DUVOL – the research findings provide insight into the complex processes impacting the country index crash risk. Investor sentiment is an important aspect that shows a negative link with the chance of a crash. It implies that a market crash is more likely when investor mood is weaker, highlighting the need to comprehend and control emotion to maintain market stability. Another significant component influencing the country index crash risk is exchange rate volatility. The inverse link suggests increased exchange rate volatility increases the possibility of a market crash. Exchange rate changes, which are a reflection of economic concerns, have the potential to incite investor fear and cause market instability. It becomes essential to monitor and control exchange rate risks to maintain stability. Additionally, the study shows that a more significant net foreign portfolio investment increases the impact of investor sentiment on the country’s index crash risk. Foreign capital inflows may amplify investor attitude’s influence on market risk. It has been discovered that a higher net foreign portfolio investment reduces the impact of exchange rate volatility and investor attitude on the likelihood of a country’s index crash risk.

The available data may restrict this study, mainly if they are incomplete or only cover a little time. Additionally, because it exclusively focuses on one market or nation, there may be less room for the findings to be applied to other markets or nations. The analysis only considers a few factors that might influence the country’s index crash risk.

Future studies should examine how other elements, such as political unpredictability, macroeconomic variables, other investment kinds, interest rates, liquidity and the PER value of each nation’s composite stock price index, affect the probability of a country’s index crash risk. In addition, use more extended periods to find market crashes from several events and conducting longitudinal studies could provide more comprehensive data and allow for the examination of long-term trends and changes.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Universitas Padjadjaran.

Notes on contributors

Lisa Kustina

Lisa Kustina, Doctoral Program in Management Science, Faculty of Economics and Business, Universitas Padjadjaran, Bandung, West Java, Indonesia.

Rachmat Sudarsono

Rachmat Sudarsono, Doctoral Program in Management Science, Faculty of Economics and Business, Universitas Padjadjaran, Bandung, West Java, Indonesia.

Nury Effendi

Nury Effendi, Doctoral Program in Management Science, Faculty of Economics and Business, Universitas Padjadjaran, Bandung, West Java, Indonesia.