EMPIRICAL TESTING OF FAMA-FRENCH ASSET PRICING MODEL IN INDONESIA STOCK EXCHANGE DURING COVID-19 PANDEMIC

The volatility of the Indonesian Stock Exchange (BEI) increased significantly during the Covid-19 pandemic period. In this period return predictability and price volatility in the stock index experienced a single structural break. There is concern among investors and academics that the asset pricing approach model that has been empirically accepted so far is unable to explain the return or excess return of an asset or investment during the Covid-19 pandemic period. This research tests the significance of the size (market capitalization), profitability, value (book-to-market), investment, and market risk premium (Rm-Rf) factors on the excess return of stock portfolios on the Indonesian Stock Exchange during the Covid-19 pandemic period. Existing studies show that the Covid-19 pandemic has affected investor sentiment, causing investors to panic and be pessimistic about their investments. In addition, there were deviations from the efficient market hypothesis during several pandemic periods in several countries so that stock prices did not fully reflect the available information. After testing, it was found that the factors size (market capitalization), profitability, value (book-to-market), investment, and market risk premium (Rm-Rf) did not have a significant influence on the excess return of stock portfolios on the Indonesia Stock Exchange during the period Covid-19 pandemic.


ABSTRACT
The volatility of the Indonesian Stock Exchange (BEI) increased significantly during the Covid-19 pandemic period.In this period return predictability and price volatility in the stock index experienced a single structural break.There is concern among investors and academics that the asset pricing approach model that has been empirically accepted so far is unable to explain the return or excess return of an asset or investment during the Covid-19 pandemic period.This research tests the significance of the size (market capitalization), profitability, value (bookto-market), investment, and market risk premium (Rm-Rf) factors on the excess return of stock portfolios on the Indonesian Stock Exchange during the Covid-19 pandemic period.Existing studies show that the Covid-19 pandemic has affected investor sentiment, causing investors to panic and be pessimistic about their investments.In addition, there were deviations from the efficient market hypothesis during several pandemic periods in several countries so that stock prices did not fully reflect the available information.After testing, it was found that the factors size (market capitalization), profitability, value (book-to-market), investment, and market risk premium (Rm-Rf) did not have a significant influence on the excess return of stock portfolios on the Indonesia Stock Exchange during the period Covid-19 pandemic.

INTRODUCTION
The Covid-19 pandemic is dynamic episode with various virus mutations and various market reactions that follow.Investor panic occurred after the emergence of various mutations of the Covid-19 virus, Delta in mid-2021, then Omicron at the end of 2021 to early 2022, as a result investors withdrew their investments on the stock exchange after witnessing an increase in the transmission of Covid-19 cases and deaths.So in 2020 and 2021, volatility in the Indonesian stock market was recorded to be very high.Rossi & Harjoto (2020) found that previous pandemics such as bird flu, SARS, swine flu, Ebola, and MERS also brought significant increases in volatility in the equity market, but Covid-19 had the strongest impact on the stock market.When comparing Covid-19 with the Great Influenza Pandemic (Spanish Flu) from 1918 to 1920 in 48 countries, it can be concluded that the impact of the Covid-19 pandemic was much greater on Gross Domestic Product (GDP), consumption and the stock market than the previous pandemic.Salisu & Akanni (2020) predicts stock returns using a fear approach.They compiled a global fear index (GFI) based on cases and deaths due to Covid-19.
They find that GFI is an effective predictor of stock returns in OECD and BRICS countries during the pandemic.An increase in the fear index causes a decrease in returns.
Apart from that, Baek et al. ( 2020) also conducted research on the impact of Covid-19 on stock market volatility and trading volume.They found that there was an increase in total risk and idiosyncratic risk due to deaths caused by Covid-19 on the stock market in the United States, while systematic risk experienced an increase and decrease based on industry.This research shows that there has been a change in stock risk (β) after the Covid-19 pandemic.
Apart from risk (β), the influence of other factors on stock returns also changes.Ramelli & Wagner (2020) examined stock returns in the United States before the pandemic and during the Covid-19 pandemic.They found that there were differences in the coefficients of profitability, book-to-market and market capitalization factors which influenced stock returns.
The changes in the influence of the factors described above are in line with the results of research by Hong et al. (2021).He found that the return predictability and volatility of the S&P 500 and DJIA indexes experienced a single structural break.A single structural break is a sudden change in the parameters of a regression model, which can cause significant forecasting errors so that the existing model is unreliable.Structural breaks can occur in extreme conditions such as the Covid-19 pandemic Cheng et al. (2022).
Based on these studies, it was found that on the United States Stock Exchange there were changes in stock beta (β) and single structural breaks during the Covid-19 period.In other words, an asset pricing model that is able to predict stock portfolio returns and excess stock portfolio returns during the prepandemic period, may not be able to predict stock portfolio returns or stock portfolio excess returns during the Covid-19 pandemic period.Likewise, vice versa, an asset pricing model that is weak in predicting stock portfolio returns and stock portfolio excess returns during the pre-pandemic period, might be used to predict stock portfolio returns and stock portfolio excess returns during the Covid-19 pandemic period.
During a financial crisis, such as during the Covid-19 pandemic, investors have a risk aversion investment strategy or a flight toward safe-haven asset classes investment strategy (Coudert & Gex, 2008).So according to A. Singh (2020), investors are paying more attention to company fundamentals in an effort to avoid the risk of falling share prices during periods of economic slowdown.
The company fundamentals that investors pay attention to are related to the internal conditions or management of a company.Even though various studies indicate that investors pay more attention to company fundamentals when investing in shares during the financial crisis, research regarding fundamental factors that influence stock portfolio returns on the Indonesian stock exchange during the Covid-19 period is still limited.Research on asset pricing models with company fundamental factors is mostly carried out on stock exchanges in developed countries.The Fama-French Five Factor asset pricing model uses many fundamental company factors.
The Efficient Market Hypothesis, introduced by Bachelier ( 2011), suggests that all information about an asset is reflected in the asset price so that it is impossible to obtain abnormal returns.Consequently, positive alpha cannot be generated using any type of analysis, neither fundamental analysis nor technical analysis.However, Grossman, S.J., Stiglitz (1980) argue that because obtaining information is expensive, investors are compensated for their efforts to gather information and discover "mispriced" assets, this information cannot be reflected in prices.This paradox is called the "Grossman Stiglitz Paradox" (Dimitrios, n.d.).
Covid-19 is an important cause of market inefficiency (Hong et al., 2021).When the economy is in bad condition, news and information will cause polarization of opinion which creates differences in investor behavior, some investors are over-reactive to news and information while some investors are under-reactive (Cujean & Hasler, 2017).Covid-19 creates better investment opportunities for investors with volatility timing abilities, especially those who have more liquidity than the general public (Hong et al., 2021).
According to Frensidy ( 2022 Covid-19 has become a symbol of new risks and concerns that are triggering anxiety among investors.However, apart from volatility and panic, stock price movements are still based on expectations of economic conditions (Wagner, 2020).In this way the public can learn about the nature of the challenges faced in these difficult times.The stock price reaction shows that various actions, including fiscal policy interventions, have the effect of avoiding further negative conditions due to the Covid-19 pandemic (Wagner, 2020) On the other hand, Engelhardt et al. (2021) argues that the amount of market volatility in reaction to Covid-19 differs between countries, depending on public trust, where volatility is lower in high trust countries.Trust in fellow citizens and in the Government are equally important factors.
In the Covid-19 period, return predictability and price volatility in the S&P 500 and DJIA stock indexes experienced a single structural break (Hong et al., 2021).
A structural break is a large change at one time in the parameters of a regression model, which can cause a very large projection error so that the model becomes unreliable.When structural break testing is carried out, it is assumed that the existing model (null hypothesis) is correct unless they find evidence to the contrary, so that it can then be concluded that the discrepancy in the results of a model is caused by a structural break (Hyndman & Athanasopoulos, 2014).
Still in the context of the crisis period, Neves et al. (2021) found that the performance of value stocks and growth stocks was different in each different period of the global financial crisis.In six countries, value stocks outperformed growth stocks in the period before the subprime crisis, and during the subprime crisis this condition continued to occur in three countries.Meanwhile, changes occurred after the crisis period.It was also found that investor sentiment has a strong significance on stock value returns and growth stock returns.
The first important asset pricing model currently used as the basis for financial theory is the Capital Asset Pricing Model (CAPM) developed by Sharpe (1964).
CAPM is a single factor model, where the only price factor is the market risk premium.This indicates that there is a positive relationship between a stock's beta and the stock's expected return.CAPM helps calculate investment risk and  (1997) found that more characteristics of factor loadings determine excess returns.Their results also show that stock values move because of investors' sensitivity to similar factors and not because of unique factors.The 3-factor model explains the value premium better than Daniel and Titman's (1997) characteristic model, in their 68 year period and there is no evidence to contradict the fact that value loadings determine expected returns.Fama and French believe that the evidence from Daniel & Titman (1997) in favor of the characteristics model is due to the short sample period.If they omitted the period examined by Daniel & Titman (1997), the regression intercept could barely approach the zero intercept.Carhart (1997) extended Fama and French's (1993) three-factor model to a fourfactor model including a momentum factor, in addition to size, value and market factors.It appears that Carhart's model explains more variation in average stock returns than Fama and French's (1993) original three-factor model.Blackburn & Cakici (2017) focused on conducting research on momentum factors and examining returns in various capital markets in developed countries.This then resulted in the discovery that returns were significant using a long strategy for long-term losers and short positions for short-term winners.These results were valid for the entire sample period and most markets.Griffin (2002) examines different versions of Fama and French's three-factor model in an international data set.He found that no model truly captured variations in average returns.However, he found that research on the threefactor Fama-French model using domestic data produced better performance than the three-factor version of the Fama-French model using international stock market data.In its dataset, Griffin has data on 23 international markets divided into four regions, North America, Asia Pacific, Europe and Japan.Griffin conducted integrated asset pricing model research in these four regions.
Novy-Marx ( 2013) identified profitability factors.They found that shares of companies with high profitability generated significantly higher returns than shares of companies with high profitability.In research, Watanabe et al. (2013) examines whether the value effect in international stock markets is consistent with results in the United States and evaluates possible economic causes of value factors.They found that the value effect existed in international stock markets and that there were large differences in this effect across the countries they studied.The value effect has a stronger impact in markets that have more efficient information.
After successfully finding a five-factor model, which explains size, B/M, profitability, and investment patterns, Fama & French (1997) tried the model internationally and they found that stock returns averaged three of the four regions they used (America North, Europe and Asia Pacific) increases as the B/M ratio and profitability increase.They also find the expected negative relationship between returns and investment.In Japan, the relationship between average investment returns is weak but the relationship between average returns and book-to-market ratio is strong.
In China, this model has also been tested.Journal entitled the five-factor asset pricing model, short-term reversal, and ownership structure -the case of China by Chen et al. (2022).The sample period is from January 2004 to December 2017.
The results of the study found that the Fama and French Five-Factor Model overall better explains excess returns than the Fama and French Three-Factor Model.
In India, similar research was also conducted by K. & French (2017) where the model has different performance from one region to another (Dimitrios, n.d.).This empirical test will examine whether the Fama -French Five Factor Model can explain the average return of portfolios prepared using large spreads on size, value, profitability and investment.The large spread referred to is that the independent variable is divided into three groups based on its size, with the value from group three reducing the value from group one.For example, SMB is arranged based on market cap, where the return from stocks with the largest capitalization (group three) is reduced by the return from stocks with the smallest capitalization (group one).

METHOD
This research aims to explain the factors that have a significant influence on the excess return of stock portfolios using the Fama -French Five Factor Model.
To see the individual significance of each Rm-Rf, SMB, HML, RMW, and CMA factor, the method used in this research is ordinary least squares (OLS).This Following research conducted by Fama and French (2015) and Ekaputra ( 2016 and B/M are as follows: For each region, the market capitalization size breakpoints are the 1st, 2nd, 3rd, 4th, and 5th quantiles of the aggregate market capitalization of all shares that was researched.For the book-to-market (B/M), operating profitability (OP), and investment (Inv) factors, use the 1st, 2nd, 3rd, 4th, and 5th quantiles of the aggregate value of each factor.To build the dependent factor, the excess return variable from period t is used.The book-tomarket ratio, operating profitability, and investment values use data at the end of period t-1.

Independent variable, right hand side
In this research, the independent variables used are market factor, size factor, value factor, profitability factor, and investment factor.Market factor is proxied by Rm -Rf, size factor is proxied by SMB, meanwhile value factor is proxied by HML, profitability factor is proxied by RMW, and investment factor is proxied by CMA in this research due to limited research time.Following research conducted by Fama and French (2015) and Ekaputra ( 2016), the calculation of independent factors in this model uses 2 x 3 portfolio sorting.A more detailed explanation for each independent variable in this research model is explained as follows: 1.The stock data studied is grouped by period, then in each period it is sorted into two capitalization categories, market capitalization size and into three categories for each book-to-market equity (B/M), operating profitability (OP), and investment ( Inv).
2. The breakpoint for market capitalization size is the median of the aggregate market capitalization of shares in one period, while the breakpoints for B/M, OP, and Inv are the 30th and 70th percentiles.Stocks with large market capitalization are stocks that are above the median IDX capitalization value in a period, while small stocks are below the median stock.
3. To build independent factors, variables from the fiscal year ending in year t-1 are used.The Rm-Rf, book-to-market ratio, operating profitability, and investment values use data at the end of period t-1, with the following explanation: e. CMA is a proxy for the company's investment factors.The data used is the quarterly increase in total assets, namely the growth in total assets at the end of the previous period (t-1) divided by total assets at the end of the two previous periods (t-2).
The summary of the formula for the independent variables is in Table 1 below:

RESULT Descriptive Statistics
In Table 2, panel A provides information on the average return for each factor.
The average monthly market return is 0.02%.On a monthly basis, the average excess return of the size factor (SMB) was 0.63%, implying an average premium of 0.63% for buying small-cap stocks over large-cap stocks.Meanwhile the average monthly excess return of the book-to-market ratio (HML) factor is 1.23%, implying an average premium of 1.23% for buying large book-to-market stocks over stocks with small book-to-market.Meanwhile, the average monthly return on the profitability factor (RMW) is 1.82%, meaning an average premium of 1.82% for buying shares of companies with strong profitability rather than shares of companies with weak profitability.Meanwhile, the average excess return from investment factors (CMA) is -2.62%, indicating that shares of companies with aggressive investment produce a higher rate of return of 2.26% than shares of companies with conservative investment.This is in line with the findings of Sutrisno & Ekaputra (2016) where all factors produce positive monthly averages except the investment factor (CMA), where investing in shares of companies that are conservative in investing produces lower returns than investing in companies that are aggressive in investing.Panel B divides the small and large components of the 2 x 3 factor.The value premium between large capitalization stocks and small capitalization stocks tends to be the same.Meanwhile, the profitability premium is lower for large shares, RMW B -2.37% compared to RMW S, 1.27% for small shares.Lower investments result in better returns on small caps CMA S= -2.12% compared.
CMA B = -3.11 on large stocks.This RMW pattern is the same as the research results of Dimitrios (2020).Meanwhile, the investment premium is negative for  both large capitalization stocks and small capitalization stocks.This indicates that the excess return on the portfolio of shares of companies that invest aggressively is higher than the excess return on the portfolio of shares of companies that invest conservatively.
Meanwhile, the t-statistic for HMLS and HMLB is positive.Meanwhile, RMWS is positive and RMWB is negative, indicating differences in the direction of influence.Meanwhile, CMAS and CMAB are negative, meaning they have a negative relationship with the average stock return.
Panel C shows the correlations between the independent variables.There is a positive correlation between RMW and HML, between CMA and SMB, and between Rm-Rf and CMA.Meanwhile there is a negative correlation between SMB and HML, SMB and RMW, SMB with Rm-Rf, HML with CMA, HML with Rm-Rf, RMW with CMA, and RMW with Rm-Rf.The correlation between RMW and CMA -0.8 means that there is a relationship between the profitability premium (RMW) and the investment premium (CMA).
The average monthly portfolio excess return pattern of the 25 Size-B/M, 25 Size-OP, and 25 Size-Inv portfolios can be seen in table 2. The excess return in table 2 is the dependent variable in the Fama -French Five Factor Model.In Panel A is a stock portfolio arranged based on size (market capitalization) and value (bookto-market).Vertically from small, 2, 3, 4, big is the order of portfolios with the smallest market capitalization to the largest.Meanwhile, horizontally, low, 2, 3, 4, high is the portfolio sequence from lowest to highest book-to-market.In Panel B is a stock portfolio arranged based on size (market capitalization) and profitability (operating profit).Vertically from small, 2, 3, 4, big is the order of portfolios with the smallest market capitalization to the largest.Meanwhile horizontally, weak, 2, 3, 4, robust is the order of portfolios with the lowest to the highest operating profit.In Panel C is a stock portfolio arranged based on size (market capitalization) and investment.Vertically from small, 2, 3, 4, big is the order of portfolios with the smallest market capitalization to the largest.
Meanwhile horizontally, conservative, 2, 3, 4, aggressive are the order of portfolios with lowest to highest investment.

Regression Results
In the regression results table below it is divided into two parts, on the left there are coefficients of the independent variables indicated by the letters a, s, h, r, c, and b.Meanwhile on the right is the t-statistic of the coefficients of the independent variables.
), although it has more or less the same impact on global finance, the crisis caused by Covid-19 is different from the previous global financial crisis in 2007-2008.During the Covid-19 crisis, the JCI continued to decline for almost three weeks, while in 1998 and 2008 there were not many lower auto rejects (ARBs) compared to 2020, where the four banks with the largest capitalization (BBCA, BBRI, BMRI and BBNI) experienced ARBs.There is no bid volume at the same time.The Covid-19 crisis has had a broad impact, affecting almost all sectors.
investment.Empirically, CAPM fails to explain abnormal stock returns, but is still used as a method for assessing the cost of capital and as a portfolio performance evaluation technique.Criticism of the CAPM is usually caused by the simplicity of the model and the linear relationship between systematic risk and the expected return of a stock.Ross (1976) proposed an alternative, The Arbitrage Pricing Theory (APT), through a multi-factor asset pricing model, showing that there is a linear relationship between the expected return of a stock and a number of macroeconomic variables.Fama & French (1993), tested the CAPM which then produced a new model known as the three-factor model.This model includes two additional market factors that can explain stock excess returns, namely the market capitalization size factor and the company's book-to-market (B/M) ratio.Fama and French found that the three-factor model is a good model for predicting portfolio excess returns.Daniel & Titman (1997) criticizedFama & French (1993) research and suggested the Characteristics Model.Fama and French show that cross-sectional variations in excess returns can only be explained by size and value factors.Daniel & Titman

For
this research period, the object of research is the stock portfolio on the Indonesian Stock Exchange during the Covid-19 pandemic period starting from the first quarter of 2020 to the fourth quarter of 2022.The portfolio used in this research comes from shares listed on the Indonesian Stock Exchange (BEI ) during the study period.The financial data used in this research is in Rupiah.The data used is panel secondary data taken via Thomson Reuters Datastream.This research uses the entire population of shares on the IDX for the period first quarter 2020 -fourth quarter 2022.This research followsFama & French (2015) andSutrisno & Ekaputra (2016) in terms of data collection criteria.The criteria are: (1) do not include financial stocks, (2) shares of the company under study must have data on operating profit, book-to-market and fixed assets; and (3) does not include shares with negative share capital (equity).Based on these criteria, a sample size of around 667 company shares was obtained.Stock data for each period will be updated every quarter.

1 Tahun 2024 DOI: http://dx.doi.org/10.31000/dmj.v8i1 ISSN (Online) 2580-2127
French model on the Indonesia Stock Exchange during the period 2000 to 2015.It was found that the Five-Factor Model had a better ability to explain the excess returns of stock portfolios on the Indonesia Stock Exchange than the threefactor model.However, in this study, investment and profitability factors had a weak influence on portfolio excess returns.From existing research, there are variations in the significance of the influence of independent variables in different geographic regions and this is proven in Fama Sutrisno & Ekaputra (2016)e title Testing Factor Models In An Emerging Market: Evidence From India with the conclusion that the five-factor model has better power to explain stock returns than the three-factor model.Fama & French (2015)tested the five factor model internationally in four regions, namely the United States, Asia Pacific, Europe and Japan in the period July 1990 to September 2014.They found that the average equity return in North America, Asia Pacific , and Europe as profitability and book-to-market increase and are negatively correlated with investment.Meanwhile for the Japanese case, the relationship between average returns and market equity ratios is strong in all size groups, but the relationship between average returns and profitability or investment is weaker.Meanwhile,Sutrisno & Ekaputra (2016)tested the performance of the five-factor Fama -

Dynamic Management Journal Volume 8 No. 1 Tahun 2024 DOI: http://dx.doi.org/10.31000/dmj.v8i1 ISSN (Online) 2580-2127 study
assessed the explanatory power with a t-test of the five-factor model.This research also tested the average adjusted R2 in the model to test the significance of the model in explaining variations in stock portfolio returns on the Indonesian stock exchange during the Covid-19 pandemic.An asset pricing model with a larger average adjusted R2 value indicates that the model is better.The equations tested are as follows: Rit -Rft = ap + bp (Rmt -Rft) + sp SMBt + hpHMLt + rpRMWt + cp CMAt + ept

Dependent variable, left hand side In
this research, the dependent variable used is excess return (Rit -Rft).In this equation, Rit is the return on security or portfolio i for period t, while Rft is the risk-free return.Where Rit is the closing price of the Indonesian Stock Exchange at the end of each quarter, and Rft is the BI-7 Day Reverse Repo Rate (BI7DRR).

Table 4 Regression for 25 Portfolios based on Size-B/M
In each sub table there are 5 x 5 rows and columns which show the regression results for each portfolio.Where for each row the first is the regression result from the smallest size stock portfolio and the fifth row is the regression result from the largest size stock portfolio.Meanwhile, the leftmost column of each sub table is the regression result of the stock portfolio with the lowest book-to-market and the rightmost column is the regression result of the stock portfolio with the highest book-to-market.