INTRODUCTION
REASON TO RESEARCH
The stock market is a crucial component of the national economy, where both institutional and individual investors aim to achieve the highest possible returns However, many lack effective methods to identify which stocks will yield the most profit Selecting the right stocks for your portfolio can be likened to soccer betting, where informed analysis of a team's skills, injury rates, and line-ups can lead to better predictions of outcomes Understanding the factors that influence stock returns is essential for making sound investment decisions, similar to how a professional bettor must grasp the dynamics of a football game Stock prices continuously adjust to reflect their market value, necessitating a thorough analysis of price volatility, risk, past performance, and potential future outcomes Thus, determining the returns of a portfolio is a critical step that investors must undertake before committing their funds.
Over the last century, numerous pricing models have emerged, beginning in the mid-1900s with the Capital Asset Pricing Model (CAPM) developed by Sharpe, Lintner, and Mossin, which measures stock returns based solely on market risk Despite its widespread use, CAPM has faced significant scrutiny, particularly in the Indian market, as highlighted by Basu's findings in 1977, and its misspecification was demonstrated by Banz in 1981 In contrast, the Arbitrage Pricing Theory (APT) introduces a multi-factor approach, allowing arbitrageurs to exploit market inefficiencies, although it does not guarantee risk-free profits Eugene Fama and Kenneth French further advanced asset pricing by introducing the three-factor model, which considers size and value factors alongside beta, gaining traction in global markets during the 1980s Subsequent research, including tests by Maroney and Protopapadakis across multiple countries, confirmed the significance of these factors In 1997, Mark Carhart expanded upon this by adding a momentum factor, resulting in a four-factor model widely used for mutual fund evaluation Further studies, such as those by Novy-Marx and Aharoni et al., revealed strong links between profits, investment, and stock returns, prompting Fama and French to develop the five-factor model in 2015, which includes Profit and Investment factors This model has garnered attention across 23 developed markets, particularly in North America, and is increasingly relevant in the Vietnamese stock market, although research remains focused primarily on the three-factor model, leaving the five-factor model's implications still under exploration.
This article aims to assess the application of Fama-French factors within the context of industrial companies in the Vietnam stock market The author anticipates that the findings will enhance understanding of the model's integration in Vietnam, ultimately assisting investors in optimizing their value in the stock market.
RESEARCH GOAL
This research aims to assess the impact of the Fama French five-factor model—comprising market, size, value, profit, and investment factors—on the expected returns of listed industrial stocks in the Vietnam stock market, specifically from HSX and HNX Utilizing analytical tools, the study seeks to clearly illustrate the fluctuations in average returns of these stocks Ultimately, the findings will provide valuable recommendations for investors, policymakers, and stakeholders to enhance portfolio selection and management.
RESEARCH QUESTIONS
To fulfill the above research‘s goal, this thesis‘s important mission is to answer these following questions:
- Do the factors: market premium, size, book-to-market ratio, profitability, and investment risk affect the returns and positive or negative correlation to the returns of the portfolio?
- The asset pricing model Fama French five-factor model is suitable model for explaining the change of the returns in Viet Nam stock market?
- How investors use research results to increase profits and decrease risks?
RESEARCH SUBJECT AND RANGE
Research subject: using factors of Fama French five-factor pricing model with 100 industrial companies on HSX and HNX stock market
- The time frame for the study is from January 2012 to January 2018 Prioritizing the objective to create an accurate analysis, any earlier return data is disregarded in this study
The sample companies must be primarily engaged in the industrial sector and classified as equity, with accessible data on stock price (market value), book value, total assets, total liabilities, shares outstanding, return on equity (ROE), and the three-month Treasury bill rate from the State Bank of Vietnam during the sample period.
This study analyzes data from publicly listed industrial companies on the HSX and HNX stock markets, specifically excluding financial sector entities such as banks, insurance companies, securities firms, and investment funds.
METHODLOGY
With the research goal is to test the Fama French five-factor asset pricing model in industrial companies of Vietnam stock market, the thesis applies mainly quantitative method:
- Use the Ordinary Least Square (OLS) method to measure the betas, check correlations between factors thus portfolios
- Use Gibbons, Ross, and Shanken (1989) GRS validation test to check the efficiency of the model whether itself has power on explanatory to listed companies below
- Use Excel Office to synthesize data and calculations then finally use Stata version 13 to run Regression models and other relevant tests
The expected return on asset i is determined by the risk-free rate of Treasury bills, along with several factors including the excess market return, the size factor (Small minus Big), the value factor (High minus Low), and the profitability factor (Robust minus Weakness).
(Conservative minus Aggressive) the investment factor The coefficients is the asset‘s sensibility, the intercept and the error term of asset i at time t.
RESEARCH CONTRIBUTION
The thesis has following contributions: xii
This thesis investigates the effectiveness of the Fama-French five-factor pricing model, aiming to clarify its components for investors and researchers seeking to predict future income rates while minimizing immediate risks The findings can be directly applied to the Vietnam stock market, enhancing investment strategies in the region.
Experimentally, by evaluating the effectiveness of the models, giving a review of the research results, thereby providing some hints to investors and individuals when selecting and managing the portfolio.
RESEARCH OUTLINE
This chapter will give an overview on reasons of the author decide to work with this topic, research goals, research subject, research range and research contribution.
LITERATURE REVIEW AND PREVIOUS RESEARCHES
Literature review
The industrial sector constitutes a significant portion of the economy, focusing on the production of material goods that are manufactured, processed, and refined to meet consumer demand and support business activities essential for daily life This large-scale economic activity is driven by advanced technology, scientific innovation, and well-developed mechanisms.
In Vietnam, the industrial sector encompasses a diverse range of companies involved in mining minerals, coal, and petroleum, as well as processing and manufacturing food products and materials Key industries also include the production of capital goods, transportation, and the generation and distribution of electricity, gas, and water Additionally, the garment industry, household appliances manufacturing, chemical processing, and various commercial and professional services play significant roles in the country's industrial landscape.
His innovative work constructed the establishment of what is now well-known as
Modern Portfolio Theory (MPT), developed by Markowitz, emphasizes the significance of the number of securities in a portfolio and their covariance relationships in enhancing portfolio performance The Markowitz curve illustrates all optimal portfolios, showcasing the balance between minimizing risk and maximizing returns Central to this model is the challenge of determining the appropriate weight for each asset to achieve the best risk-return trade-off Key contributions of the Markowitz model include the anticipated rate of return, the standard deviation of assets, and the correlation coefficients among firms in the portfolio, all of which are crucial for constructing an effective portfolio While not an asset pricing model, MPT serves as a framework for selecting the most advantageous portfolio Investors assess potential returns by calculating the expected value of returns, using probability distributions, while risk is measured by the variability of these returns, commonly represented by variance and standard deviation.
Given any set of risky assets and a set of weights that describe how the portfolio investment is separated, the general formulas of expected return for n firms is:
∑ 𝑤 is 1.0 is the return on security i and portfolio p
𝑤 is the proportion of the funds invested in security i
The difference of a portfolio mix with firms is equivalent to the weighted average covariance of the returns on its individual firms:
∑ ∑ 𝑤 𝑤 (2) Where: is correlation coefficient between the rates of return on firm or portfolio i,r i , i , and j are standard deviations of respectively
Markowitz developed modern portfolio theory by focusing on optimization through the concepts of risk, return, variance, and covariance He noted that investors typically choose from a set of Pareto optimal risk-return combinations, given the dual criteria of risk and return His mean-variance portfolio hypothesis laid the groundwork for the capital asset pricing model, which has become a crucial aspect of investment management practices.
The MPT model aims to enhance realism in financial research by incorporating factors like transaction costs, as highlighted in S Uryasev's 1999 study Additionally, alternative risk measurement approaches have been suggested, moving beyond the traditional use of standard deviation of stock returns However, Kroll, Levy, and Markowitz (1984) noted that many practitioners remain skeptical about the effectiveness of standard deviation as a reliable risk measure.
2.1.3 CAPM (Capital Asset Pricing Model)
William Sharpe developed the Capital Asset Pricing Model (CAPM) in 1964, building on Harry Markowitz's theory from 1952 CAPM illustrates the relationship between systematic risk and expected returns for financial assets, especially stocks Widely used in finance, it helps price risky firms and determine expected returns based on asset risk and capital costs The model posits that a stock's expected return is influenced by the market's expected return, demonstrating a linear and positive relationship CAPM aims to evaluate whether a stock is fairly valued by comparing its risk and time value of money against its expected return Its appeal lies in providing straightforward predictions about risk and the correlation between expected return and risk To quantify a stock's sensitivity to market returns, CAPM employs beta (β) as an indicator of systematic risk.
In the 1990s, the Capital Asset Pricing Model (CAPM) emerged as a key framework for understanding firm pricing in the market, gaining significant recognition across the industry The model's formula highlights its importance in financial analysis and investment decision-making.
( ) (2) Where: is the expected return of asset i is the risk-free rate
( ) is the market risk premium xvii is the market risk factor is the beta of the asset i
Beta is a key metric in estimating stock risk based on value instability, with calculations using a broad market index explaining about 70% of stock return volatility However, price movements carry different levels of risk, and the timeframe for assessing a stock's volatility is not standardized due to the non-normal distribution of stock returns The Capital Asset Pricing Model (CAPM) assumes a constant risk-free rate, but fluctuations in this rate can inflate capital values, potentially leading to overvaluation of stocks Additionally, the theoretical market portfolio used to determine market risk premiums cannot be directly invested in, leading investors to rely on major stock indices like the VN-Index as imperfect substitutes for the market.
A study by Chen (2003) in Taiwan indicates that the Capital Asset Pricing Model (CAPM) demonstrates a significant relationship between stock returns and beta, with a high coefficient of determination across seven industry sub-sectors, particularly noting that market beta can explain nearly 50% of equity return movements in the textiles sector Elmo (2018) highlighted that the CAPM is effective in explaining market behavior over specific time periods, particularly in Brazilian sustainability companies, where larger portfolios yield higher average returns However, despite its simplicity, the model struggles to accurately explain realized returns Dempsey (2013) cautioned that recent factor models lack the risk-return foundation of the CAPM, suggesting they may ultimately fail, while Melody Nyangara (2016) advised analysts and investors to apply the CAPM with caution.
In 1976, Ross introduced the Arbitrage Pricing Theory (APT), a groundbreaking approach to asset pricing that extends beyond existing theories APT is utilized in trading stocks, commodities, and foreign currencies across markets for arbitrage opportunities This general theory posits that the expected return of a financial asset can be modeled as a linear function of multiple macroeconomic factors or theoretical market indices, with each factor's sensitivity represented by a specific beta coefficient.
The Arbitrage Pricing Theory (APT) is a general framework for understanding expected returns on financial assets rather than a specific model It is based on two key assumptions: first, that investors can borrow or lend at a uniform risk-free rate, regardless of the amount, and second, that all investors have similar expectations, leading to a consensus on the probability distributions of future returns Stephen Ross posited that non-systematic risks can be mitigated through portfolio diversification, meaning that risk compensation applies primarily to systematic risks Systematic risk factors associated with APT include inflation, business cycles, economic growth (GNP), interest rate differentials, and exchange rates Multi-factor models are categorized into three types: macroeconomic models, which link returns to factors like employment and inflation; fundamental models, which analyze the relationship between returns and financial metrics such as earnings; and statistical models, which assess firm returns based on measurable performance The beta coefficient quantifies a firm's systemic risk relative to the overall market, with a beta of 1 indicating equal volatility to the market, a beta greater than 1 indicating higher volatility, and a beta less than 1 indicating lower volatility The expected return of a stock is thus influenced by various systematic risk factors.
The expected return on asset i can be determined by the risk-free interest rate in government bonds, the asset's beta which measures sensitivity to various risk factors, and the risk premium associated with those factors This relationship can be expressed through the formula that incorporates multiple risk factors, denoted as k, where k represents the number of factors influencing the asset's return.
The Arbitrage Pricing Theory (APT) and the Capital Asset Pricing Model (CAPM) differ significantly in their approaches to asset pricing While CAPM is a simpler, single-factor model relying on historical data to estimate the return-beta relationship, APT is a more complex, multi-factor model that does not require a market portfolio and allows for individual stock mispricing, making it applicable only to well-diversified portfolios APT can be adapted into various multifactor models, but its complexity poses challenges in identifying the correct risk factors Although CAPM is often deemed more reliable due to its straightforward calculations, numerous studies have critiqued its assumptions, leading to the development of alternative models Ultimately, APT incorporates a range of macroeconomic elements to interpret expected returns, positioning it as a forward-looking framework, despite its limitations in determining the appropriate variables for risk assessment.
The Arbitrage Pricing Theory (APT) developed by Akwimbi William in 2003 effectively explains expected returns in the Nigerian Stock Exchange (NSE), highlighting return indices as key variables in the time series of returns However, incorporating fundamental variables significantly enhances the understanding of these returns Additionally, Stefan Robert's study, "Empirical Testing of CPM and PT Models," emphasizes the value of factor analysis as a novel tool for testing the APT Despite the importance of market return, the complexities of securities' returns on the Bucharest Stock Exchange cannot be fully captured by a single factor.
2.1.5 The Fama French three-factor model
The Capital Asset Pricing Model (CAPM) has struggled to establish itself as a reliable historical asset pricing model, leading to the development of alternative theories by Eugene Fama and Kenneth French Their 1992 research demonstrated that returns are influenced not only by market risk but also by additional factors Petros Messis (2006) found that the Fama-French model significantly outperforms the Arbitrage Pricing Theory (APT) in time-series regressions, while APT shows slight superiority in cross-sectional analyses Building on the work of William Sharpe and Stephen Ross, the Fama-French three-factor model emerged, incorporating new factors derived from the CAPM and APT, specifically firm size (SMB), book-to-market ratio (HML), and stock beta Fama and MacBeth's findings from 1963 to 1990 indicated a weak relationship between beta and average stock returns, prompting further exploration for variables that could more effectively explain stock returns beyond the single market beta.
Previous researches
2.2.1 Previous researches from developed countries
Research in US of Zhu (2016):
Zhu enhanced the Fama-French five-factor model by incorporating the non-Normal error distribution of the Standardized Standard Asymmetric Exponential Power Distribution (SSAEPD) and integrating GARCH-type volatility, presenting a novel approach to asset pricing.
The article examines a new factor model based on SSAEPD error and GARCH-type volatility, aiming to determine if it can surpass the Fama-French 5-factor model Utilizing monthly U.S stock return data from July 1963 to December 2013, the empirical findings reveal that the Fama-French 5 factors maintain their positive relevance, even when accounting for GARCH-type volatilities and non-normal errors This research contributes valuable insights to the existing asset pricing literature and serves as a useful reference for investors.
Research in Australia of Chiah and partners (2016)
The study applied a five-factor Fama French model to analyze Australian companies over a 31-year period from 1982 to 2013, utilizing three portfolio types: 5x5 Size-BE/ME, 5x5 Size-OP, and 5x5 Size-Inv Findings indicate a correlation between market and size factors with rate of return, while the value, profit, and investment factors show a two-way impact that varies by portfolio Overall, in developed countries, the five-factor Fama French model provides superior explanatory power compared to the three-factor Fama French model and the CAPM model.
Research in Japan of Keiichi Kubota and partner (2017)
Keiichi Kubota and Hitoshi Takehara analyzed the pricing structure of stocks in Japan from 1978 to 2014, concluding that the Fama and French five-factor model is not the optimal benchmark for Japanese data Their findings indicate that the RMW (Robust minus Weak) and CMA (Conservative minus Aggressive) factors lack strong explanatory power, as demonstrated by generalized GMM tests using the Hansen–Jagannathan distance measure The Gibbons–Ross–Shaken test revealed a minimum test statistic of 9.447 for the five-factor model, while the simpler three-factor model performed comparably with a statistic of 9.491 Additionally, the four-factor model's statistic was higher at 9.988, closely resembling the CAPM's 10.064 Despite the four and five-factor models being suitable for the US market, their performance is inferior to that of the three-factor model when applied to Japanese data, highlighting the effectiveness of simpler models and the significance of their coefficients.
2.2.2 Previous researches in developing countries
Research in China of Grace Xing Hu and partners (2018)
This study highlights a significant relationship between stock returns and firm size in the Chinese market from 1990 to 2016, revealing that small stocks consistently outperform large stocks A long-short portfolio strategy that involves going long on the smallest stocks and shorting the largest yields a variance-adjusted average return of 1.23%, which is statistically significant Following the Fama–French methodology, the SMB factor shows a variance-adjusted average return of 0.61% per month, indicating both economic and statistical significance In contrast, the average returns of stocks do not demonstrate a clear correlation with their book-to-market (B/M) ratios, as evidenced by the HML factor, which provides a positive but statistically insignificant variance-adjusted average return of 0.23% per month Additionally, the market factor does not exhibit a significant premium, while SMB consistently outperforms both the market and HML factors in time series regressions and Fama French cross-sectional tests Among the three factors analyzed, SMB emerges as the most crucial for capturing cross-sectional variations in Chinese stock returns, with the findings suggesting that early market volatility and the limited number of stocks significantly influenced results, which diminish in significance over a longer sample period with actual volatility adjustments.
Research in India of Harshita and partners (2015)
This study employs hierarchical multiple regression analysis on data from companies in the CNX 500 index over a fifteen-year period, from October 1999 to September 2014 The findings indicate an inverse relationship between market capitalization, profitability, and investment with returns, while the book-to-market (B/M) ratio shows a direct relationship with returns Additionally, the three-factor asset pricing model by Fama and French (1993) outperforms the capital asset pricing model (CAPM) across all portfolios Furthermore, the five-factor asset pricing model from Fama and French (2015) surpasses the three-factor model when portfolios are categorized by profitability and investment Notably, the four-factor model without a specific factor demonstrates the highest explanatory power for portfolios not based on investment, whereas the five-factor model excels for investment-based portfolios.
Research in Turkey of Songul Kakilli Acaravci and partner (2017)
The study evaluates the effectiveness of the Fama-French five-factor model in Borsa Istanbul (BIST) over a 132-month period from July 2005 to June 2016, utilizing excess returns from 14 intersection portfolios based on size, B/M ratio, profitability, and investment factors The GRS-F test yielded a result of 1.00 (P 0.45), leading to the acceptance of the null hypothesis, indicating no pricing errors within the model Consequently, the five-factor model is deemed valid in the BIST, effectively explaining variations in excess portfolio returns, with an average explanatory power of 0.33.
Research of Truong Dong Loc and Duong Thi Hoang Trang (2014)
This study offers empirical support for the Fama-French three-factor model's applicability to the HOSE stock market, utilizing data from January 2006 to December 2012 The findings reveal a positive correlation between the profitability of listed companies on HOSE and market risk, company size, and the book-to-market (B/M) ratio Additionally, the market factor significantly influences the profitability across all six portfolios analyzed The size factor shows a strong positive correlation with the profitability of small companies (S), while exhibiting a negative correlation with the returns of large companies (B) Lastly, the HML factor is positively associated with high (H) and medium (M) B/M ratio portfolios, but negatively correlated with low (L) B/M ratio portfolios.
According to the results, we can confirm that the Fama-French three-factor model is appropriate in explaining the change of profitability listed on HOSE
Research of Vo Hong Duc và Mai Duy Tan (2014)
This study evaluates the Fama-French three-factor and five-factor models using data from 281 companies listed on the Ho Chi Minh City Stock Exchange, excluding financial institutions, from January 2007 to December 2015 The findings reveal that the market factor has the strongest and most consistent impact in the three-factor model, while the value factor offers better explanations but adds complexity to the model In the five-factor model, the market factor consistently shows a positive expectation, although it initially appears negative and statistically significant The size factor is positive, and while the value factor provides better explanations, it lacks statistical significance in some portfolios Among the profitability and investment factors, profitability emerges as the most influential Ultimately, the investment factor of the Fama-French five-factor model does not fully account for expected returns in the Vietnamese stock market.
Research of Nguyen Thi Thuy Nhi (2016)
This study analyzes the Fama French five-factor model and the Q-factor model by Hou, utilizing data from the HOSE and HNX stock exchanges from January 2009 to June 2015, and employs three methods for portfolio division The findings indicate that the market factor positively influences the model, while the SMB factor shows a positive correlation with small-cap stock portfolios and a negative correlation with large-cap portfolios Additionally, the HML factor is positively associated with high book-to-market portfolios, the RMW factor exhibits a positive relationship with high return on equity, and the CMA factor is positive for portfolios with low operating profits Overall, the model's explanatory power increases from 80% to 96%, leading to the conclusion that the Fama French five-factor model outperforms the Q-factor model in explaining stock returns.
Research of Huynh Ngoc Minh Tram (2017)
The study reveals that the SMB factor, along with the MRP market return factor, significantly explains the expected return of stocks, with coefficient estimates being statistically significant at the 5% level Notably, only the SMB and MRP factors are expected to remain positive, while the HML factor is nearly negative, indicating that companies with smaller sizes and lower book-to-market ratios tend to yield higher returns Conversely, other factors such as RMW and CMA are not statistically significant, suggesting that the Fama-French five-factor model does not fully capture the rate of return in the Vietnam stock market Nonetheless, MRP, SMB, and HML factors show a positive relationship with stock profits in Vietnam.
DATA AND METHODOLOGY
Data construction and processing method
This study focuses on industrial companies listed on HOSE and HNX over a six-year period from January 2012 to January 2018, utilizing quarterly data to ensure comprehensive public information for all selected companies Following the methodology of Ferson and Locke (1998), which indicates that historical averages enhance market return forecasts, the research calculates expected factor returns based on the average factor return throughout the entire data period prior to the forecast Ultimately, the analysis identifies a total of 100 industrial companies that meet the established criteria across the 24 quarters examined.
According to Trinh Minh Quang (2017), the study utilizes data from Thomson Reuters, with controlling and independent variables sourced from company annual reports and financial statements, as well as stock prices from Vietnam's official Ho Chi Minh and Hanoi stock exchange websites The primary objective of this research is to evaluate the Fama-French five-factor asset pricing model within the context of listed industrial companies on the Vietnam stock market.
Data on key model variables such as total outstanding shares, total assets, net profit after tax, stock closing prices, book value, market value, VN-Index, and Treasury bill rates are sourced from the Reserve Bank of Vietnam website, with a focus on quarterly and daily metrics over a three-month term.
During the initial phase of research, a significant volume of raw data from industrial companies was downloaded, including dead stock information and various missing data types To ensure data accuracy and reliability, it is essential to address these issues through a systematic four-step process for data cleansing and error removal.
Using the database from Step 1, I systematically calculate the rate of return for each stock, the market portfolio's return, the value factor through the B/M ratio, the size factor by multiplying market equity (ME) with total outstanding shares, the profitability factor via Return on Equity (ROE), the investment trend factor by assessing total asset growth, and the risk-free interest rate derived from Vietnamese Treasury bills.
Data collecting Data filtering Data processing
Deviding and setting up portfolios
Carrying out the tests and running regression
Analysis of research results and give conclusion xxxv
Step 4: Dividing and setting up portfolios
Based on the 4 quotas including Size, B/M ratio, OP and Inv to divide all the selected stocks into 18 portfolios which will be shown details in the following part
Utilize Microsoft Excel to analyze and quantify five key factors in the financial model: SMB (Small minus Big), which measures the return difference between small and large stocks; HML (High minus Low), reflecting the return disparity between high and low book-to-market ratio stocks; RMW (Robust minus Weak), indicating the return variation between stocks with robust and weak profitability; and CMA (Conservative minus Aggressive), representing the return contrast between conservative and aggressive investments Calculate the complements of these factors, conduct statistical descriptions, and assess the correlations among them for comprehensive insights.
Step 6: Carrying out the tests and running the regression model
Utilize Stata software to import and analyze sorted portfolios from Excel, effectively addressing multicollinearity, autocorrelation, and heteroscedasticity Execute regression models to evaluate the performance of selected companies and assess the efficiency of these models in relation to the chosen data.
Step 7: Analysis of research results and give conclusions
The regression analysis indicates the influence of various factors on the rate of return By compiling and comparing these results, we can identify the most suitable model and classification Additionally, assessing the model's appropriateness allows us to provide valuable recommendations for investors and company stakeholders.
This study mainly uses Stata version 13 in conducting data analysis and providing regression results Microsoft Office Excel is used to organize and utilize the sample data xxxvi
Model
The model for time-series regression:
The expected return on portfolio i for period t is influenced by the risk-free interest rate of government bonds, the excess market return, and the Small minus Big (SMB) common risk factor introduced by Fama and French in 1993 The SMB factor reflects the average return difference between small-cap and large-cap stocks Portfolios are constructed each quarter based on market capitalization, which serves as a proxy for size, using accounting data from the fiscal quarter ending in the previous period.
The High minus Low (HML) risk factor, introduced by Fama and French in 1993, measures the return difference between high and low book-to-market ratio stocks This factor is calculated quarterly, utilizing accounting data from the fiscal quarter ending in the previous period, to form portfolios that reflect these varying book-to-market ratios.
The Robust minus Weak (RMW) common risk factor, introduced by Fama and French in 2015, measures the return difference between high and low profitability stock portfolios These portfolios are constructed quarterly, utilizing accounting data from the fiscal quarter ending in the previous period.
The Conservative minus Aggressive (CMA) risk factor, introduced by Fama and French in 2015, measures the return difference between conservative and aggressive stock portfolios This factor is calculated quarterly, using the growth of total assets from the previous fiscal quarter, divided by the total assets at the end of that quarter.
The filtering process in this study was conducted in a manner similar the guidelines proposed by Ince and Porter (2006) in order to make the sample applicable for analysis
The process began by downloading static constituent list data from the Ho Chi Minh City (HCMC) and Hanoi stock exchange websites, ensuring that all relevant stock information was included Data was collected for all available stocks on the HSX and HNX, encompassing approximately 101 industrial stocks on the HCMC stock exchange, which includes both active and inactive stocks.
The study analyzed 120 industrial stocks, with the detailed constituent lists and variables outlined in Appendix A The initial sample was refined by filtering out static stock information from both stock markets, as detailed on the following page This process involved eliminating most investment companies and non-equity firms, whose returns are derived from financing or financial statements of other entities Step 1 documented the number of stocks remaining after filtering static variables, with a timestamp indicating the last data update Step 2 involved the removal of stocks with non-applicable time-series data, particularly those for which download requests failed to retrieve necessary variables In Step 3, the data was double-checked for stationarity throughout the sample period, leading to the exclusion of stocks with non-stationary data Finally, Step 4 presented the final sample, which comprised 100 stocks after removing duplicates.
Table 2: Number of stocks in filtering stages
Constituent List RAW Step 1 Step 2 Step 3 Step 4
Details of sorting values cleaning:
To ensure the availability of comprehensive public information, only companies listed on the stock market from the beginning of 2012 or earlier are selected for analysis This requirement mandates that companies must have been listed for at least six months (two quarters) before announcing their capital assets and financial reports, thereby minimizing the effects of new listings on the data As a result, the sample includes firms with a complete dataset spanning six years (24 quarters) from 2012 to 2018.
Financial and insurance companies are excluded from the sample due to significant differences in their operations, financial policies, capital structures, and accounting systems compared to non-financial companies.
Companies that have gone through restructuring in the time period are also eliminated because their value and size can have significant change, as consequence, affect industrial companies business performances
The methodology established by Fama and French (1993, 2015) involves defining and calculating key variables on a quarterly basis At the conclusion of each quarter (quarter t), companies are sorted and assigned to portfolios based on four critical factors: market capitalization, book-to-market equity, profitability, and investment.
Stock return refers to the average rate of return calculated over the days within a quarter It is determined by comparing the ending price of the stock at the conclusion of quarter t to its ending price at the previous quarter This metric provides valuable insights into the stock's performance over time.
𝑡 quarterly rate of return of stocks is calculated as follows:
Market return, referred to as "Market," is calculated from the VN-Index on a daily basis This metric is averaged quarterly, representing the average daily return over that period Specifically, if the VN-Index for quarter t is denoted as VN-Index(t) and the VN-Index for the previous quarter is VN-Index(t-1), the market return reflects the change in these indices.
1, the daily rate of return is calculated as follows:
The risk-free rate of return, representing the "Market," refers to the principal interest rate of a 1-year Treasury bill issued by the Reserve Bank of Vietnam from January 2012 to December 2017 This rate is calculated by dividing the quarterly risk-free interest rate by four quarters.
Market capitalization (stands for ―Size‖): the product of number of shares outstanding and the market price per share, as on the last day of each quarter t:
Book-to-market equity, often referred to as the B/M ratio, compares a company's book value, derived from its accounting records, to its market value, which is established through market capitalization This ratio is calculated by dividing the book equity by the market equity at the end of each quarter.
Profitability, represented by the term "OP," refers to the return on equity (ROE) This metric is calculated by subtracting all operating expenses, interest, depreciation, taxes, and preferred stock dividends from a company's total revenue, resulting in the remaining sales.
Investment (stands for ―Inv‖): using asset growth as a proxy for investment If the total assets in quarter t and the total assets in quarter t-1, following Cooper et al
(2008), Fama and French (2008), Gray and Johnson (2011) and Fama and French
(2014) Asset growth is defined as follows:
Beginning in January 2012, this study first categorizes stocks into five portfolios based on the cutoffs established by Brailsford et al (2012) The subsequent phase of the analysis involves creating sorted portfolios to facilitate the calculation of the Fama and French factor return series.