Testing the Validity of Capital Asset Pricing Model:
Case Study on Indonesian Stock Market
Roberta Octami Sorongan
10436081
Bachelor Thesis in Economics and Finance
Supervisor:
dhr. dr. K.B.T. Boe Thio
Faculty of Economics and Business
University of Amsterdam
Table of Content
1. Introduction ... 3
2. Literature Review ... 5
2.1 Development of Capital Asset Pricing Model ... 5
2.2 Arguments against Capital Asset Pricing Model ... 5
2.3 Arguments supporting Capital Asset Pricing Model ... 6
2.4 Emerging Markets Perspective ... 6
2.5 Previous Empirical Findings ... 7
2.6 Performance of Capital Asset Pricing Model in Indonesia ... 8
3. Data and Methodology ... 10
3.1 Data Selection ... 10
3.2 Sub Periods and Portfolio Formation ... 10
3.3 Capital Asset Pricing Model Testing Procedures... 12
4. Empirical Results and Analysis ... 13
4.1 Entire Period... 13
4.2 Sub Periods ... 14
4.3 Test of Security Market Line ... 15
4.4 Test of Non-Linearity ... 16
4.5 Test of Non-Systematic Risk ... 17
5. Conclusion ... 19
5.1 Remarks on the Effect of 2008 Crisis ... 19
5.2 Limitations to this Study ... 20
References ... 21
1. Introduction
In making investment evaluations and decisions, it has been important to provide a good estimation of expected return, price of stocks, or optimal portfolio. Investors and financial managers are also seeking to minimize the level of risk when investing in a stock. These issues can recently be solved using financial tools to determine an investment’s future orientation. Modern finance theory has provided many insights into how stock prices are formed and has provided a quantitative description for the risk structure of equilibrium expected returns (Merton, 1980). As a result, a model is developed which is referred to as Capital Asset Pricing Model, or CAPM.
CAPM was originally developed by Sharpe (1964) and Treynor (1961). In its most elementary form, the equilibrium structure is defined by the following equation:
𝐸𝐸[𝑅𝑅𝑖𝑖] = 𝑅𝑅𝑓𝑓+ 𝛽𝛽𝑖𝑖�𝐸𝐸[𝑅𝑅𝑚𝑚] − 𝑅𝑅𝑓𝑓�
where 𝐸𝐸[𝑅𝑅𝑖𝑖] and 𝐸𝐸[𝑅𝑅𝑚𝑚] respectively denote the expected rate of return on stock i and market portfolio; 𝑅𝑅𝑓𝑓 is the risk-free interest rate; and 𝛽𝛽𝑖𝑖 is the covariance of the return on stock i with the return on the market divided by the variance of return on the market. Not only this
relationship can be used on the world of securities investment, but it has also been extended to be applied in estimating a company’s cost of equity capital.
Nonetheless, CAPM has been tested and challenged empirically to evaluate its ability in explaining risk and return relationship since its development. Many have argued and come up with empirical results that indicate weak support for this model. Fama and French (2004) mentioned that the record of the model is empirically poor as well as reflecting theoretical failings as a result of many simplifying assumptions. As new financial markets emerge around the world with their differences in system, potential return and risk structures, it is getting more essential to test the validity of the model.
Earlier empirical studies of CAPM were mostly done on the US, UK, or European market, but lately many studies have also been conducted on the emerging countries market, especially in the Southeast Asia region. Garg (1998) suggested in his literature review that most studies in the emerging market have resulted in the underperformance of CAPM in explaining the risk return relationship.
However, one finding by Johnson and Soenen (1996) in Indonesian stock market for the period December 29, 1990 through the end of 1993 indicates that most of the stocks are not under- or overvalued according to CAPM. This evidence seems to support the model,
which is rather contrary to other findings in similarly emerging markets. This anomaly may be interesting to be brought to attention. Thus, the purpose of this thesis is to investigate whether the Capital Asset Pricing Model is relevant in estimating stocks return in Indonesian stock market using more recent data.
Indonesia first established its stock exchange in 1914, but it was aroused only after the deregulation actions in 1987 and 1989. Most of the foreign trades are conducted and
concentrated in the Jakarta Stock Exchange (JSX). As JSX may actually be one of Asia’s smallest bourses, it is also one of the fastest-growing (Johnson and Soenen, 1996). In this study, the test will be conducted on 38 stocks traded in LQ45 index per 2013, which is a capitalization-weighted index of the most liquid and heavily traded stocks on the Indonesia Stock Exchange (formerly Jakarta Stock Exchange). This index was launched in February 1997, and will firmly reflect the stock market condition in Indonesia as it covers at least 70% of the stock market capitalization and transaction values in the Indonesian stock market.
This thesis will be organized as follows: the next section will provide a brief summary of the literature review on fundamental background of CAPM and arguments on CAPM. It will focus more on the emerging markets point of view and previous empirical findings, including in Indonesia. Afterwards the methodology and data for the test will be discussed, and then on the next section the empirical data results and analysis will be presented. Finally we will come up with the conclusion and possibly further recommendation.
2. Literature Review
2.1 Development of Capital Asset Pricing Model
On their paper regarding CAPM, Black, Scholes, and Jensen (1972) discussed the original development of the model by Sharpe (1964) and Treynor (1961). This model was then extended and clarified by Lintner (1955a; 1965b), Mossin (1966), and Fama (1968a; 1968b), and Long (1972).
Alongside CAPM’s existence, there have been many developments, such as the portfolio evaluation models by Treynor (1965), Sharpe (1966), and Jensen (1968; 1969). These models are based on this asset pricing model or bear a close relation to it. There have also been many added assumptions, they are:
• Investors are rational, risk-averse utility of terminal wealth maximizers, and can choose between portfolios exclusively on the basis of mean and variance
• No taxes or transaction costs
• Investors have homogeneous views regarding the parameters of the joint probability distribution of all security returns, and
• Investors can borrow and lend unlimited amount at a given risk-free interest rate The last assumption regarding the risk-free borrowing and lending is the last phase in the development of the Sharpe-Lintner CAPM model (Fama and French, 2004). Under this condition, 𝐸𝐸[𝑅𝑅𝑍𝑍𝑍𝑍], the expected return on assets that are uncorrelated with market return, must be equal to 𝑅𝑅𝑓𝑓, the risk-free rate. This relation between the expected return and risk then becomes the following CAPM equation:
𝐸𝐸[𝑅𝑅𝑖𝑖] = 𝑅𝑅𝑓𝑓+ 𝛽𝛽𝑖𝑖�𝐸𝐸[𝑅𝑅𝑚𝑚] − 𝑅𝑅𝑓𝑓� (1)
Put into words, the expected return on stock i is the risk-free interest rate plus a risk premium, which is the stock’s market beta times the premium per unit of beta risk.
2.2 Arguments against Capital Asset Pricing Model
Fama and French (2004) pointed out the failure of CAPM both empirically and theoretically. They mentioned that the empirical record of the model is poor enough to be applicable, and it may reflect theoretical failings as a result of the simplifying assumptions stated previously. These assumptions are obviously not relevant in the real-world case. Lumby and Jones (2003) admit to the fact that these assumptions are unrealistic, since market inefficiencies are prevailing due to several causes such as government
In addition, Dempsey (2013) examined the CAPM theoretically and argued that the model has actually reached the point where it should have been abandoned. The continued defense of the CAPM by adding additional factors for unsystematic volatility, liquidity, momentum, and so forth are typical Kuhn’s articulation of “normal science”. This refers to how the single-factor CAPM has now become three, four, and even the latest five-factor model by Fama and French.
2.3 Arguments supporting Capital Asset Pricing Model
Despite many empirical evidences showing that CAPM has not been reliable in estimating expected returns, there are still many defenses to this risk-return relationship. Brown and Walter (2013) discussed the theoretical validity of CAPM to counter the argument of Dempsey in 2013. They explain the problems with Dempsey’s previous claim against CAPM.
First, they presume that the questionable validity is not within the model, but within the empirical evidence itself. In 1977, Richard Roll concluded that many CAPM tests were actually invalid due to the use of inefficient benchmark portfolios, whereas CAPM requires the benchmark to be efficient. Second, the suggestion that investors do not expect a compensation for unavoidable risk is contrary to the beliefs of the theorists and practitioners, namely that for the investors risk matters such that ex ante, a risk premium must exist.
Chan and Lakonishok (1993) also developed their defense to CAPM and beta as a measure of risk. They tried to evaluate if there is truly sufficient evidence to dump beta. It is rather difficult to draw any clear-cut conclusions from empirical research on stock returns, due to the noise and constantly changing environment generating stock returns.
2.4 Emerging Markets Perspective
As this study of CAPM takes perspective from an emerging market like Indonesia, it is important first to define the concept of emerging market. In the financial community, there has been a significant amount of confusion to exactly characterize an “emerging stock market”. The International Finance Corporation, World Bank, has employed a definition that is broadly accepted. It states that within emerging countries, the market is located in a low- or middle-economy, and there is a relatively low ratio of investable market capitalization to its most recent Gross National Product (GNP).
There are also many other definitions regarding emerging market. Cavusgil (1987) describe that essentially, emerging markets are high-growth developing countries that represent attractive business opportunities for the Western firms. In addition, the
opportunities for future market expansion distinguish the emerging countries from the less developed countries. The forms of economic stimulus, such as development of new technologies, foreign investment, or external participation in their commercial affairs only occur in countries with policies towards increased growth (Miller, 1998).
Up to now, Indonesia can be categorized as one of the emerging countries market as its economy characteristics are in line with the aforementioned definition. Indonesia has also been listed as a sample in many studies regarding emerging market, such as the one conducted by Hartmann and Khambata in 1993.
In its relation to CAPM, pricing risky assets in the emerging market may be rather problematic because institutional, political, and macroeconomic conditions are generally volatile. This high volatility may have considerable impacts for the test of asset pricing models. First, the parameters of both asset pricing models and expected returns are unlikely to remain constant over time. Second, the distribution of asset returns does not follow normal distribution (Brooks, Galagedera, and Iqbal, 2010).
Some of the CAPM assumptions also may raise concerns if applied to the emerging market. Harvey (2000) emphasizes that the assumption of perfect capital market is a serious problem in applying the model to emerging market. This assumption implies that markets are perfectly integrated. In contrast, evidence shows that there is segmentation of emerging markets from the global stock market, i.e. separation from the global market from a pricing point of view (Drobetz, Stürmer, and Zimmermann, 2002). Therefore, the model may not perform very well with these markets. This thesis will address this issue to Indonesia.
2.5 Previous Empirical Findings
Throughout its existence, many empirical tests have been performed to evaluate CAPM among many financial markets. In this paper we are going to review two studies that represent the results of CAPM validity testing from opposing types of financial markets. They are the developed financial market and the emerging financial market in particular. Basically these studies employ similar method of testing as what we are going to carry in this paper.
Modigliani et al. (1973) conducted the test of asset pricing model on the eight major European stock markets. They are France, Italy, U.K., Germany, Netherlands, Belgium, Switzerland, and Sweden. These countries are considered as having developed financial markets. With U.S. market as the benchmark of consistency of the model, they empirically show that on the whole, the European results are comparable with the U.S. This implies that CAPM is relevant when applied to these markets.
On the contrary, Aljinović and Džaja (2013) tested the CAPM on the emerging markets of the Central and Southeastern Europe, including returns from Croatia, Czech Republic, Hungary, Poland, Turkey, Serbia, Bulgaria, Romania, and Bosnia and
Herzegovina financial market in the sample. The test of the validity of beta, efficient frontier, as well as the cross sectional analysis suggested that the CAPM is not sufficient to assess the price of capital assets on the observed markets.
2.6 Performance of Capital Asset Pricing Model in Indonesia
In the Indonesian market itself, a survey was conducted among the companies by Leon, et. al. (2008) to investigate which capital budgeting practice is mostly used by the executives. This study reveals that only 14.7% of the respondents indicated that their companies use CAPM to estimate the cost of equity capital. This appears to be the average number compared to emerging South East Asian countries such as Malaysia with 6.2%, Singapore with 24.1%, Philippines with 24.1%, and Hong Kong with 24.1% as shown in another study conducted by Kester et. al. (1999).Not surprisingly, all these numbers are relatively low compared to the application of CAPM in other developed countries market, where it is reported that the model is used by 72.7% of the Australian companies (Kester et. al., 1999), 73% of the US and Canadian companies (Harvey, 2001), and 47% of the companies in the UK (McLaney et. al., 2004).
Despite many studies that have been conducted to test the validity of CAPM in the emerging markets, scarcely any was conducted in Indonesia. However, in the paper presented by Johnson and Soenen (1996) regarding the risk and return characteristics in the Jakarta Stock Exchange, a short test of CAPM was performed and resulted in quite unexpected conclusion. Using weekly returns from 75 leading stocks traded on the Jakarta Stock Exchange during the period December 29, 1990 through the end of 1993, they interpret that in most cases:
• alpha estimates do not differ significantly from zero; and
• investors are compensated only for bearing systematic risk and not for the non-systematic one.
These are according to what would be expected in the theory. In shorts, it indicates that most stocks are not under- or overvalued according to CAPM and this model is rather relevant to be used in the Indonesian market.
3. Data and Methodology
3.1 Data Selection
This study uses monthly stock prices for the stocks traded in LQ45 index. LQ45 will also be selected as a proxy to the market portfolio as it represents the largest companies from various economy sectors. Originally, this index consists of 45 most heavily traded stocks on the Indonesia Stock Exchange. However, only 38 out of 45 companies have the available data required for the entire testing period, which is January 2010 to December 2013. Thus we are going to narrow down the test only to these
companies1.
Here monthly returns are used instead of daily returns as conducted by Dimson (1979) and Cohen, Hawanini, and Maier (1983). The main purpose is to decrease the thin-trading effect, or the intervaling effect. The closing price of the last thin-trading day in the month is used to calculate the monthly returns based on the following equation:
𝑅𝑅
𝑖𝑖,𝑡𝑡=
𝑃𝑃𝑡𝑡−𝑃𝑃𝑡𝑡−1+𝐷𝐷𝑡𝑡𝑃𝑃𝑡𝑡−1 (2)
where 𝑅𝑅𝑖𝑖,𝑡𝑡 is the return of stock i at time t, 𝑃𝑃𝑡𝑡 is stock price at time t, 𝑃𝑃𝑡𝑡−1 is the stock price at time t – 1, and 𝐷𝐷𝑡𝑡 is the amount of dividends paid on stock i at time t. The data on these returns was retrieved from finance.yahoo.com and was already adjusted for dividends and splits.
The value of risk-free rate will be according to the BI Rate, which is the policy rate that reflects the monetary policy stance adopted by Bank Indonesia (i.e. Indonesia’s central bank) and announced to the public2. This rate is announced by the Board of Governors of Bank Indonesia in each of monthly Board of Governors Meeting.
3.2 Sub Periods and Portfolios Formation
The testing will employ the method used by Black, Jensen, and Scholes (1972). In general, the test will be done within the entire period of January 2010 – December 2013, as well as four equally divided sub periods, each containing 24 months, summarized in the
1 Sample companies are listed in Appendix 1. These are the companies which data on returns are available for the whole sets of estimation and testing period, which is from January 1, 2008 to December 31, 2013. 2 Complete risk-free rates are shown in Appendix 2. The monthly data is obtained from Bank Indonesia’s website (www.bi.go.id) and it is actually yearly rate. Therefore, in order to adjust it to monthly rate the following formula is used:
(1 + 𝑅𝑅𝑓𝑓)(1 12� )− 1
table below. The purpose of doing this division is to test the stationarity of the empirical relations.
Table 1
Beta estimation, portfolio formation, and testing periods
Sub Period 1 Sub Period 2 Sub Period 3 Sub Period 4
Beta estimation period 2008-2009 2009-2010 2010-2011 2011-2012 Portfolio formation and testing period 2010 2011 2012 2013 Number of securities 38 38 38 38
The test is based on the time series regressions introduced by Black et al (1972). We begin by estimating the coefficient 𝛽𝛽𝑖𝑖 (identified as estimate 𝛽𝛽̂𝑖𝑖) by regressing 𝑟𝑟𝑖𝑖,𝑡𝑡 to
𝑟𝑟𝑚𝑚,𝑡𝑡 for sub period 1 (2008-2009) on the following equation:
𝒓𝒓𝒊𝒊,𝒕𝒕 = 𝜶𝜶𝒊𝒊+ 𝜷𝜷𝒊𝒊𝒓𝒓𝒎𝒎,𝒕𝒕+ 𝒆𝒆𝒊𝒊,𝒕𝒕 (3)
This equation is basically obtained by assuming that the stocks are priced in the market such that equation (1) holds over each short time interval (in this case a month), then we can do the test by rearranging the traditional form of the model and adding an intercept 𝛼𝛼𝑖𝑖. 𝑟𝑟𝑖𝑖,𝑡𝑡 simply represents expected excess returns on stock i at time t, 𝐸𝐸�𝑅𝑅𝑖𝑖,𝑡𝑡� − 𝑅𝑅𝑓𝑓, while 𝑟𝑟𝑚𝑚,𝑡𝑡
represents expected excess market returns at time t, 𝐸𝐸�𝑅𝑅𝑚𝑚,𝑡𝑡� − 𝑅𝑅𝑓𝑓.
These securities were then ranked from the on the basis of estimates 𝛽𝛽̂𝑖𝑖 from highest to the lowest, which then were assigned to six equally-weighted portfolios, with each containing 6 to 7 stocks. Combining stocks into portfolio will diversify away most of the firm-specific part of returns, and therefore will enhance the precision of the beta estimates and the expected rate of return on the portfolios (Michailidis et. al., 2006). The return in each of the next 12 months (year 2010) for each of the six portfolios was calculated. This process was then repeated for the next sub periods.
The following step is to estimate the portfolio beta, 𝛽𝛽̂𝑝𝑝 according to the equation below:
𝑟𝑟𝑝𝑝,𝑡𝑡 = 𝛼𝛼𝑝𝑝+ 𝛽𝛽𝑝𝑝𝑟𝑟𝑚𝑚,𝑡𝑡+ 𝑒𝑒𝑝𝑝,𝑡𝑡 (4)
where 𝑟𝑟𝑝𝑝,𝑡𝑡 is the average portfolio excess returns at time t and 𝛽𝛽𝑝𝑝 is the portfolio beta. Once again this process is repeated for the next sub periods and the whole period.
3.3 Capital Asset Pricing Model Testing Procedures
The first test to conduct is on the ex-post Security Market Line (SML) for the testing period by regressing the portfolio excess returns (𝑟𝑟𝑝𝑝,𝑡𝑡) against the portfolio betas (𝛽𝛽𝑝𝑝) on the equation below:
𝑟𝑟𝑝𝑝 = 𝛾𝛾0+ 𝛾𝛾1𝛽𝛽𝑝𝑝+ 𝑒𝑒𝑝𝑝 (5)
Corresponding to the traditional form of the asset pricing model, it implies that the intercept 𝛾𝛾0 in (5) should be equal to zero and the slope 𝛾𝛾1 should be equal to 𝑅𝑅�𝑍𝑍, the average excess return on the market portfolio.
The next step is to run a test of non-linearity between the portfolio excess returns and portfolio betas using the following equation:
𝑟𝑟𝑝𝑝= 𝛾𝛾0+ 𝛾𝛾1𝛽𝛽𝑝𝑝+ 𝛾𝛾2𝛽𝛽𝑝𝑝2+ 𝑒𝑒𝑝𝑝 (6)
CAPM hypothesis is that the portfolio returns and betas are linearly related with each other, which means that the slope 𝛾𝛾2 should be equal to zero.
The last is to test whether the portfolio excess returns are determined solely by the systematic risk (i.e. non-systematic risk does not exist). Here we regress the portfolio excess returns, 𝑟𝑟𝑝𝑝 to the residual variance of portfolio excess returns, 𝜎𝜎2(𝜀𝜀𝑝𝑝) in the equation:
𝑟𝑟𝑝𝑝 = 𝛾𝛾0+ 𝛾𝛾1𝛽𝛽𝑝𝑝+ 𝛾𝛾2𝛽𝛽𝑝𝑝2+ 𝛾𝛾3𝜎𝜎2(𝜀𝜀𝑝𝑝) + 𝑒𝑒𝑝𝑝 (7)
Again, if CAPM holds true, then 𝛾𝛾3 should also equal to zero.
All the tests mentioned above are also repeatedly performed for both the entire period and each sub period. To sum up, concerning the validity of CAPM in Indonesia, here we are going to test whether:
• the intercept equals to zero; • average risk premium exists;
• the relation between the return and risk is linear; and • beta is the only risk variable.
Each of the hypotheses testing will be conducted using two-tailed t-tests at the 95% confidence level.
4. Empirical Results and Analysis
4.1 Entire Period
Using the monthly returns of 4-year on each of the six portfolios constructed as explained previously, we then estimate the least-squares parameters which results in 𝛼𝛼�𝑝𝑝 and 𝛽𝛽̂𝑝𝑝 in equation (4) for each of the six portfolios (p = 1, 2, …, 6) using all 4-year of monthly data. The idea is that the first portfolio contains the stocks with highest betas while the sixth portfolio contains the stocks with lowest betas. The advantage of using this approach is the unbiased and efficient properties. In this case, where the number of stocks in each sample is relatively small, the trade-off between these properties becomes crucial (Modigliani et al., 1973). The results are summarized in the following table3:
Table 2
Statistics for Time Series Tests, Entire Period (January 2010 – December 2013)
Item Portfolio Number 1 2 3 4 5 6 Market 𝜷𝜷�𝒑𝒑 1.393 1.324 1.048 1.009 1.134 0.754 1.000 𝜶𝜶�𝒑𝒑 0.019 0.012 0.008 0.017 0.019 0.021 𝒕𝒕�𝜶𝜶�𝒑𝒑� 2 2.02 1.23 2.32 3 2.56 𝒓𝒓(𝑹𝑹�, 𝑹𝑹�𝑴𝑴) 0.750 0.859 0.785 0.712 0.806 0.566 𝑹𝑹� 2.367% 1.689% 1.152% 2.114% 2.304% 2.403% 0.363% 𝝈𝝈 0.096 0.080 0.069 0.073 0.073 0.069 0.052
The estimated risk coefficients, 𝛽𝛽̂𝑝𝑝, range from 1.393 for portfolio 1 to 0.754 for portfolio 6. In general these coefficients are getting lower as we move from the first through the sixth portfolio, except for portfolio 5 in which beta is relatively higher. The significance tests of 𝛼𝛼�𝑝𝑝, given by the t-values 𝑡𝑡�𝛼𝛼�𝑝𝑝�, show that 5 out of 6 coefficients have t-values greater than 1.96, which is the critical value for 5% significance level. The
correlation coefficient between portfolio return and market return, 𝑟𝑟(𝑅𝑅�, 𝑅𝑅�𝑍𝑍), is also given in the table. The numbers appear to be lower than expected, with portfolio 2 being the highest at 0.859.
3 For complete average monthly returns on the six portfolios, see Appendix 3. These portfolios indeed have different composition of company stocks for each sub period.
However, it may be too early to derive a conclusion from this entire period result, due to a possibility of the existence of some non-stationarity in the relations, as well as the lack of more complete aggregation (Black et al., 1972). Therefore, we need to do the testing for each of the sub periods.
4.2 Sub Periods
After dividing into four equal sub periods each containing 24 months, we repeat the estimation process for each of the sub periods. The table below presents a summary of regression on equation (4) calculated using the data for each of these sub periods, as well as for each of the six portfolios4:
Table 3
Statistics for Time Series Tests, Per Sub Period
Item Sub Periods Portfolio Number 1 2 3 4 5 6 Market 𝜷𝜷�𝒑𝒑 1 1.150 1.510 1.136 1.078 0.976 0.204 1.000 2 1.455 1.239 0.690 0.807 1.064 0.971 1.000 3 1.190 1.278 1.369 1.188 0.625 0.872 1.000 4 1.643 1.223 1.276 0.762 1.597 0.650 1.000 𝜶𝜶�𝒑𝒑 1 0.021 0.007 0.003 0.056 0.038 0.061 2 0.019 0.011 0.008 0.002 0.011 0.013 3 0.042 0.027 0.000 0.000 0.015 0.027 4 0.000 -0.001 0.017 0.009 0.020 -0.007 𝒕𝒕�𝜶𝜶�𝒑𝒑� 1 1.36 0.41 0.31 3.19 2.91 2.13 2 1.09 1.68 0.60 0.27 1.63 1.15 3 1.74 1.94 0.01 -0.02 1.56 3.13 4 0.02 -0.15 1.13 0.44 1.12 -0.83 𝑹𝑹� 1 4.438% 3.712% 2.596% 7.705% 5.790% 6.537% 1.991% 2 1.562% 0.822% 0.672% 0.017% 0.837% 1.129% -0.214% 3 4.628% 3.190% 0.486% 0.391% 1.703% 3.031% 0.347% 4 -1.0715 -0.970% 0.853% 0.343% 0.887% -1.087% -0.672% 𝝈𝝈 1 0.077 0.097 0.070 0.079 0.066 0.089 0.054 2 0.105 0.078 0.061 0.054 0.068 0.070 0.060 3 0.094 0.071 0.068 0.061 0.041 0.047 0.042 4 0.107 0.071 0.083 0.075 0.101 0.042 0.052 4
Composition of the companies for each portfolio per sub period is available on Appendix 4.
From the table we can see that the data for 𝛽𝛽̂𝑝𝑝 indicates that the risk coefficients 𝛽𝛽̂𝑝𝑝 were non-stationary throughout the period. The sections for 𝛼𝛼�𝑝𝑝 and 𝑡𝑡�𝛼𝛼�𝑝𝑝� indicate that the critical intercepts were also non-stationary for portfolio number 4, 5, and 6, which contain the lower beta stocks. There seems to be no obvious patterns that can be derived from these values. However, we can see that the standard deviations of each regression are relatively small, which provide a major improvement of grouping the data into sub periods.
4.3 Test of Security Market Line
As previously explained, the traditional model of CAPM implies that the intercept 𝛾𝛾0 in equation (5) should be equal to zero and the slope 𝛾𝛾1 should be equal to the mean
excess return on the market portfolio. On the entire period, the average monthly excess return of the market portfolio was 𝑅𝑅�𝑍𝑍 = 0.145%. Therefore, the theoretical values of both the intercept and slope should be respectively
𝛾𝛾0 = 0 and 𝛾𝛾1 = 0.145%
The t-values are obtained as the following: 𝑡𝑡(𝛾𝛾�0) =𝑠𝑠(𝛾𝛾�𝛾𝛾�0 0) = 0.022 0.012 = 1.867 𝑡𝑡(𝛾𝛾�1) =𝛾𝛾1𝑠𝑠(𝛾𝛾�− 𝛾𝛾�1 1) = 0.00145 − (−0.002) 0.011 = 0.338
They appear to be relatively small and not rejected at 5% significance level (t-critical = ±1.96). This test is again repeated on every sub periods which is summarized in the following table.
Table 4
Statistics for Security Market Line Tests
Item
Time Period
Sub Periods Total
Period 1 2 3 4 𝜸𝜸�𝟎𝟎 0.074 -0.005 -0.027 -0.007 0.022 𝒔𝒔(𝜸𝜸�𝟎𝟎) 0.020 0.007 0.033 0.014 0.012 𝒕𝒕(𝜸𝜸�𝟎𝟎) 3.625 -0.825 -0.822 -0.468 1.867 𝜸𝜸�𝟏𝟏 -0.023 0.013 -0.004 0.004 -0.002 𝒔𝒔(𝜸𝜸�𝟏𝟏) 0.019 0.006 0.030 0.012 0.011 𝜸𝜸𝟏𝟏 = 𝑹𝑹�𝑴𝑴 -0.700% 3.466% 0.889% 0.067% 0.145%
𝒕𝒕(𝜸𝜸𝟏𝟏− 𝜸𝜸�𝟏𝟏) 0.826 3.436 0.450 -0.306 0.338
Hypothesis Rejected Rejected Not rejected Not rejected Not rejected
The hypotheses are rejected only for the intercept in the first and slope in the second sub period. They are not rejected in the other sub periods as well as for the entire period. In general, the hypotheses that the intercept equals to zero and that average risk premium exists are not rejected in this test, which suggests that CAPM is rather in-line with the empirical evidence.
If we compare these results with the previous findings, Johnson and Soenen (1996) found that alpha estimates are not significantly different from zero. In this test, most intercepts are also equal to zero, which suggests the same thing. In addition, they found that most beta coefficients are positive, which means that average risk premium exists.
4.4 Test of Non-Linearity
The next hypothesis is that the relationship between portfolio’s return and its systematic risk is linear. Therefore we need to conduct a test on possibility of non-linearity according to equation (6). Adding a new hypothesis to the ones on the previous test, in case of a linear relationship, 𝛾𝛾𝟐𝟐 should also be equal to zero. The result of the test is presented in the table below.
Table 5
Statistics for Non-Linearity Tests
Item
Time Period
Sub Periods Total
Period 1 2 3 4 𝜸𝜸�𝟎𝟎 0.065 0.008 -0.105 -0.034 0.079 𝒔𝒔(𝜸𝜸�𝟎𝟎) 0.032 0.033 0.148 0.059 0.058 𝒕𝒕(𝜸𝜸�𝟎𝟎) 2.057 0.242 -0.711 -0.580 1.359 𝜸𝜸�𝟏𝟏 0.010 -0.013 0.285 0.058 -0.110 𝒔𝒔(𝜸𝜸�𝟏𝟏) 0.081 0.064 0.316 0.112 0.109 𝜸𝜸𝟏𝟏 = 𝑹𝑹�𝑴𝑴 -0.700% 3.466% 0.889% 0.067% 0.145% 𝒕𝒕(𝜸𝜸𝟏𝟏− 𝜸𝜸�𝟏𝟏) -0.210 0.752 -0.873 -0.512 1.022 𝜸𝜸�𝟐𝟐 -0.020 0.012 -0.146 -0.024 0.050 𝒔𝒔(𝜸𝜸�𝟐𝟐) 0.049 0.030 0.159 0.049 0.050
𝒕𝒕(𝜸𝜸�𝟐𝟐) -0.415 0.419 -0.920 -0.484 0.993
Hypothesis Rejected Not rejected Not rejected Not rejected Not rejected
The result of this test is even less significant than the previous one. Only the
intercept in sub period one significantly differs from zero. The hypothesis is not rejected in the rest of the sub periods as well as the total period. This indicates that there is actually a linear relationship between the risk and return from the evidence, as predicted by the model.
There was no tests on the non-linearity of the model found to be conducted previously in Indonesia, particularly by Johnson and Soenen (1996). However, in most cases the risk and return relationship is found to be linear, even including the emerging stock markets. This means the quadratic version of the model as in equation (6) is not relevant enough to be applied.
4.5 Test of Non-Systematic Risk
According to the Capital Asset Pricing Model, investors are only compensated for bearing the systematic risk (i.e. not for the idiosyncratic risk). This implies that coefficient 𝛾𝛾𝟑𝟑 in equation (7) should equal to zero. The summary of the test conducted for the
non-systematic risk is presented below.
Table 5
Statistics for Non-Systematic Risk Tests
Item
Time Period
Sub Periods Total
Period 1 2 3 4 𝜸𝜸�𝟎𝟎 -0.178 -0.066 -0.043 -0.054 0.023 𝒔𝒔(𝜸𝜸�𝟎𝟎) 0.113 0.030 0.119 0.067 0.094 𝒕𝒕(𝜸𝜸�𝟎𝟎) -1.578 -2.205 -0.364 -0.810 0.245 𝜸𝜸�𝟏𝟏 0.361 0.127 0.141 0.086 -0.018 𝒔𝒔(𝜸𝜸�𝟏𝟏) 0.169 0.057 0.256 0.123 0.165 𝜸𝜸𝟏𝟏 = 𝑹𝑹�𝑴𝑴 -0.700% 3.466% 0.889% 0.067% 0.145% 𝒕𝒕(𝜸𝜸𝟏𝟏− 𝜸𝜸�𝟏𝟏) -2.178 -1.611 -0.515 -0.694 0.119 𝜸𝜸�𝟐𝟐 -0.175 -0.056 -0.080 -0.038 0.007 𝒔𝒔(𝜸𝜸�𝟐𝟐) 0.078 0.028 0.128 0.055 0.076 𝒕𝒕(𝜸𝜸�𝟐𝟐) -2.252 -2.042 -0.628 -0.705 0.096
𝜸𝜸�𝟑𝟑 2.051 0.476 0.542 0.335 0.066
𝒔𝒔(𝜸𝜸�𝟑𝟑) 0.936 0.158 0.309 0.406 0.084
𝒕𝒕(𝜸𝜸�𝟑𝟑) 2.192 3.004 1.753 0.825 0.786
Hypothesis Rejected Rejected Not rejected Not rejected Not rejected
The result shows that 𝛾𝛾�𝟑𝟑 is insignificant in sub period 3 and 4, as well as the entire period. More or less similar results are also obtained for the intercept and other slopes. When the explanatory variable unsystematic/idiosyncratic risk is introduced, the result suggests no significant relationship between these measures of risk and average portfolio returns.
The study by Johnson and Soenen (1996) also suggests a similar result in which only systematic risks have a considerable effect on Indonesian stock movements. This is in accordance with the theory, as the investor is only rewarded for taking systematic risk because non-systematic risk can be diversified away.
5. Conclusion
This study provides the investigation on the validity of CAPM when applied to the Indonesian stock market using the testing method by Black, Jensen, and Scholes (1972). Using monthly returns data of 38 companies registered in the LQ45 index for the estimation period of 2008-2012, there are four hypotheses associated with CAPM to be tested: intercept equals zero, average risk premium exists, the relation between risk and return is linear, and beta is the only risk variable.
The results provide no significant evidence to reject and rather supportive of these CAPM’s prediction, apart from the sub period 1 and most of the sub period 2 results. These are contrary to many other tests conducted in the emerging market, and suggest that CAPM actually holds in the Indonesian stock market.
5.1 Remarks on the Effect of 2008 Crisis
In general, each of the tests has shown insignificant results for the hypotheses proposed. However, it is obvious that all of these tests are in fact rejected in sub period 1, also for the Security Market Line and Non-Systematic Risk tests in sub period 2. As mentioned previously in the methodology section, the sub period 1 is using beta estimation from 2008-2009 and the sub period 2 from 2009-2010.
Just as we acknowledge, the crisis in 2008 that hit the global financial market may also affect the Indonesian stock market down to the following years after. While no separated tests results on CAPM performance are provided between crisis and non-crisis periods in this study, there are several previous empirical findings that may support this argument.
In their study, Black et. al. (1972) obtained first sub period results that mainly deviate from the latter sub periods. The first sub period of their study was using the excess returns from 1926-1930 for the beta estimation period and excess returns from 1930 for the testing period. During the 1930s, a major crisis also occured in the U.S. which may have derived those contradictory results. Thus, there may as well be an effect of the 2008 financial crisis to these Indonesian stock market results.
5.2 Limitations to This Study
Several limitations to this paper may exist and should be considered before
drawing a clear-cut conclusion to these results. First of all, the time period of estimation is rather short, which only covers 5 years returns (2008-2012). Even though longer time span may help reducing the distortion from random factors that can arise in shorter time span, there are impediments in collecting the complete data set for the whole longer period. Some historical prices are not published anywhere, even some of the companies were not established or have not launched their stock to public before 2008.
Second, the sample stocks used in this test are not randomly selected. They are taken from an index which includes only the largest and most liquid companies in the market. The purpose of using this index is that it is likely to represent the whole market. However, since the number of stocks included here is relatively small, this can lead to inefficient and/or biased tests results.
Moreover, it may be important to put emphasize on analyzing the effect of crisis to CAPM performance since the tests conducted during the period of financial crisis have shown anomalies in the results. There are possibilities that the tests yield in different conclusion when the impact of crisis is taken into account. Therefore, further investigation on this matter may be essential to conduct in another extensive study.
Despite these limitations, we can still conclude from the study that so far the results from the Indonesian market in general are consistent with the hypotheses proposed in order to test the validity of CAPM. In other words, the returns on the Indonesian stock market are relatively predictable by using the Capital Asset Pricing Model. These empirical findings, especially the distinctive results compared to other emerging markets, may be interesting for further studies and useful to the financial analysts or investors in their consideration
References
Aljinović, Z. and Džaja, J. (2013). “Testing CAPM Model on the Emerging Markets of the Central and Southeastern Europe”. Croatian Operational Research Review 4: 164-175.
Black, F., Jensen, M. C., Scholes, M. (1972). “The Capital Asset Pricing Model: Some Empirical Tests”. Studies in the Theory of Capital Markets. New York: Praeger. p. 79-121.
Brooks, R., Galagadera, D. U. A., Iqbal, J. (2010). “Testing Conditional Asset Pricing Models: An Emerging Market Perspective”. Journal of International Money and
Finance 29: 897-918.
Cavusgil, S. T. (1997). “Measuring the Potential of Emerging Markets: An Indexing Approach”. Business Horizons 40: 87-91.
Chan, L. and Lakonishok, J. (1993). “Are the Report of Beta’s Death Premature?”. Journal of
Portfolio Management 19: 51-62.
Cohen, K. Hawanini, G. and Maier, S. (1983). “Estimating and Adjusting for the Intervalling Effect Bias in Beta”. Management Science 29: 135-148.
Dimson, E. “Risk Measurement when Shares are Subject to Infrequent Trading”. Journal of
Financial Economics 7: 197-226.
Drobetz, W., Stürmer, S., and Zimmermann, H, (2002). “Conditional Asset Pricing in Emerging Stock Markets”. The Swiss Journal of Economics and Statistics 138: 507-526.
Fama, E. F. and French, K. R. (2004). “The Capital Asset Pricing Model: Theory and Evidence”. Journal of Economic Perspectives 18: 25-46.
Fama, E. F. and Macbeth, J. (1973). “Risk, return and equilibrium: Empirical tests”. Journal
of Political Economy 81: 607-636.
Garg, V. (1998). The Capital Asset Pricing Model: A Test on the Stock Exchange on
Singapore. Unpublished master’s thesis, University of Western Sydney Nepean,
Sydney, Australia.
Graham, J. and Harvey, C. (2001). “The Theory and Practice of Corporate Finance: Evidence from the Field”. Journal of Financial Economics 60: 187-243.
Hartmann, M. A. and Khambata, D. (1993). “Emerging Stock Markets: Investment Strategies of the Future”. The Columbia Journal of World Business 28: 82-104.
Harvey, C. R. (2000). Asset Pricing in Emerging Markets (Working Paper C20). Retrieved from:
https://faculty.fuqua.duke.edu/~charvey/Research/Working_Papers/W54_Asset_ pricing_in.pdf
Ho, Y. W., Strange, R., Piesse, J. (2000). “CAPM anomalies and the pricing of equity: evidence from the Hong Kong market”. Applied Economics 32: 1629-1636. Johnson. R. and Soenen, L. A. (1996). “The Jakarta Stock Exchange: Risk/Return
Characteristics”. The Journal of Investing 5: 37-46.
Kester, G. W., Chang, R. P., Echanis, E. S., Haikal, S., Isa, M. M., Skully, M. T., Kai-Chong, T., and Chi-Jeng, W. (1999). “Capital Budgeting Practices in the Asia-Pacific Region: Australia, Hong Kong, Indonesia, Malaysia, Philippines, and Singapore”. Financial
Practice and Education 9, 25-33.
Lumby, S. and Jones, M. C. (2003). Corporate Finance: Theory and Practice (7th ed). London: Thomson Learning.
McLaney, E., Pointon, J., Thomas, M., and Tucker, J. (2010). “Practitioners’ Perspectives on the UK Cost of Capital”. The European Journal of Finance 10: 123-138.
Merton, R.C. (1980). “On Estimating the Expected Return On the Market”. Journal of
Financial Economics 8: 323-361.
Michailidis, G., Tsopoglou, S., Papanastasiou, D., and Mariola, E. (2006). “Testing the Capital Asset Pricing Model (CAPM): The Case of the Emerging Greek Securities Market”. International Research Journal of Finance and Economics 4: 78-91. Miller, R. R. (1998). Selling to Newly Emerging Markets. New York: Quorum Books. Modigliani, F., Pogue, G. A., and Solnik, B. H. (1973). A Test of the Capital Asset Pricing
Model on European Stock Markets. Doctoral dissertation, School of Business,
Appendix
Appendix 1: List of Sample Companies
Company Names Ticker Symbol Sector
Astra Agro Lestari, Tbk. AALI Agriculture
Adhi Karya, Tbk. ADHI Infrastructure and Transportations
Adaro Energy, Tbk. ADRO Industrials
AKR Corporindo, Tbk. AKRA Trade, Service & Investment
Astra International, Tbk. ASII Industrials
Alam Sutera Realty, Tbk. ASRI Property and Real Estate
Bank Central Asia, Tbk. BBCA Finance
Bank Negara Indonesia, Tbk. BBNI Finance
Bank Rakyat Indonesia, Tbk. BBRI Finance
Bank Danamon Indonesia, Tbk. BDMN Finance
Sentul City, Tbk. BKSL Property and Real Estate
Bank Mandiri, Tbk. BMRI Finance
Global Mediacom, Tbk. BMTR Infrastructure and Transportations Bumi Serpong Damai, Tbk. BSDE Property and Real Estate
Charoen Pokphand Indonesia, Tbk. CPIN Basic Industry and Chemicals Ciputra Development, Tbk. CTRA Property and Real Estate
XL Axiata, Tbk. EXCL Infrastructure and Transportation
Gudang Garam, Tbk. GGRM Consumer Goods
Indofood Sukses Makmur, Tbk. INDF Consumer Goods
Indocement Tunggal Prakasa, Tbk. INTP Basic Industry and Chemicals
Indo Tambangraya Megah, Tbk. ITMG Mining
Jasa Marga Persero, Tbk. JSMR Infrastructure and Transportation
Kalbe Farma, Tbk. KLBF Consumer Goods
Lippo Karawaci, Tbk. LPKR Property and Real Estate
PP London Sumatra Indonesia, Tbk LSIP Agriculture
Malindo Feedmill, Tbk. MAIN Basic Industry and Chemicals
Multipolar, Tbk. MLPL Trade, Services & Investment
Media Nusantara Citra, Tbk. MNCN Trade, Services & Investment Perusahaan Gas Negara, Tbk. PGAS Infrastructure and Transportation Tambang Baturbara Bukit Asam, Tbk. PTBA Mining
Pakuwon Jati, Tbk. PWON Property and Real Estate
Semen Gresik, Tbk. SMGR Basic Industry and Chemicals
Summarecon Agung, Tbk. SMRA Property and Real Estate
Surya Semestra Internusa, Tbk. SSIA Property and Real Estate
Telekomunikasi Indonesia, Tbk. TLKM Infrastructure and Transportation United Tractors, Tbk. UNTR Trade, Services & Investment
Unilever Indonesia, Tbk. UNVR Consumer Goods
Appendix 2: Bank Indonesia’s Risk-Free Rate
Date Yearly Rate Adjusted for Monthly
Dec 01, 2013 7,50% 0,6045% Nov 01, 2013 7,50% 0,6045% Oct 01, 2013 7,25% 0,5850% Sep 01, 2013 7,25% 0,5850% Aug 01, 2013 6,50% 0,5262% Jul 01, 2013 6,50% 0,5262% Jun 01, 2013 6,00% 0,4868% May 01, 2013 5,75% 0,4670% Apr 01, 2013 5,75% 0,4670% Mar 01, 2013 5,75% 0,4670% Feb 01, 2013 5,75% 0,4670% Jan 01, 2013 5,75% 0,4670% Dec 01, 2012 5,75% 0,4670% Nov 01, 2012 5,75% 0,4670% Oct 01, 2012 5,75% 0,4670% Sep 01, 2012 5,75% 0,4670% Aug 01, 2012 5,75% 0,4670% Jul 01, 2012 5,75% 0,4670% Jun 01, 2012 5,75% 0,4670% May 01, 2012 5,75% 0,4670% Apr 01, 2012 5,75% 0,4670% Mar 01, 2012 5,75% 0,4670% Feb 01, 2012 5,75% 0,4670% Jan 01, 2012 6,00% 0,4868% Dec 01, 2011 6,00% 0,4868% Nov 01, 2011 6,00% 0,4868% Oct 01, 2011 6,50% 0,5262% Sep 01, 2011 6,75% 0,5458% Aug 01, 2011 6,75% 0,5458% Jul 01, 2011 6,75% 0,5458% Jun 01, 2011 6,75% 0,5458% May 01, 2011 6,75% 0,5458% Apr 01, 2011 6,75% 0,5458% Mar 01, 2011 6,75% 0,5458% Feb 01, 2011 6,75% 0,5458% Jan 01, 2011 6,50% 0,5262% Dec 01, 2010 6,50% 0,5262% Nov 01, 2010 6,50% 0,5262% Oct 01, 2010 6,50% 0,5262% Sep 01, 2010 6,50% 0,5262% Aug 01, 2010 6,50% 0,5262% Jul 01, 2010 6,50% 0,5262% Jun 01, 2010 6,50% 0,5262% May 01, 2010 6,50% 0,5262% Apr 01, 2010 6,50% 0,5262% Mar 01, 2010 6,50% 0,5262% Feb 01, 2010 6,50% 0,5262% Jan 01, 2010 6,50% 0,5262% Dec 01, 2009 6,50% 0,5262% Nov 01, 2009 6,50% 0,5262% Oct 01, 2009 6,50% 0,5262% Sep 01, 2009 6,50% 0,5262% Aug 01, 2009 6,50% 0,5262% Jul 01, 2009 6,75% 0,5458% Jun 01, 2009 7,00% 0,5654% May 01, 2009 7,25% 0,5850%
Apr 01, 2009 7,50% 0,6045% Mar 01, 2009 7,75% 0,6240% Feb 01, 2009 8,25% 0,6628% Jan 01, 2009 8,75% 0,7015% Dec 01, 2008 9,25% 0,7400% Nov 01, 2008 9,50% 0,7592% Oct 01, 2008 9,50% 0,7592% Sep 01, 2008 9,25% 0,7400% Aug 01, 2008 9,00% 0,7207% Jul 01, 2008 8,75% 0,7015% Jun 01, 2008 8,50% 0,6821% May 01, 2008 8,25% 0,6628% Apr 01, 2008 8,00% 0,6434% Mar 01, 2008 8,00% 0,6434% Feb 01, 2008 8,00% 0,6434% Jan 01, 2008 8,00% 0,6434%
Appendix 3: Portfolio Average Returns for the Entire Period
Date Portfolio 1 Portfolio 2 Portfolio 3 Portfolio 4 Portfolio 5 Portfolio 6
Dec 01, 2013 -11.141% -0.341% -1.333% 3.732% -3.293% -0.100% Nov 01, 2013 -9.723% -6.373% -7.323% -5.846% -10.302% -3.344% Oct 01, 2013 10.130% 7.730% 6.562% 7.872% 0.794% 3.732% Sep 01, 2013 5.873% 5.083% -4.445% -0.318% 4.101% -2.995% Aug 01, 2013 -12.628% -9.168% -4.413% -0.223% -16.662% -9.782% Jul 01, 2013 -20.940% -6.981% -5.518% -16.738% -3.292% -4.748% Jun 01, 2013 -6.420% -8.863% -13.563% -7.078% -7.690% -4.995% May 01, 2013 3.170% -8.254% 3.690% 11.303% 9.129% 3.200% Apr 01, 2013 1.256% 0.521% 5.556% 0.926% -0.578% 3.207% Mar 01, 2013 7.148% -2.650% 5.231% 3.135% 7.679% 1.027% Feb 01, 2013 11.478% 8.488% 14.583% 5.778% 19.080% 3.324% Jan 01, 2013 8.946% 9.173% 11.203% 1.576% 11.673% -1.573% Dec 01, 2012 -3.675% -4.580% -0.125% 7.960% 5.226% -0.590% Nov 01, 2012 12.565% 3.071% -3.454% -3.057% -7.331% 0.710% Oct 01, 2012 8.160% 4.085% -3.655% -0.885% 1.473% 3.497% Sep 01, 2012 24.836% 12.438% 9.953% 7.228% 3.516% 4.649% Aug 01, 2012 -7.555% -5.454% -5.755% -2.234% -1.174% -1.571% Jul 01, 2012 -1.022% 5.795% 6.015% 5.111% 7.549% 6.847% Jun 01, 2012 5.341% 4.471% 0.956% 0.793% 0.783% 11.386% May 01, 2012 -8.845% -12.079% -14.772% -14.277% -2.116% -6.141% Apr 01, 2012 2.295% 7.448% 5.205% -4.345% 0.043% 8.149% Mar 01, 2012 10.081% 11.003% 6.148% 3.498% 4.982% 4.229% Feb 01, 2012 3.828% 7.086% -0.389% 3.890% 5.804% 1.810% Jan 01, 2012 9.530% 4.993% 5.700% 1.008% 1.680% 3.402% Dec 01, 2011 7.022% 8.392% 8.210% 2.222% 2.702% 11.335% Nov 01, 2011 -10.475% -5.823% -1.038% 0.865% -0.651% -4.497% Oct 01, 2011 11.482% 8.887% 3.637% 6.597% 8.242% 5.280% Sep 01, 2011 -13.167% -10.356% -4.662% -9.389% -6.872% -10.649% Aug 01, 2011 2.046% -8.386% -10.199% -2.947% -5.504% -6.619% Jul 01, 2011 20.066% 8.644% 6.794% 2.656% 11.643% 5.803% Jun 01, 2011 -1.096% 4.899% -2.417% 0.653% 0.621% -2.827% May 01, 2011 -1.274% 0.033% 3.941% 0.279% 0.460% 2.574% Apr 01, 2011 4.693% 3.076% -2.352% 1.169% 3.527% 8.673% Mar 01, 2011 10.919% 8.886% 0.749% 9.018% 6.860% 5.280% Feb 01, 2011 2.948% 2.476% 10.476% -1.904% 1.716% 5.497% Jan 01, 2011 -14.418% -10.860% -5.068% -9.010% -12.696% -6.298% Dec 01, 2010 1.852% 1.157% 2.334% 0.814% 12.732% 13.727% Nov 01, 2010 4.289% -6.713% -6.438% 1.144% 0.012% 23.150% Oct 01, 2010 5.109% 5.942% 7.480% 1.067% 12.742% 13.182% Sep 01, 2010 14.625% 15.482% 9.684% 19.447% 16.434% 11.768% Aug 01, 2010 2.655% -6.388% -2.023% 7.705% 1.245% 4.201% Jul 01, 2010 9.670% 2.532% 5.114% 19.455% 0.381% 1.790% Jun 01, 2010 4.522% 7.985% 1.262% 14.465% 0.520% 2.127% May 01, 2010 -11.675% -13.030% -11.196% -4.633% -3.992% -2.409% Apr 01, 2010 15.150% 21.122% 5.817% 10.845% 9.472% -2.496% Mar 01, 2010 10.845% 11.680% 11.557% 10.355% 12.280% 9.730% Feb 01, 2010 -4.506% 4.194% -1.953% 0.734% 2.201% -8.590% Jan 01, 2010 -0.360% 0.580% 9.517% 11.062% 5.455% 12.267%
Appendix 4: Constructed Portfolios and Returns per Sub Period
Portfolio Number
Sub Periods
1 2 3 4
Company Beta Average
Return Company Beta
Average
Return Company Beta
Average
Return Company Beta
Average Return 1 BSDE 1.883 4.348% BKSL 2.313 1.562% CPIN 2.196 4.628% AKRA 2.146 -1.0708%
ITMG 1.651 BSDE 2.214 ADHI 2.064 SMRA 1.930
BBNI 1.571 SMRA 1.759 SMRA 1.867 CPIN 1.791
UNTR 1.451 BMRI 1.603 MLPL 1.805 SSIA 1.674
ADHI 1.394 BBNI 1.546 AKRA 1.646 BSDE 1.625
LSIP 1.381 ITMG 1.539 ASRI 1.633 PTBA 1.458
2 BKSL 1.298 3.712% ASII 1.381 0.822% MNCN 1.556 3.190% BBRI 1.449 -0.9696%
INDF 1.276 BBRI 1.375 BMRI 1.550 ADRO 1.408
ASII 1.263 INDF 1.350 WIKA 1.494 SMGR 1.381
BMRI 1.235 ADHI 1.324 BSDE 1.445 UNTR 1.335
ASRI 1.213 ASRI 1.313 CTRA 1.415 BMRI 1.327
BDMN 1.198 UNTR 1.176 BBRI 1.286 ASRI 1.307
3 PTBA 1.180 2.596% WIKA 1.128 0.672% LPKR 1.279 0.486% BBNI 1.245 0.8525%
AALI 1.151 MNCN 1.104 ASII 1.246 WIKA 1.205
SMRA 1.145 BDMN 1.091 BKSL 1.241 ITMG 1.171
INTP 1.134 CTRA 1.075 PTBA 1.192 MNCN 1.115
BBRI 1.125 BMTR 1.062 BBNI 1.183 MLPL 1.104
CTRA 1.078 PTBA 1.043 SSIA 1.163 ASII 1.104
4 WIKA 1.016 7.705% BBCA 0.970 0.017% INDF 1.127 0.391% LSIP 1.079 0.3432% KLBF 1.012 LSIP 0.951 ADRO 1.107 LPKR 1.062
CPIN 1.012 INTP 0.910 ITMG 1.036 INTP 1.042
AKRA 0.989 JSMR 0.837 SMGR 1.035 BBCA 0.994
GGRM 0.916 GGRM 0.826 BBCA 1.029 AALI 0.989
EXCL 0.907 MLPL 0.817 UNTR 0.980 CTRA 0.943
5 PGAS 0.906 5.790% SMGR 0.739 0.837% INTP 0.947 1.703% PWON 0.916 0.8865%
JSMR 0.816 AKRA 0.689 AALI 0.807 ADHI 0.886
PWON 0.729 PGAS 0.689 JSMR 0.760 KLBF 0.817
MNCN 0.717 CPIN 0.602 PGAS 0.755 PGAS 0.794
BMTR 0.699 KLBF 0.602 LSIP 0.725 BKSL 0.788
SMGR 0.691 UNVR 0.413 EXCL 0.698 MAIN 0.609
6 TLKM 0.675 6.537% LPKR 0.376 1.129% BDMN 0.575 3.031% JSMR 0.561 -1.0872% BBCA 0.581 AALI 0.296 BMTR 0.525 BDMN 0.523 MLPL 0.436 ADRO 0.236 GGRM 0.520 BMTR 0.489 LPKR 0.145 PWON 0.114 PWON 0.485 GGRM 0.410 UNVR 0.137 SSIA 0.091 TLKM 0.256 TLKM 0.272
SSIA 0.093 EXCL 0.004 UNVR 0.237 EXCL 0.199
