3. Chittagong Stock Market Of Bangladesh Turning Of A Weak-form Market Into An Efficient Market

SAARC Journal of Human Resource Development 2007

Chittagong Stock Market of Bangladesh: Turning of a WeakForm Market into an Efficient Market Mohammed Abdullah Raihan1 Mohammad Anwar Ullah2

Overview The Chittagong Stock Exchange (CSE) is a not-for-profit organization, formed and registered with the registrar Joint Stock Companies and Firms in Bangladesh on April 1, 1995 as a public company limited by guarantee with an Authorized capital of 15 0’000’000 divided into 500 shares of Tk. 300’000 each. The Exchange members are not its beneficiaries since they are not involved in profit sharing and taking dividend. All its surpluses are spent on the development of capital market in the country. The principal activities of the Exchange are to conduct, regulate and control the trade. Starting from a rental building, the exchange currently owns a two-storey building measuring 28’000 sft. It is the second stock exchange of Bangladesh that started its journey with the aim of offering the investors a transparent and efficient capital market. On October 10, 1995 CSE introduced a fully automated screen based trading system replacing the obsolete setup enabling its trade operations from three major cities in Bangladesh. In the backdrop of a strong desire to institute a dynamic, automated and a transparent Stock Exchange in the country, seventy reputed business personalities under the leadership of the founder president Amir Khosru Mahmud Chowdury3, , were chosen to establish and run the exchange. Only 30 securities were listed on the first day trade when market capitalization stood at US$0.2 Billion. Now CSE is facilitating investors through a fully automated exchange with screen based trading facility on countrywide communication networks. CSE started online trading system on the 30th may 2004. CSE offers buying and selling and dealing in shares securities, bonds, debentures, govt. papers and any other instruments through brokers and dealers. It is also involved in disseminating information to investors by publishing monthly portfolio and other necessary publications. The exchange is also involved in research and development activities pertaining to capital market.

1 2 3

Lecturer, Int’l Islamic University, Bangladesh. Email: [email protected] Lecturer, Eastwest University, Dhaka MP and former Minister for Commerce

26

Chittagong Stock Market of Bangladesh:

CSE at a glance Type of organization: Authorized capital: Paid up capital: Number of members: No. of listed companies: No. of listed Mutual funds: No. of listed Debentures: CSE all share Index: CSE-30 Index: Market capital: Trading days: Trading procedure: On line trading system:

A Not-for-profit public limited company by guarantee. TK. 150 Millions TK. 38.7 Millions ($0.72 millions) 129, all members are corporate bodies 212 10 04 1586.65 1457.71 TK. 56.36 Billions ($ 1.037 Billions) Sunday to Thursday. Automated Trading System. 30th may, 2004.

Abstract: The vast majority of efficient market research to date has focused on the major United States and European securities markets. A fewer have investigated the developing and less developed countries’ markets and no study has been conducted on the Chittagong Stock Exchange (CSE). The study seeks evidence supporting the existence of at least weak-form efficiency of the market. The sample includes the daily price indices of all the listed securities on the CSE for the period of 10/10/1995 to 19/01/2004. The hypothesis of the study is whether the Chittagong Stock Market is weak-form efficient. The results of parametric test – auto-correlation test, auto-regression & ARIMA model – provide evidence that the share return series do not follow random walk model and the significant autocorrelation co-efficient at different lags reject the null hypothesis of weak-form efficiency. The results are consistent in different sub-sample observations, without outlier and for individual securities. The issues are important to security analysts, investors and security exchange regulatory bodies in their policy-making decisions to improve the market condition. The study warrants the continuity of research for a conclusive analysis and synthesis concerning the level of efficiency of the less developed market. Introduction The random walk theory asserts that price movements do not follow any patterns or trends and that past price movements cannot be used to predict future price movements. There are three forms of the efficient market hypothesis: firstly, the "weak" form asserts that all past market prices and data are fully reflected in securities’ prices Or in other words, a technical analysis is of no use; secondly, the "Semi strong" form asserts that all publicly available information is fully

27

SAARC Journal of Human Resource Development 2007

reflected in securities’ prices or in other words, fundamental analysis is of no use; and lastly, the "strong" form asserts that all information is fully reflected in securities’ prices or in other words, even insiders’ information is of no use. It is usually believed that the markets in developing and less developed countries are not efficient in semi-strong form or strong form. The study seeks evidence of weak-form efficient market hypothesis (WFEMH) in a less developed emerging market like CSE. It is very much convenient to test the weak-form efficiency of the market rather than semi-strong form and strong form efficiency. The test of semi strong form and strong form efficiency is very rare in less developed countries because of the absence of sufficient data in a convenient form. The other impeding factors include: structural profile, inadequate regulations, lack of supervision and administrative slackness towards the application of existing rules. In Addition, companies’ information is released and circulated before their annual report are published and officially available. The annual reports of some of the listed companies are mistrusted and often perceived as speculative by the market. The market moved dramatically over a period of time and turned into a speculation market and then a gamble market transforming the investors into speculators. Moreover, share price indices data are available and reliable to test the weakform efficiency of the market. The empirical research on market efficiency can be divided into two broad categories: one is technical analysis, which is mainly concerned with testing for the availability of exploitable information from the past security prices that is widely used in examining the weak-form efficient market hypothesis; the other is fundamental analysis, which rests on the assumption that factors other than the past security prices are relevant in the determination of the future prices. The first category of WFEMH testing can be divided into two sub approaches: one is to determine the existence of predictability using the past return series or price information; another is to use technical trading rules if they can be exploited as profit making strategy. The aim of the study is to test the former on the CSE. The study restricts attention exclusively for WFEMH or return predictability using time-series analysis of stock return behavior in an emerging market. The remainder of the study is structured as follows: Section 1 discusses the concepts and interactions between weak-form market efficiency and emerging market; section 2 reviews the previous empirical evidences on weak-form efficiency; section 3 dilates upon the data and the research method, section 4 describes the variables used in the analysis; section 5 lays out the empirical results of the hypotheses while the last section, section 6 presents the summary and conclusions. Correlation of Weak form Market Efficiency and Emerging Markets A few studies conducted on the test of efficient market hypothesis (EMH) in emerging markets compared to the volume of studies published on the

28

Chittagong Stock Market of Bangladesh:

developed market. It is generally assumed that the emerging markets are less efficient than the developed market. The definition of emerging market highlights the potential growth as well as a rapid growth of the size of the market. However, it is likely that the market participants are not well-informed and behaving irrationally as compared to the well organized markets. The causes behind the lack of financial development, especially in capital markets are caused by certain market imperfections such as transaction costs, lack of timely information, cost of acquiring new information, and possibly a greater uncertainty about the future (Taylor, 1956; Goldsmith, 1971; Mason, 1972; Wai and Patrick, 1973). Different researchers define the emerging market in different ways. According to Samuel’s (1981): “Prices can not be assumed to fully reflect all available information. It cannot be assumed that investors will correctly interpret the information that is released. The corporation has greater potential to influence its own stock market price and there is a greater possibility that its price will move about in a manner not justified by the information available.4” Emerging markets are also defined in terms of policy-making decisions: “A realization of inefficiencies inherent in command and control policies and the tighter lending policies of international creditors have led the developing countries to re-define the role of domestic equity markets in their economies. Most countries have adopted policies that make the allocation of equity capital more responsive to market forces. These policy changes have resulted in remarkable growth in the size of the equity markets in the developing world, commonly known as ‘Emerging Stock Markets’ (ESMs ).5 ” And with this open market policy in the emerging markets, speculations are common. It is believed that the large investors can easily speculate the market. As a less organized market without market makers and timely available information, there always remains a possibility for large investors and insiders to make high profits. The ability to predict stock price changes based on a given set of information lies behind the notion of stock market efficiency: the lower the market efficiency, the greater the predictability of stock price changes.

4 Samuels, J.M and N.Yacout (1981) , “Stock Exchanges in Developing Countries”, Savings and development no.4.Page 129 5 Hussain, Fazal, (1996), “Stock price Behaviour in an Emerging Market: A case Study of Pakistan”, Ph D thesis. The Catholic University of America. abstract

29

SAARC Journal of Human Resource Development 2007

The WFEMH tests measure whether the past series of share prices or returns can be used to successfully predict future share prices or returns. The major empirical investigation of the above test measures the statistical dependence between price changes. If no dependence is found, i.e., price changes are random, then this provides evidence in support of the WFEMH, which implies that no profitable investment trading strategy can be derived based on the past prices. On the other hand, if dependence is found, for example, price increases are generally followed by further increases in the next period and vice versa, clearly indicates that this can be the basis of profitable investment rule in violation of the WFEMH assumption. However, whether any trading rule is profitable depends largely on the operating cost, such as brokerage cost, interest cost, trading settlement procedure, and whether transactions can be made at the exact prices quoted in the market. In general, the results of previous research provide evidence that the markets of developed economies are generally weakform efficient. In other words, the successive returns are independent and follow random walk (Fama1, 1965, 1970). On the other hand, the research findings on the market of developing and less developed countries are controversial. Some of the researchers found evidence of weak-form efficiency and could not reject the random-walk hypothesis in emerging markets (Branes, 1986; Dickinson and Muragu, 1994; Urrutia, 1995; Ojah and Karemera, 1999). Whereas, the others found the evidence of nonrandomness stock price behavior and reject the weak-form efficiency in the developing and emerging markets (Roux and Gilberson, 1978; Harvey, 1994; Claessens, Dasgupta and Glen, 1995; Poshakwale. S, 1996 and Nourredine Khaba, 1998). Considering the overwhelming evidence, the test of WFEMH in CSE is of interest in its own right to reach an ultimate conclusion about the level of efficiency in developing and less developed emerging markets in general. Review of Empirical Evidence on Weak-Form Efficiency The early studies started on testing weak-form efficiency of the developed market, generally agree with the support of weak-form efficiency of the market considering a low degree of serial correlation and transaction cost (Working, 1934; Kendall, 1943, 1953; Cootner, 1962; Osborne, 1962; Fama, 1965). All of the studies support the proposition that price changes are random and the past changes were not useful in forecasting the future price changes, particularly after transaction costs were taken into account. However, there are some studies which found the predictability of share price changes, for example, Fama and French, 1988; Poterba and Summers, 1988, in developed markets but they did not reach to a conclusion about profitable trading rules. Poterba and Summers (1988) suggest that the noise trading, the trading by investors whose demand for shares is determined by factors other than their expected returns provides a plausible explanation for the transitory component

30

Chittagong Stock Market of Bangladesh:

in stock prices, and they suggest constructing and testing theories of noise trading as well as theories of changing risk factors could account for the characteristics of stock returns auto-correlogram they found. Fama and French (1988) conclude that auto-correlations may reflect market inefficiency or timevarying equilibrium expected returns generated by rational investor behavior and neither view suggests, however, the patterns of auto-correlation should be stable for a long sample period. Hudson, Dempsey and Keasey (1994) found that the technical trading rules have predictive power but not sufficient to enable excess return in U.K market. Similarly, Nicolaas, (1997) also conclude that the past returns have predictive power in Australian market but the degree of predictability of return is not so high. Overall, the empirical studies on developed market show no profitability from using the past records of price series, supports the weak-form efficiency of the EMH in general. On the other hand, the research findings of weak-form efficiency on the market of developing and less developed markets are controversial. Most of the less developed market suffers with the problem of thin trading. In addition, in smaller markets, it is easier for large traders to manipulate the market. Though it is generally believed that the emerging markets are less efficient, the empirical evidence does not always support the thought. A group of researchers finds the weak-form efficiency in developing and less developed markets despite the problems of thin trading. The group is represented by Branes (1986) from the Kuala Lumpur Stock Exchange, Chan, Gup and Pan (1992) from major Asian markets, Dickinson and Muragu (1994) from the Nairobi Stock Exchange, and Ojah and Karemera (1999) from the four Latin American countries’ markets. On the other hand, another group, who offers substantial evidence claims that the developing and less developed markets are not efficient in weak-sense are Cheung, Wong and Ho, (1993), on the stock market of Korea and Taiwan; in a world bank study by Claessens, Dasgupta and Glen (1995), report significant serial correlation in equity returns from 19 emerging markets and suggest that stock prices in emerging markets violates weak-form EMH; similar findings are reported by Harvey (1994) for most emerging markets. Nourrrendine Kababa (1998) has examined the behaviour of stock price in the Saudi Financial market seeking evidence that for weak-form efficiency and find that the market is not weak-form efficient. He explained that the inefficiency might be due to delay in operations and high transaction cost, thinness of trading and illuiquidity in the market. Roux and Gilberson (1978) and Poshakwale S. (1996) find the evidence of non-randomness stock price behavior and the market inefficiency (not weakform efficient) on the Johannesburg stock Exchange and on the Indian market. In short, the review of previous studies shows that the developed markets are generally weak-form efficient. But the dynamics of emerging market equities require clarification. The comparison and needed additional information on

31

SAARC Journal of Human Resource Development 2007

equity price dynamics are important segments of the world’s emerging capital markets. So it is an interesting empirical question whether and to what extent, this is also the case with less developed market stock exchanges, and the review of the evidence gathered through the previous empirical studies raise some research questions. Is the Chittagong Stock market, as a less developed emerging market, a weak-form efficient or not? How far it deviates from idealized EMH? What return generating process drives the emerging equity market series? Is the conflicting result a function of the research methodology employed? Is it possible to build up a predictive model? What are the implications of the findings? These issues are empirically examined in the following section. Sample and Empirical Method The empirical analysis of the study uses daily market return of the Chittagong Stock Exchange for the period of 10th October 1995 to 31st December 2003 and the monthly market return for the period of October 1995 to December 2003. There are some additional models, such as auto-regression, Auto-regressiveIntegrated-Moving Average (ARIMA), which are employed to confirm the results and to build up a predictive model. Thin or infrequent trading can introduce serious bias in empirical work. In order to avoid the possible bias, we use a longer time-period, which reduces the problem of non-trading bias and increases the power of random walk test. We use corellogram test, Auto correlation test and random walk test both to compare the results considering that non-normal distribution can bias the findings. In choosing the methodology of weak-form efficiency test, we have considered the following issues: a)

The research needs triangulation between the developed and less developed market. Triangulation in research may be both theoretical and implemental through the use of different research methods, different settings, different data and improved decision making techniques.

b) The study considers both traditional, such as descriptive statistics, autocorrelation test, and dynamic time series model, such as Autoregression (ARIMA) model, which perhaps claims better findings. c)

32

Moreover, the robustness of the results is assessed in various ways: firstly, similar tests are conducted for various sub-samples of the original sample and by trimming outlying observations; secondly, the study considers individual actively traded shares return to get results free of thin trading bias; thirdly, the study includes some alternative variables, such as daily market return and monthly market return, to confirm the results; and finally,

Chittagong Stock Market of Bangladesh:

the use of different testing procedures helps to reach a conclusion of consistency in the findings. Sample Period The sample includes a total 2246 daily and 84 monthly observations for the entire sample period, 1995 to 2003. To confirm the results of the empirical analysis; we also compute the first sub-sample (1995-1999, first 5 years), the second sub-sample (2000-2003, last 4 years) and with observations excluding the outliers. Variables The monthly market returns are used as an individual time series variable. The non-availability of computerized databases has had a significantly limiting effect on market studies in our country, and consequently on the volume of published evidence. One probable solution to this problem is to use the indices of the index, which are published and readily available at a low cost. The market returns are calculated from the monthly price indices without adjustment of dividend, bonus and right issues. The monthly share price indices include all the listed companies stock. Many researchers confirm that their conclusions remain unchanged whether they adjusted their data for dividend or not. Given the general stock price index

P

t

(ignoring dividend payments), the

continuously compounded monthly percentage stock return series approximated by the following equation:

R

t

= ln

p − ln p t

R can be t

t −1

p

t stands for monthly closing CSE stock price index for month‘t’ and Where ‘ln’ stands for natural logarithm. Note that the CSE index is a market capitalization weighted all share price index in the Chittagong Stock Exchange. The primary variable examined in this paper is the monthly stock return

R

series t . The paper will discuss some important statistical properties of the monthly stock returns. These include analyses of the mean and median values, maximum and minimum values, standard deviation and studentized rage, and skewness and Kurtosis. All of these statistics will be calculated for the total sample as well as for each month of the year over the entire sample period.

33

SAARC Journal of Human Resource Development 2007

Hypotheses and Empirical Results The hypotheses of the study and the empirical results of individual tests on weak-form efficiency are described in two subsections. Hypotheses The study seeks evidence whether the Chittagong Stock market follows random walk model or the market is weak-form efficient. H01: The Chittagong stock market follows random –walk model. H02: The Chittagong stock market is efficient in weak-form. Objectives and Methodology Step-1: Check the stationarity of the return series, and, if necessary, transform the series to induce stationarity. Step-2: From the examination of the data series as well as the autocorrelation and partial autocorrelation functions of the series (transformed series for nonstationary case) choose a few ARMA specifications for estimation and testing in order to arrive at a preferred specification with white noise residuals. Step-3: Calculate forecasts over a relevant time horizon from the preferred specification. Stationarity Check Graphical analysis Figure 1 shows the graph of return indices for the period between 1995:11 and 2004:01. In the figure, what we notice at the first glance is not the presence of a linear time trend. It is very difficult to get any idea about the stationarity of the series from this graph. Figure 1: Monthly Return Index

34

Chittagong Stock Market of Bangladesh:

The Unit root test This view carries out the augmented Dickey-Fuller and Phillips-Perron unit root tests for whether the series is stationary. EViews performs two widely used unit root tests: the Dickey-Fuller (DF) and augmented Dickey-Fuller (ADF) test. In this section, we provide some theoretical background for the test. To illustrate the use of Dickey-Fuller tests, consider first an AR (1) process:

R

= α + ρ Rt −1 + ε t

t

parameter is to be estimated and

where

ε

t

α is

the vector of constant, ρ is the

is assumed to be white noise. If

ρ ≥1

,

R is a nonstationary series and the variance of R increases with time and approaches infinity. If ρ ∠1 , R is a stationary series. Thus, the hypothesis of t

t

t

stationarity can be evaluated by testing whether the absolute value of strictly less than one. The ADF test uses the modified equation:

ρ

is

[2], which suggests estimating the following

∆ Rt = ρ Rt −1 + β ∆ Rt −1 + β ∆ Rt − 2 + ...... + β ∆ Rt − q + ϑ t 2 q 11

R denotes the monthly return of the stock prices and R − R , α = ρ − 1 and the null hypothesis H : α = 0 is tested :α p 0 based on the ADF-t statistic. We will use against the alternatives, H

Where ∆ Rt =

t

t

t −1

0

1

the critical values provided by McKinnon (1996) to evaluate the null hypothesis. Observing figure 1, we include a constant as a regressor in the test equation. The lag order, 0 in this case, of the difference terms is determined by the Modified Akaike Information Criterion (MAIC). The result of the unit root test is shown below: The significance of all the coefficients and the value of DW statistics close to 2 indicate the correct specification of the test equation. The ADF test statistic is highly significant. The above results clearly reject the unit root hypothesis. Thus, we may consider the monthly return of the stock prices as stationary. So the Chittagong Stock Market is not efficient in weak-form.

35

SAARC Journal of Human Resource Development 2007

Table 1: Test of Unit Root ADF Test Statistic -6.936708

1% Critical Value* 5% Critical Value 10% Critical Value

-3.4986 -2.8912 -2.5824

*MacKinnon critical values for rejection of hypothesis of a unit root.

Augmented Dickey-Fuller Test Equation Dependent Variable: D(RETURN) Method: Least Squares Date: 06/08/04 Time: 00:09 Sample(adjusted): 1996:01 2004:01 Included observations: 97 after adjusting endpoints Variable RETURN(-1) C

Coefficient

Std. Error

-0.672234 -0.007073

R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat

0.336212 0.329224 0.118779 1.340300 70.03113 1.921998

0.096910 0.012095

t-Statistic -6.936708 -0.584801

Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion F-statistic Prob(F-statistic)

Prob. 0.0000 0.5601 -0.000677 0.145028 -1.402704 -1.349617 48.11792 0.000000

Model Specification Autoregressive Integrated Moving Average (ARIMA) models are generalizations of the simple AR model that use three tools for modeling the serial correlation in the disturbance. The first tool is the autoregressive, or AR, term. The AR(1) model introduced above uses only the first-order term but, in general, we may use additional, higher-order AR terms. Each AR term corresponds to the use of a lagged value of the residual in the forecasting equation for the unconditional residual. An autoregressive model of order ρ ,

ρ

u =ρ u +ρ u

+ ........... + ρ

uρ + ε

1 2 ρ AR( ) has the form: The second tool is the integration order term. Each integration order corresponds to differencing the series being forecast. A first-order integrated component means that the forecasting model is designed for the first difference of the original series. A second-order component corresponds to using second differences, and so on. t

1

2

t

The third tool is the MA, or moving average term. A moving average forecasting model uses lagged values of the forecast error to improve the current forecast. A first-order moving average term uses the most recent forecast error; a

36

Chittagong Stock Market of Bangladesh:

second-order term uses the forecast error from the two most recent periods, and so on. An MA(q) has the form:

u = ε +θ ε t

t

1

t −1

+ θ 2ε

t −2

+ ....... + θ q ε t − q

Following through the Box-Jenkins procedure, we first observe autocorrelation and partial autocorrelations of the monthly return series of the Chittagong Stock Exchange. The figure 2 shows the autocorrelation function (ACF) and partial autocorrelation function (PACF) (up to lag 36) for the CSE return index that covers the period 1995:11 to 2004:01. Figure 2: Autocorrelation and Partial Autocorrelation Functions of the Monthly Return of the CSE Index

The formal ADF test already rejected the nonstaionarity hypothesis. The single spike at lag 1 in the partial autocorrelation function suggests an AR (1) model.

37

SAARC Journal of Human Resource Development 2007

Figure 3: Autocorrelation and Partial Autocorrelation Functions of the Residuals

The spikes in the partial autocorrelation function suggest an MA (0) model. The goal of ARIMA analysis is a parsimonious representation of the process governing the residual. We should use only enough AR and MA terms to fit the properties of the residuals. The Akaike information criterion and Schwarz criterion provided with each set of estimates may also be used as a guide for the appropriate lag order selection. After fitting a candidate ARIMA specification, we should verify that there are no remaining autocorrelations that our model has not accounted for. In addition to the above statistical techniques, the study employs ARIMA, the dynamic time series model to examine if the stock return series depends on its past values of the return series as well as the past and current disturbance terms. Theoretically, the weak-form efficiency of the market persists when we cannot predict the share prices from its historical price information. When the share return can be predicted on the basis of data on the past returns and on the forecasted errors, together this gives rise to ARMA model (Cuthbertson, 1996). That is to mean: if the stock price is a function of its past values of stock prices or the current and past values of the disturbance term. We use ARIMA model instead of ARMA because it included the integration process. Moreover, the

38

Chittagong Stock Market of Bangladesh:

random walk model needs to fit the model ARIMA (1,0,0), where the future value of share prices can be determined on the basis of the past information. Specifically, the future share prices will depend on the past (lag) values of share prices or on the disturbance terms. The significant coefficients, different from zero, suggest dependency of the series, which violates the assumption of the random walk model and weak-form efficiency. Results of the ARIMA analysis presented on the table 2 suggest that the price index series and return series are not following random walk model. As we know that, ARIMA (1,0,0) does not support the random walk model. Table 2: Results of ARIMA (1,0,0) for the Monthly Price Index Series Dependent Variable: RETURN, Method: Least Squares Sample (adjusted): 1996:01 2004:01 Included observations: 97 after adjusting endpoints Convergence achieved after 3 iterations Variable Coefficient Std. Error t-Statistic C -0.010522 0.017941 -0.586489 AR (1) 0.327766 0.096910 3.382174 R-squared 0.107471 Mean dependent var Adjusted R-squared 0.098076 S.D. dependent var S.E. of regression 0.118779 Akaike info criterion Sum squared resid 1.340300 Schwarz criterion Log likelihood 70.03113 F-statistic Durbin-Watson stat 1.921998 Prob (F-statistic) Inverted AR Roots .33

Prob. 0.5589 0.0010 -0.01019 0.125070 -1.40274 -1.34961 11.43910 0.001046

Diagnostic Checking Having chosen a particular ARIMA model, and having estimated its parameters, we next see whether the chosen model fits the data reasonably well, for it is possible that another ARIMA model might do the job well. One simple test of the chosen model is to see if the residuals estimated from this model are white noise; if they are, we can accept the particular fit; if not, we must start all over. Through the table 2, we see that the value of Durbin-Watson Statistic is 1.921998 i.e. almost 2. Thus, we can state that the error terms of the model are not auto correlated. In other words, residuals estimated from this model are white noise.

39

SAARC Journal of Human Resource Development 2007

Figure 4: Residual, Actual and Fitted Graph for the Estimated Equation

Forecasting It is already shown that the ARIMA(1,0,0) could pass the diagnostic tests. As noted earlier, we have estimated the model using the data that cover the period between 1995:11 and 2004:01. We retain the remaining samples between 2004:02 and 2004:12 for the out of sample forecasting. In the following figures, we have shown the statistically forecasted values of the CSE index for these eleven months. Figure 5: Static Forecast for 2004:2 to 2004:12

The forecasted series roughly mimic the actual series. The forecasting evaluation based on various forecasting error criteria is summarized the above table. Theil inequality coefficient is not close to zero so the model is not a perfect fit. We also report the component of root mean squared error. The bias proportion tells us how far the mean of the forecast is from the mean of the actual series. The variance proportion tells us how far the variation of the forecast is from the variation of the actual series. The covariance proportion measures the remaining unsystematic forecasting errors. Note that the bias, variance, and covariance proportion add up to one. If the forecast is “good”, the bias and variance proportions should be small so that most of the bias should

40

Chittagong Stock Market of Bangladesh:

concentrate on the covariance proportions. Here we observe that in the above model, the bias proportion is 0% while the variance is 50%. So the model is not a good for forecasting. Table 3: Static Forecast for 2004:2 to 20004:12

Summary and Conclusion The overall results from the empirical analysis suggest that the Chittagong Stock market of Bangladesh is not weak-form efficient. However, the results presented in the study are not above limitations. The profit-making strategy, for example, was not minutely investigated by using any technical trading rules or adjusting transaction costs, such as bid-ask spread, brokerage fee, time lag of settlement procedures etc., and consequently no conclusion were drawn in this regard. Similarly, the unavailability of value weighted index considering nonsynchronous trading may bias the results. The problems of non-trading, however, were tried to overcome by considering the individual company’s daily share return series and run test. The results of individual share returns show that they are not following a random walk model. The results found in the study should be interpreted cautiously because the presence of autocorrelation violates the assumption of random walk model, which cannot be translated as inefficiency as noted by Ko and Lee, model. “If the random walk hypothesis holds, the weak-form of the efficient market hypothesis must hold, but not vice versa. Thus, evidence supporting the random walk model is the evidence of market efficiency. But violation of the random walk model need not be evidence of market inefficiency in the weak-form (1991, p.224)6. ” The possible auto-correlation found in the return series not necessarily means that the returns are predictable. It could be owing to the presence of noise traders in the market trading by investors, whose demand for stocks is determined by factors other than their expected returns, may provide an explanation for this. 6 Ko Kwang- Soo and Lee Sang-Bin, (1991), “A comparative analysis of the Daily Behavior of Stock Returns: Japan, The US and the Asian NICs”, Journal of Business Finance and Accounting, vol. 18(2), Pp. 219-234.

41

SAARC Journal of Human Resource Development 2007

However, a lower degree of efficiency in less developed countries’ markets might be caused by the common characteristics of loose disclosure requirements as well as thinness and discontinuity in trading. The other likelihood could be the institutional factors, such as illiquidity, market fragmentation, trading and reporting delays and absence of official market makers or the delay in operations and high transaction cost, thinness of trading and illiquidity in the market. Nevertheless, measures of return behavior may be useful in research on the determinants and behaviors of flows into the stock market. The major implication of the study can be pointed out as follows. Firstly, the predictive ability is interesting for investors to beat the market using trading rules. Secondly, if the auto-correlation present in the analyses not necessarily means the rejection of Weak-form efficiency is still helpful to implement the regulatory change to prevent the bias mentioned above and to improve the overall market conditions and encourage savings and investments. The need to change the appropriate index calculation method considering the infrequent trading should be a suggestion to the responsible authorities. Thirdly, the study provides the time series behaviors of a less developed market. It is also a matter that, “despite the well-documented potential benefits of investing in the ESMs, a lack of adequate information appears to be a major factor hindering the foreign investment in these markets (Hussain, 1996)7.” And finally, it is interesting to academic researchers and explores avenues for future research. Predicting model for forecasting the future based on the past and whether the deviations are large enough to exploit profitably considering transaction cost remains open question or should be an issue for further research. The study does not include the calendar anomalies and if any trading rules can make profitable investment strategy for the test of WFEMH may be a suggestion for future research. The rejection of null hypothesis that the market is not weak-form efficient can be interpreted, as that price forming information in the particular market may not be disseminated rapidly because of sophisticated communication technology, a few numbers of business journals and lack of intensive market regulations. On the whole, this is a first attempt to judge the efficiency of the Chittagong Stock Market, which shows the stock price behavior in one of a less developed market. The necessity to stock market for the development of a country might be quarry as according to Samuels and Yacout (1981) can be stated in this respect:

7 Hussain, Fazal, (1996), “ Stock price Behaviour in an Emerging Market: A case Study of Pakistan”, Ph D thesis. The Catholic University of America.

42

Chittagong Stock Market of Bangladesh:

“ . . . there are a priori reasons to believe that stock markets in developing countries are neither efficient nor perfect. If a stock market is not efficient, this does not necessarily mean that per se it is a bad thing. The crucial question is whether an inefficient stock market is better than no market at all.8”

____________________ References Branes Paul, (1986), “Thin trading and stock market efficiency: A case of the Kuala Lumpur Stock Exchange”, Journal of Business Finance & Accounting, volume 13(4) winter, pp. 609-617. Butler, K.C., and Malaikah, S.J., (1992), “Efficiency and inefficiency in thinly traded stock markets: Kuwait and Saudi Arabia”, Journal of Banking and Finance, vol. 16, pp. 197-210. Chan Kam C. Gup Benton E., and Pan Ming-shiun, (1992) , "An Empirical Analysis of Stock Prices in Major Asian Markets and United States", The financial Review, vol-27, no-2, May1992, pp-289-307. Cheung, Yan –Leung, Wong Kie-Ann, and Ho Yan-Ki, (1993), “The pricing of risky assets in two emerging Asian markets- Korea and Taiwan”, Applied Financial Economics 3, issue 4, December, pp.315-324. Chou, Y. L., (1969), Statistical Analysts. London: Holt Rinehart and Winston. Chow, V.K., and Denning, K.D., (1993), “A simple multiple variance ratio test”, Journal of Econometrics, vol., 58, pp. 385-401. Claessens Stijin, Dasgupta Susmita, and Glen Jack, (1995); “Return behaviour in emerging Stock Market”, The world Bank economic Review, vol.9, no.1, Pp. 131-151. Cootner, P., (1962), “Stock Prices: Random vs. Systematic Changes”, Industrial Management Review, vol. 3, spring, pp.24-45. Cuthbertson, K, (1996), Quantitative Financial Economics, Wiley, England. Damodar N. Gujarati, (1988), Basic Econometrics, second edition, Mcgraw – Hill Book Company, P.691. Dickinson and Muragu , (1994), “ Market Efficiency in Developing Countries: A case study of the Nairobi Stock Exchange”, Journal of Business Finance & Accounting, volume 21(1) January, pp. 133-150. Eckbo, B.E. and Liu, J. (1993), “Temporary components of stock prices: new univariate results;”, Journal of Financial and Quantitative Analysis, vol., 28, pp. 161-176. Errunza, Vihang R. and Losq, Etienne, (1985), “The behavior of stock prices on LDC markets”, Journal of Banking and Finance, vol. 9, pp. 561-575. Fama , E. (1965), "The behavior of Stock Market Prices", Journal of Business, vol 38 , p.34-105.

8

Samuels, J.M and N.Yacout (1981) , “Stock Exchanges in Developing Countries”, Savings and development no.4.

43

SAARC Journal of Human Resource Development 2007

" Efficient Capital Markets: A review of theory and Empirical Work", Journal of Finance, vol-25, no-2(May,1970), Pp.383-417. Fama, Eugune F.,and Kenneth R.French, (1988), “Permanent and temporary components of stock market prices”, Journal of Political Economy, vol. 96, pp-246-273. Fishe, R., T.Gosnell, and D. Lasser, (1993), “Good news, bad news, volume and the Monday effect”, Journal of Business Finance & Accounting 20, 881892. Ghandi, D.K. , A. Saunders and R. S. Woodward (1980) , "Thin Capital Markets: A case study of the Kuwait Stock Market " Applied Economics, vol-12, (1980).pp.341-349. Goldsmith, Raymond W. “Capital Markets and Economic Development”, National symposium on Development of Capital Markets, September 1971. Harvey Campbell R., (1994), “Conditional Asset allocation in Emerging Markets”, Working Paper, No.4623, Cambridge, MA. Hudson, Robert. Dempsey, Michael and Keasey, Kevin., “A note on the weakform efficiency of capital markets: The application of simple technical trading rules to UK Stock prices-1935 to1994”, Journal of Banking & Finance, vol.20, Pp. 1121-1132. Hussain, Fazal, (1996), “ Stock price Behaviour in an Emerging Market: A case Study of Pakistan”, Ph D thesis, The Catholic University of America. Jagadeesh, N., (1990), “Evidence of predictable behavior of security returns”, Journal of Finance, vol.45, no-3, Pp. 881-898. Kendall, M. G., (1943), The advanced Theory of Statistics, Vol.1, London: Griffin. “The analysis of Economic Time–series”, Part1, Prices, The Journal of the Royal statistical Society, 116, pp.11-25. Ko Kwang- Soo and Lee Sang-Bin, (1991), “A comparative analysis of the Daily Behavior of Stock Returns: Japan, The US and the Asian NICs”, Journal of Business Finance and Accounting, vol. 18(2), Pp. 219-234. Lakonishok, J. and S.Smidt, (1988), “Are Seasonal Anomalies Real? A Ninety Year Perspective”, Review of Financial Studies 1, pp.402-425. Laurence, Martin M., (1986), “Weak-form efficiency in the Kuala Lumpur and Singapore Stock Markets”, Journal of Banking and Finance, vol. 10, pp.431-445. Lo, A. W., and A.C. Mackinlay, (1988), “Stock market prices do not follow random walks. Evidence from a simple specification test”, Review of Financial Studies, vol.1, pp.41-66. Mason, Robert Tempest, (1972), “The creation of risk aversion by imperfect Markets”, American economic Review, vol.62, March, pp. 77Nicolaas Groenewold, (1997), “Share market efficiency: Tests using daily data for Australia and New Zealand”, Applied Financial Economics, vol.7, Pp.645-657. Nourredine Khababa , (1998), “Behavior of stock prices in the Saudi Arabian Financial Market: Empirical research findings”, Journal of Financial Management & Analysis, Volume 11(1) , Jan-June, pp. 48-55.

44

Chittagong Stock Market of Bangladesh:

Parkinson, J, M. (1987), "The EMH and the CAPM on the Nairobi Stock Exchange" , East African Economic Review, vol.3. no.2, pp- 105-110. Poterba, James M and Lawerence H. Summers, (1988), “Mean reversion of stock prices”, Journal of Financial Economics, vol. 22, pp-27-59. Poshakwale S. (1996), “Evidence on the Weak-form efficiency and the day of the week effect in the Indian Stock Market”, Finance India, Volume 10(3), September, pp. 605-616. Roux and Gilbertson, (1978), “The behavior of share prices on the Johannesburg Stock Exchange”, Journal of Business Finance and Accounting, Volume5 (2), pp. 223-232. Samuels, J. M. (1981), "Inefficient Capital Markets and their Implications", in F.G.J. Derkinders and R.L. Crum(eds.), Risk Capital Costs and Project Financing Decision (Martinus Nijhoff, Boston/ The Hauge/ London, 1981), pp.129-148. Samuels, J.M and N.Yacout (1981), “Stock Exchanges in Developing Countries”, Savings and development no.4. Sharma, J.L. and Kennedy, Robert.E., (1977), “A Comparative analysis of stock price behavior on the Bombay, London and New York Stock Exchanges”, Journal of Financial and Quantitative Analysis, September,pp.391-413. Siegel, Sidney, (1956), “Non-parametric Statistics for behavioral sciences”, New York: McGraw- Hill Company. Strong, N, (1992), “Modelling Abnormal Returns: A Review Article”, Journal of Business Finance and Accounting, vol. 19(4), June, Pp. 533-553. Taylor, Basil, (1969), “Investment: Art, Science or what?” Lioyds bank, vol. 91, January, pp.10-21. Ojah Kalu and Karemera ,(1999), “ Random walks and Market efficiency Tests of Latin Amaeracan Emerging Equity Markets: A Revisit”, The Financial Review , volume 34, pp. 57-72. Osborne, M.F.M., (1962), “Periodic Structure In the Brownian Motion in the Stock Prices,” Operation Research, (May/June) ,Pp.345-379. Urrutia, J L , (1995), “Tests of random walk and market efficiency”, Journal of Financial Research, vol. 18, Pp. 299-309. Wai, U.T. and Patrick, H.T. (1973), “Stock and Bond issues and Capital Markets in Less Developed Countries”, International Monetary Fund Staff papers, 302. Working, H., 1934, “A random-Difference Series for the use in the Analysis of Time series,” Journal of the American Statistics Association, (March) pp.11-22. Asma Mobarek and Professor Kevin Keasey, Weak-form market efficiency of an emerging Market: Evidence from Dhaka Stock Market of Bangladesh. Anisul M. Islam, University of Houston-Downtown and Kabir Hassan, University of New Orleans; Statistical properties of monthly equity returns: some preliminary results for Bangladesh Mohammed Khaled and Ainul Islam, Test of Weak form Market Efficiency: Evidence from Dhaka Stock Exchange.

45