Stylized Facts of the Statistical Properties of Risk and Return of the Dhaka Stock Exchange: 1991-2015
While the role of financial market, particularly the stock market, in promoting economic growth through efficient allocation of capital is well recognized, the investors of the developing economies have little knowledge about the return and risk of the markets they operate in. Even in the development discourse of Bangladesh the role of the security market is mostly absent largely because there is hardly any robust analysis on it. The current literature is mostly outdated and fails to ask the most fundamental first order questions such as what is the rate of return of the capital market? What is the extent of equity premium? What is the extent of trade-off between return and risk? Therefore, constructing a set of stylized facts of the capital market of Bangladesh with regard to return and risk is essential for providing feedbacks into the broader discussions.
In this study we study the behavior of stock return and its volatility. To this end, we compile daily stock market data for all listed securities from Dhaka Stock Exchange for the years 1991-2015 which contain a few million observations (1,071,312 to be exact) on the following variables: closing price of the day, total number of shares issued by the firm, bonus share, right share and cash dividend. We construct the adjusted closing prices of the stocks to calculate the compound rate of return. We identify twelve stylized facts on daily return and its volatility for the period 1991-2015 of the stocks listed in DSE.
Stylized fact 1: 2000s is the decade of high positive return, preceded and followed by negative decadal returns. The average daily rate of return was about 0.0034 %, (1.10 percent per annum) for the period 1991- 2015, it fluctuated significantly over the entire period of our study. While the daily returns fluctuated over time, variations of daily returns across shares remain more or less stable, measured by standard deviation of stock prices.
Stylized fact 2: The rate of return of manufacturing companies outperforms other sectors on average. The daily return for service and financial sectors for the entire study period was the same (-0.002 percent) but in the period of high return (2001-2010) financial sector outperforms the manufacturing sector.
Stylized fact 3: Old is gold. The old firms (which started transaction before 2000) outperformed the new ones (firms which started transaction after 2000) with daily average return of 0.011 percent since 1991 and while for new firms the average daily return was -0.01 percent.
Stylized fact 4: Categorization based on quality truly reflects the relative return. The daily average returns for A, B and Z were 0.007, 0.000 and -0.003 percent respectively, suggesting that the categorization of quality truly reflects the relative return.
Stylized fact 5: Monthly volatility of daily return has increased over time. During the period of 1991-2001, the monthly volatility of the daily return was 1.37 percent, which increased to 1.69 percent in the following decade and then further accelerated to 2.00 percent during 2011-15.
Stylized fact 6: While volatility across sectors are very similar, there is significant heterogeneity in volatility among stocks of different quality and age. The breakdown of sectors, quality and age of the stocks reveal that the volatility is 1.70 percent for both manufacturing and service sectors and for financial sector it is slightly higher (1.75 percent). Category A had much less volatility compared to other categories accompanied by higher return, the new stock has higher volatility (1.74 percent) than that of old stocks (1.35 percent).
Stylized fact 7: Distribution of daily stock return is not normal. The distribution of daily returns has asymmetric tails and the coefficients of skewness and kurtosis are statistically different from the normal distribution.
Stylized fact 8: Daily returns don’t follow random walk. The idea behind the random-walk model suggests that all information present in the market is immediately reflected in the price; price moves only with the advent of new information which is random and unpredictable, thus making price changes unpredictable and random. We perform autocorrelation test, runs test and variance ratio test to triangulate our results. All three tests suggest that the return series do not follow a random walk.
Stylized fact 9: Market is not “efficient”. We test the stock market efficiency of DSE in its weak form by applying autoregressive integrated moving average model (ARIMA). In the ARMA model the MA (1) term is statistically significant at 1 percent level which suggests that past returns have an effect on the current market price which is a violation of the weak form of the EMH i.e. violation of informational efficiency.
Stylized fact 10: There is “volatility clustering”. Our analysis suggests that period of high volatility are followed by periods of high volatility and period of low volatility tend to be followed by periods of low volatility implying that the error term is conditional heteroskedastic and can be represented by ARCH and GARCH.
Stylized fact 11: There is a risk return trade-off; one percentage point increase in return is associated with 1.6 percentage points of risk. We investigate the risk-return relationship for DSE index using GARCH-in-mean model. The coefficient of the risk-return parameter of the model came out positive (0.016) and statistically significant suggesting that the investors are compensated with high return in times of high volatility.
Stylized fact 12: There is a substantial heterogeneity in risk-return tradeoffs among stocks of different sectors, quality and age. In order to compare the risk-return trade-offs we rely on panel data estimation and run a host of fixed effects. Overall, one percentage point increase in return is coupled with 3.7 percentage points of risk (SD). Sector wise disaggregated analysis reveals that the risk-return tradeoffs are very similar for manufacturing and service sectors, 4.2 and 4.1 percentage points respectively. The tradeoff is the lowest for the financial sector for which one percentage point increase in return is associated with 2.1 percentage point increase of risk. In the case of quality of stock, “A” category stands out. The risk-return tradeoff is much higher for older stocks (3.8) compared with the new ones (3.1).