A Statistical Analysis of the Indian Stock Market. (Part I)

A Statistical Analysis of the Indian Stock Market. (Part I)

Hi, I am Alwar from Indian School of Mines. My interests include statistics, statistics and statistics!

I am currently working on a study of the Indian Stock Market using the Statistical methods, to identify if there is any correlation between the various parameters such as Share Price, No. of Shares, No. of Trades, etc.

Such correlations are critical, because we might be able to identify if there is a sudden change in the number of shares traded before the annual report is released, which might give some indication of insider trading.

If such a forecasting model could be developed then the model can be used to predict the everyday market scenario of a particular stock and to analyze whether the stock is following the trend of the past few days or is it differing from it.

Though a lot of Research Papers and study results were available on the internet regarding the relationships between the trading volumes and the share prices etc., most of them had concentrated on ways to predict the share prices of the stock in the near future, using these relationships. The variations in the trade volumes, share prices, etc. just before a quarterly or annual report release have not been discussed in detail.

Hence, my objective is to study the variations in the trade volumes, share prices, etc. just before a quarterly or annual report.

For this study, all the required data was gathered from the Bombay Stock Exchange (BSE) website http://www.bseindia.com/

1.0 Analysis: Identifying Correlations in Daily Share Data

1.1 Data Selected for Study

• SENSEX value
  
• Sectoral Index Value of the sector corresponding to the stock under consideration
  
• Share Price (in Rs.) of the particular stock under consideration
  
• High – Low (H – L) spread of the Share Price (in Rs.) in a day, of the particular stock under consideration
  
• Close – Open (C – O) spread of the Share Price (in Rs.) in a day, of the particular stock under consideration
  

All these data have been considered on everyday basis. i.e. the interval between the data points is 1 (working) day.

Company chosen : INFOSYS (BSE Scrip Code: 500209)

Sectoral Index : BSE – IT

Period of Study : 1^st April 2005 to 31^st May 2010.

Total No of Data Points : 1279

1.2 Method of Study:

With all the required data in place for the study, we conducted a regression analysis for the various parameters and try to understand their significance. In a regression analysis the value of R²
is perhaps the most important indicator. The value of R² shows us to what extent the predictor and the predicted variables are related. The Range of R² is 0≤ R² ≤ 1. If we convert this to a % value, then we can say R² % of the variation of the predicted variable can be attributed to the variation in the predictor variable(s).

1.3 Results of Study:

Predictor Variables	INFOSYS
	Share Price	No of Shares	No of Trades
BSE SENSEX	0.90%	0.21%	0.00%
BSE- IT	13.90%	0.62%	0.07%
Share Price		2.30%	2.40%
BSE SENSEX and BSE IT	39.00%	0.64%	0.17%
BSE- IT and Share Price		2.40%	2.60%
BSE SENSEX,BSE-IT and Share Price		2.90%	3.00%
Spread High – Low (H – L)	1.18%	17.34%	25.11%
Spread Close – Open (C – O)	0.07%	0.05%	0.01%
Spread H-L and Spread C-O	1.31%	17.36%	25.21%
Share Price and spread H-L		21.33%	29.65%
Share Price and spread C-O		2.39%	2.48%
Share Price, spread H-L and spread C-O		21.37%	29.82%
BSE SENSEX, BSE IT, Share Price, Spread H-L and Spread C-O		24.13%	34.36%

1.4 Conclusions

1. None of the values are more than 50%, hence the relationships between these variables are very weak.
   
2. Some of the values are very negligible, showing that they have no relation at all when everyday data is considered.
   

Further study of variation just before the report release has to be done.