Range Estimation of Stock Market Index Using Extreme Value Approach

Abstract

The heart of the financial/stock markets is the ability to predict future prices. The range i.e. the difference between the high and the low of the stock price or index in a day, could be a key to the amount of exposure and the possible gains from investment. So an extreme value approach for the range prediction is adopted to investigate the problem. The daily log ratio, i.e. log of the ratio of the high/low on one day to that of next day, is chosen as the variable of interest. Traditional approaches would consider the midrange, defined as the average of the daily high and low, as the variable of interest. The predictions from this traditional model do not portray possible extreme gains or losses available from range estimation. The results thus provide some unique insight into risk assessment and mitigation, as the approach is a breakaway from the tradition.