SGX Nifty Live Quotes

SGX NIFTY Trading Hours: Mon-Fri 6.30am - 3.40pm (India Time IST Time)



Get More Accurate SGX Chart here :> Live SGX Chart

Live World Market Futures,DowJones Future, FTSE,Nasdaq Future

Live FTSE and Wall Street ( Dow Jones) Future price)
( Auto Refresh every 10 Seconds )

If Page not open in 10 second try another link : Major World Market Future
Realtime Quotes powered by GCI Financial Ltd. - Trade these Prices Commission free - Click here!






Note :

Please wait 10 second for data upload. If Data Doesn't Upload in 10 Seconds please try another link :http://liveworldmarket.com

Sunday, August 21, 2011

THE EFFECT OF FUTURES TRADING ACTIVITY

In the previous section, we tested whether there appears to be any structural
change in the underlying market at the time of futures introduction.
In this section, we will test whether there appears to be a relationship,
after the futures have been listed, between the level of future trading activity
and the volatility of the underlying index.


Our approach is based on that of Bessembinder and Seguin (1992).
Using an autoregressive integrated moving average (ARIMA) model, they
decomposed the time series of futures trading volume and open interest
into expected and unexpected components. Bursts of trading activity
stimulated by unexpected price changes should be picked up in the unexpected
component, while the expected component should reflect the
“background” level of futures trading. They found that market volatility
was positively related to the unexpected components of volume and open
interest, reflecting the positive effect of volatility on volume, but that
market volatility was negatively related to the expected component, suggesting
an underlying stabilizing influence.

We followed a similar procedure using futures market trading volume
and open-interest data from 17 of our 25 countries for which volume and
open interest data were available. First, we analyzed the volume and openinterest
time series from each country to select an ARIMA model that
appeared to fit the data reasonably well. Restricting our attention to models
with five or less autoregressive lags and five or less moving average
lags, we selected, on the basis of the autocorrelation structure, a different
model for each time series.