Help with dummys and GARCH model

[gastonpresente]

Im trying to introduce dummy variable of newspapers (Thats a daily variable) into intraday data. And i have no idea. Because the newspapers are out at 5.am and the market is open between 10am and 3pm. Any help??
Someone recomended me "Macroeconomic News and Bond Market Volatility by Jones, Lamont, and Lumsdaine", but i couldnt get it, its not in jstor.org.

thanks a lot

Gaston

[trubador]

I believe introducing the dummy variable into your model is the least of your problems, since you are trying to mix high and very high frequencies. Quite a number of issues arise when you work with intraday data (e.g. induced positive cross correlation, induced serial correlation, spurious mean reversion, etc.). In other words, such data has unique properties that you do not observe in lower frequencies (including daily data). Therefore, I do not think that GARCH is the appropriate technique for your model. Regardless of the size of the time interval, much of your data would be without price change. Such discreteness makes it really difficult to model intraday data. If you decide to go on with the intraday data, I suggest you to build an Ordered Probit model instead. On the other hand, you can still conduct GARCH analysis with dummy variable, if you switch to daily frequency. Besides, very useful discussions have been made on this subject before. Just look into previous posts under the Estimation thread...