Regression model with hourly data?

For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

Posts: 40
Joined: Wed May 31, 2017 7:37 am

Regression model with hourly data?

Postby AfonsoRod » Wed Jan 30, 2019 3:46 am

Dear all

I am a master students in Economics at Universidade da Beira Interior - Covilhã, Portugal, and I am (trying) to perform an econometric research paper for my master thesis. All this to say that I am new at econometrics and Eviews, so I will probably ask some basic or even dumb questions. If I do, i apologize in advance.

So, in short, I am trying to prove the Merit Order Effect in Portugal and quantify it, witch basically means that I will try to see whats the impact on the wholesale electricity prices from the penetration of the renewable energy sources into the electric system.

Thus, my data has an hourly frequency for the time period of 01/01/2009 to 31/12/2018, witch gives me a total of 87648 observations. In the existent literature about this theme, authors tend to transform the data into daily averages, but I would like to avoid that.

So, my question is, how can I know the best regression model to use? Or, at least, which are the most adequate regression models to use with a hourly data with price as dependent variable?

I also know I may need to perform some pre-testing to help me decide the right model to use. So far I've only performed the ADF and PP unit root tests, and both tests told me that all my variables are I(0) - (stationary at level).

To conclude, most literature about this theme uses an OLS, but I wanted to use a better one. I initially thought to use an ARDL, but I'm afraid that having hourly data will make me need to use regressions usually used for financial matters.

Feel free to ask me for any extra pre-tests if you need me to. Thank you all for the help.

Return to “Econometric Discussions”

Who is online

Users browsing this forum: No registered users and 2 guests