EViews.com

Could you expound upon the process to use contraction? I've run across this same problem many times. Here's an example of what the data may look like...

I'll have weekly population for all of the counties in all 50 states. However, I want to look at the population for the states over time. In excel, I would use a pivot table to segment on states and time and then SUM the population. How would I do this in EViews? Also, what if I want to do "max" for some objects and "sum" for other?

Make a new page structured with dates and states. Then copy the relevant series from the existing page to the new page using paste special, match on date and state. You can specify whether the contraction is done by sum or max (among other options).

I'm assuming you have a panel workfile with date and county as the two id-series. Startz's method is correct, but if all you want to do is "look" at the population by states, the easiest thing to do would be to open up the population series, and then click on View->Stats by Classification, and then enter your state variable as the classifying variable, and sum as the statistic you're interested in, rather than making a whole new page.

Thanks Startz, this is exactly what I was trying to do. Much appreciated.

Hi
I have a problem in estimating the fixed effect model.
As in panel option/effect specification than cross section and period shows none in both cases. Which one take the fixed (Cross-section as fixed or Period as fixed) in estimating the fix model and why?

Dear,
I´m a new user (basic user) from eViews.
I'm trying to analyze a group of firms from China.
I've collected three variables from market and analyze them.
So the data are:
Income, income lagged, IFRS (dummy) and interaction IFRS and income lagged.
I have 1507 firms between 2002 and 2011.
So, when I run panel data with fixed or random effects, this error appear: Matrix size error: too many parameters for default coefficient vector.
Could you help me?
Thanks,
Jomar

Which version of EViews are you using, and do you have a panel or a pool?

I'm using 7.2. I have a pool.

Create your own coefficient vector in the workfile (type the following in the command window: Then when you estimate the pool, on the options page change the coefficient name from "c" to the one you just created ("beta" in my example).

Thank you!

Hi Gareth,

I would like your advice on how should I structure some data.

I got GDP growth and oil consumption for about 80 countries, annually. The 80 countries are split among 4 categories: low income, lower-middle income, upper-middle income, high income.
I would like to run ls regression grouping the countries according to the 4 categories. I’ll run some statistical tests (autocorrelation, unit root, etc) and I might need to do cointegration regression (probably DOLS). Eventually, I would like to use the model to do some forecasts.

My questions are:

- What is the best way to "classify" the countries according to the 4 categories? Dummies?
- Should I structure the data as pool or panel ?

I'm inclined to use panel (because of the forecasting and other estimation facilities) but I'm wondering if the classification is easier to accomplish with pooling.

I’m gonna code the whole procedure in order to facilitate updates as time goes on.

Thanks in advance,

Mara

Hello

I have a concern regarding using a pooled work file. I like to ensure I have corrected my correlation using Newey-West, but the eviews 7 SV doesn't give me the option to include it. Is there another, similar, option within eviews that allows me to correct in a similar manner? Or, is this a limitation within software/methods/theory that I am not aware of?

Thank you for any assistance, and this helpfile greatly explains a lot about the panel/pooled estimation methods within eviews.

Regards,

Ryan

Use a panel rather than a pool. Panels support NW.

Thanks! Ill give it a try.

Ryan

Many thanks, Gareth, for this insightful post.

I would add one benefit to using a Panel instead of a Pool: when dealing with many cross sections, I find its much quicker to save your workfile in Panel format than in Pool format.

I have more than 3000 cross sections in my file and it takes about 6 minutes for Eviews to save the workfile. It Panel mode it takes 2 seconds. I'm a bit paranoid about losing data so I like to save my files regularly. Waiting 6 minutes each time is a drag on my productivity.