Hello Robert Traistaru ,

It is shown as lnpop, so i think you need to transform the variables into log. Simply follow this steps if it is the case,

quick

You can also ask here: http://forums.eviews.com/

I think if you use the logarithm of the variables, this problem will be solved.

Hi Robert Traistaru ,

Please forgive me for my ignorance - kindly advise/define what an "explicative" variable is. I tried to research but couldn't find the term in econometrics terminology. I needed to contribute to answering to your enquiry above but was and am still anxious to do that without understanding what explicative variables are. Thanks beforehand!

Hi Elvis Munyaradzi Ganyaupfu, Robert Traistaru might just be referring to explanatory variable

Godwin Lebari Tuaneh Elvis Munyaradzi Ganyaupfu its an explanatory variable (ro: explicativa, hence the confusion).

Godwin Lebari Tuaneh and Georgiana Elri - many thanks for the guidance colleagues. In response to your enquiry Robert Traistaru , firstly I fully concur with Debasmita Dey that you have to transform your varibales in logarithmic form. In addition, you need to note that the values of your series/variables shouldn't be non-negative since logarithmic scales are practically defined or applied for positive numbers.

CAUTION! == In scenarios where negative values exist in the dataset, some analysts resort to adopting the following practices as solutions-:

[1]. adding an arbitrary constant value "a" to the negative observations so that the transformation becomes log(Y+a) in a manner which makes min (Y+a) a very small positive number.

[2]. contrary to the approch/practice mentioned in [1], some believe that a better way to handle negative values to enable log transformations is to use missing values for logarithms on non-posiitive values. Some programmers apply this approach to resort to loops and IF statements.

Extreme caution should be exercised in any attempt which involves any of the two above described ([1] and [2]) i situatons where some of variables/series contain negative values. Otherwise results may be rendered inaccurate and unreliable.

The above are my abstract views given the lack of information regarding the nature of your dataset, properties of your data (eg. distribution of series, etc) and study objectives, specification of your model, etc.

All the best!