Dear Lucie,

This is typically done when working with cross section data in order to control for year and country specific effects. It is quite easy to in E-views. When you set up your data as cross section (new workfile --> balanced panel) you are later given "panel options" when estimating your regression equation. Under panel options just choose "fixed" next to cross section and period and you will have country and time fixed effects. Hope this helps.

Hi,

What I see from your problem is that you don't need to have a country dummy variable because countries do not share variables, for instance, the US is on the explanatory side only and Europe is on the dependent side of your equation. However, it is important to have quarterly dummy variables because the demand for some products can be influenced by periods of the year. You need to generate three dummies to avoid dummy variable trap, each dummy variable will be compared with the constant which is taken as the base variable. You can decide to use first quarter as base variable, meaning it will take zeros. This can be easily generated in Eview's command.

I am not getting why you need country fixed effects - from what you have said there appears to be in effect only one country - Europe- and you have repeated quarterly measures on the outcome of exports to the USA. The tariff regime is then a time varying variable and to estimate its effect it must vary over the period of study! It might aslo be a lagged effect that is not instantaneous.

If you do have multiple European countries I could see the need to remove idiosyncratic things about particular counties so as to isolate the longitudinal effect and to get at the time varying effects of tariffs.

The are several ways of doing this and we argue in these papers for random rather than fixed effects as it is more flexible ( and gives the same answers if done properly) but many would disagree! We do however try and explain why you are doing it and what we hope to achieve.

Article Fixed and Random effects models: making an informed choice

Article Explaining Fixed Effects: Random Effects Modeling of Time-Se...

for some insight into the disagreement see this

https://en.wikipedia.org/wiki/Kelvyn_Jones

"has challenged the 'gold standard' that fixed effects should be the standard approach to the analysis of Panel data "

The point about random effects is that the effect of tariff regime change could be different (random that is "allowed to vary"between countries and fixed effects (with their removal of context) would not allow you to get at that this heterogeneity.

You would do this to remove unobserved heterogeneity bias (unobserved variables may be correlated with both independent and dependent variables; by not including those your estimates of the effects of your dependent variables on the independent may be biased (i.e., "WRONG") and you won't know it!). By including a dummy variable for each country, for example, you'd hope that the dummy would measure those unobserved characteristics and then your estimates would be "less wrong". But . . . this wouldn't work when you only have two countries. The other issue is that you have time-series data which presents the issue of serial autocorrelation (similar to spatial autocorrelation) which presents other issues. Sorry to be so apparently unhelpful but you should step-back and think about the nature of your data, your research question, and the accepted approach first.

Thank you all for your answers, I appreciate your effort to help me.

In the end it emerged that there was a misunderstanding, my supervisor thought I was examining panel data instead of time series data that might need addition of ARMA term types (as some of you suggested) rather than the fixed effects. Nevertheless, it brings me to next questions:

How to add ARMA term types to my model? So far, I have run a regression analysis of my data (picture 1), determined that my data is nonstationary at Level but stationary at First Difference. Also, according to DW Stat (0.124752

Kelvyn Jones Thomas J. Cooke Fulgence Dominick Waryoba

I have a question, if I'm using single country data for few years and the data is cross sectional. Reviewer asking me to do fixed effect. Is it doable?

I am unclear - how can you have a cross sectional data for a few years?

Initial Public Offerings (IPOs), only one time listing and I have collected all IPOs listed throughout the few years.

Rasidah Mohd-Rashid I hope the answer is not late, I suppose you want to add year fixed effects to control for year to year variation, this is a perfectly normal situation for event studies. Ideally, to not have a lot of dummy variables and noise, there should be few effects to be controlled, so having few years of data is fine.

Lokman Tutuncu thank you for the response. Yes, I did the same by controlling year effects.

kindly share the solution to this query with me....i couldn't find it in the given answers....thank you