VAR, Monte Carlo, Fan Charts

[Niamh]

Hello,

I have a base line equation, which is my debt path equation. From this equation, I would like to to run a VAR model, using variables which effect this debt path equation. I will use past data and my forecasted data will given by an external source. Is it correct, that if I have this data, ordered by year, for each variable, in excel, I can just import it into eviews to create a new workfile? Then all I have to do is open them as a VAR model?

Secondly, I would like to, using the variance-covariance matrix of errors from the VAR model, simulate stochastic shocks using the monte carlo simulation method and cholesky decomposition option. It is correct that all I have to do is a impulse response function, choosing monte carlo and "choelsky ordering-dof adjusted"? Should I chose to have accumulated responses?

Then, I understand, I will have lots of different values for my variables in my VAR for each year. I would like to take these values, and put the relevant ones into my debt path equation - so I will have lots of different debt paths for each year, which will center around my initial baseline equation. I do not understand the output I am shown from the impulse response function? Regarding my next step (see below) is it best that I recieve the output as a table? If so, I do not understand what this table means :(

Then, using these values for my debt path, at each year, I would like to create a frequency distribution and confidence intervals around my baseline, to eventually, create a fan chart. As I do not understand the step above, I do not know how to do this either. Do I need to create a new workfile, to put my new values into my debt path equation? However, I believe, once I have put my values back into my debt path equation, it would be best to then do the fan chart in excel?

Please can anyone help, I am really struggling and this work is for my masters dissertation.

Thank you in advance! Let me know if you have any questions or I am not being very clear.

Niamh

[wjgatt]

That's a lot of work. Better to list what you need to do in steps, though you should regularly consult your supervisor if you've never done this before, or at least follow the literature.

If I understand your query you want to:

1. Create projections for debt using the dynamics of a single equation, which you already have
2. Create paths for the forecast period for the independent variables (i.e. right hand side) that feature in your single equation using a VAR model (therefore you need to:)
3. Create a VAR, choose the lag length, and choose some method for obtaining the structural form to identify the shocks (Cholesky is the simplest)

I'm note sure whether you want to create IRFs from your VAR to understand the implied dynamics, or to use the VAR in a model object, together with your single equation, to simulate the path of debt under a stochastic setting.

Also you say the values for your forecasted data (independent variables?) will be given by an external source. If so, I don't understand why you still need the VAR to give you the forecasts for your independent variables.

Creating the fan chart is obviously the last, and easiest step, and the Eviews model object does that for you (Procs>Make Graph...), so don't worry about it for now.

Hope this helps in some way
W