"FO9. Simultaneous block is singular." Error

Use this forum to post questions about syntax problems or general programming issues. Questions on implementing a particular aspect of econometrics should go in "Econometrics Issues" below.
evankoet
Posts: 3
Joined: Mon Aug 22, 2011 4:43 pm

"FO9. Simultaneous block is singular." Error

Unread post by evankoet »

Hi,

I am trying to run the code of the paper by Fratzscher, Juvenal and Sarno called "Asset prices, exchange rates and the current account". It estimates a 7-variable VAR and then runs a Monte Carlo simulation by taking orthogonal decompositions and using sign restrictions. However, the simulation does not run and I keep on getting the following error message:

## FO9. Simultaneous block is singular. Check your equations.
The Error Occurred At Location 6693, Line 301 of loop/block

My question is whether this is related to my data, or whether this error message has something to do with the programming itself. Is there also a way to figure out where "Location 6693, Line 301 of loop/block" is, so that I can see what is wrong?

I am able to run a very similar code with a 6-variable VAR (with different sign restrictions), of which 5 variables are the same as in the 7-variable model. The code is very similar so I do not understand where the error comes from.

If you wish I can post the code online, but it is quite long so I thought I would start by asking the questions above. Thank you very much for your help.

Regards,
Evan
moderator
Site Admin
Posts: 269
Joined: Thu Oct 19, 2006 4:33 pm

Re: "FO9. Simultaneous block is singular." Error

Unread post by moderator »

evankoet wrote: My question is whether this is related to my data, or whether this error message has something to do with the programming itself.
Off hand, I'd guess it could be either. That's a "numerical" sort of an error (rather than a syntax error or other issue directly related to your code), but I can't really say whether there's a mistake in the way you've formulated the model, or you've correctly specified the model but it just doesn't work with the data.

If you want us to take a look, the best option would be to email the program and data files to us at support@estima.com. Please be sure to include your full name and your RATS serial number and version number.
evankoet wrote: Is there also a way to figure out where "Location 6693, Line 301 of loop/block" is, so that I can see what is wrong?
Sure. The Edit-Show Last Error may be able to take you to the offending line. Failing that, see "Instruction Locations-Debugging Compiled Code" starting on page 460 of the RATS User's Guide (if you have RATS 7, see the beginning of Chapter 16 in the RATS 7 User's Guide). Basically, this involves executing the code using a SOURCE(ECHO) instruction, or in batch mode, so that you can see the prefixed instruction location information, which you can then use to pinpoint the relevant instruction.

Regards,
Tom Maycock
Estima
evankoet
Posts: 3
Joined: Mon Aug 22, 2011 4:43 pm

Re: "FO9. Simultaneous block is singular." Error

Unread post by evankoet »

Hi Tom,

Thank you for your prompt reply. I will send you the code and data later today.

Regards,
Evan
TomDoan
Posts: 7814
Joined: Wed Nov 01, 2006 4:36 pm

Re: "FO9. Simultaneous block is singular." Error

Unread post by TomDoan »

evankoet wrote:Hi,

I am trying to run the code of the paper by Fratzscher, Juvenal and Sarno called "Asset prices, exchange rates and the current account". It estimates a 7-variable VAR and then runs a Monte Carlo simulation by taking orthogonal decompositions and using sign restrictions. However, the simulation does not run and I keep on getting the following error message:

## FO9. Simultaneous block is singular. Check your equations.
The Error Occurred At Location 6693, Line 301 of loop/block
In order to get a "simultaneous block", you need to have current endogenous variables on the right hand side of one of the equations. Did you add an identity to your VAR?
evankoet
Posts: 3
Joined: Mon Aug 22, 2011 4:43 pm

Re: "FO9. Simultaneous block is singular." Error

Unread post by evankoet »

The error came from the fact that I defined a variable in the VAR as i, whereas the Monte Carlo loop also used i. After changing the name of the variable it worked perfectly.
Post Reply