Expressing observed variables

Hi,

I had a question about how to express observed variables under var_obs. For example, the
interest rate in my data are the three month Treasury rate, APR, i.e., 400* r, where r= log (R/R_s) and R
denotes the model gross nominal rate of interest rate and R_s is its steady state value.

So if I write r under var_obs, that will definitely lead to errors unless I introduce another variable called r and introduce one more equation saying r=400*log(R/R_s).

How do I overcome this problem.

Also, if I try to write xls_sheet=1 and xls_range=A1:C87, it leads to errors

“line 65, cols 55-60: syntax error, unexpected QUOTED_STRING”

Thanks
Steve

For the error, I would need to see the codes.
Regarding the varobs, what exactly is your data treatment. In your post above I cannot discern what is in the model and what is the data treatment.

Hi,

I am estimating a basic CGG model (mod file attached). The observed data are (i) log, first differenced GDP (i.e, yt in the model), (ii) the interest rate (the three month Treasury rate, APR, i.e., 400 * rt, where rt = log (Rt=R) and Rt denotes the model gross nominal rate of interest rate and R is its steady state value) and (iii) inflation (log, first-difference of GDP deflator). Model data are expressed relative to their steady state value.

My question is my observed interest rate in data and my model interest rate are different.
So if I write r under var_obs, that will definitely lead to errors because it refers to r=400*log(R/R_s) which is different from the model interest rate R. The mod file and data are attached for your reference.

cggest.mod (3.2 KB)
data_q_on_q.xls (31 KB)
.

Request someone to reply to this please. Thanks.

I still don’t understand your transformations. Why is there an equal sign in the brackets of the log.

That was just a typo. It is r(t) = log (R(t)/R)

Again, why do you think your data treatment for the interest rate is incorrect? If you want to be sure, please read Pfeifer(2013): “A Guide to Specifying Observation Equations for the Estimation of DSGE Models

Many Thanks. When I did not understand the paper, I came to this forum.

Then please try to carefully explain in detail what you question is.

Thanks for your time.