Hello, I’m a new user of Dynare and I got some questions about logging data. The problem is that I directly input the linear system in the mod file and all the variables are actually variable_hat, like y_hat, but the data is about the original variables, like y. So I think there are two solutions, the first one is that I calculate the y_hat and input the data into Matlab, the second one is to define new variables, like y, and input the formula like y=ys+ys*y_hat, where ys is the steady value of y. However, no matter which method I choose, I need to know the steady values of the original variables. Take as an example a 20-year data of y, can I use the mean value of y as the steady value?
Sincerely look forward to your reply.
Why do you need the steady state? Due to homogeneity, typical models are scale invariant and results can be simply scaled up. Model information in percentage deviations can therefore be directly compared to percentage deviations in the data.
Dear professor, thanks for your reply but I’m still puzzled. As you say, typical model are scale invariant, so the meaning is that even though I input the linear model in the dynare file, I can also calculate the percentage deviations in the data, input them to the data file and directly name the data as variable_hat, which is defined in the linear model. Is that what you mean?
Dear professor, I have tried to calculate the percentage deviations in the data and input them to the data file. But when I run the program, the errors shows as following:
Error using chol
Matrix must be positive definite.
Error in gmhmaxlik (line 197)
dd = transpose(chol(CovJump));
Error in dynare_estimation_1 (line 437)
[xparam1,PostVar,Scale,PostMean] = ...
Error in dynare_estimation (line 89)
dynare_estimation_1(var_list,dname);
Error in silvia (line 538)
dynare_estimation(var_list_);
Error in dynare (line 180)
evalin('base',fname) ;I can not find the error so could you please help me with the problem?
Your syntax for estimating standard deviations is wrong, resulting in
[quote]WARNING !!!
The rank of H (model) is deficient!
sigma_j is not identified in the model!
[dJ/d(sigma_j)=0 for all tau elements in the model solution!]
sigma_u is not identified in the model!
[dJ/d(sigma_u)=0 for all tau elements in the model solution!]
sigma_Ah is not identified in the model!
[dJ/d(sigma_Ah)=0 for all tau elements in the model solution!]
sigma_An is not identified in the model!
[dJ/d(sigma_An)=0 for all tau elements in the model solution!]
sigma_Rl is not identified in the model!
[dJ/d(sigma_Rl)=0 for all tau elements in the model solution!]
sigma_R is not identified in the model!
[dJ/d(sigma_R)=0 for all tau elements in the model solution!]
[/quote]
It should be
stderr epsj,inv_gamma_pdf,3,inf;
instead of
sigma_j,inv_gamma_pdf,3,inf;
Moreover, the size of movements in R looks rather large compared to pi, suggesting a scaling problem.