Stationarity problem with log-linearization model around the

Hi all, I am working with Financial Frictions in NK-DSGE-Model with durable goods. In my earlier post, I was ask for the solution of rank condition problem. With the invaluable response, I have removed one reduntant equation and now my model is generating IRFs and some of the theoretical moments with NAN for others. If I use Drop and period option, then I am getting simulated moments for all endogenous variables. I have compared the IRFs with “stoch_simul(order=1,irf=40)” and “stoch_simul(order=1,irf=40, periods=2000, drop=200)” and these are identical. Can someone please let me know why is this happening? Is this still a serious problem? What can I do to get theoretical moments? Another question is about the nominal shock of my model. My model works well with 3 out of 4 main shocks (shock to networth, productivity shock and capital quality shock) but with nominal shock, I am getting the following message “All endogenous are constant or non-stationary, not displaying correlations and auto-correlations”. I read the various post in the forum and consult the user manual for this which talked about stationarity issue. I have log-linearized my model around steady state by hand and input the model in mod file as “model (linear)”. Can I still expect stationarity problem with log-linearization model around the steady state? What should I do to get rid of this problem?

Could you please post the mod-file.

I believe that the model that I was working with is suffering from this type of issue. What would be the necessary steps to take when you are programming a log-linearization done by hand?

I am getting the following results after log-linearizing by hand. I did look at listing 3 as a guide.

MOMENTS OF SIMULATED VARIABLES

VARIABLE MEAN STD. DEV. VARIANCE SKEWNESS KURTOSIS
y_hat 0.000000 0.000000 0.000000 NaN NaN
k_hat 0.000000 0.000000 0.000000 NaN NaN
n_hat 0.000000 0.000000 0.000000 NaN NaN
inv_hat 0.000000 0.000000 0.000000 NaN NaN
c_hat 0.000000 0.000000 0.000000 NaN NaN
w_hat 0.000000 0.000000 0.000000 NaN NaN
h_hat 0.000000 0.000000 0.000000 NaN NaN

CORRELATION OF SIMULATED VARIABLES

VARIABLE y_hat k_hat n_hat inv_hat c_hat w_hat h_hat
y_hat NaN NaN NaN NaN NaN NaN NaN
k_hat NaN NaN NaN NaN NaN NaN NaN
n_hat NaN NaN NaN NaN NaN NaN NaN
inv_hat NaN NaN NaN NaN NaN NaN NaN
c_hat NaN NaN NaN NaN NaN NaN NaN
w_hat NaN NaN NaN NaN NaN NaN NaN
h_hat NaN NaN NaN NaN NaN NaN NaN

AUTOCORRELATION OF SIMULATED VARIABLES

VARIABLE 1 2 3 4 5
y_hat NaN NaN NaN NaN NaN
k_hat NaN NaN NaN NaN NaN
n_hat NaN NaN NaN NaN NaN
inv_hat NaN NaN NaN NaN NaN
c_hat NaN NaN NaN NaN NaN
w_hat NaN NaN NaN NaN NaN
h_hat NaN NaN NaN NaN NaN

In your case, your only shock is epsa. It affects the model via

But A_hat does not enter anywhere else. That is why it has no impact on any other variable and why they are all constant.

Thanks Johannes so much for your help. I fixed it and it’s ok.

Hi all and Jpfeifer,

Could you please help me. I am replicating the Canova and Ravn (2000) RBC model. I have got the stochastic simulation and impulse responses. However, I do not get Theoretical Moments and Var Decomposition.

THEORETICAL MOMENTS
VARIABLE MEAN STD. DEV. VARIANCE
Ns_ss NaN NaN NaN
Nu_ss NaN NaN NaN
gam_ss NaN NaN NaN
ws_ss NaN NaN NaN
wu_ss NaN NaN NaN
r_ss NaN NaN NaN
y_ss NaN NaN NaN
he_ss NaN NaN NaN
hu_ss 1.0000 0.0000 0.0000
cu_ss NaN NaN NaN
yu_ss NaN NaN NaN
hs_ss NaN NaN NaN
cs_ss NaN NaN NaN
ysa_ss NaN NaN NaN
c_ss NaN NaN NaN
ks_ss NaN NaN NaN
k_ss NaN NaN NaN
xs_ss NaN NaN NaN
x_ss NaN NaN NaN
g_ss NaN NaN NaN

VARIANCE DECOMPOSITION (in percent)
epsilon_thetam epsilon_zeta
hu_ss 85.29 14.71

All endogenous are constant or non-stationary, not displaying correlations and auto-correlations

Could you give some advice, please? I do not understand how to fix the problem. I believe the problem is in the AR(1) process, I mean the shock in migration, but I am not able to see what is exactly the problem and how to fix it.
Thanks!

Mig_5mar_steadystate.m (3.7 KB)
Mig_5mar.mod (3.5 KB)

Please do not cross-post. See my reply at Stationarity problem in log-linarized model