Since JAE provides data used in their estimation, I decided to write estimation code on my own using dynare.
While writing a code, I got one thing to ask.
According to the paper, the foreign economy(pi_st, y_st, i_st) is expressed by a vector autoregressive processes of order two.
So, I coded as
- Code: Select all
pi_st = bpp1*pi_st(-1) + bpi1*i_st(-1) + bpy1*y_st(-1) + bpp2*pi_st(-2) + bpi2*i_st(-2) + bpy2*y_st(-2) + eps_ps ;
y_st = byp1*pi_st(-1) + byi1*i_st(-1) + byy1*y_st(-1) + byp2*pi_st(-2) + byi2*i_st(-2) + byy2*y_st(-2) + eps_ys ;
i_st = bip1*pi_st(-1) + bii1*i_st(-1) + biy1*y_st(-1) + bip2*pi_st(-2) + bii2*i_st(-2) + biy2*y_st(-2) + eps_is ;
One thing strange is that there is no posterior estimates about coefficients of this VAR.
Is it correct that these coefficients are also estimation target?
Also, my estimation line is as follows.
- Code: Select all
estimation(datafile=aus_data,plot_priors=0,nograph,mode_compute=4,presample=4,prefilter=1,mh_nblocks=1,mh_replic=110000,mh_jscale=0.5);
But I continuously meet
"Blanchard Kahn conditions are not satisfied: no stable equilibrium"
I think there is no unit-root in the equations but If I add "diffuse_filter" option, it starts estimation.
But soon, I meet
"Error using chol
Matrix must be positive definite."
Can you realize what is the problem?
Thank you.