i'm trying to estimate a SOE model of inflation targeting based on a IMF paper (ref :WP/08/48), and i'm having the following trouble:
there's an UIP (uncovered interest parity) condition in the model where the expected exchange rate at t+1 is expressed as follows:
z_exp = delta*z(+1) + (1-delta)*z(-1)
where:
z_exp : the expected exchange rate at t+1
z : the real exchange rate
the problem is that, when i use this expression, matlab gives this warning message during estimation : " Matrix is singular to working precision ", and estimation fails by the end (even if i use mode_compute=6).
i use this command: model_diagnostics(M_,options_,oo_), and it gives me:
Warning: Matrix is close to singular or badly scaled. Results may be
inaccurate. RCOND = 1.347220e-018.
> In model_diagnostics at 115
model_diagnostic: the Jacobian of the static model is singular
there is 1 colinear relationships between the variables and the equations
Colinear variables:
z_exp
s // s is the nominal exchange rate
z
Colinear equations
Columns 1 through 12
1 2 3 5 6 7 10 11 12 14 17 20
Columns 13 through 16
21 22 26 27
and i don't know how to deal with this colinearity problem
What is a little bit strange is that when use the following trick, estimation works perfectly !:
i've tried to declare two parameters (delta_lead & delta_lag) intsead of delta, and i have written z_exp as follows:
z_exp = delta_lead*z(+1) + delta_lag*z(-1)
however, this is not theoretically correct because delta_lead and delta_lag should sum up to 1.
So please help me, i'm really stuck and i have to deliver my project after tomorrow !
i'm using dynare 4.2.5, and i've tried the 4.2.4 and didn't work also (i thaught that may be it's a bug issue)
still have to mention that i'm an undergraduate student and i'm new to dynare
regards,
El Mehdi HAIZOUN
