steady_state_model problem!

[Glm]

A bit odd problem.
At row 71 I have

%%%Labor supply
w_flex=(eta_w_ar/(eta_w_ar-1))*(n_flex^phi)*THETA_flex/lambda_flex;
%w_flex=1.1111111*(n_flex^phi)*THETA_flex/lambda_flex;

When I switch between these two equations, equation

w_flex=1.1111111*(n_flex^phi)*THETA_flex/lambda_flex;

works perfectly even though the steady state value of eta_w_ar=10 which means that (eta_w_ar/(eta_w_ar-1))=1.11111111111 is identical. But equation

w_flex=(eta_w_ar/(eta_w_ar-1))*(n_flex^phi)*THETA_flex/lambda_flex;

generates an error with the BK conditions

Please note, eta_w_ar follows an ARMA(1) process which is why eta_w_ar=10 when its shock is set to 0. I attach the code as well.

[jpfeifer]

What matters for BK are the eigenvalues. Replacing a stochastic process with a constant of the same steady state is not equivalent. You can see this by running model_diagnostics, which returns a collinearity issue with the first equation. You need to understand what is going on when you enter the exogenous stochastic process.