IRF minimum distance estimator

[chaveslima]

Dear Sr,

I am using NOLIN command with FIND, to do a Minimum Distance Estimation in the following way:

nonlin(parmset=dsgeparms) b1 b2 nai delta1 delta2 gamma1 gamma2 gamma3
declare real ceemetric
find(parmset=dsgeparms) minimum ceemetric
dsge(a=a,f=f,model=loanmodel) y rm l rl eta
@dlmirf(a=a,f=f*.2,steps=nsteps,results=theoretical,nograph)
compute ceemetric=0.0
do h=1,nsteps
do i=1,4
compute ceemetric=ceemetric+(theoretical(i,1)(h)-results(i,1)(h))^2/sd(i,1)(h)
end do i
end do h
end find

Reading RATS Reference Manual and User’s Guide I realized that I just can have the standard errors of estimated parameters if using METHOD=BFGS on FIND. However, according to the Reference Manual, “if the function that you are maximizing isn’t a likelihood or quasi-likelihood function, the numbers reported are unlikely to be interpretable as standard errors”. Can you give any tips on how to get those standard errors considering that I do not have a case of maximum likelihood estimation?

All the best

Ricardo

[TomDoan]

What the authors did is what you would get with METHOD=BFGS and STDERRS options on the FIND instruction. However, I'm not sure what statistical theory they are basing that on, since this isn't a log likelihood function being optimized.