I attach a pdf which contains the model I am expected to test using indirect inference and also the list of the variables in detail. I also attach the impulse response function codes of the model to show how it works using dynare matlab.
If you want to know the steps for the indirect inference,
The steps for Indirect Inference are:
1. Generate single equation residuals
2. Estimate the AR(1) processes of the residuals and use the residuals of these as the innovations that shocks the model
3. Put the residuals AR(1) process in your model.
4. Bootstrap the shocks with the model to get 1000 simulations.
5. Calculate the Wald statistic using the VAR estimates from the 1000 simulations and the VAR estimate on the actual data.
Now, if you want to do Indirect Inference estimation you need to use the simulannealbnd (I think that's what it's called) function. You need to write matlab code that takes your parameters as the input and calculates steps 1-5 from above, and has the Wald as the output, then the SA algorithm will try and minimise the Wald.
In step 1 because the model has expectations we calculate a VAR of all the expected variables, and then use that to calculate E(t)Y(t+1) for example, and use that along with the actual current and lagged data to calculate the single equation residuals.
I am looking for someone that can use the Dynare toolbox in Matlab or similar to test a NKPC DSGE model for me using Indirect Inference method.