Simulated method of moments: quarterly averages of monthly artificial data

Greetings!

I am interested in estimating parameters of a search and matching unemployment model via simulated method of moments. The model is written at a monthly frequency. However, though various of the model variables are available monthly (unemployment, vacancies, etc.), output/labor productivity is not. Accordingly, I would like to take quarterly averages of simulated monthly data from the model and use these for calculating moments.

Is there a way to do this via the method of moments interface (matched_moments and method_of_moments command)?

As an aside, I know that it is possible to write the model as quarterly. But this would require workers to begin production the same period they are hired, and would also prevent us from examining monthly impulse responses.

Best and warm regards,
Mario

Do you already have example codes to test this? I would need to check what our treatment of NaN during moment computation does in case of mixed-frequency setups.