How can I estimate volatility with historical data for Monetary Policy Rate (MPR) to use in a short rate model?
I could use regular techniques like simple standard deviation or max likelihood, but the problem is the data: MPRs usually have 0 variation over months, even years and then big discrete jumps.