Downside Bias in the Stochastic Volatility Model?

[jlindbloom]

For the sake of completeness of this thread, here’s the resulting model with an inferred (nonzero) mean for the Student T distribution:

The inferred mean was positive. As you can see, this makes a huge difference!