Hi all

I am thinking about developing an asset management approach which will have a dynamic asset allocation with the aim to hold the portfolio volatility constant (the client can define a target volatility for his portfolio). The approach will be implemented particularly with ETFs. Furthermore, the potential loss in the portfolio will be reported regularly to the clients. In addition, the potential loss has to be in accordance to the preceding ALM.

Concerning this matter, I would be very pleased if I could ask you some questions about the control and measurement of volatility:

1. Which data frequency would you use to measure the volatility (high-frequency / daily / weekly / monthly)? Why?
2. Which data window would you use (how many days, weeks or months) to calculate the portfolio volatiliity? Why?
3. Wich models would you use to make a forecast for the portfolio volatility? For example: would you prefer EWMA or a typ of the GARCH familiy?
4. How would you measure the potential loss in the portfolio? For example with VaR and CVaR? Do you know other models?
5. Do you know useful instruments to generate buy- or sell-signals for the VIX (in connection with volatility forecasting)?

Thanks a lot!!!

Best regards,
Fractal