Hi David and Forum members,
I recently come across a VaR model for market risk that has an assumption that "VaR(u) of the maximum interest rate spread in year x is equal to VaR(u^(1/x)) of the interest rate spread in one year", where u is confidence level. Based on the square-root of time rule, the VaR(u) of year x should equal to VaR(u)*sqrt(x) of the one year. What do you think about this assumption? Any idea how the modeler base this assumption from? And last question. Do you think this assumption is reasonable? Thank you.
I recently come across a VaR model for market risk that has an assumption that "VaR(u) of the maximum interest rate spread in year x is equal to VaR(u^(1/x)) of the interest rate spread in one year", where u is confidence level. Based on the square-root of time rule, the VaR(u) of year x should equal to VaR(u)*sqrt(x) of the one year. What do you think about this assumption? Any idea how the modeler base this assumption from? And last question. Do you think this assumption is reasonable? Thank you.
