Hi, I was wondering if there are any advantages of using weekly returns to compute weekly volatility of commodity futures, when daily prices are available, hence daily volatility computation possible?
Thank you in advance.
No. More frequent observations allow you to estimate volatiltiy more precisely. Even if you need weekly volatiliy, you shold still use daily returns to compute weekly volatility.
Thank you Peter Molnár for your reply. Do you know if there are any advantages of having weekly volatilities over daily ones? I am trying to analyze the price discovery function of a market.
If you have daily data only, it is not possible to estimate precisely what was volatility on a particular day. On the other hand, you can estimate weekly volatility much more precisely.
The higher the frequency of returns used, the better it is. Besides, the higher frequency returns may better be used to estimate, model or forecast the lower frequency volatility. Hence, even if you wish to compute weekly volatility, better use daily returns, if not hourly...
The benefit is the advance calculation of the researchers method structure. In differential calculus, in the span time of pricing. The component time, 🔺T, is the standard normal, variable variance, in the standardization transformation. Over, stochastic calculus, formely dominated infinitesimal by the Brownian motion, the random variable differential is, square root of dt, times the diffussion component parameter. Thus, to calibrate the exact calculation value, normal for the time span say weekly, dominant of less data, the calibration will include the square root of week time the parameter. Meaning, any daily or weekly parameter initial calibrated can be used by the Euler implicit at the week-initial or explicit at the day-initial. Convergence to any day/week point, requires Newton-R or the Gauss general linear error reductions methods to resolve the estimators bias. Both are state equivalent and bias dominant in terms of bootrap with heavy data computations relevently to estimators unbais in the fine convergence or with less data for the coarse probability convergency. The advance structural calculation with dominant classes attribute is appropriate. In the illustration of the data question and the elaboration of the specific researcher' argument.
- Badges
- Science topic
- Similar topics
- Mathematics