Unfortunately, we don't have tick data for VIX. There are ETFs that seek to replicate returns of VIX futures (VXX, UVXY, TVIX), but they are not the VIX spot price.
Thanks for the information. I was interested in VIX data mainly because it was mentioned several times in Anderson's paper; and was implied to be the better predictor than VPINs; but it's okay If we do not have it.
Unfortunately, we don't have tick data for VIX. There are ETFs that seek to replicate returns of VIX futures (VXX, UVXY, TVIX), but they are not the VIX spot price.
Why is VIX spot important? Why not just estimate realized volatility, using something like stdev(log return) over several periods, or log(abs(high - low) / (0.5 * (high + low))) ? Actually, there are better estimators --- see < http://www.wilmottwiki.com/wiki/index.php?title=Volatility#Realized.2Fhistorical_volatilities >