I think I found another favorite word from the CFAI…besides nonparametric. “Asynchronism” Noun: The state of being asynchronous, or out of synchronization. CFA Use: High frequency data exhibit sensitivity to asynchronism which causes the data to produce lower estimations of correlation.
Sample 3 Spoiler** …burned me too.
This is one of the main reasons I take CFAI sample/Mock exams…to find out what fancy words they might use…
actually this was in the schweser text…i think where they talk about economic forecasting…and the problems faced in that…they mention asynchronous data… basically it refers to when you’re using a period of time… and there is data missing from certain dates …you substitute it with data eg. from the previus day…this will give you low correlation when calculating volatility …for obvious reasons…
what does nonparametric mean anyway?
Asymetrical, as in non standard, for example, historical method one advantage it has is that it is nonparametric.
smokin’hot Wrote: ------------------------------------------------------- > Asymetrical, as in non standard, for example, > historical method one advantage it has is that it > is nonparametric. Edit: I read your post wrong.
I thought Asynchronism is like calculating a SD for real estate returns valued like quarterly compared to a SD of regular market equities which is valued like daily. Isn’t that what Asynchronism mean?
Asychronism means that prices aren’t available every day so when you run a regression it seems like correlations are higher and standard deviation is artificially lower
Wouldn’t both be lower VA?
Asynchronism: A discrepancy in the dating of observations that occurs because stale (out of date) data may be used in the absence of current data. Non parametric: Involving minimal probability distribution assumptions. These are the defs out of CFAI glossary. Not sure what that second def really means about non parametric, can someone confirm for me my first definition I gave, that it is a non standard distribution?
Any other Funky words out there I may need to know about?
Oh SNAP! I forgot what that means…
For nonparametric: Basically, this is a good thing for the historical method of VAR. With the variance-covariance method you have to assume a normal distribution. With historical, you don’t have to make this assumption (and that’s what makes it nonparametric).