Serving the Quantitative Finance Community

 
User avatar
list1
Topic Author
Posts: 827
Joined: July 22nd, 2015, 2:12 pm

a question

October 3rd, 2016, 4:19 pm

I have sent several messages on economic forum and they have not appeared there?
 
User avatar
list1
Topic Author
Posts: 827
Joined: July 22nd, 2015, 2:12 pm

Re: a question

October 3rd, 2016, 5:58 pm

I have sent several messages on economic forum and they have not appeared there?
it works, thanks 
 
User avatar
tags
Posts: 3162
Joined: February 21st, 2010, 12:58 pm

Re: a question

January 2nd, 2018, 12:25 pm

I also want to ask a question.
In the OT subforum, there is the Meta OT thread.
Is there any "Meta" thread for other sub-forums? (yes, the search engine .., but...)
I mean I'm interested in the asymmetric relation between volatility and returns on the (US) equity market. More specifically, I'm looking for comments (drawn from empirical observation, and/or theory) about this relation over small timeframes. What's the most appropriate timeframe to observe this relation? To what extent does this relation hold over small timeframes?
It there already a thread discussing this topic?
Please, flag if my question just doesn't make sense. 
Merci, guys.
 
User avatar
outrun
Posts: 4573
Joined: January 1st, 1970, 12:00 am

Re: a question

January 2nd, 2018, 1:08 pm

I would do a quick search, and if you can't find the topic then start a new one. If you missed the existing topic then someone will link to it for you.
It can (will) be different on different time-scales, so you need to pick a time-scale of interest first, and then look at the empirical behavior on that time scale.

Theory is just theory IMO. Most models in finance are pulled from air because they are handy, not because they match reality. The models are handy to guard for arbitrage in the context of derivatives, but they don't need to have value in absolute terms. Interest rates were supposed to never go negative, volatility was supposed to be constant etc etc. Models of "value" of things are not easy because value is affected by endless numbers of subjective factors (if you see an icecream ad on tv then the value of icecream will increase for you) , .. economy is not physics.
 
User avatar
tags
Posts: 3162
Joined: February 21st, 2010, 12:58 pm

Re: a question

January 2nd, 2018, 1:36 pm

Hi outrun,

Thanks for your comments.

I started to read about this yesterday (disclaimers: I'm neither a quant nor in the equity space) and was quite struck that time-scale is never mentioned.

Don't laugh at me; I'm interested in 5 min time steps... Making quick charts with data for some 200 US stocks, that assumed asymmetric relation wasn't obvious. At best, charts seem to show "something" in case of extreme positive volatility changes (BTW what about causation?).
 
User avatar
Traden4Alpha
Posts: 3300
Joined: September 20th, 2002, 8:30 pm

Re: a question

January 2nd, 2018, 1:55 pm

Are you sure that scale is denominated in time? For an extremely liquid stock, 5 minutes is a lifetime but for an illiquid one, 5 minutes might be only a trade or two.

Perhaps scale in denominated in trades or orders: the 10-transaction window, 100-transaction window, etc. The shorter windows would be dominated by stochasticity in the arrival rates of buyers and sellers. The longer windows would be dominated by stochasticity in exogenous drivers of prices.
 
User avatar
outrun
Posts: 4573
Joined: January 1st, 1970, 12:00 am

Re: a question

January 2nd, 2018, 2:02 pm

Cool.
as you observe, it is what it is!
But what is the goal? You want to develop a 5min trading or risk model?
On short time scales markets show complex behavior: volume has daily patterns, autocorrelation.. The price you see is transaction prices or market orders hitting limit orders. There orders have dynamic intensities. I wouldn't assume some continuous time theoretical behavior: instead model the observations and test if any pattern you find is statistical significant and persistent?
 
User avatar
tags
Posts: 3162
Joined: February 21st, 2010, 12:58 pm

Re: a question

January 2nd, 2018, 4:43 pm

@T4A @outrun thank you. very relevant comments as always.
I'll tell you later exactly what I'm doing, otherwise you'd (yourself) find the solution to my problem, and I will have nothing left to research! I volontary stay vague for the time being.
Merci beaucoup!
 
User avatar
outrun
Posts: 4573
Joined: January 1st, 1970, 12:00 am

Re: a question

January 2nd, 2018, 5:06 pm

You know t4a can't help his deductive powers, half a vague word is all he needs.

I promise I won't disclose any preemptive solutions to hypothetical problem if he end up sharing those.