2,499 reputation
320
bio website researchgate.net/profile/…
location Vienna, Austria
age 32
visits member for 1 year, 10 months
seen yesterday

Risk Manager at an Asset Management Company

External Lecturer at Vienna University of Technology


Nov
29
revised Itô diffusion processes in finance with unknown distribution at a terminal value
added 108 characters in body
Nov
29
answered Itô diffusion processes in finance with unknown distribution at a terminal value
Nov
28
revised Basics about the scaling property of volatility
added 9 characters in body
Nov
28
comment Basics about the scaling property of volatility
The volatility of stock price is not a good starting point in my opinion. If we first consider the volatility of the return which can have nicer properties (such as stationarity) then we can draw conclusions for the price.
Nov
28
answered Basics about the scaling property of volatility
Nov
28
revised Basics about the scaling property of volatility
edited title
Nov
14
comment Interpretation of cross-correlation matrix when one sample distribution is not normal
You write "Because the distribution is not normal, I'm wondering if the cross-correlation matrix" - how could normality imply anything about cross-correlation?
Nov
14
comment Interpretation of cross-correlation matrix when one sample distribution is not normal
A comment I have to make: You write: "normal with fat tails". A normal/Gaussian distribution does not have fat tails. So in fact it is a fat tailed distribution, I guess. Could be modelled using a t-distribution (as a first attempt). Better approaches do exist but are off-topic to your question.
Nov
12
comment UAC- Unbiased Average Correlation for a Matrix of stocks
Do you have a reference for the notion of UAC? I know average correlation, what is UAC? Where is the connection to auto-correlation (AC)? Usually there is no significant AC.
Nov
12
comment Risk and Reward in practice
no problem, I like this exchange of views. For your latest comment: I am just wondering how much time today's investors give a PM to prove his abilities. Investors are nervous and have rather short horizons - and I understand that. As a result they do not give you too much time ... the business is a bit poisoned with short-horizon thinking in my mind.
Nov
12
comment Risk and Reward in practice
For your first comment: I am neither a PM nor a trader so I don't feel the need to defend any of them. You are right that it would be optimal if PMs could forecast accurately in crisis secenarios but on this planet in the big US/UK/FR/GER companies I have not seen such wizards - all just human beings ;) For the second: I agree, I must not fear underperformance but protect my clients money. But these clients might take their protected money and give it to my peers. That's the other side - it is a dilemma. Last comment from my side.
Nov
12
comment Risk and Reward in practice
I am not totally sure whether we disagree. I also don't like the concept of relative performance for a market fund. I was reather thinking of alpha or smart beta strategies. But also in your case of a market fund: it might take some time until you sell your stock and hold cash - so you have incurred losses. Then in the rebound your investors want to participate - but you are in cash. As long as we can not see the future we are in this dilemma.
Nov
12
comment Risk and Reward in practice
Don't be scared ... this is rather my personal view. If you stop a strategy after you have lost a certain amount, then when do you re-entry? I don't mean an entry signal. Just starting the strategy again in the beginning of the the new period is mostly ad-hoc .. although that's what people do. The numbers look nice .. you never lost more than $x \%$ per period. But is there anything besides this fact optimal in this procedure? What I mean is that in portfolio mgmt an easy thing such as a loss limit is quite subtle. You could underperform the peers. In trading this is different I guess.
Nov
12
comment Risk and Reward in practice
Thanks for your answer. I think I get the feeling that you want to transport: clear and supported limits that fit to the mandate. In fact we look at net/gross exposures (just as you described) and risk measures (VaR/ES taking into account all correlations) and stress tests (historial and discretionary). Loss limits are arguable in portfolio management because of the question of re-entry (when if at all and if you do not re-entry then what else do you do ;)
Nov
12
asked Risk and Reward in practice
Nov
11
revised Parameters for numerically fitting t-distribution to log-returns
added 244 characters in body
Nov
11
answered Parameters for numerically fitting t-distribution to log-returns
Nov
5
comment How to properly take averages to reduce data in regression/panel data analysis
Doesn't this rather fit to stats.stackexchange.com ?
Nov
4
comment Comparing Backtests of Value-at-Risk and Expected Shortfall
In my mind the following is a fact: risk measures have theoretical properties. They can be derived from their definitons under specific assumptions on the probability space, given distribution functions and so forth. Any statistical estimator should have these properties too (e.g. ES is sudadditive, any statistical estimator of ES should be subadditive too).
Oct
31
awarded  Popular Question