Page 1 of 1

Is model validation really that bad?

Posted: June 2nd, 2015, 11:52 am
by barny
It seems like for someone with a PhD in Physics but without an MFE and much experience in quantitative finance that model validation would be the perfect place to start. Why is everyone so negative towards it?

Is model validation really that bad?

Posted: June 2nd, 2015, 1:29 pm
by DevonFangs
how bad it is depends on the bank, but maybe in general because it's difficult to get out

Is model validation really that bad?

Posted: June 2nd, 2015, 3:36 pm
by slacker
QuoteOriginally posted by: barnyIt seems like for someone with a PhD in Physics but without an MFE and much experience in quantitative finance that model validation would be the perfect place to start. Why is everyone so negative towards it?What are your broad longer term goals coming to finance? What is the probability of attaining those goals with a career in MV?Maybe those with a negative attitude foresee that the two paths do not cross for them.

Is model validation really that bad?

Posted: June 3rd, 2015, 4:46 am
by Gamal
QuoteOriginally posted by: DevonFangshow bad it is depends on the bank, but maybe in general because it's difficult to get outAnd if want to get out from somewhere - it must be a bad place ;)You rarely validate models in model validation, that's the main problem.

Is model validation really that bad?

Posted: June 3rd, 2015, 7:58 am
by ThinkDifferent
QuoteWhy is everyone so negative towards it? who's "everyone"? FO quants? Sure they are -'ve. They work twice more, get their arses whooped by the desk, and get paid marginally better :)

Is model validation really that bad?

Posted: June 3rd, 2015, 1:22 pm
by londoner
Don't FO quants get a typical bonus of 100% versus 30% in model validation? This is not a small margin even after tax.QuoteOriginally posted by: ThinkDifferentQuoteWhy is everyone so negative towards it? who's "everyone"? FO quants? Sure they are -'ve. They work twice more, get their arses whooped by the desk, and get paid marginally better :)

Is model validation really that bad?

Posted: June 5th, 2015, 7:51 am
by Diskiss
100% typical bonus for FO quants? That is sooo 2006 :-)Try more like 100% for outperformers (top 5% in a dpt), 50% average, 0 for underperformers. So vs 30-40% in model val on average (with a lot less volatility y/y and a more stable job outlook) the gap is actually a lot closer.A friend of mine who got an offer for model val at a french bank in london was told they usually pay 50%. So? still negative? :)

Is model validation really that bad?

Posted: June 5th, 2015, 8:30 am
by Gamal
All right, you get your wages in MV. But what you do - does it have any sense?

Is model validation really that bad?

Posted: June 5th, 2015, 9:34 am
by Diskiss
I prefer to think of middle office quants or risk quants than FO model validation quants these daysRegulation is becoming more complex and regulatory compliance optimization is a new area for model generation that are not produced by the front office.i'm thinking HQLA management optimization, cashflow behaviour modelling, stress testing, RENTD modelling... Many areas for the middle office to become more than just a cost center and add value to the bank's business model while remaining independent (different line of report) to the Front Office.

Is model validation really that bad?

Posted: June 5th, 2015, 11:58 am
by Gamal
QuoteOriginally posted by: Diskissi'm thinking HQLA management optimization, cashflow behaviour modelling, stress testing, RENTD modelling... It is not what is usually called model validation.

Is model validation really that bad?

Posted: June 5th, 2015, 12:05 pm
by ThinkDifferent
QuoteOriginally posted by: GamalAll right, you get your wages in MV. But what you do - does it have any sense?we make sure that what you do makes sense.

Is model validation really that bad?

Posted: June 12th, 2015, 8:47 am
by DominicConnor
Part of the problem with MV was that it was over-sold, managers simply lied about the nature of the work, talking to newbies about how examining experienced guys work would make them the next wave of the elite.Thus people tried to leave as soon as they realised the truth, which combined with the bonus structures to make staff retention a bigger issue, which caused hiring managers to lie more.This has now evolved into changing the names of the groups and a bit of spreading the shit work around. Fact is that all types of all work have shit parts and the feedback loops mean that it is not allocated on a basis that makes sense for the whole firm, but for individual managers.

Is model validation really that bad?

Posted: July 23rd, 2015, 4:03 pm
by londoner
QuoteOriginally posted by: Diskiss100% typical bonus for FO quants? That is sooo 2006 :-)Try more like 100% for outperformers (top 5% in a dpt), 50% average, 0 for underperformers. So vs 30-40% in model val on average (with a lot less volatility y/y and a more stable job outlook) the gap is actually a lot closer.A friend of mine who got an offer for model val at a french bank in london was told they usually pay 50%. So? still negative? :)50% is high. All model validators should join a French bank then.