Forums  > Careers  > CV/interview advice: whether to mention the track record in PA and how?  
     
Page 1 of 1
Display using:  

chika


Total Posts: 9
Joined: Aug 2004
 
Posted: 2017-02-11 11:02
Hi

I have been working as a quant on the sell side. I have been actively involved with time series analysis and trading strategies for vol and correlation trading but it was mostly to manage risk from the flow and the structured business.

However, I have built up my own quant trading model combining the three strategies: carry, trend and vol, and run it with my own money (I have been working on these models during my evening and weekend time). I have had a good performance so far for the past 4 years with returns about 20% p.a. each year and vol of about 20%.

I am now keen to work as a quant researcher or a junior quant PM on the buy side and I am planning to apply to a few places. I have a strong academic background and a good programming experience so I am confident on the technical side. The only drawback is that I have no experience in professional risk/money management and no experience with the buy side, apart from my trading experience in PA.

Do you think it makes sense to mention my experience with quant trading and track record in my PA? If so, should I do it in the CV, the cover letter or during the formal interview?

Thank you

Maggette


Total Posts: 923
Joined: Jun 2007
 
Posted: 2017-02-11 12:32
IMHO it doesn't hurt as long as you make sure to communicate that you don't value that experience too high yourself and sell it as it is: a way to get your feet wet that probably won' t help too much in your new position.
If you manage to not be perceived as crackpot that considers himself the next superstar I think it is a plus that you try to educate yourself with hands on stuff an even put your money at risk.

I am not in the industry any more though, but if I have interviews in my current position to hire young data scientists and developers I look at the stuff they did on GIThub, open source projects and kaggle like competitions. To me it is important.

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

NIP247


Total Posts: 539
Joined: Feb 2005
 
Posted: 2017-02-11 17:15
Goes without saying that derivatives trades were allowed in your department and cleared by compliance. My former house did not allow p.a. trades in derivatives...

On your straddle, done on the puts, working the calls...

radikal


Total Posts: 252
Joined: Dec 2012
 
Posted: 2017-02-11 18:48
Yes, but be prepared to answer questions about it.

Sharpe, drawdowns, greek levels etc. It's a good chance for you to demonstrate that you know how to think about edge vs risk.

There are no surprising facts, only models that are surprised by facts

chiral3
Founding Member

Total Posts: 4969
Joined: Mar 2004
 
Posted: 2017-02-11 22:42
I don't know how you guys do it. My compliance questionnaire is longer than a mortgage application. There are securities that require pre-clearance that I have to go to BBG to look up because I've never heard of them. Between taxes and oversight it's become barely worth it.

Maggette gives some really good advice.

Nonius is Satoshi Nakamoto. 物の哀れ

goldorak


Total Posts: 979
Joined: Nov 2004
 
Posted: 2017-02-12 06:00
> Sharpe, drawdowns, greek levels etc

If these are the first questions you get about it, then may be you should not consider a position with them.

They may pay extremely well though...


If you are not living on the edge you are taking up too much space.

chika


Total Posts: 9
Joined: Aug 2004
 
Posted: 2017-02-12 09:35
Hi Guys

Thank you for your valuable advice.

I started investing/trading in the PA at a different bank. The conditions were pretty standard: 30 day holding rule, pre-clearance for single stocks. It was ok, I learned to cope with it. When I changed my job a couple of years ago, they first said that I can keep the existing account but then once on board they said that I need to close that account and move my money to their platform. However, that platform is very primitive intended for long-only with transaction costs prohibitive even for low frequency. At that time I already invested a considerable amount of personal time into research / back-testing these strategies and quant trading models (especially for vol) became my passion, so I moved capital to my wife account and continued trading from there.

In the end, I learned to concentrate positions and allocate capital only in the top quality stocks and investment ideas following rigorous quant analysis. The diversification comes from being exposed to different strategies that perform well in different market conditions and then to dynamically manage capital allocated to them.

For an example, my current positions and rebalancing frequency:

Carry: 11 positions (with 4 positions accounting for about 50% of the risk) with rebalancing is very rare: one insignificant trade in a month or two mostly coming from some option exercise. Last time I completely closed a position was in August last year.

Trend: 3 positions with rebalancing may be once a month but also more sensitive to market conditions. The universe is 11 asset classes as for the vol strategy.

Vol: the risk is spread across 11 asset classes/equity sectors each traded by ETFs, currently around 100 positions and some hedges. I do monthly and bi-monthly rolls and by the necessity the trading must be more frequent. I developed my relative value approach with hedging / rebalancing strategy which works ok. I have also implemented my own risk system with detailed risk reporting that provides indications at which levels I should rebalance. I live in Europe and trade US markets, so I can do most of rebalancing in the evening.

Recently I had a brief chat with a senior guy from a top-5 hedge fund who also run a bunch of systematic vol strategies. Our operational approach is actually very similar apart from the fact that they trade in hundreds of markets.

Finally, the drawdowns should be managed by exposure to different strategies that outperform in different market conditions:

Carry performs well in low vol regime, underperforms in high vol regime (especially with high vol in rates).

Vol performs well in mean-reverting / moderate vol regimes.

Trend outperforms in a trending markets with moderate/low vol.

Proper macro risk management can be achieved by dynamically sizing exposures to these strategies. In my account, I had -10% loss in November 2016 following Trump election because losses in Carry and Vol strategies mitigated by long exposure to financials in Trend strategy. However, the risk adjustment for Trend was good (by getting more exposure to financials and short in USTs) and the account was back up 10% in December as Carry and Vol rebounded as well. So far in 2017 it up about 5%. On the other hand, in June 2016 following Brexit, the Carry outperformed and compensated for losses in Vol with 10% overall gain in June.

Patrik
Founding Member

Total Posts: 1325
Joined: Mar 2004
 
Posted: 2017-02-14 17:03
> so I moved capital to my wife account and continued trading from there

Ouch.. I can tell you aren't trying to sidestep the holding rules etc here, but you do have a compliance problem either way.. Depends on firm you speak to obviously, and who you are speaking to, but at least some will probably have more trouble with the "trading in my wife's name to get around annoying rules" than they would have a positive impression from the results. So you have a tricky decision to make of how you try to deal with that.

Capital Structure Demolition LLC Radiation

chika


Total Posts: 9
Joined: Aug 2004
 
Posted: 2017-02-14 21:33
Thank you, Patrik. To make it clear, I do not feel good about it. The question is not that I needed to abide to 30 day holding period or get a pre-clearance, it was that the new firm wanted me to move my account to their platform, but that would make technically impossible to apply the strategies that I do. For me it is more about the proof of the concept, not about getting rich quick. If I have an idea and a quantitative/scientific model which I think are good, I want to put my money to prove that they work, not to publish a paper or sell it to someone else. For past few years it kept me busy and motivated in my spare time. I did bring forward some of my ideas to the desk at my work and I do believe they benefited from them as well. It is just a different type of business so why I am looking to an opportunity to align my interests with my direct responsibilities.

Patrik
Founding Member

Total Posts: 1325
Joined: Mar 2004
 
Posted: 2017-02-15 00:36
Understood. I'm just pointing out the potential problem you could face bringing up PA track record, adds a different dimension than what the original question was about.

Capital Structure Demolition LLC Radiation

peter.russell


Total Posts: 10
Joined: Dec 2016
 
Posted: 2017-02-15 09:17
Out of curiosity, is your rebalancing mechanics done automatically? or does it still rely on some level of discretionary decisions?

chika


Total Posts: 9
Joined: Aug 2004
 
Posted: 2017-02-16 22:36
Thank you Patrick, I do realize that now. I decided not to mention it at al. I know the staff, I know how it works, I could say I did it for the desk and it worked. Besides, I may have some other career options.

Peter, the rebalancing signal is generated by a model. The key is to have the same model for signal generation and capital allocation for both vol and trend strategies.

EspressoLover


Total Posts: 201
Joined: Jan 2015
 
Posted: 2017-02-17 04:43
Why not lever it up more? At 20% return/20% vol, you're way under-Kelly sized. Since you've been running it successfully for 4 years, it's not like you're worried about implementation falling short of backtests. It's also your P/A, so it's not like you have to please dipshit institutional money which fetishizes 10% vol targets.

Consider, if you scale up to 40/40 and keep similar performance, it's very unlikely you'll underperform 20/20 after ten years of (monthly/daily) compounding. And most likely you'll have multiples more ending capital.

NeroTulip


Total Posts: 991
Joined: May 2004
 
Posted: 2017-02-17 09:04
Well, full Kelly would be 100% vol... Three reasons why you should *not* do this:

- Half Kelly gives you 3/4 of the returns with 1/2 the vol, which is easier to trade from an emotional p.o.v.

- Blowup risk. People tend to think that Kelly cannot go broke. This is true with discrete distributions, but not with continuous ones. As Kelly maximizes the *expected* growth rate, not all paths avoid the absorbing barrier.

- "If you scale up [] and keep similar performance"... is a big if. Edge erosion, capacity constraints, changes in market conditions, etc... Better be conservative in estimating your forward Sharpe.

HTH

"Earth: some bacteria and basic life forms, no sign of intelligent life" (Message from a type III civilization probe sent to the solar system circa 2016)

goldorak


Total Posts: 979
Joined: Nov 2004
 
Posted: 2017-02-17 13:29
> it's not like you have to please dipshit institutional money which fetishizes 10% vol targets Applause




If you are not living on the edge you are taking up too much space.

EspressoLover


Total Posts: 201
Joined: Jan 2015
 
Posted: 2017-02-17 19:35
@NeroTulip

I agree with most of your points. We're pretty much on the same page. I think I gave the wrong impression by saying "way under-Kelly sized". It wasn't to advocate full Kelly sizing, just saying that I thought 0.2 (estimated) Kelly fraction seemed too low for OP. For all the reasons you list picking the right Kelly fraction is more art than science.

I suggested 0.4 as a fraction, which I think is justified given OP's situation. He has actual realized live data for four years, which gives a relatively tight standard error of 0.5 on the Sharpe estimate. It's also fairly long enough to reasonably expect this alpha not to disappear overnight. I would hope and presume he also has much longer backtests, and performance is in-line with live trading. Finally carry/trend/vol aren't some weird black-box strategies. They're very well known and studied, and documented to have worked for decades in nearly every single market.

> Blowup risk. People tend to think that Kelly cannot go broke. This is true with discrete distributions, but not with continuous ones. As Kelly maximizes the *expected* growth rate, not all paths avoid the absorbing barrier.

Continuous Kelly assumes continuous re-balancing, so the further away you're from this assumption the less appropriate it becomes. If you're trading illiquid stuff (or illiquid relative to your portfolio size) then it's definitely possible to become stuck in an over-leveraged position after a sharp decline.

However I really don't think this applies to OP at all. It sounds like he's trading major liquid instruments on a retail account. Daily re-balancing pretty much makes it impossible to go broke even at full Kelly size. He'd have to hit a one-day 100% decline to become insolvent. At 40% annualized vol, that's 2.5% daily vol. The insolvency scenario would require a 40-sigma downside move. Even with just monthly re-balancing, you'd still have to pull an 8.5 sigma drawdown. Maybe that's feasible in some esoteric quant-factor, but I think it's extremely unlikely in trend/carry/vol.

chika


Total Posts: 9
Joined: Aug 2004
 
Posted: 2017-02-18 01:33
Thank all you for valuable inputs. Here are some comments

> Why not lever it up more? At 20% return/20% vol, you're way under-Kelly sized.

I am actually quite leveraged (about 1 to 3 accounting for implicit leverage for short option positions) and close to an optimal Kelly. I size my position as function of the account value and the long-term estimated vol of the strategies. The key problem in multi-asset / multi-strategy portfolio is the tail correlation. You know that you run strategies that are supposed to be uncorrelated over the long scale (1 year and more) but over short scales is can be harmful. It is inevitably hard to manage drawdowns over a short period of time. But once the adjustment is made you can take more risk on strategies that suffered the most but which assume some kind of mean-reversion (like short vol or long dispersion). I think the Kelly assumes stationary chances of the success while in the real markets these chances are not stationary and it is always good to have an extra capacity to take risk when the trade-off is good.

A recent good example was the aftermath Trump elections: carry and vol strategies suffered a lot in November, but if you had capacity to roll into this strategies in December, the return over next one/two month would be spectacular. Remarkably the trend strategy (long financials, short USTs) performed since early December as well.

>> Consider, if you scale up to 40/40 and keep similar performance, it's very unlikely you'll underperform 20/20 after ten years of (monthly/daily) compounding. And most likely you'll have multiples more ending capital.

Again, for me it is all about the capacity to take risk aftermath of drawdowns or break-outs (when short term moves exceed one strandard deviation). The optimal long-term solution to outperformance is to keep a concave profile for return on your strategies in normal markets and a convex profile in trending markets . However, that implies having drawdowns during the transition phase.

>> Finally carry/trend/vol aren't some weird black-box strategies. They're very well known and studied, and documented to have worked for decades in nearly every single market.

This is very true. I have been investing in my PA since early 2009 and I have tried different strategies (including buying lottery/options, investing in small caps, trying to time the market, etc). I did pay my tuition fee, worth of fees for a top-notch MBA, to learn that carry/trend/short vol (to be presice, by short vol I mean being long mean-reversion) work most reliably compared to anything else. Most importantly, you can actually build quant models for these strategies and estimate their expected alphas and risks.

The long-term frequency of normal markets (with moves within one standard deviation) is about 67% while trending markets (with short/medium term moves above one standard deviation) are likely to occur 33% of the time. Again, by mixing strategies that outperform in different regime, one can improve the overall through-cycle performance. Note that if any strategy has a long-term alpha, the major component of this alpha is due to the cyclical risk when the strategy underperforms significantly if it is outside of its assumptions. As a result, the over-the-cycle alpha is some sort of a compensation to bear the cyclical risk. A proper diversification is then mitigating the effect of the cyclical risk of diffirent strategies, and not the effect of the intrinsic risk within a strategy.

>> Continuous Kelly assumes continuous re-balancing, so the further away you're from this assumption the less appropriate it becomes. If you're trading illiquid stuff (or illiquid relative to your portfolio size) then it's definitely possible to become stuck in an over-leveraged position after a sharp decline.

The liquidity is not only applicable for ability to manage your risk, it's also about understanding your position and how it can perform in specific market conditions. I only invest in ETFs/to large caps because my models can predict more reliably their performance in specific market conditions and there is no excess costs (illiquidity, unfair valuation, etc) to get in and out of positions. Trading systematic strategies in OTC or illiquid instruments is the biggest mistake one can get into.



Previous Thread :: Next Thread 
Page 1 of 1