Forums  > General  > Trading arrangement  
     
Page 1 of 1
Display using:  

RFMontraz
NP Italian Stallion

Total Posts: 2014
Joined: Mar 2004
 
Posted: 2017-04-26 09:49
Hi guys

An acquaintance has been offered this trading arrangement and I wonder if any of you has heard of similar set ups. Background: he is an ex prop-trading with 10+ years experience, semi-retired, had no interest in moving to the city where this firm (big commodity trader) is based (I have to stay light on details for confidentiality) and wants to keep trading his own portfolio on the side.

He will work as "advisor" with a notional allocation of 100M USD. His daily var is 500k (so AUM is 200 times var - for what is worth talking in terms of AUM anyway). His max DD is 4M USD after which he gets shut down (but to be honest he can be shut down any minute). His target his to make is var net of all fees, so 8M USD per year. His remuneration is not in terms of management/performance fees but in terms of commissions - X times higher than what they would pay normally - generated by trades (which, based on past performance, should equal to 10% PnL). I imagine this is to get around the fact that he is not an employee. Finally his trades need to be pre-approved and can only trade with limit orders (I imagine to keep conflicts with PA to the minimum).

It seems a pretty sweet deal as he is not committing any capital of his own and his remuneration is not really subject to performance (at least in the short term).

Is this credible? Do the figures above pan out in your opinion? Have you heard of similar arrangements?

Just curious

Thanks RM

ronin


Total Posts: 193
Joined: May 2006
 
Posted: 2017-04-26 10:45
It's weird.

He is supposed to generate performance, but he is paid for turnover. To be paid he needs to trade at least every few minutes, but he needs to have each trade pre-approved. He can only use limit orders (no, me neither).

The opportunity cost of saying yes to this is presumably zero, but I can't imagine something like this can last more than a few months - even with the best intentions. There are too many contradictions.


"People say nothing's impossible, but I do nothing every day" --Winnie The Pooh

RFMontraz
NP Italian Stallion

Total Posts: 2014
Joined: Mar 2004
 
Posted: 2017-04-26 18:47
To be honest I have seen this sort of compensation structure before, mainly with CTA in US. Some even offer the option to have an account which is charged man+perf fees (usually this is for a corporate account traded by different traders) OR only brokerage fees (usually this is for a personal account traded by just one specific trader), which is essence equal (based on past trading activity) in terms of charges. They do not need to "churn" the account to get paid, fees per lot are high by design, like 10 or 15 bucks a contract.

I'm skeptical about the 500K daily var equal 100M AUM. Does not make much sense, I think the max multiplier is 100 (so 50M) but I guess this was an ego trip..

ronin


Total Posts: 193
Joined: May 2006
 
Posted: 2017-04-27 11:32
I am not saying nobody does it. But you have to worry about how things evolve. Initially, the trader and the investor are 99% aligned. A few little things change, and suddenly they are only 70% aligned. Once they are 50% aligned, it's over.

The AUM number is pretty much irrelevant fwiw - it's just used for charging management fees, and there aren't any.

The numbers sound like some tight short gamma strategy.

The thing is that for short gamma you have to be able to trade when the opportunity comes. I would worry about the need to approve each trade - what happens if you need to trade, and the approver is just getting his coffee? And the second approver has diarrhea and is stuck in the toilets? And the third approver just had an emergency call from his kid's school and has to go?

I understand the thinking behind it. Presumably it is a proxy for limit on turnover. I would go back and renegotiate it to an actual limit on turnover, instead of per trade approvals.

But then a limit on turnover has other meanings when the trader is paid for turnover instead of performance. Suddenly the trader is short his own gamma, instead of being long his own gamma. This sort of thing can easily lead to various types of perverse behaviour. Which brings us back to my point 1.

So it doesn't really sound like a professional investor, to be honest. It sounds like some hnw trying to immitate a professional investor, and missing the key points. That sort of thing can not possibly be stable.


"People say nothing's impossible, but I do nothing every day" --Winnie The Pooh

RFMontraz
NP Italian Stallion

Total Posts: 2014
Joined: Mar 2004
 
Posted: 2017-04-27 16:50
Thank you Ronin.

Going back to the numbers (I cannot comment as to what he trades or how plans to make money as I simply don't know), this conversation made me think, in general given X as your daily var, what would be the correct metrics in order to express your notional AUM, budget, Max DD?

I would have said:

Notional AUM: 100*X (for the sake of conversation)
Budget: 8*X
MaxDD: 5*X
Worst Month: 4*X (4.47 to be precise)
Max DD Length <12m


The above net of "fees" (defining them as you prefer but here I include the manager compensation in all its forms, %, commissions, discretionary bonus etc..)

I guess different shops use different metrics but am I in the ballpark for a discretionary trader (no HFT and all that jazz)?

In terms of real AUM it means that with 1% daily var you try to make 8% NET a year with a max 5% DD. I imagine the whole world will say it is too conservative but I cannot see many examples of people achieving those numbers on a long enough period... Thoughts?

EspressoLover


Total Posts: 221
Joined: Jan 2015
 
Posted: 2017-04-27 22:53
I don't think those numbers add up. At 8% returns and 1% daily vol, the Sharpe is 0.5. Yet max drawdown is only 5X daily vol. Unless the arrangement's drawdown resets at the beginning of the year (which doesn't make sense), I don't think this is consistently feasible.

Intuitively the S&P has a similar Sharpe. In the past twenty-five years, MaxDD has only been <5.0*DailyVol during four years. The median ratio is 11.3, and in six of those years the ratio was north of 30. Yes, obviously the return distributions aren't the same. But the differences would need to be pretty extreme. The strategy's daily returns have to either be heavily right-tailed, or extremely mean-reverting. Otherwise you may get lucky a year, but getting stopped out in two is nearly a foregone conclusion.

If you take the numbers from the original post (with half the daily vol, but same returns and DD tolerance), it's a little more workable. Then it's 1.0 Sharpe with MaxDD of 10.0*DailyVol. So, let's hack this by bumping S&P's daily returns 4 bps to make it a 1 Sharpe process. It still gets stopped out nine years out of twenty-five. You probably survive a year, about 50/50 you survive two. But in five years there's a 90% chance you're dead.

I think the best approach to making this workable for both parties, would be to negotiate successively rising DD tolerance for every month the strategy's profitable. Tight stop losses make sense for an untested trader, but the longer he consistently delivers the more they should be relaxed. It's in both parties interest. The trader gets to stop dealing with the stress of playing Russian roulette, and the investor reaps larger returns by allowing proven traders to step up risk.

ronin


Total Posts: 193
Joined: May 2006
 
Posted: 2017-04-28 10:38
I am also not a great fan of management by numbers. Different strategies have different risk profiles. Numbers that make sense for mean reversion don't make sense for trend following etc. It's not all about minimizing the vol of daily returns.

It's not that difficult to design a strategy that (say) makes one dollar per day, and loses a million once every couple of thousand days - you can pass any tests based on Sharpe or var (in fact, you even get a Euromoney Hedge Fund award or two), but it's not necessarily something I want to be running.

@rfm, the numbers from the orignal post don't make much sense to me either. You don't need a hundred million to generate 500k worth of var. So if you need a hundred million and you are generating 500k of var, my first question is what other risks do you need the hundred million for, and why are they not showing up in var - where are they, and how big are they?

So I second @el on this one - I am happy with volatile daily returns if I like the strategy and the trader.


"People say nothing's impossible, but I do nothing every day" --Winnie The Pooh

Martinghoul


Total Posts: 847
Joined: Oct 2008
 
Posted: 2017-04-28 12:46
It really sounds mighty weird, I have to say... What happens if he doesn't make budget, but trades a lot?

Insofar as I may be heard by anything, which may or may not care what I say, I ask, if it matters, that you be forgiven for anything you may have done or failed to do which requires forgiveness...

AndyM


Total Posts: 2306
Joined: Mar 2004
 
Posted: 2017-04-28 13:07
Same...what could be the possible rationale for this sort of payout structure?

I used to be disgusted; now I try to be amused...

RFMontraz
NP Italian Stallion

Total Posts: 2014
Joined: Mar 2004
 
Posted: 2017-04-28 15:10
To be honest I have seen this payout structure (commissions on trading instead of man+perf fees) already with different CTAs, McVean and RGC to name a couple. I'm not sure the actual reasons behind it, whether is to get around some regulatory issue or something. As a matter of fact I personally invested with a CTA which had this very set up (you could invest with a team and be charged man+perf or with him only and pay brokerage at 12.5 USD per side). His AUM at the time were roughly 40M (his split of the group) + 80M personal, not huge but not irrelevant either (notional allocation - ie. without having to deposit the full amount you want him to trade - was permitted). If you feel he churns the portfolio aside from cutting him lose you can refer him to the CFTC etc... as always it is a matter of integrity/credibility, if one is dishonest he can be dishonest no matter the checks and balances. Here this risk is eliminated by the fact that the trades need to be pre-approved by the "client" (that was not my case but never had issues).

To be honest I'm surprised you find it so weird as the more I think about it, it kind of makes sense. The client has full visibility on the portfolio and keeps custody of the assets, pre-approves trades and risk (this addresses ronin's point, the clients sees the trades, not just the realized returns and Martinghoul point, it prevents churning) and can stop trading at any time hence firing him without having to worry about employment litigation issues (and I'm sure it is also cheaper from a tax/benefits perspective while the relationship still works)

Regarding the metrics the whole 100M AUM banter was clearly an ego trip on his part (they allocated me 100M!) and completely irrelevant given that there is no man fee. I wonder though, for a prop trader who gets his allocation in Var terms and not in AUM, you all think the multipliers i suggest below are wrong? If so, what are the correct multipliers in your opinion?

NB I would not mix Sharpe with Var. Sharpe is based on realized volatility. Var is ex-ante and here is understood as a limit, for the two to coincide you would need the trader to constantly "use" all his var, which is never the case (hit your var once and already questions will be asked), so annual realized volatility will be a fraction of the annualized var

Martinghoul


Total Posts: 847
Joined: Oct 2008
 
Posted: 2017-04-28 17:33
I would be really curious to get an understanding of the reasoning behind such an arrangement.

Far be it from me to suggest that I know how all such deals could be structured, but I still find it weird. I mean I am having trouble understanding how and why commissions would figure into the compensation structure at all. I am sure I am missing smth.

Insofar as I may be heard by anything, which may or may not care what I say, I ask, if it matters, that you be forgiven for anything you may have done or failed to do which requires forgiveness...

EspressoLover


Total Posts: 221
Joined: Jan 2015
 
Posted: 2017-05-01 12:23
Ahh, misunderstood the meaning of "var"... Value-at-risk, not variance... So, taking that into account puts to rest the drawdown issues, but the shoe gets put on the other foot. It makes the returns target really hard to hit. VaR is obviously defined different than vol, but to a first approximation it's still O(vol). At a minimum daily VaR is at least 2.5x average realized daily volatility. Realistically the ratio's probably even higher.

Taking the numbers from the first post, the trader has to make $8M a year while keeping daily volatility below $200K. That's 2.5+ Sharpe as a minimum target, which is really aggressive for a discretionary trader. Quantum Fund at its peak was still below 2. SAC since inception is about 2.5. John Arnold was maybe 3.0 at his best. Besides for world-famous traders, the only discretionary traders I can think of who hit targets like that are A) interacting with proprietary flows (e.g. sell-side desks or traders at oil companies). Or B) intraday click-traders who are quickly opening and closing positions (which doesn't work in this context because of the pre-approval).

The cynic in me suspects that there's a clause which says the trader doesn't get paid (or only paid a pittance) unless the fund makes it's target for the year. In which case the "swindle" seems clear: Find talented traders. Offer them a deceptively attractive deal. But set sufficiently low risk limits and high hurdle rates, such that it's nearly impossible to win. Bonus points for overly inflating the AUM, to make the target seem easier by comparison. ("All you have to do is make 8% ROI"). Harvest alpha from skilled traders while avoiding paying any incentive fees... Phase 2: ?... Phase 3: Profit.

But who knows, I'm just speculating and obviously don't know anything in particular here.

katastrofa


Total Posts: 357
Joined: Jul 2008
 
Posted: 2017-05-01 21:11
What is the HNW investor has extra-legal means of enforcing alignment?

ronin


Total Posts: 193
Joined: May 2006
 
Posted: 2017-05-02 12:30
>the trader doesn't get paid (or only paid a pittance) unless the fund makes it's target for the year. In which case the "swindle" seems clear


Which would reinforce my point that the investor is an amateur. With traders you run the winners and cut the losers. If your cunning plan is to cut winners early, it ain't gonna make much.

I also wonder about these explicit trade approvals. In London, if you pay for performance, you are obliged to do everything you can to facilitate performance. Fail to approve a trade or two, or just fail to approve them in time - the investor could easily find himself in constructive breach and become liable for all the trader's expected earnings under the contract. Courts enforce that sort of thing all the time.

Bottom line, it screams emerging markets. Very emerging markets.


@RFM, the difference between limit and realised vol isn't as large as that. If somebody isn't using their limits, it means the trading is not running efficiently. Limits cost capital.

My point was more about riskiness of a strategy vs volatility of daily returns. I much prefer volatile daily returns when I have confidence that is it as far as risk is concerned, than some toxic sh*t where I don't know what is going to sting me from the tails and when.



"People say nothing's impossible, but I do nothing every day" --Winnie The Pooh
Previous Thread :: Next Thread 
Page 1 of 1