Tcost model
For the quants out here: what tcost models do u use? Is it from vendor/in-house/broker? Feels like it’s often an under appreciated area, yet so important in optim.
Trending Content - Hedge Fund Forum
| +49 | Do Quantamental roles really exist? | 24 | 6d |
| +17 | L/S Equity Career London | 11 | 3d |
| +17 | Buy Side Credit - Seats / Guide | 3 | 2h |
| +13 | Hedge Fund Pivot | 3 | 22h |
| +11 | Mid-career shift to buy side with niche background | 3 | 1d |
| +9 | Wfh / work remotely at MMs or SMs? | 5 | 1d |
| 0 | Hedge funds | 1 | 1d |
Career Resources
Broker models. Unless you're Citadel, your broker will have orders of magnitude more transactions than you to fit their parameters. They also update and send out daily to their clients. Just ask.
a lot of quant groups build their own internal models. I mean, depends on frequency and instruments. But there are general lower bounds you should be able to beat for you tcost model to be acceptable. Eg. vwap
Depends on the kind of the strategy you are running- if it is stat arb with thousands of instruments at 20-25% turnover , a simplistic assumption such as a fraction(0.1 to 0.5) of average daily bid ask spread would do just fine .
Don't forget market impact.
The more you trade the more you move the price.
Kyle model covers that, but Citigroup came out with more research about 15 years ago.
There might be even more current analysis on this.
Remember, these are both assumptions for simulation as well as levers for controlling the optimizer, and they mean two different things depending on what you are trying to do.
Well of course there is market impact - no one can deny it , but unless you are operating at a very large scale(5B+ gross) and/or very high turnover , there is not much value in incorporating it into your models - it would only introduce noise. You can limit your impact simply by targeting no more than x%(1-5%) daily volume on any given day with VWAP execution. The extra effort to model for impact isn't simply worth it.
Point. We did try to model it off of implementation shortfall.
Guy number one is the person writing the basic alpha signals and tossing it into the vendor optimizer.
Guy number five is the portfolio engineer.
Guy number twenty is basically running the bespoke optimizer when the PM gets too pssd off at dealing with vendors.
Guy number thirty is testing new risk factors and running attribution on them.
Guy number forty is doing all of the internal crossing between different portfolios.
Guy number sixty is running the bespoke market impact modeling. And even then, he's sticking his thumb up in the air and coming back with single-digit t-values.
It is not priority #1, but it matters. And if you can herd all of the cats into a room and lock the door until the PMs can agree to an attribution model that they think is fair, there's a lot of value in a joint book or at least risk crossing and tcost optimization. With that, you've now got position crossing, risk crossing, and joint tcost optimization.
Figuring out how to do this and keeping all of your PMs and not having it be political or a PNL or power grab is what separates the mediocre managers from the terrific ones.
Quo possimus velit distinctio nihil necessitatibus aut ut. Animi voluptatibus molestiae quod ipsum animi voluptates. Maiores deleniti necessitatibus in similique deleniti natus consequatur. Vel est voluptatibus impedit. Sequi non aperiam quia at ut voluptate sit.
Fugiat non et ut earum quis animi voluptas. Esse dolorem est optio optio. In amet excepturi quibusdam blanditiis dolorum facilis.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...