wilmott.com

Can anyone lead to some papers related to measurement of market impract of a trade taking "transaction costs" into account? Desperately in need.Thanks.

ITG has papers on their website. Anant Madhavan did a lot of work in this area. Barra has a market impact model. You can get the paper on their website.

I looked for Barra's hand book but its not easy to find. What is the web site of ITG?

try itginc.com. You need to explain specifically what u need. i wouldn't worry too much about barra and itg. indeed, they are offering their price impact models to their clients. However,Barra's model is intended for extreme impact of extreme volumes, smth like half of the daily volume. It is an economic model, that has pretty questionable prediction power according even to them and was hardly tested, because you cannot test these volumes directly.Itg has two models, one they might talk about a little -- i think it's called ACT or something, another one, newer -- they won't. They won't go into factors they use, or prediction statistics, or anything like it. Their sites are basically useless for anything beyond general background and some Madahvan articles, which have purely academic bend. Ananth doesn't work for itg anymore by the way.if background is what you need i could try to dig barra book out.

I would look for any papers by Hasbrouck (http://pages.stern.nyu.edu/~jhasbrou/Wo ... rIndex.htm) and Almgren & Chriss (http://www.math.toronto.edu/almgren/optliq/ & http://www.math.nyu.edu/faculty/chriss/index.html).And take a look at this paper as a start: http://www.math.toronto.edu/almgren/optliq/optliq.pdf.I can't claim to be an expert at modelling, but in a previous life I reviewed many a "slippage" model and all of them had something to be desired: namely, accuracy. Personally, I think it's impossible to predict slippage, and you are better off focusing on your alpha model. Forgive me for ranting...it's been discussed ad nauseum before.T

Thanks folks. Will post something, if I happen to be successful.

Hi Zerdna, What am looking for is this. We have a japanese equity model and also a rough model for predicting trading costs. We have our own factors for the prediction of the portfolio return. But, the problem is that we are supposedly facing losses owing to the market impact of our own trades and hence we want to measure how much our trade impacted the price of the security, if possible taking into account the trading costs. Of course, yes, this will require the tick by tick data as proposed by most papes, but we do have access to that now. But, I need to implement a good model for predicting the impact. I had a look at the papers as proposed, but the more information I have is better. Any more suggestions are more than welcome. Thanks.

Thanks for your references. I had a look and it would be useful. -Cryptic

I think that a price impact function can be adquate, but it requires more than just a tick by tick database, I mean at least a five limite orders book...

Well, I implemented the Almgren/Chriss model of optimal execution to know the market impact. Haven't done any trading after that and hence don't know if we shall be able to get good results. In addition, we are money managers and we dont get much info from brokers like how many shares traded in the last 15 minutes, etc. These are too many details and we need to make some approximation after talkting to brokers. So, tick by tick is all that we can think of at this stage since we can lay our hands on that. Order book is a different story now and a rather difficult to obtain data. Lets see.

Hi cryptic26If you are still interested in back-testing trading strategies for Japan or other markets using market impact at the time we should talk.