Anyone understand this. Page 21 - 22, Book 5 of Schweser. Seems like there are trading costs, but they get compensated for with market beta.
Honestly, that paragraph in Schweser makes no sense…
if your implementation shortfall model says that the calculated shortfall is like in the example in the book of 87 bps,
based on beta of the stock and the way the market moved (example in the book beta is 1, Rm=1% and given that the alpha factor for daily moves would be pretty close to 0 - you have the stock going market movement = 1%
your market adjusted implementation shortfall is 87bps - 1 = -13 bps.
Trading actually accounted for 87 bps. And here you ended up with a negative shortfall - when the effect of the market is removed.
IMO,
In SCH, the calculate Implementation shortfall cost is 42 bps, being the difference btn decision price & execution price. This is a positive cost meaning in order to get manager’s wishes executed, 42 bps of the return on paper portfolio is foregone due to trading cost (includes both implicit & explicit).
Manager based on his analysis, called to buy the stock & passed buying orders to trading desk. But there crept up a difference between the price at which manager was looking to buy the stock & the price at which trading actually took place.
42 bps is + ve so we may conclude that the we our trader is not working properly.
However market would have also moved in between leading to difference in price. So we may not straight away attribute 42 BPS to trader’s kitty. We would adjust for market movement & then see our trading cost.
So shown in schweser, the stock had a beta of 1.2 & market return during the same period was 0.8%. The expected return from the stock should have been 0.8*1.2=0.96% (since alpha is almost zero over few days)
Market Adjusted IS = 0.42-0.96= - 0.54% (-ve trading cost)
Any takers?
By delaying, trader has also captured the market movements apart from the cost covered under Implementation shortfall. So actual trading cost is - ve
So conceptually, does this mean that if measured implementation costs were 42bps and the stock ralled by 1% the day the trade was executed, you actually had negative implementation costs since the stock went up?
Seems to be combining execution costs with stock performance. I don’t know what use this would be…
Implementation costs are always negative. you are comparing cost of the implementation against that of a paper portfolio.
If trading costs are 42 bps, and market went down by 1% -> your total cost of the trade = 1.42% (0.42 - (-1) = 1.42).
If however market went up 1% -> Total cost of trade = 42 bps - (1) = -58bps. Market going up took off some of the edge of your trading shortfall. So you actually did not lose as much, but you also did not gain as much as you would have if you had 0 implementation costs.
Hope I have been able to explain this.
Yep, got it. Thanks for the help thinking through that.
But implementation shortfall is computed basing on stock movements (hence we can compare the return assuming perfect execution and return based on actual execution). It already considers that the stock moved.
So if we include the beta-stock movement, wouldn’t that be double counting?
expected stock movement ( a.k.a. return ) = beta * market movement.
Trader’s skill must be able to earn excess return over the expected stock return of the stock. This will disambiguate skill and speed from normal stock price movements, especially when measured over large nnumber of trades
I still don’t get it.
If over the same period of time the stock price increased by 0.8*1.2=0.96%, isn’t that completely unrelated to the 0.42% implementation shortfall? What has the inability of the trader to execute at the decision price got to do with the market-driven price movement of the stock?
Also, why are we complicating things by treating the implementation shortfall as a positive number, from which we subtract the increase in the stock price? What is the intent here and what is the meaning of the resulting number?