Modeling Trading Costs for Daily ETF and Mutual Fund Strategies

Summary

The document considers how to represent transaction costs in a daily strategy that reallocates between a theoretical S&P 500 total-return ETF and a mutual fund earning the three-month Treasury bill rate. The strategy may also borrow at the risk-free rate to increase exposure to the risky asset. The proposed data are limited to closing index levels and daily bill rates, so the question is how much cost detail can be justified in a thesis backtest.

The answer suggests that direct trading fees for these hypothetical, highly liquid exposures are small enough to ignore for the stated purpose. It notes that mutual funds transact at closing prices, ruling out intraday rebalancing, and flags taxes as potentially more important when turnover is high. It also invokes research on factor portfolios as context for low costs in liquid S&P 500 ETFs. The reply gives no fee schedule, spread estimate, borrowing treatment, or turnover-based calculation, so its conclusion is a simplifying judgment rather than a measured cost model.

Key ideas

• The strategy reallocates daily between a risky equity exposure and a short-term Treasury bill proxy.
• The answer considers direct trading fees on the assumed liquid assets negligible for the thesis context.
• Mutual fund dealing at closing prices constrains implementation to end-of-day trades.
• Taxes may matter more than direct fees when trading frequently.
• No detailed estimate is provided for spreads, borrowing costs, or turnover.

Tags

# ETF/Mutual fund transaction costs in algorithmic trading


# ETF/Mutual fund transaction costs in algorithmic trading












For my thesis, I'm backtesting a strategy that invests in a two asset universe which consists of a theoretical ETF based on S&P 500 total return index (with zero tracking error) and a hypothetical mutual fund which pays exactly the daily yield of 3 month treasury bills (as risk free asset).

The strategy operates in a daily timeframe and invests in a dynamic weighted portfolio of this two assets. I want to incorporate the trading fees for a more accurate evaluation of my strategy. I might add that the strategy can sometimes borrow at the risk free rate and invest in the risky asset (S&P ETF in our case) as well.

Is there a simple and yet closer to reality way of handling this issue? My data consists only closing prices of S&P total returns index and daily 3M TB rates. The focus of my thesis is not on transactions costs but I want to make sure I am doing it right academic wise.

## Answer by phdstudent (score 3)

https://quant.stackexchange.com/a/81753

There's pretty much no trading fees on either of those two assets. You can safely ignore trading costs. Now mutual funds can only be traded at closing prices, so any intraday rebalancing is out of the question.

More importantly than trading costs for your assets are taxes, specially if you are doing such amount of trading.

For example, trading SMB and HML had very little trading costs. You can safely assuming that trading a liquid S&P500 ETF will have even lower trading costs even if done daily.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.