Trading Costs Across Brokers

[ajensen]

The 'Actual Retail Price' of Equity Trades is a recent working paper, covered by Bloomberg.

Abstract
We compare execution quality of six brokerage accounts across five brokers by generating a sample of 85,000 simultaneous market orders. Commission levels and payment for order flow (PFOF) differ across our accounts. We find that execution prices vary significantly across brokers: the mean account-level round-trip cost ranges from –0.07% to –0.45% excluding any commissions. The dispersion is due to off-exchange wholesalers systematically giving different execution prices for the same trades to different brokers. Overall, however, there is no evidence that PFOF harms price execution. In addition, we provide several suggestions for more informative disclosures on execution quality.

Keywords: retail trading, execution quality, bid/ask spread, market microstructure, payment for order flow, commissions, broker-dealers

Table VI
Comparison of Roundtrip Trade Costs
This table compares the average round-trip trade costs of our trades across different brokerage accounts. Average returns are reported in Panel A across two periods, including April 22 to June 9, 2022, when we traded Interactive Brokers’ free account and Fidelity. To account for market movements, costs are measured relative to the midpoint at the trade times. We also show the worst execution, where all buys would be executed at the NBO and subsequent sells at the NBB, adjusted for the midpoint. Panel B reports pairwise difference between brokers for matched trades. A negative return means that the broker reported in the column (row) has better (worse) performance. In both panels, t-values are computed using standard errors clustered by stock. Panel C reports the percentage of trades where the broker listed in the column has better execution than the broker listed in the row.

According to Table VI, the average roundtrip cost for all trades, in basis points, was

7.2 TD Ameritrade
19.7 E*Trade
23.4 Fidelity
31.4 Robinhood
44.4 IBKR PRO
46.2 IBKR LITE
61.9 NBBO (Worst Possible)

I am surprised by the high reported costs for Interactive Brokers and am a bit skeptical, since many professional traders use IB, and they are sensitive to t-costs.

Consistent with Figure 1, we find astonishingly large execution differences between brokers. TD has excellent price improvement, with 99.4% of its trades inside the NBBO and a price improvement of 47.2%. To put this in perspective, a roundtrip trade would pay only 2 × (50% − 47.2%) = 5.6% of the quoted spread. In dollar terms, TD averages 7.8 cents of price improvement, again close to the maximum amount of 8.4 cents. In contrast, IBKR Pro has price improvement on 76% of its trades, and average PI% of 18.8%. On a roundtrip trade, this translates into a cost of 62% of the spread. This is over 10 times more than TD. The average price improvement is 2.8 cents, without accounting for commissions.

In between these two brokers, we find that Fidelity and E*Trade have similar execution quality.35 Robinhood is farther behind. Over the entire period, Robinhood has on average 26.8% price improvement, which translates into a roundtrip cost of 46% of the spread. The results are similar across the two periods. IBKR Lite, however, shows slightly worse price execution. Panel B shows similar results, using only paired trades. Given the very large number of observations, the differences are highly significant, both statistically and economically. Overall, we that find order execution varies significantly across brokers.

The authors traded their own money and kept trade sizes small to save money. Would the results be different for large trades?

The target size for the order is $100. We trade full shares only, rounding the number to make the trade size closest to $100, with a minimum size of one share for higher priced stocks. When the experiment started, we also traded 26 of our stocks each day with a target value of $1,000, again using the same logic to round to the nearest share; these stocks were rotated each day. For the 26 stocks with both $100 and $1000 targets, the two trades were placed at the same time with the order randomized. We found similar results across both trade sizes. As a result, we later discontinued the $1,000 trades to reduce our transaction costs and commissions (see Appendix A).

Schwab was not included in the study since it is acquiring TD Ameritrade.

 

[mikeriley]

The authors traded their own money and kept trade sizes small to save money. Would the results be different for large trades?

Unless a trader has similar trading goals, risk management strategies, mindset, experience and a host of other trading specific dynamics it would be veritably impossible to replicate the authors results.

A better way to make that determination is test your strategy in sim mode. If your time table is shorter then just backtest.

 

[deaddog]

I am surprised by the high reported costs for Interactive Brokers and am a bit skeptical, since many professional traders use IB, and they are sensitive to t-costs.

I'm not sure that they are.
From time to time I have put money with so called professional money/portfolio managers. In most cases they were paying higher commissions than retail traders. When I inquird about this they claimed they had better access to the trading desk and got better fills.

 

[DaveV]

The authors are probably using IB's Fixed commission schedule which is significantly more expensive than the Tiered schedule that many professionals use.

 

[ajensen]

The authors are probably using IB's Fixed commission schedule which is significantly more expensive than the Tiered schedule that many professionals use.

The study looks at slippage based on trade prices, not commissions.

 

[KCalhoun]

Thx, skimming through it looks to be very well written and researched.

 

[ValeryN]

The 'Actual Retail Price' of Equity Trades is a recent working paper, covered by Bloomberg.





According to Table VI, the average roundtrip cost for all trades, in basis points, was

7.2 TD Ameritrade
19.7 E*Trade
23.4 Fidelity
31.4 Robinhood
44.4 IBKR PRO
46.2 IBKR LITE
61.9 NBBO (Worst Possible)

I am surprised by the high reported costs for Interactive Brokers and am a bit skeptical, since many professional traders use IB, and they are sensitive to t-costs.



The authors traded their own money and kept trade sizes small to save money. Would the results be different for large trades?



Schwab was not included in the study since it is acquiring TD Ameritrade.

Thanks for sharing. It’s rare so see those kind of comparison on large enough number of orders.

Noteworthy - average IB customer account is 220k, Robinhood 3k and TD 13k. Last time I looked at brokers official compliance reports, nearly all Robinhood customers used marketable orders and nearly none at IB. Point being - vast majority of their customers are not using market orders and theirs fill are much more complicated due to order or magnitude bigger average accounts. So how well 100$ orders gets executed is fairly irrelevant for that customer base.

Plus there is so much more to a broker. Including FX rates, interest on idle account, margin rates, borrow availability etc.

 

[zdreg]

I doubt costs are 6Xat IB compared to TD ameritrade.

 

[KCalhoun]

Fidelity came out well, I like them

 

[ETJ]

The actual report pretty much comes back with some really surprising data. You can draw your own conclusions.