According to the 2023 Nasdaq Technology White Paper, the institutional-level broker trade system that is connected directly via API can operate at an order execution speed of 1.2 milliseconds, while retail platforms used by retail traders have a latency of 47-300 milliseconds on average. Let’s take the case of Interactive Brokers’ PATS system. Its S&P 500 constituent stock market order execution rate is 99.8%, and 98.5% of orders are executed in 90 milliseconds. However, network latency becomes a key determinant – physical optical cable transmission between London and New York is 76 milliseconds, so the overall delay in cross-border transactions is 150-250 milliseconds. High-frequency Trading companies such as Jump Trading have used microwave communication technology to reduce the data transmission time between the Chicago Mercantile Exchange and the New York data center to 8.3 milliseconds, 42% faster than regular optical fibers. That being said, retail traders should pay attention to the difference between order types: on average, it takes 1.4 seconds for limit orders to be filled in Futu Securities’ liquidity pool, versus 0.7 seconds for market orders, though the latter may incur a slippage cost of 0.3% to 0.8%.
Volatility in the market significantly influences the execution speed. During the four circuit breakers in the US stock market in March 2020, the longest delay by TD Ameritrade in executing market orders was 3.2 seconds, 650% above the usual time, which caused the disparity between the quotations and trading prices of the S&P 500 index futures to reach 1.8%. In its 2022 Execution Quality Report, the SEC discovered Robinhood’s order routing strategy permitted only 36% of customers to receive NBBO (Nationwide Best Buy and Sell Quotes), as contrasted to Interactive Brokers at 89%. What this would mean is that under the identical conditions, an order of $1,000 may have a price impairment of $2.3. Technical metrics show that when the VIX fear index is over 30, the cancellation rate of the New York Stock Exchange orders soared from the normal 40% to 72%, and the liquidity decline widened the bid-ask spread of the mini E-mini S&P 500 futures contract to 0.5 points (the equivalent of a $25 per lot expense).
Infrastructure upgrades are changing the speed pattern. The quantum communication trial network of the London Stock Exchange, which was mounted in 2023, reduced the end-to-end encryption for trading orders from 3 milliseconds to 0.9 milliseconds. After upgrading Coinbase’s distributed matching engine in 2022, its maximum order processing quantity went up from 2 million per second to 6 million, and 99.9% of the market orders are now matched within 2 milliseconds. However, regional differences remain significant: The execution speed for Singaporean investors to trade Hong Kong stocks through Interactive Brokers is 83 milliseconds, while for the Hong Kong local broker Yiu Choi Securities, it is only 9 milliseconds. Especially, the 5G network has reduced mobile broker trade latency to 19 milliseconds (compared to 65 milliseconds of 4G), enabling the proportion of mobile orders of Kakao Pay securities in South Korea to rise from 58% in 2021 to 81% in 2023.
Government policies affect the effectiveness of implementation directly. MiFID II forces European brokers to disclose details on the quality of order execution, which has pushed market makers such as XTX Markets to narrow the spread between the US dollar and the euro to 0.6 micropoints (0.00006). However, the US SEC’s rule 606 discloses that only 54% of Q1 2023 retail orders received price improvement. The price improvement per share averaged $0.0012. Disaster recovery technical capabilities are of paramount importance under extreme market conditions: During the GameStop short-selling debacle in Jan 2021, Robin Hood’s order processing infrastructure was swamped with 400% of its capacity, causing a 78-minute service outage. As a consequence, 34% of the orders caused users to lose more than 20% of their potential gains. The development of algorithmic trading unleashes new possibility – jpmorgan Chase’s LOXM intelligent router has reduced the expense of block-trading execution by 18% with the help of machine learning and trimmed the time to execute from the usual 45 minutes of human negotiation to 3.2 seconds.