How to Avoid Being Eaten by Sharks: High-Speed Trading

Sharks in the Dark: Quantifying HFT Dark Pool Latency Arbitrage

By Matteo Aquilina, Sean Foley, Peter O'Neill and Thomas Ruf

Published By Bank for International Settlements – BIS

A very interesting paper that examines how some traders, use their fast technology to exploit delays in the prices of stocks in dark pools, and what can be done to prevent this. Let's break it down.

The paper is about how some traders use high-speed computers and networks to take advantage of delays in the prices of stocks in dark pools.

But what is Dark Pool?

Dark pools are places where people can trade stocks without revealing their orders to the public.

It is a private financial forum or exchange for trading securities. It is called a “dark” pool because the details of the trades are not visible to the public until after the trade has been completed. Dark pools allow investors to buy and sell large blocks of shares without revealing their intentions to the wider market, which can help them get better prices for their trades.

Here’s an example to help you understand how dark pools work: Imagine you are an investor who wants to buy a large number of shares in a company. If you place your order on a public exchange, everyone can see it, and this can cause the price of the shares to go up before you have a chance to buy them. This is because other investors may think that if you are buying, then the stock must be a good investment, so they start buying too, which drives up the price. But if you place your order in a dark pool, no one can see it until after the trade has been completed. This means that you can buy the shares at a lower price because other investors don’t know about your order and don’t start buying the stock before you do.

They are supposed to offer better prices and lower costs for traders, but they also have some risks.

But what are these risks and how do they affect other traders?

One of the risks is that the prices in dark pools are based on the prices in other markets, which are constantly changing. Sometimes, there is a lag between the time a price changes in one market and the time it is updated in a dark pool. This creates an opportunity for some traders, called high-frequency traders (HFTs), to use their fast technology to see the new price before others and trade at the old price in the dark pool. This is called latency arbitrage, and it means that HFTs can make profits at the expense of other traders who are slower or less informed.

But how often does this happen and how much does it cost other traders?

The paper uses data from the UK to measure how often this happens and how much it costs the other traders. The paper finds that about 4% of all trades in dark pools are at stale prices, meaning that they are not the best prices available at that time. The paper also estimates that this costs the other traders about 2.4 basis points per trade, which adds up to about 4.2 million pounds per year for all UK dark pools. The paper also shows that HFTs are almost always on the winning side of these trades and that they rarely provide liquidity (i.e., offer to buy or sell) in dark pools. Instead, they mostly take liquidity (i.e., accept existing offers) when they see a chance to exploit a stale price.

But what can be done to reduce latency arbitrage and improve liquidity in dark pools?

The paper also examines two possible solutions to reduce latency arbitrage in dark pools: randomizing the time of execution and using frequent auctions. Both of these methods aim to make it harder for HFTs to predict when and at what price a trade will happen in a dark pool, and thus reduce their speed advantage. The paper finds that both methods are effective in lowering the proportion of stale trades and improving liquidity in dark pools.

Still not sure? Let's use an imaginative example.

To illustrate this topic with an imaginative example, imagine that you want to buy a book from an online bookstore that has a dark pool feature. This means that you can place an order for a book without revealing it to other buyers or sellers, and the bookstore will match you with someone who has the opposite order (i.e., wants to sell the same book). The price of the book is based on the average price of other online bookstores, which changes frequently depending on supply and demand.

Now suppose that there is another buyer who has access to a very fast computer and internet connection, and can see the prices of other bookstores faster than you and the bookstore. This buyer can use this information to place orders in the dark pool before the price is updated and get a better deal than you. For example, if the book’s price drops from $10 to $9 in another bookstore, this buyer can quickly buy it from the dark pool for $9.50, while you are still willing to pay $10. This means that you end up paying more than you should, and the buyer makes a profit of $0.50 from each trade.

This is unfair and inefficient because it reduces your incentive to trade in the dark pool and makes you pay higher costs. To prevent this, the bookstore could use one of the two methods suggested by the paper: either randomize the time when it matches orders in the dark pool, so that the fast buyer cannot anticipate when a trade will happen; or use frequent auctions, where it collects all orders in a short period of time and executes them at a single price, so that the fast buyer cannot exploit price changes during that period.

Hope you enjoy this and learned something new.

Want to be part of a community of informed and engaged individuals?

Subscribe on LinkedIn

Disclaimer: The views expressed in this blog are not necessarily those of the blog writer and his affiliations and are for informational purposes only.

#darkpools #HFT #latencyarbitrage #trading #finance #stockmarket #stockmarkets #sharemarket #share