Replacing continuous time trading with frequent batch auctions mitigates the advantage that latency arbitrageurs, such as high-frequency traders, enjoy substantially. Discrete time trading reduces the number of races to the top of the order book, as explained before. It does not altogether reduce the advantage latency arbitrageurs have though.
With frequent batch auctions, secret bids and offers are submitted in the intervals between auctions. During these periods new price-sensitive information could emerge.
We consider here two categories of price-sensitive information pertaining to a security.
- Conventional public information. This could be company-specific announcements, for example, the release of financial results or trading updates. Alternatively, it could be general news events that affect the company’s valuation, for instance, a natural disaster like an earthquake.
- Information obtained from other security exchanges. This would be relevant in instances where:
- The same security, or versions of the security such as ADRs, trade on another exchange.
- Financial derivatives linked to the security trade on another exchange. Examples of these are futures, options or even exchange-traded funds where their prices are in some way dependent on the security.
- Related securities trade on another exchange. Examples of these are commodities or currencies whose prices affect the security’s valuation.
Relevant information about securities trading on other exchanges is, typically, the last traded price, volumes traded, bid and offer prices and sizes.
New information that impacts a security’s fair value materialises randomly. Sometimes it could arrive towards the end of the period during which the secret bids and offers are submitted for the next batch auction. In this split second, latency arbitrageurs have an advantage relative to other market participants.
For example, let us consider how frequent batch auctions would work in the shares of an oil company, say BP. Now if the oil price changes significantly on a commodities exchange this would impact the valuation of BP shares. If this change in oil price occurs in the instant before a batch auction in BP shares occurs, latency arbitrageurs will have the means to obtain the new oil price, recalculate the new fair value of BP shares and then submit their bids or offers to the BP share auction.
So far we have only considered how to curtail the advantage of latency arbitrageurs on a single exchange. Obviously, the more exchanges that run frequent batch auctions, the fewer races there are available for latency arbitrageurs in which they can compete. The number of races can be reduced even further by having the frequent batch auctions, across exchanges, occurring at exactly the same time. Given today’s; technology this would surely be possible. Refer to the blog “Are frequent batch auctions a solution to HFT latency arbitrage?“ by Rosov (2014).
Regarding our BP example above, with synchronised batch auctions across exchanges, all market participants will have the same oil price, as obtained from the commodities exchange, when they submit their bids and offers for BP. The new fair prices of oil and BP shares are determined at the same moment in their respective batch auctions.
Even though synchronised batch auctions across exchanges diminish latency arbitrageurs’ advantage, they are still able to transact on all public information pertaining to a security. For example, if there is an oil spill that involves BP and this information is disseminated in a split second before the close of the submission of bids and offers for the next BP share auction, latency arbitrageurs would have an advantage.
Latency arbitrageurs are therefore here to stay.