Frequent batch auctions clear every 100 milliseconds at a uniform price instead of trading continuously. Explain what that does to the speed race, who pays for that race today, and what the design costs a maker who quotes an ETF against a future.

Frequent batch auctions clear every 100 milliseconds at a uniform price instead of trading continuously. Explain what that does to the speed race, who pays for that race today, and what the design costs a maker who quotes an ETF against a future.

Approach: Ask what discrete uniform-price clearing does to the value of being first, trace the sniping rent through into the quoted spread, then name what continuous trading provides that a batch does not.

A uniform price batch changes the race from being first to being inside the interval, which removes the rent from picking off stale quotes and lets makers quote tighter. In continuous trading a quote can be sniped after public information moves and before the maker cancels, so the sniping loss is priced into the quoted spread and is paid by every liquidity taker, including those with no speed at all. Batching makes every order in the interval compete on price, so a stale quote is filled at the new clearing price and the rent disappears. The costs are real. A 100 millisecond batch delays hedging in a related market, so a maker quoting an ETF against futures carries 100 milliseconds of unhedged risk and widens to cover it, and venues on different batch clocks recreate the race at the boundaries between them.

Follow-up: What batch length minimises the sum of sniping cost and hedging risk for a maker quoting an ETF against a future?

Key concepts: batch auction, sniping, uniform price, quoted spread, hedging.