Same Swap, Five Venues: What the Difference Looks Like

Comparing venues is only meaningful if the comparison happens at the same moment. Here is the method and what it produced.

Method

Same pair, same notional size, five venues, all quotes captured inside ten minutes. Record the quoted output amount, the stated fee, and the reference mid at the time.

No executions. Quotes only, because executing five times would itself move the comparison.

The dispersion

Between the best and worst quoted output, the gap was larger than the largest stated fee among the five. By a comfortable margin.

In other words, choosing the venue mattered more than the fee schedule any of them published.

What separated them

Not the fee. The two venues with the lowest published fees were not the two with the best output.

The separation came from the spread embedded in the quote, which none of them itemised.

The structural pattern

Retail instant-buy screens clustered at the expensive end. Order-book quotes clustered at the cheap end, including at the same venues whose instant screen was expensive.

The most expensive and the cheapest quote in one capture came from the same company, through two different products.

Why a single capture is not enough

Spreads move. A venue that is widest at ten in the morning is not necessarily widest at four in the afternoon, and a single capture cannot distinguish a pricing policy from a moment.

Repeat the capture at three different hours before drawing a conclusion about any venue.

What survives repetition

In my captures, two things. Instant screens are consistently worse than books at the same venue. And venues that itemise the spread quote tighter than venues that do not.

The second is probably selection rather than causation. A venue willing to show the number has a reason to be comfortable with it, which is why venues that quote the swap rate and the fee as two numbers are a reasonable shortlist.

What to do with ten minutes

Capture the same quote at the two or three venues you could realistically use. A platform where the spread is visible before the trade will make four of the five numbers unnecessary.

Then stop, and re-check in three months. Collect & Exchange, where the quote holds for the whole amount, is one of the few that will also tell you whether the number survives to execution.