Markets & Trading · Module 2 of 4
The Inventory Skew
Keystone's markets desk just got handed a $200M block it can't recycle: long 20 units of a benchmark note, no buyers behind it, a directional bet the desk never chose. Decide how far to work the book down, from long toward flat, watching three things: the inventory risk you still carry (quadratic: halve the book, quarter the danger), the concession you pay to get out (convex: you walk the book, so the last units are the dearest), and the total cost that is the only score. Where is the right residual: all of it, none of it, or something in between? The ways to be wrong lie on both sides, and the quietest error is doing nothing: hold and hope, because 'it'll come back' pays the whole quadratic charge. Is getting perfectly flat any bargain? Then the danger changes: a volatility spike raises the risk charge. Does the desk that sheds more, or the one that freezes, come out ahead? The second module of Markets & Trading. Shed the risk you can't recycle, but stop before getting out costs more than staying in; score total cost, not how flat the book is.