July 14th, 2003, 9:04 pm
Well, here's my 0.02$, to get you guys started.I assume everybody is familiar with mark-to-markets of liquid exchange traded instruments (and associated margin calls, in some cases).Of course even these instruments are not always marked to markets.The question is the mark-to-market process of OTC instruments (and illiquid listed ones). Different houses have different ways of doing it. Various methods for vanilla options:-trader's whim: (i.e. the trader marks the options to whatever vol he fancies). highly contentious method... obviously. It has one advantage though: it can reduce the pathdependency of PL (e.g. always marking at same vol, say the vol the option was traded at)-direct quotes from IDB's (inter-dealer brokers): plus: they are often reliable and tradable prices. minus: they can be difficult or impossible to get on some underlyings, they can be biased, and finally can't be the only tool because brokers needs to make a living hence don't like quoting without ever trading (hence don't like quoting to the controllers, e.g. accountants, product control groups). -inter~extra polation of forwards and volsurfaces from Totem. Totem is a company which provides quotes on request (by asking contributor banks)plus: easier and more systematic than IDB.minus: much less reliable. quotes are non-tradable. enough said-mix of the above 3: the trader mantains continuosly a volsurface to the best of his market knowledge using IDB daily quotes, trades and listed option markets (and quote requests), combined with a sensible parametrization of the surface (ideally arbitrage free) and a gentle prod from some kind of independent controller.plus: should be the best approximation to "the market"minus: still open to misuse. require full time traders, with not too many underlyings to follow (how many depends on the business model), and a savvy controller (ideally and ex-trader). the qeustion of what is a sensible parametrization is also debatable and debated.For exotics: just extend/escalate the above. With the massive additional problem of modeling:there is no accepted standard for exotic pricing, and this has huge implication in marking (and pricing, and hedging) some of them, especially barriers (some houses do not take any skew into account), products that depend on forward skew (do you believe in local vol? stoch vol? or maybe constant vol! some houses traded them at flat skew, in the past few years), etc.Mark to market has serious economic impacts, the first ones I can think of are:-obviously, margin calls, either literally (from an exchange) or figurative (from a risk manager, head trader, or similar forcing the closure/reduction of a position). on long dated trades this is unavoidable.-most houses compute their hedge ratios (delta) at the same vol of the mark (there are pros and cons). this means the marking policy will have a serious impact on the hedging strategy, and thus on the pathdependence of the PL. one could think that in some cases, esp. for short dated trades this effect can become dominant, especially if the mark-to-market policy is not well thought out and creates large discontinuoties.In theory, if you had unlimited deep pockets and could take the pain for an unlimited time, you should put on the first prop trade that comes to mind, in as large a size as possible, and earlier or later you'll close it at a profit. Mark to market prevents it: hence we could say that it is a way of defining the scale (and timescale) of the acceptable losses, in other words, it defines the risk appetite. Must be time to go to sleep...