Monday, October 20, 2008

Libor fixes: credit and phase transitions

Libor fixes: credit and phase transitions

From the BBA:

LIBOR OVERNIGHT STERLING RATES FIX AT 4.76875% VS 4.68750%

LIBOR OVERNIGHT DOLLAR RATES FIX AT 1.51250% VS 1.66875%

LIBOR OVERNIGHT EURO RATES FIX AT 3.57250% VS 3.66125%

LIBOR THREE-MONTH STERLING RATES FIX AT 6.11625% VS 6.16000%

LIBOR THREE-MONTH DOLLAR RATES FIX AT 4.05875% VS 4.41875%

LIBOR THREE-MONTH EURO RATES FIX AT 4.98625% VS 5.02000%

So far no huge crashing to earth in Libor rates. There are some big movements - the reduction in three month dollar funding rates, for example, is certainly a large move (undoubtedly a result of JPM’s multi-billion interbank three months ops on Friday).

But relative to the velocity of the upward moves in Libor several weeks ago now, these are incremental baby-steps.

Indeed, what happened to Libor - or more broadly, perceptions of market risk in September (measured by the Libor-OIS spead) looked like something of a crude phase transition:

Libor

That is to say, there might be something more to the over-used “frozen credit markets” metaphor than at first meets the eye. Here’s a crude graph of how water freezes:

Phase transition graph

The dotted green line is ice. In its pure form, water makes an instantaneous transition from its liquid to its frozen state: a phase transition. More usually a seed crystal - a small piece of solid-form material (ice) develops around an impurity and tips the balance: as if by magic, the inter-molecular structure of the liquid reorganises itself to match that of the single seed crystal’s “solid” state.

Back to the market: might LEH have been something of a seed crystal to a market already cooled to a critical point?

And to flip that around, are we now seeing that suddenly frozen market only slowly melt?

More broadly - to add to this wild series of flailing stabs in the dark - how might describing markets in terms of thermodynamic physics work in terms of characterising “Black Swan” events as phase transitions?

No comments: