FX Brokerages

Interesting story about some of these brokers which went to the wall because of the Franc move on Thursday. What is ironic about this is that this may have killed them precisely because of the way these companies made money.

I caveat this is based on admittedly shaky knowledge of how these retail FX brokerage businesses worked and should be considered entirely based on hearsay and deduction, but the model as understood, was to offer huge leverage to clients, wipe their 'equity' out on some small volume based tick in the FX markets (while encouraging them to take long term positions based on research discussing 'exports' and 'non-farm payrolls'), take over their account, close it out with no real loss to the FX brokerage and pocket all the proceeds without real risk to the FX brokerage. This was possible because the 'spreads' (ie price between buy and sell) for these leveraged retail clients offered by the FX brokerages was far wider (and thus liable to their accounts being closed out) than the 'spreads' that the FX brokerages were able to get themselves with their use of the wholesale FX markets. They could take the 'maxed out' account, close it out in the market and pocket the spread.

This is what effectively killed them here, the limits were hit, they took over the accounts, found that even with the crap spreads they were offering their retail clients versus the wholesale market which the firm itself had access to, they still hadn't got anywhere near enough protection against a 25% intra-day move in a currency pair (which had shown low vol for months because of the cap) and then became exposed to the clients losses. Mostly you could go after the clients for their losses, but because of the model explained above, they couldn't because the entire model was built around closing out over-levered clients with no recourse (and taking the difference in the spreads offered when they closed out the account).

They probably had some funky VaR models which they were using to manage their risk in the overall book, because they would have had risk management, but clearly it wasn't working well enough, or it wasn't callibrated to foresee risk scenarios which according to the funky VaR models they used to manage their books, should only have happened once every 5 billion years. Or it wasn't factoring in what the clients were doing.  Maybe this could be an area of focus for the future given the new capital regime which disincentivises any regulated institution from taking any kind of risk on it's own balance sheet...