Interesting comment in FX week (sorry password required) about the problems with some banks FX feeds… “Banks must improve their foreign exchange pricing and trading infrastructure or risk losing order flow, according to currency traders….”

My take on the article, is that some banks are still streaming the illusion of ‘robust’ deep liquidity to clients, when in reality, the liquidity is not deep, and will disappear after the first client ‘hits them’, additional clients are then unlikely to be filled, and due to slippage (the price has now moved) the trade is rejected – this does not make for a happy client!

This phenomenon is appropriately known as the ‘liquidity mirage’ – from a distance (clients desktop) the liquidity looks great, but once you get up close (the banks matching engine), it’s gone, actually like a mirage, it was probably never really there!

The key factors that tend to contribute to the liquidity mirage are:

• Latency induced by inefficient web streaming engines, that are unable to cope with the volume and frequency of price updates, and/or the number of clients, and in fast volatile markets, this results in prices being old and ‘stale’ by the time they are streamed to clients.

• Banks inability to adequately aggregate their own available liquidity to produce a robust and reliable ‘core internal rates’, to which appropriate client spreads added, and firm amounts are then streamed uniquely to multiple individual clients.

In a previous blog comment internalising FX flow I touched on how internalising FX flows, can actually help to minimise the risk of a liquidity mirage, by deepening the banks own ‘internal’ liquidity pool, and thus reducing the reliance and dependence on external sources of liquidity (typically the ECNs), which themselves may be subject to similar liquidity mirages.

By internalising FX flows, banks can enable their clients to achieve better quality execution, which will contribute to driving order flow to the bank, not away from the bank.