The strange lack of technological disruption in FX

Foreign exchange (FX) is routinely described as the largest and most liquid financial market in the world. And the most recent triennial survey of FX by the Bank for International Settlements (BIS) reckoned daily turnover hit US$6.6 trillion in April 2019, up from $5.1 trillion in 2016.

The liquidity of the market, in particular currency pairs at particular times, is a lot less reliable than such a scale implies. And, despite the hordes of participants in every currency and country and the shift to from inter-bank trading to on-exchange trading, the FX markets are in reality controlled by an astonishingly small number of banks.

FX is ultimately owned by less than ten banks

Trading platforms intermediate about two fifths of all FX flows, but the banks remain the primary sources of their liquidity. The remaining three fifths of flows are still intermediated directly by the banks, but in an increasingly concentrated form.

Most banks are just consumers of FX liquidity supplied by the major banks, simply passing on prices to their underlying buy-side customers. Less than ten banks act as true liquidity providers to the FX markets today, in the sense of acting as market-makers that take principal risk rather than broking client trades.

This gives the biggest banks enormous power over what happens in the FX markets. They control access to the credit lines which enable asset managers to trade FX at all. Real money users of FX services, from large corporations to holidaymakers, are takers of the prices made by banks in every currency pair.

Exchanges have failed to break the power of the FX banks

The obvious way to break this oligopoly is by the development of large and liquid trading platforms. Exchange groups have bought or are trying to buy 360T (Deutsche Börse), FXAll (London Stock Exchange), NEX Group (CME), Hotspot (CBOE) and FastMatch (Euronext).

If exchanges could deliver sufficient savings in margin costs (notably via integrated management of margin across on- and off-exchange derivatives) to disrupt bank credit lines as the key to access to traded FX prices, and replace bank intermediation by central clearing against eligible collateral, they would indeed disrupt the FX markets.

Unfortunately, this is not happening. One reason is that real money users of FX, such as large corporations and institutional investors, use FX not to trade but to hedge currency risk and finance purchases. Banks are much better than exchanges at supplying hedging tools and executing hefty currency trades than exchanges.

As a result, real money users of FX continue to use banks for hedging and currency purchases, even though they are paying heavily for the services. Custodian banks, for example, profit mightily from executing currency hedges and purchases on behalf of buy-side clients, mainly from bid-offer spreads.

Peer-to-peer netting of FX trades could cut buy-side FX costs

It has proved near-impossible for the exchanges to replace the banks as intermediaries in trades of this kind. In theory, institutional investors could trade FX directly with each other. But a double coincidence of wants is unlikely, so netting trades before placing the net amount in the bank-dominated FX markets is a more realistic option.

If it took off, peer-to-peer netting cut the cost of FX for real money users significantly, but if the counterparties must still rely on bank liquidity for the net amounts, the banks remain in a strong position to restrict the growth of netting. It is correspondingly hard to wean to real money users of FX off what they are doing at the moment.

That said, institutional investors have already had some success in pressing custodian banks in particular to shrink the spreads they take – this followed litigation by major funds in the wake of the financial crisis of 2007-08 – and if their asset managers, now also under pressure to cut costs, do the same it could encourage the banks to look at ways of cutting their own costs.

The lavish profits of trading FX have given banks little incentive to contain their costs. But this does not mean there is not ample room to do so, and post-trade processing is the obvious area to look.

Major savings are on offer in FX post-trade

In post-trade FX, a litany of vendors, specialist providers and financial market infrastructures such as CLS and SWIFT compete to message, match, confirm, aggregate, clear, net, credit-check, novate, compress, allocate, reconcile, settle and report FX trades. Duplication and frequent and sequential rounds of reconciliation, legacy technologies and manual processes abound.

The blockchain era highlighted the benefits of all parties to a trade working off a single record of it, and FX is no exception. Estimates suggest it could cut 60-80 per cent of post-trade costs in the FX activities of banks, by eliminating the need to reconcile separate sets of data about the same trade between internal as well as external systems, conducting all post-trade functions off a single set of data and – above all -by allocating credit lines between clients more dynamically as well as more efficiently.

As always in FX, every solution must confront the fact that bank credit underpins the entire market. Which is why it is a mistake to believe that the new class of FX market makers, known collectively as non-bank liquidity providers (NBLPs), are an answer to the dominant position of the banks.

Although NBLPs account for a large share of activity in the spot market in particular, and are enormously profitable, they are far from non-bank in their modus operandi. Indeed, they are entirely dependent on a form of credit relationship with banks known as FX prime brokerage.

Indeed, it is significant that, as the FX prime brokers – all of which are owned by the small class of major FX banks – have trimmed their client lists, reserving credit lines for favoured customers only, it has actually become more difficult for smaller and wannabe NBLPs to get into the FX market or stay in it.

The dominant position of bank-owned FX prime brokers is hard to break

This has created a demand for new approaches that can match the benefits conferred by an FX prime brokerage relationship. Those benefits are threefold.

First, FX prime brokers enable NBLPs to trade anonymously and algorithmically on multiple trading venues off credit lines they advance. In other words, the counterparties of the NBLPs trade not with the NBLPs but with a bank that can be counted on to honour the trades. This greatly reduces their counterparty credit risk.

Secondly, FX prime brokers enable NBLPs to outsource all the operational complexity created by their trading activities across multiple bi-lateral counterparties and FX trading platforms to their FX prime brokers.

Thirdly – and most importantly in terms of enlarging trading capacity – users of FX prime brokers can economise on the margin they post to counterparties and trading platform clearing houses through portfolio margining, margin offsets and cross-margining of all their activities with a bank.

These benefits are not easy to reproduce. But the blockchain era has again spawned a possible solution. The alternative trading and clearing mechanisms developed by participants in the crypto-currency markets had to work out how to manage counterparty credit risk, and the solution their designers found might be robust enough to provide an alternative to FX prime brokerage.

Use of blockchain in crypto-currencies points to an alternative to FX prime brokerage

A private, permissioned blockchain network, in which assets in custody are made available for use as margin in FX trades by tokenization of the assets on to a distributed ledger creates an intriguing possibility. Once tokenized on to the ledger, the assets can be used to collateralize counterparties, eliminating the need for a bank to intermediate the credit risk.

Better still, immediately available collateral means that FX trades can be settled directly between counterparties simultaneously in real-time (what the blockchain industry knows, somewhat misleadingly, as “atomic swaps”).

In theory, the model works because the blockchain ledger cannot operate unless the assets are available in the account. In principle, “atomic swap” settlement of this kind could enable NBLPs and other FX trading firms to trade with any counterparty, anywhere, without the credit intermediation of an FX prime broker.

In practical terms, however, success depends less on the number of counterparties willing to instruct their custodians to tokenize assets to a blockchain than on the number of custodians which are prepared to join the blockchain network, since efficient access to assets held by counterparties is essential. Once again, the banks remain the arbiters of whether the disruption will succeed or not.

A “repo” market for NBLPs could provide another alternative to bank credit

Another approach may prove more effective in the short term. This is to dispense with the custodian altogether. Instead, trading firms can raise cash for use as margin by borrowing it, instantly. Cash-rich corporations (such as the FAANGs) would welcome the higher yield, especially if it also collateralised with securities.

The cost of the borrowing will erode the return on the trade but, from an operational perspective, can be built into the spread. One difficulty is that not all buy-side counterparties will be able to borrow. Another is that collateralisation requires a custodian – once again, in the absence of alternative collateral managers, a need which places the banks in a pivotal role.

Democratisation of price data is likely to be the best disruptor of the FX markets

There is one final possibility that could shake the bank oligopoly in FX. This is the most powerful of assets in digitised markets – namely, data.

The clearest symptom that a coterie of banks control the FX market is that different customers are offered different prices for the same currency pair in the same size at the same time. Banks can do this because not every participant in the FX market has equal access to price information.

This is where price transparency for all participants in the FX markets could make a difference. Access to accurate rates for all currency pairs in the spot, forward, non-deliverable forward and restricted currency markets would enable non-banks to benchmark the FX rates achieved by the banks they use, and so exert pressure to offer keener prices.

But a price service of this kind must avoid taking prices from banks, as opposed to collecting prices at which bargains were struck on trading platforms, because taking bank prices would open the rates to manipulation.

Nevertheless, data is, in the final analysis, the best solvent of unfounded arguments and indefensible practices that serve as cover for vested interests. Rigorous data analysis has already improved transparency and fairness in the equity markets, and it could do the same in FX.

After all, the dominant banks have limited incentives to change the status quo. So the most effective disruptor of the FX markets might turn out to be not start-ups armed with digital technology but digital technology in tandem with data, whose quantity was never greater. The role of digital technology will be simply to cut to nugatory levels the cost of collecting, storing, analysing and publishing data.

Written by Dominic Hobson, Co-Founder Future of Finance

September 2020