Distributed Settlement: Distributed Ledgers for Real-Time Bank Settlements

1. This post is meant to manage people's expectations of the timeline with which banks may roll out Real Time Gross Settlement using Distributed Ledger Technology (DLT).
2. Real Time Gross Settlement (RTGS) means that any payment instructions sent to a computer system is settled immediately. Payments put through RTGS process are final and considered to be risk free. 
3. Netting out is different than RTGS in that it batches settlement together into time periods rather than allowing for immediate payment. In Interbank Lending this generally happens at the end of each business day.
4. Distributed Ledger Technology can make settlement 365/24/7 whereas the legacy infrastructure of banks can not.
5. The way the current system is set up for banks, netting is preferable to RTGS. Time to settle is certainly important but not more important that system stability.
6.  In the new world of Distributed Ledger Technology (DTL), netting out will still occur, at least for a while.  This has been stated by some of the leading companies in the space.
7. Banks are extremely concerned about real-time risks and preventing another financial crisis.
8. RTGS has certain perils that distributed ledgers may exacerbate particularly when it comes to implementing two new things (RTGS & Distributed Ledger Technology) at once in a highly complex system (which financial markets are). 
9. Strategies need to be thought out regarding how distributed ledgers handle RTGS in times of systemic shocks. 

But just because you can does it mean you should?

Distributed Ledger Technology can hasten the advent of non-stop real-time interbank settlements.  Is this a good thing?  In a world without fractional reserve banking the answer is definitely. Where Assets and Liabilities match up 1:1, settling in real time is a perfect solution and would provide for the most efficient uses of capital in history. Even in good economic times, RTGS makes perfect sense. However, it's not the good times that one should be worried about, it's the bad times.  Banks operate in a world, where they loan out and put to use in varying functions, more money than they have. The average leverage used is around 30-1, during the financial crisis it was as high as 80-1.  This is done based on banks creating money supply from activities such as demand deposits, loans and interest on the loans to be paid back at some time in the future.  Banks use RTGS for high value transactions that need immediate settlement and are done with a central bank.

Most regular payments are still netted out or done in a batch processing type of way.  These fall under the function of interbank settlement using central bank reserves (Bank holdings of deposits in an account with the central bank) to operate in a closed loop. In theory, this makes it "impossible" for money to leave the loop. At the end of the day when they want to settle up, some banks pay money to other banks and some banks receive money from other banks.  Generally, this results in only a tiny fraction of the total amount of money loaned by these banks to be paid to each other at the end of the day, since they net out, based on money being circulated in this closed loop system.  Sometimes, one bank may not have enough money to make its payments. Since it's a closed loop some other bank will have more money. So that bank has more central bank reserves than it needs and will lend that to the bank that doesn't have enough and so on and so forth in a virtuous cycle. This is essentially the interbank lending market and allows banks to make new loans to new customers even if they don't have the reserves.  This is the because the bank that is short knows that at the end of the day when all the payments are netted out, another bank will be willing to lend it some reserves to settle its own payments.  This works really well most of the time and allows banks to perform vital economic functions. But then a crisis flares up.

Why?

Because this only works as long as banks are willing to lend money to each other. If Bank A thinks Bank B won't be able to pay it back it won't lend.  If banks just hold their reserves and refuse to lend them out to other banks the whole system comes to a halt. This will result in some banks failing because they are insolvent (have more liabilities than assets). This could lead to systemic failure as it did in 2008 and the central bank had to step in and print money to give to all the banks to keep them from defaulting as the banks assets starting shrinking rapidly. (In a nutshell, that's what Quantitative Easing (QE) is) In a real-time settlement system with leverage of up to 30:1 can you imagine the damage that can be inflicted without the chance to stop it? Real-time margin calls, real-time defaults and real-time systemic failure.  There is no way to reverse out negative numbers as most of the banks capital and reserves are other people's money lent out. RTGS is meant to stop credit risk in the system and instead replaces it with liquidity risk which rears its ugly head in times when the financial system suffers from extreme stress.

Time isn't just a feature that you upgrade, it's a necessary component for proper risk management in order to find solutions when extreme amounts of liquidity get unwound and credit markets tighten up.  Redundancies need to be put in place, orderly ways to sell (who is the market maker in this scenario) and orderly ways to go bankrupt and default. Even with ATM’s, when there is a queue of people clamoring to get their money out there is some order, when the money stops coming out, those at the back of the line, don’t get any.   However, ‘real-time' does not mean that everyone will now be accessing the ATM at the same time.  It means that the queue to the ATM machine becomes configured differently, such as there being more than one queue and/or more than one ATM machine.  

What happens when everyone wants their money in real time at the same time or worse yet institutions selling en masse?  Will the new configurations be able to handle and route this complexity? From the perspective of the treasury function of the bank there is no way a risk officer can analyze real time shocks and respond quick enough to “adjust” them.  With netting, if one finds themselves about to get a margin call, one can work try to work their way out of it before the next netting period and an irreparable fire sale happens.  One would need a real time dashboard with algorithms to adjust.  However, as we have found out in the past, no one is ever ready for the next crisis and risk is never measured properly for “20 sigma” (standard deviation) events.  It is highly unlikely within the banks themselves that they want real-time settlement right now.  Tightly coupled systems correlate to 1 in times of stress, meaning herd mentality kicks in.  

"The deposit channel depends on the liquidity illusion- an illusion that can survive only so long as not many people take advantage of it."

Banking had become increasingly complex leading into the financial crisis and while some new rules have been put into place (Basel III and the Volcker Rule as part of Dodd-Frank) the complexity remains.  Implementing RTGS and distributed ledger technology needs to be done slowly and tested vigorously to make sure that there will not be any real time disasters which bring the entire global financial system to the point of no return.  John Kay in his book Other People's Money writes:

"Engineers responsible for interactively complex systems have learned that stability and resilience requires conscious and systematic simplification; modularity, which enable failures to be contained; and redundancy, which allows failed elements to be bypassed. None of these features- simplification, modularity, redundancy- characterized the financial system as it had developed in 2008. On the contrary, Financialization had greatly increased complexity, interaction and interdependence. Redundancy as for example, in holding capital above the regulatory minimum-was everywhere regarded as an indicator of efficiency, not of strength."

Any solution that replaces the legacy infrastructure of a bank and tries to reduce the time of payments and settlements, thereby reduces the time available to handle a crisis.  Another lesson he mentions is "on the one hand, eschew unnecessary complexity and, on the other, to pay close attention to the management of unavoidable complexity."

Any well designed system should have as features: robustness and resilience. In an institution like a bank that is looking to deploy RTGS and distributed ledger technology these types of components could be capital and liquidity as these are needed to bolster a bank. Kay goes on to say:

"The levels of capital and liquidity envisaged are inadequate relative to the scale of resources required to protect financial institutions against panics such as a global financial crisis. More significantly, resilience of individual components is not always necessary, and never sufficient, to achieve system stability. Failures in complex systems are inevitable, and no one can ever be confident of anticipating the full variety of interactions that will be involved."

This is a theory known as Normal Accidents and was formulated by Charles Perrow. 

The point being that replacing legacy infrastructure with new technology that change how settlement and clearing is done is an inherently complex process (where everything depends on everything else and is what is known as tightly coupled.  This means the tolerance for error is low, particularly in times of extreme financial stress.  The goal for the finance industry with the help of distributed ledger technology, should be to reduce complexity, lower costs, and enhance stability. So while real time gross settlement is something that is desirable it must be implemented with caution to make sure the above goals are met.

Finally, there is an entire mathematical subject on queueing theory. Communication systems apply this theory to moving packets of data in real-time around global computer networks. The end user may think that the data they are receiving is instant, but in fact there are many distributed queues that the data must transverse in order to be received by the user. These queues are regulated and well managed by relays and routers that have been programmed to ensure that the system is not overloaded.  One would suspect this will play a part in the design of any ledger system.