Search interesting materials

Wednesday, June 20, 2012

How to achieve safety in payments

by Ajay Shah.

The technological opportunity in payments

In the old days, the field of payments was inextricably interlinked with banking. Money was only held in bank accounts; the only way to move money around was through banks.

Advances in computer technology coupled with financial innovation have changed all this. Banks are no longer the only game in town for the business of holding money. An array of innovators are now in the payments game. A few interesting examples are:
  • Paypal is a pure-play Internet company, which rides on top of bank accounts, and gives users a payments solution.
  • Western Union moves money from person to person across the globe without referencing a bank.
  • M-pesa, in Kenya, does payments over mobile phones. Money is fed into a phone as with topping up a pre-paid card. Money is then transferred to another person using an SMS.
These developments have far reaching ramifications. We can now think of payments as a distinct industry, one that is not joined at the hip with banking. Banking is primarily a risk management business, of coping with callable deposits which have an assured rate of return, even though the assets are opaque and risky. In contrast, payments is primarily a computer technology business, closer to the working of a depository or an exchange.

As Merton Miller said, banking is a disaster-prone 19th century industry. If a critical function like payments can be increasingly decoupled from banking, it would make the world safer.

There are two distinct problems in payments. The first is the systemically important payment system which is the core utility of the currency. In India, it is the RTGS. This is an entirely separate issue. The present discussion  is about the second component of the field of payments: the non-systemically important payments systems which are used by households and firms. This is an ordinary financial technology business.

The problem

When person X wishes to transfer Rs.100 to person Y, if the banking channel is used, the steps are as follows:
  1. Person X lends this money to the bank by putting it into a deposit account.
  2. He instructs the bank to send this to person Y.
  3. At the other end, it shows up as a demandable loan from person Y to the bank.
The balance sheet of the bank is inextricably tied into the payments transaction. Through this, all the problems of banking flow into the field of payments. Banks have opaque assets with 20x leverage or worse. It seems odd to place a mission-critical function such as payments in the hands of such entities. 

In the old world, it was not possible to enjoy the benefits of payments without suffering the credit risk of a bank. One solution that was mooted was for payments to get done in central bank funds. You can do this for a few special situations like the securities clearing corporation, but probably not for most other situations.

The same problem arises with a mobile phone company:
  1. When you feed money, by topping up a pre-paid account, this goes into the balance sheet of the mobile phone company.
  2. Now you have to hope that when the time comes for you to spend this money, the mobile phone company is still solvent.
Faced with this situation, conservative financial regulators have proposed a few solutions:
  1. Mobile phone companies cannot do payments; payments is the exclusive preserve of banks. (This is the state of affairs in India).
  2. Mobile phone companies must become limited purpose banks.
  3. Mobile phone companies must come under full banking regulation.
All these three solutions are unsatisfactory, because they are rooted in the old paradigm, where payments was inextricably intertwined with banking, and it was felt that this is the only way it could be. We need to look beyond this.

An alternative solution: Segregation of client funds

A remarkably clean solution has been invented in the field of asset management: Segregation of client funds.

Consider a money manager such as an asset management company (AMC). At a legal level, the AMC is a mere advisor. Client money never goes onto the AMC balance sheet. Customer money sits completely separate. If the AMC goes bust, this has zero implications for clients. In the entire history of such arrangements, there has been only one episode (MF Global) where segregation of client funds did not work, in protecting customer moneys. This is in contrast with the history of banking, where failures have been taking place across the centuries, across all countries, with a high frequency.

Under such an arrangement, client funds would always sit separately, segregated from the balance sheet of the payments provider.

Segregation of client funds requires a corresponding supervisory capacity - and MF Global shows us that this supervision can possibly fail. But it would involve a much lower failure rate when compared with the problems of banking.

Implication 1: Mobile phone company as payments provider

Suppose Vodafone is my mobile phone company. When I supply Rs.1000 into my mobile wallet, this would go sit separately in a customer trust. This would not go into the balance sheet of Vodafone. If Vodafone were to go bust, this money would be returned to me. This solves the problem of the credit risk of the payments provider.

If we could do this, it would open up an array of payments innovations. The only regulatory burden placed upon the provider would be: Never ever keep customer money on your own balance sheet. We would then need some small resolution capability to kick in when the payments firm goes bust, to take money out of the customer trust and give it back to the customer.

Implication 2: This can be done with banks also

Bank accounts can be broken up into two kinds: illiquid and liquid. (From a customer perspective, this is analogous to the Tier 1 and Tier 2 of the New Pension System; the former is illiquid and the latter is demandable). Illiquid accounts would be loans from customers to the bank (as all bank deposits today are) and have greater restrictions against convertibility. Liquid accounts would not belong to the bank. They would be segregated client funds, used for payments activities.

This would derisk customers from the problems associated with bank failure. It would greatly reduce the complexities of banking regulation and supervision. It would put banks on a level playing field when compared with other technological strategies in the field of payments.

When banks do not capture the interest income on the liquid accounts, this will force a healthy unbundling of payments and banking. Banks who engage in the payments business would have to explicitly charge for payments services. This would help ensure a level playing field between bank and non-bank players in payments.

Implication 3: How to store segregated client funds

Payment vendors could place client funds into current accounts with the central bank for riskless safekeeping. Or, they could place them into NAV-based money market mutual funds, so as to earn some return.

In this framework, there would be N money market mutual fund accounts belonging to M entities. The payments system would be a technologically diverse array of alternative competing mechanisms through which money flows from account i to account j, which generates a fee income for the payments provider.


The idea of segregated client funds, which is very well established in some areas of finance such as money management, brokerage, etc., can be usefully applied in the field of payments, to cut through the gordian knot of banks and payments.


  1. "Mobile phone companies cannot do payments; payments is the exclusive preserve of banks. (This is the state of affairs in India)."

    Really? I thought you could use Airtel Money (quote from their website - "Now send money to friends & family through your mobile. No need to visit the bank, just use your airtel money account to send money to bank accounts or to other airtel money accounts"). How does this work?

  2. This is a good article - had a somewhat unrelated question - would you be so kind as to recommend a good book/URL that describes in detail the various payment processes/their modus of working/frameworks etc?
    E.g curious to understand how ACH in the US works as against Fedwire or NETEFT as against Gross RTS in India or a simple Credit Card transaction or a Pay Pal transaction?

    1. In my knowledge, such documentation does not exist. Please write one!

  3. Hello Ajay:

    I have been following your blog for a while. I found this blog post interesting. I thought I will let you know about a couple of points that you have highlighted regarding the role of the payment intermediaries in acting as guardians of the client funds.

    As per the pre-paid instruments guidelines that have been proposed by RBI, they have classified the payment instruments into three categories - closed, semi-closed and open. Mobile payments are currently classified under the semi-closed category.

    The regulations for semi-closed specify the need for the mobile company to earmark those funds separately and park the money in an escrow account with a bank. There are restrictions on the possible uses of this escrow (i.e., lien marking, interest payouts etc). The good part is that the money is not tied to the balance sheet of the telecom company ... but the problem doesn't get solved because it is linked to the balance sheet of the bank!


  4. Ajay, regarding the second application:

    Isn't deposit insurance a way of achieving the segragation - the only difference being that the bank's client cannot choose a higher amount of "liquid" funds than the maximum that deposit insurance corporation would insure for (USD100K in the US). So the question is which model of segregation is cheaper for the economy. Your discussion suggests that in times of a crisis, a segregated accounts model would mean payments systems wont seize up, and there would not be deposit runs - but there may still be runs on illiquid accounts, with depositors preferring the pay the penalty for early withdrawal rather than lose it all in the event of a failure -the same that happens now with deposits above the insured limit.

    Also, what the segregated model would do is to make the credit growth in the economy depend on the liquidity preference of the depositors, which could potentially contribute to business cycles.
    Something else that comes to my mind is whether the same outcome can be achieved with leverage ratio limits.
    While I like your discussion, maybe you could talk a but more about the the question of which model has what costs?

    1. Dear Gurnain,

      The resolution debate is distinct from how we might choose to repackage banking.

      Under the proposal sketched in the post, there would be a resolution problem associated with payments, but it would involve zero risk since the money is always being kept segregated. That is, there would be "deposit insurance" without limit. The only thing the resolution corporation would have to do is to go pick up the money market mutual fund, find the customer, and give the money to him. It's a pure operations job, there is no financial risk management involved.

      As for the illiquid piece - policy makers would have to think about how they want to approach the deposit insurance question. Maybe the right thing is to have NO deposit insurance there. It's possible to argue this case both ways.

      About costs and benefits, as I see the tradeoff, what is going on under this proposal is that we are blocking the maturity transformation that's done by banks, where non-volatile demand deposits are used to hold long-term assets. In return for this, we get a great deal more systemic stability, and we open up the payments system to non-bank competition.

  5. Hi Ajay,

    Godd post. As long as money is controlled by a central authority the problems you mention would remain. Just decoupling the payment system out of it might not be of much impact.

    What are your view on decentralized currency like Bitcoin? (Bitcoin is also virtual). As long as we are dependent on fiat money we will have to deal with fickleness of the banking system.

    Vikas Desai
    PS: My second attempt at posting this comment!

  6. Some of the most interesting innovation in finance in India right now is in the payments space, and this is an interesting suggestion. Essentially a bank (or telco) account would transform into a money market account. Banks would then raise a lot more funds in money markets, via CDs, than they do now, and they would have to compete for the deposit funds against everyone else in the money market.

    The US experience with breaking the buck in money markets during the crisis shows these are not completely riskless markets, and India has had its own regulatory concerns with practices in the money market. So clearly some additional supervision and perhaps regulation would be needed. How to structure a new form of deposit insurance would be an open question. (As an aside, RBI bank resolution requires liquidation of bank assets, so while depositors are made whole it may take months or longer to get your funds back. Moving to an FDIC-style early resolution system makes most resolutions just a seamless acquisition from the depositor standpoint.)

    Right now, what specific activity requires a bank license? It's the right to take public deposits and put them on your balance sheet, and access to the payments system. In India I don't think there is a regulatory hurdle to non-banks putting funds in segregated accounts now (telcos might have trouble getting the requisite NBFC license, but could partner with one), but the RBI is adamant only traditional bank accounts get access to the payments system.

    Your biggest point is that there's nothing unique about a bank - especially with technological advances - that requires them to have a monopoly on access to the payments system. This is very true, and perhaps a segregated account system would help the RBI get more comfortable with relaxing the rules. Maybe they could start by allowing banks to offer segregated money market accounts with full access to the payments system.

    1. Dear Russell,

      I have routinely encountered the blanket claim: "How can client money sit with Vodafone" as the response when discussions of payment innovations come up; it is intended to be a "checkmate" on all discussions about new thinking in payments. This blog post was intended to say: "This is a very solvable problem".

      RBI's dogged opposition to innovation (whether in payments or in finance more broadly) may be rooted in conflicts of interest, organisational behaviour under low quality HR practices, turf protection, the desire to avoid accountability, etc. These deeper problems also need to be understood and solved.

  7. This could be a digression, but would like to know Ajay's views on placing limits of cash transactions. The virtues of use of cash currency are beyond question particularly in transactions involving small sums of monies. However, it is baffling to see banks, businesses, and government organizations supporting cash transactions involving crores of rupees. In the recent by-polls in the state of AP, the ECI seized about Rs. 50 crores of undocumented cash from individuals of political parties. In the recent bail scandal involving CBI court judge and mining baron G.J. Reddy, investigators seized Rs. 2.87 crores on cash.

    The point is that most of this money supporting illegal activities must have at one point come out from banking facilities in the form of big withdrawals. In the U.S., I cannot imagine going to my bank and withdrawing more than $5,000 in cash without inviting the attention of law enforcement. And I believe $10,000 is the most I can withdraw in a given day or week. In contrast, I find it amusing that banks in India dispensing crores of cash to individuals and are guided by no upper limits on such cash withdrawals or deposits. Also, most real estate transactions utilize predominantly use cash. The use of cash despite the obvious inconvenience of handling and counting only strengthens the fact that cash as a mode of payments is being misused and taken advantage of by hiding the identity and origination.

    1. I agree with you. It gets worse:

      Electronic movements of money always go with log files. When an investigation takes place, you can see every element of the movement of the money. Hence, law enforcement should be very happy when more is doing electronically.

      In India, we do this upside down: RBI blocks the rupee value of what can be done through electronic payment mechanisms, while cash flourishes unregulated. Greater analysis and thought is called for, on the part of RBI.

  8. One issue that is overlooked in the emerging cash transfer technology is their potential to launder money

  9. If we make accounts segregated into liquid and illiquid; then from the same conventional knowledge, is not bank will have less 'fund' for Credit creation??

  10. Too often, occupational accidents are considered a price of progress in developing countries. but they also occur in advanced nations. Whether we challenge nature at thirteen thousand feet underground in a South African gold mine or thirty-nine thousand feet below sea level exploring for oil in the Gulf of Mexico or at three hundred thousand kilometers in space en route to the moon—accidents can and will happen despite the best safety regulations. Human beings make mistakes, and sometimes equipment fails. Mothers often protest that they were only distracted for an instant when their child disappeared. “Pilot error” is still the leading cause of airplane crashes!

    Industrial, mining and, consumer safety should never be compromised for profits; neither should margins of safety where food and pharmaceutical products are concerned. Aircraft, space vehicle, and automobile manufacturing as well as the production of power tools and equipment are further examples of industrial safety challenges—as are consumer products, from toasters, to contact lenses, to cribs—especially imported cribs. Then there are procedural and processing standards that require constant upgrading in chemical plants and hospitals as technology advances. There are safety concerns regarding the very frontiers of science. These include the safety of our astronauts in space exploration, the safety of people downrange of sites from which military and civilian rockets are launched, and the safety of offshore and land-based drilling for oil and gas, to name but a few.

    You can buy this book now on any of the following websites:

    Strategic Book Publishing Rights Agency:

    Amazon Books:

    Barnes and Noble Books:


Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticism is perfectly okay; uncivilised language is not. We delete any comment which is spam, has personal attacks against anyone, or uses foul language. We delete any comment which does not contribute to the intellectual discussion about the blog article in question.

LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.