Serving the Quantitative Finance Community

 
User avatar
Traden4Alpha
Topic Author
Posts: 3300
Joined: September 20th, 2002, 8:30 pm

Blockchain ledgers for banks

July 19th, 2016, 1:16 pm

A startup wants banks to use a blockchain system for a transparent ledger. But it raises some interesting questions (beyond the obvious one of whether cloud computing really will be cheaper in the long run).

1. Is transparency "good"?  Yes, transparency can help the bank spot fraud/embezzlement, help tax authorities catch evaders, and help banking regulators monitor risk.  But the data can also be used to estimate and exploit price elasticities  -- knowing Bank X holds to much of asset Y and will be forced to liquidate at a low price if counterparties learn of the likely liquidation.  If one can see the ledger, one can predict the bank's future trades and price accordingly.

2.  How does one contrive the ledger so that it is reality rather than being an estimate of reality.  That is, how can each entry on the ledger be the transaction rather than being a record of the transaction.  One vulnerability in all IT systems is the gap between reality and data (e.g., fraudulent or erroneous entries into an accounting system).

3. And assuming this blockchain stuff is a good thing, I'm wondering if the "proof of work" can be contrived to do something more useful than mine random cryptographic hashes out of a sea of bits.  Might there be complex risk management, real-time balance sheet, or portfolio optimization calculations that can be embedded in the blockchain mining algorithm?
 
User avatar
ChrisGifford
Posts: 36
Joined: February 6th, 2016, 10:27 pm

Re: Blockchain ledgers for banks

July 19th, 2016, 10:59 pm

Hi Traden,

This is an area I've been researching for the last 6 months or so.  Taylor will be in good competitive company.

1. If by "good" you mean the elicitation of a reasonable rate of return for Bank X then, yes, it will depend upon who has access to the 'transparent' data.  In terms of the law, it will take a bit of time to catch up and cyber/digital law will need to be developed suitably to deal with this kind of thing, although `I envision that the non-cyber/digital law will guide most of it.  There have already been some significant fraudulent issues with bitcoin.  If I worked for Bank X then I would only allocate a certain amount of blockchain-monitored assets after doing a suitable risk analysis - the rest would be on the books elsewhere.

2. The boundaries between the record of a transaction and the transaction itself are blurred for cryptocurrency and the blockchain technology (in contrast to fiat money), although it has been claimed by one researcher recently that cryptocurrency, unlike fiat money, is not a unit of account.  Much of the distinction will come down to the design of the digital architecture for each blockchain technology.

3. That is a really nice question and something I've been thinking about with respect to some potential technology similar to the blockchain.  Proof of work is just one method of proof and yes, I think the potential for it extends at least that far.  The ASX is on top of some of the further potential use of blockchain (http://www.coindesk.com/asx-blockchain- ... l-results/).  I see a huge potential for blockchain in the future.
 
User avatar
Traden4Alpha
Topic Author
Posts: 3300
Joined: September 20th, 2002, 8:30 pm

Re: Blockchain ledgers for banks

July 20th, 2016, 2:21 pm

Hi Chris,

Thanks for your insights into this emerging technology.

1. Transparency: Yes, returns are one way to measure "good" and maybe they are the ultimate way to measure it (although there's the quibble about whether it's the customer or the bank or society who should earn a reasonable rate of return). Yet I was more concerned about the balance between beneficial forms of non-transparency (personal privacy, investment in creating proprietary knowledge, protection from kleptocracy, etc.) and detrimental forms of non-transparency (fraud, tax evasion, embezzlement, self-dealing, extortion, etc.). Does transparency throw the baby out with the bathwater or simply create different avenues for criminal action (e.g., identity theft, front-running, government seizure of assets without due process, etc.)? Or does excessive transparency constrain investor action in detrimental ways? For example, might banks become too risk-averse if they know that every individual action they take (e.g., every mortgage) will be scrutinized with hind-sight bias?

2a. Records vs. Reality: What's interesting about a bank's ledger is that it's not strictly the bank's in that so much of it encodes persistent financial relationships between the bank and other entities. And, in theory, if the ledger says the bank owes company X some amount of money on some terms (e.g., a demand deposit, time deposit, etc. ), then the ledger of company X should show the identical item in reciprocal terms. Each entry in the ledger would be an interstitial object connecting the balance sheets of two entities -- belonging to both entities at the same time which seems to be one of the interesting key features of a distributed ledger. And a transaction is then an interaction between two ledger entries -- moving some number of units of account from one entity to another. Of course, if enough ledgers move to the cloud, perhaps the accounts move with it. Might Taylor's company eliminate the need for banks?

2b. Unit of Account: I tend to agree with that researcher that BTC may not be a unit of account but it could be. More BTC users need to live their economic lives using BTC which would make BTC-denominated prices stickier and stabilize BTC FX rates. But, personally, I think Satoshi's algorithm is deeply flawed in the same way that gold standard currencies are flawed -- an inflexible money supply is not a good idea even if it forestalls the evils of central banker manipulations. No doubt there are clever people looking at alternative cryptocurrency algorithms that create a floating level of the money supply but without risks of hyperinflation, hyperdeflation, manipulation, etc.

3. Useful Proof of Work: You are right about the huge potential for blockchain in the future. Yet some are criticizing the technology -- at least the BTC implementation -- for being a prodigious user of energy & creator of carbon emissions. Clearly there are many computationally intensive useful tasks in bank with combinatoric portfolio construction, data mining, giant MonteCarlo runs, etc. but the defining characteristic of the cryptocurrency mining algorithm is that it is very hard to do but extremely easy to verify. Can one find valuable financial numerical analysis tasks with a similar property?
 
User avatar
ChrisGifford
Posts: 36
Joined: February 6th, 2016, 10:27 pm

Re: Blockchain ledgers for banks

July 20th, 2016, 10:38 pm

Hi Traden,

Thanks for this.

1. These aren't easy questions to resolve.  I think that much of this comes down to control of the flow of information and the effects of availability of information to different parties and the level of robustness of security/cryptography that goes with the transparency.  Credit and debit card numbers are secure for the level of encoding they have and, in like fashion, I anticipate that blockchain transactions will have a similar approach to encoding (prime numbers to the rescue).  In that respect the transparency will be secure (although, as with robberies there will always be criminals who will attempt to break them).  Privacy can be built into the design of a blockchain just down to which nodes on the network are responsible for verification so, in that respect, levels of transparency can be tweaked.  Banks are likely to follow their fiduciary duty to clients and respect their privacy (squids permitting).  I envision that this will develop in a heuristic manner as different banks and different clients have different needs.

2a.  Yes, that is an interesting perspective and I agree with it.  I don't think that Taylor's company will eliminate the need for banks but it (and other companies like it and banks) will revolutionise what banks are and what functions they can perform.  I was reading something a few months back about automated businesses - it's a far out prospect but not something impossible.

2b. Yes, I think it is either is or can be a unit of account.  The stabilisation of BTC is dependent upon the national currency of the BTC holders.  Much of the big moves in BTC over the last 7 months can be explained by the Chinese economy and the large number of Chinese investors who hold BTC or who have bought into it.  If an investor thinks that their country is going to purposely depreciate their own currency and they have a choice to buy in to a cryptocurrency which will not be manipulated by their government due to import/export considerations then they will buy into that cryptocurrency.
Do you mean that the gold standard currencies *were* flawed (the gold standard was replaced by Bretton Woods in '46)?
In the paper I've been working on I argue that cryptocurrency/non-fiat money is subject to economic risks like hyperinflation and hyper deflation because of the pricing mechanisms for the goods and services which can be purchased with cryptocurrency/non-fiat money (of course, some designs of a cryptocurrency and the blockchain such at BTC can but in place measures to preclude such economic risks but I think that any viable cryptocurrency-blockchain system will require the potential for something akin to QE for it to work properly).

3. Yes - I can think of tons of them ;-)
 
User avatar
Traden4Alpha
Topic Author
Posts: 3300
Joined: September 20th, 2002, 8:30 pm

Re: Blockchain ledgers for banks

July 21st, 2016, 4:07 pm

1. AFAIK, the blockchain for BTC is 100% transparent (i.e., any citizen, company, or government can inspect it) and it's only the anonymizing aspects of coins that hide the owners of the coins (although I get the sense that various methods can pierce that veil). In fact, it is the irrevocable, public nature of the blockchain that makes it impossible for any entity to repudiate a confirmed transaction or to claim ownership of a coin. I have hard time reconciling secrecy & transparency. Schemes for managing "authorized" users seem far too porous. For example, what if a bank is hacked (by any of a myriad of malware, phishing, disgruntled employee mechanisms), it's key stolen, and the entire history of the bank's ledger is published?

2a. Although Taylor's company might only allow banks to be customers, a competitor to Taylor could easily offer blockchain ledger services to large companies including the ability to do direct company-to-company transactions without the usual two intermediate banks. Consumers can already do this using cryptocurrency although I get the sense that most cryptocurrency users subscribe to a wallet service which is just a bank by another name. The point is that for basic medium-of-exchange transactions and store-of-value accumulation, cryptocurrency and distributed ledger users don't need banks. Why pay the bank's overhead for these services?

Where banks currently still provide value is in the pooling of deposits and pooling of risk-managed loans -- customers loan their money to the bank, the bank pools the money (across expected redemption time horizons) and then the bank loans the pool to suitable borrowers who pay interest which is then shared by the bank's depositors and investors. There's huge value in the pooling, time-horizon management, and risk management aspects of this. There may be various peer-to-peer lending services that could supplant this function but I don't think they do as good a job on either the risk management side (i.e., the consumer must contribute in-expert labor to the underwriting process) or the time horizon side (i.e., the consumer must decide when they'll need their money back and pick loans that match their horizon). Personally, I'd just want my money to be available when I want it and automatically earn interest while it sits. So banks may still have some advantages. I'd also think there are HUGE opportunities for alternative lending/investment structures (straddling fixed-payoff debt and variable-payoff equity) although traditional banks might be hobbled by regulation in this regard. I'd think that the biggest challenge for banks will be in stranded legacy assets which will become liabilities (bank branches, ATM networks, pensions for retired employees) in a world of cashless, online transactions, and online startup financial services.

2b Good points. The vast majority of BTC use does seem intimately connected to the BTC user's national currency. And with the supply of coins fixed by Satoshi's algorithm, fluctuations in demand resolve as significant fluctuations in the price of BTC. Moreover, given the tiny volumes of BTC relative to major national currencies, it's the price of the BTC that moves the most and the national currency moves the least. BTC will only become a bonafide unit of account if its users begin doing more transactions in BTC than they do in their national currency. Then the national currency will lose its standing as a unit of account as it does in failed countries where people start preferring USD or EUR to their country's specie.
 
User avatar
ChrisGifford
Posts: 36
Joined: February 6th, 2016, 10:27 pm

Re: Blockchain ledgers for banks

July 21st, 2016, 7:54 pm

1. Yes, it's a possibility.  That's when the lawyers get busy.

2a.  There are some recent developments which lead me to believe that it won't happen.

Yes - I agree - I was thinking about that in the background.

2b. Yes - there are a few supplementary thoughts I have on this.  I can send you a draft of my paper once it's complete if you want to have a look at it - feel free to send me a private message on here.

Thanks for the discussion.

All the best,
Chris 
 
User avatar
freddiemac
Posts: 7
Joined: July 17th, 2006, 8:29 am

Re: Blockchain ledgers for banks

July 22nd, 2016, 1:33 pm

The way I understand the application of the block chain technology in most financial applications is that it would rest not on an open ledger to all but simply an open ledger to those verifying the transactions, in this case the banks that utilise the technology. So it would be semi-centralised as opposed to the fully decentralised bitcoin system where anyone can in fact start mining and verifying transactions.

So I am not sure that the difference necessarily is so big compared to today – it depends on the market structure.
 
User avatar
ChrisGifford
Posts: 36
Joined: February 6th, 2016, 10:27 pm

Re: Blockchain ledgers for banks

July 29th, 2016, 5:27 pm

That's how I understand it.  There's a certain amount of versatility in the technology which permits different 'degrees of decentraliation'.
 
User avatar
ChrisGifford
Posts: 36
Joined: February 6th, 2016, 10:27 pm

Re: Blockchain ledgers for banks

August 2nd, 2016, 10:20 am

 
User avatar
Edgey
Posts: 14
Joined: March 23rd, 2005, 11:01 am

Re: Blockchain ledgers for banks

August 17th, 2016, 1:43 pm

1 Ledgers can be public, but (partially) encrypted.  Value could be determined by the regulator without exposing all information to all parties.  
2 If the ledger is self contained (a la bitcoin) then reality = record.  But if you want some real world value on the blockchain then you need some sort of trusted oracle service http://blog.oraclize.it/2015/11/04/orac ... al-launch/
3 Other, less wasteful options, have been proposed, and more will follow. (e.g. Proof of stake, proof of bet.)
Interesting proof of work here http://primecoin.io/
 
User avatar
ChrisGifford
Posts: 36
Joined: February 6th, 2016, 10:27 pm

Re: Blockchain ledgers for banks

August 17th, 2016, 6:20 pm

Going back to the transparency issue, Hyperledger looks like a good way to maximise it: http://www.coindesk.com/hyperledger-fir ... -explorer/ .
 
User avatar
ISayMoo
Posts: 2332
Joined: September 30th, 2015, 8:30 pm

Re: Blockchain ledgers for banks

August 19th, 2016, 12:04 pm

The biggest problem with blockhains is their energy usage.
 
User avatar
Traden4Alpha
Topic Author
Posts: 3300
Joined: September 20th, 2002, 8:30 pm

Re: Blockchain ledgers for banks

August 19th, 2016, 1:52 pm

The biggest problem with blockhains is their energy usage.
Indeed. A BTC miner will gladly burn $500 in electricity to make one $575 coin. Of course, if they can steal the electricity (e.g., malware bitcoin botnet, embezzled energy at a company, or an illegal tap into the public grid) then the amount of electricity they are willing to consume is infinite.

Yet is it any different from any other currency in which $X of metals and energy go into making $X*(1+c) units of specie? A copper penny has more than $0.01 of metal in it. How long does 1 bitcoin last before it needs to be replaced?

(And if blockchains start being used for other things such as ledgers, how much energy would a nefarious entity be willing to burn to fraudulently change a $1 billion transaction on a ledger?????)
 
User avatar
freddiemac
Posts: 7
Joined: July 17th, 2006, 8:29 am

Re: Blockchain ledgers for banks

August 22nd, 2016, 11:39 am

Bitcoin rests on decentralised verification of transactions. To get people to spend money on electricity to verify transactions they are sometimes given a newly produced bitcoin (known as mining).

But the maximum number of bitcoins is fixed. Then, what is the incentive to  verify transactions once there are no more bitcoins produced? Or more generally, if the expected return on mining bitcoins (which I guess must be falling as the amount of new bitcoins to be mined is declining) falls below the electricity costs?

The energy input needed to verify transactions increase as the number of remaining bitcoins to be mined decreases. So if the electricity price increase in the future there may not be any incentives to mine and the system would potentially break down (as I understand it)…
 
User avatar
ChrisGifford
Posts: 36
Joined: February 6th, 2016, 10:27 pm

Re: Blockchain ledgers for banks

September 29th, 2016, 12:10 am

By way of an update; http://www.jdsupra.com/legalnews/federa ... ake-91442/ .  The electronic coin now has two sides: R3 and the Fed.  Which way it lands will depend upon how it gets flipped.