Settlement and liquidity

[ben]

The Lightning network is a proposal for a Layer2 on top of Bitcoin. A similar system was discussed by Hal Finney in 2010:
https://bitcointalk.org/index.php?topic=2500.msg34211#msg34211

Finney called the agents (i.e. nodes) in such a system Bitcoin banks. That is probably more accurate than hubs. Bitcoin banks would leverage a balance-sheet to provide liquidity in the system. Otherwise the lending activity is going to be unprofitable. Lightning, like Sidechains, misses the most crucial element in Bitcoin - incentives. In the case of a credit system the incentive is related to the relative time value of money, in form of the interest rate. In the current money fiat system the interest rate is not really a market rate, but that is another topic. Bitcoin is fundamentally a cash system, without any notion of trust and credit and time based payments. The exact reverse from fiat money which is a debt instrument from the government. In Lightning the time-based contracts are a particular form of credit in the sense of a payment in the future. Such new types of credit instruments could be powerful, but it would need to be connected to the liquidity of the parties and a pricing of that time-value.

Bitcoin banks would be trust-based systems. That trust has to do with counterparty-risk. Currently banks are completely opaque and strongly tied to central banks. With denationalization of money banks would be independent from central banks and be mostly long-term and short-term credit providers. The payment channels and time-locked contracts that Lightning introduced might be useful, but there are a lot of facilities that Bitcoin misses to do these things. First and foremost Bitcoin by design is really bad at trust based systems and accounting. The entire Bitcoin system is to prevent tracing payments to people and groups. That is a good thing to increase privacy, but really not functional if one wants to assess counterparty risk and create group-based money control mechanisms, which would be needed in a Lightning type design.

There are one could say basically 4 groups in the current debate, which is not about blocksize, but really about the scale and future of Bitcoin:

1. Bitcoin will scale and there is no problem (XT and Unlimited).
2. Bitcoin can't scale and there will be new crypto-based mechanisms (Core + Lightning).
3. The financial community, which says: Bitcoin is useless, but blockchains are really great.
4. The Altcoin community which believes Bitcoin is broken and new systems should be developed (1000's of Altcoins)

I think the truth might lie between all of these, which also highly depends on political views, not merely technical fact disguised as such. There are also proposals along the lines of 2. and 3. which have been rarely considered. There are also potentially quite futuristic scenarios where central banks (nation state governments basically) would accept Cryptocurrency as collateral. I believe 1. and 2. might not work out as expected (topic for another post).

References:
Joseph Poon, Thaddeus Dryja: The Bitcoin Lightning Network: http://lightning.network
Szabo: Shelling Out -- The Origins of Money: http://szabo.best.vwh.net/shell.html
Ryan Fugger: Money as IOUs in Social Trust Networks & A Proposal for a Decentralized Currency Network Protocol: http://archive.ripple-project.org/decentralizedcurrency.pdf
Hayek: https://mises.org/library/denationalisation-money-argument-refined

 

[YarkoL]

Very good post.
In principle I think LN and similar are way out cool and might be essential if micropayments in the Internet Of Things take off.
But I don't think they are part of the core design, they should be implemented by various third parties as competing solutions. Right now Core devs are locking Bitcoin in one particular solution.

In your classification of debates, I stand firmly in the fuzzy area between 1 and 2, and don't think that I'm the only one. That is, I think Bitcoin will scale via lifting the current block size limit, but that is hardly without problems, the main one being the cost of running full nodes.

And since I come from altcoin scene, 4 is not totally accurate, there are many more motives for running alternative chains than trying to replace Bitcoin - a notion that I have always held as laughable.

 

[Peter R]

Great post, and welcome to the forum, @ben!

Your comment about the Lightning Network missing the credit component incentives hit the nail on the head. A criticism of LN is that users won't tie up funds unnecessarily--and that's probably true. But if we imagine something more like Bitcoin banks, then suddenly people can earn interest by depositing money they don't need, or pay interest to borrow money they need but don't have. At the same time, both creditor and debtor would gain access to the micropayment advantages of LN compared to on-chain TXs (assuming that those advantages materialize). It would be a sort-of 100% reserve banking system.

This actually reminds my of the "Chicago Plan" proposed by Irving Fisher and other prominent economists in the wake of the Great Depression. A key feature of the Chicago Plan was that it “called for the separation of the monetary and credit functions of the banking system,” by “requiring 100% backing of deposits,” and by “ensuring that the financing of new bank credit can only take place through earnings that have been retained” and “not through the creation of new deposits, ex nihilo, by banks.”

 

[Erdogan]

I think full reserve banks is an ideal, but I think the temptation is too big, both for banks and depositors. The free market will keep the banks in check, but can probably not ensure full reserve (including the same time conditions for deposits and loans) for all banks. A balance might arise, where some banks are full reserve, some have reserves of some safe fraction (and some will not tell their customers the whole truth), leading to sporadic bancruptcies, but not a systemic fault.

 

[ben]

In terms of scalability I think one of the more promising approaches is ColoredCoins, i.e. actual credit on top of Bitcoin. Lykke is a project I'm contributing to which has such an architecture: https://wiki.lykkex.com/colored_coin_exchange_white_paper . Also the NuBits coin has a kind of delegated issuer principle. There are many interesting possibilities with this approach. The supply of credit could be controlled through shareholders. So the bank ownership could be widely distributed. The question then becomes what would such corporations & distributed corporations look like? 99% of Bitcoin flows through Bitcoin startups and exchanges, but they are not really part of the core system. E.g. there is no proper accounting layer to integrate third party services with core. Bitcoin has been called the most important invention in accounting, since double-entry book-keeping and ironically the accounting feature was first tagged on and now removed.

I did some technical work on payment channels and Lightning, but to work with Bitcoin on this level is a pain. Peter Todd outsourced the script interpreter to python-bitcoin, but there should be more work in this area to facilitate intermediate layers. Lightning, by the way is written in low level C code, so there clearly are missing levels of abstraction. I wonder how far this tech stack will carry. The outcomes of core tech versus the investment is not that significant. But still Bitcoin has 90% market-share and can be considered very secure (although with more risk than before).

Thanks for the link on the Chicago plan. Bankrupt banks should go out of business, and they would if the accounting in the financial system would actually make sense. Ultimately there will be a future where even governments have to watch their finances. Their ability to compromise the value of money will be diminished by voters demanding sound accounting and spending. I've thought about a political system similar to a modern democracy where the budgeting process is determined by voters. Basically taxes would have an attached mandate for spending, rather than delegating arbitrary power.