Front running the central bank

There are front running opportunities that are entirely legal and can be executed at huge scale. Perhaps the mother of all front running opportunities is the so called Quantitative Easing (QE). QE is just a fancy name for the central bank buying securities in giant therefore market moving quantities. They have been buying treasury bonds till now. In anticipation of a coming restart of QE nothing seemed more obvious in 2019 than front running the central bank by buying treasuries before they do, and that is exactly what trader did at huge scale. This is observable through rising treasury prices, falling yields.

The repo market

Holding a bag of treasury bonds in anticipation of coming QE drains cash reserves, but is not usually a problem as there is a so called repo-market. Repo stands for repurchase agreement and works a bit like a mortgage loan, but that the collateral of the loan is not real estate but a treasury bond. Should a bank temporarily need more cash it can enter a repo trade with someone else whereby the other lends cash and this bank pledges the treasury bond as collateral for the loan. Most repo deals are rather short term a few days or even just over night. Since the loan is fully collateralized the interest rate is usually low. This makes it usually favorable to enter a repo deal instead of selling the speculative bond position to satisfy short term cash needs. Remember the treasury bond position was acquired in anticipation of huge profits when QE happens, it was a shame if it was liquidated before that at loss, just because of some cash shortage.

The repo market works best if there are others with offsetting needs, that is some with surplus cash. Unfortunately that is not the case now as front running the forth QE after having seen how lucrative the previous three were for those with treasury positions is a too obvious trade. Every bank is short cash and long treasuries, they are all front running the central bank.

Temporarily running out of cash would mean failing payments, something a bank can not allow to happen, therefore they turn for cash to the lender of last resort.

Lender of last resort

The central bank is never short of cash, since it creates it. Therefore it is always able to offer repo deals at any magnitude. That is what the FED engages in increasing hundreds of billions since September 2019. Doing so however sends a signal to the market, that there is not enough cash for everyone. This is like blood in the water for the predators who have thereby the confirmation that some banks struggle and either have to pay high interest rate for short term cash or liquidate their bag of treasuries at depressed prices. Both are great opportunities for those who have good standing at the moment.

Who folds next?

The FED hoped that the repo market will return to normal after some help in September, but it has not. We are witnessing a stand-off between the central bank and big banks, which could play out in two ways, none pretty:

• Some banks might fret and sell their oversize bag of treasuries and cause a spike in treasury yields, which would freak out others since their bags would be losing and would lead to a crash of the bond market and soon thereafter the economy or
• The central bank could fold and buy treasuries thereby yet again enrich those front running its operation.

The second outcome could be less noisy for now, but bodes well for a repeat of the same problem at a higher magnitude with coming QE. The predators will be only more greedy by then.

What else?

The central bank might choose to buy some other assets this time to spoil the front running bets and still inject cash into the system. Inevitably its choices will be depleted and front runner of QEn will break the system.

Bitcoin transaction fees rise if block space is scarce thereby take into account current use of a scarce resource, the block space. This is all fine but is not fully aligned with the costs implied by a transaction.

A transaction spends inputs and creates some new outputs. The new outputs become member of the UTXO set, the actual coins of bitcoin, and the spent inputs that were earlier outputs are deleted from the UTXO set.

The UTXO set is huge and ever increasing. It is not as big as the block chain but is still gigabytes with increasing trend. While the block chain can be pruned in a full node, the UTXO can not as it is the ultimate guard against double spending a coin.

Every transaction validation looks up the UTXO set and this operation is takes longer as the set grows. An output that is in the UTXO set increases processing time for all subsequent queries of the set until it is finally deleted because spent.

This means a transaction that spends an UTXO that was long standing is the final opportunity to charge for the term of its UTXO use and that is what we should do.

I suggest that miner impose a minimum fee to transactions that is not only taking into account the byte-size of the transaction but is also increasing with the age of the inputs. This does not have to be net increase of fees, but transactions that spend recently acquired coins might be even cheaper than now.

There is no fork needed for this as miner are free to implement any transaction filter, rejecting transactions that do not meet their particular taste.

I urge them to consider implementing a requirement that makes required fee increasing with the sum of spent coin’s age as this aligns fees with actual costs that the owner of those outputs imposed to the network.

Some miner might ignore this suggestion, actually I am sure most will for now, since they are focused on short term revenue that is highest if picking highest fee-per-byte transactions. Some might however recognize with time that alignment of costs with actual resource use will lead to higher revenues on the long run as it sets the right incentives on UTXO use.

A bitcoin’s rules allow that a miner who creates a valid continuation of the block chain includes a transaction that benefits the miner with new bitcoins aka. subsidy and with fees offered by transactions within the block.

The source of fees is explicit and clear. Every transaction within the block pays the difference of the sum its outputs and inputs to the miner of the block.

The source of new bitcoins is however not visible, they are just granted to miner out of nothing. New coins have significant value, where does that value come from?

Thermodynamics of wealth

A source of value is capital inflow into the bitcoin economy. Someone who already possessed wealth in an other currency buys the new coins. We see fluctuations of bitcoin’s price as a consequence of capital in- and out-flows. Those movements are rather severe and thereby hiding the answer to our question.

What if capital inflow into the bitcoin economy is zero, who pays for new coins then? This is not unlikely, we get close to it during dull days and there are many days where capital flows out of bitcoin. Mining would not stop as a consequence of no inflows, but the wealth that new bitcoins represent would have to come from within the bitcoin economy. Why? Because wealth, just like energy, can not arise from nothing. In absence of an external source someone within the bitcoin economy have to get poorer so the miner gets more wealthy.

Before we answer who pays, let’s quantify the wealth increase of the miner. Mining is a costly business to run. A miner will have to pay for electricity, amortization, labor, research, financing etc. The miner will only become wealthier if bitcoins mined worth more than the sum of these costs. Therefore we need a source of wealth that equals to miner’s profit and not revenue.

A miner who is not profitable does not get wealthier and there is no need for capital inflow or anyone to become poorer to offset it.

Above arguments are easier to follow if one considers that all costs of a miner have to be paid for with bitcoins the miner produces. This remains true even if bitcoins have to be traded for fiat currency first.

The miner trades some bitcoins with those who supply electricity, labor etc. Those resources are consumed (destroyed) in course of the production and their loss cancels out the wealth of the spent coins that are now owned by the resource supplier (or someone who gave fiat for them on the way to supplier).

In contrast wealth represented by coins that the miner kept after paying for production resources represent wealth that is not offset by resource consumption.

Dilution

Observe that number of bitcoins in circulation also might have increased in the process. This is the case until subsidy runs out.

If the miner was profitable then loss of resources consumed in production is less than the value represented by the new coins. This means the purchasing power of a single bitcoin must fall.

Thereby all bitcoin holder jointly pay for the profit of the miner.

For illustration, let us assume that it costs 3400 USD to produce a bitcoin and the current market price for it is 8400 USD. This means the miner spends only 0.40476 bitcoin to produce a bitcoin.

At current schedule, miners altogether will mine 12.5*6*24=1800 bitcoins a day. 1,800 * (1-0.40476) = 1,0714 bitcoins represent wealth in excess of burned resources.

The purchasing power of bitcoin is diluted proportional to current coins in circulation with 1,0714 / 17.89 million that is with ca. 5 USD

On the long run

Miners profit is constantly under pressure through new entrants and technology improvements, therefore profitability is expected to be low on the long run. Profit rate should still remain positive as no one assumes the associated operational risk for no return. Those who are not profitable will leave.

Consequently dilution will stay with us to some extent until there are new bitcoins available to mine.

Fees do not dilute purchasing power of holders as profits in them are paid by those who transact.

Conclusion

New coins dilute purchasing power of bitcoin holders to the extent that their production is profitable. Profitability of miner and hence dilution is expected to be low on the long run, but is rather significant nowadays.

Fees

The most obvious cost of storing wealth in bitcoin is that it requires some fee to move it in and out of a wallet. The aggregate amount of fees visible on the block chain as it is the excess income of miner above the subsidy.

These fees were paid by those who moved their coins, and were proportional to the byte-size of the moving transaction, not to the moved amount. Nevertheless they were explicit costs of storage billed for use of the ledger.

Dilution

Miners also receive new bitcoins for their work in a predefined schedule.

New coins increase the supply of coins to market and would drive their value lower if there were not other factors that dominate price movements, namely:

Capital inflow

Those who buy bitcoin for an other currency contribute to capital inflow into the bitcoin economy and those who sell drain its capital stock. The net effect of these in- and out-flows dominate price movements and hide the deterministic value decline dilution would cause.

Cost of mining

Bitcoin mining implies significant costs. Miner cover production costs by selling at least a fraction of their production. They contribute to capital inflow only to the magnitude of costs that they cover by selling bitcoins. (This thereby assumes constant bitcoin price.)

Consequently miner’s profit that was kept in bitcoin decreases wealth of holder through its dilution. This is also obvious since where else could the value come from if not covered by new inflow?

Note that if competition drives miners profit margin to zero and they cover their costs at selling bitcoin for a constant price or by paying costs entirely in bitcoins, then their activity no longer dilutes wealth of holder. This is true in the limit and temporarily at some points in bitcoin’s history but not in general.

We can observe changes to miners’ profit margin through changes of mining difficulty. If bitcoin current market price is not sufficient to sustain production then miner will turn off some devices to save electricity, which will be observable as drops in difficulty. This is certainly an imprecise measurement as it gives us a hint when miner with worst equipment and higher energy prices give up. Industry leader are likely still profitable and even cheer the shakeout of competition that will give them higher market share as price recovers.

Mining cost estimate

The most recent example where mining difficulty recovered from a longer downturn was in January 2019 at a bitcoin price of 3400 USD. I suggest to assume this as the most recent estimate of bitcoin mining costs for a unit.

Storage cost estimate

Using this mining cost estimate and current bitcoin price of 8400 USD we can assume a current mining profitability of at least 5000 USD / bitcoin produced.

This means that at current schedule of 12.5 bitcoins per block miners’ profit is ca. 9 million USD/day. This profit dilutes holders wealth at a daily rate of 9 million USD / 150.3 billion USD (market cap), that is ca. 2.1% pa.

Storage cost of wealth in bitcoin is therefore currently 2.1% annually plus transaction cost that is negligible for now. Storage cost will drop significantly with halving as that will cut into miners’ profitability.

This cost is hidden by more dominant price moves caused by short term capital in- and out-flows (speculation), but is nevertheless endured by investors, who should know about it.

Even those skeptical of Bitcoin’s monetary value recognize that it transfers its own units with unprecedented security, un-censorable and global.

Having the same utility to transfer any other asset, that is e.g. USD, Land, Equity, would be enormously useful. There were many attempts to do this with all kinds of overlay protocols and side chains linked to the Bitcoin chain.

None of these approaches managed to establish a design pattern until now that would have been as widely used as a much less secure option offered by Ethereum.

Property transfer

One could interpret a certain coin (UTXO) of Bitcoin as the right to a property and then track its movement to new owners. This is the basic Idea of most overlay designs.

There are technical issues with this, but more importantly why would anyone respect the transfer of property right expressed with the Bitcoin transaction the real world?

Transfer of property rights is usually possible through a 3rd party who maintains a register. The transfer is only valid if that registrar thinks it is valid.

Smart Contracts

There could be arbitrary rules applied by the registrar that a transfer must comply, so the registrar approves it. Attempts to capture the rules in a smart contract as advocated with Ethereum have been consistently failing for technical reasons or remained far from the complexity of the reality.

Validators

A better approach is to ask the consent of the registrar for the transfer transaction to be valid. This practically means the registrar also has to sign the transfer transaction in Bitcoin so it becomes valid. This way Bitcoin does not have to know the rules the registrar applies, just that the registrar agrees that the transaction is valid.

There could be many independent validators that together act as a registrar by contributing to a threshold signature. This allows us to implement a consent by majority opinion of the validators.

Side Chains

There are many features that would make Bitcoin more useful if they would be available there and many rules that if validated by the Bitcoin nodes would support more uses of it. Since preferences and goals differ it is very hard to reach consensus and sufficient support for a change even if it is undoubtedly useful. This made people experiment with side chains that freely define their own rules and how value is moved between Bitcoin and the side chain.

The problem with side chains is lower security. First they can not compete with Bitcoin’s POW and second it takes time until their unique design is perceived similarly trustworthy as Bitcoin’s.

Embedded Chains controlled by Covenants

It would be much more attractive from the security viewpoint if side chain was only a logical construct embedded into the main Bitcoin chain.

For this we have to find the bare minimum of backward compatible changes to Bitcoin that allow a logical isolation of transactions making up the chain and ensure additional validation of those with arbitrary added rules.

I proposed earlier the introduction of generalized covenants for the purpose of supporting full reserve banking and posted an more generic example this morning onto the developer’s mailing list how these could be used to create a a property right chain embedded into Bitcoin. There you find technical details, here I explain the concept in plain english.

The first problem to solve is that a coin (UTXO) that is considered to represent something else than Bitcoin may not escape the validations mandatory for transfers of that asset. A covenant achieves this by encumbering the coin with a restriction that it can only be sent to addresses that require a signature of the validator of its asset class. A covenant may also mandate that any coin that arises from splitting or merging of a coin associated with the covenant also inherits the covenant. This creates a transitive control of validators for the asset class.

The second problem is to ensure coins can only be split and merged such that before mentioned transitivity is well defined. That is e.g. that one should not be able to create Bitcoin transaction that would merge coins encumbered with a covenant with plain coins or coins bound to a different covenant.

I am pretty sure that both changes can be implemented and hope that fellow developer will agree that this design pattern for embedded chains constitutes the most powerful yet least invasive way to significantly extend Bitcoin’s use, so it has the chance of being deployed to the network.

The example I gave is a generic property chain embedded into Bitcoin, that could represent movement of ownership rights of just about anything provided their registrar is willing to sign transactions for the Bitcoin network.

Fractional Reserve Banking

The fiat currencies we use nowadays come into existence by someone borrowing them. To illustrate the mechanics let’s consider the case of Bob who buys a house with a mortgage loan. This implies the following series of events:

1. Bob signs a mortgage contract with Alice the banker.
2. Alice ensures that she has the right to sell the real-estate in case Bob would fail to follow terms of the contract. How she achieves this depends on jurisdiction, not relevant here.
3. Alice credits the account of the seller of the real-estate with the selling price, that is also the notional amount of the mortgage.
4. Bob regularly pays installments that cover interest and partial redemption of the mortgage until the notional amount is fully paid back through the redemptions.
5. If all goes well till end then Alice forgoes the right secured on the property, if not she uses the right to recover her loss.

This is all fine, but there is a disturbing detail in step 3. where does Alice have the money from that she debits to the seller? In our current banking system Alice simple creates it. Yes, this is completely legal until Alice does not violate a long list of checks mandated by banking regulators, mostly aimed to ensure that the amount of money Alice creates does not exceed a high multiple of her own capital and deposits under her control. This is called fractional reserve banking. The money created this way will be gradually destroyed again through Bob’s redemption or if he fails by the forced sell price.

If you think this is crazy, then consider that even the money Alice has is created the same way. Alice actually borrows that money from the central bank whereby she provides some collateral. The collateral may well be a so called ABS that re-packaged Bob’s mortgage agreement with many others.

There is no hard limit on the amount of money such system can create. The soft limits are regulated ratios and the bank’s capability to get people signing further loans that they can also honor, so the bank can afford to pay interest to the central bank. The later is not a hurdle nowadays as some central banks lend at 0% interest.

Wait, it gets even more crazy. Credits that fail, reduce the value of bank’s collateral and thereby would force them to reduce their own borrowing from the central bank which would curtail their ability to create credit which would slow down the economy and therefore cause even more credits to fail. To avoid that central banks got more creative recently and begun to buy assets that represent troubled credits at prices no one else would pay, thereby pumping or at least maintaining their value. They even invented a “Modern Monetary Theory” to justify all this.

It is evident that such a system is bound to create an ever increasing amount of debt and the measures that central banks do to underpin it only delay the inevitable collapse of the scheme at latest at the point where all prices of credit become meaningless. Guess what, we are pretty close to that. A huge fraction of government debt is trading at prices that makes them yield negative returns. That is the sum of interest and redemption does not cover the selling price.

All above was an introduction, so you understand why we need an alternate way of doing lending, so a Bitcoin economy does not end up in the same trap.

Full Reserve Banking

Full Reserve Banking is where banks are no longer allowed to extend loans in excess of their own capital. Before gold receipts were introduced, this was the way banking worked.

Not only historians have interest in full reserve banking, but is seen by many people including some economists as an alternative to fractional reserve banking that we should stop for previous reasons. The most prominent move into this direction was a recent popular vote in Switzerland to mandate full reserve banking to all Swiss banks, that achieved 24% support of voters.

Bitcoin credits

Bitcoin as is is digital cash. It does not mandate any form of banking it might be used in. Although there is a fixed supply of Bitcoins nothing yet enforces a limit on credit denominated in Bitcoins. This could lead to the very same problems in a Bitcoin economy, we currently see elsewhere.

We have seen repeatedly that e.g. Bitcoin exchanges operated a fractional reserve scheme, some collapsed as customer demanded their deposit and some are still getting away with it.

One way to keep check on Bitcoin businesses is requiring proof of reserves, that is like a public audit cryptographically proving that the Business had at a certain time point access to a number of Bitcoins.

Proof of Reserves is not much better than traditional audits as it does not give a guarantee of ongoing ordinary business, just shows a picture that could have been painted temporarily for the time point of the audit.

Credit Covenants

The best solution would be if every Bitcoin credit would carry its own proof that it is covered by reserves all the time.

It turns out that this is technically possible with implementation of the proposal I posted today to the Bitcoin developer mailing list. There you have the technical details, for which I give a higher level description here.

On the technical level Bitcoins exist as unspent transaction outputs UTXOs. Transactions merge and split UTXOs to new UTXOs. One could consider an UTXO a coin of certain amount. An coin can be spent by someone if able to satisfy any of conditions that the coin is programmed to accept.

Giving credit to someone means transferring coins to him and hoping to get them back. Such a deal requires trust, since the borrower could be dishonest or bankrupt by the time the credit should be re-paid.

It was a great option if we could programmatically ensure that the coins will be paid back. We can achieve this by assigning a condition to the coin such that the lender can take it again later. Next we have to ensure that this condition is inherited by any coin that arises from splits or merges of this borrowed coin.

The technical means to do that is to assign a credit covenant to the coin that forces descendant coins to offer the same option to be taken later by the lender.

Do coins encumbered with a credit covenant have a value? They represent exclusive but temporary access to a scarce resource. I intuitively think yes, but the answer will be given by the market. An analogy to support my intuition is land that is not sold but rented for a longer term. If land is scarce then people will pay to take over a previously agreed rent contract.

The idea of covenants was first introduced in a paper suggesting a solution for Bitcoin Vaults. I generalized the idea to be compatible with the taproot proposal and provided examples how it can be used to technically enforce full reserve banking on the mailing list.

I know hope that the community understands the importance of full reserve banking and that we developer work together to complete and deploy this proposal on the network.

Correction: the first discussion of covenants is probably this post on bitcointalk.org.

The number goes up and up. This could get inconvenient for those who got used to their omnipotence of bullying the markets in whatever direction they like, mostly up.

Yes, I am talking of central banks. Their interest rate policy became the only parameter worth trading. In this late phase of their power their actions move everything. Equities, Bonds, Commodities rise in sync every time the FED or ECB reassures to provide yet more liquidity, that is money for nothing (zero interest) or in exchange for bonds that no one else would buy at prices they do (called QE). They even invented a “Modern Monetary Theory” to justify perpetual control, mostly pumping of asset prices.

Markets breathe, even if provided with pure oxygen of abundant liquidity. Markets must sometimes exhale, that is fall for some time. This is tolerable to those in control as long as all assets do that in sync.

Otherwise things could spin out of control as owner of pumped assets would en masse switch to the asset which is not falling. The existence and prosperity of such asset would question the omnipotence of central banks. They are not new to this problem and learned some tricks.

The case of Gold

Gold used to endanger central bank’s omnipotence by offering an alternate store of value to that of the currencies they control. Some still believe this works. Decades of recent history shows, it does no longer.

Buying gold as a protection against failure of central banks is stupid, as central banks are the biggest holder of gold.

Those who own most of gold, are best positioned to control its price. They can frustrate gold bugs until they are becoming an ignorable minority, which they are.

The means of frustrating gold bugs are mainly:

• lend some gold at low or even negative interest to those willing to short sell it. Thereby they can repeatedly apply temporary downward shocks to gold market, without reducing their gold position.
• embrace futures markets and make physical holding inconvenient (e.g. delaying introduction of a Gold ETF as long as you can). Price can be bullied with futures once open interest of futures markets surpasses that of physical trading. Futures require margin only in form of any other successfully pumped assets, wich they have plenty.

The beauty of these means that they can be repeatedly executed without reducing gold holdings, thereby remain in control.

The case of Bitcoin

Should Bitcoin rise to significant value (means trillions of market cap) while other assets fall, then it could question if central banks are still in control.

Could they use the means they frustrate gold bugs to frustrate HODLer? Hardly, at least for now.

Since central banks do not own significant amount (likely none) of Bitcoin, they can not lend it to shorts. There is some danger that big pools of Bitcoin, that is e.g. exchanges might begin to lend Bitcoin so they earn interest on their balances. This could be avoided by requiring proof of reserves or using only joint custody, not letting full control (that is the only private key) at the custodian.

Bitcoin future markets are tiny compared to its “physical” markets, so the tail does not yet wag the dog. Open interest on CME over all maturities is today equivalent to 30k Bitcoin (and this is close ATH), that is ca. 20% of just Binance daily volume, so it is really insignificant.

Delaying Bitcoin ETFs is a deja vu of Gold’s case but is not that relevant as barriers of storing and moving Bitcoins are much lower than that of Gold. The Bitcoin spot market should remain more attractive even if there is widespread access to futures.

Added on 27. Jun 2019: Futures open interest on BitMEX is much higher than CME, but this is not applicable to my argument as BitMEX margins have to be funded with Bitcoin or other crypto and not with fiat collateral. Therefore those only in control of fiat can not use them to bully crypto markets.

I formulated a harsh critique of Libra’s tech as is, but hey it is software, it will evolve, and with enough persistence and enough funding it could at some point become technically usable.

Let’s assume people love it, as it is easy to use and delivers on the promise of a quite stable value if measured in traditional currencies. How big could its Libra Reserve grow?

There are many ways to estimate its magnitude, some examples:

• If people use it as pocket change, maybe as a single digit fraction of the cash they currently carry, then the value of all Libras could become about 700 billion USD (ca. 10% of global M0).
• If Libra is at least as popular to hold as Bitcoin, then it must have a total value above 150 billion USD

Above estimates are pretty conservative, if Libra works and people love it, it can easily acquire value 10 times of them.

So what, you say? We have seen a few companies surpassing the market value of 1 trillion USD recently and facebook is also worth about half of that.

There is problem with the promise that Libra’s purchasing power will be stable at those magnitudes. To understand why, let’s review how alternatives work.

Paper cash

The only guaranteed value of paper cash is that the government accepts it as a method of paying taxes. The central bank, as the issuer of paper cash, promises to exchange paper cash against a digital representation of the same, so it can be deposited into a bank account and withdrawn again from there as paper cash. That is it. Anyone else accepting paper cash and giving some value in exchange is doing so voluntarily. The ubiquitous willingness of people to accept paper cash gives stability to its purchasing power.

Bitcoin as cash

Bitcoin does not have any representation outside its ledger. Bitcoins are created on its ledger following a predetermined schedule. The number of Bitcoins in existence does only depend on the time passed since the network started. It does not matter how much energy was spent in mining or how many people find Bitcoin useful. Bitcoins value is supported by voluntary exchange only.

Since only a small subset of people is currently willing to accept Bitcoin, and its quantity does not adjust to demand, but follows time, Bitcoin’s purchasing power is fluctuating wildly.

Libra as cash

The quantity of Libras will follow demand. People will buy and sell Libra for their traditional currencies. The net of those transactions will create marginal demand or supply for Libra, which will force Libra Reserve to create or redeem Libras.

Libra Reserve will sell newly created Libra’s to those facilitating the exchange and invest the proceeds into hight quality liquid assets. Similarly it will sell the liquid assets to buy back and potentially destroy Libras in case supply exceeds demand at exchanges.

The standing offer of Libra Reserve to buy or sell Libras should keep its purchasing power stable.

The problem with scale

Above scheme of stabilizing value works, until it does not. It breaks down with scale as the magnitude of the assets backing the scheme reaches a level where eligible assets became scarce or transactions triggered by demand changes have market moving size.

It is very hard to collect and trade liquid assets in quantities of hundreds of billions or even trillions. Such pool of money has market corrupting effects only comparable to the market operations (known as QE) of central banks.

Possible remedies

There are ways Libra Reserve could deal with scale:

• Avoid frequent trading of collateral assets by also holding Libra. This actually means running fractional reserve.
• Holding cash. This is inconvenient in paper form in quantities required. Bank accounts are also not suitable for these sums as insurances would not cover losses if the bank would fail. One would rather want to have a deposit at the central bank. Access to central bank digital cash is until now restricted to big banks. Means that Libra Reserve would have to either get a banking license with all its consequences or an exception. Later would be an act that would be seen un unfair advantage by many corporations that would love to access the central bank ledger without becoming a bank.
• Free float Libra. Means widening the bid — ask spread of Libra Reserve’s offer or even suspend it. This could be attempted as willingness to accept Libra becomes ubiquitous, so it stabilizes. But what if its value drops e.g. through a vulnerability of its most popular wallet? Even if its unwarranted, panic sellers of Libra could drive the price down quickly without a standing order of Libra Reserve. Who will pick up the bill for the lost wealth in supposedly stable currency?

Consequences

The giant pool of liquid assets at Libra Reserve would qualify it either as a systemic risk or as an exceptional market participant with unfair advantages, likely both.

The preferred alternative by the incumbents is that things just stay as they are. There is little chance for this outcome.

Libra might accelerate central bank’s consideration of issuing their own digital currency adding to the strong reason I wrote about earlier.

Given that Bitcoin is bootstrapping a global currency without introducing the problems that Libra implies, we might see some surprising converts to Bitcoin maximalism.

An argument against Bitcoin’s value is that alternate crypto currencies, also known as “shitcoins”, exhibit the same digital scarcity within their own network. Pundits add up market cap of Bitcoin with shitcoins to come up with a market cap of “crypto”.

Some think the existence of shitcoins defies Bitcoin’s scarcity and show that Bitcoin can be copied and multiplied and consequently Bitcoin and all crypto is worthless.

It is pointless to argue against this with the network effect of Bitcoin, as the question is not if Bitcoin could be replaced with a better version of itself, but if scarcity can be achieved at all by a design similar to Bitcoin.

Similar to Bitcoin?

A shitcoin could be considered similar to Bitcoin for many reasons. Some consider ZuckBuck similar, because … whatever.

The similarity that really matters is the mechanism that creates scarcity. In Bitcoin’s case it is the Nakamoto consensus built on proof of work (POW).

Many shitcoins experiment with alternate consensus algorithms, such as BFT, POS, POET, governance or any combination of them. The ability to prove work is ultimately constrained by physics and available resources. Those alternatives to POW use strictly more assumptions hence the scarcity they achieve is of lower reliability, quality.

We will soon see if the quality of scarcity offered by Zuck is deemed sufficient by the masses and hence ZuckBuck manages a “flippening” against fiat in the daily uses of buying likes or in-app gadgets in facebook and related apps.

Our quest is however not for a good enough scarcity for some use case, but that of ultimate digital scarcity, which would give rise to highest value.

Proof of Work

Different qualities of scarcity are with us already. There are less guarantees for scarcity of fiat than that of gold. The marginal supply of fiat changes at the will of bureaucrats, that of Gold is limited by natural reserves and work invested.

A doubt of Gold Standard could be articulated as: Other precious metals exhibit very comparable physical properties, hence they are as good as gold, therefore supply of “hard money” is not constrained by gold supply.

The world however operated on Gold Standard and not on Precious Metal Standard. Those promoting an alternate Silver Standard experienced heavy losses in the process of consolidation.

It seems that although several precious metals are eligible through their physical properties and supply of all of them is constrained by work, only one became the standard to store value.

Scarcity through POW

There is a striking parallel between work in gold mines and in Bitcoin mines. The resulting product proves work performed. POW is not unique to Bitcoin. POW is used by countless shitcoins. Do shitcoins that also use POW undermine digital scarcity?

Digital scarcity based on POW does not require that Bitcoin, as we know it today, becomes the only one, but that only one POW coin is desired by all. Like there can only be one Highlander, one immortal swordsman.

POW Standard or Bitcoin Standard?

To keep its emission schedule Bitcoin requires the proof that miner did (busy-)wait a time span (expected 10 minutes) before producing a new block. The proof is the result of a computation.

Alternate proofs of wait, such as POET suggested by e.g. Intel could deliver a lower quality of scarcity, but POW makes no compromises. POW is a method of measuring time with computation.

The difficulty adjustments based on timestamps are only there to periodically adjust for time measurement errors, the actual tool of time measurement is the computation itself.

Time measurement with computation is only reliable if the executed computation is irreducible and all resources capable of its computation are deployed to that task.

Fragmenting computing resources across competing time chains however will not achieve the reliability of one time chain that uses all resources. The Bitcoin Standard will arise as a consequence of aiming for the best quality of time chain.

I anticipate on the long run a convergence to a single time chain, that is likely a descendant of Bitcoin as we know.

There are many choices of irreducible computation, and double-SHA256 of Bitcoin is not one. See midstate or ASICBoost. A better time chain would have a simpler POW that is provably irreducible. One that can be evaluated on the least simple machine, a cellular automaton as in NKS to exclude any chance of shortcut. A time chain with such POW would give us the most precise measurement of time with computation.

Bitcoins exist technically as unspent transaction outputs (aka. UTXO) which may be spent by their pseudo anonymous owner. The value of UTXOs vary widely. They are the actual coins of the system. As of now, there are UTXOs of a few satoshis and some with many thousands of Bitcoins.

We could use the value distribution of the UTXO set to estimate Gini coefficient, but that would only be meaningful, if every Bitcoin holder would own exactly one of them. This is certainly not the case.

We could group UTXOs by addresses, but this approach is known to be futile since most wallets nowadays do not reuse addresses.

We could work out some elaborate heuristics to cluster Bitcoins, as do companies analyzing the Bitcoin blockchain. The more external data is joined with the blockchain data the more precise a clustering would be. The downside of this approach that it is proprietary and relies on data not widely available or even secret.

Even with near perfect clustering based on elaborate research and lots of exogenous information, the amounts grouped may not represent holdings of individuals. The biggest heaps of Bitcoins are likely in hot and cold wallets of exchanges that represent wealth of many users. Although as you know: “Not your keys means not your Bitcoins”. But this is an other topic.

The approach I took instead is based only on blockchain transactions, more precisely some special but very common forms:

• A transaction with a single output (aggregation) puts those coins very likely under the control of a single owner. This assumption will hold less often with growth of the Lightning Network.
• A transaction with two outputs is very likely a payment. Whoever did that was apparently in control of the total input value. Mixing transactions that obfuscate ownership have very likely more than two outputs.

Finally I make the assumption that one makes transactions of above types in proportion of one’s wealth. Wealthy people also do small payments out of small inputs, but poor can not command large inputs. The distribution of these transaction’s aggregate input values therefore leads to an underestimation of the Gini coefficient of Bitcoin wealth.

I used a moving window of 1 million transactions of above types to calculate below Gini coefficients:

The calculated Gini coefficients show rather serious inequality, more than in any real-world country, and it is an underestimation.

I am not interested in discussing implications or desired levels of inequality, just attempt to deliver measurements here.