Profile picture
, 91 tweets, 28 min read Read on Twitter
1/ My latest article "Bitcoin's Security is Fine" where I address concerns with Bitcoin’s security model which is funded by the block subsidy + transaction fees. Part of a new content series produced by @InterchangeHQ called “The On-Ramp”

Thread 👇

blog.picks.co/bitcoins-secur…
2/ Approx. every 10 minutes, a new Bitcoin block is created, which contains newly minted Bitcoins (the “block subsidy”) plus transactions fees. The value of the newly minted coins plus transaction fees is called the “block reward.”
3/ Per Bitcoin’s hard-coded monetary policy, the amount of newly minted coins per block decreases over time, eventually reaching 0% in the year 2140. At the time of this tweet being published, over 83% of all Bitcoins that will ever exist have already been minted.
4a/ “Indeed there is nobody to act as central bank or federal reserve to adjust the money supply as the population of users grows. That would have required a trusted party to determine the value because I don’t know a way for software to know the real world value of things." - SN
4b/ If there was some clever way, or if we wanted to trust someone to actively manage the money supply to peg it to something, the rules could have been programmed for that.” — Satoshi Nakamoto
5/ Satoshi felt that setting a “proper” rate of inflation rate was impossible and that it introduced a political attack vector, so he decided to remove human decision making from the process.
6/ Each time monetary policy is changed or modified, human governance re-enters the system nullifying the certainty of monetary supply. A predictable monetary policy is key: Bitcoin’s focus on long-term stability and transparency creates confidence for investors and developers.
7/ So why does this matter? The block reward incentivizes miners to protect the network. As inflation trends towards zero, miners will increasingly obtain an income only from transaction fees. Some worry that transaction fees alone won’t provide adequate compensation. @BashCo_
8/ “In a few decades when the reward gets too small, the transaction fee will become the main compensation for [miners].” - Satoshi Nakamoto
9a/ Bitcoin’s existing UTXO set (ledger) and new blocks are protected by game theory and physics. Bitcoin uses proof-of-work (PoW) to make changes to the ledger difficult, which eliminates trust and introduces an external cost for any would-be attacker.
9b/ The miners buy hardware (capex) and electricity (opex) with the expectation of receiving their portion of the block reward based on work spent (hashes). The block reward financially incentivizes miners to behave properly.
10/ As the price of BTC increases, the value of the block reward increases (security follows price). As potential revenue increases, more hashrate comes online chasing that new profit. The higher the hash rate of a cryptocurrency network, the more expensive to 51% attack.
11a/ In the early stages of the network, Bitcoin miners are rewarded more heavily by the block subsidy than transaction fees. With Bitcoin’s disinflationary monetary policy, approximately every 4 years the block subsidy drops by 50%. This creates both volatility/price increase.
11b/ If demand remains constant/increases, the reduction in supply means demand is chasing less freshly minted Bitcoins hitting the market. This effect brings in new speculators, which is part of the beauty in it’s design, as the supply shocks bring greater awareness to Bitcoin
11c/ “As the number of users grows, the value per coin increases. It has the potential for a positive feedback loop; as users increase, the value goes up, which could attract more users to take advantage of the increasing value.” — Satoshi Nakamoto
12/ While the two are mixed into the same “security budget”, the block subsidy and transaction fees are very different. For the block subsidy, its value is both as a rational way to issue new Bitcoins and as a viral FOMO loop built into the protocol.
13/ The subsidy further stretches out the need for transaction fees to solely provide security. Hence why it’s called a “subsidy.” Over the long term as network effect becomes larger, demand for block space increases, thus decreasing the need for a block subsidy.
14/ With modeling done by @Awe_andWonder we can see that around the year 2032 transaction fees will begin to consistently represent a healthy portion of the block reward.
15/ Critics often argue that transaction fees alone won’t provide adequate security. Short answer is that we don’t know. It’s subjective since the amount of confirmations one would wait for depends on the transaction size and health of the network.
16/ Nic Carter presented several ways we could quantify an adequate security budget:

Threshold: Given USD level of security spend
Stock: Security spend should be indexed to the value of Bitcoin itself
Flow: Fees must be large relative to transactional volume
17/ I believe security is best measured as % of stock which eventually reaches a threshold level. Stock makes more sense than flow because miners are focused on long term operations as the space matures and costs shift to opex. Eventually this reaches a threshold level
18/ I hypothesize several hundred billion, in present value USD, would be an adequate security budget since it would be very difficult for a government to justify such a waste of an expense. They would also have to respond publicly for such an attack (HODLers are everywhere).
19/ Even during a 51%, transactors could simply wait to send a payment or wait for additional confirmations which isn’t all that bad. Note that a 51% attack wouldn’t “kill” Bitcoin, as you still cannot reverse historical transactions easily. bitcoin.sipa.be/19/
20/ It’s worth noting that PoW achieves several goals other than just minimizing 51% attacks, and increasing network effect:
- It makes it extremely difficult to forge units of the currency
- Accumulated historic difficulty makes it practically impossible to reverse the chain
21/ Bitcoin’s block space is a scarce and unique commodity: there is no alternative to prime real estate.
22a/ Getting a transaction mined can be seen as purchasing a portion of a block. By analogy, on average every 10 minutes, a fixed amount of land is created, people wanting to make transactions bid for parcels of this land.
22b/ The sale of this land is what supports the miners even zero-inflation environment. The price of this land is set by demand for transactions (as the supply is fixed). The basic premise is that if the network is being used then it will reward miners for protecting the network.
23a/ Some argue that altcoin-blockspace is an equal substitute. However, there are many ways we can debunk that theory. Bitcoin real estate is prime real estate. It doesn’t matter how cheap land is in Midland, Texas, it’ll never have the views or social network of San Francisco.
23b/ The unique value of Bitcoin’s block space is due to security, exchange, volatility, and coordination costs. This following section borrows heavily from @TokenHash's article “Observations About Paul Sztorc’s Bitcoin “Security Budget in the Long Run” Essay”
24/ Security costs: Bitcoin is the most secure crypto network due to the total accumulated hashes. This creates a market for high value/secure transfers in Bitcoin, e.g. central banks, governments, interbank, and other large value payment users, who will gladly pay very high fees
25/ Exchange costs: Between exchange commissions (0.1 - 4%), spreads (since alts are less liquid), and transaction fees both ways, there will be an indifference point beyond which it will be better to pay for Bitcoin fees.
26/ Volatility costs: Often people forget to consider volatility costs which depends on your holding period. Altcoins typically have higher volatility than Bitcoin which has the nasty effect of scaling with transaction size.
27a/ Coordination costs: There will be a limited competing block space market in the world. This is because our minds are limited and we will not think about 100 cryptocurrency names/fees/prices, and go about selecting the cheapest one each time we move value (@NickSzabo4).
27b/ Our brains will only support understanding the value of 2 to 3 coins at most, and we will be comfortable using them interchangeably up to a certain point (although there is a weak/unproven case to be made for obfuscating a plethora of coins behind proper UX/UI).
27c/ Additionally, since Bitcoin HODLers have a strong affinity to only transact in Bitcoin (monotheistic), multicoiners will be forced to transact in Bitcoin (Tyranny of the minority). For example: Square, Bakkt, and Fidelity are only supporting Bitcoin at this time.
27d/ “Bitcoin is money. Multicoinery is barter.” @_ConnerBrown_
28/ Finally, the Bitcoin core software is battled hardened with a mature ecosystem. This adds value to Bitcoin’s unique block space since there are more developers and businesses that examine and rely on Bitcoin’s code.
29/ With all of these considerations above, there will be a limited amount of coins in the future, with a limited amount of block space and transaction capacity.
30/ “Why would the majority of wealth and valuable apps not be secured by [Bitcoin]? As more users buy into [Bitcoin] [the price goes up]. [This] leads to increased security. From there, eventually usage, liquidity, and network efforts will compound on each other" @alexsunnarborg
31/ "The sudden multiplication of altcoins and ICOs during the last bull run was a race to mimic the wealth creation that happened in Bitcoin... not millions of people suddenly using blockchains to transfer money around the world and seeking to minimize those fees." @TokenHash
32/ A large portion of payments were likely done for novelty. Paying for something with crypto is harder, more expensive, and slower than traditional payment methods. Bitcoin's base layer is for building the strongest possible foundation for a new monetary system–not a new Venmo
33/ Transactional Demand

“I'm sure that in 20 years there will either be very large transaction volume or no volume.” - Satoshi Nakamoto
34/ Another worry is that as fees become higher users will shy away from transactions to avoid fees. However, we’ve seen empirically this is not the case: as transaction/trading volume has increased, fees follow in step.
35/ Price Elasticity

“Nobody goes there anymore. It’s too crowded.” - Yogi Berra

Consider the UX for most North American, European, and Asian participants. Any transaction fee has a higher level of friction than existing payment methods, so any fee is deemed “expensive."
36/ The price elasticity (for fees) of a Bitcoin transactor is largely due to the nature of the payment type being sent, an immutable SoV. During peak of 2017, the median fee was $38 and we briefly saw blocks with fees being greater than the subsidy. Comps in the real world:
37a/ Wiring Fiat

For US banks, the average domestic wire fee is $30-40, and $65-80 for international (both incoming and outgoing fees combined).
37b/ Offshore ($7T Market)

“The setup fee for opening an offshore bank account is between…. $1,935 to $3,745 for [a bank account and an entity filing]”
offshorecompany.com/banking/costs/
37c/ Real Estate ($250T Market)

“Buyers from China bought 40,400 units totaling $30.4 billion between April 2017 and March 2018. They spent a median of $439,100 per purchase” - National Association of Realtors. Average closing costs on a home are 2% of the value, or $8,000.
37d/ Physical Gold Delivery ($7.5T Market)

@TokenHash requested information from the Bundesbank regarding their NY Fed transfer of 300 metric tons of gold ($12 Billion at the time) from NYC to Frankfurt. It took 3 years and cost $4.8 million.
38a/ Increasing Density

Nic Carter’s MIT presentation highlighted two ways to improve transactional demand: increase economic or semantic density of transactions.

Semantic Density = other blockchains imbed data
Economic density = increasing transaction types on Layer 1:
38b/ Privacy

“In conjunction w/ techniques... Schnorr can... make it nearly impossible to pinpoint specific entities by simply looking at the blockchain while simultaneously allowing greater transaction density per block by aggregating signatures.” @LucasNuzzi
38c/ Layer 2

As Bitcoin scales (Schnorr  on Layer 1, Lightning  on  Layer 2, etc), it will become more and more efficient, driving higher on-chain usage. Jevon’s Paradox intuitively predicts this — as cars have become more fuel efficient, more miles are driven annually.
38d/ Layer 2 will support a massive number of cheap smaller transactions, whereas Layer 1 represents a more expensive settlement layer for large transactions (container ships vs cargo containers — Nic Carter).
38e/ LN boosts on-chain fees by increasing the utility of each on-chain txn (by allowing each to do the work of many txns), and by therefore making high on-chain fees more tolerable to the end user.
38f/ On-chain, block space is premium, hence transactions are charged for the space it takes to register the transfer. Off-chain, liquidity is premium, hence transactions are charged for the amount being transferred over the channels (as it requires rebalancing).
38g/ In other words, on-chain and off-chain transactions have different fee models that compliment each other. On-chain, the fee is constant despite transaction value, whereas off-chain the fee is priced as a percent of the value transfer. There is a crossover point.
38h/ We’ve already seen evidence that Lightning is increasing layer 1 usage, even in its highly experimental form. There was a block created in February 2019 which was 25% full with lightning transactions to open a channel. @Snyke
38i/ Quantum resistance

"The adoption of quantum resistant techniques will also result in larger (and more expensive) transactions. Post-quantum crypto algorithms require larger key sizes, which in turn increase the size of non-witness data in a transaction." @Lucasnuzzi
39/ Overall

We have empirical evidence that total fee revenue will slowly trend up to equal the block subsidy in the coming decades. Based on this data, fears that the transaction fee won’t replace the block subsidy are definitively overblown. Projections by @Awe_andWonder
40/ As Alex Sunnarborg pointed out, only Bitcoin and Ethereum have meaningful enough transaction fees to compensate miners in a no inflation environment. It is very unlikely that any other network will be able to compete.
41/ Security Stability

“The volatility of fees, which seem to behave nonlinearly as blocks become full. Might lead to corresponding big swings in hashrate.” - @szabo
42/ Scarce blockspace is a good thing since we will see a backlog of transactions, which demonstrates future intent to reward miners, which in turn stabilizes the system. Congestion in 2017 demonstrated that the system can create and sustain a backlog.
43/ A legitimate concern is that in a pure transaction fee security model, there will be volatility in cash flows. Transaction fees are market-centered, meaning that they go up and down adjusting to supply and demand.
44/ The base assumption is that cash flows from transaction fees will be unstable which makes the network less secure. Dan Mcardle has a good explanation as to why this won’t be an issue:
45/ “As mining becomes highly commoditized with mature corporations, miners are unlikely to play short-run games, but will rather choose to mine continuously. Taking this further, as miners will likely vertically integrate with other services (ex: OTC)" @robustus
46/ Given the worst case scenario where mining fees are unstable, it doesn’t actually undermine the system, it just makes settlement time longer until fees grow large enough for mining to turn back on. Entities, by necessity of time preference, would increase fees in response.
47/ The subsidy has already been incredibly volatile in real terms over the past few years and mining has remained strong and constant–even with 80% drawdowns in the value of the subsidy. This “instability” has not materially affected the network and mining.
48/ In the future, miners might auction space in future blocks in advance which could have a stabilizing effect on their revenue (aka crop futures). The basic premise is that if there is increasing usage of Bitcoin, the free market for future block space will price it correctly.
49/ Finally, if this is a major issue that isn’t corrected by the market naturally, there are minor changes we can make to smooth fee revenue. Ex:

“Other longer term low subsidy era ideas include fee averaging across block intervals to smooth fee revenue.” @adam3us
50a/ Modeling Security Post Subsidy

What do transaction fees in a post block subsidy world look like? I built a model that would help us think through what they might be in the year 2140 (in today’s dollars), with a few necessary assumptions that you can view for yourself below
50b/ In its most congested state in the year 2140, a transaction on layer 1 may only cost between $8 - $82 depending on Bitcoin’s market capitalization and scaling. Transacting on layer 1 will be an infrequent occurrence for most consumers, just like wiring money.
50c/ $10T market cap (You can play around with the model if you’d like to plug in your own assumptions) docs.google.com/spreadsheets/d…
50d/ $100T Market cap (You can play around with the model if you’d like to plug in your own assumptions) docs.google.com/spreadsheets/d…
50/ Assumptions

Bitcoin survives, thrives, and continues to grow in market share exponentially, as it has done the last 10 years. $10-$100T is reasonable:

Total wealth in the world: $750T
Fiat: $50-100T
Gold: $7.5T
Real Estate: $225T
51a/ Block size is constrained by latency, bandwidth, and storage. Another consideration with latency is usability with Tor. We have to look at the issues with larger blocks size holistically.
51b/ Certainly a larger block size would alleviate some of the explicit fee pressure felt by transactors. However, that offsets the cost onto node operators, decreasing decentralization in some manner, which is what makes Bitcoin valuable in the first place.
51c/ Also, this increases mining pool centralization. What level of centralization is allowable is a continuing debate in the Bitcoin community.
51d/ There is a case to be made for a market mechanism that would compensate node operators. For example: prompt relays, transaction data, SLAs, etc. However, it is mostly theoretical at the moment.
51e/ If a rate of block size increase is decided on, it should have a decreasing growth rate (similar to Bitcoin’s inflation rate). Otherwise we are going to have to agree to softfork a smaller limit in later, which is the exact opposite of the position we want to be in.
51f/ And finally, there have been some discussions around decreasing block time, which would provide the ability to further increase block size.
52/ Layer 1 Efficiency: With Schnorr signatures (soon to be implemented), it is estimated that this upgrade would reduce the use of storage and bandwidth by at least 25%. We can assume some additional efficiency gains will occur over the next 120 years (Taprooft/Graftroot)
53/ Security as % of market cap

We can see that transaction fees as % of market cap trend towards a little above 0.001%. Even with the continual decline of the block subsidy, total mining revenue (ie security) has increased exponentially as Bitcoin gain adoption @Awe_andWonder
54/ TPS

Assuming that current Layer 1 is 20 TPS max w/ Segwit. TPS depends on the byte size of the transaction but I had to choose a value to build the model.
55/ Flawed Potential Solutions

Proposals for increasing the 21M hard cap aren’t appropriate to bring up at this stage for three reasons:

- Lack of evidence that this will be an issue
- Divisiveness of the ask due to its ethical tradeoffs
- More palatable fixes
56/ I’ve addressed what an adequate level of security could be, Bitcoin’s prime block space, transactional demand, what happens when the subsidy runs out, and security stability. We have evidence that there will be proper security budget financing will equilibrate through fees.
57/ The trade off is always between inflation (block subsidy) and fees and Bitcoin is the best positioned to charge fees reliably. If there is no economic (transaction) volume in 10 years bitcoin will have failed anyway.
58/ Most well pedigreed critics predicting doom and gloom used absurd assumptions in their models. Conversely, Cypherpunks write code.
59/ Satoshi's experiment, Bitcoin, has worked for 10 years despite an intensely hostile environment - all data indicates we have reason to believe it will continue to thrive.
60/ If you’d like to know more about @interchangehq please feel free to fill out our intake form here picksco.typeform.com/to/PcdTmP
Missing some Tweet in this thread?
You can try to force a refresh.

Like this thread? Get email updates or save it to PDF!

Subscribe to Dan Hedl
Profile picture

Get real-time email alerts when new unrolls are available from this author!

This content may be removed anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!