An article exploring the trends in Bitcoin network security budgets

Written by: Paul Sztorc
Translated & Proofread by: 安仔Clint & 阿剑

This article aims to explore Bitcoin's ability to withstand 51% attacks (i.e., Bitcoin's "security budget"). The intense competition among miners prevents the system from collecting transaction fees sufficient to maintain a security budget from a single network, so we should derive transaction fees from all payment markets.

This article is a continuation of the discussions on "Two Types of Block Space Demand" and "Built on Bitcoin," analyzing the positions in a more empirical manner compared to previous writings.

1. What is the "Security Budget"

Bitcoin's "security budget" refers to the funds we pay to miners each year (in other words, the money spent on mining each year, which is essentially the same thing) (Note: In accounting, the two are not equal, but in economics, they are equivalent). When the security budget is low, a 51% attack can be initiated at a very low cost. Bitcoin's security budget in 2018 was approximately $7 million per day. From this, we can infer that the annual maintenance cost of the Bitcoin system (the never-ending arms race of 51% attacks) will not exceed $2.6 billion.

In comparison, the U.S. military's defense budget in 2017 was about $590 billion, while the FED's annual operating expenses reached $5.7 billion. By comparison, a $2.6 billion security budget is still too low.

2. Block Rewards

Fortunately, we can hope for an increase in the value of block rewards to provide us with greater security. Although the block reward halves every four years (equivalent to a yearly decrease by a factor of 0.84), as long as Bitcoin appreciates by more than 19% annually, it will be enough to offset the impact of the halving on the block reward. From a historical perspective, Bitcoin's average annual appreciation has far exceeded 19% (reaching as high as 70%), so it is certain that Bitcoin's security budget can continue to grow over time.

Of course, Bitcoin's price will eventually stop increasing. Although Bitcoin is designed quite successfully, it is likely to stagnate at an annual value growth of 1.077 [1], which is the nominal value growth rate of all currencies in the world today.

The table below shows the growth trend of the security budget and its eventual decline after reaching a turning point:

The above table depicts the trend of Bitcoin's security budget changes, with each row representing different years. The theoretical maximum exchange rate limit is derived from the Game and Watch paper. The estimated market exchange rates in the table are calculated from historical data and growth factors, with some additional adjustments made to reach the theoretical limit more quickly. The U.S. defense budget column is estimated from Wikipedia data. The "security ratio" refers to the proportion of the defense budget that can successfully destroy Bitcoin. All data is expressed in nominal U.S. dollars.

The "indifference period" refers to a time when Bitcoin is vulnerable, but only a small number of attackers are willing to make the effort to attack, while the vast majority of power is indifferent to this new phenomenon. The "healthy period" refers to a time when Bitcoin, having grown stronger, can withstand malicious attacks from large conglomerates. Finally, the "extinction period" refers to a bleak ending, where Bitcoin will ultimately be sunk by easily initiated 51% attacks.

3. Transaction Fees

I. The Ideal "Transaction Fee Pressure"

As everyone knows, the design of transaction fees is intended to solve this problem. As Greg Maxwell emphasized: Transaction fee pressure is an intentional part of the system's design and is currently the best solution, which is crucial for the long-term survival of the Bitcoin system.

He later proposed a famous viewpoint: From my personal perspective, I truly believe that market activity can provide sufficient fees to maintain security without relying on inflation.

Such a viewpoint (the demand for "transaction fee pressure") is quite common. Roger Ver has quoted similar views from other Bitcoin researchers, although his intention was to politically discredit dissenters, the quoted viewpoints are still quite accurate.

II. Duality

Since Bitcoin's inception, its duality (as a unit of currency and as a means of payment) has puzzled many.

Typically, monetarists and economists tend to overlook Bitcoin's payment attributes (criticizing Bitcoin for lacking "intrinsic value"). Merchants and bankers often ignore Bitcoin's currency unit attributes (thinking that buying Bitcoin is an incredibly naive investment), and then adopt a mindset of cashing in on the "blockchain" concept.

The duality dilemma of Bitcoin is also reflected in the current "scalability debate," where discussions focus on whether the "medium of exchange" is more important or the "store of value" use case is more significant.

At the same time, I believe that the duality dilemma of Bitcoin still exists in the long-term security budget analysis. Consider the following table:

| Source of Revenue | Block Reward (12.5 BTC) | Transaction Fee | |-------------------|---------------------|---------------------| | Resource Unit | … BTC | … block space | | Price Unit | … dollar purchasing power per BTC | … dollar purchasing power per byte of block space | | If Bitcoin Price Skyrockets… | … security budget rises | … security budget remains unchanged | | Slogan | Store of value | Medium of exchange | | Catchphrase | "Digital Gold" | "P2P Electronic Cash" |

Although both block rewards and transaction fees belong to the "security budget," they can be said to be two entirely different things. The difference between block rewards and transaction fees is like the difference between "VISA's total profit in 2017" and "the total growth of M2 in 2017." VISA's profit reflects the value of the superior service it provides to its users compared to its competitors (MasterCard, ACH, Western Union, etc.), while the change in M2 reflects other factors, including election results, public opinion, business cycles, federal policies, and so on. Comparing M2 with the yen may not be too outlandish, but comparing it with MasterCard would be quite absurd.

III. Is it Really "Using Bitcoin" to Pay Transaction Fees?

Transaction fees are priced in Bitcoin. However, unlike block rewards, the amount of fees fluctuates with Bitcoin's exchange rate. As Bitcoin's exchange rate rises, the current satoshis/byte pricing for transactions becomes excessively high, and users are unwilling to pay transaction fees at this pricing (leading to a nominal price reduction in satoshis/byte).

Thus, transaction fees are not truly "priced in Bitcoin," even though the Bitcoin protocol has consistently tried to mislead users into paying transaction fees based on this logic, market behavior has not developed in this way. Transaction fees are actually priced based on purchasing power, and currently (during the booming period before Bitcoinization) can be said to be reflected in U.S. dollars.

Therefore, for example, the statement "In December 2017, Bitcoin's transaction fee was equivalent to $28" is not problematic. Conversely, if one says that Bitcoin's transaction fee "reached 0.0015 BTC," that indicates a misunderstanding of the true pricing rules of current transaction fees. Because if Bitcoin's price appreciates tenfold [2], the transaction fee would subsequently drop to 0.00015 BTC.

IV. Incentive Products

Whenever the price of goods rises, entrepreneurs will find ways to increase production. (Holders will always find ways to transact, but this aspect is not within the scope of this article.)

It is well known that Bitcoin's supply cap is set at 21 million. Currently, Bitcoin's production rate (i.e.,) is 12.5 BTC per block, and this rate will continue until the next halving.

In addition to Bitcoin, another commodity, "bytes in Bitcoin blocks," also has a cap. Initially, the block size was limited to the widely discussed 1MB, and it is now set to approximately 2.3 MB per block.

As mentioned earlier: as blocks become more valuable, entrepreneurs will find ways to produce more blocks.

One method is to restart old mining machines with lower marginal rents. With the addition of computational power, the block production rate can temporarily accelerate. Of course, as new difficulty adjustments occur, the block output rate will revert to normal (one block every 10 minutes). As another avenue, entrepreneurs may mine altcoins.

V. Altcoins as Substitutes

Altcoins are poor substitutes for Bitcoin. Every form of currency is inevitably in a life-and-death competition with other currencies: currency has strong network effects; recognizable attributes promote the widespread use of currency in a super-linear manner; token conversions caused by exchange rates during transactions can be quite inconvenient. Therefore, what people truly need is Bitcoin, and everyone wants to throw other currencies into the dustbin of history!

However, considering the high prices of transaction fees and "bytes in Bitcoin blocks," a subtle twist emerges: the block space of altcoins can be considered a good substitute for Bitcoin's block space. However, this demand does not prevent people from continuing to pursue Bitcoin; only a small number of users are willing to use Bitcoin as a means of payment when making purchases. This image from the 2013 FINCEN congressional inquiry may clarify this point:

The coins sent through blockchain payment methods can always be adjusted; people may send LTC worth "20 dollars"; they may also send Dogecoin worth "1 BTC"; or EOS worth "the price of a sandwich." With the help of "exchanges" (e.g., Coinbase, ShapeShift, SideShift, BitPay, LocalBitcoins, multi-currency wallets, CC ATMs, etc.), people can easily use different forms of currency for payments.

Furthermore, even those hardcore Bitcoin maximalists occasionally clearly acknowledge this (real) hypothesis [3], that altcoins' payment methods are substitutes for Bitcoin's payment methods. At the end of 2017, when transaction fees skyrocketed, many people discussed this viewpoint:

• Samson Mao

• Francis Pouliot

• "Digital currencies as payment methods"

VI. Demand for Competitive Payment Methods

Currently, the "transaction fee pressure" that should play a core role has deviated.

The following chart shows transaction fees priced in Bitcoin (source from this webpage):

This chart shows transaction fees priced in U.S. dollars (source from this webpage):

We can see that the effect of transaction fee pressure is fragmented. Currently, a typical Bitcoin transaction costs about 30 to 40 cents—much cheaper than a VISA transaction.

Comparing historical data, using a 90-day moving average…

… the following two charts compare:

We can see that Bitcoin crossed the "1 dollar transaction fee red line" in May 2017, a trend that coincided with the rise of altcoins. We can also see the rapid disappearance of "transaction fee pressure" at the end of 2017, followed by a slight reappearance. Ultimately, it can be observed that this release of pressure coincided with an unprecedented drop in transaction volume on the Bitcoin network.

For me, the above data proves that the theory that users are willing to pay high Bitcoin transaction fees is incorrect. In fact, they have only reluctantly paid high transaction fees, and only during a brief bubble period (when people feared missing out).

If the theory that people would pay high Bitcoin transaction fees is falsified, then, priced in U.S. dollars, future transaction fees will not be significantly higher than today's transaction fees.

According to data from blockchain.info, the total transaction fees over the past 12 months amounted to $70 million. (12 months ago, this figure was $770 million)

Looking back at the charts above, the total amount of transaction fees is negligible. After all, a total of $70 million in transaction fees, when expressed in the units of the above chart (billions), is only 0.07.

If users are sensitive to the price of transaction fees and only willing to pay as little as possible, how can we increase the total transaction fees?

VII. Other Sources of Transaction Fees

a. Lightning Network

The Lightning Network (if successfully implemented) can pack many "real-world transactions" into two on-chain transactions.

The direct impact of this achievement is to lower on-chain transaction fees, but from a long-term perspective, it can actually increase the total amount of transaction fees. The Lightning Network increases the utility of each on-chain transaction (one on-chain transaction completes the work of multiple transactions), making every user willing to pay more transaction fees for on-chain payments.

So how high can the total transaction fees of the Lightning Network go?

Currently, any viewpoint is merely speculation. I believe that the Lightning Network cannot increase the total transaction fees by more than two orders of magnitude.

First, it still requires creating on-chain transactions and periodically maintaining the Lightning Network. Therefore, Lightning Network users who still need to pay on-chain transaction fees will find ways to reduce this expenditure. At the same time, altcoins can completely establish their own Lightning Networks (after all, copying the Lightning Network is no different from copying other code). The competition between these Lightning Networks is essentially the same as the competition between different blockchains.

It should be noted that, by definition, the transaction fees paid to Lightning Network relay nodes [4] are not paid to miners. Therefore, the envisioned scenario of multiple Lightning Network transaction fees "aggregating" into a single large on-chain transaction fee will not occur (because it does not align with economic efficiency).

Second, the user experience of the Lightning Network will always be inferior to that of on-chain users. The Lightning Network requires interaction, meaning each user must be online and perform operations such as [signing transactions] to receive payments. This also means that partners in the Lightning Network may cause inconveniences (e.g., they may go offline and not respond, or their computer may literally catch fire), even directly disrupting the user's normal usage. Although the clever design of the Lightning Network minimizes the risks introduced, it ultimately still affects the user experience. Naturally, users who find it troublesome are likely to shift from Lightning Network transactions on the main chain to on-chain transactions using altcoins.

b. Merged Mining of Sidechains

Merged mining sidechains can fulfill all the functions of altcoins without requiring the purchase of new tokens. Therefore, the exchange rate risk under this scheme is extremely low and can provide a good user experience.

In addition, merged mining of sidechains will send all received transaction fees to Bitcoin miners. Using merged blind mining technology, this process does not require users or miners to run additional sidechain node software.

A set of sidechains composed of large blocks can handle a large number of transactions. In the next section, I will assume that a certain sidechain network completely replaces VISA (only VISA) and thus receives all of VISA's transaction fee profits. VISA is just a small part of the entire payment market (including checks, Western Union, Apple Pay, etc.), but it can illustrate the point.

VIII. VISA Transaction Fee Revenue

I initially thought that VISA primarily generated revenue through the interest on funds paid by users, but upon researching relevant information, the reality is quite different.

Page 40 of their recent annual report shows:

Our operating profit primarily comes from the use of VISA payment methods when users purchase products or services using VISA, that is, from the transactions processed within the VISA network. We do not, and cannot, sacrifice our credit to earn interest or profits from the funds placed by users in VISA products.

Thus, VISA's main profit still comes from transaction fees. This makes the following comparison quite logical.

VISA's total profit grew from $11.778 billion in 2013 to $18.538 billion in 2017, with a growth rate of about 12% per year.

Assuming this momentum continues, we can obtain the following chart:

Click the link to view the Excel table.

The above table adds a new column to the "security budget table" mentioned earlier: VISA transaction fees. These transaction fees, combined with the basic block rewards, yield the final new security budget.

This security budget is more secure from a long-term and overall perspective.

Conclusion

Withstanding 51% attacks requires Bitcoin to have a high "security budget." Today's transaction fee revenue is far from sufficient; we must ensure that the total transaction fees in the future can increase to a level that can cover the "security budget."

While raising prices (for example, increasing the satoshis/byte transaction fee rate) can somewhat increase revenue, unfortunately, due to the competition among various chains, blindly raising transaction fee rates will only provoke a market backlash, achieving the opposite effect.

A better approach is to swallow the entire payment market and capture all transaction fee revenues. By utilizing merged mining of sidechains, this goal can be achieved without compromising decentralization.

Footnotes

[1]. This figure is calculated using 1.077 = (25.94/5.85)\^(1/20). It can be seen that 1.077 is lower than the required "stagnation rate" of 1.19.
[2]. I mean if the price of USD/BTC directly skyrocketed more than tenfold during the "bubble" at the end of 2017. In other words, if Bitcoin was only $9,000 in January 2017 and then directly rose to $190,000.
[3]. I clearly remember many examples, but (before I gave up) I only retrieved the following. If you find or remember other examples, please be sure to message me. If there are truly no more examples, I will eventually delete this section.
[4]. "Transaction fees paid to Lightning Network relays," here refers to the fees users (off-chain) pay to Lightning Network nodes that maintain payment routes in the Lightning Network.