Examining Censorship Resistance and Mining Concentration Across Public Blockchains
A few weeks ago I published a post about Analysing the Costs and Benefits of Public Blockchains, this post makes more sense if you read that one first. But for those of you who prefer the TL;DR, I argued that
The cost to maintain and secure a blockchain has a direct, and positive correlation with the ability of a network to secure against real or perceived censorship, and;
Economically rational uses for public blockchain technology will be confined to instances where the real or perceived threat of censorship by external actors justifies the transactional cost required to maintain and secure a blockchain.
The bottom line is that it turns out that the functional operations that Bitcoin or Ethereum offer could be run for ~$100–200 per month on Amazon AWS, whereas the networks effectively required $275,000,000-$400,000,000, respectively, per month in March 2018. What’s the reason for this — not inconsequential— difference? You ask. Well essentially, you’re paying strangers on the internet to not stop or reverse your transaction (presumably because you don’t trust Amazon) by performing a function called mining. I’m not going to get back into details on this, as there are many people who can explain it better than I (you could also read my post), but essentially the more mining is spread across independent parties the less likely it is that your transaction will be censored. As a result you should be ready to pay more at times when mining is more decentralised, and in networks that have more decentralised mining. In my last post I showed a Herfindal-Hirschmann Index (HHI) for Bitcoin mining on a 2016 block (~2 week) basis versus the price of Bitcoin, for reference:
As a reminder, the lower the number in the HHI the more evenly the distribution of mining.
Of course this should not only be reflected in the price of the currency itself but in the cost of making a transaction, which again was shown in the post and looked like this for HHI on a 1008 block (~1 week) basis:
My takeaway in the post was that
periods of lower concentration of mining have tended to coincide with higher aggregate transaction fees being paid to confirm transactions.
Whilst, I absolutely acknowledge that correlation ≠ causation, and that other forces are at play (including expanding the monetary base through seigniorage of new BTC) it is interesting to see.
All fascinating of course, but this got me to thinking, how does this compare to other public blockchains? So I painstakingly pulled the same dataset for Ethereum (ETH), likely violating T&Cs for several services in the process; as well as for Bitcoin Cash (BCH) (which was much easier, thanks again Blockchair!). I’ve decided to only focus on the relation between prices and mining concentration in this post, but stay tuned for more on the cost of transactions across different public blockchains in the future.
Because I was also not sure of the veracity of comparing HHI across differing periods, and in particular between BTC+BCH and ETH, I opted to look at everything over a fixed period of one day (~144 blocks for BTC and BCH, and ~2900–6300 for ETH). If anyone reading this is aware of any work done on this I would be grateful if you could share that. I also will preface comparisons between chains with the caveat that the identity of miners comes from blockchair for BTC and BCH and etherscan for ETH, which offered varying granularity and consistency. This makes a 1:1 comparison between ETH and the others less reliable at a fixed time, but should not produce a notable difference when comparing within ETH over time. This said, I still feel it worthwhile to look at the comparison of mining concentration and relative price, in particular between BTC and ETH.
First things first, restating the price and the HHI of BTC on a daily basis since Jan 2013 looks visually very similar to longer periods (albeit more reactive to changes), and can be seen below:
Shifting to Bitcoin Cash (BCH), from the split on 1 August 2017, let’s start with similarities. Although time frames were massively compressed compared to BTC, the first few days of BCH mining were extremely centralised with the first few days and weeks arguably being subject to the whims of a handful of miners. As the price of BCH began to rise in late 2017, the distribution of mining began to occur. Presumably would-be BCH miners also needed time to retool from BTC mining to join the BCH chain. Current levels of mining centralisation as measured by HHI still flirt with, or even occasionally exceed 0.25, which is the same as saying that one miner (or mining pool) receives 50% of the blocks in one day. I’ll pause here to reiterate that I am not taking a view on the technical decisions of either BTC or BCH, but am looking at price and the distribution of mining, I have friends who are strong supporters on either side and I respect their reasons and philosophies for supporting their view on the technical decisions. Given the level of mining concentration in BCH, when compared to BTC, which otherwise has many economic similarities (e.g., number of coins in circulation, emission schedule, etc…), it could be argued that the price of BCH is currently overvalued. Whilst this is overly simplistic and ignores many important valuation aspects, including the number of BCH coins which are not in free circulation due to having not been accessed since the chain split in August 2017 (which in itself deserves further discussion), it is somewhat counter intuitive that the market valued the cost of censorship resistance offered by BCH in March 2018 at $57 million (but alas, that is just my humble opinion).
Now, if we line BTC and BCH miner concentration measures up next to each other with their relative market caps (I use relative market caps rather than price to account for a difference in the number of units in the chain), we can see that the drop and subsequent stabilisation of the HHI around December 2017/January 2018 proceeded a spike in the relative value of BCH.
We saw a similar anti-correlation in October/November with the relative value of BCH dropping and BCH mining concentrating, which could be correlated with the relative profitability of each chain on that given day. I have heard anecdotes of “die-hard” BCH miners who would continue to follow the BCH chain despite the relative profitability, which may go some way towards explaining the relative richness of the BCH price discussed previously.
Moving our attention to Ethereum (ETH), we can see some interesting differences from both BTC and BCH. Before delving in though, let’s again mention that the identification of miners (pools) comes from a different source and may be more or less accurate than that of BTC and BCH. It’s also worth noting that whilst BTC and BCH have fairly similar economic structures, ETH is very different. For this post I have ignored uncle rewards, and the dramatic difference in block times (10 minutes vs 15 seconds). The results can be seen below:
What I first noticed about this chart was the relative positive correlation between the HHI and the price. This is very different from what we saw in BTC and BCH. In fact, very noticeably in 1Q2016 as the price moved from $1-$10 and miner concentration came very close to the critical 0.25 level.
Drilling in on that period, I found that DwarfPool, one of the first large pools to support mining on ETH came very close to achieving 50% of the blocks mining, moving back down towards the middle of 2016. Let’s hang on there for a moment, at the time when a single mining pool controlled nearly 50% of the mining, the price of ETH gained 10x, this seems at odds with the maxims of decentralisation and censorship resistance.
If we lay this out as a X-Y plot, we can see that there is no strong relationship between the price of ETH and the concentration of miners.
Contrast this for a moment with the same graph for BTC, below:
Again, while not perfect, we can general see a trend towards higher prices during periods when miner concentration is low.
If we go quickly back to BCH, we can also see this general trend of more even distribution of mining occurring during periods when prices are higher.
Now comparing ETH and BTC, and with the previous caveats in mind, we can see that ETH and BTC concentration largely tracked each other since the launch of the ETH network in 2015, below:
Given that ETH mining and BTC mining are not easily interchangeable substitutes (compared to BTC and BCH mining), this might suggest that ETH prices and mining, a) may not be as strongly driven by censorship resistance, and b) may historically have been more correlated with BTC price rises.
In conclusion, I will again reiterate that correlation does not mean causation, and that these datasets are not perfectly comparable but do give some interesting insight into the way that a market values cryptocurrencies in relation to their censorship resistance at a mining level. Of course censorship resistance has many other facets, including on more qualitative measures such as the influence of community leaders, and governance of the client code base. While these are beyond the scope of this post, it is an interesting thought experiment to consider whether these qualitative measures will demonstrably change over timeto favour greater distribution of control if the market does not respond to changes in the quantitative measure, most fundamentally in miner decentralisation…
Vini, vidi, FUDi.
