Crypto Asset Investment Behaviour

Crypto assets can be seen as entirely new forms of economies whose primary institution is its blockchain which provides the reference for truth. Some of the most interesting work that has been done analyzing these economies has been research around holding behaviour – attempts to quantify the length of time over which agents hold onto a given crypto asset and how this changes. Some of the most promising work has been done on Bitcoin’s holding behaviour by Unchained Capital, Coin Metrics, Ikagi Capital, and Adaptive Capital. Our research builds on the current literature by using the UTXO Age Distribution, Realized Capitalization, and Coin Days Destroyed metrics for inter-crypto asset comparisons, as opposed to an intra-crypto asset comparison as has been the case in previous work. Doing this allows us to better understand the varying economic functions different crypto asset fulfil; for example, we can show that on-chain data supports the commonplace thesis that Bitcoin is a superior store of value to an inflationary crypto asset such as Dogecoin.

Unspent Transaction Outputs (UTXOs)

In order to properly understand the implications of the metrics, it is important to first explain how transactions work in an unspent transaction output (UTXO) crypto asset like Bitcoin. Each block in the Bitcoin blockchain consists of a series of transactions, each transaction spends the bitcoin previously received in another transaction earlier. Therefore, it is always the case that the input of one transaction is the output of a previous transaction. The diagram below helps explain how this transaction model works.

Each output of a given transaction can only be spent once, so all outputs can be classified as either unspent transaction output (UTXOs) or spent transaction outputs.

UTXO Age Distributions

Unchained Capital’s work on UTXO age distribution builds on the research done by @jratcliff and is currently the best way to visualize the changes in holding behaviour of users of crypto assets. The metric builds on the fact that Bitcoin is a transaction-centric crypto asset compared to an account-centric crypto asset like Ethereum. As such, each UTXO shares a timestamp with the transaction and block in which it was created. Moreover, all bitcoin in existence will always be contained in some UTXO – allowing us to always assign an age to each existing bitcoin which measures when it was last used in a transaction. This metric simply looks backwards and analyses the age distribution of UTXOs over time – grouping different UTXOs using age bands. Below we show the UTXO age distribution of Bitcoin as Unchained Capital did in their article:

Data: Unchained Capital

There are several interesting things to note about this graph; firstly, the presence of distinct “HODL waves” throughout time. HODL waves are periods “when a large amount of bitcoin transacts on the way up to and through a local price high, becoming recent BTC (1 day — 1 week old), and then slowly ages into each later band as its new owners HODL”. There are several distinct HODL waves and we show two below: The Genesis HODL and The Great HODL.

The Genesis HODL (January 2009 – June 2011)

Data: Unchained Capital

The Great HODL (December 2013 – December 2017)

Data: Unchained Capital

An open question is whether this phenomenon is visible in other crypto assets and if so, in what way are cross-crypto asset HODL Waves similar and different. We will explore UTXO age-distribution graphs for other crypto assets in the coming weeks.

Realized Capitalization

The Realized Capitalization (or Realized Cap for short) was formulated by the Coinmetrics team as a way to account for the large fractions of crypto assets which tend to get lost, but yet continue to be counted within the Market Capitalization calculation. Simply put, the metric can be seen as a historically-weighted UTXO market capitalization calculation for a given crypto asset. The metric becomes more meaningful as a measure to better understand the behaviour of crypto asset users when computed as a ratio against the crypto asset’s market capitalization. Below we plot the metric Market Value to Realized Value for the UTXO-based crypto assets from late 2016 onwards.

Data: Coin Metrics

The most notable thing we see by comparing Bitcoin to two other quite different crypto assets is that Bitcoin’s MVRV is generally higher than other– this is the case for all other UTXO-based crypto assets aside from Bitcoin Cash SV. The ratio can be used to understand the relationship between users of a given crypto asset with high time preferences and those with low time preferences. Bitcoin’s prominence as the most popular crypto asset and common base currency on many exchanges may help explain why its MVRV generally seems to be higher than other crypto assets.

Coin Days Destroyed

The Coin Days Destroyed metric introduced by Ikagi Capital relies on the concept of Bitcoin Days which are defined as: the quantity of bitcoin multiplied by the number of days since said bitcoin were moved. This allows us to segment circulating crypto assets into different groups based on the timeframes in which they were last moved. The Bitcoin Days Destroyed metric naturally follows from this as moving bitcoin which has been idle for some period of time, destroyed the number of days that it had accrued. The chart below plots the Bitcoin Days Destroyed metric over time.

Source: Blockchair

In order to allow for comparisons of coin days destroyed across different crypto assets, we can use the Adjusted Coin Days Destroyed metric which is simply Coin Days Destroyed divided by the crypto asset’s circulating supply.

Source: Blockchair

The metric can be understood as showing what proportion of old holdings of a given crypto asset are transferred or sold over time compared to the crypto asset’s circulating supply. Bitcoin generally had the lowest Adjusted Coin Days Destroyed value of the group of crypto assets suggesting that a smaller proportion of long-term holdings tended to be transferred for Bitcoin compared to other assets.

Conclusion

This research report has demonstrated some of the analytic tools available for comparing the holding behaviours of different crypto assets. Both Realized Cap and Coin Days Destroyed are useful for inter-crypto asset comparisons, whilst the UTXO age distribution metric better visualizes how trends in holding behaviour changes over time for a given crypto asset. All the metrics seem to suggest the fact that users seem to have lower time preferences for holding bitcoin compared to other crypto assets – a fact that was anecdotally accepted as true up until this point. We believe that more work will allow more refined comparisons of all aspects of economic behaviour on different crypto assets and this research is just the start.