What Drove the Bitcoin Flash Crash?

Bitcoin and the wider crypto market experienced a flash crash over the weekend, wiping over $400mn off the total crypto market cap. At the macro level, a combination of factors including Omicron and Fed hawkishness contributed to the moves. Meanwhile, the micro level saw substantial deleveraging and poor liquidity. Several hacks in the crypto space have also weighed on sentiment. Currently, bitcoin is trading near $50,000 and drawing down 26% from previous highs (Chart 1), the total crypto market cap sits at $2.45tn, and bitcoin dominance has fallen to 37.9%.

What’s Behind the Flash Crash?

Flash crashes are essentially dramatic drops in prices that are due to significant short-term imbalances in buy-and-sell orders. We know there was sharp deleveraging as seen by the drop in open interest in futures contracts on crypto exchanges around the flash crash (Chart 2). We also know that market liquidity was poor as the price drop occurred during the weekend when trading activity is lower. But the more challenging question is why did the flash crash occur when it did, rather than (say) a few days or weeks earlier.

We can try to understand any common factors behind crypto flash crashes by looking at earlier episodes. Starting from 2019, when bitcoin started to become more ‘mainstream’. We define a flash crash as any episode that saw an intraday price drop greater than or equal to the most recent drop (-16%).  We identify seven flash crashes (including the most recent). One in 2019, another during the initial COVID outbreak in 2020, three in May this year and then the most recent one.

Common Factors Behind Flash Crashes

Looking at how flow metrics and macro factors behaved in the 20 days leading up to each flash crash, we find the following:

• On flow metrics, we find that in the run-up to flash crashes, typically (Table 1):
  • Younger coins are sold (i.e., coin days destroyed (CDD) falls)
  • More flows on to exchanges
  • No clear pattern from miners (hash rate and miner revenue)
  • More bitcoin holders are experiencing mark-to-market losses (PSIP)
  • Number of whales (1K of bitcoin or more) is falling, but super-whales (10K+) are increasing
  • Leverage is picking up (open interest increases).

We have a small sample size, so these patterns may not be reliable. The latest episode did see earlier younger coins being sold, more market-to-market losses, super-whales increasing and more leverage. But we didn’t see more flows on to exchanges, nor the number of whales falling

• On macro factors, we find that in the run-up to flash crashes, typically (Table 2):
  • No consistent pattern with bond yields
  • Equity markets are falling
  • Dollar is weakening
  • Gold is rising

The latest flash crash did occur with weak equity markets, but the dollar was rising, not falling.

Finally, we note that after a flash crash there is no consistent pattern on subsequent trend. On half of occasions, bitcoin fell, and the other half it rose.

What Now?

Our flow metrics are continuing to give a neutral signal for bitcoin. We have three bullish signals: decreasing exchange supply, longer term HODLers maintaining conviction to hold, and hash rate closing in on new all-time highs. We have three bearish signals: ETF outflows, futures open interest plunge, and reduced profitability of the overall coin supply.

We therefore maintain a neutral view on bitcoin. Here is the full rundown of our metrics:

Institutional Demand: Bearish Bitcoin

Our preferred metric to track institutional demand is flows into bitcoin ETFs (Appendix). Throughout the correction, inflows have been decreasing, and the past week has seen renewed outflows (Chart 4). We view this as bearish for bitcoin.

News around bitcoin and crypto shows mixed momentum. Grayscale published a report that found 26% of US investors already own bitcoin, and 55% of bitcoin investors started in 2021. Facebook will no longer ban cryptocurrency companies from running ads on the social media platform. The third-largest bitcoin holder added over $150mn worth of bitcoin to their holdings, apparently ‘buying the dip’. Barry Sternlicht, the chairman of Starwood Capital Group, argued bitcoin could rise to $1mn per coin and said owning bitcoin was a ‘smart hedge’. On the other hand, crypto exchange Bitmart was hacked for $196mn and BadgerDAO – a decentralised autonomous organisation focused on bridging bitcoin with decentralised finance (DeFi) – was hacked for $120mn. Together, outflows from ETFs and mixed news momentum are bearish bitcoin.

Demand for Liquidity and Exchange Activity: Bullish Bitcoin

The past two weeks saw pockets of increased inflows onto exchanges. However, the magnitude of these inflows has been smaller than those during previous corrections. In the days immediately after the flash crash, outflows increased: 30,909 coins came off exchanges in the last two days alone (Chart 5). The percentage of total bitcoin supply held on exchanges continues to decrease. Currently, 2.4mn coins are held on exchanges (12.8% of the total supply) – down 7% year-to-date (Charts 6 and 7). Increased outflows and decreasing exchange balances suggest investors are willing to hold bitcoin in a more illiquid form without the need to easily sell – this is bullish bitcoin.

Futures Activity: Bearish Bitcoin

The big story on the flash crash was the huge deleveraging event played out on futures exchanges. Aggregate open interest has been declining on average since registering all-time highs in November (Chart 8). However, it has remained historically high for some time. November saw muted liquidations until prices began to draw down from the highs, leaving room for a large flush out were prices to drop again, which they inevitably did over the weekend. A surge of long liquidations was triggered, causing open interest to drop a staggering $5.4bn in just hours.

Perpetual futures have no expiry date, so never settle in the traditional sense. Therefore, exchanges set funding rates to prevent a lasting divergence in the price of the futures contract and the underlying. We can consequently interpret funding rates as the cost of holding bitcoin via perpetual futures. They have been decreasing on average throughout November, recently falling into negative territory (Chart 9). 

Positive funding rates imply longs pay shorts and vice versa. We can use it as a proxy for trader sentiment since a positive funding rate implies traders are paying a premium to keep open long positions. The (average) funding rate had been positive for all the previous bull run, and the flip to negative rates suggests a transition of sentiment from greed to fear. The Fear and Greed (F&G) Index also captured this. It uses various data sources to quantify emotions and sentiment around bitcoin. The index has been decreasing since prices began to capitulate in November and now signals ‘extreme fear’ (Chart 10). Historically, periods of extreme fear have presented good buying opportunities.  

HODLers: Bullish Bitcoin

This week, we have split HODLers into those that have held for greater than or equal to six months, and those for under six months. The longer-term HODLer vintage (6m+) has been steadily increasing since late July (Chart 11), whereas the shorter-term vintage (<6m) has been continually decreasing. Throughout the correction, there has been no material shift between the two vintages. This suggests longer-term HODLers have maintained their conviction to hold.

To evidence this further, we note both the coin days destroyed metric and the 1y+ revived supply metric have been falling on average throughout the correction, after peaking at the new all-time highs (Charts 12 and 13). Overall, longer-term HODLers are not spending their coins – this is bullish bitcoin.

Investor Profit and Loss: Bearish Bitcoin

Public blockchains allow us to calculate three P&L-related measures: percent supply in profit (PSIP), net unrealised profit and loss (NUPL) and the spent output profit ratio (SOPR). (The Appendix details each.)

The share of the supply in profit (PSIP) is currently 78%, down 7pp week on week (Chart 14). The size of the unrealised profits (NUPL) is currently 52% of the market cap, down 55pp week on week (Chart 15). This is the first time NUPL has tested 0.5 since the September correction – a key level to watch as dropping below this could signal further downside. SOPR dropped below one on numerous occasions throughout the crash, signifying loss taking. However, it has since rebounded to one (Chart 16).

PSIP and NUPL are at monthly lows. We expect this during a flash crash, but SOPR has also signified some realised losses, likely by late buyers at the new highs. Overall, the profitability decrease is bearish bitcoin.

Mining Activity: Bullish Bitcoin

We track the hash rate for bitcoin. A higher rate means more computing power is available to maintain the network, deliver more security (resistance to attacks), and facilitate more transactions. We view this as a bullish sign (Appendix).

Amid the correction, bitcoin’s hash rate has made solid gains and is nearing new all-time highs. It is up 12% since the start of November and 15% week on week (Chart 17). This counters the normal relationship between price and hash rate: usually they move in tandem. However, the hash rate has diverged from prices during the correction, and new all-time highs are possible before the new year. This speaks to the robustness and dynamic nature of the bitcoin network. We can say the same for miner revenue (Chart 18).

The increasing hash rate adds positive light to the negative flash-crash backdrop. And judging by price action, the market seems yet to have appreciated it. Watch the hash rate closely as it approaches new all-time highs. This is bullish bitcoin.

Bottom Line

We have introduced a framework for understanding the flow and microstructure dynamics of bitcoin markets. The key metrics are:

1. Institutional demand: ETF outflows and mixed news momentum. Bearish bitcoin.
2. Liquidity demand: outflows from exchanges/exchange supply decreasing. Bullish bitcoin.
3. Futures activity: open interest plummeting, and perpetual funding rates become negative. Bearish bitcoin.
4. HODLer behaviour: overall, longer-term HODLers are not spending. Bullish bitcoin.
5. P&L of investors: realised loses, and reduced profitability in the coin supply. Bearish bitcoin.
6. Mining activity: hash rate closing in on all-time highs. Bullish bitcoin.

On balance, the metrics are bearish bitcoin in the short term.

Appendix

Institutional Demand

Perhaps the largest institutional vehicle for bitcoin is the Grayscale Bitcoin Trust, with over $27bn in assets. It invests solely in bitcoin, and so many investors, notably institutional, who cannot hold bitcoin directly can get exposure through investing in Grayscale. Consequently, if the trust trades at a premium to bitcoin prices, it may imply ‘excess’ demand from institutions, but ‘excess’ supply if it trades at a discount. Alternatively, investors may be using other vehicles to get exposure such as ETFs or holding bitcoin directly. We put more weight on ETF flows than the Grayscale premium.

Liquidity Demand

Another measure of cryptocurrency bullishness is whether investors are willing to hold it in illiquid form (e.g., a private wallet) or prefer a liquid form (e.g., on an exchange). The former would suggest investors are bullish, as they are comfortable with being unable to sell easily. Conversely, holding it in liquid form would suggest investors are bearish, as they prefer being able to sell easily.

Therefore, large flows onto crypto exchanges would suggest investors want to convert their holdings to a more liquid form, possibly implying more bearishness.

HODLers

In our introductory bitcoin flow framework, we explained ‘HODLers’ and ‘HODLing’. HODLing refers to buy-and-hold strategies in the context of bitcoin and other cryptocurrencies. Those who HODL for long periods are die-hard adherents.

We can categorise HODLers by the length of time they have held BTC. We define long-term or staunch HODLers as those who bought BTC five or more years ago and have held it ever since, medium-term HODLers as those who bought 6-12 months ago, and short-term HODLers as those who bought 3-6 months ago. We can break this down further into those who have held bitcoin from the very early days (7-10 years ago and 10+ years ago).

Profit and Loss

• The percent supply in profit (PSIP). This tracks the share of circulating BTC supply in profit. That is the percentage of circulating BTC whose current price is higher than when it was last transacted (movement).

• Net unrealized profit and loss (NUPL). This is the ratio of unrealised profits over total market capitalisation. While PSIP just focuses on whether BTC coins are in profit or not, the NUPL focuses on the size of profits. So, we could have a situation where the PSIP is low – that is, a low share of supply is in profit – but the NUPL could be high if the size of those profits is very large.

• Spent output profit ratio (SOPR). While PSIP and NUPL focus on unrealised profits or mark-to-market, this measure focuses on realised profits. SOPR is the realised value of a transaction divided by the value at initiation (or creation) – more simply, price sold divided by price paid. If SOPR is above one, investors in aggregate have realised profits, while below one means they have realised losses. In broad uptrends, SOPR spends a significant amount of time above one, whereas the opposite is true for broad downtrends.

When SOPR is rising, sellers are increasingly realising profits. The opposite is true when it is falling. A price rally with a flatter SOPR trend indicates investors are not yet realising their profits with the rally. The reluctance of investors to sell and realise a profit may be because they believe the price will increase further, which would be bullish. At the same time, more profit taking could precede a correction. Typically, buying BTC as SOPR moves around one during bullish periods has proven to be a profitable strategy.

Mining Activity

Computing power is central to the crypto market. Miners use advanced computing hardware to solve complex problems that confirm BTC (and other coin) transactions on the public ledger or blockchain. The miners are rewarded with new coins for their efforts. A measure of the complexity of the problems and so the computing performance required to solve them is the hash rate. The higher this rate, the more computing performance is needed to maintain the blockchain. The rate can fluctuate depending on demand for crypto.

Dalvir Mandara is a Quantitative Researcher at Macro Hive. Dalvir has a BSc Mathematics and Computer Science and an MSc Mathematical Finance both from the University of Birmingham. His areas of interest are in the applications of machine learning, deep learning and alternative data for predictive modelling of financial markets.