Wednesday, September 4, 2024

Understanding Key Factors Influencing Bitcoin's Price Movements

Hi everyone, please see below for a non-comprehensive, but detailed write-up that explains several key factors / trends influencing BTC prices over the last year. Let me know if anyone has any questions, I will try to respond in the comments, and for any detailed responses or follow-ups, can always DM. 

This report is purely educational / informative and is not trying to sell you on a bullish or bearish indicator for BTC.

Hope you all enjoy!

All charts are available in the imgur link below (please refer to the link to view full report with figures / charts).
https://imgur.com/a/Zqy1yHw

PURPOSE:
(1) Discuss where we potentially are in the overall Bitcoin (BTC) “supercycle”; (2) Discuss multiple factors that are influencing Bitcoin’s price 

1 | Where Are We in the BTC Super Cycle? The Four Illustrative Stages of Bitcoin

There are theoretically four illustrative stages of BTC during a super cycle that seems to emerge every ~4 years and conveniently correspond with the BTC halving cycle. It’s a cycle that has played out several times now – most notably in 2017 and recently in 2021.

(1) Stagnation / Accumulation; (2) Ascent; (3) Euphoria; (4) Collapse 

To understand where we currently are in the super cycle (as of September 1, 2024), we can look at a simple, historical analysis of rolling averages that conceptualize Bitcoin “momentum” and categorize each stage of the BTC supercycle by a range of momentum values (i.e. the rolling average) as shown below. Momentum values above 100 represent upward momentum and values below 100 represent downward momentum. To note, the distinction of momentum ranges is entirely arbitrary and a back-wards looking analysis, but shown for illustrative purposes / for reference only.

After the last peak in 2021, 2022 saw a collapse of BTC’s price with 2023 primarily being a consolidation phase. BTC momentum in early 2024 was high but has since declined as BTC is setting lower highs since the peak in March 2024.

BTC Momentum - Current BTC Price (14-day average) / Rolling Average (90-day average)
BTC halving dates shown as light blue bars

Interestingly, based on a rolling average, the analysis implies we are currently in the “Collapse” phase relative to historical momentum trends – meaning the recent decline in BTC prices have occurred in an extended enough time period that the momentum indicator index is now below 100%.

Google Trends Data
Google Trends data have yet to show a meaningful uptick in Google searches (relative to a ~8Y lookback period). Although the Trends index for this cycle peaked in March 2024, it was a fraction of where it was at previous cycle peaks in the last two halvings. Part of the reason we aren’t seeing a strong uptick in Trends data is that Trends is mostly driven by consumer or retail interest in BTC (i.e. non-institutional), which has likely been absent in this cycle.

For context, in the past two cycles, we saw retail interest (and leverage) led to large price spikes just before the cycle peak. Even outside BTC, there are plenty examples of market bubbles driven by surging retail interest (i.e. Dot Com bubble, SPAC bubble, Cannabis bubble, etc). However, it appears retail traders have largely sat out during the rally. Whether we see robust retail interest return is unknown, but retail interest in BTC (similar to other assets) has generally increased with price (i.e. as BTC price increases, retail interest increases), and we will keep an eye out for changes in Trends data to see if this holds true.

2 | Muted Impact of ETF Inflows (BTC & ETC)

So, if it wasn’t retail interest that drove BTC’s price to an all-time high earlier this year, what could it be? One well-known theory that was perpetuated by mainstream media was BTC ETF inflows drove BTC price. To a certain extent, this seems like a fairly straightforward cause and effect – if BTC had large net inflows, the excess demand would result in upwards price movement. However, what most investors seem to forget is that inflows / outflows from these ETFs represent an extremely small portion of overall spot trading volume per day. Since mid-January, the top ETFs have collectively seen a net inflow of ~$18B, and represents just ~1.6% of the total BTC market cap.

Even when looking at this from a supply perspective and taking into account only ~10.5mm are available to the general public after excluding BTC lost forever, owned by institutions, owned by law enforcement, etc., assuming the net inflow of ~$18B of Bitcoins were purchased at an average price of $50,000 – we only get a net inflow of ~350K bitcoins or ~3.4% of the estimated total supply available to the public. Regardless of how you cut the data, the impact of BTC ETFs seems to be small / shouldn’t have an outsized impact on BTC’s price.

We can further support this theory by looking at regressions of BTC and ETH 1-day price performance and the 1-day inflow / outflow.

Refer to linked PDF for regressions

Both regressions have very low r-squared values, meaning that the data points don’t fit the trendline very well and any correlation between ETF inflows and price movement across BTC and ETH seems to be mostly random.  

The counter argument to this could be the fact that ~68-76% of BTC accounts have been inactive for 1+ year, indicating the majority of BTC holders are longer-term holders that wouldn’t participate in buying/selling to institutional funds. So, the ~350K BTCs that were bought as part of the new demand created by the BTC ETFs has the potential to account for ~10% of all “active” holdings. However, this percentage is still relatively small and also represents an optimistic upper-bound vs. the lower bound being 2-3% as discussed earlier. Ultimately, the view that ETF inflows drove BTC price higher / is correlated is likely incorrect.

3 | Do Liquidation Events Produce Outsized Effects?

For context, there are two types of liquidation events: short and long. In a short liquidation, investors who are shorting BTC (betting the price will decline) are forced to close their position as the price of BTC increases. Rather than let the bet continue and potentially result in infinite losses, the position is liquidated. In the traditional equity market, this is known as a short squeeze. While short liquidations are primarily a function of Bitcoin’s price and not a driver (i.e. a short liquidation is the result of Bitcoin’s price increasing that force shorts to close positions), it can result in a potential flywheel effect where price increase drives both short liquidation and buying pressure, which further increases the price of the asset, increasing short liquidations, and so on.

The following series of charts explain:
(1) Potential correlations between short, long and net liquidations vs. BTC price changes
(2) Examining the outsized role net liquidation events have on overall BTC movement

Figure 3A: Short only liquidations in the past year (09/01/2023 – 09/01/2024)
Figure 3B: Long only liquidations in the past year (09/01/2023 – 09/01/2024)

Refer to linked PDF for charts

Figures 3A and 3B shows that there is a slight r-squared significance between short only and long only liquidation events and BTC price movement, although a r-squared of ~0.2 is relatively low. This is mainly because short and long liquidations should not be looked at in isolation, and a more significant metric to analyze is net liquidations.

Figure 3C: Net liquidations in the past year (09/01/2023 – 09/01/2024)

Refer to linked PDF for charts

Figure 3C shows a dramatically improved r-squared, suggesting the net liquidation data fit the linear trendline relatively well. While this doesn’t imply a cause-and-effect relationship between these two variables, it does help contextualize that BTC price movements and net liquidation movements occur mostly in the same direction and magnitude.

Figure 3D: Net liquidations above $50mm in the past year (09/01/2023 – 09/01/2024)

Refer to linked PDF for charts

Figure 3D shows an even higher r-squared value when narrowing the data set for only large net liquidation events, defined as >$50mm in either direction.

In addition to looking at the relationship between liquidation events vs. BTC price performance (%), we can look at the dollar amount of BTC price movement around each liquidation event. We find that large, short only liquidation events account for 23% of all upwards Bitcoin price movements in the past year. Similarly, large, long only liquidation events account for 37% of all downwards Bitcoin price movements during this period. While accounting for ~30% of price movement, these large liquidation events only represent ~18% of the trading days in the past year. In terms of large, net liquidation events, there is a similar trend and on average, each net liquidation events have been associated with a price move of ~$3k (5%).

Figure 3E: Net liquidations above $50mm in the past year vs. BTC performance in next trading day (09/01/2023 – 09/01/2024)

Refer to linked PDF for charts

It's important to remember, while this data appears highly correlatory, we cannot make any further conclusions (i.e. claim that BTC price is a result of liquidation events) as there’s simply no data to support that argument. What we can try to do is see if there are any tail effects that occur post-liquidation. That is, does BTC price continue to be driven higher post a short liquidation, or vice-versa for a long liquidation. As we see from Figure 3E above, it appears that we cannot extrapolate large, liquidation events into the next trading day. 

4 | Halving Cycles

The saying “buy the rumor, sell the news” appears to fall apart when looking at halving cycles, which are a known phenomenon that occurs every ~4 years where the supply of new bitcoin becomes more limited as the reward for miners (“suppliers”) decreases therefore making supply more inelastic. If markets were highly efficient and incorporated this known data, price increases from this supply-side shock would occur prior to the halving date, and Bitcoin’s price after this date would lag the preceding period. However, as we see in the charts below, that thesis isn’t true.

 2016 Halving Cycle
Halving Date = 7/16/16
Peak = $2,958 on 6/11/17 (337 days after the halving)
Percent Increase Prior to Halving (1-year): 128% | Percent Increase After to Halving (1-year): 287%
Refer to linked PDF for chart

2020 Halving Cycle
Halving Date = 5/11/20
Peak = $65,503 on 4/13/21 (337 days after the halving)
Percent Increase Prior to Halving (1-year): 23% | Percent Increase After to Halving (1-year): 559%
Refer to linked PDF for chart

Creative Data Can Tell a Powerful Story
For such a widely known / predictable event, it seems strange that the results are relatively consistent where the price increase leading up to the halving is significantly outpaced by the price performance following the halving. However, a 365-day lookback / lookforward from the halving date represents a 2-year period in total, adding significant noise around the true relevance of the halving impact. While it could be that there is a delayed halving effect on supply, it’s not as clear as articles online make it out to be.

The below data looks at price performance in intervals of 25 days. The 2016 halving cycle breaks the trend shown above, where price performance before the halving outperformed the same period after the halving for the majority of all periods. In 2020, the opposite was true. However, remember there were multiple rounds of stimulus checks that started in April 2020, one month prior to the Bitcoin halving. BTC is largely thought to have appreciated in late 2020 and early 2021 due to Americans buying crypto with excess stimulus money, making the 2020 cycle unique and likely not comparable to the 2024 halving cycle.

Refer to linked PDF for table

If you are wondering what the data would be if we ran a 1-day increment period, 266 of 365 periods (or ~73%) in the 2016 halving cycle had price performance before the halving greater than after the halving, consistent with the data shown above. The purpose of showing this is to highlight that timing plays an important role when purchasing BTC, and depending on when you purchase you can either underperform or outperform others. To counter this effect, one could pursue a dollar-cost average (DCA) strategy that (at least for traditional equities) tends to outperform over a long investment horizon vs. buying/selling and trying to time the market.

Notes & Sources
See linked PDF

General Note
This report is focused on analyzing trends with no bullish / bearish implications on BTC; more generally, it’s impossible to determine the direction BTC will go in the near-term

DISCLAIMER: The content below is exclusively based on publicly available information and presents opinions that are also formed on public materials. This is not investment advice and shouldn’t be considered as any form of advice. The figures / charts mentioned below are for illustrative purposes only and may contain errors. The information below should not be used for any investment decision and is shared with you for educational and recreational purposes only*. No buy or sell recommendations have ever, or will ever be provided in these reports. For disclosure purposes, I will provide ownership information if I own any of the assets mentioned in the report. This disclosure does not constitute an investment recommendation and should not be interpreted as such.* 

Disclosure: Own BTC: Yes; Own ETH: No


Should I worry about this?