Back to Research
Last Updated:  
October 22, 2024
8 min read

Bitcoin's Macro Correlations

Bitcoin has long been touted as digital gold. However, despite both rallying to all-time highs in 2024, the correlation of BTC returns has been closer to US equities than gold. Now, following the first in the Fed's cycle of interest rate cuts, we re-examine the macro drivers of BTC spot returns and suggest a reason for BTC and gold's underperforming correlation that may have been resolved over the summer of 2024.

Returns and Surprises in CPI

As we have previously highlighted, Bitcoin has become increasingly correlated with macroeconomic assets. The relationship has oscillated since the outbreak of the Covid pandemic, and has grown stronger after the launch of the BTC Spot ETFs in January 2024. Since then, macro has played a significant role in BTC’s price trajectory, range bound activity, and recent breakout. That much is clear from its relationship with the SPX at times close to the release of macroeconomic data that has caused market participants to revise their expectations of monetary policy.

Figure 1.  Linear regression of BTC (orange, R^2=4%) and SPX (red, R^2=3%) returns on CPI release day onto the amount by which the CPI figures surprised relative to the median estimate by Bloomberg economists. Source: Bloomberg, Block Scholes

The chart above measures this by plotting the relationship between surprises (the difference between the actual headline CPI figure year-over-year, and the Bloomberg median forecast) in the U.S. Consumer Price Index (CPI) reports, a key macro data release, and the same-day returns of the S&P 500 index and BTC.

The negative slope indicates a negative relationship: when CPI surprises to the upside (when actual inflation is faster than expected), BTC tends to drop in price over the next day. This is potentially because a higher-than-expected inflation figure raises the forecast for how long restrictive monetary policy will last, as was the case from March 2022 and July 2023. The S&P 500 exhibits a less volatile reaction than BTC, but a stronger (higher R^2) statistical relationship.

While a single, static regression over the entire time period already showcases a moderately strong negative relationship, this analysis ignores the effect of changing market dynamics. To capture this, we perform a regression on a rolling window of 24 months of CPI surprises and their impact on BTC returns over the same day.

Figure 2. Coefficient (blue) and R^2 (red) of rolling 24-month linear regression of BTC returns on CPI release day onto the amount by which the CPI figures surprised relative to the median estimate by Bloomberg economists. Source: Bloomberg, Block Scholes

Between 2020 and early 2022, the strength of the sensitivity (as measured by the R^2 of the regression) of BTC returns to CPI surprises was low, indicating that BTC returns on the day of the CPI report release had little statistical relationship to the number that was announced. However, from mid-2022 onwards, we see the CPI surprise begin to increase its explanatory power of same-day BTC returns.

Since that date in mid-2022, we have observed a negative correlation between CPI surprises and same-day BTC returns. Larger-than-expected CPI figures have tended to result in a drop in BTC price while downside surprises have led, on average, to a price increase (higher possibility of accommodative monetary policy to boost liquidity). At the peak of the relationship, over 20% of the variance in BTC’s inter-day returns on the day of the release were captured by surprises in CPI.

The S&P 500 again exhibits a similar trend with an even higher R^2, suggesting that (at the relationship’s peak) CPI surprises explained as much as 40% of the variation in SPX returns on release day.

Figure 3. Coefficient (blue) and R^2 (red) of rolling 24-month linear regression of SPX returns on CPI release day onto the amount by which the CPI figures surprised relative to the median estimate by Bloomberg economists. Source: Bloomberg, Block Scholes

The strongest impact of CPI on both equities and BTC was recorded in late-2023, after the Federal Reserve's final hike in July 2023. Markets had by then turned their attention towards signals that the Fed would begin their cycle of cutting interest rates, such as signs of weakness in the labour market and growth in GDP. The importance of CPI was relegated slightly as the Fed gained further confidence that disinflation was well underway.

This has recently been amplified by the pivot in the Fed’s monetary policy, confirmed in Jerome Powell’s Jackson Hole speech at the end of August. Powell stated “the balance of risks to our two mandates has changed” and so “the time has come for policy to adjust”. This was an endorsement that rate cuts were now imminent as inflation seemed steadily on path to the 2% target, but the labour market had “unmistakably” begun to cool down.

BTC Spot ETFs and Institutional Demand

While macroeconomic releases like inflation figures have had an increased impact on price action throughout this cycle, the release of spot BTC ETFs in January has allowed institutional investors to express their macro views directly on BTC for the first time. This includes risk appetite driven by expectations for the future path of monetary policy. As a result, flows into and out of spot ETFs have become a key indicator of institutional demand for bitcoins.

Figure 4. Linear regression of BTC (orange, R^2=14%) onto the same-day net inflows into US Spot BTC ETFs. Source: Farside Investors, Block Scholes

We see evidence of that indicator of demand holding a positive correlation to price action too. Net daily inflows to spot BTC ETFs are an indicator of positive same-day returns. While the reverse relationship also appears to hold (net outflows correlate with negative returns) the relationship is not quite as strong. Confirming this as a statistical pattern is slightly more difficult however, as we have only 10 months worth of data since the listing of the ETFs in January, and we find evidence that the correlation has varied over time.

Figure 5.  BTC spot price (orange) and net inflows to US Spot BTC ETFs (red), with estimated daily supply by miners in dollars (white). Source: Bloomberg, Farside Investors

The major rally that BTC and crypto enjoyed in Q1 of 2024 corresponded with very strong net inflows to ETFs as their release opened BTC up to a new wave of institutional demand. When these inflows abruptly stopped, so did the rally in price and BTC has traded in a relatively tight range since. The abrupt halt in the pace of capital going into ETFs was likely due to demand saturating – those institutional investors who wanted exposure to BTC did so around launch, and no new demand was driving up spot price.

Since then, the relationship between ETF flows and returns has not been quite so clear – as captured by regressing BTC spot price returns onto daily ETF flows on a rolling window. We observe significant periods where ETF flows explain almost none of the variability in spot price returns.

Figure 6. Coefficient (blue) and R^2 (red) of rolling 24-month linear regression of BTC returns on same-day net inflows to US Spot BTC ETFs. Source: Farside Investors, Block Scholes

However, the R^2 of this regression has grown steadily since mid-July, indicating that once again ETF demand is a good explainer of daily spot price returns. Daily BTC Spot ETF inflows have reached their highest since July.

According to the estimated linear relationship that we model here, each daily inflow of $1BN would result in an extra 7.2% increase in same-day BTC returns on average. We can interpret this via the supply and demand dynamics of bitcoin. Daily ETF flows are frequently 10-15x greater than the amount of bitcoins produced by mining in one day. Since the halving of the block reward in mid-April, increased institutional demand has faced an even tighter supply squeeze – and the rolling regression shows that the impact of extra ETF inflows is stronger now than it was after the launch of the ETFs in January.

China, Gold, and Bitcoin

Bitcoin is often touted as digital gold. However, in the early part of 2024, the daily returns of the two assets held very little correlation despite both rallying to all-time highs. However, both have rallied strongly in late September following the Fed’s decision to cut the FFR by 50bps on Sep 18th.

Figure 7. BTC price vs gold price since July 2023. Source: Bloomberg

Historically, the upward momentum in gold prices has moved in lockstep with ETF gold accumulation. Divergences away from this relationship have historically been led by increases in the price of gold that later resulted in an increase in the amount of gold held by ETFs.

However, over the past two years this relationship has changed. ETFs have been net selling gold, while gold prices have simultaneously rallied, surging past previous all time highs and hitting several new highs. This suggests there may be other parties involved in accumulating gold during this period, offering demand that can account for the buy-pressure that pushed prices higher.

Figure 8. Troy  ounces of gold held by ETFs (blue) and gold price per ounce (red). Source: Bloomberg
Figure 9. Troy ounces of gold held by ETFs (x-axis) and gold price per ounce (y-axis). Darker colours indicate older datapoints. Source: Bloomberg

The gold price rally lines up with the accumulation of gold by China, who were the largest purchasers of gold in 2023. China bought a total of 316 tonnes of gold in an accumulation period which began from 2022 and ended in May of 2024. Following China’s pause, gold prices then began to dip which also coincides with the bottom in ETF gold holdings.

Figure 10. Troy ounces of gold held by ETFs (blue), gold price per ounce (red), and gold held by China (green). Source: Bloomberg
Figure 11. Tonnes of gold purchased by China since 2022. Source: Bloomberg

Since then, the price of gold has continued to rally, reaching new all-time highs in early October. But this time the rally in demand cannot be attributed to China’s reserve stockpiling. Instead, the positive pattern of correlation between ETF accumulation and gold price action has returned, albeit at a higher price than historically indicated by the stockpile of gold held by ETFs.

Figure 9. Troy ounces of gold held by ETFs (x-axis) and gold price per ounce (y-axis), darker colours indicate older datapoints. Source: Bloomberg

The restart in Gold ETF accumulation and gold price rally also coincides with an increase in correlation between BTC returns and gold returns. If the demand for gold from ETF investors is the reason for the September rally in gold prices, we could see a similar pool of institutional investors who are both looking for exposure to BTC via the BTC Spot ETFs as well as gold – while buyers like China are far less likely to express their demand for a store of value in digital gold as well as physical gold.

Figure 13. Rolling 180-day correlation of BTC returns and gold returns. Source: Bloomberg

We see early indications of exactly that, as the correlation between BTC and gold returns has increased, beginning from the same mid-July date that gold ETF holdings began to rise alongside gold price. Accumulation via ETFs (which we note are available to institutional and retail investors alike) seem to be driving both gold and BTC returns once again, leading both to rally in early October. At the same time, the correlation between those returns has increased. Following the beginning of the Fed’s rate cutting cycle in September, gold has reached all-time high prices and BTC has returned close to the top of its recent range bound price action.

Share this post
Copy URL
www.blockscholes.com/premium-research/bitcoins-macro-correlations

Returns and Surprises in CPI

As we have previously highlighted, Bitcoin has become increasingly correlated with macroeconomic assets. The relationship has oscillated since the outbreak of the Covid pandemic, and has grown stronger after the launch of the BTC Spot ETFs in January 2024. Since then, macro has played a significant role in BTC’s price trajectory, range bound activity, and recent breakout. That much is clear from its relationship with the SPX at times close to the release of macroeconomic data that has caused market participants to revise their expectations of monetary policy.

Figure 1.  Linear regression of BTC (orange, R^2=4%) and SPX (red, R^2=3%) returns on CPI release day onto the amount by which the CPI figures surprised relative to the median estimate by Bloomberg economists. Source: Bloomberg, Block Scholes

The chart above measures this by plotting the relationship between surprises (the difference between the actual headline CPI figure year-over-year, and the Bloomberg median forecast) in the U.S. Consumer Price Index (CPI) reports, a key macro data release, and the same-day returns of the S&P 500 index and BTC.

The negative slope indicates a negative relationship: when CPI surprises to the upside (when actual inflation is faster than expected), BTC tends to drop in price over the next day. This is potentially because a higher-than-expected inflation figure raises the forecast for how long restrictive monetary policy will last, as was the case from March 2022 and July 2023. The S&P 500 exhibits a less volatile reaction than BTC, but a stronger (higher R^2) statistical relationship.

Returns and Surprises in CPI

As we have previously highlighted, Bitcoin has become increasingly correlated with macroeconomic assets. The relationship has oscillated since the outbreak of the Covid pandemic, and has grown stronger after the launch of the BTC Spot ETFs in January 2024. Since then, macro has played a significant role in BTC’s price trajectory, range bound activity, and recent breakout. That much is clear from its relationship with the SPX at times close to the release of macroeconomic data that has caused market participants to revise their expectations of monetary policy.

Figure 1.  Linear regression of BTC (orange, R^2=4%) and SPX (red, R^2=3%) returns on CPI release day onto the amount by which the CPI figures surprised relative to the median estimate by Bloomberg economists. Source: Bloomberg, Block Scholes

The chart above measures this by plotting the relationship between surprises (the difference between the actual headline CPI figure year-over-year, and the Bloomberg median forecast) in the U.S. Consumer Price Index (CPI) reports, a key macro data release, and the same-day returns of the S&P 500 index and BTC.

The negative slope indicates a negative relationship: when CPI surprises to the upside (when actual inflation is faster than expected), BTC tends to drop in price over the next day. This is potentially because a higher-than-expected inflation figure raises the forecast for how long restrictive monetary policy will last, as was the case from March 2022 and July 2023. The S&P 500 exhibits a less volatile reaction than BTC, but a stronger (higher R^2) statistical relationship.