1. What drives Bitcoin prices?

Bitcoin supply is designed to become increasingly constrained

As it is the case in all other commodities, supply and demand drive Bitcoin prices. Thus, to understand why Bitcoin prices move today and how they could behave in the future, any analysis has to start with a deep look at the underlying supply and demand conditions of this commodity. For starters, it is worth noting that Bitcoin supply is capped and designed to become increasingly constrained. Also, by construction, the cost of Bitcoin mining features prominently (Exhibit 1) in price discovery, particularly energy. A quick look at the dramatic increase in prices in the past year quickly points to a major supply/demand imbalance. In fact, Bitcoin's rise from late 2020 to now has made the previous run-up in 2017 look miniscule. Plus this time around the run up in Bitcoin prices has been joined by a surge in ether prices too (Exhibit 2).

Miner rewards halve every 4 years and there is a 21mn BTC cap

How is Bitcoin supply constrained? When Bitcoin was first launched in 2009, its supply was designed to be maxed out at 21mn coins. The view that unlimited money printing by central banks during the financial crisis would contribute to devalue fiat currencies featured prominently in the initial stage of Bitcoin. A similar view may have contributed to the recent hype too, as some investors argued that large scale monetary easing could lead to rising inflation. Yet, it is crucial to understand that Bitcoin is in fact "the fuel" that keeps the decentralized cryptocurrency system moving. New coins have to be issued constantly to "clear and settle" transactions on the blockchain. By design, the creators of Bitcoin allowed more than 100 years to get to this maximum level of 21mn coins (Exhibit 3). So given the hard cap, mining rewards have to halve every 4 years, a factor that has positively impacted prices in recent years by artificially reducing supply growth rates (Exhibit 4).

Institutional demand has become more active in the past year…

With supply capped by design and supply growth built to halve every so often, the various swings in Bitcoin demand in recent years have been key to price changes. While there is no single way to split Bitcoin demand, we would differentiate between retail, whale/institutional, and illicit activity demand for crypto assets. For example, institutional announcements in the past year have led to major price increases (Exhibit 5), as a number of household names like Tesla, Square, and PayPal have incorporated Bitcoin into their businesses either as a treasury holding or as a means of payment. Still, reported institutional holdings generally comprise of a small portion of total Bitcoin supply, led by the Grayscale Bitcoin Trust (GBTC) (Exhibit 6).

…in contrast to the retail flows that drove the 2017-18 run up

In section 2, we provide a detailed breakdown on whale/institutional demand for some of the largest accounts. We label it whale/institutional as it is unclear whether some these billion dollar plus accounts are held by a single individual or by an institution. But it seems to us that, while a flurry of retail money to Coinbase and other crypto wallets drove prices higher in 2017 (Exhibit 7), the run up in Bitcoin prices in the past year has been more broadly driven by institutional money flows. In part, this is because institutional scale Bitcoin purchases are hard to accommodate in such a supply constrained market. For instance, quarterly trading volume and change in ETF holdings relative to new Bitcoin supply show that institutional demand is overwhelming supply (Exhibit 8) at present, a factor that has surely contributed to rising prices. So who are the major institutional players entering the Bitcoin market?

Major institutional announcements and reward halving….

Microstrategy was probably the first billion+ dollar market cap corporation to announce Bitcoin holdings in August 2020 (Exhibit 9). The company is now holding over $3.5bn in crypto, or about half its market cap, and plans on issuing a $600m convertible senior note to purchase even more. Another notable transaction in the past year was Tesla's net purchase of $1.5bn worth of Bitcoin in 4Q20. Separately, some financial companies have moved to add crypto services. For instance, Square has offered Bitcoin trading on their CashApp since 2018, and recently purchased $50mn Bitcoin in October 2020 for their own balance sheet. On the other end, Paypal has not announced any purchases and only launched support for crypto transactions. While Paypal and Square may need to hold some crypto inventory, it's more probable they will operate like a broker dealer: a lot of the volume will likely be washed out. And last but not least, Mastercard recently announced support for crypto later in 2021, although nothing has been launched yet.

…have positively impacted Bitcoin prices, on average

A key point to note is that the steady purchases by some of these institutional buyers are a very sizeable share of the market. For instance, we calculate that the current position of Microstrategy amounts to about 0.40% of total Bitcoin supply. For Tesla, we estimate that the company purchased 44079 BTC, or about a quarter of a percent of total supply. Given the large scale of these purchase and the steady reduction in incremental supply, it is no surprise that Bitcoin prices have surged. In fact, when we look at the first three Bitcoin reward halvings, we note that each of these events led to price increases on average (Exhibit 10). Also, our work shows that Bitcoin prices have surged on average after major institutional adoption announcements (Exhibit 11).

We find Bitcoin prices have been correlated to Google searches…

While it is clear that reductions in supply growth and increases in demand are behind the Bitcoin price increases of recent years, we also note that external signals seem to have provided some warning of impending movements in prices. In particular, some investors have pointed to internet searches as a potential leading indicator of crypto-currency moves. The last two large Bitcoin rallies have been accompanied by a large increase in on-line interest in crypto, as Google trends show (Exhibit 12). A similar pattern emerges with Ethereum and we show that interest has been even higher this time around than before (Exhibit 13).

…but our work also shows Google searches lagged Bitcoin prices

Yet correlation does not necessarily imply causation, and fortunately there are a few statistical tools that can help us establish the direction of this causality. For instance, the cross-correlations of Google trend searches and Bitcoin prices show that search trends lagged Bitcoin prices (Exhibit 14). Similarly, a Granger-causality test confirms the hypothesis that trends lagged Bitcoin prices (Exhibit 15), and not the other way around. Put differently, people do not generally search for Bitcoin and Ethereum on the internet to try to understand whether these crypto assets do or do not make good investments. Instead, statistical analysis shows people are looking for information on Bitcoin because they see prices going up.

Grayscale (GBTC) is now the largest public holder of Bitcoin

So what other public information can provide some clues regarding Bitcoin's price direction? Just like it is the case with other financial instruments, investor flows into publicly listed vehicles can also be of great use. For this reason, we looked at various financial instruments with direct exposure to crypto. We note that the market capitalization of Grayscale Bitcoin Trust has skyrocketed and the fund has accumulated substantial BTC (Exhibit 16), with total assets now likely at 700k BTC or $31bn or 3.5% of total supply. In that regard, we also show that the Trust has been steadily buying Bitcoin over 2020, especially in 4Q20 (Exhibit 17) and has become one of the 5 largest holders of the crypto-currency.

In contrast to searches, we do find GBTC flows lead Bitcoin returns?

While we have argued that Google searches provide little insight into the direction of Bitcoin prices, we do find instead that GBTC flows lead Bitcoin returns? To be precise, flows into Grayscale Bitcoin Trust as measured by changes in shares outstanding appear to have lead weekly Bitcoin returns (Exhibit 18). A similar pattern appears when analyzing cross-correlations in BTC and Granger causality tests (Exhibit 19). This is not an uncommon occurrence in the world of commodities, with physical gold and silver ETFs showing similar patterns. Inflows into GLD and SLV also lead prices of these precious metals. Unlike the price impact of money flows into futures markets, where every long position is matched by an exact short position, physically-backed ETFs reduce actual supply to the market. So as financial players mop up "physical" supply of gold, silver, copper, or even Bitcoin in this case, the spot market tightens and prices inevitably go up. In the next section, we provide a more detailed analysis of the GBTC and other crypto-financial instruments.

Daily Bitcoin transactions have remained steady as prices rose

We first discussed the liquidity properties of Bitcoin three and a half years ago. Back then, we believed the surge in Bitcoin trading liquidity was perhaps one of the most important features of this emerging asset. Yet, much of the growth in crypto in recent years has not translated into increased Bitcoin liquidity. Rather, the number of daily Bitcoin transactions has averaged 287k in the past four years (Exhibit 25). While renewed interest is apparent, much of the increased institutional focus has translated into rising prices (and thus transactional values) and not into rising Bitcoin liquidity and not even in new Bitcoin addresses (Exhibit 26).

Bitcoin transactions reflect higher fees and bigger players

With the resurgence in Bitcoin interest, are there any features in transactions that are particularly striking? For starters, transaction fees have soared in 2021, with the 30-day average close to $20 per transaction and median fees just north of $10, and quickly closing in on peak levels seen in 2017-2018 (Exhibit 1). Taken as the ratio of total daily transaction fees to number of transactions, this suggests that even with more activity, participants are paying more to join the party. Furthermore we find that while the number of trades per account has dropped, total output (amount of transaction including change) per transaction jumped late 2020 and has remained at high levels akin to 2017-2018 (Exhibit 2). Total output includes the amount of transactions and change, the amount of excess Bitcoin sent in a transaction above the transaction value that is returned following settlement¹. Parsing out change, estimated average transaction value (total output net of change) remains low, suggesting that average transactions have declined in size. However because average total output remains high, accounts making these transactions are likely facilitating them with larger positions, resulting in larger change and thus higher output despite estimated transaction values being low.

Futures volumes have been growing at a relatively rapid pace

The same cannot be said of other financial instruments that support crypto assets. Bitcoin futures volume has been on the rise for several years now and we show that it tends to increase towards the end of the month as it approaches settlement. Even then, the notional amount of futures trading is still magnitudes below actual spot BTC volume (Exhibit 27). Note that in other commodity markets, the reverse is true. Daily futures and swaps liquidity in copper, gold, or oil is a multiple of the traded spot market value of these commodities. And as a side note, Bitcoin futures appear to be net short by the speculative community, suggesting that these instruments may in fact be used more for hedging than to establish fresh long Bitcoin positions (Exhibit 28).

Dollar crosses increasingly dominate Bitcoin transactions…

Another important aspect of price formation relates to the currency that Bitcoin is traded in. It is often said that a weaker USD helps gold prices rise and vice-versa. In the case of Bitcoin and Ethereum, dollar crosses increasingly dominate Bitcoin transactions. Nowadays, about 80% of Bitcoin is exchanged in dollars (Exhibit 29), while other currencies like the EUR, the JPY, and the KRW account for 19%, .01%, and .43% of Bitcoin trading, respectively. Another way to look at liquidity is to analyze the Amihud ratio, a key trading metric that represents the illiquidity of an asset as proxied by the absolute value of daily returns scaled by trading volume. For most assets, especially the most illiquid (high Amihud ratio), we expect higher illiquidity to translate to higher returns as buyers are compensated for an illiquidity premium. However with Bitcoin, we are finding that increased liquidity is actually supporting prices, with the largest impact coming from dollar and KRW liquidity, the opposite of traditional assets (Exhibit 30).

…with Coinbase market share increasing over other exchanges

One last point to note on our supply/demand analysis of Bitcoin is where it trades and where it is held. From that viewpoint, we show that the share of Coinbase as an exchange has increased to 40% in recent years (Exhibit 31) as the US company has become the leading wallet and exchange. Again, running a similar analysis using the Amihud ratio, we find that Coinbase volumes are now adding to Bitcoin liquidity and subsequently supporting prices (Exhibit 32).

 2. How does Bitcoin compare to traditional portfolio assets?

About 95% of Bitcoin is controlled by just 2.4% of the accounts…

In our view, one of the main challenges that Bitcoin faces to gain mainstream adoption going forward is its high concentration ownership ratio. As a reference point, we note that about 95% of total Bitcoin is owned by the top 2.4% of addresses with the largest balances (Exhibit 33). We also show that the distribution is heavily skewed towards the largest accounts (Exhibit 34). By comparison, the latest Fed data suggests that the top 1% of Americans control about 30.4% of all household wealth in the US. In our view, the fact that such a small percentage of Bitcoin accounts hold most of the BTC in circulation makes this instrument impractical as a payments mechanism or even as an investment vehicle. It can also create social and governance issues, as we will discuss in the section 3.

…and large holders have kept accumulating through the rally

An often misunderstood detail on Bitcoin is that many whale/institutional account holders are actually buying, rather selling BTC. Looking at detailed blockchain records, we find that the largest addresses have not been selling in aggregate (Exhibit 35) since the pandemic began. Also, we note that Bitcoin has a large number of dormant accounts, a factor that likely drains liquidity from the crypto-asset. For instance, the top 100 dormant Bitcoin addresses over the past year comprise of over 7% of total supply (Exhibit 36). Who are these dormant holders? It is hard to say, although recent news reports for example (New York Times: Lost Passwords Lock Millionaires Out of Their Bitcoin Fortunes) identified a holder of a lost a crypto key with $220mn worth of Bitcoin. As much as 20% of existing Bitcoin appear to be in lost or in stranded wallets, according to crypto data firm Chainalysis.

There is simply no other major asset that shares Bitcoin's features

Bitcoin is clearly a unique asset class by way of its supply mechanics, ownership concentration, and even historical performance. Ownership concentration is particularly severe. A small set of total accounts owns a massive portion of Bitcoin, leaving the asset vulnerable to sharp price swings on account of movements in these "whale" accounts. Indeed looking at the income distribution, we find that Bitcoin inequality is unprecedented and easily eclipses countries plagued with the most income inequality as measured by its remarkably high Gini coefficient (Exhibit 41). However even with susceptibility to large movements on account of ownership concentration and limited supply, the asset has produced relatively high risk-adjusted returns despite having one of the highest volatilities in the cross-asset space (Exhibit 42).

Even after a decade of trading, Bitcoin remains very volatile…

Another feature that makes crypto-assets impractical as a payments or store of wealth mechanism is their persistently high volatility. As an example, Bitcoin volatility is well above that of FX, gold, and even silver. Bitcoin volatility in 2021 is already the second highest in history year-to-date (Exhibit 43). The persistently high volatility levels are a big concern, as they suggest that even after 10+ years of trading and major episodes of price appreciation, Bitcoin has failed to develop some price stability. If anything, Bitcoin volatility has been rising with prices during recent months. And if we had to make an observation on Bitcoin's realized volatility, it is that it appears to have had some cyclicality too (Exhibit 44).

…and Bitcoin volatility still exceeds that of many EM FX...

Just like Bitcoin volatility is multiples of the vol observed in other commodity markets, we also note that it vastly exceeds that of many Emerging Market FX. For instance, Bitcoin is more volatile than currencies with severe capital controls (Exhibit 45), which tend to be EM countries with high inflation (Exhibit 46). Thus, we believe the claim that crypto could replace some volatile emerging market currencies sits on weak grounds and is only supported by the massive price appreciation observed in recent years-- appreciation which is a function of the major supply and demand imbalance discussed in section 1 of this note.

… and rising $ volumes are increasing vol and supporting prices

Having said all of that, there are some positive developments on the liquidity front when it comes to Bitcoin. For instance, the notional dollar value of Bitcoin trading volume has risen to over 10x that of GLD in the past 2 years (Exhibit 47), even if mostly driven by price appreciation. So large transactions are increasingly common in the crypto space. Also, a rolling Amihud liquidity analysis shows that Bitcoin liquidity is now helping speed up price gains as participants find the market more readily accessible and legitimate (Exhibit 48). Note that this same analysis provided similar conclusions during the retail-driven rally of 2017-18.

The market value of Bitcoin has been rising relative to gold

Another way to look at the importance and growth trajectory of Bitcoin is to compare it to the size of other major physical and financial assets. In particular, we believe assets like the Fed's balance sheet or the price of gold are relatively good reference points for Bitcoin. In that regard, we note that Bitcoin market capitalization is still less than 10% of gold and the Fed's balance sheet even after years of steady growth (Exhibit 49). Even then, the value of Bitcoin is enormous relative to the largest silver ETF (SLV) and the physical stock of silver in vaults (Exhibit 50), underscoring the diminishing importance of silver as a reserve asset.

Bitcoin no longer appears uncorrelated to other major risk assets…

Perhaps the biggest change in Bitcoin pricing dynamics since we last wrote about them in 2017 is the change in correlations to financial assets. For starters, Bitcoin no longer appears uncorrelated to other major risk assets. In fact, we find the crypto-currency more positively correlates with equities and commodities, while it is neutral/slightly negatively correlated to haven assets such as the dollar and US treasuries (Exhibit 51). And as a new development in the crypto space since the pandemic, Bitcoin correlations with risk assets have tended to move in lock step even across asset classes (Exhibit 52).

...although it is still highly correlated to other crypto assets…

Digging down a little deeper, we find that correlations are high with most risky FX, while have become negative with havens like the CHF and JPY (Exhibit 53). Note that this pattern is rather common for risk assets and ultimately subtracts from the diversification argument for owning Bitcoin and other cryptocurrencies. Importantly, Bitcoin has also started to correlate well with volatile and cyclical commodities as of late (Exhibit 54), again supporting the idea that Bitcoin diversification benefits are falling, not rising. Lastly, we would also note that Bitcoin correlations to other crypto assets remain relatively high, as they did years ago (Exhibit 55), suggesting once more that there may be limited benefits to owning a diversified basket of crypto in a portfolio.

…so Bitcoin portfolio diversification benefits have come down…

A frequent argument advanced by investors in the past year is that Bitcoin could protect portfolios against inflation. In fact, we find that Bitcoin has been positively correlated with TIPS as of late. Meanwhile, higher rates have been associated with higher Bitcoin prices and we now are at multiyear troughs in this relationship (Exhibit 56). Broadly, we find that Bitcoin has not been particularly compelling as an inflation hedge as commodities and even equities provide better correlations to inflation (Exhibit 57). As such, we think the main portfolio argument for holding Bitcoin is not diversification, declining volatility, or inflation protection, but rather sheer price appreciation, a factor that depends exclusively on Bitcoin demand outpacing supply on a forward basis.

… and inflation hedging benefits are not particularly apparent

Looking year by year, we find that Bitcoin has been positively correlated with CPI inflation in 5 out of the 9 past years, with the largest correlations in 2014 and 2018 (Exhibit 58). On the bright side, the 2018 performance is encouraging as this was the last time that US CPI inflation was above 2.5% for a sustained period of time. However when looking at correlations with inflation surprises since 2011, we find that Bitcoin has among the lowest co-movements, lagging most asset classes such as commodities, TIPS, and EM FX in particular (Exhibit 59).

Vol adjusted returns have been high compared to other assets

True, total Bitcoin returns this year are already among the highest in its short history (Exhibit 60) and investors have noticed. In 2020 Bitcoin returns neared 160%, while so far this year Bitcoin returns are up 420%. The lowest annualized return for Bitcoin in its nearly 13 year history is 101%, while the highest return was actually 525%, with the average annual return exceeding 180% when including 2021 year to date. Having said that, these high returns have also come with high volatility. So has Bitcoin delivered good risk-adjusted returns? When adjusting Bitcoin returns for volatility and comparing them to other major asset classes, we still find that while Bitcoin risk-adjusted returns may look attractive from a historical perspective (Exhibit 61), they are not overwhelmingly larger than that of other risk assets.

Crucially, money inflows have a huge impact on Bitcoin prices

What has created the enormous upside pressure on Bitcoin prices in recent years and, particularly, in 2020? The simple answer: modest capital inflows. After a detailed analysis, we find that Bitcoin is extremely sensitive to increased dollar demand. For example, we estimate a net inflow into Bitcoin of just $93mn would result in price appreciation of 1%, while the similar figure for gold would be closer to $2bn or 20 times higher (Exhibit 62). In contrast, the same analysis for the 20y+ Treasuries shows that multi-billion money flows do not have a significant impact on price, pointing to the much larger and stable nature of the US Treasuries markets.

 3. How does Bitcoin score on ESG?

 The E in Bitcoin: major environmental risks

Bitcoin prices are directly correlated to its energy consumption…

In contrast to the low carbon intensity of central bank digital currencies, we believe ESG-minded investors have to pay attention to the enormous environmental costs of Bitcoin, as well as its social and governance flaws. In some ways, the rising complexity that underpins Bitcoin is its biggest asset. In some others, rising complexity is the biggest flaw of the entire system. On the one hand, rising Bitcoin prices bring in rising miner profits and growing network complexity to capture them, essentially a mechanism aimed at preventing a hack. On the other, growing network complexity pushes up the hash rate and ultimately the energy consumption of the entire system. As such, the Bitcoin network hash rate continues to grow as the network adds mining hardware and improves efficiency (Exhibit 63). And even if the recent growth in Bitcoin's market value has outpaced the growth in its energy consumption (Exhibit 64), the skyrocketing prices have propelled Bitcoin's CO2 emissions higher.

…as the hash power required to solve blocks increases with price

Ultimately, new crypto miners keep joining the highly competitive crypto space in the hope of huge profits. With more miners around, the difficulty to solve each block goes up and, in turn, so does the hash rate. With renewed interest and volume in the past year, we can confidently say that Bitcoin's mining difficulty and blocksize are back to all-time highs (Exhibit 65). From an environmental standpoint, this development is not good news. Given the relatively linear relationship between Bitcoin prices and Bitcoin energy use, it is perhaps no surprise that Bitcoin's estimated energy consumption has grown over 200% in the past two years (Exhibit 66), creating large environmental risks.

Hardware becomes more efficient over time, but profits vanish

True, miners do not use old hardware or waste energy voluntarily, as this is a key input into Bitcoin mining costs and thus critical to profit margins. Bitmain Antminer, a leading hardware provider, has gotten more energy efficient with each release (Exhibit 67). Yet, Bitmain Antminer profit margins quickly decline as more efficient cards come to market (Exhibit 68), so the rising complexity of the system creates ultimately a vicious environmental cycle of rising prices, rising hashpower, rising energy consumption and, ultimately, rising CO2 emissions. And with limited amounts of capital pushing prices higher, as we discussed in section 2, rising Bitcoin prices can quickly lead to astronomical CO2 emissions.

Today Bitcoin consumes as much energy as a country like Greece…

Our analysis shows that Bitcoin's annual energy consumption now rivals that of some small developed countries (Exhibit 69) like the Netherlands and Czech Republic. Of course, in each of these countries live 17 and 10 million people, respectively. And while Bitcoin may have some undetermined social value, it is hard to argue that it provides a greater value to society than the lives of millions of people. From an institutional perspective, we find that Bitcoin's energy consumption is comparable to that of major US corporations like American Airlines, the largest airline in the world that flies over 200mn people a year, or even the Federal US government (Exhibit 70), which employs more than 2mn people. The US military also emits about 60mn MT of CO2.

…and, at $50,000, it uses about 0.4% of global energy consumption

But as we noted at the start of this section, Bitcoin's biggest energy design flaw is the same rising complexity that helps prevent a hack. One of the most concerning datapoints across the crypto space is that Bitcoin emissions have soared by over 40mn tons in the last two years (Exhibit 71), the equivalent of 8.9mn Internal Combustion Engine vehicles. With increasing Bitcoin prices, the carbon footprint of Bitcoin may rise further into the future. Again, that's what the system is designed to do. Bitcoin supply is constrained, so any excess demand pushes prices higher. In turn, higher Bitcoin prices encourage more mining activity and would push CO2 emissions to even higher levels (Exhibit 72).

Yet, rising prices could push Bitcoin energy burn up dramatically

While the 40mn ton increase in CO2 emissions in the past two years is concerning, we find even more worrying the price sensitivity of Bitcoin to modest dollar capital inflows. In fact, a back-of-the envelope calculation suggests that rising Bitcoin prices could push Bitcoin CO2 emissions to levels rivalling some of the largest economies in the world (Exhibit 73). Again, the network is operating exactly as it was designed to do. But this comes at an enormous environmental cost with arguably a limited social purpose. No surprise then that the network hash rate has continued to increase as miners have consistently been able to cover variable costs due to continued price appreciation (Exhibit 74).

Ultimately, three quarters of hash power is concentrated in China…

Perhaps the most environmentally concerning development of all when it comes to Bitcoin is where most of the world's hash power sits. Geo-location data from mining pools suggests that about three quarters of global Bitcoin mining occurs in China (Exhibit 75). So fresh capital into Bitcoin breeds increased use of Chinese electricity networks. And what source of energy does China use the most? Nearly 60% of Chinese electrical generation is from coal fired power plants, with less than 20% coming from natural gas or renewables (Exhibit 76). In other words, the main input into Bitcoin mining is coal, not exactly the cleanest source of energy on planet Earth.

…largely because miners are looking for cheap electricity sources

Why are miners so concentrated in China? Perhaps it is because the Chinese government has encouraged Bitcoin mining to gain increased control to global crypto networks. Or perhaps it is because China has some of the lowest electricity costs in the world, an extremely attractive feature for Bitcoin miners (Exhibit 77). Or perhaps it is because China has overbuilt its electricity generation capacity over the years and needs to find a use for all those power plants. After all, most of the variable costs of Bitcoin mining comes from energy. Net, we find that Bitcoin's cost of supply has risen to nearly $20,000/BTC as supply became more limited and mining competition increased (Exhibit 78), with energy representing about 40% of that figure.

Half of China Bitcoin mining happens in coal-heavy Xinjiang province

For those that have followed the logic behind the our arguments, it should come as no surprise that China's Bitcoin industry is not located around the Three Gorges Dam, the largest hydropower plant in the world, or around Qinghai Province, next to the second largest solar park on the planet. Rather, the Chinese province of Xinjiang is home to nearly half of all Bitcoin mining in China (Exhibit 79). In Xinjiang, more than 80% of the electricity is generated from thermal sources and it is mostly coal-based (Exhibit 80). Granted, Sichuan province, also an important Bitcoin mining center, does run along the Yangtze River and benefits from access to hydropower. But Bitcoin today is mostly a derivative of coal, with some of China's hydropower being used seasonally in the summer months. While that may change in the future, and China has recently announced plans to be carbon neutral by 2060, the pace of appreciation of Bitcoin has been too fast to handle and comes with significant adverse environmental consequences.

A $1bn Bitcoin purchase may have a CO2 footprint of 5.4mn tons

Investors looking into this asset class should keep in mind that modest amounts of capital inflows into Bitcoin can have an outsized environmental impact. We estimate that a $1bn fresh inflow increases Bitcoin prices on average by 11%, which in turn results in a carbon footprint of 5.4mn tons (Exhibit 81). Comparatively, we do not find many other human activities that have a higher carbon footprint per dollar of inflow (Exhibit 82). As a reference point, we calculate that a $1bn dollar inflow into Bitcoin is equal to 1.2mn cars driven over the course of a year or 12.7mn barrels of oil. Looked at differently, a single Bitcoin purchase at a price of ~$50,000 has a carbon footprint of 270 tons, the equivalent of 60 ICE cars.

Other cryptos like Ethereum also suffer from huge energy intensity

While Bitcoin's carbon footprint is concerning, other crypto assets are only slightly better. For instance, we estimate Ethereum's energy consumption doubled from 8 to 17 tera-watt hours in less than six months as prices jumped to nearly $2,000 (Exhibit 83), pushing its carbon footprint from 4.4 to 9.5mn tons of CO2. At present, Ethereum's energy footprint appears comparable to that of a small economy like Cuba (Exhibit 84), home to 11 million people, but its rapid price appreciation presents meaningful environmental risks down the line. True, the move away from proof-of-work to proof-of-stake could reduce Ether's carbon footprint. But in turn the new proof-of-stake model will increase both the social and governance risks of this crypto-currency. Is there a silver lining? As renewable energy production increases over the next 20 years, quantum computers reduce energy usage, and new, more efficient crypto assets continue to emerge, the cryptocurrency space could eventually find ways to reduce its carbon footprint. But a rapid surge in adoption of Bitcoin presents a major risk, and thus drives Bitcoin's low Environmental score.

 The S in Bitcoin: anonymity can be good and bad

From a social perspective, the key negatives, including money laundering, tax evasion, terrorist financing, and cyber-attacks, could potentially be tackled through more stringent regulation.

On the positive side, Bitcoin, and particularly blockchain technology, has significant potential to contribute to long-term social sustainability vis a vis supply chain traceability, renewable energy distribution, proxy voting, and even prevention of some illicit activities.

• The democratisation and decentralization of money: A key premise of Bitcoin is that of democratizing the financial markets. As Satoshi Nakamoto wrote in his white paper from 2008: 'A purely peer-to-peer version of electronic cash would allow online payments to be sent directly from one party to another without going through a financial institution'. Removing intermediaries presents some social advantages.
• Furthering UN's Sustainable Development Goal target 17.3: lower costs for remittances would benefit developing nations and more disadvantaged communities.
• Small business benefit: lower payment transaction costs would be helpful for small businesses (i.e. no card payment fees, which eat 1-5% of product value).
• Anonymity (also a negative, see below): helpful for individuals who live in oppressive societies, where Bitcoin can provide financial security independent of potentially corrupt domestic financial systems.

Negatives:

• Technological requirements: Bitcoin transactions presume access to computers and internet, which are particularly challenging for the developing countries. According to UN's International Telecommunication Union data, c.51% global population are using the Internet as of end-2019.
• Currency volatility: As discussed in the section on How do cryptocurrencies compare to traditional portfolio assets?, Bitcoin volatility is well above that of FX, gold, and even silver. But continued momentum in digitalisation, as well as wider adoption, could reduce volatility over time.
• Anonymity: the negative side of anonymity is the cover it provides for illicit activities including terrorism financing, money laundering, tax evasion, and even evasions of currency regimes.

• • Research from Copenhagen Business School and Coller School of Management uses forensic accounting and machine learning on millions of crypto transactions as a way to confirm a link between terrorism and cryptocurrency payments. The researchers find strong positive cumulative abnormal value responses in the week preceding attacks, followed by a reversal in the week after. This pattern is consistent with terrorist financing.
• • Researchers from the University of Sydney in 2019 estimated that around $76bn of illegal activity per year involves Bitcoin, which is close to the scale of the US and European markets for illegal drugs².
• • A systematic literature review by researchers from the University of Ottawa (2020) points to a growing number of cyber-attacks through Bitcoin and other cryptocurrencies (most commonly Ethereum and monero). Some of the most common cyber-attacks are: high-yield investment programs, ransomware, pump and dump schemes, and crypto jacking. Furthermore, CipherTrace's 2020 Cryptocurrency Crime and Anti-Money Laundering Report reveals that in 2020, major crypto thefts, hacks, and frauds totalled $1.9 billion.

• More regulatory oversight needed: concerns about Bitcoin point to increased need for regulation and, more importantly, compliance, which is costly and onerous. Illicit activities remain a challenge for both crypto and non-crypto channels of payments.

 Mixed reviews on G

Bitcoin presents mostly negative G consequences from the perspective of decentralised finance, where key stakeholders (particularly the exchanges) remain relatively concentrated and lack regulatory oversight. This concern resonates strongly with the centralised ownership structure discussed in the previous sections (i.e. about 95% of Bitcoin is controlled by just 2.4% of the accounts).

With wider adoption of Bitcoin, there is scope and necessity for more robust governance / regulatory mechanisms to protect customers. An additional regulatory issue is a lack of international cooperation and heterogeneous AML (anti money laundering) /CTF (counter terrorist financing) regimes globally.

Unregulated exchanges

Bitcoin has highly concentrated control over the mining process (by a small number of mining pools), transactions / payments (through cryptocurrency exchanges), and maintenance of the code (the 'Bitcoin Core' group). For prospective investors / users who would invest in or transact with Bitcoin, the most problematic aspect is exchanges.

Crypto exchanges provide liquidity and facilitate price discovery, just like traditional exchanges. Furthermore, many crypto exchanges have expanded their business models into upstream activities (i.e. mining) and downstream (i.e. payment). This, in turn, allows them to play a larger role that would normally be shared among trading platforms, custodians, banks, and clearinghouses.

The concentration of power is not necessarily negative per se (the economic literature would support some benefits from centralisation), but the modus operandi of some exchanges is concerning. According to researchers from Cornell, Newcastle, and Tsinghua universities, unregulated exchanges are active in market manipulations. Based on the sample of 29 exchanges and millions of transactions from 2019, the researchers concluded that the wash trading volume on average is as high as 77.5% of the total trading volume on the unregulated exchanges analysed. Currently, a majority of exchanges are unregulated.

Institutional adoption of Bitcoin

The current cryptocurrency regulatory framework is insufficient, as reflected in European Central Bank President Christine Lagarde's call for regulation to address 'funny business' associated with cryptocurrencies. However, regulation is ever-evolving and is likely to speed up with more institutional adoption of Bitcoin.

According to research by our US banking analysts, 21% of US banks under coverage already have blockchain technology in use: JPM, C, WFC, UBS (an early investor in R3), PNC, FITB, and SBNY. Of the regionals not currently using blockchain technology, MTB and CFG appeared to be the most open to engaging third parties at some point to provide the technology (for more details see here).

 ESG analytics mostly fail to capture Bitcoin risks

Most ESG data providers do not provide a comprehensive framework for how to assess large Bitcoin / cryptocurrency portfolio holdings or exposure. This theme is likely to become more salient if corporates begin to purchase large amounts of Bitcoin (most recently, Tesla). Such purchases could gain more traction through the existing ESG metrics of materiality or controversies.

Reprisk, a leading data provider on risk scores related to environmental, social and governance related issues, found 181 companies that have an association between ESG risks and crypto-currency, 60 of these incidents occurring in 2020 - a demonstrable acceleration in crypto-currency related risks. The "S" and "G" issues most frequently associated with Bitcoin are money laundering, corruption, bribery, fraud, and breaches of data privacy, with US, UK and Japan as the top 3 countries by incident.

Case study: A peek under the hood of Tesla

Favored by passive ESG, not by active ESG: Tesla currently holds the top weighting in the MSCI USA SRI index, one of the largest ETFs that we track, and generally carries a 60% overweight in the average passive ESG fund. But active ESG managers are less bullish, and only one in ten funds own the stock (versus one in three passive funds) and the average active ESG manager is ~60% underweight the stock. Why the difference? TSLA scores Low in our ESGMeter scoring system, with Low scores on all components of E S and G. The low scores were impacted by lack of disclosure (more below) as well as incentive compensation structure, lack of proxy access, and board characteristics (staggered, negative votes on directors).

Most data providers rank Tesla as above average on environmental aspects - after all, the company produces electric cars. But TSLA fails to disclose Scope 1, 2 or 3 emissions, where peers like General Motors (GM) do disclose this information. A lack of emissions disclosure currently carries a median 3 point P/E multiple discount by for companies within Consumer Discretionary.

Furthermore, by applying our Bitcoin emissions analysis to TSLA's recent investment of $1.5bn of reserves in Bitcoin, the indirect carbon intensity of the company if disclosed would jump by a significant amount.

From a Social and Governance perspective, the diversity of thought leadership and workforce statistics at Tesla are sparse, but we do note that average age of board members is considerably younger at 53 than Consumer Discretionary peers (60yo average age) indicating perhaps greater comfort with new asset classes like Bitcoin, but also perhaps shorter experience with boom/bust cycles for speculative themes.

 4. Are CBDCs Kryptonite for crypto?³

Central bank digital currencies ("CBDC") would represent a major break with the status quo; in this, they are like cryptocurrencies. However, we think they are likely to differ technologically from cryptos, and are in part motivated by central banks' wanting to defend their territory from cryptos and their relatives.

 Giving retail access to central bank assets

At the moment, when retail users (you and I) use any electronic money, be it via phone, card or direct payment, we use some form of commercial bank money. Our credit risk is the risk of the bank. Typically, when banks settle between each other, they do so using central bank money, that is, money backed by the central bank of where they are settling. In Europe, securities trades are also settled in central bank money, via T2S. When we use cash, though, we use central bank money.

A lot of the structure of modern finance derives from this structure. For example, we maintain bank accounts with commercial banks in part because central banks don't provide current account services. If they did, they would probably be highly valued for their credit quality. Commercial banks, in turn, typically use retail deposits to fund lending activity.

The idea behind a CBDC is that, in some form, it would give retail users access to a central bank asset. This is a highly radical decision, we think. The authors of this section (Philip Middleton, Alastair Ryan and Adithya Metuku) have explored the consequences of this in the "Digital Money" series.

It's important to understand why central bankers have decided to move towards a CBDC. The clearest presentation here is the ECB's.

 What's the ECB doing?

The ECB is currently reviewing the replies to a consultative exercise about CBDC. It has said it will announce whether or not it will proceed to a trial of a CBDC by the middle of the year. We think it is likely that they do. The decision whether or not to proceed to a production version will, we presume, depend on the experience of the trial as well as other macro developments.

Why CBDC?

A good starting point here is a speech by ECB Executive Board member Fabio Panetta⁴. Mr Panetta writes that:

We are working to safeguard the role of sovereign money in the digital era: we want to be ready to introduce a digital euro, if needed. A digital euro would combine the efficiency of a digital payment instrument with the safety of central bank money. It would complement cash, not replace it. Together, these two types of money would be available to all, offering greater choice and access to simple, costless ways of paying.

This represents a positive agenda for CBDC. However, Mr Panetta also sets out a number of concerns which may weigh on central bankers.

Big tech, China

He talks about the potential for big techs to become major players in payment - he appears to be against this. "If not properly regulated, big techs may pose considerable risks from an economic and social perspective and they may restrict, rather than expand, consumer choice." Also, "Monetary sovereignty could be threatened if foreign central bank digital currencies became widely used in the euro area".

Eurocentric approach

He also points out that cards used across Europe rely on "agreements with international card schemes. As a result, people mostly use international schemes for cross-border card payments, and the European market for card payments is dominated by non-European schemes." Presumably, this includes MasterCard and Visa.

This suggests that major reasons for the ECB being interested in a digital Euro include the worry that the Facebook-led "Diem" digital currency might become a dominant currency, and that Europeans might progressively adopt another central bank's digital currency (such as the Chinese Yuan).

 What about crypto?

Mr Panetta's commentary on CBDC doesn't explicitly engage with cryptocurrencies. ECB President Christine Lagarde has made a range of comments which touch on cryptos, though. For example, she said recently in an interview on BFMTV⁵: