The article examines volatility in cryptocurrency markets and how they relate to global stock markets and U.S. monetary policy. The researchers identified a single price component, dubbed the "crypto factor," that explained 80 percent of cryptocurrency price movements, and showed that its correlation with the stock market increased with the timing of institutional investor entry into the cryptocurrency market. match. The researchers also documented a similar phenomenon to equities, where tightening monetary policy from the Federal Reserve reduced the impact of crypto factors through the risk-taking channel, contrary to the notion that crypto assets provide a hedge against market risk. Finally, the researchers show that a sample heterogeneous agency model with time-varying overall risk aversion can explain their empirical results and highlight the potential for crypto markets to transmit risk to equity markets if institutional investor participation becomes large.

introduce

Cryptoassets differ significantly in design and value proposition, yet their prices exhibit common cyclical fluctuations. It surged from $20 billion in 2016 to nearly $3 trillion in November 2021, before plummeting below $1 trillion in the recent crypto "winter". Phases of exponential returns attract the attention of retail and institutional investors (Benetton and Compiani, 2022; Auer and Tercero-Lucas, 2021; Auer, Farag, Lewrick, Orazem, and Zoss, 2022), while subsequent crashes attract attention from politicians and growing attention from regulators. These swings in the crypto market may also be increasingly in sync with those of other asset classes: Bitcoin provided a partial hedge against market risk to some extent until 2020, but its correlation with the S&P 500 has gradually faded since then. Enhancement (Adrian, Iyer, and Qureshi, 2022).

However, little is known about the common drivers that affect cryptoasset prices and the factors that influence the correlation between cryptoassets and stock markets, including U.S. monetary policy. This article attempts to shed light on these issues by answering the following questions. To what extent are there common cycles among cryptoassets? Is the crypto market increasingly in sync with the global stock market? If yes, why is this happening? Given that U.S. monetary policy is considered a key driver of global financial cycles (Miranda-Agrippino and Rey, 2020), does U.S. monetary policy similarly affect the cyclicality of cryptoassets? If yes, through which channels?

We answer these questions by using dynamic factor models to identify dominant trends in cryptoasset prices. Using a daily price panel of seven tokens created prior to 2018, which collectively account for about 75% of the total crypto market cap, we decompose their volatility into asset-specific noise and AR(q) common components . We found that the resulting "encryption factor" explained about 80% of the variance in encrypted price data. This is much larger than the figure of 20% for global stocks calculated by Miranda-Agrippino and Rey (2020), and reflects the concentration of market capitalization of the largest cryptoassets relative to the largest stocks. This figure is robust across various lag orders q, and we find similarly high correlations when expanding the panel to include more cryptoassets.

In a second step, we examine the relationship of this crypto factor to a set of global equity factors constructed using stock indices of the largest countries by gross domestic product (GDP) (drawing on Rey, 2013; Miranda - Thoughts from Agrippino and Rey, 2020). We find positive correlations across the sample, particularly strong correlations since 2020. This growing co-movement is not limited to just between Bitcoin and the S&P 500, but involves crypto and global equity factors more broadly. In the segment of the equity market, we find that since 2020, the crypto factor has the strongest correlation with the global technology factor and the small cap factor, while the correlation with the global financial factor is surprisingly low.

The increased correlation between crypto assets and stocks coincides with the increased participation of institutional investors in the crypto market since 2020. Although institutions' exposure is small relative to their balance sheets, their absolute trading volume is much larger than that of retail traders. In particular, the trading volume of institutional investors on crypto exchanges increased by more than 1700% between the second quarter of 2020 and the second quarter of 2021 (from about $25 billion to more than $45 billion) (Auer et al., 2022 ). As institutional investors trade equities and cryptoassets, this leads to a gradual increase in the correlation between marginal equity and crypto investor risk allocations, which in turn leads to an increased correlation between global equity and crypto factors. Following the factor motion decomposition of Bekaert, Hoerova, and Lo Duca (2013), we find that the correlation between the overall effective risk aversion of cryptoassets and stocks can explain most (up to 65%) of the correlation between these two factors .

The increased correlation between cryptoassets and stocks coincides with the growth of institutional investor participation in the crypto market since 2020. Although institutional exposures are small relative to their balance sheets, their absolute trading volumes are far greater than retail traders. In particular, institutional investor trading volumes on crypto exchanges increased by more than 1700% between Q2 2020 and Q2 2021 (from roughly $25 billion to over $45 billion) (Auer et al. 2022). As institutional investors trade equities and cryptoassets, this leads to an incremental increase in risk profile correlations between marginal equities and crypto investor risk allocations, which in turn leads to an increased correlation between global equities and crypto factors. Following the factor-movement decomposition of Bekaert, Hoerova, and Lo Duca (2013), we find that the correlation of the overall effective risk aversion of cryptoassets and stocks can explain most (up to 65%) of the correlation between these two factors.

As U.S. monetary policy affects global financial cycles (Miranda-Agrippino and Rey, 2020), the high correlation between stocks and crypto suggests that it may have a similar impact on crypto markets. We test this hypothesis using a daily (vector autoregressive model), which includes the shadow federal funds rate (SFFR) proposed by Wu and Xia (2016), to account for the significant role of balance sheet policy over our sample period. We identify the impact of monetary policy shocks through a Cholesky decomposition, in the following order: SFFR; Treasury 10-year and 2-year spreads, reflecting expectations of future growth; U.S. dollar index, oil and gold prices, as international trade, credit and A proxy for commodity cycles; the VIX index, which reflects uncertainty about future expectations; and finally, stock and crypto factors. In this setup, endogeneity is less likely to be an issue, as the Fed is less likely to adjust its monetary policy in response to fluctuations in crypto prices, and such adjustments are less likely to occur on a daily level.

We find that U.S. monetary policy affects cryptocurrency cycles in the same way that it affects global equity cycles, in stark contrast to the claim that crypto assets provide a hedge against market risk. A one percentage point increase in the federal funds rate (SFFR) leads to a sustained 0.15 standard deviation drop in the crypto factor and a 0.1 standard deviation drop in the equity factor over the following two weeks. Interestingly, like the global financial cycle (Rey, 2013), we find that only the Federal Reserve’s monetary policy works, while those of other major central banks do not, likely reflecting the high use of the U.S. dollar by the crypto market.

We find that the risk-taking channel of monetary policy is an important channel driving these outcomes, similar to what Miranda-Agrippino and Rey (2020) find for global equity markets. In particular, we find that monetary tightening leads to a reduction in the crypto factor, accompanied by a surge in proxy measures of overall effective risk aversion in the crypto market. In other words, restrictive policies make investors’ risk positions less sustainable, so they reduce their exposure to crypto assets. When splitting the sample in 2020, we find that the effect of crypto market risk aversion is only significant in the period after 2020, consistent with the inclusion of institutional investors enhancing the transmission of monetary policy transmission to crypto markets. In a more formal test, we find the same results when testing the hypothesis using the smoothing transformation proposed by Auerbach and Gorodnichenko (2012), where the transformation variable is the share of institutional investors.

Next, we rationalize our results in a model that includes two classes of heterogeneous investors, namely crypto and institutional investors. The former are retail investors who only invest in crypto assets, while the latter can invest in stocks and crypto assets. The point is, crypto investors are risk averse, while institutional investors are risk neutral but face value risk constraints. We can rewrite the equilibrium return of cryptoassets as a linear combination of their variance and the covariance of stock returns, multiplied by the ratio of overall effective risk aversion. The latter can be interpreted as the average risk aversion of investors, weighted by their wealth weights. This means that the higher the relative wealth of institutional investors, the more similar the overall effective risk aversion of the crypto market is to its risk appetite, and the greater the correlation between crypto and stock markets. Since the presence of institutional investors in crypto markets reduces overall effective risk aversion, we explain the increased response of crypto prices to monetary tightening, which reflects the greater sensitivity of leveraged investors to economic cycles (Coimbra, Kim, and Rey, 2022; Adrian and Shin, 2014). Finally, we note that even in our simple framework, crypto-to-stock spillovers can emerge: If institutional crypto holdings become large, a collapse in crypto prices reduces equilibrium returns to stocks.

Overall, our findings highlight a remarkable synchronization between cryptocurrency cycles and global equity markets, with similar responses to monetary policy shocks. While many explanations exist for the value of cryptoassets, such as serving as an inflation hedge or providing more means of economic conversion, our findings suggest that U.S. monetary policy affects the cyclicality of cryptomarkets.

encryption factor

To summarize the volatility of the crypto market as a single variable, we use dynamic factor modeling, which is a dimensionality reduction technique. This allows us to decompose a set of prices into its specific components and a common trend. Specifically, we start with the daily prices of the largest cryptoassets created before January 2018 (excluding stablecoins). This leaves us with seven crypto assets representing 75% of the total market capitalization in June 2022. We then represent this cryptographic price panel as a linear combination of an AR(q) common factor ft and an asset-specific perturbation εit (the latter in turn follows an AR(1) process):

where L is the lag factor,

is the vector of order q of the factor loadings for asset i. Estimating this system using maximum likelihood, choosing q using the information criterion, yields our common factors. It is also possible to specify multiple factors that affect the price differently, and we use this latter specification when we consider several different subclasses of cryptoassets.

Figure 1 shows the crypto factor and the underlying price series from which we extract it. The crypto factor effectively captures well-characterized crypto market phases, such as the early 2018 decline, the ensuing "crypto winter," the latest boom in Bitcoin and Dogecoin, and the 2022 decline in Terra and FTX without being overly influenced The impact of isolated spikes like Ripple and TRON.

Figure 1 Encryption factor

Note: This graph shows the crypto factor (blue) and the normalized crypto price (grey) to construct it, generated using a dynamic factor model.

To more systematically assess the importance of this factor, we regress each price series on the encryption factor in turn. On average, 80% of the variation in the underlying series can be explained by our encryption factor. This figure is above 68% for all seven assets, underscoring the high degree of co-movement over our sample period. As a comparison, the global equity factor calculated by Miranda-Agrippino and Rey (2020) explains only 20% of global equity prices, highlighting the greater common movement and concentration of market capitalization in crypto markets. Thus, our findings strongly support the existence of a single crypto factor that drives prices across the crypto market.

Given the limited range of assets used to calculate our factors, we also confirmed that our crypto factors reflect more recent trends in newer assets. To do this, we examined a broader sample of assets, grouped into five categories: first-generation tokens (Bitcoin, Ripple, and Dogecoin), smart contract platform tokens (Ethereum, Binance Coin, Cardano, Solana, and Polkadot), DeFi tokens (Chainlink, Uniswap, Maker, and Aave), Metaverse tokens (Flow, Ape Coin, the Sandbox, Decentraland, and Theta Network), and IoT tokens (Helium, Iota, IoTex, and MXC). We then estimate a new model with five different factors, where each factor affects only one category. The results are shown in Figure 2, along with the general encryption factor estimated above. All categories are highly correlated with the general crypto cycle, validating our focus on common trends.

Figure 2 Encryption sub-factors

NOTE: The figure shows the overall encryption factor and five encryption subfactors, normalized and smoothed. These sub-factors are constructed from the following assets: first generation tokens - Bitcoin, Ripple and Dogecoin; smart contract platform tokens - Ethereum, Binance Coin, Cardano, Solana and Wave Card points; DeFi tokens – Chainlink, Uniswap, Maker, and Aave; Metaverse tokens – Flow, Ape Coin, the Sandbox, Decentraland, and Theta Network; and IoT tokens – Helium, Iota, IoTex, and MXC.

Finally, consistent with the case evidence, the crypto factor is associated with a proxy variable for crypto market leverage. Figure 3 plots the relationship between crypto factors and crypto leverage, defined here using the total value locked (TVL) in decentralized finance ("DeFi") contracts on a normalized benchmark of total crypto market capitalization. This shows a small relative leverage effect in the system until the end of the crypto "winter" of 2018-2019, after which leverage increased significantly and the correlation with the overall crypto factor increased.

Figure 3 Decentralized financial leverage

Note: This graph shows the overall crypto factor and an alternative metric representing total DeFi leverage, defined as the total value locked (TVL) in decentralized finance contracts, normalized to the total crypto market capitalization benchmark. TVL data from

Cryptocurrencies and the Global Financial Cycle

We now turn to the relationship between crypto factors and global equities. Iyer (2022) has documented an increase in the correlation between Bitcoin and S&P 500 returns since 2020. Therefore, we suspect that the crypto market has become more consolidated and in sync with the stock cycle. To assess this, in this section we compute a global equity factor and then study its relationship to the crypto factor.

We construct global equity factors using an all-equity index of the largest fifty countries ranked by GDP from Eikon/Thomson Reuters. We then compute as in the previous section: using aggregate factors for all major equity indices, factors for small-cap stocks, and individual factors for the technology and financial sectors. Figure 4 presents the stock and crypto factors. Like the crypto factor, the equity factor has reliably replicated the dynamics of global markets, including the sharp decline during the COVID-19 shock, the subsequent recovery, and the decline in early 2022. Overall, the correlation between the two series was relatively low until 2020, and then gradually increased from the second half of 2020 onwards. More formally, in Table 2, we regress changes in the encryption factor on changes in each of the other factors. Model (1) shows that, in general, the correlation between crypto factors and equity factors is very significant, while models (2) and (7) specifically emphasize that this relationship is driven in part by technology and small cap.

Figure 4 Cryptocurrency and Equity Factors

Note: The figure shows normalized time series for crypto and equity factors, derived from a wide range of crypto price and equity indices, respectively, using dynamic factor models, as described in Section 2.

Given the importance of institutions, we now investigate their role in changing the risk profile of marginal crypto investors. To investigate this empirically, we decompose factor changes into two components, following the methods of Bekaert et al. (2013) and Miranda-Agrippino and Rey (2020): (i) changes in market risk, and (ii ) changes in the market's attitude towards risk, that is, "overall effective risk aversion", defined as the weighted average risk aversion of investors. We use the proxy variable to achieve market risk (i), namely the 90-day variance of the MSCI World index measured according to the method of Miranda-Agrippino and Rey (2020), and then the remaining term of the logarithmic regression can be obtained by the following

(as its inverse function) to estimate (ii):

The same is true for cryptocurrencies:

in:

is a factor estimated using the method in equation (1) above; we repeated the MSCI World term in the crypto regression to control for overall global market risk; Similar proxy variables for market risk.

The effective equity risk aversion extracted in Equation (2) is consistent with other proxy variables for investor risk taking in the literature. The correlations between 90-day equity risk aversion and the square of the intermediary capital ratio and intermediary leverage ratio proposed by He, Kelly, and Manela (2017) (in Table A.4 of Appendix A) are -0.292 and 0.434, respectively. These proxy variables are explained as follows: When negative shocks affect the equity of intermediaries, their leverage ratios increase; thus, their risk-taking capacity is affected and effective risk aversion rises. These correlations are relatively high considering that He et al. (2017) used a very different approach and that we are comparing diurnal measurements. In fact, their proxy variable is constructed using only the capital ratios of the major dealers of the New York Fed, not from global equity prices (computed from global equity prices) (see Equation 6 of their paper).

Figure 5 shows the results for marginal crypto investors overall effective risk aversion, as well as the crypto factor. We identify two main phases, before and after the end of 2019. At the beginning of our sample, the effective risk aversion of crypto investors is more volatile and tends to increase slightly. Notably, this coincides with a “crypto winter,” an extended period of relatively flat or negative returns. After 2020, effective risk aversion declined relatively steadily, with crypto factors showing large returns and high volatility. Interestingly, since the May 2022 crash of Terra/Luna, the crypto factor has almost mirrored effective risk aversion, meaning that crypto prices are primarily driven by changes in crypto investors' risk appetite. Finally, we note that the decline in effective risk aversion corresponds to an increase in the participation of institutional investors, who can take on more risk than retail investors, thereby changing the risk appetite of marginal crypto investors.

Figure 5 Overall effective encryption risk avoidance

Note: The figure shows the cryptocurrency factor and the overall effective risk aversion in the cryptocurrency market, according to estimates by Bekaert et al. (2013) and Miranda-Agrippino and Rey (2020), as described in the text. Both variables are standardized.

Overall, our findings support the hypothesis that the entry of institutional investors is the main factor driving the increased correlation between the crypto market and the stock market. At the same time as many traditional financial institutions entered the crypto market, marginal crypto investors tended to be more risk-averse than marginal equity investors, and this correlation could in turn explain a significant portion of the correlation between crypto and equity factors .

Overall, our findings support the hypothesis that the entry of institutional investors is the main factor driving the increased correlation between the crypto market and the stock market. At the same time as many traditional financial institutions entered the crypto market, marginal crypto investors tended to be more risk-averse than marginal equity investors, and this correlation could in turn explain a significant portion of the correlation between crypto and equity factors .

in conclusion

Cryptoassets differ significantly in design and value proposition, yet their prices exhibit common cyclical fluctuations. A single crypto factor can explain 80% of crypto price movement, and since 2020, its correlation with the global financial cycle has strengthened, especially with technology and small-cap stocks. We provide evidence that this correlation is driven by the increased presence of institutional investors in the crypto market, which leads to similar risk profiles for marginal equity and crypto investors. Additionally, the crypto market is very sensitive to U.S. monetary policy, and monetary tightening would significantly reduce the crypto factor, similar to how global stock markets respond.

We outline a minimal theoretical framework capable of explaining our empirical results. We show that crypto returns can be expressed as a dynamic overall risk aversion function in crypto markets, which in turn is influenced by changes in the crypto investor base mix. With the increasing share of institutional investors in the crypto market, the risk-taking characteristics of crypto marginal investors tend to be similar to those of the equity market. A rise in the risk-free rate reduces returns, whereas if institutional investors hold a larger share of the crypto market and more leveraged proxies are more sensitive to economic cycles (Adrian and Shin, 2014; Coimbra et al., 2022), this This effect will become more and more significant.

Our findings also contribute to policy discussions on cryptoassets. We find that these assets do not provide a hedge against the economic cycle; instead, our estimates suggest that they are more sensitive than equities. Additionally, the increased correlation between crypto and the stock market, coupled with institutional investors trading both crypto assets and equities, means there could be potential spillover effects that could ultimately raise systemic risk concerns. In particular, our framework implies that in a possible future world where crypto makes up a substantial portion of institutional investors' portfolios, a crypto market crash could have important negative effects on equity markets. For these reasons, policymakers can take advantage of the fact that institutional investor exposure to crypto remains limited to develop and implement a more robust regulatory framework.