Title page 1

Contents 1

Abstract 2

Non-Technical Summary 3

1. Introduction 5

2. Data 9

2.1. Summary statistics 11

3. Drivers of crypto ownership 13

3.1. Addressing rare event bias 15

4. Payment usage of crypto-assets 16

4.1. A two-stage model for crypto adoption and payment use 16

4.2. Identification strategy 18

4.3. Results from the two-stage model 19

4.4. Discussion of results 21

5. Revisiting the relationship between holding fiat cash reserves and crypto ownership 22

5.1. Multiple instruments IV approach 22

5.2. Addressing the binary endogenous variable problem 24

5.3. Results 26

5.4. Discussion of results 28

6. Conclusion 30

References 33

Acknowledgements 39

Tables 11

Table 1. Descriptive statistics of crypto-asset ownership and usage 11

Table 2. Determinants of Crypto Ownership 14

Table 3. Robustness of Sample Selection Model to Identification Strategy 19

Table 4. Determinants of Crypto Ownership and Payment Usage 20

Table 5. IV models results: Effect of use of cash as store of value on selected instruments Model 27

Figures 12

Figure 1. Younger cohorts hold more crypto, with high country variability 12

Figure 2. Younger generations lead crypto-asset holdings 12

Figure 3. Coefficients from probit, standard logit, and Firth logit. Similar coefficients of logit and Firth logit indicate that bias-reduced... 16

Figure 4. COVID-19 shocks used as instruments for cash-holding behavior. Supply-side constraints: reduced ATM access and merchant... 23

Figure 5. Aggregate EUR currency demand and Bitcoin market cap (2018-2021). The inverse relationship during the March 2020... 29