ECB-比特幣與法定貨幣交易的全球和本地驅(qū)動(dòng)因素-英

WorkingPaperSeriesGlobalandlocaldriversofBitcointradingvis-à-visfiatcurrenciesPaolaDi

Casola,MaurizioMichaelHabib,David

Tercero-LucasNo2868Disclaimer:

Thispapershouldnot

bereportedasrepresentingtheviewsof

theEuropeanCentralBank(ECB).

TheviewsexpressedarethoseoftheauthorsanddonotnecessarilyreflectthoseoftheECB.AbstractWe

analyse

the

drivers

of

Bitcoin

transactions

against

44

?at

currenciesin

the

largest

peer-to-peer

crypto

exchanges.

Momentum

and

volatility

inthe

cryptoasset

market,

as

well

as

volatility

and

liquidity

in

global

?nancialmarkets

do

matter

for

Bitcoin

trading.

There

is

suggestive

evidence

of

aglobal

crypto

cycle

driven

by

speculative

motives.

However,

in

emergingand

developing

economies

(EMDEs),

Bitcoin

seems

to

o?er

also

transactionalbene?ts,

since

trading

increases

when

the

value

of

the

domestic

currency

isunstable.

Proxies

of

banking

depth

and

digitalisation

are

negatively

correlatedwith

the

currency

loadings

on

the

global

factor,

indicating

that

crypto-assetsmay

o?er

a

speculative

alternative

to

traditional

?nance

when

this

is

notavailable,

especially

in

EMDEs

where

the

share

of

younger

risk-prone

populationis

higher.

Our

results

clearly

point

to

potential

?nancial

stability

risks

fromcryptoisation

in

EMDEs

with

low

levels

of

?nancial

development

and

unstable?at

currencies.JEL

Classi?cation:

E42

F21

F24

F32

F38

G15

O33Keywords:

digital

currencies,

Bitcoin,

peer-to-peer

exchanges,

?nancialdevelopment.ECBWorkingPaperSeriesNo28681Non-technical

summaryThe

popularity

of

Bitcoin

and

other

cryptocurrencies

has

not

been

con?ned

tofew

economies,

but

it

has

morphed

into

a

global

phenomenon,

rapidly

spreadingto

economies

with

disparate

levels

of

economic

development

and

?nancial

literacy.Notably,

Emerging

and

Developing

Economies

(EMDEs)

are

at

the

forefront

ofcrypto

adoption.

There

a

number

of

potential

reasons

behind

the

growing

popularityof

Bitcoin

and

other

cryptocurrencies

in

EMDEs.

First,

cryptocurrencies

may

beused

as

speculative

assets,

which

may

be

particular

attractive

to

investors

fromcountries

where

the

portfolio

choice

of

investment

assets

is

restricted

by

regulatoryor

institutional

factors.

Second,

even

though

prices

have

been

very

volatile,

thesecryptocurrencies

may

represent

a

better

store

of

value

with

respect

to

the

domesticcurrency

of

countries

where

in?ation

is

high

and

the

exchange

rate

tends

to

depreciate.Third,

residents

from

EMDEs

may

use

cryptocurrencies

as

a

means

of

payment

incross-border

transactions

to

circumvent

capital

controls

or

to

lower

the

cost

ofreceiving

remittances

from

abroad.

How

this

range

of

explanations

and

drivers

ofcrypto

adoption

maps

into

the

cross-section

of

economies

and

their

characteristicshas

so

far

received

limited

attention,

largely

due

to

data

constraints

and

the

inherentdi?culty

to

track

the

?nal

owners

of

cryptocurrencies.In

this

study,

we

overcome

the

obstacle

of

limited

country-by-country

informationby

looking

at

?at

currency

transactions

against

Bitcoin.

The

implied

assumption

isthat

those

trading

currencies

that

are

not

major

international

currencies,

in

particularcurrencies

of

EMDEs,

are

residents

of

the

countries

issuing

that

currency.

We

supportthis

assumption

for

those

transactions

where

data

about

the

residence

of

the

tradersof

Bitcoin

are

available.

Compared

to

previous

studies,

we

analyse

transactionstaking

place

in

peer-to-peer

(P2P)

exchanges,

outside

the

blockchain

network

(i.e.they

are

o?-chain).

These

P2P

exchanges

have

an

important

peculiarity:

they

targetmainly

small

retail

users.

In

these

P2P

exchanges,

bid-ask

spreads

tend

to

be

large

sothat

these

exchanges

are

usually

not

a?ected

by

the

problem

of

market

manipulation,such

as

wash

trading,

typical

of

centralized

exchanges,

thus

making

the

transactionswe

analyse

more

reliable.

The

growth

in

cryptocurrency

markets

in

recent

years

hastaken

place

particularly

through

o?-chain

transactions

and,

in

EMDEs,

especiallyin

P2P

exchanges.1

Speci?cally,

we

analyse

trading

volumes

of

Bitcoin

versus

thecurrencies

of

14

Advanced

Economies

(AEs)

–

excluding

the

US

dollar

due

to

its1While

on-chain

transactions

occur

on

the

blockchain

network

and

need

to

be

validated

byminers,

o?-chain

transactions

are

conducted

outside

the

blockchain

network,

making

them

-

ingeneral

-

faster

and

cheaper.

O?-chain

transactions

may

take

place

in

centralised

exchanges

thatact

as

an

intermediary,

or

peer-to-peer

exchanges

that

only

match

o?ers

from

buyers

and

sellersbut

do

not

act

as

intermediaries.ECBWorkingPaperSeriesNo28682special

status

–

and

the

currencies

of

30

EMDEs.

Data

are

obtained

from

the

largestP2P

exchanges,

namely

LocalBitcoins

and

Paxful,

over

the

period

January

2018

-April

2022,

on

a

weekly

basis.Our

results,

overall,

reinforce

the

hypothesis,

currently

prevailing

in

the

literature,that

Bitcoin

trading

is

driven

by

speculative

motives.

In

this

paper,

we

show

thatthis

is

truly

a

global

phenomenon.

There

is

suggestive

evidence

of

a

global

cryptocycle

in

Bitcoin

trading

against

?at

currencies,

with

transactions

across

currenciesand

users

around

the

world

moving

in

tandem

with

?uctuations

in

the

Bitcoin

price.Similarly

to

other

risky

assets,

momentum

in

the

crypto-asset

market,

past

Bitcoinprice

volatility

as

well

as

global

?nancial

market

volatility

do

matter

for

Bitcointrading

against

di?erent

?at

currencies.However,

Bitcoin

seems

to

o?er

also

speci?c

transactional

bene?ts,

in

particularin

EMDEs.

The

depreciation

of

the

domestic

currency

of

EMDEs

–

notably

not

ofthe

currency

of

AEs

–

induces

more

Bitcoin

trading,

in

particular

after

the

COVID-19pandemic.

This

indeed

suggests

that

Bitcoin,

despite

its

wide

price

?uctuations,might

have

been

appreciated

also

as

a

store

of

value

or

medium

of

exchange

incountries

which

experienced

a

loss

in

the

the

purchasing

power

of

their

domesticcurrency.

In

turn,

this

implies

that

macroeconomic

instability

may

potentially

spurgreater

cryptoasset

usage.

This

result

is

important

for

the

asset

pricing

theory

ofcryptoassets,

suggesting

that

the

fundamental

value

of

Bitcoin

may

be

substantiallydi?erent

between

AEs

and

EMDEs,

since

its

transactional

services

are

probably

moreelevated

in

less

developed

economies.

Moreover,

we

?nd

that

proxies

of

bankingdepth

and

digitalisation

are

negatively

correlated

with

the

extent

to

which

eachcurrency

loads

on

the

global

common

factor

in

Bitcoin

trading

volumes,

indicatingthat

crypto-assets

may

o?er

a

speculative

alternative

to

traditional

?nance

whenthis

is

not

available,

in

particular

in

EMDEs

where

the

share

of

younger

risk-pronepopulation

is

higher,

another

important

?nding

of

our

analysis.Our

?ndings

clearly

point

to

potential

?nancial

stability

risks

in

EMDEs

withlow

levels

of

?nancial

development

and

unstable

?at

currencies.

The

intrinsic

pricevolatility

of

Bitcoin

may

discourage

its

use

as

a

store

of

value

or

means

of

payment;however,

in

the

future,

other

crypto

assets,

such

as

stablecoins

that

pledge

to

ensurea

parity

to

the

US

dollar

or

other

reserve

currencies,

might

become

more

widely

usedby

individuals

and

?rms

in

order

to

compensate

for

the

lack

of

?nancial

alternatives.ECBWorkingPaperSeriesNo286831

IntroductionTo

what

extent

is

Bitcoin

usage

a

global

phenomenon

driven

by

speculative

demand?To

what

extent

can

country-speci?c

factors

explain

the

use

of

Bitcoin?

What

drivesthe

adoption

of

an

unbacked

digital

currency

like

Bitcoin?

These

are

importantquestions

that

so

far

have

received

only

partial

answers,

largely

due

to

the

di?cultyto

trace

who

owns

and

trades

cryptocurrencies.In

this

study,

we

overcome

the

obstacle

of

limited

country-by-country

informationon

cryptocurrency

use

by

looking

at

?at

currency

transactions

against

Bitcoin.1Compared

to

previous

studies,

we

analyse

transactions

taking

place

in

peer-to-peer(P2P)

exchanges

that

perform

transactions

outside

the

blockchain

network

(i.e.

theyare

o?-chain)

and

in

a

decentralised

manner

(see

Section

2).

These

P2P

exchangeshave

an

important

peculiarity:

they

target

mainly

small

retail

users.2

In

these

P2Pexchanges,

bid-ask

spreads

tend

to

be

large

so

that

these

exchanges

are

usually

nota?ected

by

the

problem

of

market

manipulation,

such

as

wash

trading,

typical

ofcentralised

exchanges,

thus

making

the

transactions

we

analyse

more

reliable.3

Asshown

in

Figure

1,

the

growth

in

cryptocurrency

markets

in

recent

years

has

takenplace

particularly

through

o?-chain

transactions

and

in

emerging

and

developingeconomies

(EMDEs),

especially

in

P2P

exchanges

(see

Section

2).Speci?cally,

we

analyse

trading

volumes

of

Bitcoin

versus

the

currencies

of

14advanced

economies

(AEs)

–

excluding

the

US

dollar

due

to

its

special

status

–

andthe

currencies

of

30

EMDEs.

Data

are

obtained

from

the

largest

P2P

exchanges,namely

LocalBitcoins

and

Paxful,

over

the

period

January

2018

-

April

2022,

ona

weekly

basis.

First,

we

study

the

impact

of

a

number

of

crypto-speci?c

drivers,global

drivers

and

local

drivers

on

Bitcoin

transactions

in

a

?xed

e?ect

dynamicpanel

model

in

order

to

understand

the

motivations

of

Bitcoin

trading.

In

particular,we

investigate

whether

Bitcoin

transactions

have

been

driven

by

(i)

trends

that

arespeci?c

to

the

crypto-market

and

may

be

ascribed

to

the

demand

factors

highlightedby

Biais

et

al.

(2023);

(ii)

trends

that

are

related

to

the

traditional

?nancial

system,such

as

developments

in

global

?nancial

markets

and

liquidity,

global

macroeconomic1The

implicit

assumption

is

that

those

trading

currencies

that

are

not

major

internationalcurrencies,

in

particular

currencies

of

emerging

and

developing

economies,

are

residents

of

thecountries

issuing

that

currency.

We

shall

support

this

assumption

for

those

transactions

where

dataabout

the

residence

of

the

traders

of

Bitcoin

are

available.2See

Chainanalysis

(2021),

page

10.

Graf

von

Luckner

et

al.

(2023)

calculate

that

the

averagetrade

size

in

one

of

the

P2P

platforms

we

study,

Paxful,

is

around

USD

150.3Wash

trading

is

the

problem

of

having

investors

simultaneously

selling

and

buying

the

same?nancial

assets

to

create

arti?cial

activity

in

the

marketplace,

which

is

known

to

distort

price,volume,

and

volatility,

and

reduce

investors’

con?dence

and

participation

in

?nancial

markets.According

to

Cong

et

al.

(Forthcoming),

analysing

a

sample

of

unregulated

exchanges,

wash

tradingis

a

serious

problem,

with

the

reported

volumes

in?ated

on

average

by

over

70%.ECBWorkingPaperSeriesNo28684Figure

1:

Bitcoin

price,

on-chain

and

o?-chain

volumes.Source:

CoinMetrics

and

CoinMarketCap.conditions

or

geopolitical

events,

similarly

to

analyses

of

foreign

exchange

tradingvolumes

(Cespa

et

al.,

2022);

and,

?nally,

(iii)

country-speci?c

variables

that

re?ectthe

weakness

of

the

institutional

and

macroeconomic

framework,

which

may

in?uencethe

transactional

services

of

Bitcoin

and

its

fundamental

value.

Second,

as

the

panelanalysis

?nds

that

there

is

a

large

share

of

variation

in

Bitcoin

transactions

that

iscommon

across

di?erent

currencies,

we

use

a

static

factor

model

to

identify

commonfactors

in

Bitcoin

trading

against

di?erent

?at

currencies

and

analyse

their

features.Finally,

we

turn

our

attention

to

EMDEs,

including

in

our

analysis

a

large

numberof

economic

and

institutional

variables

that,

according

to

the

literature,

might

beassociated

with

Bitcoin

usage,

studying

which

country

features

correlate

with

thecurrency

loading

on

the

main

global

component

of

Bitcoin

trading.We

?nd

that

momentum

and

volatility

in

the

crypto-asset

market

as

well

as

global?nancial

market

volatility

and

liquidity

do

matter

for

Bitcoin

trading

against

di?erent?at

currencies.

However,

in

a

panel

setting,

a

global

component

of

Bitcoin

tradingstill

needs

to

be

identi?ed,

since

our

crypto-assets

and

global

drivers

fail

to

capturethe

full

extent

of

co-movement

in

Bitcoin

trading

across

currencies

and

over

time.Indeed,

the

factor

analysis

identi?es

a

single

global

factor

that

on

average

explainsup

to

around

40%

of

the

variance

of

Bitcoin

trading

across

di?erent

currencies

in

theCOVID-19

period.

There

is

therefore

suggestive

evidence

of

a

global

crypto

cycle

inBitcoin

transactions,

echoing

the

?ndings

of

the

global

?nancial

cycle

literature

forECBWorkingPaperSeriesNo28685traditional

asset

markets

(Miranda-Agrippino

and

Rey,

2022).

This

global

factor,in

turn,

is

correlated

with

the

Bitcoin

price,

as

shown

in

Figure

2.

Trading

acrosscurrencies

and

users

around

the

world

moves

in

tandem

with

?uctuations

in

theBitcoin

price,

suggesting

that

Bitcoin

is

largely

used

as

a

speculative

investmentasset

across

advanced,

emerging

and

developing

economies.Figure

2:

A

global

factor

in

Bitcoin

trading

volumes

against

?at

currencies

iscorrelated

with

the

Bitcoin

priceThe

?gure

shows

the

?rst

factor

extracted

from

the

trading

volume

of

Bitcoin

transactions

against?at

currencies

(blue

line)

and

the

change

in

the

Bitcoin

price

(yellow

line).

See

section

4.2

forfurther

details.

Trading

volumes

and

the

Bitcoin

price

are

detrended

with

the

log

di?erence

withrespect

to

the

past

15-week

moving

average.However,

Bitcoin

seems

to

o?er

also

transactional

bene?ts,

in

particular

inEMDEs.

The

trading

of

Bitcoin

against

?at

currencies

of

EMDEs

is

somewhatdi?erent

from

that

against

currencies

of

AEs.

The

impact

of

momentum

in

thecrypto-asset

market

is

particularly

strong

in

EMDEs

in

the

period

of

the

COVID-19pandemic.

Importantly,

one

local

driver,

the

depreciation

of

the

domestic

currencyversus

the

US

dollar,

plays

a

role

in

encouraging

more

Bitcoin

trading

against

thecurrencies

of

EMDEs,

again

in

particular

since

the

start

of

the

COVID-19

pandemic.This

?nding

suggests

that

Bitcoin

might

be

used

as

a

store

of

value

or

mediumof

exchange

in

countries

which

experience

a

loss

in

the

purchasing

power

of

theirdomestic

currency.

Finally,

the

extent

to

which

each

currency

of

EMDEs

loads

onthe

global

factor

–

the

factor

more

closely

associated

with

the

Bitcoin

price

–

isnegatively

correlated

with

the

number

of

ATMs

or

the

di?usion

of

digital

paymentsECBWorkingPaperSeriesNo28686in

the

countries

issuing

the

respective

currency

and

positively

with

a

higher

share

ofyounger

population.

Possibly,

the

lack

of

a

developed

?nancial

system,

providinge?cient

payment

rails

and

the

ability

to

diversify

the

?nancial

portfolio,

encouragesthe

youngest

cohorts

of

the

population

in

EMDEs

to

take

the

risk

of

a

cryptoassetinvestment.Our

results

point

towards

?nancial

stability

risks

associated

with

cryptocurrencyspeculation.

This

is

exacerbated

by

the

limited

consumer

protection

and

high

levelof

opaqueness

in

cryptomarkets,

as

suggested

by

evidence

on

price

manipulationand

insider

trading

(Gandal

et

al.,

2018;

Gri?n

and

Shams,

2020).

Moreover,

therather

high

price

volatility

of

the

cryptoasset

markets

has

been

punctuated

by

marketcrashes,

while

the

link

with

the

mainstream

?nancial

system

has

been

increasing.4Notably,

our

results

indicate

that

?nancial

stability

risks

could

be

more

pronouncedin

EMDEs

with

low

levels

of

?nancial

development

and

unstable

?at

currencies.

Inthese

countries,

Bitcoin

–

or,

more

likely,

other

crypto-assets

like

stablecoins

in

thefuture

-

might

become

widely

used

by

individuals

and

?rms

for

ordinary

transactionsor

as

a

store

of

value,

in

order

to

compensate

for

the

lack

of

?nancial

alternatives.5These

developments

raise

the

risk

of

cryptoisation

(IMF,

2021)

–

i.e.

the

substitutionof

the

domestic

currency

with

a

cryptocurrency

in

the

similar

fashion

as

the

USdollar

is

widely

used

in

countries

with

high

in?ation

–

and

represent

a

threat

tothe

implementation

of

capital

?ow

management

policies

in

these

countries

(He

etal.,

2022).

To

a

large

extent,

our

results

and

policy

implications

concern

EMDEs,as

decentralised

P2P

exchanges

found

fertile

ground

in

less

developed

economies,whereas

residents

of

AEs

tend

to

use

o?-chain

centralised

exchanges,

which

arebeyond

the

scope

of

this

study.

However,

?nancial

stability

risks

stemming

fromcryptocurrencies,

in

general,

are

not

con?ned

to

EMDEs,

as

US-based

evidencepoints

towards

substantial

spillover

e?ects

from

cryptocurrencies

on

the

real

economythrough

consumption

and

investment

into

other

asset

classes

(Aiello

et

al.,

2023).Our

study

contributes

to

a

growing

literature

on

the

drivers

of

Bitcoin

and

othercrypto

assets

usage.

Bitcoin

was

created

with

the

aim

of

providing

an

alternativepayment

system

that

would

operate

in

a

decentralised

way,

free

of

the

control

ofa

third

party

or

an

authority.

Notably,

like

many

other

crypto-assets,

Bitcoin

isnot

backed

by

any

real

asset

or

any

governmental

claims

(Halaburda

et

al.,

2022).64For

instance,

the

crash

of

the

algorithmic

stablecoin

TerraUSD

in

May

2022,

analysed

in

Uhlig(2022).

Iyer

(2022)

provide

evidence

on

the

increased

interconnection

between

cryptoasset

andequity

markets

across

economies

over

time,

while

Karau

(2023)

shows

that

Bitcoin

prices

respondto

monetary

policy

shocks

similarly

to

stock

prices

since

the

COVID-19

pandemic.5Stablecoins

are

digital

assets

designed

to

minimise

price

volatility

typically

against

a

single

?atcurrency

like

the

US

dollar

(or

a

basket

of

?at

currencies

or

reserve

assets).6According

to

the

de?nition

of

the

Financial

Stability

Board,

crypto-assets

like

Bitcoin

are

atype

of

private

sector

digital

asset

that

depend

primarily

on

cryptography

and

distributed

ledger

orECBWorkingPaperSeriesNo28687Considering

the

apparent

lack

of

a

fundamental

value,

the

exponential

growth

inthe

volume

of

transactions

involving

Bitcoin

and

in

its

price

is

certainly

surprising(see

Figure

1).

This

in

turn

has

generated

a

lively

and

fast-growing

debate

amongeconomists

about

the

motivations

behind

the

use

of

Bitcoin.

After

an

initial

usageby

a

small

community

of

experts,

the

usage

of

Bitcoin

has

been

subsequently

drivenby

the

black

market

of

illegal

goods

and

services

and

by

gambling

(Foley

et

al.,

2019;Marmora,

2021).

More

recently,

however,

the

soaring

popularity

of

cryptocurrencyexchange

markets

and

Bitcoin

might

have

also

been

driven

by

speculative

motives(Baur

et

al.,

2018)

or,

possibly,

by

its

potential

use

for

cross-border

transactions

andtransfers

(Graf

von

Luckner

et

al.,

2023;

Makarov

and

Schoar,

2021),

although

thehigh

volatility

of

its

price

makes

Bitcoin

impractical

as

a

medium

of

exchange

(Baurand

Dimp?,

2021).7The

popularity

of

Bitcoin

and

other

cryptocurrencies

has

not

been

con?ned

tofew

economies,

but

it

has

morphed

into

a

global

phenomenon,

rapidly

spreadingto

economies

with

disparate

levels

of

economic

development

and

?nancial

literacy.Notably,

EMDEs

are

at

the

forefront

of

crypto

adoption.

According

to

Chainanalysis(2022),

among

the

top

20

countries

with

the

highest

crypto

adoption

index,

thereare

only

two

AEs,

the

United

States

and

United

Kingdom,

whereas

the

remainingcountries

are

all

EMDEs

from

Asia,

Africa,

Europe

or

Latin

America.8There

a

number

of

potential

reasons

behind

the

growing

popularity

of

Bitcoin

andother

cryptocurrencies

in

EMDEs.

First,

cryptocurrencies

may

be

used

as

speculativeassets,

which

may

be

particular

attractive

to

investors

from

countries

where

theportfolio

choice

of

investment

assets

is

restricted

by

regulatory

or

institutionalfeatures.

Second,

even

though

prices

have

been

very

volatile,

these

cryptocurrenciesmay

represent

a

better

store

of

value

or

medium

of

exchange

with

respect

to

thedomestic

currency

in

countries

where

in?ation

is

high

and

the

exchange

rate

tendssimilar

technology

(see

/work-of-the-fsb/?nancial-innovation-and-structural-change/crypto-assets-and-global-stablecoins/).

Crypto-assets

transactions

that

are

recorded

ona

distributed

ledger

-

the

blockchain

-

are

public

and

rely

on

consensus

mechanisms

instead

oftrusted

parties.

The

transactions

are

not

tied

to

real-world

entities

but

rather

crypto-addresses

(i.e.account

numbers)

whose

owners

are

not

explicitly

identi?ed,

thereby

ensuring

(pseudo)

anonymity.7Leveraging

on

a

recent

US

survey

of

consumers,

Auer

and

Tercero-Lucas

(2022)

show

thatcryptocurrency

investors

tend

to

be

highly

educated,

young,

male

and

digital

natives,

and

thatdistrust

in

regular

?nance

is

not

the

key

driver

of

investment

in

cryptocurrencies.

Using

data

froma

German

online

bank,

Hackethal

et

al.

(2021)

?nd

that

cryptocurrency

investors

are

active

traderswho

are

prone

to

investment

biases

and

hold

risky

portfolios,

tilting

their

portfolios

toward

evenmore

risky

securities

after

cryptocurrency

usage.

Weber

et

al.

(2023)

also

rely

on

survey

data

onUS

households

to

report

that

cryptocurrency

holders

tend

to

be

young,

white,

male

and

morelibertarian

relative

to

non-crypto

holders.8In

September

2021,

Bitcoin

was

even

adopted

as

legal

tender

in

El

Salvador

and

Central

AfricanRepublic.

However,

Alvarez

et

al.

(2022)

?nd

that

the

use

of

Bitcoin

for

everyday

transactions

inEl

Salvador

was

low,

based

on

a

national

survey

conducted

soon

after

the

adoption

of

Bitcoin

aslegal

tender.ECBWorkingPaperSeriesNo28688to

depreciate.

Third,

residents

from

EMDEs

may

use

c