區(qū)塊鏈應(yīng)用的經(jīng)濟(jì)價(jià)值-英

Thisarticlewasdownloadedby:[2001:da8:d800:b048:d50e:2864:5976:2131]On:26January2024,At:04:37Publisher:InstituteforOperationsResearchandtheManagementSciences(INFORMS)INFORMSislocatedinMaryland,USAManagementSciencePublicationdetails,includinginstructionsforauthorsandsubscriptioninformation:TheEconomicValue

ofBlockchainApplications:EarlyEvidencefromAsset-BackedSecuritiesXiaChen,QiangCheng,Ting

LuoTo

citethisarticle:XiaChen,QiangCheng,Ting

Luo(2024)TheEconomicValue

ofBlockchainApplications:EarlyEvidencefromAsset-BackedSecurities.ManagementScience70(1):439-463./10.1287/mnsc.2023.4671Full

termsandconditionsofuse:/Publications/Librarians-Portal/PubsOnLine-Terms-and-ConditionsThisarticlemaybeusedonlyforthepurposesofresearch,teaching,and/orprivatestudy.

Commercialuseorsystematicdownloading(byrobotsorotherautomaticprocesses)isprohibitedwithoutexplicitPublisherapproval,unlessotherwisenoted.For

moreinformation,contactpermissions@.ThePublisherdoesnotwarrantorguaranteethearticle’saccuracy,

completeness,merchantability,fitnessforaparticularpurpose,ornon-infringement.Descriptionsof,

orreferencesto,productsorpublications,orinclusionofanadvertisementinthisarticle,neitherconstitutesnorimpliesaguarantee,endorsement,orsupportofclaimsmadeofthatproduct,publication,orservice.Copyright?2023,INFORMSPleasescrolldownforarticle—itisonsubsequentpagesWith12,500membersfromnearly90countries,INFORMSisthelargestinternationalassociationofoperationsresearch(O.R.)andanalyticsprofessionalsandstudents.INFORMSprovidesuniquenetworkingandlearningopportunitiesforindividualprofessionals,andorganizationsofalltypesandsizes,tobetterunderstandanduseO.R.andanalyticstoolsandmethodstotransformstrategicvisionsandachievebetteroutcomes.FormoreinformationonINFORMS,itspublications,membership,ormeetingsvisitMANAGEMENT

SCIENCEVol.

70,

No.

1,

January

2024,

pp.

439–463ISSN

0025-1909

(print),

ISSN

1526-5501

(online)/journal/mnscThe

Economic

Value

of

Blockchain

Applications:

Early

Evidencefrom

Asset-Backed

SecuritiesXia

Chen,a

Qiang

Cheng,a,*

Ting

Luoba

Singapore

Management

University,

Singapore

178900;

b

Tsinghua

University,

Beijing,

100084,

People’s

Republic

of

China*Corresponding

authorContact:xchen@.sg

(XC);

qcheng@.sg,

/0000-0001-5905-1740

(QC);

luot@,/0000-0002-8648-2578

(TL)Received:

April

30,

2021Revised:

December

16,

2021,

April

6,

2022Accepted:

April

25,

2022Published

Online

in

Articles

in

Advance:January

26,

2023Abstract.

In

this

paper,

we

evaluate

the

economic

value

of

a

blockchain

application.

In

thecontext

of

asset-backed

securities

(ABS)

issuance

in

China,

where

some

ABS

are

issuedwith

blockchain

technology

and

others

are

not,

we

?nd

that

the

use

of

blockchain

signi?-cantly

reduces

the

coupon

yield

at

issuance.

Compared

with

other

ABS,

those

issued

usingblockchain

technology

experiencea

decrease

of

31.4

basis

points

in

the

yield

spread,

whichcorresponds

to

a

relative

decrease

of

13%.

We

further

document

that

the

effect

of

block-chain

is

more

pronounced

for

ABS

deals

rated

by

less

reputable

credit

rating

agencies

andagencies

that

rely

more

on

issuers

for

their

rating

business,

for

revolving

ABS,

and

for

ABSwith

a

larger

number

of

underlying

assets.

We

also

?nd

that

the

use

of

blockchain

canreduce

the

level

of

retained

interest

and

number

of

credit

enhancement

mechanisms.

Thispaper

contributes

to

the

literature

by

providing

a

small-sample

analysis

of

the

economicvalue

of

a

blockchain

application

in

?nancial

markets./10.1287/mnsc.2023.4671Copyright:

?

2023

INFORMSHistory:

Accepted

by

Brian

Bushee,accounting.Funding:

X.

Chen

and

Q.

Cheng

acknowledge

funding

provided

by

the

Lee

Kong

Chian

Professorshipat

Singapore

Management

University.

This

work

was

supported

by

Singapore

Ministry

of

Educa-tion

[Grant

MOE-T2EP40120-0005].Supplemental

Material:

Data

are

available

at

/10.1287/mnsc.2023.4671.Keywords:

blockchain

asset-backed

securities

information

asymmetry??1.

Introductiontechnology

in

other

areas.

Examples

include

producttracing

by

e-commerce

?rms,

property

registration

andrecord

keeping

by

local

governments

as

in

the

states

ofArizona

and

Delaware

in

the

United

States,

and

the

pro-cessing

of

equity

transactions

as

done

by

the

AustraliaStock

Exchange

(e.g.,

Organisation

for

Economic

Cooper-ation

and

Development

2018,

Bayly

2019).The

bene?ts

of

blockchain

indicate

that

the

technologyis

well-suited

to

security

offerings,

registration,

and

set-tlement.

As

a

result,

many

governments

and

stock

ex-changes

are

studying

the

use

of

blockchain

technology

inthese

areas

(e.g.,

Caytas

2017).

Thus,

understanding

theeconomic

value

of

blockchain

applications

in

?nancialmarkets

is

of

utmost

importance.

The

Organisation

forEconomic

Cooperation

and

Development

(2018,

p.

31)comments

that

“further

research

and

dissemination

onresults

of

pilot

initiatives

using

blockchain

technology

isessential

to

give

regulators

and

policy

makers

a

strongerposition

to

assess

blockchain’s

potential

for

equity

mar-ket

activities.”

However,

some

experts

question

the

eco-nomic

value

of

blockchain

applications

(Cheng

et

al.2019,

McKinsey

and

Company

2019).

For

example,

tech-nology

expert

Tim

Bray,

formerly

of

Amazon,

commen-ted

in

2019

that

“if

[blockchain

is]

actually

generallyThe

development

of

blockchain

technology

has

fosteredmuch

excitement

about

its

potential

applications.

Becauseof

the

bene?ts

of

using

blockchain—the

accuracy

andreliability

of

data

maintained

on

blockchain—many

cor-porations

and

governments

are

exploring

blockchainapplications.

However,

given

the

high

development

andimplementation

costs

of

blockchain

technology,

it

is

criti-cal

to

empirically

investigate

the

economic

value

of

itsapplications.

Thus

far,

few

studies

have

doneso

in

a

com-mercial

environment

or

?nancial

markets

largely

becauseof

a

lack

of

data.

In

this

study,

we

exploit

the

context

ofasset-backed

securities

(ABS)

issuance

in

China,

wheresome

ABS

are

issued

using

blockchain

technology

andothers

are

not,

to

provide

empirical

evidence

on

the

valueof

a

blockchain

application.1

More

speci?cally,

we

investi-gate

whether

ABS

issued

using

blockchain

technologyhave

lower

yields

than

other

ABS.Blockchain

is

a

cryptographically

secure

ledger.

Itmaintains

areliable,

sharable,

and

time-stamped

recordof

transactions,

ownership,

and

rights

(e.g.,

Tinn

2017).Although

cryptocurrencies

are

the

most

well-knownapplications

of

blockchain,

corporations

and

govern-ments

have

either

begun

using

or

plan

to

use

this439Chen,

Cheng,

and

Luo:

The

Economic

Value

of

Blockchain

Applications440Management

Science,

2024,

vol.

70,

no.

1,

pp.

439–463,

?

2023

INFORMSuseful,

someone

somewhere

should

be

getting

dramati-cally

good

returns

on

these

investments…

I

went

look-ing

for

real

actual

production

applications,

and

lookedhard,

and

what

I

found

was

laughable.”2

Thus,

deter-mining

whether

ABS

issued

using

blockchain

technologyhave

lower

yields

than

other

ABS

can

contribute

to

thisdebate.We

test

this

prediction

in

the

setting

of

a

blockchainapplication

for

ABS

issuance

in

China.

Because

of

thegrowth

of

e-commerce

and

investment

in

blockchaintechnology

by

large

corporations,

some

Chinese

?rmsuse

blockchain

when

issuing

ABS.4

Our

sample

con-sists

of

10,287

tranches

of

4,390

ABS

deals

issuedbetween

January1,

2015,and

June30,

2021.

Bysearch-ing

for

ABS

issuance

news

online,

we

identify

14

ABSdeals

issued

using

blockchain

technology,

consistingof25tranches.Tranche-levelanalysesindicatethattheuse

of

blockchain

in

ABS

issuance

can

signi?cantlyreduce

the

yield

spread

of

ABS

tranches

(measured

asthe

yield

of

the

tranches

minus

the

yield

of

the

T-billswith

the

closest

maturity).

In

the

analysis,

we

controlfor

ABS

characteristics

that

might

affect

the

yieldspread

and

the

issuer

and

issuance

year–month

?xedeffects.

The

effect

is

economically

signi?cant

withtranches

of

blockchain-based

ABS

deals

having

anaverage

yield

spread

31.4

basis

points

lower

than

othertranches,

which

corresponds

to

a

relative

decrease

of13%

from

the

sample

mean.

This

decrease

translatesinto

a

saving

of

RMB9.6

million

in

interest

paymentsover

the

life

of

an

average

ABS

deal

withaprincipal

ofRMB1,181

million.5We

conduct

a

number

of

sensitivity

tests

to

ensurethat

the

results

are

robust.

First,

to

ensure

that

the

dif-ferences

in

characteristics

between

ABS

issued

usingblockchainand

otherABSdonotdrive

ourresults,weuse

an

entropy

balancing

approach

and

obtain

thesame

inferences.

Second,

we

obtain

the

same

infer-ences

using

a

sample

of

ABS

tranches

from

issuersthat

have

issued

ABS

deals

using

both

blockchain

andother

methods.

Third,

the

results

are

similar

when

wecontrol

for

the

time-variant

issuer

characteristics

and?xed

effects

of

the

underlying

asset

types.

Fourth,

it

ispossible

that

the

reduction

in

yield

spread

is

con-foundedbymarketparticipants’irrationalreactionstothe

application

of

blockchain

to

ABS

issuance.

UsingBitcoin

returns

to

capture

investor

hype,

we

?nd

thatthe

effect

of

blockchain

on

yield

spread

does

not

varywith

Bitcoin

returns.

This

result

is

inconsistent

withtheinvestorhypeargument.Issuing

ABS

with

blockchain

technology

can

reducethe

yield

because

blockchain

technology

can

reduce

theinformation

asymmetry

between

issuers

and

investorsregarding

the

quality

of

underlying

assets.3

In

a

typicalABS

issuance

process,

the

issuer

bundles

the

underlyingassets

and

issues

bond

securities

backed

by

these

assetsto

potential

buyers,

who

are

primarily

institutional

inves-tors.

Because

the

ABS

prospectus

usually

only

includessummary

information

about

underlying

assets,

the

qual-ity

of

these

assets

is

not

well-understood

by

marketparticipants,

leading

to

information

asymmetry

betweenissuers

and

investors

about

the

quality

of

underlyingassets,

which

is

the

main

determinant

of

ABS

yield.

Toreduce

this

information

asymmetry,

issuers

pay

creditrating

agencies

to

independently

evaluate

the

quality

ofthe

underlying

assets.

However,

rating

agencies

can

onlypartially

reduce

information

asymmetry

because

theyonly

evaluate

the

quality

ofa

random

sample

of

underly-ing

assets.

In

addition,

issuers

pay

for

these

third-partyservices,

leading

to

agency

problems

such

as

rating

in?a-tion

or

rating

shopping

(e.g.,

He

et

al.

2011,

2012).A

blockchain,

which

constitutes

a

private

chain

man-aged

by

the

issuer,

can

effectively

reduce

the

infor-mation

veri?cation

cost

(Catalini

and

Gans

2020)

andinformation

asymmetry

when

used

for

ABS

issuance.First,

unlike

traditionally

issued

ABS,

blockchain-basedABS

have

all

of

the

underlying

assets

recorded

on

theblockchain.

As

such,

market

participants

with

access

tothe

ledger

stored

on

the

blockchain

have

informationabout

the

underlying

assets

and

can

evaluate

their

qual-ity.

Second,

any

modi?cations

of

information

on

theblockchain

can

be

traced.

Therefore,

the

issuers

cannotchange

information

about

the

assets

without

the

knowl-edge

of

other

parties.

This

mechanism

increases

themarket

participants’

con?dence

in

information

aboutthe

underlying

assets.

Third,

the

ability

of

market

parti-cipants

to

conduct

their

own

due

diligence

regardingthe

quality

of

underlying

assets

mitigates

the

agencyproblem

faced

by

rating

agencies.

Given

that

marketparticipants

have

information

about

the

underlyingassets,

the

rating

agencies

are

more

likely

to

make

objec-tive

evaluations

of

the

quality

of

underlying

assets,

thusimproving

the

accuracy

of

ratings.Issuers

decide

whether

to

use

blockchain

for

ABSissuance.

The

factors

that

affect

this

decision

mightalso

affect

the

yield

spread

of

ABS.

Although

control-lingforissuer?xedeffectsandthevariablesthataffectthe

yield

spread

can

address

the

concern

of

correlatedomittedvariables,weexplicitlyaddresstheendogene-ity

concern

by

using

an

instrumental

variable

(IV)approach.

The

IV

we

use

is

the

number

of

?rms

in

theissuer’s

industry

that

are

included

on

the

“List

ofCompanies

with

Blockchain

Digital

Services,”

whichis

maintained

by

China’s

Cyber

Security

and

Digitiza-tion

Committee.

This

variable

should

only

affect

theyield

spread

of

ABS

tranches

through

its

impacton

the

issuers’

choice

of

using

blockchain

for

ABSIn

summary,

using

blockchain

to

issue

ABS

can

effec-tively

address

the

information

asymmetry

between

is-suers

and

investors.

As

information

asymmetry

is

akeydeterminant

of

the

yield

of

ABS,

we

predict

that,

ceterisparibus,

ABS

issued

using

blockchain

technology

havelower

yields

than

other

ABS.Chen,

Cheng,

and

Luo:

The

Economic

Value

of

Blockchain

ApplicationsManagement

Science,

2024,

vol.

70,

no.

1,

pp.

439–463,

?

2023

INFORMS441issuance.

We

obtain

the

same

inferences

when

we

usethis

IV.

In

addition,

it

is

possible

for

issuers

to

useblockchain

to

issue

ABS

with

higher

underlying

assetquality,

thereby

reducing

the

yield

spread

for

suchABS.

We

address

this

concern

by

examining

the

qual-ity

of

the

underlying

assets.

We

do

not

?nd

any

differ-ences

in

the

underlying

asset

quality

of

ABS

issuedusing

blockchain

and

other

ABS

with

the

underlyingasset

quality

proxied

by

the

default

rate

and

delin-quency

rate

of

the

underlying

assets

and

the

likelihoodof

the

rating

downgrade

of

ABS

tranches

after

issu-ance.

However,

despite

these

tests,

we

acknowledgethat

we

cannot

completely

rule

out

the

possibility

thatour

results

are

affected

by

potential

endogeneity.We

conduct

cross-sectional

analyses

to

provide

addi-tional

insights.

First,

because

credit

rating

agencies

playa

key

role

in

ABS

deals,

we

investigate

how

the

effect

ofblockchain

on

the

yield

spread

varies

with

the

charac-teristics

of

credit

rating

agencies.

The

reputation

ofcredit

rating

agencies

can

increase

the

accuracy

of

theirratings,

whereas

their

reliance

on

issuers

for

rating

busi-ness

can

reduce

it

(e.g.,

He

et

al.

2011,

2012).

Therefore,we

focus

on

these

two

characteristics

and

?nd

that

thenegative

effect

of

blockchain

on

yield

spread

is

less

pro-nounced

for

ABS

rated

by

more

reputable

credit

ratingagencies

but

more

pronounced

for

ABS

rated

by

creditrating

agencies

for

which

the

issuer

is

a

more

importantsource

of

rating

business.

Second,

not

all

ABS

tranchesare

the

same.

Given

that

the

use

of

blockchain

canreduce

the

information

asymmetry

about

the

quality

ofunderlying

assets,

the

effect

of

using

blockchain

shouldbe

greater

when

the

level

of

information

asymmetry

ishigher

as

is

the

case

with

revolving

ABS

and

ABS

witha

larger

number

of

underlying

assets.

The

results

areconsistent

with

these

predictions.Finally,

we

examine

the

effect

of

using

blockchaintechnology

in

ABS

issuance

on

the

level

of

retainedinterest

and

number

of

credit

enhancement

mechan-isms

at

the

ABS

deal

level.

Retained

interest

and

othercredit

enhancement

mechanisms

protect

investors

inthe

event

of

default,

but

they

are

costly

for

issuers.Therefore,

we

expect

issuers

tobe

lesslikely

torely

onthese

mechanisms

when

they

have

other

means

ofreducing

information

asymmetry.

As

expected,

we?nd

that

the

useof

blockchain

inABS

issuancesigni?-cantly

reduces

both

the

level

of

retained

interest

andnumber

of

credit

enhancement

mechanisms

with

rela-tive

reductions

of

20%

and

25%,

respectively,

fromtheirrespectivesamplemeans.Overall,

our

?ndings

indicate

that

blockchain

canreduce

the

cost

of

issuing

ABS

and

increase

the

ef?-ciency

of

capital

allocation.

To

the

best

of

our

knowl-edge,

this

study

is

the

?rst

to

empirically

examine

theeconomic

value

of

blockchain

applications

in

securi-ties

issuance.

Our

?ndings

are

timely

as

manycompaniesandgovernmentsarenowexploringblock-chain

applications.

Our

results

can

help

them

betterunderstand

the

role

of

blockchain

in

securities

issu-anceandassessfutureinvestmentsinblockchaintech-nology(Consensys2019).6Our

?ndings

are

also

signi?cant

and

timely

for

ABSissuance.

Information

asymmetry

about

the

quality

ofunderlying

assets

is

a

critical

issue

in

the

ABS

market.This

information

asymmetry

prevents

investors

fromconducting

their

own

due

diligence

and

results

in

anoverreliance

on

credit

rating

agencies.

Recognizingthis

problem,

the

Securities

and

Exchange

Commis-sion

(SEC)

issued

Regulation

AB

II,

effective

from

Novem-ber

23,

2016,

which

requires

issuers

of

certain

types

of

ABS(e.g.,

those

backed

by

auto

loans

or

mortgages)

to

providedetailed

information

about

speci?c

underlying

assets

in

aprescribed

format.

Neilson

et

al.

(2022)?nd

that

followingthe

implementation

of

this

requirement,

the

ABS

yieldspread

and

ratings

better

re?ect

the

quality

of

underlyingassets.

However,

such

mandatory

disclosure

is

costly

for7issuers

and

can

discourage

them

from

issuing

ABS.8Our

?nding

that

ABS

issued

using

blockchain

tech-nology

have

lower

yields

than

other

ABS

indicates

thatblockchain

represents

an

effective

alternative

approach

toalleviate

information

asymmetry

and

reduce

the

cost

ofcapital.

The

use

of

blockchain

to

disseminate

informationhas

three

advantages

over

the

one-size-?ts-all

disclosurerequirement.

First,

issuers

can

choose

to

use

blockchainwhen

the

dissemination

of

asset-level

information

has

agreater

bene?t.

Second,

unlike

Regulation

AB

II,

whichrequires

asset-level

information

to

be

publicly

disclosed,only

ABS

issuance

participants

have

access

to

the

informa-tion

on

blockchain,

reducing

the

proprietary

cost

of

disclo-sures.

Third,

blockchain

is

valuable

for

ABS

issuance

asmost

countries

do

not

have

asset-level

disclosure

require-ments

and

Regulation

AB

II

does

not

apply

to

all

types

ofABS

issuance,

such

as

those

backed

by

credit

card

receiv-ables

or

student

loans.9We

note

the

following

caveats

to

our

research.

First,the

application

of

blockchain

technology

is

still

in

itsinfancy;

the

number

of

ABS

deals

issued

using

block-chain

technology

is

small,

potentially

reducing

thepower

of

the

tests

and

the

reliability

of

the

results.Second,

we

provide

evidence

on

the

economic

valueof

blockchain

applications

in

ABS

issuance,

but

we

donot

consider

the

cost

of

such

applications,

which

mayvary

signi?cantly

across

settings.

Each

ABS

deal

has

asmall

number

of

market

participants,

typically

fewerthan

200

institutional

investors.

The

cost

of

blockchainapplications

with

a

signi?cantly

larger

number

ofmarket

participants

can

be

substantially

higher

andmay

outweigh

the

bene?ts.

We

leave

the

investigationof

the

economic

value

of

blockchain

applications

inothersettingstofutureresearch.10Chen,

Cheng,

and

Luo:

The

Economic

Value

of

Blockchain

Applications442Theremainder

ofthispaper

isorganized

asfollows.Section2providesadiscussionofthebackgroundanddevelops

the

hypotheses.

Section

3

presents

the

sam-ple,

data,

and

research

design.

Section

4

presents

themain

analyses,

and

Section

5

presents

the

additionalanalyses.Section6concludesthepaper.Management

Science,

2024,

vol.

70,

no.

1,

pp.

439–463,

?

2023

INFORMSsuch

that

ABS

investors

have

a

poor

understandingof

the

quality

of

these

assets.

Other

research

showsthat

issuers

and

underwriters

misrepresent

informa-tion

about

underlying

assets

and

it

is

dif?cult

for

in-vestors

to

detect

such

misrepresentation.

For

example,Piskorski

et

al.

(2015)

?nd

that

investors

receive

falseinformation

about

the

quality

of

loans

underlyingmortgage-backed

securities

as

some

originators

andunderwriters

do

not

disclose

secondliens

for

mortgageloans.

Similarly,

Ertan

et

al.

(2017)

discuss

the

agencyproblems

associated

with

loan

securitization.

Suchinformation

asymmetry

can

lead

to

adverse

selectionwhen

originators

decide

which

assets

to

securitize.Downing

et

al.

(2009),

Agarwal

et

al.

(2012),

and

Jianget

al.

(2014)

?nd

that

originators

are

more

likely

tosecuritize

lower

quality

assets

and

retain

higher

qual-ity

assets.To

address

the

problem

of

information

asymmetry,issuers

hire

credit

rating

agencies

to

assess

the

qualityof

underlying

assets.

Issuers

then

price

ABS

tranchesbased

on

these

ratings.

However,

reliance

on

creditrating

agencies

can

lead

to

issues

such

as

rating

in?a-tion

and

rating

shopping

(e.g.,

Bolton

et

al.

2012,

Bon-sall

et

al.

2017,

Baghai

and

Becker

2020)11

as

it

is

theissuers

and

not

investors

who

pay

the

credit

ratingagencies.

Rating

agencies

may

issue

more

favorableratings

to

obtain

future

business

from

issuers.

Forexample,

He

et

al.

(2011,

2012)

?nd

that

rating

agen-cies

provide

more

favorable

ratings

for

larger

issuersthan

smaller

ones

as

the

agencies

are

more

likely

toobtainfuturebusinessfromlargerissuers.Inaddition,issuers

can

“shop”

for

more

favorable

ratings

on

themarket

when

deciding

which

rating

agencies

to

hire.He

et

al.

(2016)

provide

evidence

consistent

with

rat-ing

shopping

by

issuers.

This

is

also

recognized

byregulators.

For

example,

SEC

ruling

33-9638

statesthat

“complex

assets

that

are

dif?cult

to

rate

and

thatare

likely

to

generate

differences

in

ratings

can

createincentives

for

issuers

to

shop

for

ratings

and

discloseonlythoseratingsthatarehigh”(p.38).2.

Related

Research

and

HypothesisDevelopment2.1.

Review

of

the

Related

Literature2.1.1.

Prior

Research

on

Blockchain

Technology

inFinancial

Markets.

Prior

research

on

blockchain

tech-nology

in

?nancial

markets

primarily

focuses

on

cryp-tocurrencies,

including

the

underlying

mechanisms

ofblockchain

technology

and

related

consensus

genera-tion.

For

example,

Biais

et

al.

(2019)

study

mininggames

involving

proof

of

work,

whereas

Saleh

(2021)explores

proof

of

stake

as

an

alternative

consensuspr