Given that enforcement efficacy is determined by credibility and agility of regulatory regimes, we compared three antitrust regimes actively reigning in big tech in 2020-2021—the United States (US), the European Union (EU), and China—and assessed their regime impacts with the random effects model.

1. Credibility and agility of
antitrust crackdown:
A cross-country analysis of
big tech
數位治理：韌性、AI、規管】研討會
Dec. 15, 2023
周韻采
王立達

2. Introduction

3. Big tech dominance
1. 3 economic features: 1. network effects; 2. multi-sidedness; 3.
multi-homing.
2. big techs have incentive to engage in strategic practices in safeguarding
their dominance amid the volatility and blurred boundaries of digital
business and multi-homing.
3
12/30/2023
Digital platform
(big data + algorithm)
Consumer
base
Advertisement
market
Network effects

4. Annual revenue of big tech firms from 2017 to 2022
2017 2018 2019 2020 2021 2022
Amazon 177.87 232.89 280.52 386.06 469.82 513.98
Apple 229.23 265.6 260.17 274.52 365.82 394.33
Google 109.65 136.22 160.74 181.69 256.74 279.80
Alibaba 22.99 56.15 71.99 109.48 134.57
Meta 40.65 55.84 70.7 85.97 117.93 116.61
Tencent 21.90 45.56 54.08 73.88 85.84
Netflix 11.69 55.84 20.16 25.00 29.70 31.62
PayPal 13.09 15.45 17.77 21.45 25.37 27.52
Baidu 13.03 14.88 15.43 16.41
19.54
4
(unit: billion USD)

5. Big tech dominance
簡報標題 5
• big tech firms have already obtained considerable market powers in a couple of major platform services.
PLATFORM COMPANY MARKET SHARE
APP MARKET Apple Store 62.4%
Google Play 33.3%
SEARCH ENGINE Google 86.2%
ONLINE AD Google (search) 86%
Meta (social media) 90%
Youtube (video) 59%
E-COMMERCE Taobao (Alibaba) 15%
Tmall (Alibaba) 14%
Amazon 13%
STREAM VIDEO Netflix 20%
Prime Video (Amazon) 14%
Tencent Video 12%
STREAM MUSIC Spotify 32%
Apple Music 16%
Amazon Music 13%

6. Waves of antitrust crackdown
• Their maneuvers of dominance demonstrate a high level of
resemblance amid market dynamics and political-institutional
variance.
• Tech firms’ anticompetitive behaviors are hardly deterred by
traditional antitrust rules.
• We compared three political entities actively reigning in big
tech—the United States (US), the European Union, and
China—and assessed their enforcement efficacy.
簡報標題 6

7. Antitrust enforcement

8. Eight anticompetitive conducts
• 1. Self-preferencing (demoting): google search, FB marketplace, Amazon e-
commerce
• 2. Tying of services (Bundling): google Andorid + google play; google search +
apple iOS, Amazon e-commerce + buy box (prime), app store + payment service
• 3. Exclusivity contract: Google employed restrictive agreements with browser and
phone partners such as Apple, Mozilla, Samsung and Verizon, making Google the
default search engine on phones. Google’s separate agreements with Android-
based mobile-device manufacturers forbid pre-installing or promoting rival search
engines if they opt to take a cut of Google’s search revenue.
• 4. Non-interoperability: Taobao’s block of access to WeChat in 2013. In retaliation,
WeChat denied access to Taobao and Alipay. WeChat also blocked access to
Uber’s red-envelope marketing event while providing direct access to DiDi (in
which Tencent invested) in 2015. WeChat and QQ (another social media platform
owned by Tencent) blocking access to TikTok in 2018.

9. Eight anticompetitive conducts
• 5. Unfair collection and use of data: Data scale improves the ability of
algorithms to learn by trial-by-error (Stucke et al., 2016:184). Data
concentration would constitute significant entry barriers for competitive rivals
(Stucke & Grunes, 2016:36-40).
a). small search engines often struggle with less frequent tail inquires; b).
Netflix tracks its subscribers’ viewing habits that can predict consumer
preferences; c). Amazon reported that “30 percent of sales were due to its
recommendation engine that uses personal history of purchases.
• 6. Algorithmic discrimination.
• 7. Anticompetitive price-related conducts.
• 8. merger & acquisition (M&A): FB+Whatsapp+IG, FB+Within Unlimite,
MS+ATVI

10. US antitrust cases, 2020~2021
簡報標題 10
Firm Plaintiff/date Allegation Status
Google DoJ+12 states
(Oct. 20, 2020)
engaging in anticompetitive behavior by
paying Apple between US$8 and 12 billion to have
Chrome set as the default search engine on
iPhones.
Non-jury trial began
on Sep 12, 2023.
Meta FTC
(Dec. 9, 2020)
illegal monopolization of the social
networking market by acquiring Instagram and
WhatsApp. FTC requested the divesture of
Instagram and WhatsApp from Meta.
Dismissed on Jun. 28,
2021, revived on Jan. 11,
2022. Still pending.
Google Texas-led 10 states
(Dec. 16, 2020)
illegal digital advertising monopoly and
negotiated with Meta for preferential treatment.
Pending.
Google 40 states
(Dec. 17, 2020)
manipulating its search results to ensure its
own products and services were ranked higher
than those of their rivals.
Google denied
destruction of evidence.
Status conference to be
held on Aug. 24, 2022.
Google DC, Texas,
Washington & Indiana
(Jan. 24, 2022)
making misleading promises about its users’
ability to turn off location tracking during
movement from 2014-2020.
Google agreed to
US$391.5 million
settlement with 40 states.
(Nov. 16, 2022)

11. EU antitrust cases, 2020~2021
FIRM DATE ALLEGATION
APPLE 2020.6.16 Forcing app developers to use Apple Pay, which
constitutes unfair competition.
AMAZON 2020.11.10 Using nonpublic data gathered from eight million
third-party sellers to unfairly compete against them.
APPLE 2021.4.30 Abusing control over the distribution of music-
steaming apps, including Spotify.
META 2021.6.5 Unfair competition against digital advertisers.
GOOGLE 2021.6.22 Anticompetitive business practices, including ad
brokerage and sharing of user data with advertisers.

12. China’s antitrust cases, 2020~2021
Firm Date Allegation
Alibaba 2021.4.12 ¥18.28 billion fine for “choose one from
two” contracting with online sellers
Alibaba, Tencent,
Meituan, DiDi
2021.7.7 Illegal M&As from big platforms
Tencent 2021.7.10 M&As with e-sports platforms Huya and
Doyu rejected
Tencent 2021.7.13 M&A with search engine Sohu approved
Tencent 2021.7.24 Exclusive copyright licensed to Tencent
Music (music platform) rescinded
Meituan 2021.10.8 ¥3.42 billion fine for “choose one from
two” contracting with online sellers

13. CASE SELREF TYING EX INTER DATA ALGO PRICE M&A YFINE
GOOGLE 15 2 9 7 2 4 0 1 0 5
AMAZON 5 0 1 2 0 2 1 1 0 3
FACEBOOK 5 0 1 1 0 3 0 0 1 1
APPLE 13 0 10 1 0 0 1 2 0 3
SUBTOTAL 38 2 21 11 2 9 2 4 1 12
AVERAGE 9.5 0.5 5.25 2.75 0.5 2.25 0.5 1 0.25 3
ALIBABA 12 1 3 5 4 3 4 2 2 4
TENCENT 21 1 1 4 6 7 3 2 5 4
MEITUAN 5 1 3 3 1 2 2 1 1 2
DIDI 4 1 1 1 1 2 2 1 5 2
BAIDU 3 0 0 1 1 0 1 0 1 1
SUBTOTAL 45 4 8 14 13 14 12 6 11 13
AVERAGE 9 0.8 1.6 2.8 2.6 2.8 2.4 1.2 2.2 2.6
TOTAL 83 6 29 25 15 23 14 10 12 25
Statistics of enforcement, 2020~2021

14. Policy effectiveness results from a
“goodness of the fit” of the regulatory
system with a country’s institutions
“
”
~ Levy & Spiller
legal certainty, speed of intervention, and flexibility
increase policy effectiveness
”
“
The EC’s impact assessment report for
the DMA ~

15. olitical
institutions
• Enforcement credibility
enhances efficacy because it
decreases the costs of
implementation and
enforcement. Credibility could
be created when policymakers
abide by law to exercise
discretionary administration.
• Enforcement agility refers to
agencies’ capability of making a
policy shift in accommodating
market and political-societal
contingencies.
• There may be a trade-off
between credibility and agility.
簡報標題 15

16. Political institutions
簡報標題 16
The United States EU China
Digital Acts Numerous drafts DMA & DSA Administrative
guidelines
Antitrust
approach
Ex post enforcement Ex ante compliance Ex ante compliance
Enforcement
credibility
High credibility due to
separation of power
limiting administrative
discretion
Moderate credibility
due to imperfect
separation of power
Low credibility due to
lack of separation of
power, adequate
governance and high
political risks
enforcement
agility
Low agility
compromised by
inflexible enforcement
Moderate agility due
to strong administration
responding to
contingency
High agility due to
strong administration
swiftly adapting to
contingency

17. ata
• Because all 9 big tech firms are listed
in the Nasdaq market, we collected
their daily trading data from 2020 to
2021 on Nasdaq, containing daily high
price, low price, open price, close
price, turnover, volume, turnover ratio,
and the release date of financial
reports.
• Our database contains 4,679
observations on 496 trading days. We
identified 23 competition rulemakings
and 83 firm-related enforcement cases,
totaling 106 unique antitrust events.
• We inputted values for discrete
enforcement variables based on our
reading of news articles and official
notices or documents.
簡報標題 17

18. Variable Value/unit Definition
ANTITRUST EVENT
event_flag {1, 0} 1 = antitrust event,
0 = none
Eight antitrust remedies
(reference, tying, exclusivity, interoperability,
data, discrimination, abusive price, M&A)
{1, 0} 1 = antitrust event,
0 = none
18
fine ≥ 0 The recorded fine/settlement charge for a
given antitrust event
POLITICAL INSTUTITION
WGI {100, 0} WGI percentile for each political institution
imposing antitrust enforcement or rulemakings
event_WGI {100, 0} event_flag x WGI
USW {100, 0} US * WGI
EUW {100, 0} EU * WGI
CNW {100, 0} CN * WGI

19. odel
• The event study considers a given
law enactment, enforcement, or a
court decision as the event that
impacts a firm’s performance.
• The performance can be measured by
both financial and non-financial
indicators.
• stock price is often selected for firms’
financial performance indicator in the
event study.
• The random walk process assumes
the difference in a firm’s share prices
as a form of stationary time series.
• To correct the heteroscedasticity
inherent in cross-sectional
observations, we used random effects
model. 19
𝑷𝑪𝒊𝒕 = 𝜶 + 𝜷𝑿𝒊𝒕 + 𝜸𝑹𝑬𝑮𝒊𝒕 + 𝜹𝑺𝒊𝒕 + 𝝆𝑾𝑮𝑰𝒊 + 𝝈𝑻𝒕 +
𝝂𝒊 + 𝜺𝒊𝒕
𝑃𝐶𝑖𝑡: The percentage change in stock price; :𝑋𝑖𝑡: a firm’s financial
information; 𝑅𝐸𝐺𝑖𝑡: a set of enforcement remedies and the fine
amount; 𝑆𝑖𝑡: the macroeconomic factors affecting the share prices,
such as GDP, interest rates, unemployment rate, and Nasdaq
Composite index. 𝑊𝐺𝐼𝑖: political-institutional variable that denotes
each country’s level of enforcement credibility and agility; 𝑇𝑡 is a
time-series variable, controlling for the autocorrelation problem
inherent in longitudinal studies.

20. Fine is the most effective remedy
when reining in big tech
“ Results ”

21. Variable
% changes in
share prices
(A)
event_flag
(discrete)
(B) (C)
event_flag
(discrete)
ANTITRUST EVENT -1.637***
[0.393]
Omitted
SELF-REFERENCING 0.757
[0.464]
TYING 0.820
[0.467]
EXCLUSIVITY CONTRACT -0.722
[0.569]
NON-INTEROPERABILITY 0.174
[0.392]
USE OF NONPUBLIC DATA -1.156***
[0.252]
DISCRIMINATION -0.375
[0.497]
ANTICOMPETITIVE PRICE 0.238
[0.598]
MERGER -0.302
[0.413]
(% changes in) FINES -0.506*
[0.193]
WGI 0.023***
[0.006]
0.002
[0.003]
-0.034
[0.030]
The effect of
antitrust
enforcement on
big tech firms’
stock returns

22. Antitrust enforcement does matter!
• 1) Each enforcement or rulemaking occurrence causes a firm’s stock return to
decrease by 0.8%.
• 2) Among the eight anticompetitive remedies, fixing big tech’s unfair data use
significantly decreases the firm’s stock return by 0.7%.
• 3) a one percent increase in the penalty amount will cause a selloff in the stock
return by 0.5%.
• Q1: Why is refraining big tech from unfair data use the effective remedy?
•  Because data collection and use involves both unfair competition against rivals
and the breach of personal privacy, it is more noticeable to the public and likely to
be disciplined than other misconducts.
22

23. Q5: Why do fines become an effective tool of
enforcement compared to other remedies?
Variable
% changes in share prices
(D) (E)
US 0.006
[0.004]
Omitted
EU 0.001
[0.004]
-0.009
[0.015]
CN -0.015***
[0.004]
0.086
[0.043]
(% changes in) FINES -0.648**
[0.179]
financial reports 0.445
[0.262]
Omitted
(% changes in) trading volume -0.251***
[0.022]
-0.512*
[0.231]
trading date -0.001***
[0.000]
-0.023**
[0.007]
(% changes in) Nasdaq index 0.822***
[0.021]
1.673***
[0.285]
GDP growth rate 0.014
[0.008]
0.766
[0.489]
(% changes in) interest rates 0.022*
[0.010]
-0.541
[0.506]
(% changes in) the unemployment
rates
-0.041**
[0.016]
-1.125*
[0.381]
sample size 4,276 20
The institutional
effect on big
tech firms’ stock
returns

24. Chinese regime reduces stock returns
• 4) Only Chinese regime generates effective antitrust enforcement on big tech.
• 5) a one percent increase in the penalty amount will cause a selloff in the stock by
0.65%. The enforcement/institutional impacts are absorbed monetary penalty when
estimated jointly.
• Q2: Why is the Chinese enforcement effective while the U.S. and the EU regimes
fail to generate effective remedies?
•  Because of the low credibility and high agility of Chinese regime, investors
anticipate non-credible but enforceable antitrust crackdown. The crackdown is then
detrimental for the big platformers, resulting in a selloff for their stock returns 24

25. Fine does matter!
• Q2: Why is the Chinese enforcement effective while the U.S. and the EU regimes
fail to generate effective remedies?
•  investors have difficulties in anticipating the outcome of the US and EU remedies
due to the uncertainty associated with court decisions on either continent. Lagged
effects may emerge after court verdicts are delivered
• Q3: Why does monetary penalty cause a selloff in the stock return?
•  Because fine is easily carried out with few monitoring costs, it sends investors a
strong signal regarding enforcement effectiveness. Investors then anticipate a selloff
for the penalized firms’ stock returns.

26. Conclusion
onclusion

27. 如何達到目標
27
簡報標題
effective With fine imposed
antitrust remedy
Ban on unfair data use
Insignificant
Political institution China Insignificant
fine √

28. Interpreting the results
簡報標題 28
Chinese regime
When antitrust regulation
lacks credibility but is
enforcement, its effect on
firms is detrimental.
Fine
Fine becomes the most
effective instrument amid
political-institutional
differences.
US
Trustbusters there are advised
to prioritize monetary
punishment during
crackdown

29. Q&A
Yuntsai Chou
ychoutotochu@gmail.com

30. 每季績效
簡報標題 30
4.5
3.5
2.5
4.3
2.8
1.8
4.4
2.4
5.0
3.0
2.0 2.0
-
1.0
2.0
3.0
4.0
5.0
6.0
第 4 季 第 3 季 第 2 季 第 1 季
數列 3 數列 2 數列 1

31. 時間表
簡報標題 31
協同可調整規模的電子商務
20XX 年 9 月
發佈標準化指標
20XX 年 11 月
協調電子商務應用程式
20XX 年 1 月
培養整體卓越的方法
20XX 年 3 月
部署具有令人信服之電子商務
需求的策略網路
20XX 年 5 月