Digital markets and platform competition

Detail

1. Why digital markets are different from perfect competition

• Perfect competition (NCERT Class 12) has many small firms, identical products, free entry and exit, and firms that are price takers (they must accept the market price). Digital markets break almost every one of these conditions.
• The OECD lists the main features of digital markets [5][6]:
• they are multi-sided (they serve more than one group of users);
• many services have a zero price;
• they have network effects;
• they have economies of scale and scope (costs per unit fall as a firm grows, and as it offers more products);

• access to user data, and earning money from it, matters a great deal.

• The result: a few firms become very large. The market moves towards monopoly or oligopoly (a market with a few large sellers), even when the firms started out competing fairly.

• This links to market failure (Class 9, Class 7). When one firm controls a market, society does not get the best result. So the state steps in to regulate, as it already does with monopolies.

2. How platforms work

Network effects

• Network effect: a product becomes more valuable to each user as more people use it.
• Direct network effects work within one group.

• Example: more WhatsApp users make WhatsApp more useful to each user.

• Indirect network effects work across two groups.

• Example: on a ride-hailing app, more riders attract more drivers.

• More drivers mean shorter waiting times, so even more riders join.

• Worked example (why value grows faster than users):

• With n users, the number of possible user-to-user links is n(n − 1) ÷ 2.
• 10 users → 10 × 9 ÷ 2 = 45 links.
• 100 users → 100 × 99 ÷ 2 = 4,950 links.
• Users grew 10 times, but possible links grew about 110 times. So a big network pulls ahead fast, and a small rival cannot catch up.

Two-sided markets

• Two-sided market: a platform that serves two separate user groups. Each group gains only if the other group also joins.
• Often one side pays nothing. The platform pays for this "free" side by charging the other side.

Platform	Free (or subsidised) side	Paying side
Search engine	Users	Advertisers
Social media	Users	Advertisers
UPI apps	Merchants (zero merchant discount rate, i.e. no fee on each payment)	Mostly paid for by government support and the wider business
Food delivery	Customers (partly)	Restaurants pay a commission

• Exam point: a zero price is not proof that there is no market power. The platform may earn its money on the other side.

Switching costs and data lock-in

• Switching costs: the cost or hassle of changing to another supplier. Examples:
• moving your data;
• losing your contacts or reviews;

• learning a new app.

• High switching costs lock users in. Data lock-in makes the leading firm (the incumbent) even stronger. A new rival cannot take users away, even with a better product.

Winner-takes-all markets and "tipping"

• Winner-takes-all market: a market where strong network effects or scale economies let one firm take most of the market share and profits. Examples: search, app stores.
• Tipping: the point at which a market turns decisively towards one firm. After that point, rivals cannot recover.

Why ex-post enforcement is too slow

• Ex-post enforcement: the regulator acts after the harm has happened. This is the model of the Competition Act, 2002. It covers anti-competitive agreements (Section 3) and abuse of dominant position (Section 4).
• Ex-ante regulation: the rules are set in advance. Large firms must follow them before any harm happens.
• The problem with ex-post enforcement:
• a case takes years;
• by the time an order comes, the market has already tipped;

• competition cannot be brought back.

• Example: the Google Android case took roughly 3 years at the Competition Commission of India (CCI), followed by appeals.

• The OECD's 2024 study of G7 countries looks at how ex-ante and ex-post tools can be used together to control the market power of large platforms [6].

3. Platform conduct (the practices that worry regulators)

Self-preferencing

• A dominant platform favours its own products over those of rival businesses that depend on the platform.
• Examples: search rankings; marketplace listings.
• The platform is both the referee and a player in its own market.

Anti-steering

• These are platform rules that stop business users from pointing customers to cheaper offers, or to other payment channels outside the platform.
• Example: app-store billing rules. An app developer may not tell users, "Pay on our website and save 20%."

Deep discounting

• Deep discounting: very large discounts, often below cost, paid for with investor money, used to capture market share.
• Kiranas (small neighbourhood shops) and distributors make this charge against e-commerce and quick-commerce platforms. For example, the AICPDF (All India Consumer Products Distributors Federation) did so in 2024-25.
• The debate:
• It is predatory pricing if the aim is to drive rivals out and later recover the losses by raising prices.

• But it benefits consumers today.

• Worked example:

• A product costs a platform ₹100 to deliver. It sells the product at ₹70, a loss of ₹30 per unit.
• It sells 1 crore units a year. The yearly loss is ₹30 × 1 crore = ₹300 crore, paid for by investors.
• A kirana cannot survive a ₹30 loss on every unit, so it leaves the market.
• Later the platform raises the price to ₹130 and earns ₹30 per unit. This earns back the ₹300 crore loss in about one year. That recovery is what makes the pricing "predatory".

Killer acquisition

• Killer acquisition: a large incumbent buys an innovative start-up mainly to shut down the start-up's products and head off future competition.
• The old merger-review tests looked only at a firm's assets and turnover. Start-ups have small assets and turnover, so these deals were not reviewed.
• This is why a deal value threshold was added. Merger review is now triggered by the price paid for a firm, not only by the size of the target.

4. India's response

(a) Standing Committee on Finance (53rd report, December 2022)

• Report title: "Anti-Competitive Practices by Big Tech Companies" [2][3].

• It listed ten anti-competitive practices: 1. anti-steering; 2. self-preferencing; 3. bundling and tying (forcing a buyer to take one product in order to get another); 4. data usage; 5. pricing and deep discounting; 6. exclusive tie-ups; 7. search and ranking preferencing; 8. restricting third-party apps; 9. advertising policies; 10. mergers and acquisitions.

• It recommended an ex-ante law.

(b) Committee on Digital Competition Law (CDCL) and the draft Digital Competition Bill

• The Ministry of Corporate Affairs (MCA) set up the CDCL in February 2023. It followed the Standing Committee's 53rd report. The Secretary, MCA chaired it [2].
• The report and a draft Digital Competition Bill were released on 12 March 2024 [2][3].
• Approach: an ex-ante framework that focuses on preventing anti-competitive conduct [3].
• SSDE (Systemically Significant Digital Enterprise): a large firm that provides a core digital service and so must follow advance obligations.

• Examples of core digital services: search engines, social networking services, operating systems, web browsers [3][4].

• How an SSDE is designated:

• Financial test (significant financial strength): India turnover ≥ ₹4,000 crore, or global market cap ≥ US$75 billion. The measures used include turnover, gross merchandise value and market capitalisation [4].
• User test: ≥ 1 crore end users, or ≥ 10,000 business users in India [3].

• Qualitative test: if the numbers are not enough, the regulator can also look at the firm's resources and the volume of data it holds [4].

• Associate Digital Enterprises (ADEs): group companies that provide core digital services are designated along with the SSDE [4].

• Obligations on an SSDE [3][4]:
• no self-preferencing (favouring its own products or those of related parties);
• no use of business users' non-public data to compete against those same users;
• no restrictions on users' use of third-party apps;
• no tying (forcing users of one service to use its other services);

• no anti-steering.

• Penalty: up to 10% of global turnover. It is a civil penalty, not a criminal one [4].

• Status: reported in 2025 to be on hold pending a market study (verify current).

(c) The model: the EU Digital Markets Act (DMA)

• The DMA's obligations have applied since March 2024.
• Gatekeeper platform: a large platform that controls access between business users and consumers.
• Designated gatekeepers: Alphabet, Amazon, Apple, ByteDance, Meta, Microsoft and Booking.
• First DMA fines (April 2025): Apple €500 million; Meta €200 million.
• The OECD keeps a G7 inventory of new digital-market rules. It compares ex-ante rules by their status, scope, institutions and content [6].

(d) Related Indian tools

Tool	What it does	Competition logic
FDI policy, Press Note 2 (2018)	Foreign-owned e-commerce firms may run only a marketplace. They may not hold inventory (their own stock) and cannot control sellers' prices.	Stops a platform from being both the marketplace owner and a seller on it
ONDC (Open Network for Digital Commerce)	An open network that separates buyer apps from seller apps	Reduces platform lock-in; sellers can reach buyers on any app
NPCI 30% UPI market-share cap	No single UPI app may handle more than 30% of UPI transaction volume. The compliance deadline is extended to 31 December 2026 (verify current).	Stops the UPI market from tipping to two apps
CCI market studies	E-commerce (2020); AI and competition (see below)	Help the regulator understand a market before it acts

• Worked example (30% cap): suppose there are 2,000 crore UPI transactions in a month. The most any one app may handle is 0.30 × 2,000 = 600 crore transactions.
• CCI Market Study on AI and Competition:
• It was carried out through the Management Development Institute Society (MDIS) and completed in October 2025 [7][8].
• Global AI market size: US$103.6 billion (2020) → US$288.8 billion (2024) [7].
• India's AI market: US$2.97 billion (2020) → US$7.63 billion (2024) [7].
• Findings: AI brings large gains, such as higher efficiency and better market access for MSMEs. But possible anti-competitive conduct must be watched [9].
• The CCI issued a guidance note with a self-audit method. Firms' boards and managers can use it to check that their AI systems do not produce hidden anti-competitive results [9].

Prelims Hooks

• Direct network effect works within one user group (WhatsApp). Indirect network effect works across groups (riders ↔ drivers).
• In a two-sided market, one side is often charged zero. So a zero price does not mean there is no market power.
• The Standing Committee on Finance's 53rd report (December 2022), "Anti-Competitive Practices by Big Tech Companies", listed ten practices and recommended an ex-ante law [2].
• The CDCL was set up by the Ministry of Corporate Affairs, not MeitY. It was chaired by the Secretary, MCA. Its report and the draft bill came out on 12 March 2024 [2].
• SSDE user test: ≥ 1 crore end users or ≥ 10,000 business users in India [3]. Maximum penalty: 10% of global turnover [4].
• Trap: the "gatekeeper" label comes from the EU DMA. India's draft bill uses the term SSDE.
• Killer acquisitions slipped past asset and turnover thresholds. This is why the deal value threshold was added.
• Press Note 2 (2018): foreign-owned e-commerce firms may use only the marketplace model. The inventory model is not allowed.
• ONDC separates buyer apps from seller apps. The UPI 30% market-share cap is set by NPCI, not the RBI.
• CCI's AI market study (October 2025): India's AI market was US$7.63 billion (2024) [7].

Mains Points

• Ex-ante vs ex-post regulation:
• Cases are slow, as the roughly 3-year Google Android case at the CCI shows. Markets tip before any order comes. This supports ex-ante duties for SSDEs.

• Against: strict early rules may slow innovation, raise compliance costs for Indian start-ups, and copy the EU model without Indian evidence. This is why the bill was put on hold pending a market study [4][6].

• Deep discounting debate:

• For consumers now: lower prices and more convenience.
• Against the long run: it may be predatory pricing that damages kiranas and distributors (AICPDF, 2024-25). Once rivals leave, consumers may face higher prices later.

• Policy therefore has to weigh static efficiency (low prices today) against dynamic efficiency (keeping competition alive tomorrow).

• Using market design, not only law:

• India also shapes digital markets through public digital infrastructure: ONDC, the UPI 30% cap and FDI rules (Press Note 2).
• These tools make markets open to all (interoperable) and so reduce lock-in.

• This complements the CCI's enforcement work. It is a GS-III theme: digital public infrastructure used as competition policy.

• Data and AI as new entry barriers:

• Control of data, computing power and AI models may concentrate markets further.
• The CCI's AI study (October 2025) and its self-audit guidance show a "soft law first" approach. The regulator watches the market before it makes hard rules [7][9].