Algorithmic Liquidity Formation and Exchange Dependency Structures

1. Introduction

Algorithmic liquidity formation refers to the way automated trading systems, matching algorithms, market-making software, smart-order-routing systems, and data-driven pricing tools collectively create, allocate, and sometimes concentrate liquidity in financial or digital markets.

Exchange dependency structures arise where traders, brokers, market makers, investment firms, or competing platforms become dependent upon a particular exchange or exchange-like infrastructure for:

• access to order books;
• real-time market data;
• matching engines;
• clearing and settlement;
• liquidity pools;
• price discovery;
• co-location facilities;
• APIs and connectivity;
• smart-order-routing infrastructure;
• benchmark or reference prices; and
• execution algorithms.

Competition law becomes important when an exchange or platform is not merely facilitating transactions but controls an infrastructure on which effective participation depends.

A useful analytical distinction is:

Liquidity creation itself is generally pro-competitive; control over the mechanisms through which liquidity is created, accessed, or redirected may raise competition concerns.

The law therefore examines whether algorithmic systems improve market efficiency or instead create exclusion, discriminatory access, information asymmetry, self-preferencing, foreclosure, or coordinated conduct.

Importantly, there is not yet a single universally recognised legal doctrine called the "algorithmic liquidity formation doctrine." The concept is assembled from established doctrines concerning dominance, essential facilities, refusal of access, discriminatory conditions, information exchange, market foreclosure, exchange mergers and algorithmic coordination.

2. Meaning of Algorithmic Liquidity Formation

Liquidity traditionally depends upon the presence of buyers and sellers willing to transact at competitive prices.

Algorithmic systems can radically change this process.

Traditional liquidity formation

A simplified structure is:

Buyer → Broker → Exchange → Seller

Algorithmic liquidity formation

The structure may become:

Market Data → Algorithm → Order Generation → Matching Engine → Execution → Feedback → New Algorithmic Orders

The system therefore creates a continuous feedback loop.

For example:

1. An algorithm receives order-book information.
2. It detects a shortage of sell-side liquidity.
3. It increases the bid.
4. Other algorithms observe the new price.
5. Market makers modify their quotes.
6. Liquidity migrates toward the exchange offering the best execution probability.
7. Increased liquidity attracts additional traders.
8. Additional traders generate still more liquidity.

This creates liquidity network effects.

The exchange with the deepest liquidity may therefore become increasingly attractive, even if competing exchanges offer technically similar infrastructure.

3. Exchange Dependency

Exchange dependency occurs when effective participation in a market increasingly requires access to one particular infrastructure.

There are several forms.

A. Data dependency

A trading algorithm may depend upon:

• real-time order-book data;
• historical tick data;
• benchmark data;
• proprietary feeds;
• latency-sensitive information.

B. Execution dependency

Traders may need access to a particular matching engine because most counterparties are located there.

C. Liquidity dependency

A trader may technically be able to move to another exchange but receive materially worse execution because the alternative exchange lacks comparable liquidity.

D. Clearing dependency

A trading venue may be linked to a particular clearing infrastructure, creating additional switching barriers.

E. Technological dependency

Algorithms may be specifically designed around:

• an exchange's API;
• order types;
• latency architecture;
• market-data format;
• co-location arrangements.

F. Network dependency

The more market participants use an exchange, the more attractive the exchange becomes to additional participants.

This creates a self-reinforcing liquidity loop:

More traders → More orders → Greater liquidity → Better execution → More traders.

4. Competition-Law Significance

The principal competition-law concern is not simply that an exchange is large.

The critical question is:

Has control over algorithmically generated liquidity become a mechanism for restricting effective competition?

Potential theories include:

1. Abuse of dominance
2. Discriminatory access
3. Refusal to deal
4. Essential-facilities-type conduct
5. Margin squeeze
6. Self-preferencing
7. Exclusionary technological design
8. Information asymmetry
9. Algorithmic coordination
10. Anti-competitive mergers between liquidity centres

5. Algorithmic Liquidity as a Network Effect

Liquidity markets are particularly susceptible to network effects.

Suppose Exchange A has 80% of relevant trading activity and Exchange B has 20%.

A market maker may prefer A because:

• spreads are narrower;
• execution probability is higher;
• more counterparties are present;
• price discovery is stronger;
• algorithms can execute larger orders;
• adverse-selection risk may be lower.

But the market maker's decision itself reinforces A's liquidity advantage.

Thus:

Existing liquidity → attracts algorithms → algorithms generate additional liquidity → liquidity attracts more participants.

This can create path dependency.

Competition law must therefore distinguish between:

• efficiency-based liquidity concentration, and
• artificially maintained liquidity concentration.

The latter may become problematic if created through discriminatory access, exclusionary rules, manipulation of data feeds, or restrictions on interoperability.

6. Co-location and Latency Dependency

High-frequency trading makes physical and technological proximity extremely valuable.

Co-location permits trading systems to operate close to an exchange's matching infrastructure.

The economic significance is straightforward:

Lower latency → faster information reception → faster order submission → higher probability of execution.

The competition issue becomes more complicated where:

• access is selectively granted;
• data is disseminated unequally;
• some users obtain systematically earlier information;
• technological architecture favours particular participants;
• alternative participants cannot reproduce the advantage.

The Indian NSE co-location litigation provides an especially relevant example.

7. Case Law

Case 1 — Manoj K. Sheth v. Secretary, Competition Commission of India & National Stock Exchange (NCLAT, 2026)

This is one of the most directly relevant recent Indian authorities.

The case concerned allegations surrounding NSE co-location facilities, algorithmic trading and preferential access to market information. The appellant alleged that selected brokers could obtain faster access to tick-by-tick information and consequently obtain latency advantages over other market participants.

The proceedings discussed:

• Direct Market Access;
• algorithmic trading;
• high-frequency trading;
• co-location;
• latency;
• order-book information;
• market liquidity;
• price discovery;
• preferential access; and
• Section 4 of the Competition Act, 2002. 

The tribunal ultimately did not treat the provision of co-location services itself as an established abuse of dominance on the facts before it. It emphasised, among other things, that the service was made available to eligible trading members subject to stated conditions.

Legal significance

The case is important because it demonstrates that:

Technological advantage is not automatically equivalent to anti-competitive foreclosure.

Competition analysis requires examination of the actual access conditions, discrimination, relevant market, and effects.

It is particularly useful for analysing the boundary between legitimate algorithmic liquidity enhancement and preferential technological access.

Case 2 — Jitesh Maheshwari v. National Stock Exchange of India Ltd. (CCI)

The NSE co-location controversy also generated earlier CCI proceedings concerning whether alleged discriminatory access to exchange infrastructure warranted competition-law intervention.

The later litigation records the relationship between the CCI proceedings and SEBI's regulatory investigation into co-location arrangements. The issues included unequal access, information dissemination and the effect of exchange architecture on trading participants.

Legal significance

The case illustrates an important principle:

Exchange infrastructure can have competition-law significance even where the conduct simultaneously falls within financial-market regulation.

This is particularly relevant to algorithmic liquidity because technological design can affect:

• execution speed;
• information availability;
• trading opportunities;
• market access; and
• competitive equality.

Case 3 — Samir Agrawal v. Competition Commission of India (CCI/NCLAT)

Although this case concerned ride-hailing rather than securities exchanges, it is an important Indian authority concerning algorithmic pricing and competitive autonomy.

The allegation was that Ola and Uber's algorithms effectively determined prices for drivers and passengers, allegedly replacing independent price competition.

The CCI concluded that algorithmic pricing alone did not establish a hub-and-spoke cartel because the necessary agreement or meeting of minds had not been demonstrated. The decision recognised that algorithms can independently calculate prices using large quantities of information relating to demand, traffic, time and other variables.

Legal significance

The case establishes an important distinction:

Algorithmic interdependence ≠ automatically unlawful coordination.

For algorithmic liquidity structures, this means that the fact that numerous trading algorithms react to the same market information does not by itself establish a cartel.

Evidence concerning:

• common instructions;
• information sharing;
• coordination;
• common algorithmic architecture;
• communication between competitors; or
• intentional reduction of competitive uncertainty

may become critical.

Case 4 — Clearstream Banking AG v. European Commission, Case T-301/04

Clearstream concerned securities clearing and settlement infrastructure.

The European Commission found that Clearstream had abused its dominant position through conduct involving access to cross-border clearing and settlement services and discriminatory pricing. The General Court dismissed Clearstream's action against the Commission's decision.

Legal significance

This case is highly relevant to exchange dependency.

A modern trading ecosystem does not end when an order is matched.

It involves:

Trading → Clearing → Settlement → Custody

If one undertaking controls an indispensable part of this chain, its conduct can influence downstream competition.

Thus, algorithmic liquidity may be undermined not by the matching engine itself but by control over post-trading infrastructure.

Case 5 — Slovak Telekom v. European Commission, Case C-165/19 P

The Court of Justice considered access to an incumbent's local-loop infrastructure and the conditions under which access restrictions could constitute abuse under Article 102 TFEU.

The Court dealt with:

• access;
• indispensability;
• regulatory obligations;
• margin squeeze; and
• conditions imposed upon competitors. 

Legal significance

The case is important by analogy to exchange infrastructure.

An exchange could potentially control an infrastructure that competitors need to compete effectively.

However, not every refusal or restriction of access constitutes an abuse.

The legal analysis depends upon factors including:

• whether access is indispensable;
• whether alternative infrastructure exists;
• whether the undertaking is under a regulatory access obligation;
• whether the conduct excludes equally efficient competitors; and
• whether the conduct produces anti-competitive effects.

The case therefore provides a framework for analysing exchange-access dependency.

Case 6 — Oscar Bronner GmbH & Co. KG v. Mediaprint, Case C-7/97

Bronner is a foundational European authority concerning the essential-facilities doctrine.

The Court imposed demanding conditions before a dominant undertaking can be required to provide access to infrastructure controlled by it.

The relevant considerations include:

1. the facility must be indispensable;
2. there must be no realistic alternative;
3. refusal must be capable of eliminating effective competition; and
4. there must be no objective justification for the refusal. 

Application to exchanges

A dominant exchange's matching engine or liquidity pool should not automatically be treated as an essential facility.

The claimant would need to establish something much stronger than:

"This exchange is more liquid."

The question becomes whether effective competition is realistically possible without access to the relevant infrastructure.

Case 7 — United States v. Deutsche Börse AG and NYSE Euronext

The United States challenged the proposed merger between Deutsche Börse and NYSE Euronext as a horizontal merger involving securities and commodity exchanges.

The DOJ identified the transaction as a horizontal merger in the securities and commodity-exchange sector.

Legal significance

Exchange mergers have special competition significance because exchanges do not merely sell a conventional product.

They control ecosystems involving:

• liquidity;
• order flow;
• market data;
• clearing;
• trading technology;
• market participants;
• execution services.

A merger between major liquidity centres can therefore alter the structure of competition itself.

This provides an important precedent for understanding liquidity concentration as a merger concern.

8. Algorithmic Information Exchange

Algorithmic liquidity formation can also create an information-exchange problem.

Suppose several competing market makers submit information to the same algorithmic intermediary.

The intermediary receives:

• prices;
• inventories;
• spreads;
• demand estimates;
• trading intentions;
• execution data.

It then generates recommendations for those same competitors.

The resulting structure can be represented as:

Competitor A → Algorithmic Hub ← Competitor B

followed by:

Algorithmic Hub → A + B

The danger is that the algorithm becomes an information intermediary capable of reducing strategic uncertainty.

The OECD has noted that competition authorities increasingly examine situations where common algorithms combine commercially sensitive information from competing firms and use it to generate pricing or strategic recommendations.

9. Algorithmic Liquidity and Tacit Coordination

Algorithmic markets may make coordination easier because algorithms can:

• monitor competitors continuously;
• respond within milliseconds;
• detect deviations;
• punish aggressive price reductions;
• replicate market signals;
• optimise around competitors' behaviour.

However, parallel algorithmic behaviour alone is not necessarily unlawful.

The distinction is:

Independent adaptation

"My algorithm reacts to publicly observable market conditions."

versus

Coordinated conduct

"My algorithm uses competitively sensitive information supplied through a common mechanism to coordinate my conduct with competitors."

The second situation creates substantially greater competition-law risk.

10. Liquidity Steering

A dominant exchange may potentially influence where liquidity appears.

For example:

Exchange A

• preferential order routing;
• rebates;
• proprietary data advantages;
• superior API access;
• better execution priority.

may cause algorithms to direct orders disproportionately toward Exchange A.

If these advantages result from genuine efficiency, they may be legitimate.

If they arise from exclusionary conduct, concerns can include:

• foreclosure;
• discrimination;
• self-preferencing;
• tying;
• refusal of access;
• discriminatory technical standards.

11. Smart Order Routing and Exchange Dependency

Smart-order-routing algorithms complicate the analysis.

An algorithm may continuously decide:

Where should the next order be sent?

It may consider:

• price;
• available liquidity;
• execution probability;
• fees;
• latency;
• queue position;
• market depth.

Consequently, an exchange may become dependent upon algorithmic routing preferences, while traders become dependent upon the exchange's liquidity.

This creates a two-sided feedback structure:

Exchange liquidity → Routing algorithms → Order flow → Greater exchange liquidity

This is a powerful network effect.

12. Liquidity Fragmentation

Multiple exchanges can sometimes increase competition.

Fragmentation may create:

• competing fees;
• better execution;
• innovation;
• specialised market-making;
• alternative trading venues.

But excessive fragmentation can also increase:

• search costs;
• technological complexity;
• latency;
• data expenses;
• routing costs.

Research on multiple markets and algorithmic trading has found that the relationship between algorithmic activity, fragmentation and liquidity can be complex rather than uniformly positive or negative.

Therefore, competition law should not automatically assume:

"More exchanges = more competition."

Nor should it assume:

"One liquid exchange = anti-competitive."

The actual competitive mechanism must be examined.

13. Relevant Competition-Law Tests

A. Relevant market

Possible markets include:

• exchange trading services;
• securities trading;
• derivatives trading;
• algorithmic trading services;
• co-location services;
• market-data services;
• clearing and settlement;
• execution technology.

Market definition must reflect the actual competitive constraint.

B. Dominance

Indicators may include:

• trading volume;
• liquidity share;
• number of active participants;
• market-data dependency;
• switching costs;
• network effects;
• execution quality;
• technological infrastructure.

High market share is important but not conclusive.

C. Access discrimination

Authorities may investigate whether:

Participant A receives → superior data / latency / connectivity

while

Participant B receives → inferior access

without objective justification.

D. Foreclosure

The question becomes whether the exchange's conduct makes it materially harder for competing venues or participants to compete.

E. Essential facility

The Bronner framework requires particularly strong evidence of indispensability and absence of effective alternatives.

F. Margin squeeze

Where an exchange controls an upstream infrastructure and competes downstream, the relationship between:

wholesale access price + downstream price

may become relevant.

Slovak Telekom provides the major analytical framework by analogy.

14. Algorithmic Liquidity as a Potential Competitive Moat

The combination of:

Data + Algorithms + Liquidity + Network Effects + Infrastructure

can create a particularly strong competitive moat.

For example:

More data → better algorithms → better execution → more traders → more liquidity → more data.

This is a data-liquidity-algorithm feedback loop.

A competition authority may therefore need to examine whether the incumbent's advantage is:

Efficiency-based

or

Artificially protected.

The distinction is fundamental.

15. Possible Anti-Competitive Practices

16. Regulatory Overlap

Algorithmic liquidity is particularly complicated because competition law interacts with financial-market regulation.

In India, for example, the relevant institutional framework may involve:

• Competition Commission of India;
• Securities and Exchange Board of India;
• recognised stock exchanges;
• clearing corporations;
• depositories;
• securities-market regulations.

The NSE co-location proceedings demonstrate the importance of considering both competition-law principles and securities-market infrastructure regulation.

17. Evidence in Algorithmic Liquidity Cases

Traditional evidence may be insufficient.

Authorities may need to examine:

Technical evidence

• source code;
• API architecture;
• matching-engine design;
• latency measurements;
• order logs;
• server-location data.

Economic evidence

• liquidity concentration;
• bid-ask spreads;
• execution probability;
• switching costs;
• market shares;
• price impact.

Algorithmic evidence

• decision rules;
• training data;
• parameter changes;
• automated responses;
• routing instructions.

Communication evidence

• internal emails;
• developer instructions;
• agreements with brokers;
• algorithm-provider contracts.

This makes digital forensic evidence increasingly important.

18. Defences Available to an Exchange

An exchange may argue that the challenged practice:

1. improves liquidity;
2. reduces transaction costs;
3. improves price discovery;
4. increases execution efficiency;
5. reduces latency;
6. improves market resilience;
7. is objectively necessary;
8. is available on equal terms;
9. responds to regulatory requirements; or
10. is subject to effective competition from alternative venues.

The NSE litigation illustrates the significance of arguments that co-location can enhance algorithmic trading, execution and liquidity rather than necessarily restrict competition.

19. Core Legal Principle

The emerging legal framework can be expressed as:

Algorithmic efficiency

↓

Liquidity formation

↓

Network effects

↓

Exchange concentration

↓

Participant dependency

↓

Potential competition concern

But the final step is not automatic.

Competition law intervenes where the concentration or dependency is connected with conduct that produces an unlawful competitive harm, such as:

• exclusion;
• discrimination;
• coordinated conduct;
• foreclosure;
• exploitative conditions; or
• anti-competitive merger effects.

20. Conclusion

Algorithmic Liquidity Formation and Exchange Dependency Structures represent an important emerging competition-law problem at the intersection of financial-market infrastructure, algorithmic decision-making and network effects.

The central legal issue is not whether algorithms create liquidity. They clearly can. The deeper question is whether an exchange can use control over liquidity, data, matching technology, connectivity, clearing or algorithmic access to make itself indispensable and thereby weaken competitive constraints.

The principal case-law lessons are:

1. Manoj K. Sheth v. CCI/NSE — co-location, algorithmic trading, latency and equal access must be examined carefully; technological advantage does not by itself establish abuse. 
2. Jitesh Maheshwari v. NSE — exchange infrastructure and discriminatory-access allegations can raise competition-law questions alongside securities regulation. 
3. Samir Agrawal v. CCI — algorithmic pricing or parallel algorithmic conduct does not itself prove collusion without the necessary agreement or coordination. 
4. Clearstream v. Commission — control over financial-market infrastructure and discriminatory or delayed access can have Article 102 implications. 
5. Slovak Telekom — access restrictions and margin-squeeze principles provide an important framework for infrastructure-dependent markets. 
6. Bronner — essential-facilities intervention requires a high threshold of indispensability and lack of effective alternatives. 
7. United States v. Deutsche Börse/NYSE Euronext — concentration of exchange infrastructure and liquidity can be relevant to horizontal merger analysis. 

Thus, the emerging doctrine can be summarised as:

Control over algorithmically generated liquidity is not inherently anti-competitive; the competition-law concern arises when control over liquidity becomes a mechanism for discriminatory access, exclusion, coordination, foreclosure, or the artificial creation of exchange dependency.