Skip to main content

The Future of Liquidity Provision

The market-making function has evolved continuously since the days of floor traders shouting prices on the New York Stock Exchange. Today, the industry faces a period of substantial transition shaped by technological innovation, regulatory pressure, competitive consolidation, and the emergence of new market structures like decentralized finance. Understanding these trajectories matters for investors because changes in how liquidity is provided will directly impact the trading costs and market stability you experience. The future of market making will determine whether capital allocation in markets becomes more efficient or less, whether retail investors gain or lose relative to institutions, and whether systemic risks increase or decrease.

Quick definition​

The future of liquidity provision encompasses emerging trends, technologies, and market structures that will reshape how market makers operate, including algorithmic intelligence, regulatory evolution, alternative trading venues, decentralized finance protocols, and potential shifts toward market-maker obligations or restrictions.

Key takeaways​

  • AI and machine learning are enhancing market-maker models: Firms are moving beyond traditional statistical models toward deep learning approaches that recognize patterns in order flow and market microstructure at scales human analysts cannot.
  • Regulatory pressure is mounting: The SEC and Congress are considering reforms that could restrict PFOF, mandate best execution benchmarking, or impose liquidity obligations on designated market makers.
  • Consolidated market structure creates fragmentation: As exchanges proliferate and dark pools expand, market making has become more fragmented, with multiple venues requiring separate quoting and complex routing algorithms.
  • Decentralized finance challenges the liquidity provision model: DEX protocols like Uniswap provide liquidity through algorithmic market makers (AMMs) rather than human-directed traders, potentially disrupting traditional market-making economics.
  • Regional markets and emerging economies offer growth: As equity markets in Asia and emerging markets grow, opportunities for market makers expand but with different structural constraints than mature markets.
  • Volatility regimes may shift: If central banks maintain higher-for-longer interest rates or inflation volatility persists, the stable-low-volatility regime of 2010-2019 may not return, challenging market-maker profitability models.
  • Technology costs are escalating: Investment in AI, infrastructure, and data infrastructure is rising faster than market-making revenue, potentially pushing smaller competitors out and accelerating industry consolidation.

Future of Liquidity Provision: Key Drivers​

Regulatory Reforms on the Horizon​

The regulatory environment for market makers is tightening. Multiple proposals have been floated that could reshape market-making economics:

Payment for Order Flow Restrictions​

One of the most significant potential changes is banning or restricting payment for order flow. Congressional pressure has mounted, particularly after the 2021 GameStop retail trading episode and subsequent retail trading boom. Critics argue that PFOF distorts routing incentives and disadvantages retail traders.

If PFOF were banned, market-making economics would shift dramatically:

  • Brokers would need alternative revenue models: They could charge commissions again (returning to the pre-2015 model), charge bid-ask spreads directly, or implement subscription models. Most retail brokers will likely resist commissions, fearing customer migration.

  • Retail order flow would become less valuable: Without PFOF payments, brokers would have less incentive to direct retail orders to particular market makers, and retail orders would flow to the venues offering the best prices.

  • Retail execution might improve: Without PFOF incentives distorting routing, retail traders would likely receive better average execution prices. Academic research suggests PFOF restrictions would improve retail execution by 1-2 basis points, worth billions annually to retail traders collectively.

  • Market makers would adjust business models: Firms might increase focus on proprietary trading, reduced spreads due to less favorable order flow, or consolidation as smaller firms exit.

The SEC has held hearings on PFOF, and several congressional bills have proposed restrictions. However, brokers have lobbied extensively to preserve PFOF, arguing it enables commission-free trading. As of 2024, no ban has been implemented, but the regulatory probability is non-negligible.

Best Execution Enforcement​

The SEC has proposed amendments to Rule 10b-1 (Best Execution) that would tighten oversight and disclosure requirements. Under the proposed rule, brokers would be required to:

  • Demonstrate quantitatively that they achieved best execution, with specific benchmarks and metrics.
  • Disclose execution quality metrics publicly, making inter-broker comparisons transparent.
  • Evaluate market makers' execution quality more rigorously and adjust routing dynamically based on performance.

If implemented, these changes would create measurable pressure on market makers to narrow spreads and improve pricing. Firms known for poor execution would lose order flow. Market-maker profitability would likely compress, particularly among firms quoting poor prices.

Potential Liquidity Obligations​

Some regulatory proposals have suggested imposing explicit liquidity obligations on market makers during market stress. Currently, market makers can withdraw quotes during crises. Some proposals would require market makers to maintain minimum liquidity (e.g., quotes at reasonable spreads) even during volatility spikes.

Such obligations would:

  • Stabilize markets during crises: If market makers were forced to maintain quotes, the flash crashes and volatility spirals driven by withdrawal could be prevented.

  • Increase market-maker losses during crises: Market makers would be forced to trade at unfavorable prices, absorbing losses during exactly the periods when losses are most severe.

  • Potentially reduce entry to market making: Firms would demand higher spreads to compensate for crisis-period losses, increasing normal-period spreads and transaction costs.

Regulators have been reluctant to impose strict liquidity obligations because they would increase costs and potentially reduce market-making participation. However, after each crisis (2010, 2020, 2021-2022), momentum builds for such proposals, though momentum typically fades once markets stabilize.

Artificial Intelligence and Machine Learning Advancements​

One of the most significant trends in market making is the deployment of artificial intelligence and machine learning to improve trading models.

Traditional market-making models rely on:

  1. Econometric time-series models: ARIMA, GARCH, and similar statistical models that assume specific forms of autocorrelation and volatility clustering.

  2. Volatility estimation: Historical volatility, realized variance estimators, and implied volatility from options prices.

  3. Order-flow models: Techniques to infer large orders hidden in market data and predict their impact.

  4. Hedging algorithms: Rules-based systems for managing inventory and delta exposure.

These approaches work but have known limitations: they assume specific functional forms, require careful parameter tuning, and can be disrupted by regime changes.

Machine learning and AI approaches promise improvements in several dimensions:

Deep Learning for Pattern Recognition​

Large neural networks can identify patterns in high-dimensional data that traditional statistical models miss. For example, a deep learning model could be trained on millions of historical one-second intervals of order flow, trades, and prices to predict the next price move at a microsecond resolution. The model would discover interactions between different order types and price movements that humans wouldn't hypothesize.

Firms like Citadel Securities and Virtu are known to be investing heavily in deep learning for trading. The competitive advantage, if real, could be substantial: even a 0.1-basis-point edge in predicting the next price move could generate millions in annual profits.

However, the advantage may be temporary. As more firms deploy similar models, the edge could erode through competition. Additionally, machine learning models trained on historical data can be disrupted by regime changes (e.g., changes in market structure, participant composition, or market conditions).

Reinforcement Learning for Strategy Optimization​

Reinforcement learning trains agents to maximize rewards through trial-and-error learning. A reinforcement learning algorithm could be trained to manage a market-making portfolio by simulating millions of market scenarios and optimizing quoting behavior to maximize profit while constraining risk.

This approach could potentially discover novel quoting strategies that traditional models miss. However, reinforcement learning is computationally expensive and requires careful design of reward functions to avoid converging on unstable or deceptive strategies.

Natural Language Processing for News and Information Extraction​

Market-making can be disrupted by news events. A machine learning model that processes news in real-time and extracts quantitative implications (e.g., "this earnings miss implies the stock should be down 5%") could allow market makers to adjust quotes preemptively. Firms are investing in natural language processing to feed news-driven signals into quoting algorithms.

Challenges and Limitations​

Despite the promise of AI, several challenges remain:

  1. Data snooping and overfitting: Machine learning models can appear to discover patterns that are actually noise. Unless models are carefully validated and tested on out-of-sample data, they can be profoundly wrong.

  2. Adversarial adaptation: As other traders (institutional and retail) learn about successful AI strategies, they adapt. A strategy that works for 6 months may break when enough competitors copy it.

  3. Regulatory uncertainty: If AI-driven strategies are deemed to constitute market manipulation, they could be banned. Regulators are still developing frameworks for AI trading.

  4. Explainability: Machine learning models can be "black boxes" whose decision-making logic is opaque. Regulators may be reluctant to permit trading strategies they can't explain.

Overall, AI and machine learning are likely to improve market-maker efficiency, potentially narrowing spreads and improving pricing for all traders. However, the advantages will likely be temporary as the technology diffuses.

Fragmented Market Structure and Complexity​

Market-making has become increasingly fragmented due to the proliferation of trading venues:

  1. Multiple stock exchanges: NYSE and NASDAQ used to be the only venues; now there are over 13 U.S. stock exchanges including EDGX, EDGA, BATS, CBOE, IEX, and others.

  2. Dark pools: Venues like Citadel Securities' Apogee, BIDS Trading, and Instinet's CrossFinder allow investors to trade without public quotes visible to other traders.

  3. Broker-dealer crossing networks: Brokers operate internal crossing networks where customer orders cross against each other.

  4. International venues: U.S. market makers can access international venues, and vice versa.

This fragmentation creates both opportunities and challenges:

Challenges:

  • Complexity increases: Market makers must quote and monitor positions across multiple venues simultaneously, requiring sophisticated technology.

  • Liquidity fragmentation: Order flow is split across venues, reducing liquidity on each individual venue and widening spreads.

  • Latency arbitrage opportunities: Small price discrepancies across venues reward high-speed traders, increasing technology arms races.

  • Regulatory complexity: Each venue has rules; market makers must navigate different quoting requirements, position reporting, and enforcement regimes.

Opportunities:

  • Arbitrage profits: Prices can differ across venues, allowing alert market makers to profit from buying on one venue and selling on another.

  • Order routing optimization: Market makers can choose which venue(s) to quote based on inventory levels and order flow, optimizing across the fragmented structure.

  • Less competition: Certain venues or security segments might have fewer market makers, allowing wider spreads.

The fragmentation trend seems likely to continue as new exchange models emerge and alternative trading venues grow. This will increase the technological demands on market makers and likely further concentrate the industry among firms that can invest in sophisticated routing and inventory management systems.

Decentralized Finance and Algorithmic Market Makers​

A fundamentally different approach to liquidity provision has emerged through decentralized finance (DeFi) and automated market makers (AMMs).

In traditional markets, liquidity comes from human market makers quoting prices. In DeFi, liquidity comes from liquidity pools—smart contracts that hold reserves of two or more assets and allow traders to exchange one asset for another against the pool.

The canonical example is Uniswap: a decentralized exchange where any trader can deposit equal values of two assets (e.g., $1 million of ETH and $1 million of USDC) and earn a percentage of trading fees. The AMM uses a simple formula: the product of the quantities of the two assets remains constant (or nearly so). If a trader wants to buy ETH, they deposit USDC and withdraw ETH, causing USDC to increase and ETH to decrease in the pool. The ratio of prices shifts to incentivize more deposits of ETH to restore balance.

Key differences from traditional market making:

  1. No active management: Liquidity providers deposit capital and earn fees passively. They don't actively manage positions or hedge risks; the AMM formula determines pricing.

  2. Transparent pricing: Pricing is determined by the formula and pool composition, not by human market makers' discretion. This creates fairness: everyone gets the same prices and can calculate expected execution prices in advance.

  3. Lower barriers to entry: Any trader can become a liquidity provider by depositing assets. Capital requirements are the only constraint, not technology or expertise.

  4. Impermanent loss risk: Liquidity providers face "impermanent loss"—if the price of one asset in the pool diverges significantly from the deposit price, the provider's capital value declines. This is the risk compensation for earning fees.

Implications for traditional market making:

  • Competition in DeFi markets: DeFi markets are growing rapidly. As more trading volume migrates to decentralized exchanges, traditional market makers lose order flow and profitability.

  • Hybrid models emerging: Some firms are building technology to arbitrage across traditional and decentralized venues, creating new profit opportunities while reducing price discrepancies.

  • Innovation pressure: The transparency and accessibility of AMM-based liquidity provision creates pressure on traditional markets to become more transparent and accessible.

  • Systemic risk: As more trading volume migrates to DeFi, the stability of DeFi systems becomes critical. DeFi liquidity is sometimes shallow and vulnerable to flash loan attacks and other exploits.

Currently, DeFi markets remain far smaller than traditional markets. However, the trajectory suggests meaningful migration to DeFi could occur over the next decade, particularly for certain asset classes like cryptocurrency pairs and tokenized commodities.

Market-Making in Emerging Markets and New Asset Classes​

As equity markets in emerging economies grow and new asset classes (cryptocurrencies, tokenized assets, etc.) emerge, market-making opportunities expand.

Emerging market challenges:

  1. Less capital availability: Smaller market capitalizations and lower trading volumes mean market makers face higher inventory risk and wider spreads.

  2. Currency risk: Market makers in foreign markets face currency exposure that requires hedging.

  3. Regulatory uncertainty: Emerging markets have less established regulatory frameworks, creating legal risk for market makers.

  4. Infrastructure constraints: Trading technology, data feeds, and settlement infrastructure may be less developed.

New asset class opportunities:

  1. Cryptocurrency pairs: As cryptocurrencies mature, market makers specialize in providing liquidity in crypto pairs. Margins can be wider than equities due to higher volatility and less liquidity.

  2. Tokenized assets: As real-world assets (real estate, commodities, equities) become tokenized on blockchain networks, market makers will specialize in these markets.

  3. Synthetic assets: Platforms like Synthetix allow traders to buy synthetic exposure to various assets. Market makers provide liquidity by accepting the other side of these synthetic trades.

These new markets and asset classes offer growth opportunities but also require market makers to develop new expertise and manage new types of risks.

Changes in Volatility Regimes​

Market-making profitability is highly sensitive to realized volatility. The relationship is complex:

  1. Very low volatility: If realized volatility is very low, market makers profit from theta decay and the bid-ask spread. However, if realized volatility is below implied volatility (market expectations), options market makers lose money.

  2. Moderate volatility: Market makers can adjust spreads to compensate for expected losses from gamma and vega exposure.

  3. Very high volatility: Market makers face enormous gamma losses, and spreads become too wide to allow trading.

The 2010-2019 period had unusually low and stable realized volatility, which was favorable for market-maker profitability. The VIX (implied volatility index) averaged around 13-15 during this period.

If volatility regimes shift higher due to geopolitical tensions, inflation persistence, or recession risks, market-maker profitability would be squeezed:

  • Higher spreads required: To compensate for increased gamma and vega losses, market makers would quote wider spreads.

  • Higher transaction costs: Investors would face higher trading costs.

  • Reduced market efficiency: Wide spreads reduce the efficiency of capital allocation.

Conversely, if a new low-volatility regime emerges, market-making spreads could tighten further, improving outcomes for traders.

Example 1: AI-Driven Volatility Modeling: A major market maker implemented a deep learning model trained on 20 years of options data to predict implied volatility movements across all strikes and expirations. The model identified subtle patterns in how the volatility smile responds to market stress, allowing the firm to adjust positions preemptively before volatility actually spiked. The edge was estimated at 2-3 basis points annually on their options volume, worth tens of millions in profits.

Example 2: Uniswap's Market Impact: As Uniswap volumes grew from $1 billion daily to $2+ billion daily, the implicit spreads in major Ethereum pairs tightened as liquidity providers entered to capture trading fees. Traditional market makers operating in the same currency pairs faced competition from AMM liquidity, pushing some to focus on emerging tokens with less AMM depth.

Example 3: IEX's Market Structure Innovation: IEX Exchange implemented an intentional communication latency ("speed bump") to reduce latency arbitrage opportunities. Market makers had to adjust strategies and brokers had to redirect order flow. IEX captured market share from larger exchanges by offering a fairer market structure, demonstrating that regulatory and structural innovation can shift competitive dynamics.

Common Mistakes About the Future of Market Making​

Assuming AI will eliminate human traders entirely: AI will enhance human-directed trading and decision-making but is unlikely to fully replace discretionary judgment. Machine learning models are vulnerable to overfitting, regime change, and adversarial adaptation.

Believing spreads will perpetually narrow: While technology generally tightens spreads, periods of stress or uncertainty cause spreads to widen. The long-term trend may be tighter spreads on average, but the volatility of spreads may also increase.

Thinking decentralized finance will immediately replace traditional markets: DeFi is growing but has vulnerabilities (security risks, liquidity slippage, smart contract risks) that limit adoption. Traditional markets will coexist with DeFi for many years.

Overlooking regulatory changes as exogenous shocks: Regulatory actions can dramatically alter market-making economics. A ban on PFOF, stricter best execution requirements, or new circuit breaker rules would reshape the landscape.

Assuming market structure remains stable: Market structure changes (new venues, new order types, fragmentation) happen gradually but have profound effects on market makers' economics.

FAQ​

Will AI replace human market makers?

AI will enhance market makers' models and decision-making but is unlikely to fully replace human market makers. Human judgment, risk tolerance, and strategic decision-making remain important, particularly during novel market conditions.

Is payment for order flow likely to be banned?

Congressional interest in PFOF restrictions is growing. However, brokers' lobbying efforts and lack of clear consensus on alternatives make a complete ban unlikely in the next 2-3 years. Restrictions and enhanced disclosure are more probable.

Will decentralized finance replace traditional stock markets?

DeFi will grow and coexist with traditional markets. However, DeFi faces challenges (security, scalability, regulatory uncertainty) that prevent wholesale replacement of traditional markets in the near term. A hybrid model with some trading in DeFi and some in traditional venues is most likely.

What would happen to market making if a liquidity obligation were imposed?

Market makers would require wider spreads during normal periods to compensate for losses during stress periods when they'd be forced to trade at unfavorable prices. Overall market efficiency would likely worsen.

How might shifts in market structure affect retail traders?

Fragmentation increases routing complexity, potentially worsening retail execution unless brokers invest in sophisticated order routing. Regulatory changes could improve or worsen retail execution depending on the specific rules. Technology improvements could tighten spreads, benefiting all traders.

Will market-making margins compress further?

Long-term trend toward tighter spreads is likely due to technology and competition. However, margin compression may be offset by increased volumes and new markets as economies grow.

What new business models might emerge in market making?

Possible future models include: market makers becoming more like hedge funds with significant proprietary trading, hybrid models combining market making with market-structure optimization, specialized market makers focusing on narrow segments (e.g., cryptocurrency or emerging markets), and technology-focused firms licensing market-making algorithms to smaller brokers.

Summary​

The future of liquidity provision is shaped by several powerful trends. Regulatory pressure to restrict PFOF and improve best execution is mounting, which could improve retail execution but compress market-maker margins. Artificial intelligence and machine learning are enhancing market-making models, potentially tightening spreads through improved pricing models and competition. Market fragmentation across multiple exchanges and dark pools is increasing complexity and widening spreads for traders. Decentralized finance and automated market makers represent a fundamental alternative to human-directed liquidity provision, though they currently remain smaller than traditional markets. Emerging markets and new asset classes offer growth opportunities for market makers but with different risk and regulatory profiles. Volatility regime shifts could substantially alter market-making profitability and market efficiency. Overall, the future landscape will likely be more technologically sophisticated, more regulated, and more competitive than the present, with implications for both market makers' profitability and traders' execution quality. Investors should monitor regulatory developments (particularly on PFOF and best execution) and be aware that market structure is continuously evolving in ways that affect their trading costs and market stability.

Next​

Investor Mistakes About Market Makers