Copyright Project Syndicate
CAMBRIDGE – Economists around the world are rightly focused on how AI will reshape labor markets. But the next decade’s most consequential shift may unfold in a different market altogether: equity. By transforming how data are generated and shared, AI could sharply reduce the costs of information, monitoring, compliance, and market-making. In doing so, it could turn equity from an elite privilege into a major source of financing, especially for small and young firms. Outside a handful of countries like the United States, the United Kingdom, Canada, and Singapore, vibrant equity markets are scarce. And even in such cases, the “long tail” of public companies is remarkably thin. Few firms go public at valuations below $10 million, and when they do, liquidity is negligible, and the fixed costs of being public are prohibitive. In much of the world, including Western Europe, “financial development” has come to mean “banking development.” While banks are indispensable, they sell only one product: debt. A healthy balance sheet, however, depends on a balanced mix of debt and equity – the riskier the venture, the more equity it requires. And therein lies the rub: while debt markets are robust, equity markets are not. In practice, banks lend in proportion to the equity cushion a firm already has. But that equity typically comes from two sources: friends-and-family money or retained earnings. The first is limited and unequally distributed; the second constrains growth. As a result, firms that rely on self-financing can grow only as fast as their profits, rather than their opportunities. That may suffice for a bakery, but not for a digital platform that depends on network effects or a manufacturer that must invest heavily before it generates any revenue. Many transformative firms – from e-commerce giants like Amazon to ride-hailing pioneers like Uber – operated at a loss for years. A banking-centric system cannot fund that kind of growth. The scarcity of equity can be largely attributed to asymmetric information. While lenders can secure collateral, demand fixed payments, and seize assets if things go wrong, equity investors are last in line and get whatever remains after everyone else has been paid. This makes disclosure, due diligence, and governance critical to the viability of equity markets. Yet for small firms, the costs often exceed the potential gains. Add illiquidity, high listing costs, limited analyst coverage, and weak protections for minority shareholders, and you have a ladder with no bottom rungs. AI can help build them, starting with disclosure. Today’s small issuers face compliance systems built for giants, but AI can lower costs by generating machine-readable financials and cross-checking invoices, bank statements, tax filings, and payroll records in real time – and flagging inconsistencies before they become scandals. Standardized, machine-readable data would enable investors (or their bots) to make instant, apples-to-apples comparisons across thousands of firms, not just the few big names analysts already cover. Over time, AI-powered monitoring could make the traditional annual audit seem as outdated as dial-up internet. Moreover, natural-language models can review contracts, permits, litigation histories, and environmental disclosures, while time-series models can reconcile orders, shipments, and cash flows. Tasks that once required armies of associates can now be completed in minutes, and the output will be cheaper, more consistent, and fully auditable. This shift could give rise to a new business model: the AI underwriter. Instead of collecting hefty fees from a few large clients, such underwriters could take dozens of smaller issuers public each week using standardized disclosures, automated checks, and real-time risk alerts. Another opportunity involves liquidity. Algorithmic market makers – already standard among large public firms – could be extended responsibly to smaller issuers once disclosures are standardized and monitoring becomes continuous. AI-driven research could help improve pricing and liquidity by increasing information availability and trust, while matching algorithms could align investor mandates and issuer profiles far more precisely than today’s crude groupings. In many emerging economies, related-party transactions, asset tunneling, and sudden dilutions have long eroded investor trust. Here, too, AI-enabled monitoring systems can help enforce minority-shareholder protections by tracking transactions, board minutes, procurement records, and trading patterns in real time, alerting regulators and investors to potential abuses. Meanwhile, programmable governance – charters that automatically enforce preemptive rights, protections, and dividend triggers – can turn legal text into enforceable code, making it far harder for insiders to dilute or disadvantage outside investors. Lastly, standardized digital disclosures and robo-advisers could help democratize investment in smaller companies, building diversified portfolios of shares in small and medium-size enterprises (SMEs) tailored to investors’ risk tolerance, location, and goals. Pension funds and insurers, which are largely absent from this segment, could be allowed to allocate modest portions of their assets to AI-curated indices, thereby eliminating the need for costly in-house teams. But for this vision to materialize, policymakers will need to establish a small public equity regime built on four pillars: continuous, streamlined reporting, instead of bulky periodic filings; liability protections for issuers that adopt AI-based verification systems; simplified listing requirements; and open-data frameworks that allow third parties to add value by analyzing disclosures. After decades of subsidizing credit, public development banks should channel some of their resources into equity investments. For example, they could support firms that meet AI-based disclosure and verification standards, provide first-loss coverage for diversified SME equity funds, and promote shared oversight infrastructure. To be sure, AI can hallucinate, models can be gamed, and insiders will always know more. But the test is relative advantage, not perfection. If AI cuts transaction costs by 50-90%, the impact would be transformative: though some frauds will inevitably slip through, many more legitimate businesses would become viable candidates for outside investment. To broaden ownership, spur innovation, and accelerate growth, we should direct AI toward the higher-return challenge of making equity abundant. The goal, crucially, is not a casino but a safer, cheaper, data-rich marketplace where risk is borne by those best able to bear it. After a century spent perfecting the plumbing of credit, AI offers a chance to complete – at last – the financial system’s other half.
