Managing Herding Risks in the Republic of Korea’s Robo-Advisory Market

Since their introduction in 2016, robo-advisory services in ROK have expanded rapidly, with both client numbers and assets under management (AUM) rising sharply. Growth has been particularly strong in discretionary services, including retirement pension accounts, which now represent a big share of the market. By early 2026, more than 380,000 clients were using robo-advisory platforms, managing over KRW 1.25 trillion^[1]($885 million) in assets.

This expansion reflects the appeal of algorithm-based portfolio management, which offers customized investment strategies and automated rebalancing. Yet regulatory changes, such as quarterly ex-post reviews of approved algorithms, have prompted some asset managers to discontinue certain models due to costs and complexity. Despite these challenges, overall AUM has continued to grow, underscoring the sector’s resilience and potential.

Figure 1: Growth in Robo-Advisory Clients and AUM

Source: Koscom.

At the same time, evidence suggests that robo-advisors may contribute to herding behavior—where algorithms converge on similar investment decisions during periods of market stress. Herding can arise when investors imitate one another or when algorithms independently react to shared information, leading to synchronized trading patterns. While AI tools can reduce human biases, they may also rationally follow early movers, amplifying volatility.

Figure 2: Trends in Robo-Advisory Services for Retirement Pensions

Source: Koscom.

Empirical studies across major economies show that herding intensifies during downturns and high-volatility regimes. ROK’s own data supports this: analysis of daily returns^[2] from robo-advisory algorithms reveals that the increase in return dispersion slows during market shocks, a clear sign of herding. This convergence of strategies may reduce diversity in market views, heighten price swings, and propagate systemic risk.

The continued expansion of AI-driven robo-advisory services in ROK highlights the need for stronger safeguards to protect financial stability. While these platforms broaden access to investment management, their tendency to converge on similar strategies during market stress raises systemic concerns. If left unchecked, synchronized trading could amplify downturns, distort price signals, and undermine investor confidence.

International experience offers useful lessons. In the United Kingdom, regulators have warned that rapid AI integration in financial services could heighten systemic risks, including cyber vulnerabilities, reliance on a few technology providers, and algorithmic herding. The Bank of England and the Financial Conduct Authority have been urged to conduct stress tests to assess potential AI-induced disruptions. Similarly, the United States Securities and Exchange Commission has prioritized oversight of automated investment tools, with particular attention to risks facing retail and elderly investors.

It is important to establish streamlined oversight mechanisms that balance innovation with resilience, reducing compliance burdens while maintaining robust monitoring. Such a framework would enable ROK to sustain the growth of its robo-advisory market while safeguarding against systemic vulnerabilities. The challenge is not to slow innovation, but to ensure that technological progress strengthens, rather than weakens, financial stability.

^[1] The figures are drawn from statistics covering only firms that have passed Koscom’s robo‑advisor testbed.

^[2] CSAD_t = ∑^N_i=1 | R_i,t - R_m,t | 1/N, R_i,t denotes the daily return of robo-advisors, while R_m,t represents the average return across all robo-advisors at time t.

Korea Institute of Finance website.