Do Swedish Robo-advisors outperform active funds during market stress?

Sammanfattning

As automated investment services continue to grow, understanding their performance relative to traditional portfolio management becomes increasingly relevant, particularly during turbulent market conditions. Our study examines whether Robo-advisors represent a superior investment strategy compared to traditional actively managed funds during periods of high market volatility. We have applied a quantitative research approach based on historical return data for Swedish Robo-advisors and actively managed funds over the period 2018-2025 in this study. Performance is compared during periods of high volatility. Our analysis focuses on three dimensions: maximum drawdown, risk-adjusted returns measured by the Sharpe ratio, and alpha after fees measured by Jensen’s alpha.The results found show that Robo-advisors consistently experienced lower maximum drawdowns than actively managed funds, indicating stronger downside protection during volatile periods. However, this advantage did not translate into higher risk-adjusted returns. Actively managed funds generally performed equally well or better on a risk-adjusted basis, particularly among equity funds during periods of increased volatility. For alpha after fees, the results provide little evidence that actively managed funds generated significantly higher alpha than Robo-advisors in most cases. Both groups also produced negative alpha during turbulent market conditions. Overall, the findings suggest that Robo-advisors offer greater resilience in terms of limiting losses, while actively managed funds may provide stronger risk-adjusted performance in more extreme market environments. However, neither strategy is able to consistently outperform the market after fees during periods of both increased and extreme volatility. The results should be interpreted with some caution due to the limited sample of Swedish Robo-advisors included in the study, which likely affected the robustness tests and reduced the reliability of the statistical findings.