<p>This paper investigates the herding behavior among investors in the market of Japanese real estate investment trusts (J-REITs). Using a survivorship bias-free dataset of daily returns, we capture the emergence of herding effects through measures that assess the convergence of individual returns towards the market consensus. We document that the results concerning the existence of herding behavior during the full sample depend on the choice of the respective measure. Sub-sample analysis and rolling window regressions show that the herding dynamics vary substantially over time, with the COVID-19 pandemic leading to the intensification of the phenomenon. The episodes of herding and anti-herding are concentrated in the initial part of the dataset, aligning with improvements in J-REIT market efficiency over time. Additionally, the deterioration of investor sentiment is identified as a significant driver of herding in J-REITs.</p>