A conditional factor model for real estate investment trusts returns
研究发现，用Instrumented Principal Component Analysis模型识别出的少数潜在系统性因子能解释美国REIT收益的很大一部分，且定价误差小于标准因子模型。REIT特征（规模、杠杆、股息率、动量、物业类型）是因子暴露的主要决定因素。
Abstract We find that a small set of latent systematic factors identified using the Instrumented Principal Component Analysis model explains a substantial share of the cross‐section of US real estate investment trust (REIT) returns. These factors deliver markedly smaller pricing errors than standard REIT and equity factor models, both in‐sample and out‐of‐sample. We also find that REIT characteristics—size, leverage, dividend yield, momentum, and property‐type indicators—are the primary determinants of factor exposures. Evidence from correlation analysis suggests that the latent factors reflect aggregate risk, financing conditions, liquidity/visibility, sector rotation between property types, and a time‐varying dividend‐income premium. Together, these findings point to a unified and economically interpretable factor structure for REITs.