Search

Purpose: The purpose of this study is to examine the return sensitivity of public farmland and timberland real estate investment trusts (REITs) to private-equity farmland, timberland and real estate, long-term corporate bonds and large- and small-cap stocks. The study also examines time-dependent contributions of selected asset classes to farmland and timberland REIT volatility. Design/methodology/approach: The authors use a multi-factor asset pricing model under a seemingly unrelated regression framework to evaluate farmland and timberland REIT returns, and a state-space model with the Kalman filter to evaluate the time-dependent contributors of farmland and timberland REIT volatility. The authors first perform orthogonalized regressions to obtain pure independent factors, and then decompose volatility into individual asset components. Findings: Significant loadings on financial assets are found for both farmland and timberland REITs, suggesting that they are generally driven by some common state variables. Large-cap stocks are found to be the major contributor of farmland and timberland REIT volatility, despite some differing patterns over time. Originality/value: Empirical analysis of farmland REIT is very scarce. The authors compare the risk-return characteristics of farmland and timberland REITs under a state-space framework with the Kalman filter. This study can improve the understanding of the roles of farmland and timberland REITs in a multi-asset portfolio. [ABSTRACT FROM AUTHOR]

This study introduces two Markov chain time approaches, time-homogeneous and nonhomogeneous models, for analyzing farm credit risk migration as alternatives to the traditional discrete-time (cohort) method. The Markov chain models are found to produce more accurate, reliable transition probability rates using the 3 x I migration measurement method used by farm lenders. Compared to corporate bond ratings migration results, this study obtained larger mean differences in singular value decomposition between the cohort matrix and each of the Markov chain matrices. This finding suggests that the omission of transient, indirect migration activities under the cohort method is more costly when applied to farm credit analysis. This discrepancy could lead to understated transition probability estimates which, in turn, could produce misleading indicators of farm loan portfolio quality. [ABSTRACT FROM AUTHOR]