A New Metric for Evaluating Performance of REITs.

Joseph CHENG, Alfred LAM

Research output: Journal PublicationsPolicy or Profession paperProfessionpeer-review


It has been suggested that conventionalearnings are a poor metric for evaluating the Real Estate Sector which comprises mainly Real Estate Investment Trusts (REITs) because conventional accounting practice “implicitly assumes that the value of real estate assets diminishes predictably over time.” Thus, Funds From Operations (FFO) recently has been proposed as an alternative metric for evaluating REIT performance. This article sets forth a proposition that REITs with debts might be better evaluated by the Unlevered EPS (earnings per share), which is derived by purging out the portion of earnings contributed by financial leverage. Furthermore, relative performance of REITs might be better evaluated by Unlevered Funds From Operations per Share (FPSU), whose portion attributed to debt is also purged out. These unlevered earnings metrics, which are Unlevered EPS (EPSU), and Unlevered Funds From Operations per Share (FPSU), may provide a clearer focus on earnings that are purely performance related. This article describes an interesting comparative study for these metrics which shows that the relative performance of REITs bears a stronger correlation with
standardized percentage change in unlevered earnings metrics than with the standardized percentage change in standard accounting earnings metrics, thereby highlighting the merit of disclosing the Unlevered EPS and Unlevered
FPS to stakeholders as an additional information for a more precise evaluation of performance of REITs.
Original languageEnglish
Pages (from-to)125-132
Number of pages8
JournalReal Estate Finance
Issue number3
Publication statusPublished - Jan 2020


  • REAL estate investment trusts
  • FINANCIAL leverage
  • INTEREST rates
  • ACCOUNTING standards
  • LONG-term debt


Dive into the research topics of 'A New Metric for Evaluating Performance of REITs.'. Together they form a unique fingerprint.

Cite this