A Quadrant Approach to Commercial Real Estate Investing

What drives investment performance? The investment types are placed into quadrants, which are distinct but interdependent—more than a classification scheme. While physical property is just one component (or quadrant) of real estate, it is the feed stock of the other quadrants. The other quadrants represent alternative ways to package, price and trade property. Figure 1 shows estimated market capitalizations provided by the Pension Real Estate Association.

Figure 1: Market Capitalization by Quadrant ($ trillions)

The quadrants are neither homogeneous nor independent; they are inseparably bound. The third quadrant ($1.1 trillion) of REITs is impacted by property performance and wider capital markets forces. Property which comprises the fourth quadrant ($1.8 trillion), is often, but not always, leveraged with senior and sometimes mezzanine debt. The second quadrant ($3.8 trillion) includes only private debt, primarily senior mortgages. These mortgages are securitized and used to create our first quadrant ($1.7 trillion) of CMBS notes. The value of mortgage-backed securities depends on the priority with which losses and income are assigned to the various tranches or classes. The more senior the class, the less risky its returns. To the extent that property markets weaken, or loan-to-value ratios change, for example, the performance of all quadrants, not just one, changes.

To analyze how the quadrants are interlinked, we selected time-series datasets that are most representative of each quadrant’s investment characteristics (Table 1).

Table 1: Data Sources by Quadrant

Figure 2: Quarterly Total Return Time-Series by Quadrant (%)

In addition to pure real estate data, we link our CRE performance models with macroeconomic and financial variables. Figure 3 represents how our models leverage these variables. In Tables 2 and 3, we provide summary statistics and a correlation matrix of the variables we will be using in future Viewpoints, respectively. We leverage different corporate bond metrics, stock market variables, and CRE transaction volume. In Table 3, the highest and lowest correlations are assigned dark green and red, respectively.

Figure 3: Capital Markets Road Map

Table 2 highlights the differences between publicly and privately traded assets. While the standard deviations of NPI returns are less than that of stocks, NPI has greater skewness and kurtosis than either stocks or bonds. The Jarque-Bera statistic, which is a measure of normality, indicates that property distributions are not normal. The serial correlation for NPI is much greater than the serial correlation for publicly traded assets. While the measured or apparent standard deviation of NPI is much less than the standard deviation of publicly traded assets, the true standard deviation of property is comparable to that of stocks, REITs and AAA-rated corporate bonds. The serial correlations of BBB-rated CMBS and BBB-rated corporate bonds are comparable and in excess of their respective AAA-rated CMBS and corporate bonds. The reason is that BBB-rated securities are less liquid.

Table 2: Descriptive Statistics, 2000 - 2023

Table 3: Correlation Matrix, 1997 – 2023

The bivariate correlation may obscure the true relationship between two variables due to omitted variables. For example, a scatter plot of interest rates and cap rates shows no apparent relationship, but in a multivariate context, where we can hold NOI growth statistically constant and control for credit risk, there is a strong positive relationship between interest rates and cap rates. In practice, supply-demand imbalances can swamp the interest rate-cap rate relationship. Hence, readers should use bivariate correlations with care.

The correlation matrix indicates the AAA-rated CMBS has a near-zero bivariate correlation with property, whereas BBB-rated CMBS has a higher, but still low, correlation with leveraged or unleveraged property. In general, BBB-rated bonds should be more sensitive to real estate fundamentals than AAA-rated bonds. The G-L mortgage return has a near-zero correlation with property returns. Equity REITs have a low correlation with property but are highly correlated with the S&P 500 and small cap stocks. The BBB-rated corporate bond has a relatively high correlation with the AAA-rated CMBS tranche. Transactional sales volume growth is highly correlated with property. We believe that transactions volume, which is very volatile over the cycle, impounds important information regarding the underlying property markets. The G-L mortgage return is highly correlated with AAA- and BBB-rated CMBS. NOI growth exhibits modest correlation with leveraged and unleveraged property but low correlation with senior or subordinate CMBS.

In this Viewpoint, we have defined and described the quadrants for CRE investing, estimated their market capitalizations, and their historical investment returns. In future work, we will explore each quadrant in more depth. Finally, we will lay out the complex web of influences linking the quadrants and ways to glean information for them.

In Figure 4, we present a preview of the work to come. Arrows indicate the flow of information from one quadrant to another. For instance, there is an arrow from public debt to private equity. This is because we include CMBS returns in our models to explain property returns. Future Viewpoints will examine this web in further detail.

Figure 4: Flow of Information Between Quadrants

CBRE Econometric Advisors invited Dr. Randall Zisler to co-author this paper on the real estate quadrants. His unique approach reflects a multifaceted career as a Princeton University professor, Goldman Sachs research director, pension fund consultant, and investment banker.