Intelligent Investment

A Quadrant Approach to Commercial Real Estate Investing: Private Equity

February 5, 2025 5 Minute Read

Download PDF
Aerial view of the city

Executive Summary

Receive EA Insights Directly in your Inbox

  • Property returns exhibit serial correlation, making them partially predictable based on past returns. This needs to be considered when comparing private and public property returns.
  • This serial correlation masks the risks from private CRE equity. Failure to account for this hidden risk would lead to overallocation relative to the other quadrants.
  • Our property models explain more than 80% of the historic levered and unlevered property (NPI) returns. These models include lagged REIT returns, BBB CMBS returns, NOI growth, and lagged property returns.

Introduction

This Viewpoint is the fifth in a series that explores the relationship between the CRE quadrants of public debt, private debt, private equity and public equity. Specifically, we will examine the historic performance of directly held property and CRE equity funds in relation to the quadrants.

There are significant differences between the characteristics of private market and public market return data, which impact risk and public-private analyses. Unlike publicly traded stocks (such as REITs) and some fixed income securities, private assets do not trade in continuous auction markets.

Private market pricing relies on backward-looking methods like appraisals, resulting in a lagged reaction to public market shocks. Hence, public markets can help predict the performance of private assets due to these differences. Additionally, property returns exhibit serial correlation or smoothing, meaning that past returns can predict future returns to some extent. On the other hand, heavily traded stocks like the S&P 500 or public REITs show little to no serial correlation.

Public markets are also more transparent and more liquid; by contrast, private markets are illiquid and transactions costs are higher. Private returns are serially correlated or smoothed because of their backward-looking nature, while public markets are forward-looking and exhibit near-random fluctuations. The randomness of public prices does not mean, however, that public markets defy or lack causality. Quite the contrary. Public markets are just more efficient at impounding news; prices fluctuate as if they are random.

The impact of serial correlation on returns is an important distinction between private real estate and public real estate markets because it poses some empirical challenges, especially when comparing risk-return performance, such as REITs with property, or when determining the optimal asset allocation in a portfolio of widely traded stocks, bonds, property and mortgages. Serial correlation masks risk, leading investors to incorrectly perceive property as less risky than stocks.

Correcting for serial correlation shows that property’s true return volatility is similar to that of traded equity REITs. We compare past and current total quarterly returns of traded public shares—S&P 500 and all equity REITs—with unleveraged and leveraged property (the NCREIF all-property leveraged index).

A scatter of the public assets, S&P 500 and equity REITs indicates that there is no statistically significant relationship between past and current returns (see Figures 1 and 2 and Models 1 and 2). T-statistics and adjusted R-squared are insignificant. The coefficient on each variable, lagged one quarter, is essentially zero statistically.

Figure 1: Past S&P 500 Returns Do Not Predict Current S&P Returns

a-quadrant-approach-to-commercial-real-estate-investing-private-equity-Fig01

Source: Morningstar.

Past S&P 500 returns do not predict current returns, which is a characteristic of an efficient market. The coefficients on the lagged-dependent variables are not significantly different from zero:

Models 1 and 2: S&P 500 and REIT Returns Explained by Lagged Returns

Coefficients S&P 500 Total Return REIT Total Return
(T-Statistics)    
1-Quarter Lagged Total Return 0.0207
(0.27)		 0.088
(1.174)
2-Quarter Lagged Total Return 0.00348
(0.05)		 -0.112
(-1.484)
3-Quarter Lagged Total Return -0.0509
(-0.68)		 -0.058
(-0.773)
Constant 3.218***
(4.43)		 3.417***
(4.397)
Observations 179 179
Adjusted R-squared -0.014 0.008
Durbin-Watson 1.988 2.00
Notes: * p < 0.05, ** p < 0.01, and *** p < 0.001.

Models 1 and 2: S&P 500 and REIT Returns Explained by Lagged Returns

Coefficients S&P 500 Total Return REIT Total Return
(T-Statistics)    
1-Quarter Lagged Total Return 0.0207
(0.27)		 0.088
(1.174)
2-Quarter Lagged Total Return 0.00348
(0.05)		 -0.112
(-1.484)
3-Quarter Lagged Total Return -0.0509
(-0.68)		 -0.058
(-0.773)
Constant 3.218***
(4.43)		 3.417***
(4.397)
Observations 179 179
Adjusted R-squared -0.014 0.008
Durbin-Watson 1.988 2.00
Notes: * p < 0.05, ** p < 0.01, and *** p < 0.001.

Figure 2: Past REIT Returns Do Not Predict Current REIT Returns

a-quadrant-approach-to-commercial-real-estate-investing-private-equity-Fig02

Sources: Morningstar, NAREIT.

Property exhibits serial correlation as discussed in the introduction and indicated by the linear trendline in Figure 3 and the regression analyses. The coefficients on the lagged variables are significant. The past three quarters explains about 64% of the variation in current leveraged property returns. If we compare property and REIT returns, for example, then we must correct for serial correlation.

Figure 3: Past Property Returns Predict Future Returns

a-quadrant-approach-to-commercial-real-estate-investing-private-equity-Fig03

Sources: NCREIF, CBRE Econometric Advisors.

Models 3 and 4: Unleveraged and Leveraged NPI Returns Explained by Lagged Returns

Coefficients Unleveraged NPI Total Return Leveraged NPI Total Return
(T-Statistics)    
1-Quarter Lagged Total Return 0.736***
(9.98)		 0.722***
(9.45)
2-Quarter Lagged Total Return 0.255**
(9.98)		 0.340***
(3.70)
3-Quarter Lagged Total Return -0.225**
(-2.99)		 -0.316***
(-4.07)
Constant 0.488***
(3.26)		 0.588*
(2.53)
Observations 179 1591
Adjusted R-squared 0.622 0.641
Durbin-Watson 1.963 1.967
Notes: * p < 0.05, ** p < 0.01, and *** p < 0.001.

Models 3 and 4: Unleveraged and Leveraged NPI Returns Explained by Lagged Returns

Coefficients Unlevered NPI Total Return Levered NPI Total Return
(T-Statistics)    
1-Quarter Lagged Total Return 0.736***
(9.98)		 0.722***
(9.45)
2-Quarter Lagged Total Return 0.255**
(9.98)		 0.340***
(3.70)
3-Quarter Lagged Total Return -0.225**
(-2.99)		 -0.316***
(-4.07)
Constant 0.488***
(3.26)		 0.588*
(2.53)
Observations 179 1592
Adjusted R-squared 0.622 0.641
Durbin-Watson 1.963 1.967
Notes: * p < 0.05, ** p < 0.01, and *** p < 0.001.

NOI Growth

In the next phase of our exploration, we look at the relationship between property returns and NOI growth. NOI growth affects returns directly through income and appreciation returns, but also indirectly in that expectations for future NOI growth affect yields. This statistically significant but complex relationship between property returns and NOI growth deserves examination. For example, a bivariate regression indicates that NOI growth is a significant factor in explaining unleveraged returns. However, on its own, NOI growth explains only about 6.4% of the variation in returns. Hence, we include other variables in our models to avoid bias in the estimated relationship between growth and returns. The inclusion of these additional variables increases the adjusted R1 from just 6.4% to 83.9%, a substantial increase in the explanatory power of the model. The NOI growth coefficient decreases from 0.33 to 0.09, a reduction in the return’s sensitivity to NOI growth.

1 The difference in the number of observations is because the leveraged series begins in 1983 whereas the unlevered NPI begins in 1978.

Models

Figure 4: Leveraged NPI and Unleveraged NPI Total Returns (%)

a-quadrant-approach-to-commercial-real-estate-investing-private-equity-Fig04

Source: NCREIF.

Our property model is the centerpiece of this report. We model leveraged and unleveraged property total returns. In Figure 4, we report historical total returns for NCREIF’s leveraged and unleveraged NPI index. These indices include many of the same properties, but the unleveraged index reports returns on an unlevered basis. Leveraged and unleveraged returns are highly correlated, but the fit is less precise between 2002 and 2015. The properties included in the leverage index have loan-to-value (LTV) ratios of around 50% to 60%, as shown in Figure 5. The coefficients on equity REIT returns are positive for both leveraged and unleveraged property. REITs are a leading indicator of property performance. However, most of the coefficients in the leveraged model are twice as great as the coefficients in the unleveraged model, indicating increased risk associated with leverage.

NOI growth is a strong determinant of returns. The CMBS BBB variable is not lagged; lagged versions reduced statistical performance. The leveraged property model explains 84% of the variation in the dependent variable and the t-statistics are significant.

Figure 5: NCREIF Loan-to-value Ratio for Leveraged Property

a-quadrant-approach-to-commercial-real-estate-investing-private-equity-Fig05

Sources: NCREIF, CBRE Econometric Advisors.

Models 5 and 6: Unleveraged and Leveraged NPI Returns Explained

Coefficients Unleveraged NPI Total Return Leveraged NPI Total Return
(T-Statistics)    
1-Quarter Lagged REIT Total Return 0.0228*
(2.39)		 0.0518**
(2.99)
2-Quarter Lagged REIT Total Return 0.0187
(1.94)		 0.0607***
(3.46)
3-Quarter Lagged REIT Total Return 0.0146
(1.48)		 0.0505**
(2.78)
4-Quarter Lagged REIT Total Return 0.0185
(1.84)		 0.0460*
(2.43)
BBB CMBS Total Return 0.0930***
(7.60)		 0.150***
(6.69)
NPI NOI Growth 0.0945*
(2.27)		 0.183*
(2.40)
1-Quarter Lagged Total Return 0.779***
(15.18)		 0.729***
(14.46)
Constant 0.0579
(0.45)		 -0.186
(-0.89)
Observations 106 106
Adjusted R-squared 0.839 0.846
Durbin-Watson 2.27 2.44
Mean Variance Inflation Factor Value 1.23 1.24
Notes: * p < 0.05, ** p < 0.01, and *** p < 0.001.

Models 5 and 6: Unleveraged and Leveraged NPI Returns Explained

Coefficients Unlevered NPI Total Return Levered NPI Total Return
(T-Statistics)    
1-Quarter Lagged REIT Total Return 0.0228*
(2.39)		 0.0518**
(2.99)
2-Quarter Lagged REIT Total Return 0.0187
(1.94)		 0.0607***
(3.46)
3-Quarter Lagged REIT Total Return 0.0146
(1.48)		 0.0505**
(2.78)
4-Quarter Lagged REIT Total Return 0.0185
(1.84)		 0.0460*
(2.43)
BBB CMBS Total Return 0.0930***
(7.60)		 0.150***
(6.69)
NPI NOI Growth 0.0945*
(2.27)		 0.183*
(2.40)
1-Quarter Lagged Total Return 0.779***
(15.18)		 0.729***
(14.46)
Constant 0.0579
(0.45)		 -0.186
(-0.89)
Observations 106 106
Adjusted R-squared 0.839 0.846
Durbin-Watson 2.27 2.44
Mean Variance Inflation Factor Value 1.23 1.24
Notes: * p < 0.05, ** p < 0.01, and *** p < 0.001.

Conclusion

In conclusion, REITs, BBB CMBS notes, lagged returns and NOI growth can largely explain property returns. Fundamentals and capital markets drive property returns. Property investors can derive valuable insights from following the other quadrants. In future work, we will summarize the complex relationship between the 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.

Explore Other Viewpoints in This Series

Item 0 Item 0 Item 0 Item 0

Related Services

Related Insights

Contacts

Item 0 Item 1 Item 2 Item 3