During times of market uncertainty — such as a financial crisis, pandemic or spike in interest rates — the illiquid nature of private equity real estate (PERE) leads to questionable valuations. Appraisals rely on recent transactions of comparable properties to estimate the most probable selling price for a property. But commercial property transactions typically drop significantly following market shocks, making appraisals more challenging. Today real estate professionals are discussing if appraisals appropriately reflect the impact of COVID and recent interest rate spikes due to limited transactions and a wide bid-ask spread. Certainly, lower income growth expectations (or even declines) combined with higher borrowing and discount rates should result in compound value reduction. But this change in market conditions has been slow to impact appraised values.

Some experts suggest listed REITs serve as a direct proxy for property valuations. Others counter that REIT security prices are too volatile and incorporate many non-property features, such as business entities, development functions and expansion expectations. Despite these concerns, equity REIT securities are supported by real estate assets and are continuously priced by the market, yielding a return in virtually every environment.

The same can be said about real estate debt securities. The delinquency rates of commercial mortgage-backed securities (CMBS) offer insights into the health of the commercial real estate sector because they rise when property performance fails to meet financing obligations. Compared with other sources of debt, CMBS loans typically have stricter covenants in their contracts, which can result in quicker delinquencies during a market downturn. The volume of CMBS issuance serves as a barometer of market sentiment, reflecting the appetite for commercial real estate financing. In times of economic uncertainty, a decline in CMBS origination volume may indicate a cautious approach by lenders, constraining access to capital and potentially signaling dampening PERE returns. Conversely, surges in origination volume during periods of economic expansion may suggest heightened investor optimism and increased transaction activity, presenting opportunities for PERE investors.

Potential leading indicators

Using regression analysis, we measured the information in equity REIT returns that can be used to predict tomorrow’s property returns (valuations). Then we explored real estate debt securities as possible prediction information providers. Both equity and debt securities are continuously priced, and both are overlays on property assets. In total, we considered the PERE return prediction power offered by three data series: equity REIT returns, CMBS issuance volume and CMBS delinquency rate.

Our data comprised 24 years of quarterly data for PERE, REIT and CMBS issuance, and 17 years of CMBS delinquency rates. We took the natural log of the CMBS volume variables to normalize the volatility and simplify interpretation of results. The contemporaneous performance of the four variables, with PERE being the subject variable, is shown in Figure 1. As expected, the REIT series is far more volatile than the PERE series. It is the rapid response of market pricing that we hope to exploit. Natural logarithm — Ln(issuance) and Ln(delinquency) — variables offer both a continuously positive and continuously negative series but neither seems too closely related to the PERE series, except for issuance and PERE during the financial crisis. Following this visual inspection, the correlation matrix in Figure 2 reveals low concurrent movement between the variables with two exceptions. There is a modest correlation between PERE returns and CMBS issuance of 0.33 (recalling that 1.00 is perfect correlation that a variable has with itself). CMBS issuance and CMBS delinquency are significantly correlated at (0.63) indicating that they typically move in opposite directions. The other relationships are rather uncorrelated with coefficients below 0.20.

Sources: Bloomberg, Green Street Advisors, NCREIF and NAREIT as of May 2024

The last step in analyzing the concurrent relationship between PERE returns and our independent variables is a series of regressions measuring each independent variable’s relationship with PERE returns in the same quarter. Three statistics from each model will establish the strength of the relationship between variables: adjusted R-squared, coefficient, and p-value. In the first model listed in Figure 3, PERE returns were regressed on REIT returns and the model produced an adjusted R-squared (Adj R2) of 0.3% indicating REIT returns explain less than 1% of the variation in same-period PERE returns. The coefficient appears positive but according to the p-value is insignificant. The p-value indicates the probability that the coefficient might be zero based on the error terms. When this value is 5% or less the coefficient is significantly different than zero.

The second model tests the relationship between CMBS issuance and current PERE returns with a result that is positive and significant. Issuance volume can explain 17% of PERE return volatility using a coefficient of 0.0151 that has a zero probability of being zero — in other words it is very significant. The third model, employing CMBS delinquency, yielded the expected negative sign (delinquency goes up when the environment is bad, so property returns go in the opposite direction), but the explanatory power (Adj R2) is very low and the p-value of 95% indicates that the coefficient is likely zero. There is no information in this model. Of the three indicators, only CBMS issuance offers any insight into current property returns.

Upcoming expectations

By shifting the quarterly data series (lagging the independent variables) the relationships between each indicator and future PERE returns become a measure of predictability. The results shown in Figure 4 indicate that REIT returns are more informative about the next quarter’s PERE returns than concurrent returns. The model suggests there is some predictive power in equity REIT returns that explain 4.2% of PERE return variation with a coefficient that is significantly different than zero at a 3% level. We are rather confident that lagged REIT returns exhibit a significant positive relationship, and that relationship explains a small amount of the PERE return changes.

The lagged debt securities perform similarly to the contemporaneous debt performance. CMBS issuance explains 14% of next quarter’s PERE return with a zero p-value indicating extreme significance in the estimation. We are very confident in this statistical result. CMBS delinquency by itself is as uninformative about future PERE performance as it is about current performance. The coefficient is not significantly different than zero and R-squared is very low. This might be a surprising result since CMBS delinquency rates are intimately intertwined with market conditions and offer insights into the health of the commercial real estate sector. Also, CMBS loans are transferred to a special servicer immediately upon delinquency forcing quick response to weak conditions. It is reasonable to expect that market conditions driving the quick response in CMBS delinquency rates would also shape the investment landscape for PERE. Yet the data tell a different story.

Continuing the search for predictability, Figure 5 offers three models with pairs of independent variables regressed on PERE returns. We combined REIT returns with CMBS issuance, which produced a model that explains 19% of PERE returns with both coefficients being highly significant. This combination offers more information about upcoming property performance than any of the single variable models. Adding delinquency to the REIT model had very little effect on explanatory power (Adj R2) or significance (p-value). At this point, it seems that delinquency data continues to have little to offer. However, the combination of issuance and delinquency produced a model that offers the best predictive power yet. In this model, the combination of variables explains 24.6% of PERE return variability and both coefficients are highly significant. This result is likely linked to the high negative correlation between issuance and delinquency shown in Figure 2. Given this correlation, it is difficult to interpret the coefficients, but the overall model performance is valid.

Putting it all together

In the final step we combined all three indicators in a multiple regression model. These results are shown in Figure 6. Using all three indicators, the model explains 24.8% of PERE return volatility which is slightly greater than the combination of debt securities. This model suggests that a 1% change in REIT returns has a 6 basis-point impact on PERE returns, issuance has a 2 basis-point impact and delinquency has a 1 basis-point impact. However, the REIT returns come with greater variance as indicated by the slightly elevated p-value. The CMBS variables are more significant.

Our exploration into predicting private equity real estate returns uncovered varying results. The most highly expected predictability variable (REIT returns) proved to offer some insight into the single variable lagged model. From our analysis, real estate debt securities proved more capable of predicting property performance than equity REITs. CMBS delinquency data failed to yield the significant predictive power that we expected, but promising results from CMBS origination data demonstrated significance across most models. Across all models we evaluated, the greatest portion of PERE return variance explained by lagged securities data was less than 25%. There is some information in the traded real estate market about the illiquid property market, but it does not tell the whole story.

If investors are looking to predict PERE performance and anticipate appraised value changes, then CMBS origination data would be a useful tool to have in their tool belt. Its demonstrated predictive power across various model specifications underscores its value in informing investment decisions within the private real estate market. Simply put, equity REIT returns might provide some insight to illiquid property value direction, but CMBS performance is at least as powerful in this endeavor.