Are High Cap Rate Properties Better Investments?

Transcription

1 Are High Cap Rate Better Investments? Eli Beracha Florida International University David Downs Virginia Commonwealth University Greg MacKinnon Pension Real Estate Association The authors thanks the Real Estate Research Institute (RERI) for providing funding for this project, and the National Council of Real Estate Investment Fiduciaries (NCREIF) for allowing us access to the data. NCREIF Conference, March 2017

2 Value Investing Interest in value investing for at least 100 years (back to Benjamin Graham, probably longer) Value Investing = Some version of buying assets with low relative prices Extensive academic research on value effect in equity markets going back to at least the 1980s More recently: Asness, Moskowitz, Pedersen (2013) ( Value and Momentum Everywhere ) find evidence of a value effect across several asset classes and geographies However their everywhere does not include real estate

3 Value and Real Estate In real estate, value can be related to the cap rate High cap rate = value Plazzi, Torous, Valkanov (RFS, 2010) higher cap rates predict higher returns (except for office) Analysis at property type/metro index level Downs and Beracha (JPM, 2015) - high cap rate metros outperform low cap rate metros Analysis at the MSA (metro) level (also done by property type) Useful for investors deciding on general strategy, e.g. which markets to target But, actual investment decisions in CRE made at an asset level each asset unique

4 Value and Real Estate Property level data: Plazzi, Torous, Valkanov (JPM, 2011) optimal portfolios should tilt towards high cap Property level data: Peng (WP, 2016) higher acquisition cap rates predict higher future returns over holding period We use NCREIF property-level data to classify individual as value (or not) and then examine performance of aggregated portfolios Does a typical high cap rate property outperform a typical low cap rate property?

5 What is Value in Real Estate? In real estate, value generally thought of as high cap rate We use trailing one-year income return as proxy But, what is a high (or low ) cap rate? Depends on time period (market conditions) Depends on location (metro) Depends on sector (property type)

6 Data Data from NCREIF, property level, 1Q1978 to 4Q2015 All observations from NPI 481,052 property-quarter observations 23,981 different Data appears fairly clean Likely due to only NPI Do not apply filters for data errors Use medians to control for outliers

7 Data We do apply some filters for number of observations: Each property must have sufficient data to classify as high/mid/low cap rate and to calculate subsequent return (1 year trailing performance and 5 years forward performance) CBSA must have at least 20 consecutive quarters of data Each quarter-cbsa must have at least 5

8 Data Do everything separately for each property type. Final sample: Office: 1264 ; 22,894 property-quarters; start in Q Apartment: 804 ; 13,900 property-quarters; start in Q Retail: 617 ; 10,948 property-quarters; start in Q Industrial: 2093 ; 41,123 property-quarters; start in Q2 1979

9 Overview of Process Each quarter, calculate trailing one-year income return for each property in database (= cap rate) In each quarter in each metro, take median cap rate across market-adjust cap rate of each property each quarter Property Cap Rate Median Cap Rate in CBSA to the extent that growth prospects, risk, liquidity are metro determined, controls for these

10 Overview of Process Across all nationally, each quarter categorize into: highest 30%/middle 40%/lowest 30% by marketadjusted cap rate Note: forming a national portfolio, but each property judged based on its local (metro) median cap rate at that time Each quarter, form portfolios, skip one quarter, and then measure total returns over following five years also look at income and appreciation returns separately

11 Overview of Process Measure cap rate, mkt adjust, form portfolios Observe total returns on portfolios t-4 t-1 t t+1 t+20 Repeat this whole process (categorize, measure 5 year returns going forward) every quarter With 5-year horizons, have starting points each quarter from 2Q1979 to 1Q2011

12 Results Return and Risk Exhibit 1: Mean annualized total returns by cap rate strategy High Cap Rate Mid Cap Rate Low Cap Rate t-stat for diff. btw. high/low Office 5.48% 5.16% 4.73% 3.27*** Apartment 11.07% 9.94% 8.95% 10.93*** Retail 11.09% 10.90% 9.27% 4.59*** Industrial 9.76% 9.14% 8.20% 13.16*** Exhibit 4: Standard deviation of annualized returns by cap rate strategy High Cap Rate Mid Cap Rate Low Cap Rate F-stat for diff. btw. high/low Office 7.24% 6.80% 6.98% 1.08 Apartment 4.11% 3.53% 3.55% 1.34 Retail 5.69% 5.40% 5.16% 1.21 Industrial 4.87% 5.10% 4.70% 1.07

13 Results Source of Performance Exhibit 2: Mean annualized income returns by cap rate strategy High Cap Rate Mid Cap Rate Low Cap Rate t-stat for diff. btw. high/low Office 8.18% 7.70% 6.88% 12.31*** Apartment 7.63% 7.28% 6.89% 10.64*** Retail 8.64% 7.89% 7.16% 17.56*** Industrial 8.67% 8.19% 7.72% 22.28*** Exhibit 3: Mean annualized appreciation returns by cap rate strategy High Cap Rate Mid Cap Rate Low Cap Rate t-stat for diff. btw. high/low Office -2.60% -2.42% -2.21% -2.32** Apartment 3.18% 2.51% 1.92% 8.75*** Retail 2.22% 2.61% 1.94% 0.75 Industrial 0.90% 0.73% 0.40% 4.85***

14 Results Office Performance Overtime 25% 20% 15% 10% 5% 0% -5% % high cap rate low cap rate

15 Results Apartments Performance Overtime 20% 15% 10% 5% 0% -5% high cap rate low cap rate

16 Results Retail Performance Overtime 35% 30% 25% 20% 15% 10% 5% 0% -5% high cap rate low cap rate

17 Results Industrial Performance Overtime 20% 15% 10% 5% 0% -5% high cap rate low cap rate

18 Results Performance Range Exhibit 14: Mean of total return percentiles High Cap Rate Mid Cap Rate Low Cap Rate Office: 75 th percentile 9.64% 9.85% 9.28% Median 5.48% 5.16% 4.73% 25 th percentile 0.36% -0.11% -0.18% Apartments: 75 th percentile 13.98% 12.46% 11.18% Median 11.07% 9.94% 8.95% 25 th percentile 8.30% 7.38% 6.48% Retail: 75 th percentile 13.82% 13.48% 11.95% Median 11.09% 10.90% 9.27% 25 th percentile 8.05% 7.86% 6.17% Industrial: 75 th percentile 13.76% 12.91% 11.92% Median 9.76% 9.14% 8.20% 25 th percentile 5.79% 5.51% 4.52%

19 Results City vs. Property Exhibit 17: Office mean total returns, Interaction between median metro cap rate and property cap rate relative to metro High relative cap rate Mid relative cap rate Low relative cap rate High cap rate CBSAs Mid cap rate CBSAs Low cap rate CBSAs 6.22% 5.25% 4.46% 5.99% 4.03% 3.72% 5.30% 4.11% 3.15% Exhibit 18: Apartment mean total returns, Interaction between median metro cap rate and property cap rate relative to metro High relative cap rate Mid relative cap rate Low relative cap rate High cap rate CBSAs Mid cap rate CBSAs Low cap rate CBSAs 12.01% 11.46% 10.99% 10.67% 10.22% 9.72% 9.66% 9.06% 9.00%

20 Results City vs. Property Exhibit 19: Retail mean total returns, Interaction between median metro cap rate and property cap rate relative to metro High relative cap rate Mid relative cap rate Low relative cap rate High cap rate CBSAs Mid cap rate CBSAs Low cap rate CBSAs 10.64% 9.44% 9.77% 10.30% 9.63% 9.77% 9.56% 8.19% 7.41% Exhibit 20: Industrial mean total returns, Interaction between median metro cap rate and property cap rate relative to metro High relative cap rate Mid relative cap rate Low relative cap rate High cap rate CBSAs Mid cap rate CBSAs Low cap rate CBSAs 9.77% 9.88% 10.04% 9.10% 9.22% 9.25% 7.85% 8.55% 7.92%

21 Conclusion Value (high CAP) outperforms growth on a raw and risk adjusted basis. Outperformance is statistically significant & economically meaningful. The outperformance is robust: - Over time - Across property types - Across and within cities - Throughout the value and growth performance range.