private equity real estate risk the good, the bad, and the ugly
TRANSCRIPT
www.northinfo.com
Richard Gold
May 13, 2014
Private Equity Real Estate Risk The Good, the Bad, and the Ugly
www.northinfo.com
Goals for Today’s Presentation
• What is the problem facing real estate investors? • What are the current solutions that most investors are employing today?
• Fixed “artificial allocation” • “My consultant will tell me!” • Smoothing appraisal-based indices
• What are the drawbacks to these solutions? • What is Northfield’s solution? • Applications for risk management
• Teaser for tomorrow’s Webinars
• Literature review is at the end of the presentation
www.northinfo.com
What is the Problem?
• There is no arm’s length price data within long periods of time for illiquid assets; interim appraisals are required
• This clashes with basic principles of MPT and arbitrage-free pricing • Appraisal bias leads to serial correlation
• Serial correlation is correctable but creates other challenges
• “Location, Location” as a factor does not integrate well with the classical equity or fixed income factors such as value/growth, size, yield curve factors, etc. • Fama-French” is not common jargon in real estate investment departments • Unlisted investment experts tend to think of risk in terms of first distributional
moment, i.e. return, not volatility
• Therefore, given this intractable problem how do you determine how much real estate to include in your portfolio?
www.northinfo.com
Public Versus Private Property Returns
-20%
-15%
-10%
-5%
0%
5%
10%
15%
20%
25%
1994:1 2000:3 2007:1 2013:3
QUA
RTER
LY R
ETU
RN
NCREIF VERSUS NAREIT 1994Q1 - 2013Q3
NAREIT
NCREIF
Source: NCREIF & NAREIT
www.northinfo.com
Appraisal Bias Remains an Issue
.0 .1 .2 .3 .4 .5 .6 .7 .8 .9
LAG 1
LAG 2
LAG 3
LAG 4
LAG 5
Return Persistence Lagged NCREIF Correlations
NCREIF: 00Q1 TO 11Q3
NCREIF: 78Q1 - 11Q3
2000Q1-2011Q3 1Q Lagged Correlations NCREIF = .85 NAREIT = .19
www.northinfo.com
The “Global” Total Portfolio Risk Problem
• Multiple portfolios with diverse characteristics
• Across countries, across asset classes
• Asset classes such as “direct-owned” property and infrastructure investments have no visible pricing, return, or risk information
www.northinfo.com
Segregated Solutions
• One approach: model asset class separately and aggregate risk only through covariance matrix:
• Overall covariance matrix resembles a chessboard where each square is a sub-model segment of the covariance matrix
• As aggregate number of factors increases relative to the limited number
of observations, the matrix becomes unstable
• Fixing the problem by extending the historic observation period of the sub-models discounts the importance of the dynamics in the market place embedded in more recent observations
www.northinfo.com
Illiquid Assets Compound the Total Portfolio Problem
• In the absence of reliable risk metrics, practitioners default to a naïve global geographical diversification scheme - “distance is diversification!”
• Current real estate modeling practice produces yet another silo factor set in the total portfolio covariance matrix
• Current models de-smooth appraisal indices to arrive at “actual” asset returns, then apply MPT techniques to these derivative time series within a risk model: • This extra layer of estimation error compounds across the modeled global markets and
exponentially decreases effective confidence of the risk metrics from 90% in one market to just 35% in ten markets
• Unlike individual security econometric estimation errors which diversify, factors errors compound across the portfolio
www.northinfo.com
Fully Integrated Approach
• Employ a single parsimonious factor model across all asset classes across all geographies
• All investable assets are related to the same consistent set of factors, so interrelationships are easily observed and understood
• Limited number of factors allows for stable estimation of factor relationships, and fluid regime shifts
www.northinfo.com
Fully Integrated Approach (con’t)
This solution addresses:
• Appraisal bias • Volatility and pricing levels
• Structural factor differences between liquid and illiquid assets • Covariances automatically are corrected as a result
www.northinfo.com
What is the solution?
Factor Model: Generic Factors
• Regions • Sectors • Investor Confidence • Currencies • Energy Costs • Transient (Blind) Factors • Interest Rates / Yield Curve
www.northinfo.com
The Solution (con’t)
How does a factor model apply to Illiquid assets? We look at an illiquid asset as a “composite asset” : • Risks based on “steady-state” cash flow assumptions for existing and
expected leases • Uses lease structure, renewal, credit quality of tenants, vacancy dynamics, revenue
and expense schedules
• Risks related to mortgage financing (if any) • Takes into consideration floating rate, fixed rate, interest-only, balloon clauses,
prepayment behavior, etc.
• Risks of future fluctuations in market rents • Takes into consideration the combined impact of lease rollover, vacancy, renewal,
and market volatility of rents
Each of the 3 components has risk exposures to common risk factors plus idiosyncratic risks
www.northinfo.com
Real Estate Property Risk Approach
• Main Inputs To Model: Current Operating Income and Expenses
Current and structural vacancy Renewal rate and down-time between leases
Growth of rent and expense cash flows over time
Useful life of building as well as Leases Contract Length
Main / Anchor tenants Rent time series
Financing terms – rate, floating fixed, balloon, amortization, etc.
www.northinfo.com
Real Estate Model Structure MORTGAGE FINANCING
(SHORT)
STEADY STATE CASH FLOW
(LONG)
RENT VOLATILITY
TIME VALUE OF
MONEY
CREDIT RISK
CHANGE IN RENT
GLOBAL RISK MODEL
PROPERTY/ PORTFOLIO RISK
RISK FACTORS RISK FACTORS RISK FACTORS
www.northinfo.com
Components of Portfolio Risk
Rent Risk Interest Rate Risk
Credit Risk
TOTAL RISK
Idiosyncratic Risk Still Present But
Largely Diversified Away as Portfolio Size Increases
www.northinfo.com
Model Results – A Sample Portfolio: 15% Gearing
PORTFOLIO PROFILE Metro Apartment Office Industrial Retail
Berlin 1
Budapest 1
Frankfurt 1
Rome 2
London 1 2 1
Marseilles 1
Amsterdam 1
Paris 1
Bucharest 1
Stockholm 1
Slide 16
www.northinfo.com
Model Results – Property Risk by Source
Total Risk
15.9%
Apartment Building – Romania (No Leverage)
Interest Rate Risk Rent Risk Credit Risk
13.7% 1.2% 6.9%
Slide 17
www.northinfo.com
Model Results: Output
Single Property: Bucharest Apartments (Leverage Removed)
Factor PortExp BenchExp ActiveExp FactorVar VarContr ENERGY MINERAL SECTOR 0.00 0 0.00 235.60 0.00 INDUSTRIAL SECTOR 0.00 0 0.00 247.12 0.01 TECHNOLOGY&HEALTH SECTOR 0.00 0 0.00 161.43 0.05 INTEREST RATE SENSITIVE SECTR 0.00 0 0.00 204.85 0.06 NON-ENERGY MINERALS 0.00 0 0.00 423.90 0.06 DEVELOPING MARKET 0.01 0 0.01 115.72 0.13 VALUE/GROWTH -0.02 0 -0.02 6.47 0.13 CONSUMER SECTOR 0.00 0 0.00 134.21 0.29 CONTINENTAL EUROPE 0.28 0 0.28 208.25 25.97 TREASURY CURVE FACTOR1 -17.88 0 -17.88 0.30 121.76 TREASURY CURVE FACTOR2 -221.58 0 -221.58 0.00 145.07 TREASURY CURVE FACTOR3 -1590.87 0 -1590.87 0.00 -79.82 Factor Tracking Variance 212.94 Stock Specific Tracking Variance 39.97 Total Tracking Variance 252.92 Tracking Error 15.90 Total Risk of Portfolio 15.90 Total Risk of Benchmark 0.00 R-Squared 0.00
Slide 18
www.northinfo.com
Model Results – Geospatial Risk
Country Risk (All Leverage Removed)
Country Risk Property Count Land Use
Romania 15.9 1 Residential
Sweden 16.4 1 Retail
Germany 17.6 2 Residential, Office
France 19.8 2 Retail, Office
Netherlands 19.9 1 Industrial
U.K. 20.3 4 Office(2), Industrial, Residential
Italy 20.4 2 Industrial(2)
Hungary 20.7 1 Retail
Slide 19
www.northinfo.com
Model Results –Risk by Property Type
15% 16% 17% 18% 19% 20%
Residential
Retail
Office
Industrial
All Leverage Removed
Slide 20
www.northinfo.com
INCREMENTAL TOTAL RISK
Model Results: Incremental Risk with Additional Acquisitions
15
16
17
18
19
20
21
22
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Unlevered Levered
0
1
2
3
4
5
6
7
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Levered
Unlevered
INCREMENTAL SPECIFIC RISK
Slide 21
www.northinfo.com
Effects of Leverage on Equity Risk – Single Property
0
20
40
60
80
100
120
140
160
0.2 0.4 0.6 0.95 1.5
Stan
dard
Dev
iatio
n
NYC Office Building: Risk & LTV
Debt as a Percent of Equity
For every percent increase in leverage over 60%, risk increases by over 3.2% until 95% LTV.
Leverage > 95% results in little chance of covering debt resulting in lower risk: no equity to loose and insufficient cash flow just waiting to hand back keys.
www.northinfo.com
Reality Check How to perform a check on the reasonableness of results when there is not objective way to measure individual property risk results any time between buy and sell points, which is a long time: • One general way is to compare with other asset classes and see if the
ranking corresponds to our intuition.
• Another way is to de-smooth real estate appraisal-based indexes using a time horizon of a typical property holding period and compare standard deviations
• A third way is to use real estate expert qualitative risk rankings and compare to model predicted risk rankings
• The acid test: does the methodology appeal to economic intuition?
www.northinfo.com
Reality Check (con’t):
12 Month Ex Ante as of November 2013 per Northfield’s Everything Everywhere Model
0%
5%
10%
15%
20%
25%
Granular RE Port Generic RE Port S&P500 Barclay's Bond Index
Expected Annualized Risk by Asset Class – November 2013
Last 5 Years Ex Ante Risk
www.northinfo.com
Reality Check (con’t) Historical Standard Deviation - Last 20 Quarters Ending November, 2013
7.8%
11.7%
11.4%
15.3%
17.4%
14.1%
21.3%
20.7%
26.5%
30.2%
NCREIF
NCREIF OPEN-END FUNDS
TOWNSEND CORE
TOWNSEND VALUE ADDED
TOWNSEND OPPORTUNISTIC
ADJUSTED
UNADJUSTED
Source: NCREIF
No Leverage
~ 24% Leverage
www.northinfo.com
Reality Check (con’t)
17.53
13.52
13.46
10.47
10.57
11.86
11.12
11.8
11.5
12.44
Sample Property 1
Sample Property 2
Sample Property 3
Sample Property 4
Sample Property 5
Sample Property 6
Sample Property 7
Sample Property 8
Sample Property 9
Average
Annual Model-Predicted Standard Deviation
www.northinfo.com
Applications in Risk Management
Lessons learned from the experience of major North American pension fund using this type of model approach: – The results were immediately useful to top level management by identifying
economic exposure to interest rates, inflation, sectors, resources, geographies, and credit
– The riskiness rankings produced by this technique were reinforced by a blind study of internal expert property risk rankings where such opinions existed. Where no such opinions existed the model became the de factor “expert” providing insight to regions, markets, and properties where such expert opinions could not be formed on a qualitative basis alone.
www.northinfo.com
Applications in Risk Management (con’t)
– The good news is that the risk management flexibility provided by Northfield’s granular private equity property model and its factor model methodologies provides users with unprecedented flexibility in designing tailored risk analysis.
– The bad news is that you will have to wait until tomorrow’s Webinar to hear the details.
www.northinfo.com
Summary The Methodology:
• Provides integrated and consistent risk measurement and management capability, comparable to other asset classes
• Poses manageable data requirements, but yet does not cut corners with respect to granularity, locality, and other difficult aspects of the analysis
• Features a fundamental, bottom-up approach which integrates intuitively the economic forces that drives returns for private equity real estate
• Uses broad risk drivers that allow identification of the types of risks that can be hedged using liquid market instruments (e.g. discount rate risk = interest rate risk)
• Is fully flexible with regards to risk platform use
www.northinfo.com
The Ugly, the Bad, & the Good The Ugly:
• Traditional appraisal-based indices fail as risk metrics significantly underreporting volatility due to serial correlation
The Bad: • Index-based approaches with minimal appetite for fundamental
property-level data at first glance appear to be a solution
• The statistical confidence of the aggregate model declines dramatically when correcting for serial correlation. It does not diversity away!
• When you have only real estate factors such as regions and property types you have limited options. Not integrated with other asset classes and therefore operate in silos.
The Good: • Treat a property as a composite asset with “steady state” cash
flows, rent changes, and financing risks.
• Fully integrated, bottom-up approach using a global factor model using a parsimonious set of risk factors.
www.northinfo.com
References • DiBartolomeo, D., Gold, R., Baldwin, K., Belev, E. “A New Approach to Real Estate Risk.” Northfield Working Papers .
http://www.northinfo.com/documents/191.pdf. 2005
• Belev, E., Gold, R.. “The Pitfalls of an Index-Based Approach to Managing Real Estate Investment Risk”. Northfield Working Papers . http://www.northinfo.com/Documents/589.pdf. 2014
• Geltner, D. and B. Kluger, “REIT-Based Pure-Play Portfolios: The Case of Property Types”, Real Estate Economics, 1998, 26:4, pp. 581-612.
• Giliberto, S. M., ” Measuring Real Estate Returns: The Hedged REIT Index”, Journal of Portfolio Management”, 1993.
• Giliberto, S. Michael “Equity Real Estate Investment Trusts and Real Estate Returns”, Journal of Real Estate Research”, 1990, 5, pp. 259-264.
• Kim, E., “REIT-Based Pure-Pure Play Portfolios: The Case of Property Types and Geographic Locations”, MIT Master of Science in Real Estate Development Thesis, 2004.
• Horrigan Holly, Case Brad, Geltner, David, and Pollakowski, Henry, “REIT-Based Property Return Indices: A New Way to Track and Trade Commercial Real Estate”, The Journal of Portfolio Management, 2009.
• Lee, Stephen, “Is There a ‘Case for Property’ All the Time?”, Presented at the 9th Annual European Real Estate Society (ERES) Meeting, June 2002.
• Ross, Stephen A., “The Arbitrage Theory of Capital Asset Pricing”, Journal of Economic Theory, 1976, 13, pp. 341-360.
• Sharpe, W., “Capital asset prices: A theory of market equilibrium under conditions of risk”, Journal of Finance, September 1964.