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.