Risk Analysis and Project Evaluation

Risk Analysis and Risk Analysis and Project EvaluationProject Evaluation

Campbell R. HarveyDuke University

andNational Bureau of Economic Research

Project Appraisal and Risk Management (PARM)Duke Center for International Development at the Sanford Institute May 27-28, 2002

1. Cash Flow versus Discount Rate2. Approaches to Cost of Capital Measurement3. Recommended Framework4. Comparison of Methods5. Conversion of Cash Flows 6. Project Specific Adjustments7. Conclusions

Risk Analysis and Project EvaluationRisk Analysis and Project EvaluationPlan

Basic Project Evaluation:• Forecast nominal cash flows• Currency choice (assume US$)• Decide what risks will be reflected in cash

flows and those in the discount rate– Beware of double discounting

Risk Analysis and Project EvaluationRisk Analysis and Project Evaluation1. Cash Flow vs. Discount Rate

Simple example:• Assume a simple project with expected

$100 in perpetual cash flows• If located in the U.S., the discount rate

would be 10% and Value= $100/0.10= $1,000


Simple example:• However, project is not located in the U.S.

but a risky country• If we reflect the country risk in the discount

rate, the rate rises to 20% Value = $100/0.20 = $500


Simple example:• If we reflect the country risk in the cash

flows, the value is identical Value = $50/0.10 = $500


Our approach• We will propose methods that deliver

discount rates that reflect country risk.• As our example showed, it is a simple

matter of shifting the country risk from the discount rate to the cash flows.


Our approach• Indeed, we will often do this.

– That is, we will use quantitative methods to get a measurement of country risk in the discount rate.

– Use the country risk adjustment in the cash flows (and adjust discount rate down accordingly).

– Use Monte Carlo methods on cash flows rather than cash flows and discount rate.


Many different approaches:1. Identical Cost of Capital (all locations)2. World CAPM or Multifactor Model (Sharpe-

Ross)3. Segmented/Integrated (Bekaert-Harvey)4. Bayesian (Ibbotson Associates)5. Country Risk Rating (Erb-Harvey-Viskanta)6. CAPM with Skewness (Harvey-Siddique)

Risk Analysis and Project EvaluationRisk Analysis and Project Evaluation2. International Cost of Capital

Risk Analysis and Project EvaluationRisk Analysis and Project Evaluation 2. International Cost of Capital

7. Goldman-integrated sovereign yield spread model

8. Goldman-segmented9. Goldman-EHV hybrid10. CSFB volatility ratio model11. CSFB-EHV hybrid12. Damoradan

Identical Cost of Capital

• Ignores the fact that shareholders require different expected returns for different risks


Identical Cost of Capital

• Risky investments get evaluated with too low of a discount rate (and look better than they should)

• Less risky investments get evaluated with too high of a discount rate (and look worse than they are)

• Hence, method destroys valueAvoid


World CAPM• Sharpe’s Capital Asset Pricing Model is the

mainstay of economic valuation• Simple formula• Intuition is that required rate of return depends on

how the investment contributes to the volatility of a well diversified portfolio

Risk Analysis and Project Evaluation Risk Analysis and Project Evaluation 2. International Cost of Capital

World CAPM• Expected discount rate (in U.S. dollars) on

investment that has average in a country = riskfree + x world risk premium• Beta is measured relative to a “world” portfolio• OK for developed markets if we allow risk to

change through time (Harvey 1991)


World CAPM

• Strong assumptions needed• Perfect market integration• Mean-variance analysis implied by utility

assumptions• Fails in emerging markets


Returns and Beta from 1970

R2 = 0.013

-0.1

0

0.1

0.2

0.3

0.4

0.5

-0.5 0 0.5 1 1.5 2 2.5 3

Beta

Ave

rage

ret

urns

Should be a positive relation, with higher risk associated with higher return!But perhaps we should look at a more recent sample of data.


Returns and Beta from 1990

R2 = 0.0211

-0.1

0

0.1

0.2

0.3

0.4

0.5

-0.5 0 0.5 1 1.5 2 2.5 3

Beta

Ave

rage

ret

urns

Still goes the wrong way - even with data from 1990!


World CAPM

• OK to use in developed markets• May give unreliable results in smaller, less liquid

developed markets


Segmented/Integrated CAPM

• CAPM assumes that markets are perfectly integrated– foreign investors can freely invest in the local market– local investors can freely invest outside the local market

• Many markets are not integrated so we need to modify the CAPM



• Bekaert and Harvey (1995)• If market integrated, world CAPM holds• If market segmented, local CAPM holds• If going through the process of integration, a

combination of two holds



Estimate world beta and expected return= riskfree + w x world risk premium

Estimate local beta and expected return= local riskfree + L x local risk premium



• Put everything in common currency terms• Add up the two components.

CC= w[world CC] + (1-w)[local CC]• Weights, w, determined by variables that proxy for

degree of integration, like size of trade sector and equity market capitalization to GDP



• Weights are dynamic, as are the risk loadings and the risk premiums

• Downside: hard to implement; only appropriate for countries with equity markets

• Recommendation: Wait


Ibbotson Associates(Recognized expert in cost of capital calculation)

• Approach recognizes that the world CAPM is not the best model

• Ibbotson approach combines the CAPM’s prediction with naïve prediction based on past performance.


Ibbotson Associates• STEPS1 Calculate world risk premium=U.S. risk premium

divided by the beta versus the MSCI world 2 Estimate country beta versus world index3 Multiply this beta times world risk premium


Ibbotson Associates

4 Add in 0.5 times the ‘intercept’ from the initial regression. “This additional premium represents the compensation an investor receives for taking on the considerable risks of the emerging markets that is not explained by beta alone.”


Ibbotson Associates

• Gives unreasonable results in some countries• Only useful if equity markets exist• Ibbotson Associates does not even use itRecommendation: Do not use this version.

Ibbotson has alternative methods available.


CAPM with Skewness

• For years, economists did not understand why people spend money on lottery tickets and horse betting

• The expected return is negative and the volatility is high

• Behavioral explanations focused on “risk loving”


CAPM with Skewness

• But this is just preference for positive skewness (big positive outcomes)

• People like positive skewness and dislike negative skewness (downside)


CAPM with Skewness

• Most are willing to pay extra for an investment that adds positive skewness (lower hurdle rate), e.g. investing in a startup with unproven technology


CAPM with Skewness

• Harvey and Siddique (2000) tests of a model that includes time-varying skewness risk

• Bekaert, Erb, Harvey and Viskanta detail the implications of skewness and kurtosis in emerging market stock selection


CAPM with Skewness

• Model still being developed• Skewness similar to many “real options” that are

important in project evaluationRecommendation: Wait


Goldman-Integrated*

• This model is widely used by McKinsey, Salomon and many others.

• Addresses the problem that the CAPM gives a discount rate too low.

• Solution: Add the sovereign yield spread


*J.O. Mariscal and R. M. Lee, The valuation of Mexican Stocks: An extension of the capitalasset pricing model to emerging markets, Goldman Sachs, June 18, 1993.

Goldman-Integrated

• The sovereign yield spread is the yield on a U.S. dollar bond that a country offers versus a U.S. Treasury bond of the same maturity

• The spread is said to reflect “country risk”


Goldman-Integrated

STEPS• Estimate market beta on the S&P 500• Beta times historical US premium • Add sovereign yield spread plus the risk free


Goldman-Integrated-EHV Hybrid

• Goldman model only useful if you have sovereign yield spread

• Use Erb, Harvey and Viskanta model to fit ratings on yield spread


Real Yields and Institutional Investor Country Credit Ratings from 1990 through 1998:03

R2 = 0.8784

0.00%2.00%4.00%6.00%8.00%

10.00%12.00%14.00%

0 20 40 60 80 100

Rating

Rea

l Yie

lds



• You just need a credit rating (available for 136 countries now) and the EHV model will deliver the sovereign yield



• Even adding this yield spread delivers a cost of capital that is unreasonably low in many countries

• While you can get the yield spread in 136 countries with the EHV method, you can only get risk premiums for those countries with equity markets


Goldman-Segmented

• Main problem is the beta• It is too low for many risky markets• Solution: Increase the beta


Goldman-Segmented

• Modified beta=standard deviation of local market return in US dollars divided by standard deviation of the US market return

• Beta times historical US premium • Add sovereign yield spread


Goldman-Segmented

• Strange formulation. The usual beta is:

• Using volatility ratio implies that the Correlation=1 !!

World

iWorldiWorldi devStd

devStdnCorrelatioBeta

..

,,


Goldman-Segmented

• No economic foundation for modification• No clear economic foundation for method in

generalRecommendation: Not recommended


CSFBE[ri]=SYi + i{E[rus-RFus] x Ai} x Ki

• SYi = brady yield (use fitted from EHV)

• i = the beta of a stock against a local index


L. Hauptman and S. Natella, The cost of equity in Latin American, Credit Swisse FirstBoston, May 20, 1997.

CSFBE[ri]=SYi + i{E[rus-RFus] x Ai} x Ki

• Ai =the coefficient of variation (CV) in the local market divided by the CV of the U.S. market) where CV = /mean.

• Ki =“constant term to adjust for the interdependence between the risk-free rate and the equity risk premium”


CSFB

• No economic foundation• Complicated, nonintuitive and ad hocRecommendation: Avoid


Damodaran

• Idea is to adjust the sovereign spread to make it more like an equity premium rather than a bond premium


A. Damodaran, Estimating equity risk premiums, working paper, NYU, undated.

Damodaran

Country Sovereign Equity std. dev. equity = yield x ------------------premium spread Bond std. dev.


Damodaran• Advantage: Recognizes that you just can’t

use the bond yield spread as a plug number in the CAPM

• Disadvantage: Assumes that Sharpe ratios for stocks and bonds must be the same in any particular country.


Country Risk Rating Model

• Erb, Harvey and Viskanta (1995)• Credit rating a good ex ante measure of risk• Impressive fit to data

Risk Analysis and Project Evaluation Risk Analysis and Project Evaluation 3. Recommended Framework

C.B. Erb, C. R. Harvey and T. E. Viskanta, Expected returns and volatility in 135 countries,Journal of Portfolio Management, 1995.

Country Risk Rating Model

• Erb, Harvey and Viskanta (1995)• Explore risk surrogates:

– Political Risk, – Economic Risk, – Financial Risk and – Country Credit Ratings


Country Risk Rating ModelSources• Political Risk Services’ International Country Risk Guide• Institutional Investor’s Country Credit Rating• Euromoney’s Country Credit Rating• Moody’s• S&P


Political risk. International Country Risk Guide% of

Individual % ofPolitical Points Index CompositeEconomic expectations vs. reality 12 12% 6%Economic planning failures 12 12% 6%Political leadership 12 12% 6%External conflict 10 10% 5%Corruption in government 6 6% 3%Military in politics 6 6% 3%Organized religion in politics 6 6% 3%Law and order tradition 6 6% 3%Racial and nationality tensions 6 6% 3%Political terrorism 6 6% 3%Civil war 6 6% 3%Political party development 6 6% 3%Quality of the Bureaucracy 6 6% 3%

Total Political Points 100 100% 50%


Financial risk. International Country Risk Guide

FinancialLoan Default or unfavorable loan restructuring 10 20% 5%Delayed payment of suppliers’ credits 10 20% 5%Repudiation of contracts by governments 10 20% 5%Losses from exchange controls 10 20% 5%Expropriation of private investments 10 20% 5%

Total Financial Points 50 100% 25%


Economic risk. International Country Risk Guide

EconomicInflation 10 20% 5%Debt service as a % of exports of goods and services 10 20% 5%International liquidity ratios 5 10% 3%Foreign trade collection experience 5 10% 3%Current account balance as a % of goods and services 15 30% 8%Parallel foreign exchange rate market indicators 5 10% 3%

Total Economic Points 50 100% 25%

Overall Points 200 100%


International Country Risk Guide Risk Categories

Risk Category Composite Score RangeVery High Risk 0.0-49.5High Risk 50.0-59.5Moderate Risk 60.0-69.5Low Risk 70.0-84.5Very Low Risk 85.0-100.0


Institutional Investor’s Country Credit Ratings

OECD Emerging Rest of World1979 1994 1979 1994 1979 1994

Economic Outlook 1 1 2 3 3 4Debt Service 5 2 1 1 1 1Financial Reserves/CurrentAccount

2 3 4 4 4 3

Fiscal Policy 9 4 9 7 6 6Political Outlook 3 5 3 2 2 2Access to Capital Markets 6 6 7 9 8 9Trade Balance 4 7 5 5 5 5Inflow of Portfolio Investment 7 8 8 8 7 8Foreign Direct Investment 8 9 6 6 9 7


Ratings are correlated:

0102030405060708090

100

Inst

itutio

nal I

nves

tor C

CR

AA

+

AA

AA

-

A+ A A-

BB

B+

BB

B

BB

B-

BB

+

BB

BB

-

B+ B

NR

S&P Sovereign Ratings



0102030405060708090

100

Eur

omon

ey C

CR

AA

+

AA

AA

-

A+ A A-

BB

B+

BB

B

BB

B-

BB

+

BB

BB

-

B+ B

NR




0102030405060708090

100

ICR

G C

ompo

site

AA

+

AA

AA

-

A+ A A-

BB

B+

BB

B

BB

B-

BB

+

BB

BB

-

B+ B

NR




Risk Measure ChangesII CCR ICRGC ICRGP ICRGF ICRGE

II CCR -0.03 0.01 0.03 -0.09ICRGC 0.35 0.79 0.54 0.43ICRGP 0.30 0.83 0.25 0.06ICRGF 0.26 0.60 0.35 0.05ICRGE 0.10 0.52 0.24 0.25

Risk Measure Levels


ICRG ratings predict changes in II ratings:

Attribute Coefficient T-Stat R-SquareICRGC 0.2120 7.59 5.0%ICRGP 0.1244 5.67 2.8%ICRGF 0.0956 5.69 2.8%ICRGE 0.0833 4.65 1.9%


Ratings predict inflation:

00.10.20.30.40.50.60.70.80.9

1

0 20 40 60 80 100

II Rating September 1996

Infla

tion

expe

ctat

ions

for 1

997


Ratings correlated with wealth:

$0

$5,000

$10,000

$15,000

$20,000

$25,000

0 20 40 60 80 100

II ratings for 74 countries

Per

cap

ita re

al G

DP


Time-series of ratings:

0102030405060708090

100

Switzerland Italy Kuwait Argentina


Returns and Institutional Investor Country Credit Ratings from 1990

R2 = 0.2976

-0.1

0

0.1

0.2

0.3

0.4

0.5

0 20 40 60 80 100

Rating

Ave

rage

ret

urns

Fit is as good as it gets - lower rating (higher risk) commands higherexpected returns. Even in among US firms, our best model gets about 30% explanatory power.


Credit Rating Model

• Intuitive• Can be used in 136 countries, that is, in countries

without equity markets• Fits developed and emerging markets


Country Risk Rating ModelSTEPS:

EVR = risk free + intercept - slope x Log(IICCR)• Where Log(IICCR) is the natural logarithm of the

Institutional Investor Country Credit Rating


Easy to use:

0%

10%

20%

30%

40%

50%

60%

70%0 10 20 30 40 50 60 70 80 90 100

Rating

Hur

dle

rate

ICRGC IICCR:84 IICCR:79


Also predicts volatility:

R2 = 0.5033

0%

10%

20%

30%

40%

50%

60%

70%

0 20 40 60 80 100Institutional Investor Country Credit Rating

Ann

ualiz

ed V

olat

ility


Fitted volatility:

0%10%20%30%40%50%60%70%80%

Rating

Exp

ecte

d vo

latil

ity

IICCR:84 IICCR:79


And correlation.

R2 = 0.6809

-20%

0%

20%

40%

60%

80%

100%

0 20 40 60 80 100

Institutional Investor Countyr Credit Rating

Cor

rela

tion

with

MSC

I AC

Wor

ld


Fitted correlation.

-100%-80%-60%-40%-20%

0%20%40%60%80%

Rating

Exp

ecte

d co

rrel

atio

n w

ith w

orld

IICCR:84 IICCR:79


Asian Crisis.

0102030405060708090

100

Jan-97

Mar-97

May-97

Jul-97

Sep-97

Nov-97

Jan-98

Mar-98

May-98

Jul-98

ICR

G ra

ting

China Hong Kong India IndonesiaKorea Malaysia Pakistan PhilippinesSingapore Taiwan Thailand Russia


Asian Crisis.

60

65

70

75

80

85

90

Jan-97

Mar-97

May-97

Jul-97

Sep-97

Nov-97

Jan-98

Mar-98

May-98

Jul-98

Sep-98

ICR

G ra

ting

Korea Malaysia Russia

Beginning of crisis


Value of US$100

020406080

100120140160180200

Jan-97

Mar-97

May-97

Jul-97

Sep-97

Nov-97

Jan-98

Mar-98

May-98

Jul-98

Sep-98

Val

ue o

f $10

0


Beginning of crisis


Value of local currency (indexed at 100)

0

20

40

60

80

100

120

Jan-97

Mar-97

May-97

Jul-97

Sep-97

Nov-97

Jan-98

Mar-98

May-98

Jul-98

Sep-98

Val

ue o

f $10

0


Beginning of crisis



• September 11 impacted the way that business is conducted all over the world (cannot be diversified away)

• It is reasonable to expect that investors demand a premium to compensate them for new investment in ventures that are now deemed riskier.


950970990

10101030105010701090111011301150

S&P 500 September 2001

September 11


950100010501100115012001250130013501400

September 2001

S&P 500 2001


80

280

480

680

880

1080

1280

1480

1680

September 2001

S&P 500 1980-2002


• Impact not as substantial as one might think in advance.

• Nevertheless, risk increased.• Initially, people thought more terror would

be soon to come.• As time elapsed, the probability of

additional terror decreased.


ICRG Political Risk Rating

60.065.070.075.080.085.090.095.0

Apr

-01

May

-01

Jun-

01

Jul-0

1

Aug

-01

Sep

-01

Oct

-01

Nov

-01

Dec

-01

Jan-

02

Feb-

02

Mar

-02

United States

World


• More impact on U.S. than average of other countries.

• Implies a small increase in the risk premium in the U.S. (10bp) and a smaller increase in world premium (2bp).


• Graham-Harvey survey of the risk premium during September 11 crisis.


Pre-Sept. 11 Post-Sept. 1110-year premiumMean premium 3.63 4.82Disagreement volatility 2.36 3.03

0.00%

5.00%

10.00%

15.00%

20.00%

25.00%

30.00%

35.00%

Argentina Mexico Thailand

CAPMIbbotson EHVGS-EHVGS-SegCSFB-EHV

68%

Risk Analysis and Project Evaluation Risk Analysis and Project Evaluation 4. Comparison of Methods

-20.00%-15.00%-10.00%-5.00%0.00%5.00%

10.00%15.00%20.00%25.00%30.00%

Slovakia Pakistan United States

CAPMIbbotson EHVGS-EHVGS-SegCSFB-EHV

537%


Excel version


Risk Analysis and Project Evaluation Risk Analysis and Project Evaluation 5. Conversion of Cash Flows

Forward Rate• Intuitive (expected exchange rate levels)• Works fine for developed countries• In emerging markets, there are two problems

– Data not readily available– Will reflect a risk premium


Forward Rate• Risk premium in forward rate will lead to “double

discounting”• Think of the forward rate as the difference between

two interest rates (local and U.S.).– This difference will tell us something about

inflation expectations– But the local interest rate also reflects a default

probability (sovereign risk)


Purchasing Power Parity• Simple theory: The exchange rate will depreciate

by the difference in the local inflation rate and the U.S. inflation rate.

• Empirical evidence shows this assumption works well in emerging markets (but not that well in developed markets)


Purchasing Power Parity• To operationalize, we need multiyear forecasts of

inflation in the particular country as well as the U.S.• The difference in these rates is used to map out the

expected exchange rates• The expected exchange rates are used to convert cash

flows into US$• We then apply the US$ discount rate to US$ cash

flows

Risk Analysis and Project Evaluation Risk Analysis and Project Evaluation 6. Project Specific Adjustments

Project Risk Analysis• Operating Risk

– Pre-completion– Post-completion– Sovereign

• Financial Risk


Operating Risk• Precompletion

– Resources available (quality/quantity)– Technological risk (proven technology?)– Timing risks (failure to meet milestones)– Completion risk

Handle in cash flows


Operating Risk• Post-completion

– Market risks (prices of outputs)– Supply/input risk (availability)– Throughput risk (material put through plus

efficacy of systems operations)– Operating cost

Handle in cash flows


Operating Risk• Sovereign Risk (Macroeconomic)

– Exchange rate changes– Currency convertibility and transferability– Inflation

Handle through discount rate


Operating Risk• Sovereign Risk (Political/Legal)

– Expropriation• Direct (seize assets)• Diversion (seize project cash flows)• Creeping (change taxation or royalty)

– Legal system• May not be able to enforce property rights



Operating Risk• Sovereign Risk (Force Majeure)

– Political events• Wars• Labor strikes• Terrorism• Changes in laws

– Natural catastrophes• Hurricanes/earthquakes/floods



Financial Risks• Probability of default

– Look at debt service coverage ratios and leverage through life of project

• Check to see if internal rate of return is consistent with (at least) the financial risks



Conclusions• Project evaluation in developing countries is

much more complex than in developed countries• Critical to: accurately identify risks and to

measure the degree of mitigation – if any.• Each risks need to be handle consistently –

either in the cash flows or the discount rate, not both.

Documents

Risk Analysis and Project Evaluation