H U LT

I NTER NATI ONAL

BUSINESS
Session 2:
SCHOOL

Portfolio The Equity Risk Premium (ERP).

Management
S
F r é d é r i c C h a r ti e r
MFIN Program
Spring 2020
 Equity Risk Premium (ERP):
Determinants, Estimation, and Implications*
   Central
component of many risk-return models (CAPM, APT, Multifactor models).
 Key input in estimating cost of equity and determining firm and asset valuation.

 should include all risky assets. ERP often used as proxy.

 S&P 500 most often used as ERP.
Helps to determine the “price” of risk.
ERP = Premium demanded of the average risky asset.
 Key component of expected rate of return
 Key component of the discount rate applied to expected cash flows.

* Aswath Damodaran, Stern School of Business, NYU

 Consequences of Poor Choice of ERP

 ERP is a “market-wide” number, not company-specific or asset-specific.

 Yet it will affect the expected return on all risky investments.
 Huge consequences on the firm-specific cash flows, growth rates, and/or betas.

 Consequences of poor choice of ERP:

 Can affect amounts set aside by corporations or governments for future pensions.
 Can impact cost of capital determinations.
 Can affect judgement on how much to set aside for retirement.
 Can impact asset allocation.
 Determinants of ERP
1) (Collective) risk aversion of investors.
 Risk aversion ↑, ERP ↑ .
 Variables that influence risk aversion:
a) Investor age.
b) Preference of current consumption: ERP ↑ when preference for CC ↑.

2) Economic Risk:
 ERP ↓ in an economy with predictable macro variables (inflation, interest rates, etc.)
 Lower ERP in the US in late 1990’s, higher in 2008.
 Not necessarily the high level (e.g., inflation), more the uncertainty.
 Determinants of ERP (2)
3) Information.
 More complex relationship.
 Intuitively, ERP↓ with more precise information.
 Problem with too much information (noise).
 Firms try to beat analysts' forecasts  bend/violate accounting standards.

4) Liquidity:
 If investors must accept large discounts of high transaction costs to liquidate positions,
they will pay less for equities today  ERP↑.
 Determinants of ERP (3)
5) Catastrophic risk.
 Sudden drop in economic output or default by government on debt.
 Great Depression (1929-1930).
 Collapse of Japanese equities (1989-1990).
 ERP must reflex that risk.
 Globalization may have increased the frequency and probability of disasters.

6) Government Policy:
 Uncertain government policies  ERP↑.
 Extends to issues such as unstable government or incompetent bureaucracy  ERP↑.
 ERP Estimation Approaches
3 Approaches:
1) Surveys of investors, manager, and academics
2) Assess historical equity returns relative to riskless investment returns;
Use the resulting premium as expectation.
3) Estimate forward-looking premium based on current market rates or prices
(“implied premium”) .
 Surveys
1) Investors: Individual investors vs. Investment professionals
a) Individuals: Yearly surveys.
 2014 survey shows median expected stock market return
of 12.8%.
 This implied an 8.8% ERP.
b) Investment professionals: Monthly surveys.
2007-2014: Expected ERP fluctuated between 3.5% and 4.5%.

 Practitioners reluctant to use these results:

 highly responsive to market movements.
 dependent on how the question is formulated.
 Dependent on who receives the survey.
 Surveys (2)
2) Managers: Corporate Finance (hurdle rates, cost of equity, cost of capital).
 Surveys to CFO’s: 2014, ERP of 3..73%, down from 4.27% a year earlier.
 Average over 14 years: 3.54%.
of 12.8%.


3) Academics:
 Lack real world impact (rarely directly involved).
 Wide range when surveyed.
 Estimations tend to be higher (6.7% average).
 Historical Premiums

 Most widely used approach.

 Equity returns – Government security returns.
 Results range from 3% to 12%.
 Differences due to various choices of:
 Time periods;
 Risk-free rate;
 Market indices;
 How returns are averaged.
Time Periods: Short or Long?

 Ibbotson Associates goes back to 1926 .

 Many analysts use shorter periods (50, 20, 10 years).
 Reason for shorter periods: risk aversion changes over time.
 Shorter periods provide a more updated number.
 But greater “noise”.
 Size of standard error increases with shorter periods.
 Larger periods are preferable.
Risk-Free Rate: T-Bill or T-Bond?

 Yield curve is generally upward-sloping.

 Choosing T-Bill over T-Bond will increase the ERP.
 T-Bond has interest rate risk, but T-Bill has repricing risk.

 Most practitioners and academics use the “long-term, default-free,

government bond rate”.
 Choice of Market Index

 “Market” should include all risky investments . Not practical.

 Should be as broad as possible.
 Should be market-weighted.
 Should be free of “survivor bias”
 Nominal vs real returns? Not important since inflation component
Is in both market return and risk-free rate.

 Most US practitioners and academics default to S&P 500.

veraging Approach: Arithmetic or Geometric Mea

 
Geometric mean preferred by most practitioners in corporate finance.
 Stock returns tend to be “negatively correlated”.
 Increases the AM, which overstates the ERP.
-1
 Conclusion of Historic Premiums

 
Preferred choice:
 S&P 500;
 T-Bond;
 Geometric Mean;
 Long period: 1926 – latest full year.

 ERP (1928-2014) = 4.60%

 Global ERP Estimates

 ERP for other countries more difficult to estimate,

especially more emerging markets .
 Less market history.
 Dominance of a few large corporations even in more mature
stock markets (e.g., France, Italy, Germany).
 Table shows a wide rage of ERP’s (1.69% for Canada vs. 15.25% for Chile).
 Estimation for global premium from 17 countries is 3.20% (2015).
 Country Risk Premium
 General belief: should be a premium for investing in emerging countries
Due to higher political and economic risks.
 Is that risk diversifiable?
 Yes and no: some risks country-specific, but others are not (increasingly
globalized markets.
 Country risk has a “market risk” that is not diversifiable and requires
Premium.
 Increased correlation with globalization.
 Correlation increases further during periods of crises (contagion effect).
 E.g., Asian Crisis (1997-1998) or Global Financial Crisis (2008-2009).
 Correlation is higher in bear markets than in bull markets.
 Impact on cost of capital for integrated companies vs. companies in
Segmented domestic companies.
 Country Risk Premium (2)
 Most research suggest to incorporate country risk in the discount rate.
 The residual country risk that cannot be diversified away.

 Global Asset Pricing Model: Adapting CAPM.

 Use a global equity premium (e.g., based on MSCI World Index).
 Use betas that capture county risk (“country beta”).
 Problem: larger markets (US) have larger weights in the index.
 Higher correlation  end up with higher betas.
 Estimating Country Risk
  𝐸𝑅𝑃(𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑋 )=𝐸𝑅𝑃 (𝑀𝑎𝑡𝑢𝑟𝑒 𝑀𝑎𝑟𝑘𝑒𝑡𝑠)+𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑅𝑖𝑠𝑘 𝑃𝑟𝑒𝑚𝑖𝑢𝑚(𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑋 )

 3 Approaches to estimate country risk:

1) Sovereign Ratings.
2) Services (agencies) that provide county risk estimates.
3) Market-based measures.
 Sovereign Ratings
  𝐸𝑅𝑃(𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑋 )=𝐸𝑅𝑃 (𝑀𝑎𝑡𝑢𝑟𝑒 𝑀𝑎𝑟𝑘𝑒𝑡𝑠)+𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑅𝑖𝑠𝑘 𝑃𝑟𝑒𝑚𝑖𝑢𝑚(𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑋 )

 3 major rating agencies: Standard & Poors (S&P), Moody’s, Fitch.

 Currency, budget balance, trade balance, political uncertainty.
 Problems with this measure:
 Ratings lag market movements.
 Default risk obscures other risks that can affect equity markets.
 Many countries don’t have ratings (e.g., many Sub-Saharan countries).
 Country Risk Scores

 Companies such as Political Risk Services (PRS) or the Economic Intelligence

Unit (EIU) provide a more comprehensive view of risk (for a fee).
 PRS publishes a yearly reports and provides a grade (0 to 100) to each,
Country with 0 meaning no risk.
 International County Risk Guide (ICRG) at www.prsgroup.com .
 EIU provides extensive country analyses.
 Part of The Economist group (www.eiu.com).
 Market-Based Measures

1) Bond-default Spreads:
 For countries that issued government bonds denominated in USD,
Euros, or Yen.
Example (January 2015):
10-Yr USD Brazilian government bond = 3.87%
10-Tr US T-bond = 2.17%
 Brazil’s sovereign bond default spread = 3.87%-2.17% = 1.70%
 Market-Based Measures (2)

2) Credit Default Swap (CDS) Spreads:

 Measures default risk.
Example (January 2015):
CDS for US = 0.31%
CDS for Brazil= 3.17%
 CDS spread for Brazil = 3.17%- 0.31% = 2.86%
 CDS spreads higher than bond spreads.
 Counterparty risk on the CDS transaction.
 Other risks taken into account.
 Market-Based Measures (3)

3) Stock Market Volatility:

 Measures by the standard deviation of returns of local stock prices.

Advantage of Market-based numbers: they are constantly updated.

 Estimating the Country Risk Premium

 Linking a country risk measure to the country risk premium.

1) Default Spreads.
2) Relative Equity Market Standard Deviation.
 Estimating the Country Risk Premium
1. Default Spreads

Default Spreads: Spread on sovereign bond or CDS market.

a) Current default spread:
Ex: Jan. 2015 default spread between 10-Yr Brazilian vs US T-bond = 1.70%.
b) Average default spread:
To reduce volatility, take average spread over 5 years = 1.65%.
 Default Spreads(cont’d)
c) “Imputed” or “synthetic” spread:
 Both earlier approaches can only be used if emerging country
has a sovereign bond issued in USD, Euro, or Yen.
 For those who don’t, find a country with similar sovereign rating
Or CDS default spread.
 Example: Bulgaria has the same rating as Brazil (Baa2).
  Apply Brazil’s default spread to Bulgaria.
 For those who don’t have a sovereign rating, look for similar PRS score
And apply same default spread.
 Cuba and Cameroon have the same PRS score. Cuba = Caa1.
Apply Cuba's’ default spread.
 Default Spreads(cont’d)
 Some analysts add the default spread to the US ERP and
multiply by company beta.
 Others add default spread to cost of equity and cost debt
of the company traded in that country.

 Others will prefer using the spread between the average return
of 10 emerging market ADR’s with the US S&P return.
 Estimating the Country Risk Premium
2. Relative Equity Market Standard Deviations
 Equity risk is measured by standard deviation of equity returns.
  𝑆𝑇𝐷𝐸𝑉 𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑋
𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑆𝑇𝐷𝐸𝑉 𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑋 =
𝑆𝑇𝐷𝐸𝑉 𝑈𝑆
 

   Example: Annualized STDEV S&P 500 (2013-2014) = 10.85%.


ative Equity Market Standard Deviations (cont’d)

 Problems with this method:

 Market Liquidity: the most liquid markets tend to be the most volatile.
Panama and Costa Rica have smaller STDEV than the US
 their ERP’s are lower than the US ERP.
 Currency differences: STDEV measured in local currency.
 Need to combine relative currency volatility
with equity market volatility.
 Estimating the Country Risk Premium
Default Spread + Relative Equity Market STDEV
 Combine both previous methods.
  𝜎 𝐸𝑞𝑢𝑖𝑡𝑦
𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑟𝑖𝑠𝑘 𝑝𝑟𝑒𝑚𝑖𝑢𝑚= 𝐶𝑜𝑢𝑛𝑡𝑟𝑦 𝑑𝑒𝑓𝑎𝑢𝑙𝑡 𝑠𝑝𝑟𝑒𝑎𝑑 𝑥
𝜎 𝐺𝑜𝑣𝑒𝑟𝑛𝑚𝑒𝑛𝑡 𝑏𝑜𝑛𝑑
   Example: Brazil (Jan.2015).
Default spread = 1.70% (with US)

 x = 3.25%
Add to US ERP: 5.75% + 3.25% = 9.00% (Brazil’s ERP)
 Implied Equity Premiums
Historical premiums are relatively easy to calculate
but they are “backward looking”.

 DCF Model-Based Premiums.

 DCF-based Premiums
 
Price of a financial asset
= Discount of future cash flows @ required/expected rate of return.

 
1) Stable growth DDM Premiums:

Value of Equity =
where,
 Value of Equity = Current market index
 Growth rate (g) = (1 - Payout ratio)(ROE)
 Expected dividend next period = Expected earnings next period x Payout ratio.
 Stable Growth DDM Premiums (cont’d)
Assumptions: In the long-run,
 Growth rate in dividends = risk-free rate.
 Companies pay all earnings in dividends

 
In that case, equation becomes:

Value of Equity =

 Dividends/Value of Equity =

 Dividend Yield = Equity Risk Premium

 Stable Growth DDM Premiums (cont’d)
Using Earnings instead of Dividends:
  (Expected) Growth Rate (g) =
= (1-Payout Ratio)(ROE)

Value of Equity =

If we assume that ROE = Required return on Equity,

Value of Equity =

Required return on Equity =

 Using S&P 500
In-Class Exercise:
December 31, 2007, S&P 500 index closed at 1,468.36.
Dividend yield on Index = 1.89%
Estimated future annual earnings growth = 5% for next 5 years.
Long-term growth rate (g) = Risk-free rate or return.
10-Year US T-Bond yield = 4.02%

What is the implied equity risk premium?

 Using S&P 500
The focus on dividends only may be understating the premium,
Because many companies use cash for stock buybacks.

Assume that:
(Dividend + Stock Buyback) yield on Index = 4.11%
Other data unchanged.

What is the implied equity risk premium?