LEARNING MODULE 2: CAPITAL MARKET EXPECTATIONS,

PART 2: FORECASTING ASSET CLASS RETURNS

LESSON 1: FORECASTING FIXED INCOME RETURNS

LOS: Discuss approaches to setting expectations for fixed-income returns

Applying Discounted Cash Flow (DCF) to Fixed Income

The discounted cash flow (DCF) approach to setting expectations assumes that the value of an asset is the present value of its
cash flows discounted at investor’s the required rate of return. Through the analysis of yield curves, the approach works well for
setting return expectations for fixed-income securities. It also allows detailed return decomposition and attribution, as discussed
in the curriculum sections on fixed income and performance attribution.

The starting point for the return expectations of a bond is its yield to maturity (YTM), the single discount rate equating the
present value of the bond’s cash flows to its price. The YTM is only actually earned if the following three conditions are met:

1. The cash flows of the bond are received in full and on time.
2. The bond is held to maturity. If the bond is sold prior to maturity, the investor will underperform the YTM if interest rates
have risen over the investment horizon, causing bond prices to fall.
3. Cash flows are reinvested at the YTM. The investor will underperform the YTM If interest rates fall causing early cash
flows received on the bond to be reinvested at a rate lower than that of the original YTM of the bond.

Generally, the effects of conditions 2 and 3 above offset each other, and if the investment horizon is equal to the (Macaulay)
duration of the bond or portfolio, they will roughly cancel out. If the investment horizon is less than the duration of the portfolio,
then condition 2 will dominate and the investor will underperform the YTM when interest rates rise. Conversely, if the
investment horizon is greater than the portfolio’s duration, condition 3 will dominate and the investor will underperform the
YTM when interest rates fall.

Note that the timing of an expected interest rate move does matter, since earlier shifts in longer investment horizons will have a
larger impact on reinvestment return.

The Building Block Approach

The building block approach breaks the expected return for a fixed-income security into the following four components: the
short-term nominal default-free rate, a term premium (to compensate investors for duration risk), a credit premium, and a
liquidity premium.

Short-Term, Nominal Default-Free Rate

The short-term nominal default-free rate is usually proxied by the government zero-coupon bill with the shortest maturity (e.g.,
three months). This is the only directly observable component of expected return and is closely tied to the central bank’s policy
rate, which will rise and fall in line with the business cycle. Short-term default-free rates that reflect a negative policy rate are
likely to underrepresent investment risk and may require normalization to more usual non-negative values. Alternatively, other
risk premiums may be increased to reflect elevated willingness to pay for safety.

Term Premium
The term premium is based on the instrument’s horizon—it is usually positive, increases with maturity, and varies over time. The
term premium is driven by:

1- Level-dependent inflation uncertainty—Arguably the main driver of term premium. High levels of inflation tend to lead
to higher levels of inflation uncertainty. Hence, high inflation increases the term premium due to both the high expected
inflation and an increased inflation risk (i.e., less accuracy in forecasting inflation).
2- Supply and demand—High demand and undersupply for longer-dated bonds will lead to lower long-term yields and a
flatter yield curve.
 3- Cyclical effects—The yield curve slope varies substantially over the business cycle—steep at the trough of the cycle and
flat or even inverted around the peak. This is mainly due to short-term policy rates moving in line with the business cycle.
4- Ability to hedge recession risk—If a type of asset performs well when the economy is weak, then its term premiums will
be lower. A demand shock is likely to see risky assets underperform and bonds act as a safe haven, implying a relatively
low term premium. However, a supply shock to economic activity would likely be inflationary, leading to higher interest
rates and lower bond prices. In this situation a high term premium would be warranted.

Credit Premium (Default Risk)

The credit premium is based on issuer creditworthiness, including both expected losses and additional risk of unexpected
losses. Primary drivers are:

• Downgrade bias (the chance of downgrade exceeding the chance of upgrade) for AAAand AA bonds
• Cyclical risk affecting spreads for A and BBB bonds
• Default risk for below-investment-grade bonds

Although a longer maturity would suggest greater opportunity for loss and therefore larger credit premiums, historical evidence
suggests that shorter-dated credit premiums are surprisingly attractive relative to longer-dated credit premiums. This may be due
to lower prices resulting from illiquidity at the short end, since old issues become less liquid as they approach maturity. It may
also be related to the “event risk” of large proportional losses due to default remaining on short-dated instruments with only a
few years of returns left as compensation. Either way, the high credit spreads at the shorter end of the yield curve have given rise
to the credit barbell trade, which involves taking credit exposure at the short end oft he curve and duration exposure through
longer-maturity government bonds.

Liquidity Premium
The liquidity premium is based on the ability to trade a bond with little market impact and transaction cost. The most liquid
bonds are recently issued sovereign bonds, current-coupon mortgage-backed securities (MBS), and the largest high-quality
corporate bonds. Outside of these issues, liquidity depends on the inventories of securities held by bond dealers, which will be
higher for bonds that are of higher quality, more recently issued, larger in issue size, priced closer to par, issued by a well-known
issuer, or simple in structure.

LOS: Discuss risks faced by investors in emerging market fixed-income securities and the country risk analysis techniques used to
evaluate emerging market economies.

Fixed-Income Securities in Emerging Markets

Investors in emerging market debt face the same fixed-income risks as those in developed markets, (e.g., duration and default),
plus the risks outlined below, which are broadly categorized as “ability to pay” and “willingness to pay.”

Economic Risks/Ability to Pay

Relative to developed markets, emerging markets’ sovereign fixed-income markets may have:

• A less diverse tax base due to greater concentration of wealth

• Vulnerability to capital flight during a crisis (i.e., capital being withdrawn from the economy)
• Less educated/skilled workforce; limited infrastructure, and/or less technological sophistication
• Greater dependence on commodities and agriculture (i.e., cyclical industries with low pricing power)
• Small, less sophisticated financial institutions and markets
• Restrictions on trade, capital flows, and exchange rates
• Poor monetary discipline and fiscal control
• Reliance on foreign borrowing, especially from developed countries with hard currency
Political and Legal Risks/Willingness to Pay

• Unstable coalition governments unwilling to honor promises by previous/current regimes.

• Sovereign immunity can make it difficult or impossible to get the government to pay debt
• Weak property rights laws and enforcement, leading to the inability to enforce claims or recover capital

Warning signs of potential inability or unwillingness to pay might include:

• Debt/GDP ratio greater than 70–80%

• Foreign debt greater than 50% of GDP or 200% of current account receipts
• Annual real growth less than 4%
• Persistent current account deficits greater than 4% of GDP
• Foreign exchange reserves less than 100% of short-term debt

Access to the International Monetary Fund (IMF) or the World Bank may slightly reduce concerns about ability or willingness
to pay.

Analysts should consider looking at how the market has historically responded to problems associated with this list,
understanding that past performance is no guarantee of future results.

LESSON 2: FORECASTING EQUITY RETURNS

LOS: Discuss approaches to setting expectations for equity investment market returns.

Historical Statistics Approach

Sample averages, even over very long histories, provide imprecise estimates for mean return unless the variance of the data is
small relative to the mean. Different countries have different characteristics, but equity returns generally do not have small
variance relative to the mean. Confidence intervals for the true mean returns of different countries therefore largely overlap,
meaning there is little statistical support for believing that true mean equity returns differ much across countries.

Discounted Cash Flow (DCF) Approaches to Equity Returns

The Gordon growth model can be used to isolate the expected return in equity markets in order to set capital market
expectations. This model links long-term changes in equity market value to changes in economic growth. The relatively noisy, in
terms of volatility, impact of the P/E ratio and the ratio of earnings-to-GDP can fortunately be ignored in this model.

The Grinold-Kroner model extends the Gordon growth model to consider the effects of share repurchases and changes in
valuation levels through the forecast horizon. Under this model, the expected return on an equity market, E(Re), is given by the
formula:

𝑫
E(Re) = + (%ΔE - %ΔP/E
𝑷
Some points of note:

1- The term - %ΔS is referred to as the “rate of net share repurchases” and represents income from company buybacks.
Hence, the “income” component of expected return is D/P - %ΔS.
2- Expected capital gains are composed of the nominal earnings growth rate, %ΔE, plus the repricing return, %ΔP/E.
3- The term (%ΔE - % ΔS ) represents the estimated rate of change of earnings per share.
Earnings cannot grow faster than the economy in perpetuity since this would lead all wealth flowing to company earnings, which
is not a realistic assumption. Also, it is unreasonable to assume in perpetuity that valuations (P/E) can expand forever, or that
companies can issue or redeem shares (S) forever. Hence, the only very long-run assumptions that make sense for this model are
%ΔE = Nominal GDP growth, %ΔS = 0 and %ΔP/E = 0. However, analysts may assume any value for these inputs that is consistent
with their view over their shorter-term investment horizon.

Most of the inputs to the Grinold-Kroner model are readily available as estimates from research published by supranational
organizations, such as the IMF, or investment banks. The estimation of change in P/E, however, is more complicated and often
involves the assumption that extreme values revert to a longer-term mean. To remove the impact of the business cycle when
gauging the extremity of valuation levels, analysts could use the cyclically adjusted P/E(CAPE). The CAPE measures price relative
to average real earning over the prior 10 years, thereby providing a more reliable signal with regard to whether valuations are
currently too high/low.

Risk Premium Approaches to Equity Returns

Whereas DCF methods and the Grinold-Kroner model reflect the supply of equity returns because they outline sources of
returns, risk premium approaches reflect the demand for equity returns. The equity premium may start from a base that can
take two forms:

• A premium over default-free short-term bills or

• A premium over default-free bonds.

The latter includes not only the nominal risk-free rate, but a term premium as well, so the first description should be used as a
building block for the equity premium. However, since equity returns are significantly more volatile than bills or bonds, the same
estimation issues relating to accuracy arise for estimating these risk premiums as exist for directly estimating the equity returns
themselves.

Equilibrium Approaches to Equity Returns

The Singer-Terhaar method extends the basic CAPM model to incorporate the impact of individual markets having incomplete
integration with the global market.

The model begins with a calculation of the risk premium appropriate for an individual market under CAPM, which assumes the
market is fully integrated with the global market (recall that a key assumption for CAPM is that there are no barriers to capital
flows).

Applying the CAPM framework for the expected return of a market “i” given its level of systematic risk relative to the global
market portfolio (GM) gives:

E(Ri) = Rf + βi GM RP GM
Where Rf is the risk-free rate, RPGM is the risk-premium of the global market portfolio, and βi,GM is the sensitivity of
market “i” to systematic global market risk. This equation shows that the risk premium (i.e., excess return of Rf) of
market “i” assuming 100% integration with the global market, (RPi), is given by:
RPi = βi,GM RPGM
When the degree of integration is some value φ between 0% and 100%, the appropriate risk premium can be calculated as a
weighted average of RPi and RPiS as follows:

RPi = φRPGi + (1 – φ) RPSi)

Highly integrated markets will have a range of 0.75–0.90 degree of global integration whereas emerging markets will have a
range of 0.50–0.75. A liquidity premium would then be added to the equity risk premium; less integrated markets are less liquid
and would therefore require a greater liquidity premium.
Equity Securities in Emerging Markets

LOS: Discuss risks faced by investors in emerging market equity securities.

Emerging market equities present many of the same risks as emerging market fixed-income securities, such as:

• more fragile economies,

• lower degree of informational efficiency,
• less stable political and policy frameworks, and
• weaker legal protections.

However, whereas emerging market debt analysis considers ability and willingness to pay, emerging market equity investors need
to focus on a variety of risks beyond the traditional credit and counterparty risk, especially in times of macroeconomic and
political distress.

For example, the majority of listed companies in an emerging market might be “closely held,” meaning that the stock is held and
controlled by a small number of investors with only a minority of shares listed on the exchange. This could give rise to corporate
governance issues, where company decisions are made to prefer the dominant shareholders over minority investors. In some
cases, similar issues may arise when the government has a stake in the business. These challenges vary from country to country,
so country effects typically outweigh industry effects.

LESSON 3: FORECASTING REAL ESTATE RETURNS

LOS: Explain how economic and competitive factors can affect expectations forreal estate investment markets and sector returns.

Historical Real Estate Returns

Real estate is inherently quite different from equities, bonds, and cash. It is a physical asset rather than a financial asset. It is
heterogeneous, indivisible, and immobile. It is a factor of production that produces a return. The heterogeneity, immobility, and
illiquidity of real estate pose a problem for historical analysis. Individual properties trade infrequently and erratically over time.
Even in more developed real estate markets, there is a tendency for market transactions to occur in properties with lower to
moderate price growth.

Real estate owners/investors generally rely on appraisals instead of transactions to value properties. These appraisals tend to
reflect slowly moving averages of past market conditions. Returns calculated from appraisals don’t, in general, bias the mean
return. But they do underestimate the true volatility of returns.

Capitalization Rates
The capitalization rate—the standard metric for valuing commercial real estate—equals net operating income (NOI) divided by
the property’s value. This is analogous to EBITDA as a percentage of EV (reciprocal of EV/EBITDA). While the capitalization rate
represents an income yield, it isn’t a cash flow yield given that a portion of the operating income may be reinvested into the
property. Investors can add long-term expected growth in income to this set of return expectations.

E(R re) = Cap rate + g NOI

The long-term NOI growth rate, g NOI, should be close to the nominal GDP growth rate. The NOI growth rate may be decomposed
into real NOI growth and an inflation component, in which case real NOI growth should be close to real GDP growth. Short-term
forecasts can incorporate changes in valuation through estimated changes in the cap rate, rather than through modifying the
expected long-term growth rate. Note that for a fixed NOI, a fall in real estate values will imply an increase in the cap rate:

E(Rre) = Cap rate + gNOI - %Δ Cap rate

Higher cap rates tend to result from greater risk to the NOI stream, such as for properties in hotels versus apartments, secondary
cities versus top-tier cities, and lower-quality low-productivity malls versus high-productivity malls.

Capitalization rates represent discount rates, hence, might be expected to rise and fall in a pro-cyclical manner in line with
interest rates. However, the cap rate is also sensitive to credit spreads which are countercyclical (i.e., rise when economic activity
is low and vice versa), which lowers the cyclicality of cap rates.

Real estate investments usually involve substantial leverage, hence, an increase in the availability of debt financing should
improve the liquidity of real estate markets and lower liquidity premiums.

Risk Premium Approaches to Real Estate Returns

Real estate has a high duration (compared with stocks and bonds) because it is a long-lived asset. As such, real estate must earn a
high term premium. In addition, income-earning properties must earn the credit premium associated with their rent-paying
tenants. Properties must earn a significant equity risk premium to compensate owners for uncertainties related to changes in
property value, rent growth, lease rollover/termination, and vacancy rates. Because real estate has both bond-like properties
(term and credit premiums) and stock-like properties, the risk premium should be between those of bonds and stocks.

Liquidity risk, defined for direct investors in real estate as the inability to sell except at infrequent, relatively random points in
time, is significant and warrants a liquidity premium. For a given property, the appropriate liquidity premium should be inversely
related to the frequency with which the property is traded.

Equilibrium Approaches to Real Estate Returns

It makes sense to consider real estate in a global equilibrium framework such as Singer-Terhaar because it is so important in
global portfolios, but there are important considerations:

1- Appraisal-based data should be unsmoothed in order to remove the downward bias in volatility and correlations with
other asset classes.
2- A liquidity premium should be added to the required returns of equilibrium models since these models assume assets
are liquid.
3- It should be recognized that real estate returns may rely on local factors rather than global financial markets.

Public versus Private Real Estate

Because real estate is expensive, smaller investors may not be able to properly diversify a portfolio of properties. Indirect
ownership may be more appropriate for smaller investors, but comparing risk and return characteristics is complicated by
heterogeneity, return smoothing, and variations in leverage.

To make sensible comparisons between the returns of direct real estate and REITs, the analysis should use unlevered transaction-
based returns for direct investments, de-lever REIT returns, and make sure that the underlying properties under consideration
are of a similar nature.

Once these adjustments are made, REIT returns appear superior to direct real estate returns, which suggests investors do not
forgo performance for the extra liquidity of REITs versus direct real estate investments. REITs tend to behave like equities in the
short run but behave more like direct real estate over the long run.

For most individuals, residential real estate is the greatest source of exposure to real estate investment. Residential real estate
outperformed equity markets worldwide with lower volatility over the very long term (1870–2015), however, it has
underperformed equities since World War II.
LESSON 4: FORECASTING EXCHANGE RATE RETURNS

LOS: Discuss major approaches to forecasting exchange rates

The exchange rate of a currency is affected by a multitude of different forces, including, but not restricted to, government policy,
central bank action, financial systems, trade flows, interest rates, inflation, investment flows, and even expectations of future
exchange rate changes. Because of this there is no single forecasting rule for exchange rates that works in all markets at all times.
The theories discussed in this lesson should be viewed as ideas that may explain why an exchange might behave in a certain way
when certain forces are dominating the supply and demand of the currency.

Goods and Services, Trade, and the Current Account

There are three main ways that trade in goods and services can impact an exchange rate: trade flows, quasi-arbitrage of prices
(purchasing power parity [PPP]), and competitiveness and sustainability. We consider each in turn.

Trade Flows
The net trade flows (Exports – Imports) of a country are unlikely to drive exchange rates, as long as there is an offsetting
supply/demand for the currency for finance/investment purposes on the capital account. For example, a country that runs a
trade deficit (Imports > Exports) will not see its currency weaken if there is an offsetting surplus on the capital account
representing financing/investment flows into the country. A large trade balance relative to finance/investment flows, however,
could be an early warning sign of a large exchange rate change and a potential currency crisis.

Purchasing Power Parity

PPP is based on the idea that the same goods and services should cost the same in different currencies. Therefore, if there is high
price inflation expected in a country, then the currency should be expected to depreciate such that the real exchange rate stays
the same. PPP is a poor exchange rate forecasting tool over short to intermediate horizons; however, it has been shown to hold
over long-time horizons.

The monetary approach to exchange-rate forecasting makes two assumptions: first, that PPP holds, and second, that inflation is
driven by the money supply.

Competitiveness and Sustainability of the Current Account

As noted above, the impact of the current account balance (primarily driven by net trade flows) on exchange rates is likely to be
small if offset by flows on the capital account. Restricting capital flows, therefore, makes a country’s exchange rate more sensitive
to a negative current account balance. The size and source of the imbalance determine whether it will be persistent and
sustainable. Negative current account balances of 2% or less of GDP generally pose no threat to exchange rates and may be
sustainable for many years. A larger but temporary negative current account may also pose no threat. A country may continue
running current account deficits if profitable domestic investment attracts international capital from its trading partners.

Whereas temporary trade imbalances arise from business cycles, longer-term structural trade imbalances leading to persistent
current account deficits result from:

1. Persistent fiscal imbalances

2. Profitable investment opportunities for innovation and capital deepening
3. Institutional characteristics affecting savings decisions (e.g., demographics and preferences)
4. Availability of important natural resources
5. Prevailing terms of trade

Current account imbalances may be slow to change due to inertia in changes in consumption, savings, and production patterns.
Changes to exchange rates also may be slow due to people updating their exchange-rate expectations slowly over time.

Capital Flows
Since adjustments to the current account tend to occur only gradually, the primary driver of shorter-term movements in
exchange rates is more likely the capital account (i.e., financing/investment flows), assuming capital can flow freely between
countries.
There are three main theories regarding how capital flows (financing/investment flows) can impact an exchange rate:

1. Dornbusch overshooting mechanism

2. Uncovered interest rate parity
3. Portfolio balance, composition, and sustainability issues. We consider each in turn.

The Dornbusch Overshooting Mechanism

Capital is assumed to flow toward the market with the highest available investment return. Returns can be higher in an economy
for a combination of many reasons including a higher risk-free rate, term premium, credit spreads, equity risk premium, or
liquidity premium. Assuming perfect capital mobility (i.e., no restrictions to capital flows), the Dornbusch overshooting
mechanism assumes that capital flowing into a higher-returning domestic economy will instantly strengthen the currency up to
the point where, looking forward, the currency is expected to weaken by the return advantage of the domestic economy.

For example, assume that Country A has an overall investment return advantage (from a combination of risk-free rate, term
premium, credit spreads, equity risk premium, or liquidity premium) of 10% over Country B. The Dornbusch overshooting
mechanism assumes

• Capital flows into Country A to earn the 10% investment return advantage, instantaneously strengthening the currency of
Country A versus Country B.
• This instantaneous strengthening of currency A occurs up to the point where, looking forward, the currency of Country A
is expected to depreciate by 10% versus currency B.

The idea that a stronger performing economy is expected to see currency weakness may seem contradictory to the idea that
capital is expected to flow toward higher-returning economies; however, critical to the understanding of this model is that
forecasts are based on the second stage, not the first. The first stage is assumed to happen instantaneously, and the model is
looking forward and projecting currency returns as per Stage 2.

Uncovered Interest Rate Parity

Uncovered interest rate parity (UIP) states that expected exchange rate changes are only related to nominal interest rate
differences (i.e., the first term in the previous equation). It is a highly simplified version of the overshooting mechanism which
states that countries with higher nominal interest rates have currencies which are expected to weaken by their nominal interest
rate advantage over time.

If uncovered interest rate parity held universally across currency markets, then it would be impossible to earn excess returns
from the carry trade, a trade where low interest rate currencies are borrowed in order to deposit in high interest rate currencies.
The empirical observation that the carry trade can work is evidence that uncovered interest rate parity does not hold—high
interest rate currencies do not tend to weaken according to their interest rate advantage over other currencies.

Once again, this empirically observed strengthening of the currency with superior returns is likely a reflection of the first stage of
the overshooting mechanism playing out over time, rather than happening instantaneously, as capital gradually flows toward the
higher interest rate currency (referred to as the “hot money effect”). Hot money inflows can be problematic for central banks
since they:

• interfere with effective monetary policy

• can be an unstable source of funds for domestic firms
• Strengthen the currency and make domestic exporters less competitive.

The central bank can combat the third issue by selling domestic currency to the market in order to stop domestic currency
appreciation. If the central bank were concerned about the inflationary impact of the subsequent increase in the money supply,
this could be sterilized (i.e., mitigated) by the central bank simultaneously selling bonds to financial institutions to remove
liquidity from the financial system, thereby simultaneously lowering the money supply.
Portfolio Balance, Portfolio Comparability, and Sustainability Issues
The global market portfolio is made up of each country’s unique mix of assets. Changes in exchange rates automatically adjust
the relative size of each country’s assets in the global portfolio to reflect investors’ desire to hold them.

Investors may desire tactical adjustments to recognize opportunities brought about by relative economic strength and related
policy measures. For example, capital will likely flow into economies entering the growth phase of the business cycle as growth
opportunities emerge, particularly if economic expansion is driven by investment in productive assets.

Longer term, the allocation to a country in the global market portfolio depends on that country’s relative long-term trend growth
rate and current account balance. A country experiencing strong economic growth will increase as a proportion of the global
market portfolio, which may lead to investors rebalancing through selling the assets of the country in order to maintain the same
strategic asset allocation, causing currency weakness. However, this effect may be mitigated by:

• Home country bias in the fast-growing economy—Domestic wealth will increase as the economy grows, and domestic
investors are likely to be inclined to overweight domestic assets, thereby absorbing the growth in the country’s assets.
• Productivity-driven growth—both foreign and domestic investors will fund growth driven by increases in productivity
over the longer term.
• Small initial allocations of the fast-growing economy in the global portfolio—the share of local currency–denominated
assets likely has room to grow in the global portfolio before investors rebalance and cause currency depreciation.
A currency with persistently high current account deficits will eventually experience downward exchange rate pressure,
although there are mitigating factors here, too:
o Deficit source matters—Deficits due to profitable investment spending can usually be financed without
downward exchange rate pressure. Deficit s resulting from a lack of fiscal discipline or a low savings rate are
more likely to result in depreciating currency.
o Reserve currency exceptions— some currencies of larger countries are considered safe havens for debt
investment; that is, international trade occurs in the currency and governments are unlikely to renounce their
sovereign debt obligations. Small deficits are welcome because they provide liquidity to the global economy.

In frictionless markets, everyone would be satisfied to hold the same portfolio composition, regardless of wealth transfer impacts
from a country with a current account deficit to a country with a current account surplus. Due to home country bias, however,
demand for the current-account surplus country’s assets and currency appreciates while the current-account deficit country
currency depreciates. Asset price adjustments resulting from changes in interest rates and risk premiums should take the
pressure off exchange rates; however, they do not.

Certain types of flows are more conducive to maintaining rates than others. Direct foreign investment, including in private equity
and real estate, indicates a long-term commitment from abroad, even if the flows do not create new assets or expand the
economy. Government debt incurred for public investment in economically productive assets might be serviced from investment
returns and suggests less stress on the exchange rate. The latter may give rise to debt sustainability concerns; however, as the
debt service becomes more burdensome on an economy. Large or rapidly accumulating short-term borrowing may be a warning
sign.

LESSON 5: FORECASTING VOLATILITY

LOS: Discuss methods of forecasting volatility

Forecasting a Variance/Covariance Matrix

Capital market expectations required as inputs for asset allocation models include individual variances of assets and their
covariance with each other. These data are often presented in matrix form, referred to as a variance-covariance (VCV) matrix.

We consider three methods of VCV forecasting: historical sampling, multi-factor modeling, and shrinkage estimating.

Historical Sampling
The simplest and most popular method used to estimate a VCV matrix is to base forecasts on historical sample data. There are
three important issues associated with this process:
 1. Sample size (number of historical time periods used): needs to be significantly larger than the number of assets in the
VCV matrix, or some portfolios will incorrectly appear to be riskless.
2. Sampling error: can be substantial if sample sizes are not large enough.
3. Cross-sectional inconsistency: means that each pairwise correlation is estimated without regard to the other pairs.

The first two problems relate to small sample issues and generally may be resolved by using 10 times the number of periods in a
sample as the number of assets in the matrix.

Multi-Factor Modeling
As mentioned above, data requirements for VCV forecasting are significantly higher than the number of assets in the VCV matrix.
Data requirements, and the number of variance/covariance forecasts required, can be reduced significantly through the use of
multi-factor models to explain asset returns.

For example, by mapping the returns of 100 assets to just 5 systematic factors, the analyst has the considerably less onerous task
of estimating the pairwise covariance of just the 5 factors rather than the covariance of all 100 assets in the portfolio. This should
significantly reduce estimation error.

A multi-factor model also has the advantage of imposing cross-sectional consistency on the covariance forecasts. Cross-sectional
consistency ensures that for any three assets in a portfolio, A, B, and C, the estimates of pairwise covariance (A vs. B, B vs. C, and
A vs. C) are mathematically consistent with each other. This is not necessarily a feature of a VCV matrix derived from pairwise
historical samples.

The disadvantages of multi-factor models are:

• Multi-factor models are never perfect representations of reality and are always subject to some level of misspecification
error. As a result, a multi-factor model will be biased in its estimation.
• Misspecification will also cause the undesirable trait of statistical inconsistency—as more data is used, covariance
estimates do not converge to true underlying covariance.

Shrinkage Estimating
As discussed above, historical sampling estimates the true VCV matrix with high imprecision, whereas a multi-factor model has
high precision but is biased.

A shrinkage estimator aims to improve the overall efficiency (i.e., to lower both bias and estimation error simultaneously) of a
forecast through combining both approaches. It is a weighted average of VCV forecasts derived from historical samples and
model-based forecasts (such as the multi-factor model discussed above).

Unsmoothing Appraisal-Based Returns

It is common in illiquid asset classes, such as real estate and private equity, to use appraisal-based valuation rather than
transaction price. Appraisal-based valuations are well known to lag true underlying valuations due to infrequent appraisals, the
use of stale data, and psychological anchoring to previous appraisal values. This smoothing of returns dampens volatility and
correlations with other assets, understating risk and overstating the diversification benefits of the asset class. For this reason, an
analyst should unsmooth appraisal-based returns using an unsmoothing model.

A simple, widely used unsmoothing model assumes the current observed (smoothed) return, Rt, is a weighted average of the
current, unobservable true return, rt, and the previous observed return Rt-1.

Rt = (1 – λ)rt + λRt-1

Time-Varying Volatility: ARCH Models

The forecasting methods considered so far in this chapter assume that variances and covariance are constant through time.
However, asset returns are shown to exhibit volatility clustering, defined as periods of low or high volatility over time.
Autoregressive conditional heteroscedasticity (ARCH) models can be used to model this time-varying nature of volatility.
LESSON 6: ADJUSTING A GLOBAL PORTFOLIO

PORTFOLIOLOS: Recommend and justify changes in the component weights of a global investment portfolio based on trends
and expected changes in macroeconomic factors.
Even a portfolio with a stable and client-appropriate long-term strategic asset allocation will need to make short-term tactical
changes to reflect opportunities that evolve in global markets.

These opportunities to reallocate the portfolio will likely revolve around the following macroeconomic factors:

• Countries with increasing trend growth rates: Increase allocation to equities due to the likely increase in corporate
earnings growth. Decrease allocations to bonds, due to potential monetary tightening leading to higher interest rates
• Countries with increasing levels of global integration: Increase allocation in general to these markets, as increased
integration will likely lead to rising prices as risk premiums in the market decline.
• Countries at the trough of the business cycle: Increase allocation to equities due to the likely attractive valuations.
Decrease duration in the bond portfolio allocation due to expected rising rates but increase exposure to credit as credit
spreads are likely to decrease and credit-risky asset prices rise. The opposite allocations should be made for economies
that are approaching the peak of the cycle.
• Monetary and fiscal policy: Allocate toward countries with positive structural shifts in policy that are not already
reflected in market prices.
• Current account: Current account deficits need to be funded through a surplus on the capital account. In order to attract
capital, it is likely required returns will rise and prices will fall. Therefore, allocate away from countries with secular
current account deficits and toward those with surpluses.
• Currencies: Analysts should be wary of countries with attractive domestic returns but a strong currency that
“overshoots” and is therefore expected to depreciate going forward.

While making judgments relating to the above factors, a portfolio manager must always be aware of what information is
currently reflected in asset prices and returns. For example, a manager would only consider increasing the equity allocation
to a market with superior trend growth if this growth were not already priced into the market.

A process for setting capital market expectations that brings together all the tools of this section is outlined below:

• Estimate the VCV matrix for all asset classes.

• Estimate equilibrium (i.e., fair) asset class returns using the VCV matrix and the Singer-Terhaar model.
• Estimate actual expected equity market returns using the Grinold-Kroner model,incorporating analyst views on growth,
valuation, and income.
• Estimate fixed-income returns using the building block approach, incorporating assumptions regarding economic cycle
and monetary/fiscal policy.
• Estimate currency movements using relative investment opportunities and the potential for overshooting.
• Incorporate potential currency movements into expected equity and bond returns.
• Combine equilibrium (fair) expected returns from the Singer-Terhaar model with actual expected returns from the
remaining steps using the Black-Litterman