Harvard Management Company – Inflation-linked bonds
Deals with Strategic Asset Allocation decision: how to set long-term, strategic allocation to different
asset classes (equities, bonds, or alternatives).
Short term: less than five years in future.
o Volatility is bigger risk short-term than long-term: if drop in market occurs, end date
will be too close for portfolio to recover.
o Holding investment in cash or cash-like vehicles: money market funds, certificates of
deposit.
Intermediate term: five to ten years in future.
o Time until end far enough to permit degree of volatility.
o Balanced mutual funds (mix of stocks and bonds).
Long-term: more than ten years in future.
o Stocks offer greater potential rewards. Entail greater risk, but more time to recover
from loss.
Asset allocation tools:
1. Black-Litterman Model: combines market-neutral expected returns with views.
2. Parametric Portfolio Policy: asset characteristics determine active weights.
Harvard Management Company manages endowment fund of Harvard University, total assets under
management is 19 billion USD (2000), 68% managed internally, 32% by outside asset managers.
Real return growth objective between 6% and 7%.
Role of policy portfolio for HMC:
1. Long-run (strategic) asset allocation of endowment over main asset classes.
2. Benchmark for performance measurement (endowment, managers)
3. Determined by quantitative techniques.
Strategic Asset Allocation: mean-variance analysis.
1. Objective function specification (utility function).
2. Specify asset classes.
3. Capital market assumptions (ER, volatility,
correlation).
4. Optimization.
Backtest by simulation, implementation (portfolio manager
selection), and follow up (performance measurement).
Rules for portfolio building blocks:
1. Don’t neglect equities:
a. Equities have consistently outperformed T-bills and bonds.
b. Conventional wisdom: mean reversion makes volatility to decrease with horizon.
c. Dampening effect of mean reversion is reversed by positive effect that uncertainty
of mean return has on long-run volatility stocks are riskier in long run (Pastor and
Stambaugh, 2012).
d. Need unrealistically high risk aversion to motivate small stock allocations,
𝛾
optimization problem: max 𝑤 𝑇 𝜇 − 2 𝑤 𝑇 Ω𝑤, solution: 60% equity, 40% bonds. Even
𝑤
for unrealistically high risk aversion, still get 20% allocation to equities. Combination
of two below can explain large equity premium and overpricing of options.
i. Loss aversion: investors feel worse about losses than they feel good about
gains avoid assets with large negative tails, such as stocks.
, 1. If you monitor investments frequently, especially during stress
periods, displeasure form temporary losses may drive you out of
equity (some argue that equity premium is exactly so high to
compensate for displeasure).
ii. Probability overweighting: investors tend to overweight extreme events
(relative to objective or observed probability).
Combination of loss aversion and probability overweighting can explain large equity
premium and overpricing of options (Baele et al., 2015).
iii. Ambiguity aversion: investors will have preference for asset with least
uncertainty considering assets with same distribution, ambiguity is larger for
stocks than for bonds. Reasonable levels of ambiguity/ambiguity aversion
lead to portfolios with zero equity allocation.
2. Diversify across many stocks.
a. Portfolios with large amounts of idiosyncratic risk are inefficient.
i. Stock index has volatility of 20% and beta of 1 (by construction).
b. Diversifying between stocks within same country may be sub-optimal:
i. All stocks within country may be exposed to same local economic shocks.
ii. Local index may be concentrated in limited number of industries.
iii. There is indirect international diversification through multinationals.
3. Diversify internationally, including emerging and frontier markets. International equity
market returns have become more and more correlated:
i. Economic integration: economies have become more and more
economically interlinked, and hence also firms’ cash flows.
ii. Financial integration: cross-country cash flows are discounted at same
global discount rate, set by global representative investor.
iii. Effect of financial integration stronger than that of economic integration
(Baele & Soriano, 2010).
b. Emerging markets and frontier markets interesting, because of additional
diversification at moderate risk, additional risk premiums (political risk, liquidity
risk). But also correlations of frontier markets seems to be rapidly rising.
4. Diversify across alternative asset classes (hedge funds, real estate, private equity).
a. Endowment model: theory and practice of investing that is characterized by highly-
diversified, long-term portfolios that differ from traditional stock/bond mix in that
they include allocations to less-traditional and less-liquid asset categories (private
equity, real estate) as well as absolute return strategies. Yale endowment fund is
most successful, very high proportion in alternative asset classes.
b. Difficult to invest in hedge fund (huge amounts), performance not spectacular.
c. Private equity (nonlisted equity): equity in companies that is directly sold to
investors (rather than on an organized exchange).
i. Illiquid with horizon of at least 10 years.
ii. Performance hard to measure and hence easy to manipulate
(underperformance, high costs of transaction). Most evidence: PE doesn’t
outperform listed equity.
iii. PE offers not as many diversification as some people may think.
d. Gold: lousy asset on long-term perspective (no innovation, no dividend).
i. Hedge against high inflation (gold grows equally with inflation):
1. No relation between inflation and gold.
ii. Hedge against equity market drops.
1. Drop in equity market: 20% increase in gold price, 17%
decrease in gold price.
iii. Correlation of commodity returns with equity have increased.