The abCs of Commodity Investing

1
The abCs of Commodity Investing

• Claude Erb
• Trust Company of the West
• Campbell Harvey
• Duke University

2
Background

• Original research completed in May 2004

May 2004
Source Bianco Research, Market Facts, vol. 12,
no. 10 How Big Are Commodity Index Funds?
3
Why?

• At least two types of research
• That which supports a predetermined conclusion to
  launch a product
• That which tries to weigh the opportunities and
  pitfalls of an investment
• Our focus are commodities an appealing
  alternative?

Internet Bubble
Commodities Relatively Cheap
Death of Equities August 1979
Stocks Relatively Cheap
4
Some Questions Worth Asking

• Are commodities an asset class?
• What is the composition of the commodity
  market?
• What is the expected beta return of
  commodities?
• Is there a risk premium?
• How much should be allocated to commodities?
• Is there commodity alpha?

5
Are Commodities an Asset Class?

• Yes
• If one defines commodities as an asset class
• No
• If one does not define commodities as an asset
  class
• Asset class definitions are fairly arbitrary
• Asset class definitions seem to follow the
  Justice Potter Stewart rule
• Hard to define, but I know it when I see it

6
The Market Portfolio

• Modern finance suggests that an assets return
• Is driven by its covariance with the market
• The Market is supposed to be the sum of all
  asset values

7
Are Commodities an Asset Class?

• Greer (2000) proposed an asset class
  classification scheme
• The independence of valuation methodologies
• Potentially leads to a portfolio diversification
  benefit

8
Are Commodities an Asset Class?

• Kritzmans (1999) four-step asset class test
• Primarily focuses on expected utility and
  correlation

9
Are Commodities an Asset Class?

• Swensens (2000) asset class checklist
• Distinguishes between core and non-core assets

10
Are Commodities an Asset Class?

• A more pragmatic definition of an asset class
• Any investment you are willing to stake your
  reputation on
• Renaissance to Launch New Fund
• By Deborah Brewster in New York, Financial
  Times
• Published July 5 2007 0446
• Renaissance Technologies, one of the
  worlds biggest and best performing hedge fund
  groups, plans to launch a managed futures fund
  with a capacity of 25bn, an unusual foray into
  an investment strategy that has been
  underperforming.
• If it walks like a duck and looks like a duck,
  its probably a duck

11
The Coffee Can Portfolio

• Imagine burying 1000 worth of gold
• in a coffee can in your backyard
• Imagine burying 1000 worth of your favorite
  stocks
• in a coffee can in your backyard
• Retrieve the coffee can, one hundred or two
  hundred years in the future
• Which coffee can portfolio will be worth more?

12
What Affects the Value of Coffee Can Portfolios?

• For a coffee can portfolio, the longer the
  investment time horizon
• The greater the sovereign and bankruptcy risks

13
What is the Composition of the Commodity Market?

• Whatever you want it to be
• Unlike the stock and bond markets, there is no
  natural market cap

14
What is the Expected Return?

• What role should the study of the past play in
  assessing the future?
• It is common to warn of the limitation of naïve
  historical extrapolation
• Past performance is not a guarantee of future
  results
• If past history was all there was to the game,
  the richest people would be librarians
• History is a pack of lies about events that
  never happened told by people who weren't there
• It is also common to caution against ignoring
  history
• The four most dangerous words in investing are
  this time its different
• Those who cannot remember the past are condemned
  to repeat it
• Is the past, or intuition, prologue?

15
Can Commodities Have Equity-Like Returns?

• Yes
• Of course this begs the question
• Whats the forward looking return of equities?
• If equities have low returns in the future
• Then who cares if commodities have equity-like
  returns
• If, instead, one asks Can commodities have
  attractive returns?
• Then the answer is Yes
• Whatever attractive means

16
Commodities may be Equity-like

• Focusing on volatility, commodities may seem
  equity-like

GSCI
SP 500
Long Treasuries
Intermediate Treasuries
T-Bill
17
Naive Historical Extrapolation

• The SP/GSCI index is a broadly followed
  commodity index

18
Naive Historical Extrapolation

• The original GSCI started trading in 1991
• Imagine an investor in 1992 using history to
  estimate future returns
• Subsequent excess returns declined by about 50

19
Naive Historical Extrapolation

• Some assert that more historical data is better
  when forecasting returns
• There is more historical data for stocks than for
  commodities
• Is more data better?

Ibbotson and Sinquefield Stocks, Bonds, Bills
and Inflation January 1976
20
Capital Supply and Demand Historical and Future
Returns

• Historical returns could be high when the supply
  of capital is low
• As soon as someone writes up the discovery of
  past high returns
• It is possible that an increased supply of
  capital will drive down future returns

More Attractive
Historical Return
Future Return
Less Attractive
High
Low
21
Buying High and Selling Low

• Another view is that returns fluctuate around
  some average level
• A challenge is to avoid the Lake Wobegon trap
• And to think that above average is average

Death of Equities August 1979
The Internet Age October 1999
22
Commodity Futures Return

• An investor in commodities typically invests in
  commodity futures/swaps
• This requires buying, and later selling, a
  futures contract

December 31, 2002
January 31, 2003
Step 2 January 2003 Sell March 2003 33.51 Nearby

Step 1 December 2002 Buy March 2003 30.59 Next
Nearby
Futures Return 33.51/30.59
9.54
23
Components of Return

• The return on a futures contract can be
  decomposed into two components
• One component is the spot price return
• The period to period change in the price of the
  nearby futures contract

December 31, 2002
January 31, 2003
January 2003 Nearby contract March 2003 33.51
December 2002 Nearby contract February
2003 31.20
Spot Return 33.51/31.20
7.40
24
Components of Return

• The other component is the roll return
• The passage of time return

December 31, 2002
January 31, 2003
December 2002 February 2003 31.20 Nearby
contract
December 2002 March 2003 30.59 Next nearby
Roll Return 31.20/30.59
1.99
25
Commodity Beta Investing

• Many investors have added broad-based commodity
  exposure to their portfolios
• Many of these investors have chased the commodity
  risk premium
• Many of these investors have learned that index
  exposure
• has a dividend yield that can be positive or
  negative

Roll 20
Roll -20
26
Forward-Looking Expected Return

• When thinking about individual commodities, focus
  on two questions
• What is the expected spot return for a commodity
• What is the expected roll/term structure return
  for a commodity
• When thinking about a portfolio of commodity
  exposures
• Focus on the impact of following a buy-and-hold
  strategy
• Focus on the benefit of rebalancing

27
Inflation and Spot Prices

• Received wisdom says that commodity prices should
  rise with inflation
• But want does this mean?
• Spot prices for commodities may exceed, lag, or
  match inflation
• Spot price returns will also have substantial
  inflation tracking error
• Commodities may be uncorrelated with inflation in
  the short-run
• And, perhaps, cointegrated with inflation in the
  long-run

28
Historical Real Price of Gold

• There is no one right historical time period
• Over the long-run, golds real return has been
  volatile

Great Bullion Famine
29
Relative Price Fluctuations

• Does history shed any light on the long run
  performance of other commodities?
• The price of wheat relative to gold has zigged
  and zagged
• However, the long-run relative price of wheat to
  gold is largely unchanged

Napoleonic Wars
Great Bullion Famine
Black Death
30
Full Carry and Equities

• Full carry is the dominant driver of the price
  of SP 500 futures contracts
• As a result, investors are largely indifferent as
  to which contract to own
• Each futures contract should have the same
  expected return
• Each futures contract should have the same
  expected volatility, and
• All contracts should zig and zag with a
  correlation of approximately 1.0
• Full carry forward prices are not forecasts of
  prices in the future
• Full carry forward prices do not incorporate a
  risk premium

Data as of December 31, 2006
31
Which Return?

• Say an investor wants to invest in the common
  equity of Google
• There is only one security to buy
• If one wants to invest in crude oil, there are
  many contracts one can buy
• When full carry does not hold, it matters which
  contract you invest in

Traditional commodity indices invest in the
next nearby contract
32
Commodities May Be Bond-like

• Bonds have a term structure
• In this chart, the three-month T-Bill is the
  nearby contract
• Few investors think that the three month T-Bill
  is the bond market
• Commodities have a term structure
• As a result, the nearby contract is not the
  market

33
Crude Oil Futures Return and Risk

• Historically, the longer the maturity of a crude
  oil futures contract
• The less volatile the return
• Unlike the full carry case, different
  maturities have different returns

Six Month
Three Month
One Year
Two Years
Average
Three Years
Four Years
Five Years
Nearby
34
Crude Oil Futures Return and Risk

• Looking at geometric returns, not arithmetic
  returns, traditional indices
• Seem to have identified the exposure with the
  largest variance drain
• Geometric Return Arithmetic Return Variance/2

Six Month
One Year
Three Month
Two Years
Average
Three Years
Four Years
Five Years
Nearby
35
Beta Investing and Maturity Selection

• It is easy to illustrate the impact of contract
  maturity selection
• On the performance of a portfolio, such as the
  GSCI

36
Beta Investing and Maturity Selection

• Slicing and dicing returns illustrates the
  historical importance
• Of contract maturity choice

37
The Joy of Rebalancing

• Commodity prices may track inflation in some long
  term sense
• Short-term spot price returns for many
  commodities tend to be largely uncorrelated
• However, long-term prices may be cointegrated
  with one another

38
The Joy of Rebalancing

• Relative to natural gas, sometimes the price of
  nickel has been high
• And sometimes the price of nickel has been
  relatively low
• The price of nickel relative to natural gas has
  been very volatile
• Buy-and-hold portfolio have a tendency to
  overweight expensive assets
• And to underweight cheap assets

75 Nickel 25 Natural Gas
Annualized Relative Price Volatility 69
Ideally, underweight nickel
28 Nickel 72 Natural Gas
Ideally, underweight natural gas
39
The Joy of Rebalancing

• In a mean reverting world, buy-and-hold
  portfolios tend to
• overweight expensive securities and underweight
  inexpensive securities
• Buy-and-hold portfolios, on average, have
  price/value ratios greater than one
• On average, rebalanced portfolios have
  price/value ratios of one

40
The Joy of Rebalancing

• What works for pairs of securities works for more
  broadly diversified portfolios

41
The Joy of Rebalancing

• Rebalancing is a naïve value oriented
  investment strategy

42
Focusing on the Invisible

• Sometimes normal backwardation is mentioned as
  the reason
• That investors should consider investing in
  commodities
• Keynes coined the phrase normal backwardation
  in the 1920s
• Keynes never observed normal backwardation, he
  simply believed in it
• Normal backwardation, like the CAPM, is
  unprovable and unobservable
• The ideas of economists and political
  philosophers, both when they are right and when
  they are wrong, are more powerful than is
  commonly understood. Indeed, the world is ruled
  by little else. Practical men, who believe
  themselves to be quite exempt from any
  intellectual influences, are usually the slaves
  of some defunct economist.
• -John Maynard Keynes
• Nothing sways the gullible more than arguments
  they can't understand
• -Cardinal de Retz

43
More Focusing on the Invisible

• Another popular nebulous concept is the
  convenience yield
• The convenience yield story says that the return
  of a commodity investment
• Is driven by the need of certain agents to hold
  inventory
• In a sense the story seems reasonable
• The convenience yield is similar to the economic
  idea of utility
• However, it is hard to measure ones own utility
• It is quite challenging making an investment
  based on a bet
• That you can measure someone elses
  utility/convenience

44
Real Return Expectations

• A framework for thinking about possible future
  returns

45
Commodities in a Portfolio Context

• It is common to look at commodities as a
  diversifier of
• A stock and bond portfolio
• What if one invests in more than just stocks and
  bonds?

Private Equity
Assume Top Quartile Returns
Emerging Stocks
Real Return
US Stocks Developed Stocks
Absolute Return
Commodities
Bonds
Other than commodities, all return and risk
estimates from the Yale Endowment 2006 annual
report
46
Optimization

• Commodities seem to find a place in
• Lower and moderate risk portfolios

Commodities
Real Assets
Private Equity
Absolute Return
Fixed Income
Emerging Equity
47
Optimization

• What if the volatility of fixed income was 5 per
  year, not 10
• Commodities still seem to find a place in
• Lower and moderate risk portfolios

Commodities
Real Assets
Private Equity
Fixed Income
Absolute Return
Emerging Equity
48
Commodities in a Portfolio Context

• Optimization precisely answers vaguely formed
  questions
• What if all investment opportunities had similar
  Sharpe ratios?

Private Equity
Emerging Stocks
US Stocks Developed Stocks Commodities
Real Return
Absolute Return
Bonds
49
Optimization

• Similar Sharpe ratios, larger commodity allocation

Commodities
Real Assets
Absolute Return
Private Equity
Fixed Income
Emerging Equity
50
What to do After Making the Beta Decision

• The historical performance of the GSCI has
  influenced
• The view many investors have of possible
  commodity returns
• The GSCI was recently sold to Standard and Poors
• Goldman Sachs now focuses on alpha strategies
• How long will it be before other beta index
  providers
• Distance themselves from their beta indices?
• What is the appeal of the alpha story?

51
Overcoming Poor Index Construction

• Currently the hurdle for commodity alpha is
  currently fairly low
• Largely because of the poor design of existing
  commodity beta indices
• Alpha is supposed to be the pay-off
• To value-added investment activity
• Usually alpha comes from investing in a subset of
  a larger universe
• Yet it is a peculiarity of the commodities
  market that alpha exists
• Just from investing in a more broadly diversified
  portfolio

52
Alpha and DispersionJanuary 1999 to May 2007

• The greater the cross-sectional dispersion of
  returns
• The greater the opportunity to produce alpha
• Historically, the dispersion of nearby
  commodity futures returns
• Has been similar to that of stocks

53
Alpha and DispersionJanuary 1999 to May 2007

• There are two sources of nearby cross-sectional
  return dispersion
• Systematic dispersion, and
• Non-systematic dispersion

Data Source Bloomberg, FactSet Research Systems
54
Alpha and Term Structure DispersionJanuary 1999
to May 2007

• There is also a within commodity term structure
  dispersion story

55
Alpha and Term Structure DispersionJanuary 1999
to May 2007

• Some commodities trade in full carry, others
  dont
• Commodities that do not trade in full carry can
• Possibly provide an opportunity to produce alpha

56
Alpha and Term Structure DispersionJanuary 1999
to May 2007

• The nearby contract has been the workhorse of
  commodity indices
• Yet it is not the only choice for investors
• Historically, the higher the cross-sectional term
  structure dispersion
• The higher the average alpha relative to the
  nearby contract

57
Summary

• Are commodities an asset class?
• Yes, unless you want to believe otherwise
• Can commodities have attractive returns?
• Yes
• Are commodities an appealing portfolio
  diversifier?
• Yes
• Are there interesting alpha opportunities?
• Yes