Chapter 5.

Hedging with Forwards and Futures

Rangarajan K. Sundaram

Stern School of Business

New York University

Derivatives: Principles & Practice 1 Rangarajan

c K. Sundaram
Outline

Introduction

The Minimum Variance Hedging Strategy

Examples

Implementation

Conclusion

Derivatives: Principles & Practice 2 Rangarajan

c K. Sundaram
Introduction

I Hedging is perhaps the most important function served by futures and

forward contracts.
I Hedging: Reduction of risk in cash flows associated with market
commitments.
I Both forwards and futures can be used for hedging cash flow risk.
I However, standardization creates some problems in using futures contracts
(and perhaps also forward contracts).

Derivatives: Principles & Practice 3 Rangarajan

c K. Sundaram
A Motivating Example

I It is currently January and a gold-wire manufacturer anticipates a need for

10,000 oz. of gold in April.
I Hedging strategy:
1. Go long 10,000 oz. of April gold futures on COMEX.
Current futures price for April delivery: F .
2. Close out the futures position in April.
Futures price at close-out: FT .
3. Buy 10,000 oz. spot in April.
Spot price in April: ST .
I Net cash outflow from the strategy:

10, 000 ST − 10, 000 (FT − F ) = 10, 000 F − 10, 000 (FT − ST ).
I But futures contract is at maturity in April, so FT = ST .
I Net cash flow = 10, 000 F , a certainty amount, so hedge is “perfect.”

Derivatives: Principles & Practice 4 Rangarajan

c K. Sundaram
The Basis

I The term basis refers to the futures-spot differential F − S.

I In the example, the “perfectness” of the hedge depends on the time-T
basis being zero with certainty.
I At least two reasons why, in general, the basis may be risky and non-zero.
1. Commodity mismatch: The grade of the commodity underlying the
futures contract may not be the one being hedged.
I This is commodity basis risk.
2. Delivery mismatch: The standardized maturity dates of the futures
contract may not match the desired hedging dates (e.g., there is no
COMEX gold contract expiring in May).
I This is delivery basis risk.
I In the presence of basis risk, a perfect hedge is impossible.
I So what is the best we can do?

Derivatives: Principles & Practice 5 Rangarajan

c K. Sundaram
Basis Risk in Forward Contracts

I Basis risk can also arise with forward contracts.

I Cross-hedging: Hedging strategy where forward contract on one
underlying (e.g., USD/EUR) is used to hedge the spot exposure on
another underlying (e.g., USD/CZK).
I Typically used because there is no active forward contract on the
underlying risk being hedged (the Czech koruna in this case) so a
forward on a closely-related asset (the euro) is used instead.
I Commodity basis risk is obviously present here.
I So again, what is the best hedging strategy to be followed?

Derivatives: Principles & Practice 6 Rangarajan

c K. Sundaram
Handling Basis Risk: The Questions

I First, what do we mean by the “best” hedging strategy?

I The one that minimizes cash-flow variance.
I Three questions in implementing a hedge:
1. Choice of futures contract?
2. Size of the futures position to be taken in that contract?
3. Long or short?

Derivatives: Principles & Practice 7 Rangarajan

c K. Sundaram
Some Mathematical Preliminaries

I Let X and Y be random variables with variances σX2 and σY2 , respectively.
I Let E (·) denote the expectation of a random variable.
I We have the following definitions:
1. The covariance of X and Y , denoted Cov(X , Y ), is given by
Cov(X , Y ) = E (XY ) − E (X )E (Y ).
2. The correlation between X and Y , denoted Corr(X , Y ), is given by
Cov(X , Y )
Corr(X , Y ) = .
σX σY
3. For any constants a and b, the variance of (aX − bY ) is

Var(aX − bY ) = a2 σX2 + b 2 σY2 − 2ab Cov (X , Y ).

Derivatives: Principles & Practice 8 Rangarajan

c K. Sundaram
Outline

Introduction

The Minimum Variance Hedging Strategy

Examples

Implementation

Conclusion

Derivatives: Principles & Practice 9 Rangarajan

c K. Sundaram
Notation

I Assume a futures contract has already been chosen for hedging purposes.
I We discuss later the criteria that should govern this choice.
I The first step in the analysis is to obtain an expression for the cash flow
from a hedged position.
I Notation:
I S: spot price of asset being hedged.
I F : current futures price.
I T : date of market commitment.
I Q: size of market commitment.
I H: size of futures position.
I ST : time-T price of asset being hedged.
I FT : time-T futures price.

Derivatives: Principles & Practice 10 Rangarajan

c K. Sundaram
Comments on the Notation

I The asset being hedged and the asset underlying the futures contract
need not be the same.
I That is, there could be commodity basis risk.
I The date of spot market commitment need not be the maturity date of
futures contract.
I That is, there could be delivery basis risk.
I Therefore, we need not have ST = FT .
I Note also that H and Q need not be the same. Indeed, the central
question of interest we are looking to address is:
I What is the variance-minimizing hedge ratio h∗ = H/Q?

Derivatives: Principles & Practice 11 Rangarajan

c K. Sundaram
Cash Flow from Hedging a Long Spot Commitment

I To hedge a commitment to buy at T , the strategy is:

1. Long futures position of size H at futures price F .
2. Close out futures at date T at price FT .
3. Buy Q units spot at date T at price ST .
I Note that H < 0 is not precluded (futures may be a short position).
I Cash flows:
I Outflow of QST from spot purchase.
I Inflow of H(FT − F ) from futures resettlement.
I Net cash outflow: QST − H(FT − F ).
I Thus, investor with long spot commitment wishes to minimize variance of

QST − H(FT − F ).

Derivatives: Principles & Practice 12 Rangarajan

c K. Sundaram
Cash Flow from Hedging a Short Spot Commitment

I To hedge a commitment to sell at T , the strategy is:

1. Short futures position of size H at futures price F .
2. Close out futures at date T at price FT .
3. Sell Q units spot at date T at price ST .
I Again, H < 0 is not precluded (futures may be a long position).
I Cash flows:
I Inflow of QST from spot sale.
I Outflow of H(FT − F ) from futures resettlement.
I Net cash inflow: QST − H(FT − F ).
I Thus, investor with short spot commitment wishes to minimize variance of

QST − H(FT − F ).

Derivatives: Principles & Practice 13 Rangarajan

c K. Sundaram
Hedged Cash Flow: Summing Up . . .

I In either case, the objective is to minimize the variance of

QST − H(FT − F ).
I We examine this minimization problem in two steps:
I When there is no basis risk.
I When basis risk may be present.

Derivatives: Principles & Practice 14 Rangarajan

c K. Sundaram
The Case of No Basis Risk

I No basis risk implies:

I No commodity basis risk.
I No delivery basis risk.
I Therefore, we must have ST = FT .
I If ST = FT , the cash flow from the hedged position is

QST − H(FT − F ) = ST (Q − H) + HF .
I Only unknown at inception: ST .
I Solution: Set H = Q. Then, the first term drops out, leaving the certainty
cash flow HF = QF .
I Minimized variance = 0. Cannot be improved.

Derivatives: Principles & Practice 15 Rangarajan

c K. Sundaram
No Basis Risk: Summary

I In summary, if there is no basis risk:

I The minimum variance hedge ratio is h∗ = 1.
I The variance of cash flows under the optimal hedging strategy is
zero.

Derivatives: Principles & Practice 16 Rangarajan

c K. Sundaram
The Case of Basis Risk

I Need no longer be true that ST = FT . So what do we do?

I First, we express the hedged cash flow in terms of price changes.
I Add and subtract Q S to the cash flow: we obtain

QST − QS + QS − H(FT − F ) = QS + Q(ST − S) − H(FT − F ).

I Let
I ∆S = ST − S denote the change in spot prices.
I ∆F = FT − F denote the change in futures prices.
I Then, the hedged cash flow may be written as

QS + Q∆S − H∆F .
I Letting h = H/Q denote the hedge ratio, we can reexpress this as

QS + Q(∆S − h∆F ).

Derivatives: Principles & Practice 17 Rangarajan

c K. Sundaram
The Variance of the Hedged Cash Flow

I Let
I σ 2 (∆S ) denote the variance of ∆S .
I σ 2 (∆F ) denote the variance of ∆F .
I Cov(∆S , ∆F ) denote the covariance of ∆S and ∆F .
I ρ denote the correlation between ∆S and ∆F .
I The hedged cash flow is QS + Q(∆S − h∆F ).
I The variance of this cash flow is:
h i
Q 2 σ 2 (∆S ) + h2 σ 2 (∆F ) − 2h Cov(∆S , ∆F ) .

I This is the variance we are to minimize.

Derivatives: Principles & Practice 18 Rangarajan

c K. Sundaram
Minimizing the Variance of the Hedged Cash Flow

I We are to minimize the variance of

h i
Q 2 σ 2 (∆S ) + h2 σ 2 (∆F ) − 2h Cov(∆S , ∆F ) .

I To minimize the variance, we

I Take the derivative of this cash flow with respect to h.
I Set the derivative equal to zero.
I This gives us:
h i
Q 2 2h σ 2 (∆F ) − 2 Cov(∆S , ∆F ) = 0.

and so

h σ 2 (∆F ) − Cov(∆S , ∆F ) = 0.

Derivatives: Principles & Practice 19 Rangarajan

c K. Sundaram
The Minimum Variance Hedge Ratio

I Therefore, the minimum-variance hedge ratio is

Cov(∆S , ∆F )
h∗ =
σ 2 (∆F )

I We can also express h∗ in terms of the correlation between ∆S and ∆F .

I Since Cov(∆S , ∆F ) = ρσ(∆S )σ(∆F ), we can write

σ(∆S )
h∗ = ρ
σ(∆F )

Derivatives: Principles & Practice 20 Rangarajan

c K. Sundaram
The Role of Correlation - I

I Note that, ceteris paribus, the minimum variance hedge ratio increases
with |ρ|.
I This is intuitive:
I If ρ = 0, the futures contract only adds another source of
uncertainty. Minimizing cash flow variance mandates a hedge ratio of
zero, i.e., no use of the futures contract.
I If |ρ| = 1, futures and spot move in lockstep, so one can “fully”
hedge spot price movements with futures contracts.
I In general, as |ρ| increases, we should use a greater hedge ratio to
exploit the greater ability to offset spot cash flow risk with futures.

Derivatives: Principles & Practice 21 Rangarajan

c K. Sundaram
The Role of Correlation - I

Var(CF): Low Corr

Cash Flow Variance Var(CF): High Corr
Var(CF): Perfect Corr

Hedge Ratio

Derivatives: Principles & Practice 22 Rangarajan

c K. Sundaram
The Role of Correlation - II

I The sign of ρ determines the sign of h∗ .

I If ρ > 0, then h∗ > 0. That is:
I A long spot exposure is hedged by a long futures position
I A short spot exposure is hedged by a short futures position.
I If ρ < 0, then h∗ < 0. That is:
I A long spot exposure is hedged by a short futures position
I A short spot exposure is hedged by a long futures position.
I Intuition?

Derivatives: Principles & Practice 23 Rangarajan

c K. Sundaram
The Role of Correlation - III: Minimized Cash-Flow
Variance

I What is the variance of the hedged position using the hedge ratio h∗ ?
I Substituting h∗ into the expression for cash flow variance of the hedged
position, and simplifying, we obtain

Q 2 σ 2 (∆S )(1 − ρ2 ).
I This is the lowest possible value for for cash flow variance.
I This variance depends on ρ2 , decreasing as ρ2 increases. In particular:
I Minimized cash flow variance depends only on |ρ|, not on the sign of
the correlation.
I It is zero in general only if ρ2 = 0, i.e., |ρ| = ±1.
I That is, it is zero only if futures and spot are perfectly correlated, a
situation that holds only when basis risk is not present.
I Put differently, if basis risk is present, there is always some residual
cash flow uncertainty even after hedging.

Derivatives: Principles & Practice 24 Rangarajan

c K. Sundaram
The Role of Correlation - IV: Choice of Futures Contract

I Since the minimized cash flow variance depends only on |ρ|, the choice of
futures contract is simple:
I Ceteris paribus, pick the contract for which |ρ| is maximal.
I This will minimize cash flow variance from the hedged position.

Derivatives: Principles & Practice 25 Rangarajan

c K. Sundaram
How Much Uncertainty is Removed?

I The minimized cash flow variance from optimal hedging is

Q 2 Σ2 (∆S )(1 − ρ2 ).
I The cash flow variance from not hedging at all (h = 0) is

Q 2 σ 2 (∆S ).
I So optimal hedging reduces cash flow variance by a factor of ρ2 .

Correlation Variance Removed

0.90 81%
0.80 64%
0.50 25%
0.20 4%

Derivatives: Principles & Practice 26 Rangarajan

c K. Sundaram
Hedging One-for-One

I What happens if, ignoring the imperfect correlation, we use h = 1?

I If h = 1, the hedged cash flow variance is

Q 2 σ 2 (∆S )(1 − ρ2 ) + Q 2 [σ(∆F ) − ρσ(∆S )]2

I Thus, using h = 1 instead of the optimal h∗ increases cash-flow variance
by

Q 2 [σ(∆F ) − ρσ(∆S )]2

I Indeed, hedging one-for-one may be worse than not hedging at all! This
happens if
σ(∆F )
> 2ρ.
σ(∆S )

Derivatives: Principles & Practice 27 Rangarajan

c K. Sundaram
Outline

Introduction

The Minimum Variance Hedging Strategy

Examples

Implementation

Conclusion

Derivatives: Principles & Practice 28 Rangarajan

c K. Sundaram
Example 1: Cross-Hedging with Currencies

I Setting: US exporter
I Will receive NOK 25 million in thee months.
I Decides to use USD/EUR forward contract to hedge.
I Basis risk exists: we are hedging USD/NOK exchange rates with
USD/EUR forwards:
I Spot asset: NOK.
I Asset underlying forward contract: EUR.
I Need variance-covariance information to determine optimal hedge.
I Suppose we are given:

σ(∆S ) = 0.005, σ(∆F ) = 0.025, ρ = 0.85.

Derivatives: Principles & Practice 29 Rangarajan

c K. Sundaram
Example 1: The Minimum-Variance Hedge Ratio

I The minimum-variance hedge ratio is

0.005
h∗ = 0.85 × = 0.17.
(0.025
I That is, to hedge the exposure of NOK 25 million, we need forward
contracts calling for the delivery of

EUR(0.17 × 25, 000, 000) = EUR4, 250, 000.

I Short this be a short or long forward position?
I Why is the minimum variance hedge ratio so “small” (i.e., only 0.17?).

Derivatives: Principles & Practice 30 Rangarajan

c K. Sundaram
Example 2: Cross-Hedging with Equities

I Data:
1. Underlying exposure: S$P 100 portfolio worth $80,000,000.
2. Current level of the S&P 100 index is 800.
3. Futures contract used for hedging: S&P 500 index futures.
4. Current level of S&P 500 index futures: 960.
5. One S&P 500 futures contract is for 250 times the index.
I There is basis risk:
I Underlying asset: S&P 100 index.
I Given S&P 100 index level of 800, it is as if we own 100,000 units of
the index.
I Futures contract: S&P 500 index futures.
I One futures contract calls for delivery of 250 units of S&P 500 index.
I So suppose we are also given:

σ(∆S ) = 60, σ(∆F ) = 75, ρ = 0.90.

Derivatives: Principles & Practice 31 Rangarajan

c K. Sundaram
Example 2: The Minimum-Variance Hedge Ratio

I Then, the minimum-variance hedge ratio is

„ «
σS 60
h∗ = ρ = (0.90) = 0.72.
σF 75
I Since our exposure is 100,000 units of the spot asset, we should hedge
with

0.72 × 100, 000 = 72, 000

units of the futures contract.

I Since one futures contract is for 250 units of the index,
72, 000
No. of futures contracts = = 288.
250
I Should this be a long or short futures position?

Derivatives: Principles & Practice 32 Rangarajan

c K. Sundaram
Outline

Introduction

The Minimum Variance Hedging Strategy

Examples

Implementation

Conclusion

Derivatives: Principles & Practice 33 Rangarajan

c K. Sundaram
Identifying h∗ from Historical Data - I
I Suppose we have information on daily price changes:
I σ 2 (δS ): Variance of daily spot price changes δS .
I σ 2 (δF ): Variance of daily futures price changes δF .
I ρ(δS , δF ): Correlation between δS and δF .
I Suppose also that
I Price changes are i.i.d.
I There are K days in the hedging horizon.
I Since the price change over K days is the sum of the price changes over
each day, we have

σ 2 (∆S ) = K σ 2 (δS ), σ 2 (∆F ) = K σ 2 (δF ), ρ = ρ(δS , δF )

I So the minimum variance hedge ratio is identified from daily price-change
data as
σ(δS )
h∗ = ρ(δS , δF )
σ(δF )

Derivatives: Principles & Practice 34 Rangarajan

c K. Sundaram
Identifying h∗ from Historical Data - II: Regression

I An alternative way to identify h∗ from historical data on daily price

changes is to run the regression

δS = a + bδF + 
I Let b
a and b
b be the regression estimates. Then, we have

h∗ = b
b.
I That is, the minimum-variance hedge ratio is precisely the regression
estimate bb!
I Intuition?

Derivatives: Principles & Practice 35 Rangarajan

c K. Sundaram
Hedging with Multiple Futures Contracts

I No reason for using only one futures contract to hedge a given spot risk.
I E.g., a portfolio of junk bonds may be better hedged by a
combination of equity futures and interest-rate futures than by either
alone.
I Let F1 , F2 , . . . , Fn denote the n futures contracts to be used for hedging.
I Consider the regression

δS = a + b1 δF1 + b2 δF2 + · · · + bn δFn

I Then, the regression estimates b

b1 , . . . , b
bn are precisely the
minimum-variance hedge ratios h1∗ , . . . , hn∗ associated with the n futures
contracts.
I Should we use more than one futures contract in general?

Derivatives: Principles & Practice 36 Rangarajan

c K. Sundaram
Hedging Multiple Risks with a Given Futures Contract

I Consider a firm exporting to areas with several different currencies and

using a single futures contract to hedge all of the currency risks. How
should it choose the total hedge size.
I Simple:
1. Compute the minimum-variance hedge size for each risk separately.
2. Add them all up.
The result is the minimum-variance hedging strategy for the portfolio of
risks.

Derivatives: Principles & Practice 37 Rangarajan

c K. Sundaram
Hedging with Futures: Resettlement Cash Flows

I Thus far, we have treated the futures contract as if it is a forward

contract, i.e., it is marked-to-market only once at the end of the contract.
I What happens if we take daily marking-to-market into account?
I Let K be the number of days in the hedging horizon.
I Let R denote the per-day gross rate of interest.
I Then, the adjusted optimal hedge ratio is

h∗∗ = g (R, K ) × h∗ ,

where
1. h∗ is the optimal hedge ratio ignoring daily resettlement, and
2. g (R, K ) is a “tail factor” given by
1 + R + R2 + · · · + RK
g (R, K ) =
1 + R 2 + R 4 + · · · + R 2K

Derivatives: Principles & Practice 38 Rangarajan

c K. Sundaram
Tailing the Hedge

I How important is tailing the hedge?

I For an example, suppose interest rates are 5% on an annualized basis.
I Then, the tail factor is given by
I 0.9993, if K = 10
I 0.993, if K = 100
I 0.932, if K = 1, 000
I In practice, tailing the hedge is important for long horizons (or very high
interest rates), but the tail factor is to small to make a practical difference
over short horizons.

Derivatives: Principles & Practice 39 Rangarajan

c K. Sundaram
Outline

Introduction

The Minimum Variance Hedging Strategy

Examples

Implementation

Conclusion

Derivatives: Principles & Practice 40 Rangarajan

c K. Sundaram
Conclusion

I Basis risk arises when the futures/forward contract used to hedge a spot
exposure is not perfectly matched with the spot exposure.
I Commodity basis risk: Arises from commodity mismatch.
I Delivery basis risk: Arises from delivery date mismatch.
I This chapter has examined hedging with futures and forwards in such
circumstances.
I The optimal (i.e., cash flow variance minimizing) strategy is one that
takes the correlation between spot and futures price changes into account,
increasing in size as this correlation increases.
I We have also examined the implementation of this strategy using daily
price data, and several extensions of the strategy to cover
I Hedging multiple risks simultaneously.
I Hedging a single risk with multiple futures contracts.
I “Tailing” the hedge to take into account daily resettlement in futures
markets.

Derivatives: Principles & Practice 41 Rangarajan

c K. Sundaram