Securities Markets : Interest Rates

Lecture 5

Seunghyup Lee
September 19, 2024

Chung-Ang University
Review

 Is the arithmetic mean or the geometric mean the right way to measure an
investor’s typical return experience? How do they compare?
 If an asset pays off in states where marginal utility is high it will be more valuable,
other things equal. Why is this?
 Investors’ desire to allocate wealth to states with high marginal utility pushes
up the prices for those states.
 Assets that pay off in those states command higher prices.
 Does the history of bills, bonds, stock, and gold returns make sense in this context?
 Stocks pay off when times are good and gold pays off when times are bad,
consistent with their average returns.
 How does the stochastic discount factor (SDF) approach price assets?
 The price of any asset is the average of its payoffs, weighted by state
probabilities and marginal utilities.
 Equivalently, it is the expected product of the SDF and the payoff, where the
SDF is proportional to marginal utility.

1/16
Key Questions

 What is the definition of a forward interest rate?

 Suppose the expectations hypothesis is true and the yield spread is unusually high.
What does this imply for future yields on long-term and short-term bonds?
 Suppose the pure expectations hypothesis is true and the 10-year-ahead 1-year
forward rate is 8%. What is the expected short rate in 10 years?
 How well does the expectations hypothesis work in the data?

2/16
Forward Rate

 Bonds of different maturities can be combined to guarantee an interest rate on a

fixed-income investment to be made in the future. This guaranteed interest rate is
called a forward rate.
 The m-period-ahead 1-period forward rate is
 The rate guaranteed at time t
 for a one period investment to be made at time t + m,
 which pays at time t + m + 1.
 Consider the investment strategy that locks in a rate in the future
 Buy one unit of (m + 1)-period bond: Costs Pm+1,t now and pays 1 at time
t + m + 1.
 P
Sell m+1,t
Pm,t
P
units of m-period bonds: Receive −Pm,t m+1,t
Pm,t
P
now and pay m+1,t
Pm,t
at
time t + m.
 The forward rate is then

1 (1 + Ym+1,t )m+1
1 + Fm,t = =
Pm+1,t /Pm,t (1 + Ym,t )m

3/16
Forward Rate

 This formula is easier to use if we take logs. The m-period ahead 1-period log
forward rate is:

fm,t = pm,t − pm+1,t

= (m + 1)ym+1,t − mym,t
= ym,t + (m + 1)(ym+1,t − ym,t )

 The above equation illustrates that the forward rate curve lies above the yield curve
when the yield curve is upward-sloping, and below it when the yield curve is
downward sloping.
 The two curves intersect when the yield curve is flat.

4/16
Forward Rate

Zero-Coupon Yield and Forward Rate Curves on 9/19/2022

Annual percentage points for each maturity in years.

5/16
Expectation Hypothesis

 Although longer-term bonds may have higher yields, this does not mean that they
have higher returns over any fixed holding period. The expectations hypothesis
says that bonds of all maturities have the same expected returns.
 Pure Expectation Hypothesis Exactly the same return
 Expectation Hypothesis Different only by a constant (that remains the same over
time).

 If the yield curve is steep, you know you can earn a higher annualized return
holding a 30-year bond for 30 years than holding a 1-year bond for 1 year. But it
does not give us a fair comparison.
 But it’s possible that holding a 30-year bond and selling it after 1 year gives you the
same return as holding a 1-year bond for 1 year (short-term view).
 It’s also possible that rolling over 1-year bonds for 30 years gives you the same
return as holding a single 30-year bond for 30 years (long-term view).

6/16
PEH: Short-Term View

 How can a long-term bond have the same expected return, over one period, as a
short-term bond with a lower yield? It’s easiest to do the calculation in logs.
 Example
 Suppose an investor faces 4% log yield on a 1-year bond and 7% log yield on a
30-year bond.
 Log holding period return for the 1-year bond is 4%.
 Log holding period return for the 30-year bond is

y30 − 29 × (y29,next year − y30 )

|{z} | {z }
This is 7% Expectation needs to be ≃ 0.1%

 Thus, PEH holds if the yield on the 30-year bond is expected to rise from 7% to
approximately 7.1% over the next year.
 This will generate a capital loss on the 30-year bond just enough to offset the
higher initial yield.

7/16
PEH: Short-Term View

 PEH: 1-period log holding returns are equal in expectation:

y1t = Et [rm,t+1 ] = ym,t − (m − 1)Et [ym−1,t+1 − ym,t ]

 This can be rearranged as:

ym,t − y1,t = (m − 1)Et [ym−1,t+1 − ym,t ]

Short Term View The yield spread forecasts short-run changes in the yield of the
long-term bonds.

 A high yield spread must predict an increase in the yield on the long bond, just
enough to generate a capital loss to offset that bond’s higher yield.
 This seems counterintuitive: When the long bond yield is already high, it is
expected to rise further!
 Does this mean that the yield spread just keeps getting bigger and bigger?
 To resolve the puzzle, let’s also consider the long-term view.

8/16
PEH: Long-Term View

 Consider the above example with 4% log yield on a 1-year bond and 7% log yield
on a 30-year bond.
 This time focus on 30-year returns.
 Log 30-year holding period return for the 30-year bond is 7%.
 Log 30-year holding period return for the 1-year bonds:

4% + 29 × (Average of expected short rates over years 1 to 29)

30

 Thus, the PEH holds if the future short-rates are expected to be 7%×30−4%
29
≃ 7.1%
on average.
 This is higher than the long rate of 7%, and much higher than the current short-rate
of 4%.
 This expected increase in future short rates ensures that both strategies are equally
attractive.

9/16
PEH: Long-Term View

 PEH : m-period log holding returns are equal in expectation:

y1,m + Et [y1,t+1 ] + · · · + Et [y1,t+m−1 ]

ym,t =
m

 This can be rearranged as

"m−1 #
1 X
ym,t − y1,t = Et (y1,t+s − y1,t )
m
s=0

Long Term View The yield spread forecasts long-run changes in yields on short-term
bonds.

 According to the PEH, a high yield spread:

 Predicts a small near-term increase in the long bond yield (from 7% to 7.1% in
the example)
 A larger longer-term increase in the short yield (from 4% to an average of 7.1%
in the example)
 In this way, the yield spread will return to normal, resolving the earlier puzzle.
10/16
PEH and EH

 The more general version of the EH allows for a constant difference between
expected returns of short-term and long-term investments (i.e., it allows for a risk
premium that remains constant over time).
 But it generates predictions similar to the PEH.
 Example: The 30-year bond might have an average return 3% higher than the
1-year bond.
 Then, a yield spread of 3% does not require any expected changes in interest
rates.
 But an unusually high yield spread (i.e., higher than 3%) generates the same
predictions as above.
 Thus, for some empirical tests, the distinction between PEH and EH is not that
important.

11/16
PEH : Forward Rate View

 A third equivalent way to state the PEH is that it says that forward rates equal
expected future short rates.
 Consider the example: 30-year bond has log yield 7%, 1-year bond has yield 4%.
 Suppose also that the 29-year bond has log yield 6.9%. This implies that the
log forward rate is:

f29 = 30 × 7% − 29 × 6.9% = 9.9%

 Applying the long-term view twice (for years 0-30 and for years 0-29), the
forward rate must also satisfy:

f29 = (sum of expected short yields for years 0 − 29)

− (sum of expected short yields for years 0 − 28)
= (expected short yield for year 29)

 In general, fm,t = ym+1,t + (m + 1)(ym+1,t − ym,t ) = Et [y1,t+m ]

 Interpretation Under PEH, forward rates capture the market’s prediction of future
short rates.
12/16
Empirical Evidence

 For any bonds with maturity m, the holding period return and excess return are

rm,t+1 = pm−1,t+1 − pm,t = −(m − 1)ym−1,t+1 + mym,t = ym,t − (m − 1)(ym−1,t+1 − ym,t )

rm,t+1 − y1,t = (ym,t − y1,t ) − (m − 1)(ym−1,t+1 − ym,t )

 Note that expected excess returns are always zero for any m under PEH, because
ym,t − y1,t = (m − 1)Et [ym−1,t+1 − ym,t ].

13/16
Empirical Evidence

 Regressions of short-run changes in long yield (short-term view) and long-run

changes in short yield (long-term view) on yield spread.

ym−1,t+1 − ym,t = α + βm (ym,t − y1,t )/(m − 1)

X
m−1
y1,t+s /(m − 1) − y1,t = α + γm ((m − 1)/m)(ym,t − y1,t )
s=1

 When the yield spread is unusually high, the yield on the long-term bond tends to
fall, not rise as predicted by the EH.
 When the yield spread is unusually high, short-term interest rates do tend to rise,
but not as much as predicted by the EH.
14/16
Empirical Evidence

 The short-term view fails disastrously, but the long-term view fails modestly. When
the yield curve is steep, riding the yield curve!
 Long yields are generally above short yields. When the pattern reverts, this is
usually followed by falling interest rates and a recession, consistent with the
long-term view of the EH.

15/16
Empirical Evidence

 The empirical tests suggest an eclectic view.

 Some movements in long-term interest rates are driven by expected future
movements in short rates.
 Others are driven by changes in the perceived risk of long-term bonds (rational
view) and/or mispricings (behavioral view).

16/16