Vol.

2, 2008-28
September 19, 2008

Special Issue "Using Econometrics for Assessing Economic Models"

Editor: Katarina Juselius

Level, Slope, Curvature: Characterising the Yield Curve

in a Cointegrated VAR Model

Julia V. Giese
Nuffield College, University of Oxford

Abstract:
Empirical evidence on the expectations hypothesis of the term structure is in-conclusive and its
validity widely debated. Using a cointegrated VAR model of US treasury yields, this paper extends a
common approach to test the theory. If, as we find, spreads between two yields are non-stationary, the
expectations hypothesis fails. However, we present evidence that differences between two spreads are
stationary. This suggests that the curvature of the yield curve may be a more meaningful indicator of
expected future interest rates than the slope. Furthermore, we characterise level and slope by deriving
the common trends inherent in the cointegrated VAR, and establish feedback patterns between them
and the macroeconomy.

JEL: C32, E43, E44

Keywords: Yield curve; term structure of interest rates; expectations hypothesis; cointegration;
common trends

Correspondence: Julia Giese, Nuffield College, University of Oxford,Oxford, OX1 1NF, U.K.
[email protected];
The author thanks David Hendry, Katarina Juselius and Bent Nielsen for the invaluable advice and
guidance extended. In addition, thanks are due to the two anonymous referees, Heino Bohn Nielsen,
Søren Johansen, Jeremy Large, and participants at seminars and conferences for providing helpful
comments. Financial support from the ESRC (PTA-030-2003-00194), Nuffield College and the
Studienstiftung des deutschen Volkes is gratefully acknowledged.

www.economics-ejournal.org/economics/journalarticles
© Author(s) 2008. This work is licensed under a Creative Commons License - Attribution-NonCommercial 2.0 Germany
2 Economics: The Open-Access, Open-Assessment E-Journal

1 Introduction
Investigating relations between yields of different maturities was one of the first
applications of cointegration analysis (see, for example, Engle and Granger (1987),
or Campbell and Shiller (1987)). The initial bivariate approach was extended to
the multivariate case by Hall, Anderson, and Granger (1992) among others. The
study concentrated on a set of short-term maturities and found one common trend.
This was believed to corroborate the expectations theory of the term structure,
which says that a longer-term bond rate is just the average of expected one-period
rates for the duration of the bond plus some constant term premium. According to
this hypothesis, the spreads between different maturities make up the cointegrating
vectors. Given only one common trend, it was concluded that the term premia
therefore must be mean-reverting if not constant.
Shea (1992) examined a broader set of yields, including long-term maturities
up to twenty-five years. His results support the findings of Hall, Anderson, and
Granger (1992) for the short end of the yield curve but reject stationarity of the
spreads between longer-term maturities. Building on Shea (1992), other researchers
find up to three common trends when including yields of longer maturities. Zhang
(1993) demonstrates this for US data while Carstensen (2003) looks at German data.
They argue that their findings suggest that term premia are in fact non-stationary
and that additional common trends have interpretations familiar from factor models
of the yield curve.
This paper seeks to extend past understanding of driving forces behind the yield
curve by employing a cointegrated vector auto-regression (CVAR) model on monthly
data of US treasury zero-coupon yields over the period 1987 to 2000.1 We show that
there is strong evidence for two common trends, implying that not all independent
spreads can be stationary.
However, weighted differences between pairs of spreads are found to be station-
ary, and hence two term premia cointegrate. This suggests that while investors’
preferences with respect to a certain maturity vary over time without reverting back
to a mean, their relative preferences between two maturities are stationary. A con-
clusion from this finding is that we should look at the curvature of the yield curve
(approximated by the weighted difference between two spreads) if we are interested
in the interest rate expectations embodied in the term structure. It enables policy-
makers to deduce whether the rate of change in interest rates is expected to diminish
or increase in the long run compared with the medium run. The finding may also
be interesting for traders trading on mean-reversion properties of the yield curve.
Our analysis of the yield levels’ non-stationary common trends through the
Granger-Johansen representation, introduced in Engle and Granger (1987) and ex-
tended by Johansen (e.g. Johansen (1996)), confirms the results. It is, as far as we
know, a novel application to the term structure of interest rates: both Zhang (1993)
and Carstensen (2003) arrive at their conclusions using factor representations. We
1
The data was kindly provided by Diebold and Li and uses the unsmoothed Fama-Bliss method-
ology to construct the zero-coupon series. See Bliss (1997), and Fama and Bliss (1987), for details.
To my knowledge, this particular series has not been updated to include more recent years. We use
it nonetheless because the way it was constructed is particularly suited for the purposes at hand.

www.economics-ejournal.org
Economics: The Open-Access, Open-Assessment E-Journal 3

find that one common trend acts on the level of the yield curve and the other on
the slope, giving them interpretations of a level and slope factor.
Hence, this paper shows empirically that the common trend analysis is related
to common factor models, often used in financial economics to model the yield
curve. This literature finds up to three factors to be sufficient to explain the yield
curve’s shape, often level, slope and curvature.2 Rather than relying on assessments
of explanatory power, cointegration theory provides powerful and thoroughly un-
derstood methods for doing inference on the number of cointegrating relations and
thus directly on the number of common trends. In addition, the analysis of the
Granger-Johansen representation allows us to characterise the driving forces behind
each common trend, and to link them with macroeconomic variables.
The paper is structured as follows: In Section 2 we develop the theoretical model
based on Hall, Anderson, and Granger (1992). Section 3 introduces the CVAR and
presents the cointegration analysis. Section 4 concludes. The computations were
made using CATS and PcGive.3

2 Theoretical Framework
Define bm t as the yield at time t of a zero-coupon bond with maturity m, m =
1, 2, 3, ... . Similarly, let the forward rate at time t of period j be ftj , giving the
linearised no-arbitrage condition4
m
1 X j
bm = f . (1)
t
m j=1 t

The term premium, ltj , is defined as

¡ ¢
ftj = Et b1t+j−1 + ltj , (2)
and is paid for the release from risk in contracting for a future debt immediately.
P
m
Combining (1) and (2) and setting m1 ltj = Lm
t , we get
j=1

m m
1 X£ ¡ 1 ¢ ¤ 1 X ¡ 1 ¢
bm
t = Et bt+j−1 + ltj = Et bt+j−1 + Lm
t . (3)
m j=1 m j=1

Equation (3) can be interpreted in terms of the expectations hypothesis. Its pure
version would suggest that the term premium, Lm t , is zero, allowing the yield to ma-
turity only to be determined by expectations of future short-term yields. A constant
2
See among others Litterman and Scheinkman (1991), Knez, Litterman, and Scheinkman
(1994), Nelson and Siegel (1987), as well as Duffee (2002), for discussions of different factor mod-
els. The former two only place structure on the factors and not the loadings, e.g. using principal
components. Nelson and Siegel (1987) introduce a latent factor model where the factors are unob-
served but the loadings represent level slope and curvature. The latter uses an affine latent factor
model, imposing a no-arbitrage condition.
3
See Dennis (2006) on CATS and Doornik and Hendry (2007) on PcGive.
4
The relationship in (1) is an approximation derived from taking logs of bm t =
£ 1 2 m m
¤1
(1 + ft )(1 + ft )...(1 + ft ) − 1.

www.economics-ejournal.org
4 Economics: The Open-Access, Open-Assessment E-Journal

term premium is consistent with a less strict interpretation, while a stationary term
premium is more flexible still. We will concentrate exclusively on the latter, most
generous, version of the hypothesis and find evidence in the data to reject even it.
Rearranging Equation (3) gives a convenient representation in terms of the spread
between yields of different maturities which is tested empirically in this paper,
i.e.
m−1
1 1 X ¡ 1 ¢
bm
t − b1t 1
= ( − 1)bt + Et bt+j + Lm
t
m m j=1
1
= Et (∆b1t+1 + ∆b1t+2 + ∆b1t+1 + ∆b1t+3 + ∆b1t+2 + ∆b1t+1 + ...
m
+∆b1t+m−1 + ∆b1t+m−2 + ∆b1t+m−3 + ...) + Lm t
m−1
1 X ¡ ¢
= (m − j) Et ∆b1t+j + Lm t . (4)
m j=1

Since bond yields are well approximated by processes integrated of at most order one
(I(1)), their differences are integrated of order zero (I(0)) and the first term on the
right hand side of Equation (4) is stationary. If the term premium was stationary, we
would expect the spreads to be I(0) because a process determined by two stationary
processes is itself stationary. On the other hand, non-stationary spreads found in
the data would imply a non-stationary term premium.
Extending the framework to weighted differences between spreads, Equation (5)
shows that if we find the spreads to be pairwise cointegrating, the weighted differ-
ences between the term premia of differing maturities have to be stationary:5
¡ n ¢
(bm
t − bn
t ) − c b t − b 1
t =

m−1 n−1
1 X ¡ 1 ¢ 1+cX ¡ ¢
(m − j) Et ∆bt+j − (n − j) Et ∆b1t+j + Lm n
t − (1 + c)Lt , (5)
m j=1 n j=1

where c is a constant weight.

A deviation in a mean-reverting process is informative because we can judge the
observation against its long-run equilibrium while we do not have a point of reference
for a non-stationary process. In the case of non-stationary spreads driven by a
non-stationary term premium, the stationary part of the process is indeterminable.
However, given cointegrated term premia, a deviation from the typical curvature
may well reflect changes in future interest rate expectations, rather than changes in
preferences. In this case, the yield curve’s curvature may be used as an indication
of expectations on the future path of interest rates.
5
A curve going through three points A, B and C with A ≥ B ≥ C, can be described by
a quadratic whose curvature, the second derivative, is given by − 12 ((A − B) − (B − C)) if the
distance between A and B, AB, is the same as that between B and C, BC. However, when
distances between points are not equal, as is the case for maturities considered here, the curvature
1
is the weighted expression M +N (N (A − B) − M (B − C)) where M is AB and N is BC, i.e. a
weighted rather than exact difference between spreads should be characteristic of the curvature of
the yield curve.

www.economics-ejournal.org
Economics: The Open-Access, Open-Assessment E-Journal 5

In the following analysis we show how a CVAR can be used to test the theoretical
models in terms of stationarity and thus assess the expectation theory of the term
structure in its conventional notation and the extension discussed above.

3 A CVAR Model
Our model consists of monthly end-of-period yields for US treasury zero-coupon
bonds of five different maturities, namely for the one-month, three-month, eighteen-
month, four-year and ten-year bonds. The choice of variables reflects the structure
of the yield curve with very short-term as well as medium- and long-term maturities.
Zhang (1993) includes 19 yields of different maturities in the initial analysis. Given
the nature of VAR models, however, a smaller set with still many parameters should
suffice here. Future work should examine the robustness of the results with respect
to the dimension of the system and the choice of maturities included in the analysis.

3.1 Properties of the Data

The subsequent analysis will focus on the time span of Alan Greenspan’s chair-
manship of the Fed from August 1987 to the end of our sample in December 2000.
Samples that reach further back exhibit problems of non-constancy in the parameters
when analysed in a linear CVAR framework. This is in line with Baba, Hendry, and
Starr (1992) who find in their study of US money demand that even risk-adjusted
spreads may change when yields reach previously unknown levels as they did under
the “new operating procedures” in the late 1970s and early 1980s. Giese (2006)
extends the present analysis to a sample beginning in January 1970, by fitting a
Markov-switching model that allows for regime changes. Results from that model
very closely resemble the ones presented here.
A graphical analysis of Figure 1 suggests the zero-coupon series of the one-month,
eighteen-month and ten-year yields differ not so much in levels – although the ten-
year yield is typically higher than yields of short-term maturities – but more in
their differences which show that the long-term yields are less variable. According
to the theory presented above, longer-term yields can at least partly be explained by
the average expected spot interest rates of all periods to maturity and thus contain
information on the shorter end of the yield curve. This aggregation implies that
long maturities are less affected by temporary shocks and that the plots of the long-
term yields look smoother than those of short-term maturities. The bond yields also
appear to be highly persistent, well approximated by I(1) processes.

3.2 Cointegration Analysis

We begin the empirical analysis with a formal definition of the statistical concept
to be used, i.e. the CVAR or vector equilibrium correction model (VECM(k)):
µ ¶
xt−1
∆xt = (Π, µ0 ) +Γ1 ∆xt−1 +...+Γk−1 ∆xt−k+1 +φDt +εt , εt ∼ iidNp (0, Ω),
1

www.economics-ejournal.org
6 Economics: The Open-Access, Open-Assessment E-Journal

10

b1 b18
4 b120

1990 1995 2000

−1
∆b 1 ∆b18
∆b120

1990 1995 2000

Figure 1: Levels and differences of yields for three maturities, 1987:08-2000:12

with p endogenous variables xt = (b1t , b3t , b18 48 120 0 m

t , bt , bt ) where bt represents the yield
at time t of a zero-coupon bond with m months to maturity. The constant µ0 is
restricted to the cointegrating space and Dt is a vector of three impulse dummy
variables to allow for one-off shocks unexplained by the variables in the model and
not reconcilable with the assumption of normality in the residuals. The dummies
take the value one in the months February 1989, December 1990 and May 2000.6
Since yield levels are non-stationary, Π is of reduced rank r, and we can write
(Π, µ0 ) = α(β 0 , β0 ).

3.2.1 Links to the Theoretical Model

The theoretical framework presented in Section 2 is tested by determining r which
defines the numberPof stationary cointegrating relations β 0 xt and non-stationary
0 t
common trends α⊥ i=1 εi , and by testing explicit hypotheses on parameters in these
relations. In our case where p = 5, rank r = 4 is equivalent to four cointegrating
relations and p−r = 1 common trend. This is implied by the expectations hypothesis
in its weakest form: according to Equation (4), a stationary term premium can only
hold if the four independent spreads between five yields are stationary and thus form
cointegrating relations. The studies cited in the Introduction test for this case.
If r = 3, however, there would be two common trends which implies that all
6
Given the length of the sample, we include dummies for standardised residuals which are greater
than 3.5. This value is found by determining the critical value corresponding to the probablity of
having an outlier in a particular sample size.

www.economics-ejournal.org
Economics: The Open-Access, Open-Assessment E-Journal 7

Table 1: LR tests for lag determination and restricted short-run structure

Lag deletions Test statistic p-value

k=5→k=4 χ2 (25) = 17.022 [0.881]
k=5→k=3 χ2 (50) = 85.363 [0.001]
Γ-restrictions ∆b1 ∆b3 ∆b18 ∆b48
∆b120
Γ1 1 1 0 0 0
Γ2 0 0 1 1 0
Γ3 1 1 1 1 1

four spreads cannot be stationary, and the expectations hypothesis even loosely
defined fails. Nevertheless, interesting conclusions can still be drawn, for example,
the weighted differences between two spreads may be stationary, as formulated in
Equation (5). The analysis is thus extended to examine the stationarity of the
yield curve’s derivatives. While the cointegrating relations can be used to assess the
stationarity of the level, slope and curvature, so can the common trends: We find
that they represent the non-stationary derivatives of the yield curve. Thus, in the
case of three common trends (r = 2), even the curvature could be non-stationary.

3.2.2 Specification of the Statistical Model

Before determining the rank r of Π, we choose the lag length k of the model. Nielsen
(2006) shows that the location of the characteristic roots is of no consequence when
testing for the lag order. The choice of lag length k reflects the persistence of short-
run effects, and is determined by a likelihood ratio (LR) test that compares two
hypotheses on different lag lengths according to residual auto-correlation. The test
statistic is given by

LR(Hk |Hk+i ) = −2 ln Q(Hk /Hk+i ) = T (ln |Ω̂k | − ln |Ω̂k+i |), (6)

where Hk is the null hypothesis that there are k lags, while Hk+i is the alternative
hypothesis that k +i lags are needed. Ω̂k is the estimated variance-covariance matrix
of the residuals. The results obtained for our model are presented in Table 1. They
give strong evidence for k = 4, which means that three Γi matrices need to be
estimated in the VECM.
Since k = 4 involves many parameters, we set insignificant columns in the Γi s
to zero. The resulting short-run structure is also shown in Table 1, where an entry
of 1 stands for a fully estimated column, and an entry of 0 for a column with all
entries restricted to zero. These restrictions are rejected in an LR test with a p-
value of 0.001 (χ2 (30) = 59.144), but not at the 1-percent level in an F-test (p-value:
0.012, F(30,518) = 1.705). In addition, the Schwartz, Hannan-Quinn and Akaike
information criteria support the restricted model, and hence we continue with the
restrictions on Γi imposed.
Furthermore, we test the model for mis-specification, in particular whether the
residuals are consistent with the errors behaving according to εt ∼ iid Np (0, Ω). A

www.economics-ejournal.org
8 Economics: The Open-Access, Open-Assessment E-Journal

Table 2: Vector mis-specification tests

Lag Test statistic p-value

LM tests for no auto-correlation:
1 χ2 (25) = 24.444 [0.494]
F(25,484) = 0.977 [0.497]
2 χ2 (50) = 79.368 [0.005]
F(50,573) = 1.483 [0.020]
3 χ2 (75) = 107.75 [0.008]
F(75,578) = 1.329 [0.041]
4 χ2 (100) = 133.57 [0.014]
F(100,565)= 1.234 [0.075]
Test for normality:
χ2 (10) = 24.514 [0.006]
LM tests for no ARCH-effects:
1 χ2 (225) = 261.746 [0.047]
2 χ2 (450) = 474.927 [0.201]
3 χ2 (675) = 700.646 [0.240]
4 χ2 (900) = 989.998 [0.019]

discussion of parameter constancy is left until after the system has been identified.
Table 2 shows that there may be a problem with auto-correlation: the null hypothesis
is rejected for lags 1-2 and 1-3 in the χ2 but not the F form. However, none of the
single-equation tests (not reported) is rejected, suggesting that the problem is not
serious. The vector test for normality, also shown in Table 2, suggests that there
may be a minor problem with normality, but for individual yields normality is not
rejected (test results not shown here). Also, cointegration results have been found to
be quite robust to moderate degrees of excess kurtosis (see Gonzalo (1994)). ARCH
effects can be rejected, broadly supporting the null hypothesis of homoskedasticity.

3.2.3 Determination of the Cointegration Rank

The trace test seeks to determine which eigenvalues correspond to stationary and
which to non-stationary relations. A small eigenvalue indicates a unit root and
thus at least a very persistent and possibly non-stationary process. In Table 3 we
report the test results, where starred trace statistics and p-values are corrected by
the Bartlett factor for small sample size and λr+1 denotes the smallest eigenvalue of
rank r + 1.
Our economic prior of r = 4 is not rejected, implying that there is at least one
eigenvalue – 0.025 – that is not statistically different from zero. However, the next
smallest eigenvalue has a magnitude of only 0.047, which the test also finds to be not
statistically different from zero, supporting r = 3. To investigate further whether
the data include one or two non-stationary trends, we use information from other
indicators.

www.economics-ejournal.org
Economics: The Open-Access, Open-Assessment E-Journal 9

Table 3: Trace test of the cointegration rank

p−r r λr+1 trace trace* 5% cv p-value p-value*

5 0 0.433 184.577 163.903 76.813 0.000 0.000
4 1 0.306 95.453 85.634 53.945 0.000 0.000
3 2 0.156 38.150 32.911 35.070 0.022 0.086
2 3 0.047 11.597 10.224 20.164 0.494 0.624
1 4 0.025 4.015 3.805 9.142 0.422 0.454

Table 4: Modulus of the five largest roots

r=5 r=4 r=3 r=2 r=1

0.970 1.0 1.0 1.0 1.0
0.939 0.939 1.0 1.0 1.0
0.939 0.939 0.878 1.0 1.0
0.735 0.734 0.735 0.879 1.0
0.735 0.734 0.735 0.682 0.731

The roots z of the characteristic polynomial Π(z) = Ip − Π1 z − Π2 z 2 − ... − Πk z k

are shown in Table 4. Three roots are close to one, suggesting r = 2. However,
imposing r = 3 changes the roots and we observe that the 3rd root is now more
moderate. With monthly data a root of 0.88 may well imply slow adjustment rather
than non-stationarity.
Furthermore, we may graphically assess the stationarity of the cointegrating
relations. The relations β̂30 xt and β̂40 xt , shown in Figure 2, support r = 3 since β̂40 xt
appears non-stationary and β̂30 xt stationary.
Finally, coefficients in columns 4 and 5 of the unrestricted α̂ (t-values in brackets)
are largely insignificant, implying that we would not learn anything about adjust-
ment if we included more than three cointegrating relations, while coefficients in the
first three columns contain information on adjustment (t-values ¿ 2.6 in bold face):

 
0.116 0.074 −0.030 −0.016 −0.003
 [7.283] [4.608][−1.873] [−0.978] [−0.164]

 −0.008 0.082 −0.007 −0.021 −0.002
 [−0.617] [6.500] [−0.576] [−1.690] [−0.165] 
 
 0.039 −0.052 −0.041 −0.010
α̂ = −0.053
u
[1.989] .
 [−2.728] [−2.652] [−2.106] [−0.520] 
−0.056 0.038 −0.060 −0.036 −0.022
 [−2.618] [1.766]

 [−2.797] [−1.676] [−1.040] 
−0.022 0.010 −0.028 −0.034 −0.028
[−1.109] [0.495] [−1.428] [−1.766] [−1.446]

Overall, the results suggest r = 3, or p − r = 2 common trends. Not all four

spreads can form cointegrating relations and the term premium is non-stationary.

www.economics-ejournal.org
10 Economics: The Open-Access, Open-Assessment E-Journal

β^3’x t

0.2

0.0

−0.2

1990 1995 2000

1.0
β^4’x t
0.5

0.0

−0.5

−1.0

−1.5
1990 1995 2000

Figure 2: Plots of the estimated βˆ30 xt and βˆ40 xt

3.2.4 Identification of the Long-Run Structure

In the unrestricted model, α and β 0 are not uniquely identified but their product Π
is. Given r = 3, we firstly test some non-identifying hypotheses for the structures
of β and α, and secondly impose over-identifying restrictions on β.
Table 5 shows that the spreads including a constant are far from stationary
except for the spread between the two short-term yields.7 The restrictions in each
case are formulated as
Hc (r) : β c = (H1 ϕ1 , ψ),
where ψ and ϕ1 are (p + 1) × (r − 1) and 2 × 1 matrices, respectively, of unrestricted
estimates, and H1 a (p + 1) × 2 known matrix of the form (for the short-term spread)
µ ¶
0 1 −1 0 0 0 0
H = .
0 0 0 0 0 1
The restrictions are tested for each spread using an LR test, and the results are
supported by Figure 3.
If a variable is not equilibrium correcting, it is “weakly exogenous with respect
to β”. The hypothesis of a zero row in α is given by
Hαc (r) : α = Hαc or equivalently Hαc (r) : R0 α = 0 (7)
7
Not only are spreads with respect to the one-month yield not stationary, but also spreads
between any other combination of yields with different maturities (not shown). Moreover, non-
exact spreads, i.e. where (1, -1) is not imposed but coefficients are estimated, are also found to be
non-stationary implying that they share more than one common stochastic trend.

www.economics-ejournal.org
Economics: The Open-Access, Open-Assessment E-Journal 11

1.5 b 3−b 1 ∆(b 3−b 1)

1

0.5 0
−1
1990 1995 2000 1990 1995 2000
4 b18 −b 1 ∆(b18 −b 1)
1
2
0
0 −1
1990 1995 2000 1990 1995 2000
b48 −b 1 ∆(b48 −b 1)
4 1

2 0
−1
0
1990 1995 2000 1990 1995 2000
5.0 b120 −b 1 ∆(b120 −b 1)
1
2.5 0

0.0 −1

1990 1995 2000 1990 1995 2000

Figure 3: Levels and differences of yield spreads, 1987:08-2000:12

Table 5: Hypothesis tests on β and α for r = 3, [p-value]

Tests on β b3t − b1t b18

t − bt
1
b48
t − bt
1
b120
t − b1t
Stationary spreads, χ2 (4) 4.791 20.983 21.668 21.971
[0.310] [0.000] [0.000] [0.000]
Tests on α b1t b3t b18
t b48
t b120
t
Weak exogeneity, χ2 (r) 56.375 32.694 14.346 14.156 2.888
[0.000] [0.000] [0.003] [0.003] [0.409]
2
Unit vector, χ (2) 0.324 0.747 20.353 13.009 9.271
[0.851] [0.688] [0.000] [0.002] [0.010]

www.economics-ejournal.org
12 Economics: The Open-Access, Open-Assessment E-Journal

where H is a p×s matrix, αc a s×r matrix of non-zero α-coefficients (s = p−number

of restrictions = 4), and R = H⊥ . The test results in Table 5 suggest that only the
ten-year yield can be regarded as non-equilibrium correcting. If we set a row in
0 0
α to zero, this is translated to a unit vector in α⊥ as α⊥ α = 0, implying that the
cumulated residuals from a weakly exogenous variable form a common trend on their
own.
Imposing a unit vector on α is equivalent to setting the corresponding entry in
0
α⊥ to zero, i.e. a variable that has a unit vector in α is purely adjusting and shocks
to such a variable have no permanent effect on any of the variables in the system.
The hypothesis can be expressed as follows

Hcα : αc = (a, τ ) ⇐⇒ α⊥ = Hα⊥

c
(8)

where a is a p × 1 vector and τ a p × (r − 1) matrix. Table 5 shows that the null

hypothesis is not rejected for the one- and three-month yields.
Since α and β are jointly determined, we cannot impose all restrictions discussed
above at once. In identifying β, we make use of the stationary spread between the
short-term maturities and the weighted differences between spreads as discussed
theoretically in Section 2, in addition to the weak exogeneity of the ten-year yield
in α. The LR test statistic for these restrictions on β and α is 8.309 and follows a
χ2 (8) (p-value: 0.404). The normalised long-run relations are given by (t-values in
brackets):

β̂10 xt = (b3t − b1t )−0.214; (9)

[−10.056]

β̂20 xt = 0.759 (b18 1 120

t − bt ) −0.241 (bt − b18
t )−0.381; (10)
[38.975] [−12.376] [−7.808]

β̂30 xt = 0.467 (b48 18 120

t − bt ) −0.533 (bt − b48
t ). (11)
[−43.137] [−49.314]

Equation (9) shows that the slope of the yield curve is stationary for short-term
maturities, while Equations (10) and (11) indicate that the medium and long ends
of the curve are characterised by an approximately stationary curvature. While the
weights in (10) are close to those suggested by the discussion in Footnote 5, the
same is not true for (11). In practice, however, the curvature of the long end may
be best approximated by equal weights on spreads due to problems of discounting
time to maturity: perceptions of time may be compressed.
Besides the restrictions on β, the row in α corresponding to the ten-year yield
was set to zero (t-values in brackets):
 
0.498 0.140 0.012
 [3.117] [1.432] [0.053]

−0.469 0.434 −0.609
 [−3.669] [5.533] [−3.478] 
 
 0.170 −0.726
α̂ =  −0.288 .
 [−1.418] [1.367] [−2.614] 
 −0.256 0.150 −0.712
 [−1.158] [1.106] [−2.350] 
 
0 0 0
[N.A.] [N.A.] [N.A.]

www.economics-ejournal.org
Economics: The Open-Access, Open-Assessment E-Journal 13

LR test statistic 5%CV=15.5

15

10

5
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

Test stat λ 1 Test stat λ 2

Test stat λ 3 Test stat all λ
1.0 5%CV=1.36

0.5

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Figure 4: Recursive test statistics of the LR test of over-identifying restrictions and

fluctuation test of transformed eigenvalues.

We note that the one-month yield reacts only to the first cointegrating relation.
The three-month yield reacts to all three relations, the eighteen-month and four-
year yields only to the third, the long-end relation.
Recursive tests suggested by Hansen and Johansen (1999) and discussed in
Juselius (2006), Ch. 9, show that coefficients in the restricted CVAR are stable,
and that the restrictions are valid over the entire sample. The upper panel in Figure
4 shows the recursively calculated LR test statistic of the over-identifying restric-
tions, providing evidence for their validity. The recursively computed fluctuation
test is given in the lower panel of Figure 4 where we look at the r-largest trans-
formed eigenvalues and their weighted average.8 At no point is the null hypothesis
of recursively estimated eigenvalues being the same as the full sample estimates
rejected. Hence, the eigenvalues and corresponding cointegrating relations seem
reasonably stable over time.
The finding that the spreads are not stationary by themselves whereas linear
combinations of the spreads are, is useful for an improved understanding of how
expectations on interest rates are formed and hence for monetary policy. See discus-
sions in Sections 1 and 2. Investors making bets on the yield curve may find it even
more profitable to bet on the mean reversion of the weighted differences between
spreads – also called butterfly spreads – than on mean reversion of simple spreads.
The results here suggest that equilibrium mean reversion is fast for differences be-

8
Transformed eigenvalues are given by log(λ̂i ) − log(1 − λ̂i ).

www.economics-ejournal.org
14 Economics: The Open-Access, Open-Assessment E-Journal

tween spreads, whereas deviations from spreads are persistent making the timing of
bets more difficult.

3.2.5 Identification of Common Trends

The moving average (MA) or Granger-Johansen representation of the CVAR is given
by
t
X
xt = C (εi + φDi ) + C ∗ (L)(εt + φDt ) + X0 , (12)
i=1
0
Pt
where C = β⊥ (α⊥ Γβ⊥ )−1 α⊥
0 0
≡ β̃⊥ α⊥ with Γ = I − Γ1 − ... − Γk−1 . α⊥0
i=1 εi defines
the p − r non-stationary common trends loaded by coefficients in β̃⊥ , while C ∗ (L)εt
denotes the stationary part of the process. X0 is a function of initial conditions. See
Johansen (1996), Ch. 4, for derivations. For calculating the coefficients, we use the
model with the above final restrictions imposed on β and α.
Factor models of the yield curve interpret the driving forces of the curve in terms
of level, slope and curvature as noted above. We have found two common trends and
use the Granger-Johansen representation to interpret them. Equation (13) shows
the decomposition of Ĉ into β̃ˆ⊥ α̂⊥
0
multiplied with the vector of cumulated residuals:
 1    Pt 1

bt 2.008 −3.956 P i=1 ε̂i
 b3t  2.008 −3.956 µ ¶  ti=1 ε̂3i 
 18    1  
 bt  = 1.744 −2.740 0 0 0 0  Pt ε̂18  + ...
 48    −0.020 0.065 −1.000 0.963 0  Pi=1 i 
 bt  1.302 −0.697  t ε̂48 
Pti=1 120 i
b120
t 0.914 1.090 i=1 iε̂
(13)
For simplicity of notation, we suppress the other terms of Equation (12). The first
two variables have identical β̃ˆ⊥ -coefficients because they were restricted to cointe-
grate. Since the second common trend in (13) has tiny coefficients on the cumulated
residuals of the one- and three-month yields, it is useful to reformulate the system
as xt = (b1t , b3t , b18 48 18 120 0 48 18
t , bt − bt , bt ) , and test whether bt − bt is weakly exogenous
conditional on the same over-identifying restrictions on β. The LR test statistic
for the restrictions on β and α is 8.637 and follows a χ2 (11) (p-value: 0.655) and
the additional restrictions seem acceptable. The Granger-Johansen representation
simplifies to
   
b1t 2.001 −3.751  Pt

 b3t   2.001 −3.751 ε̂120
    i 
 bt 18  =  1.744 −2.583  i=1 
 48     P 48 P 18  + ... ,
t t
bt − b18 t
  −0.435 1.921  ε̂i − ε̂i
b120
t 0.921 1.013 i=1 i=1

or equivalently
   
b1t 2.001 −3.751  Pt

 b3t  2.001 −3.751 120
ε̂i
 18    
 bt  = 1.744  
−2.583  P i=1  + ... .
 48   t Pt 
 bt  1.309 −0.662 ε̂48
i − ε̂18
i
b120
t 0.921 1.013 i=1 i=1

www.economics-ejournal.org
Economics: The Open-Access, Open-Assessment E-Journal 15

The coefficients in β̃ˆ⊥ are interpreted as the weights attached to the common trends,
or in the language of finance models as the loadings of the factors with respect to
each variable. The loadings on the first common trend, the cumulated residuals
of the ten-year yield, are in an interval between just under one and two for all
variables which implies that it affects all yields similarly, giving it the interpretation
of a level factor. This is a plausible result because the ten-year yield contains
most information and we should expect shocks to it to influence all yields similarly.
The 2nd column of β̃ˆ⊥ shows coefficients decreasing with maturity, suggesting an
interpretation as a slope factor. This interpretation is further strengthened by noting
that the second common trend is the spread between the cumulated residuals of the
four-year and eighteen-month yields. A positive shock to the spread increases the
slope and therefore has a negative effect in particular on the short end of the curve,
while a negative shock flattens the yield curve.
In conclusion, our model not only shows that the yield curve is explained by
a level and slope factor but gives meaning to them by identifying the cumulated
residuals that drive them: the long end of the curve determines the level and the
medium sector the slope. The short end does not contain information on yields of
other maturities.

3.2.6 Towards a Structural Interpretation

The CVAR residuals used in the determination of the common stochastic trends in
(13) are seldom uncorrelated and, therefore, not structurally unique. To attach a
structural interpretation to the shocks in the model it is common practise to or-
thogonalise the residuals. However, Hendry and Mizon (2000) warn against such
interpretation because a structural shock should be invariant to changes in the in-
formation set. But, unless the model coincides with the data-generating process,
residuals are generally not invariant to extensions and omissions. We do not claim
that the conditions for a structural interpretation are granted in our case but choose
to orthogonalize the residuals as a robustness check of the results in the previous
section. In case the conclusions remain reasonably robust we believe this to lend
credibility to a causal interpretation of the common trend results.
We use the Granger-Johansen representation but pre-multiply by a rotation ma-
trix B such that the shocks ut = Bεt are uncorrelated:9
t
X
−1
xt = CB (ui + BφDi ) + C ∗ (L)B −1 (ut + BφDt ) + X̃0 (14)
i=1

from

B∆xt = Bαβ 0 xt−1 + BΓ1 ∆xt−1 + BΓ2 ∆xt−2 + BΓ3 ∆xt−3 + BφDt + Bεt , (15)

where BΩB 0 = I for orthogonality. The impact matrix CB −1 = β̃⊥ α⊥ 0

B −1 has r
zero columns due to transitory shocks, and p − r non-zero columns corresponding to
µ ¶
9 α0−1
B is chosen such that B = S −1 G, where G = 0 and S is found through Choleski
α⊥
decomposition of GΩG0 . Then V ar(ut ) = BΩB 0 = Ip .

www.economics-ejournal.org
16 Economics: The Open-Access, Open-Assessment E-Journal

the number of common trends. P It reveals how the different variables in the system
react to the permanent shocks ti=1 uji , j = 1, 2. However, due to the rotation of
the residuals, the permanent shocks may not have straightforward interpretations.
We have two permanent shocks and to identify the impact matrix need to restrict
one entry to zero.10 We choose the ten-year yield because based on the previous anal-
ysis we believe it to be influenced least. The normalised impact matrix is estimated
as  
0 0 0 1 1
0 0 0 1 1 
ˆ  
Ĉ B̂ = β̃⊥ α̂⊥ B̂ = 0 0 0 0.992 0.761
−1 0 −1 
,
0 0 0 0.978 0.357
0 0 0 0.967 0

where the first three columns reflect impacts from transitory shocks, and the 4th and
5th impacts from permanent shocks (the equal coefficients for the one- and three-
month yields are again due to them cointegrating, and to normalisation). The first
cumulated permanent shock has nearly identical loadings, indicating a level factor,
while the second one is a slope factor. To gain understanding on how to interpret
the independent shocks, we examine the rotation matrix given by
 
1.000 −0.008 −0.653 0.828 −0.587
−0.390 1.000 −0.753 0.858 −0.446
 
B̂ = 
 0.061 −0.576 0.083 1.000 −0.816 ,
 0.008 0.144 −0.418 −0.043 1.000 
0.023 −0.061 1.000 −0.988 0.226

where the first three rows are due to transitory shocks and rows 4 and 5 are due
to permanent shocks. The rotation matrix matches the relations found for the
common trends above. The 4th row has small coefficients for all variables except
the ten-year yield (and possibly the eighteen-month yield), while the 5th row is
essentially the spread between the four-year and eighteen-month yields. Hence, the
permanent structural shocks correspond closely to the linear combination of the
CVAR residuals that defined the previously estimated common trends, suggesting
that they might have been approximately orthogonal from the outset. Therefore,
the estimated common stochastic trends might be given a structural interpretation.

3.3 The Common Trends and Macroeconomic Variables

In this section, we use cointegration to determine relationships between the macroe-
conomy and the common trends from Section 3.2.5 thereby providing a further link
to the common factor literature where combining yield factors with macroeconomic
variables is a relatively new strand.11 As demonstrated below this is also possible
within the CVAR framework.
10
Since restrictions on α̂ are removed when estimating the structural MA, the two formulations of
xt presented in the previous section give identical results up to linear combinations of coefficients.
Results presented here are for xt = (b1t , b3t , b18 48 120 0
t , bt , bt ) .
11
See inter alia Diebold, Rudebusch, and Aruoba (2006) and Dewachter and Lyrio (2006).

www.economics-ejournal.org
Economics: The Open-Access, Open-Assessment E-Journal 17

CT1 CT2
0.0
0.05

−0.2 0.00

−0.05
−0.4

1990 1995 2000 1990 1995 2000

dlcpi 12−month MA(dlcpi) tcu

0.75 84

0.50
82

0.25
80
0.00
1990 1995 2000 1990 1995 2000

Figure 5: Level plots of the common trends and macroeconomic variables

Variables included
Pt in the new CVAR are the two common
Pt trends
Pt from the yield
120 48 18
model (CT 1 = i=1 ε̂i , the level factor, and CT 2 = i=1 ε̂i − i=1 ε̂i , the slope
factor, both scaled down by 12), the monthly inflation rate based on the log of the
consumer price index (dlcpi, scaled up by 100), and total capacity utilisation (tcu,
in percent).12 The series are shown in Figure 5.
The model is estimated with a lag length of 2 as suggested by LR tests. Mis-
specification tests reveal problems only with normality, and three impulse dummies
are accordingly included, for January 1990, August 1990 and April 1999. Based on
the criteria discussed in Section 3.2.3, the rank is set to r = 2. Recursive estimation
suggests that parameters are stable over the sample. Together with long-run weak
exogeneity of both CT 1 and tcu in α, the following over-identifying restrictions on
β are accepted with an LR test statistic of 7.313 (χ2 (5), p-value: 0.198):

β̂10 xt = CT 1 − 2.211dlcpi + 0.768; (16)

β̂20 xt = CT 2 − 0.297dlcpi + 0.030tcu − 2.396. (17)

The first common trend from the yield model, CT 1, is positively related with infla-
tion in (16), i.e. shocks to the long-term interest rate shifting the yield curve seem
associated with inflationary shocks. The second cointegrating relation (17), which
involves the second common trend CT 2, includes inflation and capacity utilisation.
12
The data for all macroeconomic variables was obtained from FRED, the database of the Federal
Reserve Bank of St. Louis, and are seasonally adjusted. Alternative monthly measures of economic
activity like the number of housing starts and of help-wanted advertising in newspapers gave
qualitatively equivalent results, but are not presented here.

www.economics-ejournal.org
18 Economics: The Open-Access, Open-Assessment E-Journal

A steepening of the slope is associated with an increase in inflation, a decrease in

activity or both.
The Granger-Johansen representation of the macro model gives an indication of
feedback directions between the yield curve and the macroeconomy. Coefficients of
the C-matrix (t-values in brackets) are estimated as
 
1.221 0.000 0.000 0.018
 [7.652] [0.000] [0.000] [0.496] 
−0.039 0.000 0.000 −0.041
 [−0.417] [0.000] [0.000] [−1.962] 
Ĉ = 
 0.552 0.000 0.000 0.008  ,

 [7.652] [0.000] [0.000] 
[0.496] 

6.724 0.000 0.000 1.430
[1.941] [0.000] [0.000] [1.860]

where the zero columns are due to CT 2 and inflation being purely adjusting, and the
non-zero columns to CT 1 and capacity utilisation being weakly exogenous. This rep-
resentation provides evidence that CT 1 is determined largely by its own cumulated
residuals, while CT 2 is driven by the cumulated residuals of the activity measure.
The inflation rate is driven only by the cumulated residuals of CT 1, while capacity
utilisation depends on its own cumulated residuals as well as CT 1.13
Hence feedback in both directions exists between the macroeconomy and the
yield curve. The level of the yield curve positively influences the inflation rate and
activity measure, but is itself independent of the macroeconomy. This is in contrast
to the slope which reacts to the activity measure, but exerts no influence on other
variables. Given that CT1 captures unexplained elements of the long-term yield,
it may represent inflation expectations. These may then determine inflation itself
by being partly self-fulfilling, and should therefore be useful in forecasting inflation.
CT2 may be interpreted as a term premium following Equation (4). We find the
spread to be non-stationary and according to (4) the non-stationary part is the
term premium. Since CT2 is made up of the spread between unexplained parts of
the two medium-term yields, this interpretation appears plausible. Related to the
macroeconomy, we have that the term premium increases if inflation increases and
activity decreases. Its non-stationary component derives from the common trend
associated with the activity measure (as shown in the estimated C-matrix), and
macro factors may thus prove helpful in forecasting it.

4 Conclusion
The approach introduced in this paper considers the stationarity of the yield curve’s
derivatives. Past literature has focused on testing the hypothesis of stationary
spreads in accordance with the expectation theory but not, in case of rejection,
examined the differences between spreads. This extension to the theoretical frame-
work allows us to test the stationarity of weighted differences between spreads in
a CVAR of US treasury yields which is accepted. Two term premia of different
maturities therefore cointegrate and the curvature of the yield curve may allow a
more meaningful assessment of future interest rate expectations than the slope.
13
Where significance is only borderline it becomes stronger when removing the weak exogeneity
restrictions.

www.economics-ejournal.org
Economics: The Open-Access, Open-Assessment E-Journal 19

In addition, the Granger-Johansen representation of the CVAR proved a powerful

tool for characterising the non-stationary components of the yield curve, identifying
them as level and slope. The cumulated residuals of the ten-year yield make up a
common trend associated with a level shift, while the second common trend is the
spread between the cumulated residuals of the four-year and eighteen-month yields
impacting on the slope of the yield curve. When considering relations between these
common trends and macroeconomic variables, we find causation running from the
level of the yield curve to inflation, and from an activity measure to the slope of the
yield curve.

References
Baba, Y., D. F. Hendry, and R. M. Starr (1992). The demand for M1 in the
U.S.A., 1960-1988. Review of Economic Studies 59, 25–61.
Bliss, R. R. (1997). Testing term structure estimation methods. Advances in Fu-
tures and Options Research 9, 197–231.
Campbell, J. Y. and R. J. Shiller (1987). Cointegration and tests of present value
models. Journal of Political Economy 95, 1062–1088.
Carstensen, K. (2003). Nonstationary term premia and cointegration of the term
structure. Economic Letters 80, 409–413.
Dennis, J. G. (2006). CATS in RATS - Cointegration Analysis of Time Series,
Version 2. Evanston, Illinois: Estima.
Dewachter, H. and M. Lyrio (2006). Macro factors and the term structure of
interest rates. Journal of Money, Credit and Banking 38, 119–140.
Diebold, F. X., G. D. Rudebusch, and S. B. Aruoba (2006). The macroeconomy
and the yield curve: a dynamic latent factor approach. Journal of Economet-
rics 131, 309–338.
Doornik, J. A. and D. F. Hendry (2007). Modelling Dynamic Systems Using Pc-
Give 12: Volume II. London: Timberlake Consultants Press.
Duffee, G. (2002). Term premia and interest rate forecasts in affine models. Jour-
nal of Finance 57, 405–443.
Engle, R. F. and C. W. J. Granger (1987). Co-integration and error correction:
representation, estimation and testing. Econometrica 55, 251–276.
Fama, E. F. and R. R. Bliss (1987). The information in the term structure. Amer-
ican Economic Review 77, 680–692.
Giese, J. V. (2006). Characterising the yield curve’s derivatives in a regime-
changing cointegrated VAR model. Manuscript.
Gonzalo, J. (1994). Five alternative methods of estimating long-run equilibrium
relationships. Journal of Econometrics 60, 203–233.
Hall, A. D., H. M. Anderson, and C. W. Granger (1992). A cointegration analysis
of Treasury bill yields. The Review of Economics and Statistics 74, 117–126.

www.economics-ejournal.org
20 Economics: The Open-Access, Open-Assessment E-Journal

Hansen, H. and S. Johansen (1999). Some tests for parameter constancy in the
cointegrated VAR. Econometrics Journal 2, 306–333.
Hendry, D. F. and G. E. Mizon (2000). Reformulating empirical macro-
econometric modelling. Oxford Review of Economic Policy 16, 138–159.
Johansen, S. (1996). Likelihood-based Inference in Cointegrated Vector Auto-
regressive Models. Oxford: Oxford University Press.
Juselius, K. (2006). The Cointegrated VAR Model: Econometric Methodology and
Macroeconomic Applications. Oxford: Oxford University Press.
Knez, P. K., R. Litterman, and J. A. Scheinkman (1994). Explorations into factors
explaining money market returns. Journal of Finance 49, 1861–1882.
Litterman, R. and J. A. Scheinkman (1991). Common factors affecting bond re-
turns. Journal of Fixed Income 1, 51–61.
Nelson, C. R. and A. F. Siegel (1987). Parsimonious modeling of yield curves.
Journal of Business 60, 473–489.
Nielsen, B. (2006). Order determination in general vector autoregressions. In H.-
C. Ho, C.-K. Ing, and T. Lai (Eds.), Time Series and Related Topics: In
Memory of Ching-Zong Wei. IMS Lecture Notes and Monograph Series, 52.
Shea, G. S. (1992). Benchmarking the expectations hypothesis of the interest-rate
term structure: an analysis of cointegration vectors. Journal of Business and
Economic Statistics 10, 347–366.
Zhang, H. (1993). Treasury yield curves and cointegration. Applied Economics 25,
361–367.

www.economics-ejournal.org
Please note:
You are most sincerely encouraged to participate in the open assessment of this article. You
can do so by either rating the article on a scale from 1 (bad) to 5 (excellent) or by posting your
comments.
Please go to:
www.economics-ejournal.org/economics/journalarticles/2008-28

The Editor

www.economics-ejournal.org

© Author(s) 2008. This work is licensed under a Creative Commons License - Attribution-NonCommercial 2.0 Germany