The financial crisis and the

changing dynamics of the yield curve1

Morten L Bech2 and Yvan Lengwiler3

Abstract

We present evidence on the changing dynamics of the yield curve from 1998 to 2011. We
identify four different phases. As expected, the financial crisis represents a period of elevated
yield volatility, but it can be split into two distinct periods. The split occurs when the Federal
Reserve reached the zero lower bound. This bound suppressed volatility in the short end of
the yield curve while increasing volatility in the long end – despite lower overall volatility in
financial markets. In line with previous studies, we find that announcements with regard to
the Federal Reserve’s large scale asset purchases reduce longer term yields. We also
quantify the effect of widely observed economic news, such as the non-farm payrolls and
other items, on the yield curve.
Keywords: Term structure of interest rates, financial crisis, interest rate dynamics, LSAP,
unconventional monetary policy
JEL classification: E43, E52

1
We thank Torsten Ehlers, Jacob Gyntelberg and Peter Kugler for helpful discussions and we thank Andrew
Filardo for his useful comments. The views expressed in this paper are those of the authors and do not
necessarily reflect those of the Bank for International Settlements.
2
Bank for International Settlements.
3
Dept. of Economics, University of Basel, Peter Merian-Weg 6, P.O. Box, CH–4002 Basel, Switzerland.
Corresponding author. Email addresses: [email protected] (Morten L. Bech), [email protected]
(Yvan Lengwiler).

BIS Papers No 65 257

1. Introduction
The yield curve on U.S. Treasury securities is one of the most closely watched data of the
global economy. Understanding its dynamics is a preoccupation of many financial market
participants as well as academics. In this paper, we investigate how the dynamics of the yield
curve were affected by the financial crisis and the subsequent policy responses using the
“intelligible factors” framework of Lengwiler and Lenz (2010).
We identify four different phases of yield curve dynamics since 1998 (Section 3). After a
“normal” phase ending mid-2004 we observe a period that is characterized by a conspicuous
absence of volatility in yields. This “moderation” phase ends with the beginning of the
financial crisis in August 2007. The first part of the crisis, which we label “liquidity crisis”, was
characterized by money market turmoil and liquidity problems. Accordingly, we observe huge
volatility in the short and medium maturity spectrum of the yield curve. This pattern abruptly
changes in December 2008, after the Federal Reserve reached the zero lower bound. Since
then, we observe a lack of perturbations at short maturities, but unusually large volatility in
the long maturity spectrum of the yield curve. Reaching the zero lower bound appears in our
analysis to be a significant event that has quantitatively changed the dynamics of the yield
curve.
Our second result (Section 4) concerns the identification of the most important shocks. We
quantify and locate in the maturity spectrum the most significant shocks, e.g. 9/11, the
Lehman collapse, the rescue of AIG, or the increase of the large scale asset purchases
(LSAPs) in March 2009.
Our third result (Section 5) concerns the measurement of the effect of surprises in key
macroeconomic data on the yield curve. In particular, we measure how deviations of
published indicators, such as non-farm payrolls, jobless claims, and other items, from
expected values affect the yield curve over the whole maturity spectrum. We find that these
surprises do indeed correlate with yield curve shocks, but the connection has become
weaker in the crisis.

2. Intelligible factors
We use the decomposition of the term structure into “intelligible factors” developed by
Lengwiler and Lenz (2010). We have M maturities that we observe on T days. Let rt  m 
denote the interest rate for a zero bond at time t which matures at time t  m . The cross
section of interest rates is described by three factors,
rt  m   k1  m  1,t  k2  m  2,t  k3  m  3,t   t  m  , (1)

where k M  3 matrix  are the loadings and  3  T  matrix  are time-varying factors. 
and k are constructed together so that they have certain desirable properties. Firstly,
constraints are imposed on the loadings k , such that they load on different parts of the
maturity spectrum, as can be seen from Figure 1. The first factor is the only one that loads on
the very long end of the maturity spectrum, so we call 1 the long factor. The second factor is
the only one that loads on the very short end of the maturity spectrum, so we call 2 the short
factor. The third factor has zero loading at the short and the long end of the maturity

258 BIS Papers No 65

spectrum, but it is normalized in such a way that it achieves unit loading somewhere in the
middle. We call this the curvature factor.4
Secondly, the dynamics of the factors Á is described by a vector auto-regression (VAR),
t  D0  D1t 1  ...  Dpt  p  ut (2)
,
where t  1,t ,2,t ,3,t  ' and D0 ,..., Dp are the coefficient matrices of the VAR. We set p large
enough so that the factor innovations ut become serially uncorrelated. As described in
Lengwiler and Lenz (2010), the shape of the loadings k is adjusted in such a way that the
factor innovations u are also uncorrelated with each other. As a result, the covariance matrix
of the innovations, E uu ' , is diagonal, and the VAR is structural in that sense.

Figure 1
Loadings of the three factors

4
Note that these loadings differ from the more common loadings “level”, “slope”, and “curvature”, which
have become custom in applications of principal component analysis (Litterman and Scheinkman, 1991)
or in the specification of Nelson and Siegel (1987).

BIS Papers No 65 259

The result of this procedure is a set of loadings that describe the long end, the short end, and
the curvature of the yield curve. The dynamics of these factors are described by a structural
VAR model.
We use the constant maturity yield curve data produced by the U.S. Treasury. These
estimates are generated from secondary market quotes of U.S. Treasury debt, and
interpolated with splines to yield estimates at given, constant times to maturity.5 We use
observations at three and six months, and one, two, three, five, seven, ten, and twenty years.
We use daily observed data from January 2, 1998 to November 8, 2011 (worth
3468 business days). We repeat the estimation presented in Lengwiler and Lenz (2010) with
this expanded data set. We find that we need thirty lags in the VAR to remove serial
correlation of the innovations. The estimated factors are shown in Figure 2.
The innovations u are uncorrelated white noise random variables by construction. They drive
the dynamics of the factors and thus of the term structure. Through the VAR dynamics, an
innovation into one factor has the potential to ultimately affect all the factors as time passes.
However, as was already discussed in Lengwiler and Lenz (2010), it is an important stylized
fact of the intelligible factors decomposition that innovations into the short and the long factor
essentially only affect themselves: there is very little spillover on the other factors.
Innovations into the curvature factor, in contrast, are the main drivers of movements of the
curvature and the short factor. As a result, curvature innovations are by far the most
important source of the overall yield curve dynamics.

Figure 2
Estimated intelligible factors

This stylized fact is also true in the extended data sample. Figure 3 depicts the variance
decomposition, i.e. the parts of the variance of the interest rates that are due to the innovations
into the three factors, u . The variance decomposition firstly reveals that the model captures
the second moment of the yields – the term structure of interest rate variance – very well, and
secondly confirms that most of the yield curve movements have their source in innovations into
the curvature factor.

5
See www.treasury.gov/resource-center/data-chart-center/interest-rates/Pages/yieldmethod.aspx for a description
of the methodology.

260 BIS Papers No 65

 Figure 3
Variance of yields of different maturities and shares
explained by the three types of innovations

3. Four phases of term structure dynamics

Visual inspection of the innovations reveals that their volatility has not been constant
throughout the sample. In order to measure this, we compute the “local volatility” of the factor
innovations; see Figure 4.6 The financial crisis is clearly visible in this graph, but we can
distinguish two phases. Beginning in August 2007, the volatilities of the short and the
curvature innovations explode, and stay high until the end of 2008. After that they go back to
pre-crisis levels. The volatility of long factor innovations also increases in 2007, but becomes
particularly large at the end of 2008. It remains high until the end of the sample. Today, the
long factor innovations appear much more volatile than before the crisis. The same is not
true for the short and curvature innovations.
This pattern becomes even clearer if we focus attention only on the largest innovations. To
that avail, we compute standardized innovations, i.e. we divide the three factor innovations
by their unconditional standard deviations. Figure 5 depicts those standardized innovations

6
“Local volatility” is a non-parametric measure of the second moment. It is essentially a Nadaraya-Watson
kernel regression on the squared innovations; details are explained in Appendix A.

BIS Papers No 65 261

that are significantly different from zero at 95% confidence (so greater than 1.96 in absolute
terms). We can distinguish four phases.
The first phase begins at the start of our sample and ends roughly at mid-2004 (we chose
end of July). In this phase we see approximately what we would expect to see. In fact, the
three standardized innovations are independent and serially uncorrelated random variables
with unit variance. If they are normally distributed, for each individual series, 5% of the
observations should be significantly different from zero. During this first phase, this is more or
less what we observe: 3.7% of the long innovations, 2.6% of the short innovations, and 4.7%
of the curvature innovations are significantly different from zero. One might label this period
the “normal phase”.
The second phase begins August 2004 and ends August 2007. We call this the “moderation
phase”. The exact timing between the normal and the moderation phase is difficult to
pinpoint. The end of the moderation phase, however, is connected to an important event,
namely BNP Paribas’s announcement that it was freezing three funds invested in sub-prime
securities, which is commonly taken to mark the beginning of the financial crisis. This second
phase is characterized by the marked absence of large innovations. Only 0.7% and 0.5% of
the innovations into the long and the short factor are significantly different from zero. For the
curvature, the number is 1.3%. Thus, this phase has very low volatility, and thus the
standardized innovations turn out to be small and statistically insignificant.
This has dramatically changed with the financial crisis, which we can split into two separate
phases. Phase number three, which we call the “liquidity crisis phase”, begins on August 9,
2007 and ends on December 16, 2008. This is the date when the Federal Open Market
Committee lowered the target for the effective federal funds rate to a 0 to 25 basis points
(bps) range and effectively reached the zero lower bound. This phase was characterized by
the freezing of the interbank money market and substantial liquidity interventions by the
Federal Reserve; in particular later in the period. Accordingly, we observe 23.4% of the
innovations into the short factor that are significantly different from zero. For curvature, the
number is also very large, 19.5%. Long factor innovations are also more volatile than before,
but to a lesser extent: 8.9% of the days feature a long factor innovation that is significantly
different from zero in this phase.
The fourth, the “zero lower bound phase”, begins after the Federal Reserve has reached the
zero lower bound and lasts to the end of the sample. With no room downward on the federal
funds rate, and traditional instruments of monetary policy exhausted, the volatility in the short
and the curvature innovations vanishes. Only 0.3% and 2.5% of the innovations of these
factors are significant. In contrast, 12.2% of the long factor innovations are now significantly
different from zero.
We ran breakpoint tests (Bai and Perron, 1998, 2003) for the long, short, and curvature
innovations, respectively. The tests for the short and the curvature innovations both find a
break in early August 2007. All tests find a break in late 2008, but the dates differ. For the
long innovations a third break is found in late 2009. For the period before the financial crisis,
no consistent breaks are found across the three series.

262 BIS Papers No 65

 Figure 4
Innovations and local volatilities
In basis points

BIS Papers No 65 263

 Figure 5
Large innovations into the long, short, and curvature factors,
measured in multiples of unconditional standard deviations

264 BIS Papers No 65

Table 1 reports similar information to Figure 5 but focuses on the joint distribution of the
innovations across days. Counting just significant or non-significant innovations, eight
combinations are possible on any given day. The most likely possibility is that none of the
innovations is significant. Theoretically, that should happen with probability 0.953 =86% . The
long innovation should be significant while the short and the curvature innovations are not with
probability 0.05  0.952 = 4.5% , etc. The least likely case is that all three innovations are
significant on the same day. This event should be observed only in 0.053  0.01% of the days.
The theoretical values for the cases with at least one significant innovation are reported in the
first column of Table 1. The remaining columns contrast this with the actual measurement in
the four phases. We observe more or less what we should observe if the shocks are
independently and normally distributed in the “normal phase”. In the “great moderation phase”
there are clearly too few significant innovations. In the “liquidity crisis phase” we observe way
too many short and curvature innovations. In particular, we also find nineteen days where we
observe significant contemporaneous short and curvature innovations. Theory would have
predicted zero or one such day. There are even five days where all three innovations are
significant. In the “zero lower bound phase”, finally, significant short innovations have
completely vanished and significant curvature innovations are far below the theoretical
expectation. Instead, there is a large density of significant long factor innovations.

Table 1
Significant shocks to the yield curve during the four phases
Theoretical normal great liquidity zero lower
phase moderation crisis bound
# cases share # cases share # cases share # cases share
long only 4.51% 49 2.98% 5 0.66% 144 14% 77 10.6%
short only 4.51% 30 1.82% 4 0.53% 50 14.8% 0 0.00%
curv only 4.51% 57 3.46% 10 1.32% 36 10.7% 4 0.55%
long & short only 0.24% 1 0.06% 0 0.00% 5 1.48% 0 0.00%
long & curv only 0.24% 9 0.55% 0 0.00% 6 1.78% 12 1.65%
short & curv only 0.24% 10 0.61% 0 0.00% 19 5.62% 2 0.27%
all three 0.01% 2 0.12% 0 0.00% 5 1.48% 0 0.00%

The shift of the location of the innovations during the financial crisis, and in particular to the
longer part of the maturity spectrum when the zero lower bound became binding, also
manifests itself in the variance attribution. We compute the variance of the yields separately for
the four phases; see Figure 6 and compare with Figure 3 for the whole sample. The differences
are striking. First of all, the overall variance of the yields has dramatically decreased for shorter
maturities in the “zero lower bound phase”. This is a direct corollary of the fact that the zero
lower bound does not allow rates to decrease further, and the Federal Reserve has not allowed
the short rates to increase, hence volatility in this duration spectrum has vanished. As a result,
all the volatility that remains is at longer maturities. The volatilities of the ten- and twenty-year
yields is more or less unchanged for the two subperiods. Yet, because no further movements
at the short end are possible, and the major innovations now occur in the long factor, almost all
of the term structure of yield variance can be attributed to long innovations.

BIS Papers No 65 265

4. (Reverse) event study
In this section we relate the largest innovations that we measure to identifiable events. We
rely on a variety of sources. For Federal Reserve news we use press release information
from the website of the Federal Reserve Board and the Federal Reserve Bank of New York.
For market news, in the early years of our sample period we rely on next day summaries of
financial market activity from the New York Times – with a particular focus on the Treasury
market. After September 2004, we use daily press summaries from Wrightson ICAP. These
press summaries are produced towards the end of the business day and are made available
for clients before the close of business. They contain so-called “wraps” for different financial
markets (including Treasuries) as well as a list of the news stories that are likely to make the
headlines the following day. For the part of the sample that covers the height of the financial
crisis, we also use the financial crisis time line of the Federal Reserve Bank of St. Louis for
robustness.7 In addition, we also check whether announcements by the Treasury department
concerning its funding needs or issuing strategy might be related to our yield curve
innovations.8 We find, however, no evidence that they contain relevant information.

Figure 6
Variance of yields of different maturities and shares
explained by the three types of innovations, in the four phases

Tables 2, 3, and 4 collect the twenty-five largest innovations (in absolute terms) for the long,
short, and curvature factor, respectively, and also report potentially related economic or
financial events. To better gauge the size of these innovations we also divide them, in the
seventh column, by the unconditional standard deviation, by the local volatility estimate for

7
http://timeline.stlouisfed.org.
8
http://www.treasury.gov/resource-center/data-chart-center/quarterly-refunding/Pages/default.aspx.

266 BIS Papers No 65

that day (the eighth column) and the conditional GARCH volatility estimate (the ninth
column). We also report simple first differences of some key interest rates.
Some dates are particularly noteworthy. We measure a –37 bps shock in the short and a
–49 bps shock in the curvature factor on the day the markets reopened after the
9/11 attacks. These are 5.6 and 11.0 standard deviation events, respectively. The greatest
short factor innovation, however, is measured the day of the AIG bailout (–88 bps).
The Lehman bankruptcy on September 15, 2008 shows up as a large innovation in all three
factors simultaneously: we measure innovations into the long factor (–23 bps), the short
factor (–32 bps), and the curvature (–26 bps) on that day. This amounts to shocks between
3.5 and 5.9 standard deviations of the respective innovation series.
Notable is also the (perverse) effect of the S&P downgrade of U.S. government debt on
August 8, 2011. On that day we measure a large negative innovation in the long factor
(–25 bps). A possible interpretation might be that the downgrade has produced a flight for
safety (“Europe will be next”) and thus increased the demand for U.S. debt.

Table 2
Twenty-five most important innovations to the long factor. The second column reports the size of the innovation
in basis points (bps), and, in parentheses, relative to the unconditional and the local volatility of that day,
respectively. For instance, the largest absolute long innovation is measured on March 18, 2009. We measure a
–53 bps shock. This is 8.0 times the unconditional standard deviation of the long innovation series, and it is
12.3 times larger than the local volatility of that day.

date innovation event

2011-10-31 –22 [3.4, 2.0] Greek PM Papandreou announces referendum on Eurozone debt
deal
2011-10-27 +23 [3.4, 2.1] Euro summit on Greek debt
2011-09-22 –23 [3.4, 2.1] One day after Operation Twist 2 was announced
2011-08-24 +21 [3.2, 1.9] French government unveils a EUR 12 billion deficit cutting
package
2011-08-11 +25 [3.7, 2.2] Bad bond auction three days after downgrade and two days after
Fed’s forward guidance
2011-08-09 –22 [3.3, 1.9] “Forward guidance”: Low federal funds rate through mid-2013
2011-08-08 –25 [3.9, 2.3] Downgrade of U.S. government debt by S&P
2010-12-14 +25 [3.7, 2.5] Confirmation of reinvestment policy and purchase of $600 billion
of longer term Treasuries; little likelihood of increase of QE 2
2010-12-07 +23 [3.5, 2.3] (No relevant news)
2009-06-01 +30 [4.6, 2.8] Surprisingly strong data sapped the safe-haven appeal of
government debt
2009-05-27 +24 [3.6, 2.2] Concerns about the growing supply of bonds
2009-03-18 –53 [8.0, 4.3] QE 1 enlargement: Additional $750 billion agency MBS and
$100 billion agency debt; $300 in longer term Treasuries
2009-02-17 –29 [4.3, 2.3] Worries about European banks spurred investors to seek safety in
U.S. government debt
2008-12-01 –25 [3.7, 2.1] Bernanke: Fed could purchase Treasuries
2008-11-25 –27 [4.0, 2.3] QE 1: Initial large scale asset purchase announcement:
$500 billion agency MBS and $100 billion agency debt
2008-11-20 –26 [4.0, 2.3] Jobless claims reach new record
2008-09-15 –23 [3.5, 2.8] Lehman bankruptcy
2004-04-02 +25 [3.7, 3.9] (No relevant news)
2003-01-02 +26 [3.9, 3.8] (No relevant news)
2002-11-07 –22 [3.4, 2.7] One day after 50 bps cut
2001-12-07 +25 [3.7, 2.8] (No relevant news)
2001-11-15 +22 [3.3, 2.5] Dimmed hopes for further rate cuts due to positive news
2001-01-03 +27 [4.1, 4.0] 50 bps cut
1998-10-09 +23 [3.5, 2.8] (No relevant news)
1998-10-08 +25 [3.8, 3.1] Rumors of unwinding of a carry trade by a hedge fund

BIS Papers No 65 267

 Table 3

Twenty-five most important innovations to the short factor; see Table 2 for explanation.

date innovation event

2008-10-20 +46 [6.9, 2.2] Government measures show signs of reviving the frozen money
market, causing an exodus out of ultrasafe short-dated Treasuries
2008-10-16 +32 [4.8, 1.4] (No relevant news)
2008-10-10 –35 [5.3, 1.6] Early close ahead of Columbus day. Flight to safe haven
2008-09-23 –43 [6.5, 1.2] Bernanke supports TARP
2008-09-19 +78 [11.8, 1.8] Asset-Backed Commercial Paper Money Market Mutual Fund
Liquidity Facility (AMLF) and ABCP MMMF Liquidity Facility
2008-09-17 –88 [13.3, 2.1] AIG bailout
2008-09-15 –32 [4.8, 0.8] Lehman bankruptcy
2008-03-24 +57 [8.7, 2.3] FRBNY announces that it will provide term financing to facilitate
JPMorgan Chase’s acquisition of Bear Stearns
2008-03-19 –31 [4.7, 1.3] One day after 75 bps cut. Reduction of required capital for Fannie
Mae and Freddie Mac
2008-03-18 –36 [5.5, 1.5] 75 bps cut
2008-01-22 –49 [7.4, 2.4] 75 bps cut
2007-12-24 +36 [5.5, 2.2] (No relevant news)
2007-09-04 +51 [7.7, 2.1] Money market turmoil
2007-08-29 –36 [5.4, 1.1] Money market turmoil
2007-08-27 +51 [7.7, 1.5] Money market turmoil
2007-08-24 +37 [5.5, 1.0] Money market turmoil
2007-08-21 +46 [7.0, 1.2] Money market turmoil
2007-08-20 –70 [10.6, 1.9] Money market turmoil
2007-08-15 –49 [7.4, 1.5] Money market turmoil after BNP Paribas writedown
2001-09-13 –37 [5.6, 2.2] Market reopens after terrorist attacks, Fed will “provide whatever
liquidity might be needed”
2000-12-26 +69 [10.5, 2.2] (No relevant news)
2000-12-21 –43 [6.6, 1.6] Speculation that Federal Reserve may lower interest rates before
scheduled meeting at the end of January
1998-10-19 +40 [6.1, 1.8] Two days (!) after 50 bps rate cut, reversing move of short factor a
day earlier
1998-10-16 –49 [7.4, 2.0] One day after 50 bps rate cut
1998-10-08 –31 [4.8, 1.9] Rumors of unwinding of a carry trade by a hedge fund

268 BIS Papers No 65

 Table 4

Twenty-five most important innovations to the curvature factor; see Table 2 for explanation.

date innovation event

2009-06-05 +29 [6.5, 3.6] Smaller than expected drop in non-farm payrolls
2008-12-17 +17 [3.8, 3.1] One day after rate cut to 0–0.25%. FOMC statement mentions the
possibility of purchases of longer maturity debt
2008-10-20 +18 [4.1, 2.2] Government measures show signs of reviving the frozen money
market, causing an exodus out of ultrasafe short-dated Treasuries
2008-10-02 –19 [4.2, 1.5] Rise in jobless claims and worse than expected factory orders
2008-09-29 –23 [5.1, 1.8] Fed: Expansion of FX Swap lines. The U.S. House of
Representatives rejects legislation submitted by the Treasury
Department requesting authority to purchase troubled assets from
financial institutions
2008-09-19 +30 [6.7, 2.3] Asset-Backed Commercial Paper Money Market Mutual Fund
Liquidity Facility (AMLF) and ABCP MMMF Liquidity Facility
2008-09-17 –21 [4.7, 1.6] AIG bailout
2008-09-16 +17 [3.9, 1.4] Rate unchanged
2008-09-15 –26 [5.9, 2.1] Lehman bankruptcy
2008-06-12 +15 [3.5, 1.6] (No relevant news)
2008-06-09 +31 [7.0, 3.2] Better looking housing data
2008-03-18 +19 [4.3, 2.2] 75 bps cut
2008-03-11 +22 [5.1, 2.5] Joint statement of central banks of the United States, England,
Japan, Canada, Switzerland, and Sweden, and the ECB. Federal
Reserve action: Term Securities Lending Facility (TSLF), Fed
lends up to $200 billion of Treasury securities against agency
debt, agency MBS, and non-agency AAA/Aaa-rated MBS
2008-01-22 –26 [6.0, 3.5] 75 bps cut
2007-11-15 –16 [3.7, 2.1] Bad job claims report
2004-08-06 –17 [3.8, 3.8] Bad non-farm payrolls report
2004-05-07 +19 [4.3, 3.4] Unexpectedly strong employment report
2002-08-14 +16 [3.7, 2.4] (No relevant news)
2002-03-08 +21 [4.7, 3.4] Chairman Greenspan provides a positive outlook, saying that an
expansion is already “well under way”
2001-12-07 –16 [3.6, 2.1] (No relevant news)
2001-11-29 –20 [4.5, 2.4] Correction following Enron and Japan downgrade
2001-09-13 –49 [11.0, 3.9] Market reopens after terrorist attacks, Fed will “provide whatever
liquidity might be needed”
2001-01-05 –19 [4.3, 2.3] Weaker than expected non-farm payrolls
2001-01-02 –18 [4.1, 2.2] (No relevant news)
1998-10-16 –24 [5.4, 3.0] One day after 50 bps rate cut

Overall, we note that the larger volatility of long factor innovations in the “zero lower bound”
phase is only partly due to announcements concerning unconventional monetary policy
measures. Of the twenty-five events reported in Table 2, thirteen occurred in the “zero lower
bound” phase. Only four of these large innovations are related to announcements of the Fed
concerning unconventional monetary policy. These are the enlargement of QE 1 (March 18,
2009, –53 bps), the confirmation of the reinvestment policy (December 14, 2010, +25 bps),
one day after Operation Twist 2 was announced (September 22, 2011, –23 bps), and finally
the “forward guidance” announcement (August 9, 2011, –22 bps). All other major shocks

BIS Papers No 65 269

were due to business cycle surprises or are related to the crisis of the European currency
union. This point is nicely illustrated in Figure 7.

Figure 7
Major long factor innovations in the zero lower bound phase. The last five months of the sample are depicted
separately in the right-hand panel because there is more action there.

5. News
It is well-documented that economic news releases and in particular surprises from market
expectations move Treasury yields (see e.g. Fleming and Remolona, 1999). Not surprisingly,
a similar relationship holds for our factors. A natural question is whether or not the changing
dynamics of the yield curve can (in part) be explained by changing dynamics in terms of
economic news surprises. Table 5 shows the results of regressing innovations into our long,
short, and curvature factors on day-of-release surprises for a range of economic indicators.
The indicators we consider include the Conference Board Consumer Confidence Index®, the
Institute for Supply Management (ISM) purchasing managers index (PMI), the advance GDP
print, the unemployment rate, industrial production, retail sales, housing starts, one-family
houses sold, the personal consumption expenditure price index, capacity utilization, initial
jobless claims, the leading economic indicator index, and the federal funds rate target. We
measure surprises as the difference between the actual value released and the median value
from an “expectations” survey among Wall Street economists conducted by Bloomberg News
prior to the release. To put the surprises on a common scale we standardized them by their
standard deviation over the sample. Moreover, we switch the sign of some surprise
variables, so that a positive surprise is “good news”. For instance, non-farm payroll surprises
are measured as actual release minus median expectation, whereas jobless claims are
defined the other way around. In addition, we control for non-linear effects by including the
squared standardized surprises as additional regressors.
Besides the results for the entire sample, we also split our sample in two with a view to
investigating whether or not the impact of economic news has changed with the financial
crisis. Our “pre-crisis” sample runs from 1998 to August 8, 2007 (“normal” and “great
moderation” phases) and the “crisis” sample covers the remainder of our sample (“liquidity

270 BIS Papers No 65

crisis” and “zero lower bound” phases). Furthermore, because our left-hand variables exhibit
clustered volatility (see Section 3 and Figure 4 in particular), we use the EGARCH
specification. In order to capture general market volatility we add the VIX as an exogenous
variable to the variance equation. We also add dummies for our phases that we identified in
Section 3 to the variance equation.
Consistent with previous literature we find that the non-farm payrolls are among the most
informative signals. This was the case before the crisis, and has remained so: non-farm
payroll surprises (linear and squared) have highly significant effects on all three yield curve
factors. PMI surprises used to be significant predictors of all three innovations before the
crisis; since the crisis they contain information only on long innovations. Surprises about
retail sales and about capacity utilization used to contain information on long and curvature
innovations before the crisis; in the crisis, retail sales surprises seem to no longer affect the
curvature, and capacity utilization has lost its connection to the yield curve completely.
Surprises about jobless claims used to affect curvature innovations before the crisis, but now
affect long innovations instead. Surprises concerning the FOMC’s federal funds target rate
used to be highly significant with respect to short factor innovations; they have lost their
explanatory power during the crisis. Prior to the financial crisis, our economic surprise
indicators could account for 4.5%, 1.5%, and 6.6% of the variation in the long, short, and
curvature factors (as measured by the R2-statistic). In our crisis sample, the comparable
numbers are 3.2%, 0.3%, and 0.9%.
The phase dummies in the variance equation partially verify our partition of the sample into
four phases. Interestingly, the “great moderation” dummy is not significant. The general
reduction of the volatility of our innovations between 1998 and the beginning of the financial
crisis seems fully captured by the VIX. The two other phase dummies, however, come in as
significant, as expected. The “liquidity crisis” dummy measures a higher volatility for short
and curvature innovations, but is not significant in the variance equation of the long
innovations. The “zero lower bound” dummy, on the other hand, measures a significantly
higher volatility of long factor innovations, but significantly lower volatility of the short factor
innovations. With respect to curvature innovations, this coefficient is either negative (pointing
to a reduced volatility of curvature shocks in this phase) or statistically insignificant.

6. Conclusions
The financial crisis has deeply affected financial markets as well as the economy as a whole.
This has also affected the yield curve and its dynamics. We document these changes in this
paper. Our main results can be summarized as follows: Firstly, we divide the dynamics of the
yield curve into four phases. The first two phases occur prior to the financial crisis. The
second phase is characterized by substantially less volatility of the yields compared to the
first phase. However, this is well explained by the simultaneous decline of overall financial
market volatility during that period as measured by the VIX.
The third and the fourth phase comprise the financial crisis. This means that we can divide
the crisis into two distinct subperiods. In the first subperiod, the yield curve experienced very
strong shocks in the short and medium maturity spectrum due to the freezing of the money
market and subsequent emergency measures taken by the Federal Reserve. The second
part of the financial crisis began when the federal funds rate reached the zero lower bound.
From that point forward, we find an absence of shocks hitting the yield curve at low and
medium maturities. Instead, the longer end of the curve experiences greater disturbances
than before.
Secondly, we perform a (reverse) event study in which we match the greatest shocks to the
yield curve with headline news. We find that some large shocks are associated with

BIS Papers No 65 271

announcements by the Federal Reserve. However, a significant number of shocks in
particular in the recent past are due to international developments.
Thirdly, we identify and quantify the informational content of well-known macroeconomic
surprise data. We find that that some, but not all of these variables have lost significance in
the crisis. The overall information content of these news variables with respect to the yield
curve, however, is small.

Table 5
Effects of macroeconomic news on the yield curve

variable long innovations short innovations curvature innovations

sample pre-crisis crisis all pre-crisis crisis all pre-crisis crisis all
constant –0.033 –0.090 –0.065 0.235 0.530 0.328 0.206 –0.195 0.122
standardized surprise in:
Capacity
utilization 2.205** 0.813 2.186** –0.740 –0.457 –0.575 1.233** –0.224 0.808**
Consumer
confidence 1.350** –0.311 1.135** –0.358 0.191 –0.135 0.449 0.224 0.360
Initial jobless
claims 0.975 2.955** 1.620** –0.027 –0.160 –0.174 1.005** –0.553 0.661**
Federal funds
target –2.689 –7.409 –1.334 –12.83*** –1.362 –10.40*** –0.471 3.792 –1.001
Advance GDP –0.668 –2.094 –0.809 0.018 0.697 –0.038 0.528 –0.654 0.571
One-family
houses sold 0.996** 0.711 0.999** –0.129 –0.076 –0.102 –0.047 0.502 0.010
Housing starts –0.407 2.167 –0.179 –0.103 –1.067 –0.264 0.055 1.795* 0.126
Industrial
production –0.990 2.364 –0.694 0.439 0.098 0.334 –0.480 0.537 –0.199
ISM PMI 1.597*** 3.136** 1.941*** –1.305*** 0.348 –0.483* 1.370*** –0.620 0.888***
LEI 0.274 0.175 0.175 –0.833 –0.287 –0.502 –0.708 –0.574 –0.754**
Non-farm
payrolls 3.578*** 4.786** 3.761*** –2.009*** –2.390*** –2.206*** 3.117*** 1.938** 2.948***
PCE price
index 0.356 –3.064* –0.482 –1.324* –0.351 –0.724 –0.141 –0.345 –0.124
Retail sales 1.445** 5.530*** 2.110*** –0.570 0.933 –0.143 1.176*** –1.490* 0.581*
Unemployment
rate 0.783 1.477 0.961* –0.542 –0.524 –0.644** 2.065*** 0.837* 1.477***

272 BIS Papers No 65

 Table 5 (cont)
Effects of macroeconomic news on the yield curve
variable long innovations short innovations curvature innovations
sample pre-crisis crisis all pre-crisis crisis all pre-crisis crisis all
constant –0.033 –0.090 –0.065 0.235 0.530 0.328 0.206 –0.195 0.122
squared standardized surprise in:
Capacity
utilization –0.655 –1.615 –0.968* 0.190 –0.049 0.049 –0.657* 0.809** –0.052
Consumer
confidence –0.119 –1.337** –0.286 –0.391 –0.379 –0.390** –0.085 0.67** 0.133
Initial jobless
claims 0.101 0.096 0.184 –0.296 0.018 –0.232* 0.197 0.024 0.149
Federal funds
target –0.331 11.666 1.339 –7.455** –6.158 –6.005** 4.580 –11.21 3.450*
Advance GDP 0.715 0.118 0.681 –1.056** 0.155 –0.518 0.459 –0.468 0.455
One-family
houses sold –0.090 1.541 –0.063 –0.082 –0.962 –0.137 0.083 0.463 0.047
Housing starts –0.098 –1.663 –0.143 –0.086 –0.704 –0.142 –0.174 0.323 –0.169
Industrial
production 0.220 1.513 0.627 –0.180 –0.008 –0.015 0.740* –0.315 0.179
ISM PMI 0.809*** –0.025 0.626*** –0.290 –0.377 –0.354** –0.199 0.213 –0.062
LEI 0.668 0.594 0.580 0.198 0.253 0.314 0.204 –0.066 0.195
Non-farm
payrolls 0.782*** 1.031 0.821*** –0.622*** –1.396*** –0.774*** 0.326** 1.989*** 0.296**
PCE price
index –0.826 –0.677 –0.747 1.235* 1.079* 1.086** 0.202 0.276 0.304
Retail sales 0.223 0.078 0.102 0.023 0.135 –0.056 0.023 –0.167 0.180
Unemployment
rate 0.066 0.477 0.145 0.000 0.670** 0.490** –0.456 –0.900*** –0.438**
Variance equation
constant 0.005 2.658** 0.043 –0.086** 0.284** –0.001 –0.067** 0.179 –0.046**

  1 /   1 0.077*** 0.092 0.098*** 0.260*** 0.271*** 0.298*** 0.144*** 0.251*** 0.178***
  1 /   1 0.013 0.034 0.012 –0.061*** –0.112*** –0.072*** –0.024 –0.003 –0.021

log    1 
2
0.969*** –0.022 0.952*** 0.931*** 0.887*** 0.907*** 0.965*** 0.891*** 0.956***
great moderation
dummy 0.002 –0.006 –0.004 –0.039* 0.000 –0.014
liquidity crisis
dummy 0.015* 0.145*** 0.051***
zero lower
bound dummy 0.415** 0.026** –0.288*** –0.086*** –0.166** –0.010
VIX 0.002*** 0.043** 0.002*** 0.004*** 0.001 0.003** 0.002** 0.002 0.001**
2
R 0.045 0.032 0.035 0.015 0.003 0.007 0.066 0.009 0.033
Note: *,**,*** denotes significance at the 10%, 5%, and 1% level, respectively.

BIS Papers No 65 273

 Appendix A.:
Spot and local volatility

This is a purely technical appendix which is not necessary to understand the economic
content of the paper. It explains the concept of “local volatility” that is used in some places in
the main part of the article.
We aim at quantifying the changing volatility of the innovations u . Simple visual inspection
suggests heteroscedasticity. We fully acknowledge that this feature of the data is not in line
with the specification of the model. After all, we assumed normally distributed homoscedastic
innovations when estimating the loadings and the VAR with maximum likelihood. Taking the
heteroscedasticity fully into account at the estimation stage of the model seems very
challenging. Being aware of this inconsistency, here we simply aim to measure the stochastic
volatility of the innovations as they present themselves from the model that was estimated
assuming homoscedasticity.
Consider a continuous-time diffusion,
dX t  t dt   t dWt , (A.1)

where Wt is a standard Brownian motion.  t2 is the spot variance process, which is not
observed. Instead, we observe only X t at discrete points in time, t1  t 2  ...  t n . Based on
 2
work by Bandi and Phillips (2003), Kristensen (2010) establishes that one can estimate
as

 
K h  t i 1    X t i  X ti 1 
n 2

ˆ
2
 i 1
, (A.2)
 K h  t i 1   
T
i 1

where K h is a kernel function with bandwidth h . This is a Nadaraya-Watson kernel

regression on the squared first difference of X . Spot volatility is simply the square root of the
estimated spot variance. In our application, X is one of the factor innovations in the VAR
model, i.e. uf for f  1,2,3 .

In order to estimate spot volatility, two choices need to be made, namely the specification of
the kernel function K and the selection of the bandwidth h . To select the bandwidth, we use
the cross-validation technique; that is, we minimize the mean squared “leave-one-out”
residuals. The kernel function K is symmetric around zero if it weighs observations in the
future the same way as observations in the past. Most popular kernel functions have this
property. Symmetric kernel functions have, however, the disadvantage that for  close to the
edge of the sample, they assign positive weights to observations outside the available
sample, which biases the estimation. This is a well-known problem in non-parametric
econometrics.
One way to address this problem is to use a locally adapting kernel function. Such a function
was for instance proposed by Brown and Chen (1999) and Chen (2000). Their kernel
function, based on the beta-function, automatically adapts to the boundaries of the sample:
for  close to the first observation t1 , the kernel is right-sided, for  close to the last
observation t n , the kernel is left-sided. We have experimented with this kernel but found it to
give unsatisfactory estimates in our application. The volatility measure has significantly more
high-frequency variability close to the edge of the sample than in the interior, which suggests
that the precision of the estimate deteriorates close to the edge, or that the bandwidth
becomes too small.

274 BIS Papers No 65

For this reason, we resort to an older, simpler idea that was proposed by Schuster (1991). It
consists of reflecting observations close to the edge of the sample to the other side. So,
..., X t , X t , X t  are appended in reverse order as  X t , X t , X t ,... , and then the
 n 3 n 2 n 1   n 1 n  2 n3 
symmetric kernel function is applied to these synthetically expanded observations. We use
the popular (symmetric) Epanechnikov specification as the kernel function.
This procedure as described so far is, however, not very successful in our application. The
estimated spot volatilities turn out to be much too large on average. Only about 1% of the
absolute innovations are greater than 1.96 times the estimated spot volatilities. It is not
completely clear why this is the case. It may be due to the fact that equation (A.1) is not the
correct model for the innovations u . After all, these are residuals and they have, by
construction, no drift, so t  0 .9

Because the spot volatility does not appear reasonable, we compute a slightly simpler and
maybe more transparent measure. We apply the Nadaraya-Watson kernel regression on the
squared innovations directly instead of on the squared first differences,

 K t    X
n 2

  i 1
2 h i ti
. (A.3)
 K t   
T
i 1 h i

This is the same as the approach suggested by Carroll (1982) and Hall and Carroll (1989).
They consider a model where the mean can be parametrically estimated but the variance
cannot. In our case, t is zero by definition, so the setting is simpler.

Figure A.8
Curvature innovations and 50% confidence interval using local volatility estimate

We use the same kernel function and reflection technique as before and perform the cross-
validation bandwidth optimization. The result is an estimate of the volatility that seems much
more reasonable. We call the square root of 2 the local volatility, in order to distinguish it

Consider E  ut  ut 1    E ut2   E ut21   2E ut ut 1  . In our case, u is serially uncorrelated by

9 2
 
construction, so E ut ut 1  0 . Consequently, the spot variance overestimates the variance of u by a factor

of two, E  ut ut 1    E ut2   E ut21 .

2
 

BIS Papers No 65 275

from the spot volatility. The size of this volatility measure appears more appropriate: 4.7% of
the absolute innovations into the long factor are greater than 1.96 times the estimated local
volatility of this factor innovation. For curvature, the corresponding number is 4.8%,
reasonably close to the 5% one might expect. Only for the short factor innovations do we find
that only 3.3% of the innovations are in absolute terms greater than 1.96 times the estimated
local volatility. Still, this is much better than the 1% we get when using spot volatility. The
optimized bandwidths are 47.0 days for the long factor innovations, 9.7 days for the short
factor innovations, and 23.5 days for the curvature innovations. Figure A.8 depicts, as an
example, the innovations into the curvature factor, as well as a 50% confidence band using
the local volatility estimate.

References
Bai, J., Perron, P., 1998. Estimating and testing linear models with multiple structural
changes. Econometrica 66, 47–78.
Bai, J., Perron, P., 2003. Computation and analysis of multiple structural change models.
Journal of Applied Econometrics 18, 1–22.
Bandi, F., Phillips, P., 2003. Fully nonparametric estimation of scalar diffusion models.
Econometrica 71, 241–283.
Brown, B., Chen, S., 1999. Beta-Bernstein smoothing for regression curves with compact
support. Scandinavian Journal of Statistics 26, 47–59.
Carroll, R., 1982. Adapting for heteroscedasticity in linear models. The Annals of
Statistics 10, 1224–1233.
Chen, S., 2000. Beta kernel smoothers for regression curves. Statistica Sinica 10, 73–91.
Fleming, M., Remolona, E., 1999. Price formation and liquidity in the U.S. Treasury market:
The response to public information. Journal of Finance 54, 1901–1915.
Hall, P., Carroll, R., 1989. Variance function estimation in regression: The effect of estimating
the mean. Journal of the Royal Statistical Society, Series B 51, 3–14.
Kristensen, D., 2010. Nonparametric filtering of the realized spot volatility: A kernel-based
approach. Econometric Theory 26, 60–93.
Lengwiler, Y., Lenz, C., 2010. Intelligible factors for the yield curve. Journal of
Econometrics 157, 481–491.
Litterman, R., Scheinkman, J., 1991. Common factors affecting bond returns. The Journal of
Fixed Income 1, 54–61.
Nelson, C., Siegel, A., 1987. Parsimonious modeling of yield curves. The Journal of
Business 60, 473–489.
Schuster, E., 1991. Incorporating support constraints into nonparametric estimators of
densities. Communications in Statistics: Theory and Methods 14, 1123–1136.

276 BIS Papers No 65