1

The Global Financial Crisis, Commercial Paper Market,

and Regulatory Responses


Mark D. Griffiths
Visiting Professor of Finance
New York University

Vladimir Kotomin*
University of Wisconsin Eau Claire

Drew B. Winters
Lucille and Raymond Pickering Chair in Finance
Texas Tech University



February 2010










*Corresponding author:
Department of Accounting and Finance
University of Wisconsin Eau Claire
105 Garfield Ave.
Eau Claire, WI 54702-4004
Phone 715-836-3535, Fax 715-836-3582
Email kotomin@uwec.edu
2


Abstract

Money markets have been stressed during the global financial crisis of 2007-2009. The two main
explanations for the crisis in the money markets are (1) credit concerns and (2) liquidity
concerns. Credit risk and liquidity risk are intimately related, especially in the money markets,
and either can lead to similar behavior by market participants, such as hoarding of cash. We
study the U.S. commercial paper (CP) market to draw insights about the nature of the crisis. The
crisis of 2007-09 has had a profound impact on the CP market with the amount of outstanding
CP shrinking from the peak of $2.18 trillion in early August 2007 to $1.27 trillion in early July
2009. The CP market is not homogenous in terms of credit quality, maturities, and types of
issues. We show that not all CP issuers suffered equally during the crisis. Our results suggest that
the crisis in the money markets is related more to increases in perceived counterparty (credit)
risk, with liquidity concerns being secondary.

3

The Global Financial Crisis, Commercial Paper Market,
and Regulatory Responses


Im not a financial engineer; Im a financial
pathologist (t)he reason pathologists study the
dead, though, is . . . that from what we learn, we can
help the living.

Stephen A. Ross,
Forensic Finance: ENRON and
Others,
Fourth Angelo Costa Lecture.

1. Introduction:

The financial crisis of 2007-2009 provides financial pathologists with many and varied
opportunities to study what market mechanisms went wrong, so appropriate policies can be put
in place going forward. Many of the opportunities are in: real estate, mortgages, lending
practices, financial institutions, securitization, shadow banking, and money markets. Initial
analyses of the then existing policies include, among others, Adrian and Shin (2008) and Mizen
(2008).
This paper contributes to the understanding of the problems in the money markets the
market for short-term borrowing and lending with original maturities of one year or less that are
designed to provide liquidity funding for the global financial system. Instruments traded in this
market include Treasury bills, commercial paper, bankers' acceptances, certificates of deposit,
federal funds, and short-term mortgage-backed and asset-backed securities. Specifically, we
examine the effects of the recent financial crisis on components of the commercial paper market.
It is generally accepted (Taylor and Williams, 2009) that the current financial crisis
commenced in earnest on August 9, 2007 when BNP Paribas, a large French bank temporarily
halted redemptions from three of its funds which held assets backed by US subprime mortgage
4

debt. As a result, international money market traders found that the interest rate on term loans
increased dramatically to the point where rates on term lending such as one- and three-month
Libor appeared to become disconnected from the overnight right. Cecchetti (2008) provides a
succinct description of the situation:
Normally less than 10 basis points, on Thursday 9 August 2007 the 3-month
Libor-OIS
1
jumped to 40 basis points. And normally stable, with a standard
deviation of several basis points, the spread fluctuations between 25 and 106 basis
points through the fall. Something is clearly wrong. There should be an arbitrage
that allows a bank to borrow overnight, lend for three months, and hedge the risk
that the overnight rate will move in the federal funds futures market leaving only
a small residual level of credit and liquidity risk that accounts for the small spread
we see before the crisis. (pg 14)

As Taylor and Williams (2009) state, (i)t was as if banks suddenly demanded more liquidity or
had grown reluctant to lend to each other, perhaps because of fears about the location of newly
disclosed losses on sub-prime mortgages.
On August 16, 2007 Countrywide Financial (the largest US mortgage lender at the time)
announced it had borrowed $11.5 billion from a group of 40 banks. The loan was necessitated
by Countrywides inability to borrow the funds in the commercial paper market, as it had done
regularly in the past.
2
Concerns that numerous sub-prime mortgage loans Countrywide had
made were lowering the quality of its assets resulted in Countrywide losing its access to the

1
Overnight Indexed Swap rate represents the expected interest rate that would accrue from repeatedly rolling over
an investment at the overnight rate for three months.

2
Commercial paper is generally an unsecured promissory note with a fixed maturity of 1 to 270 days issued by large
banks and corporations to obtain funds to meet short term obligations, and is backed by the issuer's promise to pay
the face amount on the maturity date. Since it is not backed by collateral, only firms with excellent credit ratings
from recognized rating agencies will be able to sell their commercial paper at reasonable prices. CP rates fluctuate
with market conditions, but are typically lower than banks' rates. With the introduction of wide-spread securitization
of loans and receivables in the mid-1980s, a new class of collateralized commercial paper appeared: asset-backed
commercial paper (ABCP). We discuss this further in section 2.2.

5

commercial paper market.
3
The Countrywide announcement was one of the first of many
financial news stories about the inability of borrowers to access the commercial paper market.
Additionally, there were stories about the inability to borrow in the repurchase agreement (repo)
market.
4
Both sets of stories were generally accompanied by comments of a liquidity crisis. The
Federal Reserve agreed (at least initially) with the consensus opinion and began undertaking
remedial measures to treat such a crisis. McAndrews, Sarkar, and Wang (2008) and Wu (2008)
implicitly assume and Anderson and Gascon (2009) suggest that the crisis was mostly a liquidity
crisis. Bernanke (2009) discusses both liquidity and credit issues in his 13 January 2009 speech
but seems to focus more on liquidity. However, Taylor (2009) and Taylor and Williams (2009)
suggest that the Federal Reserve treated the symptoms but not the cause. They suggest that the
problem in the money market was a credit crisis and not one of liquidity.
Unlike the earlier studies which were concerned with a macro-finance or monetary policy
point of view, we examine the crisis from a micro-perspective by studying the markets for
commercial paper of various terms and quality. By doing so, we can determine whether the
problems in specific segments of the commercial paper market were related to liquidity or credit
issues. Also, we avoid many of the controversies that abound in the earlier studies. Specifically,
we are able to avoid issues dealing with allegations of mis-reporting of Libor
5
or with the most
appropriate method of entering the timing for liquidity injections or Libor reporting.
6


3
See Adrian and Shin (2008) for a detailed explanation of contagion in the subprime mortgage market.

4
A repurchase agreement is a short-term collateralized loan where a security is exchanged for cash with the
agreement that the parties will reverse the transaction at a pre-arranged price at a specified time in the future.

5
Mollenkamp, Carrick, Bankers Cast Doubt on Key Rate amid Crisis, Wall Street Journal, April 15, 2008.

6
See, Taylor and Williams (2009)

6

Our results suggest that an increase in perceived credit (counterparty) risk was the
primary force shaping the commercial paper market during the crisis with liquidity concerns
playing a secondary role. Therefore, simple injections of liquidity into the financial system
would not be very effective as they do not address the root cause of the crisis.
7
On the other
hand, central banks and other regulators taking on the functions of market makers and credit risk
insurers may have had a positive impact on the situation in the financial markets.

2. Background
We begin with a discussion of the economic role of the money market and the primary
characteristics of all money market securities. We continue by covering essential details of the
commercial paper market. From this discussion, we present the testable hypotheses. Next, we
summarize the actions of the Federal Reserve in addressing a financial crisis. The background
on the Feds actions is the basis for any policy discussion that might stem from our empirical
results.

2.1. The Economic Role of the Money Markets
Kidwell, Blackwell, Whidbee, and Peterson (2006) state (t)he money markets are where
depository institutions and other businesses adjust their liquidity positions by borrowing or
investing for short periods of time. (p. 168). They further state (t)he most important economic
function of the money markets is to provide an efficient means for economic units to adjust their
liquidity positions. (p. 170). Thus, a reasonable definition of the economic role of the money
markets is to provide for the trading of liquidity.

7
The authors recognize that liquidity injections have historically shortened the length of recessions following
financial crises (Bruner and Carr, 2009). It may be too early to determine whether this applies equally to the current
crisis. Nonetheless, we concentrate on the proximate cause of the crisis.
7

Mishkin and Eakins (2009) highlight the temporary nature of investments in the money
markets in the following description of the purpose of the money markets:
The well-developed secondary market for money market makes
the money market an ideal place for a firm or financial institution
to warehouse surplus funds until they are needed. Similarly, the
money markets provide a low-cost source of funds to firms, the
government, and intermediaries that need a short-term infusion of
funds. (p. 214)

The temporary nature of investments in the money markets requires two primary
characteristics of money market securities: (1) high liquidity and (2) low default risk. High
liquidity is achieved through the use of short-term debt contracts and by trading through dealers.
Low default risk is achieved by lending to (investing in) only high credit-quality borrowers.
Clearly, liquidity is of primary concern. However, the temporary nature of all money
market investments makes credit quality equally important. A potentially confounding issue is
that banks fearing a write-down in their own assets may decline to lend regardless of the credit
quality of the counterparty in order to preserve funds for their own needs (Mizen, 2008).

2.2. Commercial Paper Market and Testable Hypotheses
Commercial paper (CP) is a class of short-term debt of private corporations with
maturities between one and 270 days. Most of CP is unsecured (backed only by general
creditworthiness of an issuer), with the exception of asset-backed CP.
Asset-backed CP arises from the securitization of loans and receivables. Briefly, the
lender makes the initial loan and instead of holding it on its balance sheet, it puts it into a pool
containing a large number of similar assets. This pool is the collateral for asset-backed securities
which may take the form of commercial paper. The owners of these asset-backed securities
8

receive the payments from the borrowers whose loans are in the pool. The most extensive use of
these pass-through securities was for mortgages in particular, for higher risk subprime
mortgages that did not qualify for insurance from such government-sponsored entities (GSEs) as
the Federal National Mortgage Association and the Federal National Mortgage Corporation.
Greenlaw, Hatzius, Kashyup and Shin (2009) report that from 2001 to 2006 the percent of
mortgage originations, as measured by value, fell in GSEs, while the percent of mortgages in the
lower quality non-GSE category rose from 9.7% to 33.5% of mortgages issued.
8

Because high credit quality is essential in money markets, only the highest-rated
companies have access to this market: between 95 and 99% of all commercial paper outstanding
has historically fallen into the highest two CP rating categories, AA and A2/P2. Although the
number of CP issuers is limited (fluctuating between 500 and 1,200 over the past two decades),
the size of the market exceeded $1 trillion throughout the 2000s. Because short-term funds are
usually cheaper than long-term credit, some companies have financed long-term projects by
rolling over their commercial paper (i.e., refinancing maturing CP issues with new ones). This
strategy is exposed to risk if short-term interest rates increase or access to the CP market dries up
for the firm (Milne, 2009). Some large, sophisticated issuers, especially financial institutions,
place their paper directly, although most issuers resort to dealers for placement purposes.
Historically, while most investors in this market have bought and held commercial paper to
maturity, a handful of dealers that help issuers place their paper also stand ready to buy and sell,
ensuring the liquid secondary market.

8
Quoting from Bate, Bushweller, and Rutan (2003, The fundamentals of asset-backed commercial paper
Moodys Special Report), Asset-backed commercial paper (ABCP) is a form of senior secured, short-term
borrowing, in contrast to corporate commercial paper, which is senior unsecured short-term corporate debt. Asset-
backed commercial paper programs offer low-cost financing to companies that could not otherwise directly borrow
in the commercial paper markets. These firms could not borrow directly in the CP market due to their lower credit
quality. The rating of a partially-supported ABCP program depends on the performance of the programs assets. The
rating of a fully-supported ABCP program is directly linked to the credit strength of the guarantor. See the report for
more details (available on Ian Giddy's webpage at http://pages.stern.nyu.edu/~igiddy/ABS/moodysabcp.pdf)
9

We collected rates and amounts outstanding for different classes of commercial paper
which vary by type and credit quality of the borrower. We have CP issued by [1] financial
institutions (financial CP), [2] non-financial firms (non-financial CP), [3] asset-backed
commercial paper (ABCP), [4] high quality issuers (AA), and [5] lower quality issuers (A2/P2).
If the problems in the commercial paper market were related primarily to liquidity, we
expect to find that all term commercial paper rates increase by similar amounts and that the
amounts outstanding of all commercial paper decline. If the observed problems relate to credit
quality, we expect that rates increase more and amounts outstanding decline more for the lower
credit quality classes than for the higher credit quality classes of commercial paper. Similarly, if,
as has been generally accepted, the quality of sub-prime mortgage collateral was the motivating
concern for the crises, then we would expect rates of asset-backed CP to increase more and that
the amounts outstanding decline more than for non-financial CP.

2.3 The Federal Reserves Response to the Problems in the Money Markets
Rather than relying on financial news reports that the Federal Reserve was attempting to
solve the liquidity crisis in the money markets, we review the chronology of Fed actions based
on two forms of evidence on the Feds actions during the financial crisis: (1) changes in the
policy target funds rate and (2) reports on Federal Open Market Committee (FOMC) meeting
discussions.
The Fed raises the target rate to signal restrictive monetary policy and lowers the target
rate to signal easier availability of funds. The target funds rate was 5.25% entering the financial
crisis. As the crisis progressed, the Fed continually reduced the target funds rate eventually
setting it in the range of 0-25 basis points on December 16, 2008 which has been maintained
10

through the end of 2009. The decline of the target rate throughout the financial crisis is strong
prima facie evidence that the Fed policy was to add monetary reserves to the financial system
and this action is consistent with treating a liquidity shortage in the money markets.
Second, we examine the FOMC meeting minutes from December 2006 through
December 2008 for Fed actions and policy discussion on the financial crisis. In Appendix A, we
provide a brief outline of the FOMC discussions during the period, focusing on the money
markets. Our interpretation of FOMCs discussions and actions is that the Fed clearly discussed
both liquidity and credit quality in its deliberations during the financial crisis. However, when
the Fed acted, it primarily strived to increase liquidity with little action to address credit quality
concerns until after the bankruptcy of Lehman Brothers.

3. Data
The primary data are daily commercial paper yields from the Federal Reserve (Fed). The
Fed reports yields for 1-day, 7-day, 15-day, 30-day, 60-day, and 90-day CP. The yields are
reported for four categories of CP: AA asset backed, AA financial, AA non-financial, and A2/P2
non-financial, as well as for the CP market total. While we examined all available maturities, we
focus on 30-day and 1-day maturities in the remainder of the paper to avoid complications
arising from time-varying default premia.
9
These daily data are available starting on January 2,
2001 and our data set continues through July 2, 2009. We use two alternative controls for the
general level of short-term interest rates: the target federal funds rate and the constant maturity
one-month T-bill yield. The T-bill yields are determined by the market, while the target federal
funds rate is set by the Fed.

9
Fama (1986) shows that default premia in the money markets decline with maturity and are higher during
recessions while term premia increase with maturity but humps and inverses occur in recessions.

11


4. Empirical Analysis
We conduct our empirical analysis of the commercial paper market in four parts. We
begin with plots of the commercial paper rate and spread series to identify extreme changes.
When we observe such extremes, we examine the rates around the events to determine if all
commercial paper rates increase or if the increases are limited to certain maturities or segments
of the market. We then extend this analysis to amounts outstanding. Third, we extend our
analysis into a multivariate framework to assess statistical significance after using appropriate
controls. Finally, we conduct a brief robustness analysis.

4.1. Rate and Spread Plots
Figures 1 and 2 plot 30-day and one-day CP rate series, respectively, from January 2,
2007 through July 2, 2009. Panel A in each figure contains yields of AA financial and AA non-
financial CP (non-collateralized and higher credit quality), while Panel B contains yields of AA
asset-backed (collateralized) and A2/P2 non-financial CP (lower credit quality).
10
Both figures
also contain the target federal funds rate
11
and yield on one-month T-bills (constant maturity).
12

The figures also contain labels for three newsworthy events: (1) the BNP Paribas announcement
on August 9, 2007 that it had suspended the valuation of its asset-backed funds, (2) the failure of
Bear Stearns on March 14, 2008, and (3) the bankruptcy of Lehman Brothers on September 14,

10
Anderson and Gascon (2009) describe financial and non-financial CP as unsecured short-term corporate debt
issued by highly rated companies. They describe asset-backed CP as a securitized claim issued by a conduit where a
manager selects the portfolio of assets held and CP issued against the portfolio.

11
We recognize as does Taylor and Williams (2009) that there is a timing issue with respect to the target federal
funds rates. Close examination of the data reveal (not presented) that decreases in the target federal funds rate lags
changes in the effective fed funds rates.

12
To construct the plots and later to run regressions, we set the target federal funds rate equal to 0.25% during the
period when it was specified as a range between 0 and 25 basis points.

12

2008 which also led to the Reserve Primary money market fund breaking the buck (i.e., having a
negative return on investment, an extremely rare occasion for money market funds).
13
We do not
analyze these events specifically. Instead, we analyze the commercial paper markets behavior
around these three events.
14

Panel A of Figure 1 reveals several interesting features. First, around each of the three
events the yield on the one-month T-bill declines dramatically. This is consistent with either a
flight to liquidity or a flight to quality. Second, the target funds rate declines across the crisis
consistent with the Fed treating a liquidity crisis. Third, from January 2, 2007 through the failure
of Bear Stearns, AA financial and AA non-financial CP rates track each other closely.
Following the Bear Stearns event, the spread between the two series becomes positive; the
spread then increases dramatically following the Lehman bankruptcy. The spread between the
CP rates in Panel A following Bear Stearns is approximately 50 basis points.
15
While we do not
conduct statistical tests at this point, we note that a 50 basis point spread is an attractive
investment opportunity in the money market under normal market conditions (Stigum (1990)).
16

The series with the higher rates is the financial CP series. Following the Bear and Lehman

13
Quoting the BNP press release on August 9, 2007, The complete evaporation of liquidity in certain market
segments of the U.S. securitization market has made it impossible to value certain assets fairly regardless of their
quality or credit rating. Armantier, Krieger, and McAndrews (2008) commenting on this matter observe: This
public statement of a condition that many firms had been grappling with uncertainty about the true value of
various financial securities created from residential mortgages affected the interbank money markets in Europe
and the United States like a match in a dry forest.

14
While dozens of events may have influenced money markets during the crisis, these three appeared to have had
the most profound impact on the commercial paper market. Accordingly, we focus on these events throughout the
paper. Afonso, Kovner, and Schoar (2009) use similar events in their analysis of the federal funds market during the
crisis.

15
Following Bear Sterns failure, the 15-day spread (not reported) was often around 20 basis points. Overall, term
yields and spreads other than 30-day ones (e.g., 15-day or 60-day) in all categories behave similarly to 30-day yields
and spreads. Overnight yields and spreads behave differently in some cases, and are discussed later in the paper.

16
Stigum (1990) notes that money market traders consider 10 to 20 basis points of abnormal spread an attractive
opportunity.

13

events, the market generally had concerns about the viability of many financial institutions.
Accordingly, the higher rates on financial CP are consistent with a counterparty risk concerns in
the CP market during the financial crisis.
Panel B of Figure 1 plots 30-day rates of higher-risk CP categories AA asset-backed
and A2/P2 non-financial paper.
17
We carry forward the plots of the target rate and one-month T-
bill yield and continue the use of the three events as reference points. First, Panel B of Figure 1
shows that following the BNP announcement the rates on asset-backed and A2/P2 CP increase
dramatically. The lower risk CP rates plotted in Panel A did not spike at this time suggesting this
reaction was related to credit issues and not liquidity issues. Second, 30-day rates in both asset-
backed and A2/P2 CP spike in December 2007. Panel A shows only a small change before the
2007 year-end in AA financial and AA non-financial rates. Panel B of Figure 1 suggests that
even with the Fed increasing liquidity, investors exited the higher risk CP prior to the end of
Year 2007.
18
Third, following the Bear Stearns event, the rates on asset-backed and A2/P2 paper
remained near 3 percent while the target funds rates and the rates on the lower-risk CP in Panel
A fell between 75 and 100 basis points. Again, the higher risk classes paid substantially higher
rates indicative of a credit quality issue. Finally, following the Lehman bankruptcy, the CP rates
in Panel B increase approximately 300 basis points to approximately 6 percent. The maximum
rate in the lower risk classes following Lehman is 4 percent (Panel A). The most persistent rate

17
We state that AA asset-backed CP is more risky than AA non-financial and AA financial CP while all three
clearly have the same rating. The securitization process allows the asset-backed CP to receive an AA rating.
However, the uncertainty of the specific assets held by the conduit and the uncertainty of the quality of the assets
held by the conduit caused investors to view AA asset-backed CP as more risky that AA non-financial and AA
financial CP.

18
Griffiths and Winters (2005a) show year-end rate increases across one-month money market securities and
conclude the effect is related to year-end liquidity preferences. Griffiths and Winters (2005b) examine different
classes of commercial paper and find a larger year-end effect in high risk classes. They conclude that investors with
year-end liquidity needs exit higher risk classes first.
14

increase is in the A2/P2 paper, the highest risk class. The Lehman bankruptcy thus appears to
have caused an increase in perceived credit risk.
19

In a liquidity crisis, lenders become unwilling to commit their funds for more than a few
days.
20
If lenders move toward overnight maturities, the supply of term funding decreases while
the supply of overnight funding increases. In a static analysis, lenders moving toward overnight
terms would increase the slope of the yield curve through both an increase in the term rates and a
decrease in the overnight rate. To gain insight at the beginning of the yield curve, we create
Figure 2 which plots AA Financial and AA non-financial overnight (one-day) CP rates in Panel
A and AA asset-backed and A2/P2 non-financial overnight rates in Panel B, with the same
reference rates, the target federal funds rate and the one-month T-bill yield.
Panel A of Figure 2 shows that lower-risk CP overnight yields (AA Fin and AA Non-Fin)
follow the target fed funds rate closely until after Lehman, when they experience a couple of
short-lived spikes.
21
Panel B of Table 2 shows that overnight yields in higher-risk CP categories
(asset-backed and A2/P2 non-financial) diverge from the lower risk CP overnight rates at the

19
The Federal Reserve Board approved a number of new facilities. On September 19, it announced the Asset-
Backed Commercial Paper Money Market Mutual Fund Liquidity Facility (AMLF), which extends nonrecourse
loans at the primary credit rate to U.S. depository institutions and bank holding companies to finance the purchase of
high-quality asset-backed commercial paper (ABCP) from money market mutual funds. On October 7, the Board
announced the creation of the Commercial Paper Funding Facility (CPFF) to provide a liquidity backstop to U.S.
issuers of highly rated commercial paper through a special-purpose vehicle that purchases three-month unsecured
commercial paper and ABCP directly from eligible issuers. On October 21, the Money Market Investor Funding
Facility (MMIFF) was created, under which the Federal Reserve Bank of New York provides funding to facilitate an
industry-supported initiative to finance the purchase of certain highly rated certificates of deposit, bank notes, and
commercial paper from U.S. money market mutual funds. According to FOMC reports, the AMLF, CPFF, and
MMIFF were intended to improve the liquidity in short-term debt markets and ease the strains in credit markets
more broadly.

20
Lenders may not be willing to commit funds to term CP because they fear that either that (1) the CP market will
not function properly and they will not be able to sell their CP before maturity if needed or (2) they will not be able
to raise short-term funds for themselves when needed. These concerns are related, with the latter being broader. On
the other hand, if perceived credit risk of CP issuers increases, lenders would either exit the CP market altogether or
require higher yields, even on overnight CP.

21
One-day financial CP may be a close substitute for a federal funds loan, and the Fed injected large amounts of
liquidity after the Lehman bankruptcy to keep the fed funds rate close to the target.
15

BNP announcement and remain higher through the remainder of the analysis period. The
combination of rates on the lower risk CP remaining clustered together around the target rate
with higher rates on the higher risk CP continues to suggest a credit risk problem in the term CP
market during the financial crisis.
To complement our rate plots, Figure 3 graphs various spreads. Specifically, Figure 3
Panel A presents the credit spreads between A2/P2 non-financial and AA non-financial CP (in
30-day and one-day maturities) and Figure 3 Panel B presents term spreads (30-day minus one-
day yield) in each CP category used in previous figures.
A spread between A2/P2 and AA non-financial CP rates (Panel A) of the same term
should reflect differences in perceived credit risk. Prior to the BNP announcement, the spread
between the two risk classes of non-financial CP was near zero. This minimal spread reflects the
unusually low price of risk common across the financial markets prior to the financial crisis.
Following the BNP announcement, the spread jumps to approximately100 basis points in both
maturities and then drifts down to 50 basis points prior to the Bear event with the exception of a
spike in the 30-day spread to 125 basis points in December of 2007.
22
After the Bear Stearns
event both credit spreads increase, suggesting increased concerns about credit risk in the non-
financial CP market. However, the 30-day credit spread increases more than the overnight credit
spread which can be viewed as either a liquidity concern or a short horizon credit risk problem.
Finally, the spreads jumped by 300 to 400 basis points following the Lehman bankruptcy
suggesting major credit quality concerns. The credit spread between the 30-day maturities
separated from the overnight spread and remained at a high level and even drifted upward
throughout the end of Year 2008. The cause was the 30-day A2/P2 yield remaining high while

22
Griffiths and Winters (2005b) discuss the existence of year-end preferred habitats for liquidity in CP and find that
when liquidity is constrained lenders exit higher risk CP which they view as an issue of credit risk.
16

AA non-financial rates declined with the general level of short-term rates (see Figure 1). Again,
having the larger spread persist in term CP suggests investors are concerned about committing
their funds over time which can support either a desire to remain liquid or concerns about
borrower solvency beyond a few days. We attempt to disentangle these two possible
explanations later using regression analysis. That the bankruptcy of Lehman affected non-
financial CP suggests that the event was a concern to the broader markets.
Panel B of Figure 3 plots term spreads in each category. If lenders have liquidity
concerns and move toward shorter terms then the spread increases as the term rates rise while the
overnight rate falls. Given that none of the overnight rates in Figure 2 is substantially below the
target rate and some are substantially above, we have no evidence of a migration of lenders into
the overnight instruments. Thus, for these term spreads to represent liquidity concerns we need
(ceteris paribus) all the term spreads to increase. Alternatively, if only the term spreads in higher
risk CP increase then this would suggest credit concerns. First, we see that all term spreads
become more volatile after the BNP announcement, consistent with increased uncertainty.
Volatility of term spreads remains elevated throughout the rest of our data period. Second,
allowing for increased volatility, the non-financial CP term spread remains around zero
following both BNP and Bear events and then increases dramatically following the Lehman
bankruptcy. The lack of a term spread increase in non-financial CP following BNP and Bear
suggests that these events were not seen as liquidity events. However, the increase in the non-
financial term spread following Lehman accompanied by increases in all the other term spreads
suggests that Lehman may have a liquidity component. Third, the spreads widen temporarily
again prior to the end of the Year 2007 in every category excluding AA non-financial CP
consistent with the Griffiths and Winters (2005b) finding that lenders exit risky CP during year-
17

end preferred habitats for liquidity. Fourth, the Bear Stearns failure is followed by a widening in
the term spreads in the same three categories, indicating credit concerns. The Lehman
bankruptcy signaled increases in the term spreads in all four categories, of different magnitudes
and durations. The lowest credit quality category, A2/P2 non-financial CP, experienced the
largest and longest increase, followed by AA financial, asset-backed, and non-financial CP.
Liquidity in the commercial paper market may have suffered after the Reserve Primary money
market fund broke the buck due to its exposure to the Lehman-issued commercial paper, after
which many money market funds started avoiding CP. While the financial press often implied
the entire CP market suffered to the same extent, Figure 3B clearly shows that different
categories of CP exhibit different degrees of suffering. The categories with the largest and most
prolonged term spread increases are those whose credit quality raised most concerns (lower-rated
and financial paper) suggesting an increase in perceived counterparty (credit) risk. The
preponderance of evidence on the term spreads suggests a significant increase in credit risk
following each event with some short-lived liquidity effects following Lehman. We note that the
lower-rated CP, A2/P2, was not eligible for purchases by the Commercial Paper Funding Facility
(CPFF) established by the Fed in October 2008, which may have caused elevated term yields in
this category through the remainder of the calendar year.
In summary, Figures 1, 2, and 3 reflect three important points on money market rates
around significant financial news events. First, the Federal Reserve lowered its policy target rate
throughout the financial crisis, consistent with the Fed adding liquidity to the financial system.
Second, following each of the three events, one-month T-bill yields declined, consistent with
both a flight to liquidity and a flight to quality. Finally, following the three events, the higher the
credit risk of the CP class, the more the rates and the term spreads within each category increase.
18

The behavior of rates and spreads in each instance is consistent with credit quality concerns in
the CP market being primary and liquidity concerns being secondary during the financial crisis.

4.2. Amounts Outstanding Plots
The financial news on the commercial paper market during the crisis suggested that the
market had virtually stopped functioning (Milne, 2009), in other words, that the commercial
paper market was experiencing a major crisis in which no one could borrow. Even a Treasury
official suggested that the CP markets stopped functioning at some point.
23
In Figure 4, we plot
amounts outstanding in the different classes of commercial paper to separate the hype from the
reality. If the amounts outstanding dramatically decline across all classes, then there was a
broad-based liquidity crisis in commercial paper. However, if only the riskier classes show
declines in the amounts outstanding, then the issue is one of credit quality concerns.
Figure 4 provides several important insights for our analysis. First, the maximum amount
of total commercial paper outstanding (which occurs just prior to the BNP announcement) is just
short of $2.2 trillion. Second, following the BNP announcement, the total amount of CP
outstanding declined by about $400 billion, with almost the entire decline coming from asset-
backed CP, the highest risk class for which we have data available. Having the decline occur
almost entirely in asset-backed CP is consistent with this decline being the result of credit quality
issues. This result is also consistent with our earlier conclusions drawn from the rate and spread
plots. Third, the amounts outstanding in CP remained roughly the same after the Bear Stearns

23
Quoting from the Interim Assistant Secretary for Financial Stability Neel Kashkari Review of the Financial Market
Crisis and the Troubled Assets Relief Program Neal Kashkari, interim : As a result, credit markets froze. The
commercial paper market shut down, 3-month Treasuries dipped below zero, and a money market mutual fund
"broke the buck" for only the second time in history, precipitating a $200 billion net outflow of funds from that
market. The savings of millions of Americans and the ability of businesses and consumers to access affordable credit
were put at serious risk. The speech was given at the McDonough School of Business at Georgetown University on
January 13, 2009; available at http://www.treas.gov/press/releases/hp1349.htm.
19

failure, indicating that lenders did not withdraw en masse from the CP market at this time.
Combining this information with the information from our rate plots suggests that lenders did not
alter the amount of CP in which they invested following Bear, but did require higher yields for
term CP with higher credit risk. Fourth, around the Lehman bankruptcy the total amount of CP
outstanding again declined dramatically by about $400 billion. More than $200 billion of this
decline is from a decline in financial CP, with another $100 billion in asset-backed CP.
24
The
Lehman bankruptcy signaled that the US government was willing to let some (or at least one)
large financial institutions fail making lending into the financial CP market generally riskier. A
reinforcing event may have been the Reserve Primary money market fund breaking the buck
due to its exposure to the Lehman-issued CP, leading money market funds to shun CP and fly to
more secure government securities.
Consistent with our rate plots, Lehman appears to be a major credit-risk related event that
also led to liquidity problems. Finally, throughout the period, the amount outstanding in non-
financial AA-rated CP remains relatively constant with no significant declines around any of the
three events leading to the conclusion that the problems in the commercial paper market during
the financial crisis are not broad-based liquidity problems as generally believed and/or reported.
Instead, the data suggest that the problems in the commercial paper market were primarily and
directly related to credit quality, with liquidity concerns being secondary.




24
Total is not the sum of categories due to unclassified (other) paper. Also, there is no data on outstandings of
A2/P2 non-financial CP available to us. The closest approximation to A2/P2 paper is tier-2 paper. Its outstanding
amount is small relative to the entire CP market (between $50bln and $70bln throughout the crisis period) and do
not change much in response to the major events.

20

4.3. Summary of the Plots
Table 1 reports means and standard deviations of the overnight and 30-day AA non-
financial CP yields as well as the spreads of asset-backed, financial, and A2/P2 non-financial CP
rates over the AA non-financial CP rates of the same maturities. As we have seen, the AA non-
financial CP has been the most stable category throughout the crisis. We chose it as the base rate
in Table 1 because (1) it is market-determined (unlike the federal funds target rate) and (2) did
not demonstrate extreme rate swings (unlike T-bill yields).
25
Table 1 reinforces our conclusions
based on Figures 1 through 3.
Here we provide a summary of findings from Figures 1-4 and Table 1:

Category & Event 30-day
yield
Overnight
yield
Term
Spread
Outstandings Preliminary
Conclusion
A2/P2, BNP Up Up Up n.a. Credit event
A2/P2, Bear Up Up Up n.a. Credit
A2/P2, Lehman Up Up Up n.a. Credit
AA AB, BNP Up Up Unch. Down Credit
AA AB, Bear Up Unch. Up Unch. Primarily Credit
AA AB, Lehman Up Up Up Down Primarily Credit
AA FIN, BNP Unch. Unch. Up Unch. Some Credit
AA FIN, Bear Up Unch. Up Unch. Primarily Credit
AA FIN, Lehman Up Up, briefly Up Down Primarily Credit
AA NonFIN, BNP Unch. Unch. Unch. Unch. No effect
AA NonFIN, Bear Unch. Unch. Unch. Unch. No effect
AA NonFIN, Lehman Unch. Up, briefly Unch. Unch. Some Liquidity effect

The CP yields in the summary above are reported compared to the target federal funds
rate, e.g., AA asset-backed CP yields did not change much after the Bear Stearns failure but they
did increase relative to the target federal funds rate. Primarily Credit means that liquidity
concerns appear secondary.



25
Using the target federal funds rate or T-bill yield as the base results in qualitatively similar patterns.
21

4.3. Multivariate Analysis
To assess the statistical significance of rate and spread changes, we run a GARCH (1,1)
model of the following specification:
t
j
t j j
i
t i i t
Control Subperiod DV

= =
+ + + =
4
1
,
6
1
, 0
* *
(1)

t j
j
j
i
t i i t t t
Control d Subperiod c b b
,
4
1
6
1
,
2
1 1
2
1 1 0
2
* *

= =

+ + + + = (2)
where (1) and (2) are the mean and the variance equations, respectively.
The variables are defined as follows:
DV
t
a dependent variable (the spread of a 30-day or one-day CP rate in a given category
over the base rate or the term spread (30-day minus overnight rate) in a given CP category)
on day t;

Subperiod
1
a dummy variable that equals 1 on 1/2/2007 through 8/8/2007 (Year 2007
before the BNP announcement) and 0 otherwise,

Subperiod
2
a dummy variable that equals 1 on 8/9/2007 through 3/14/2008 (between the
BNP announcement and Bear Stearns failure) and 0 otherwise,

Subperiod
3
a dummy variable that equals 1 on 3/17/2008 through 9/12/2008 (between the
Bear Stearns failure and the Lehman bankruptcy) and 0 otherwise,

Subperiod
4
a dummy variable that equals 1 on 9/15/2008 through 10/31/2008 (between the
Lehman bankruptcy and the end of October 2008) and 0 otherwise,

Subperiod
5
a dummy variable that equals 1 on 11/3/2008 through 12/31/2008 (the last two
months of the Year 2008) and 0 otherwise (most of the Feds facilities or other programs
were in place by the beginning of this subperiod),

Subperiod
6
a dummy variable that equals 1 on 1/2/2009 through 7/2/2009 (Beginning of
2009 through the end of the sample period) and 0 otherwise;

t
is the error term, assumed to be normally distributed.
22

The control variables include:

QEND a dummy variable equal to 1 starting two days before the maturity of a given
commercial paper spans the end of quarter through the third-to-last business day of the
quarter, other than the end of year,

YEND a dummy variable equal to 1 starting two days before the maturity of a given
commercial paper spans the end of year through the third-to-last business day of the year,

YEND2007 defined similarly to YEND for the end of the Year 2007 only,

YEND2008 defined similarly to YEND for the end of the Year 2007 only.
26


The control variables are only applied in the regressions where 30-day CP yields or spreads
are used. The regressors are the same in the mean and variance equations.
The results are reported in Table 2. For the ease of exposition and in the interest of
brevity we do not report the estimated control variable coefficients and variance equations
coefficients (available upon request). The reference period, not covered by the subperiod
dummies, is 2001 through 2006.
Panel A of the table reports the output of the regressions with the dependent variable
being the spread between a given CP rate and the target federal funds rate. Since the target rate is
not market-determined, we have an alternative set of results in Panel B, with the AA non-
financial yields used as the base rates. Both sets of results are qualitatively similar. Panel C
reports the estimated coefficients from the regressions of the term spreads. We use the Wald test
to determine whether a given subperiods estimated coefficient is different from the previous
subperiods coefficient (or the intercept in the case of the first subperiod). The control variables
coefficients (not reported) are often significant and improve the models fit.

26
Griffiths and Winters (2005a) show that term commercial paper yields increase two days before its maturity starts
to span the end of the calendar year and return to normal on the second-to-last trading day of the year. Thus, we
control for this effect. E.g., YEND for 30-day CP in 2005 equals to one on November 29 through December 28. We
also include a control variable for possible quarter-end effects, QEND, which is turned on for ends of the first three
quarters only. Year-ends 2007 and 2008 are covered by additional dummy variables (overlapping with regular
ones) to control for the year-end effects during the crisis period.

23

The regression results confirm the findings based on the analysis of the plots.
27
Based
upon the regresssion results, the crisis in the CP market appears to be driven primarily by
counterparty risk. While the Fed focused on injecting liquidity into the financial system in the
first year of the crisis, our results suggest that liquidity concerns were important but secondary.

4.4. Robustness Checks
The Federal Reserves open market operations (OMO) during the crisis were designed to
improve liquidity in the interbank lending market and may also have influenced money markets
in general. We downloaded the data on daily temporary OMO, including the size of the
operations, maturity, and type of collateral (Treasury, agency, or mortgage-backed securities),
and used these data as control variables in Equation (1).
28
The results are not qualitatively
different from those reported in Table 2, and OMO variables are not statistically or economically
significant in the majority of categories and subperiods (not reported in the interests of brevity).
It is consistent with the crisis being driven by increased counterparty risk, which means that
simple injections of liquidity into the financial system would not be very effective as they do not
address the root cause of the crisis. We do not report these results for the sake of brevity.
Christensen, Lopez, and Rudebusch (2009) suggest that the liquidity facilities helped
lower the liquidity premium in term interbank lending rates. Taylor & Wiliams (2009) adhere to
the opposite view.

27
The results are not qualitatively sensitive to the models functional form, base rate, or control variables.
28
Permanent OMO are designed to cause permanent changes in the federal funds rate. Since the actual fed funds rate
followed the target quite closely, the effect of these operations is already reflected in the target fed funds rate.

24

The Federal Reserve has used some unorthodox tools to address the financial crisis.
29
We
study the effect of the following tools on the commercial paper market:
1) Increased (easier) access to the discount window (August 10, 2007),
2) Term Auction Facility (TAF, announced December 11, 2007, first auction held December
17, 2007),
3) Term Securities Lending Facility (TSLF, implemented March 11, 2008),
4) Asset-Backed Commercial Paper Money Market Mutual Fund Liquidity Facility (AMLF,
announced September 19, 2008),
5) Commercial Paper Funding Facility (CPFF, announced October 7, 2008, started
functioning October 27, 2008),
6) Zero target rate policy (keeping the target rate in the 0-0.25% range and focusing on
using the Feds balance sheet instead, such as purchasing agency MBS, announced
December 16, 2008).
We estimate the following GARCH model:
t
j
t j j
i
t i i t
Control Facility Subperiod DV

= =
+ + + + =
4
1
,
6
1
, 1 1 0
* *
(3),
t j
j
j
i
t i i t t t
Control f Facility d Subperiod c b b
,
4
1
6
1
, 1 1
2
1 1
2
1 1 0
2
* *

= =

+ + + + + = (4),
where (3) and (4) are the mean and the variance equations, respectively.
The dependent variable is the spread between commercial paper rate of a given category
and maturity and the target federal funds rate. The facility dummies are defined as follows:
Facility
1
= 1 from August 10, 2007 through December 16, 2007, the period when the Fed
had eased access to the discount window and worked on injecting liquidity into the
financial system in general, before the implementation of TAF;
Facility
2
= 1 from December 17, 2007 through March 10, 2008 (TAF, before TSLF);
Facility
3
= 1 from March 11, 2008 through September 18, 2008 (TSLF, before AMLF);

29
The details of new (and traditional) policy tools are available at the Feds Board of Governors site at
http://www.federalreserve.gov/monetarypolicy/default.htm
25

Facility
4
= 1 from September 19, 2008 through October 26, 2008 (AMLF, before CPFF);
Facility
5
= 1 from October 27, 2008 through December 15, 2008 (CPFF, before the zero
target-rate policy);
Facility
6
= 1 from December 16, 2008 through the end of the sample (zero-rate policy,
focusing on the Feds balance sheet instead of adjusting the target rate).
The control variables are the same as in Equations (1) and (2); they are only used in the
30-day regressions.
The coefficients of the facility variables may be interpreted as cumulative effects of the
facilities/measures, while the differences between the adjacent coefficients represent the
differential effects of each new facility. These dummies, in some cases, cover very similar
periods to the subperiod dummies; e.g., the TSLF dummy covers a period that starts one day
before the Bear Stearns failure and ends three days after the Lehman bankruptcy, practically
coinciding with the Subperiod
3
dummy. This would introduce multicollinearity. Accordingly, we
have dropped the subperiod dummies, with the exception of Subperiod
1
(Year 2007 before the
BNP announcement). If a facility worked as intended, then the estimated coefficient for a CP
yield or spread would be either not different from or smaller than the previous subperiods
coefficient.
Table 4 reports the first eight estimated coefficients from Equation (3). Injections of
liquidity (including easier access to the discount window) after the BNP announcement (and
before TAF) did not seem to help the lower credit quality CP. The TAF appears to have at least a
temporary calming effect (or coincide with the period of relative tranquility) which, however,
does not extend past the Bear Stearns failure (TSLF). The AMLF funded purchases of asset-
backed CP by banks and other financial institutions from money market funds. While this
facility, implemented a few days after the Lehman bankruptcy, may have helped the money
26

funds to meet redemptions by investors, it does not seem to have helped the issuers of asset-
backed CP to reduce the cost of funds. It appears that of the Feds six facilities/measures
examined herein, the CPFF (commercial paper funding facility that buys CP from qualified
issuers, thus guaranteeing them access to funds) had the intended affect across the CP market.
The CPFF may also have sent a signal to CP investors who may have withdrawn from the market
after the Lehman bankruptcy that the Fed believed the credit quality of many CP issuers was
high enough for it to hold their CP on its balance sheet.
Of course, we may not know what would have happened without the facilities like TAF,
TSLF and AMLF which, according to our results, do not seem to work as intended in the CP
market. It is possible, as we point out in footnote 7, that the situation would have been much
worse. However, it is clear that the initial measures undertaken by the Fed to treat the money
markets crisis the measures directed almost exclusively at increasing liquidity did not, at least
immediately, succeed in achieving the desired effect. Injections of liquidity did not necessarily
lead to increased lending/investing.

5. Conclusion
We study the U.S. commercial paper (CP) market to draw insights about the nature of the
crisis in the money markets and its possible solutions. The CP market is not homogenous in
terms of credit quality, maturities, and types of issues. We show that not all CP issuers suffered
equally during the crisis. Commercial paper of the highest perceived credit quality (AA-rated
non-financial CP) appears to have been influenced the least by the crisis, as measured by changes
in yields relative to other rates, term spreads, and the amount outstanding. Other categories
27

(A2/P2 non-financial, AA asset-backed, and AA financial) experience increased yields and/or
term spreads for at least some time during the crisis. While term spreads may be indicative of
liquidity issues, the fact that term spreads increase more for categories with lower perceived
credit quality suggests that the crisis in the money markets is related more to increases in
perceived counterparty (credit) risk, with liquidity concerns being important but secondary.
Therefore, simple injections of liquidity into the financial system would not be very effective as
they do not address the root cause of the crisis.
28

Appendix A.
Outline of the FOMC discussions, December 2006 December 2008
2006
December 12, 2006: Economic growth had slowed over the course of the year, partly reflecting
a cooling of the housing market. Going forward, the economy seemed likely to expand at a
moderate pace. The Committee judged that some inflation risks remained.
Keep the target federal funds rate (TFFR) at 5.25%
2007
March 20-21, 2007: At the January meeting, available data suggested that housing demand was
stabilizing. However, a tightening of standards for subprime borrowers in recent weeks seemed
likely to restrain home sales.
Keep TFFR at 5.25%
August 7, 2007: Conditions in markets for subprime mortgages and related instruments,
including segments of the asset-backed commercial paper market, deteriorated sharply toward
the end of the period. Credit conditions for speculative-grade corporate borrowers tightened
substantially. The markets for securities backed by subprime and other non-traditional
mortgages had become illiquid, and originations of new subprime mortgages had dropped
sharply.
Keep TFFR at 5.25%,
August 10, 2007 conference call: The Federal Reserve issued a statement announcing that it
was providing liquidity to facilitate the orderly functioning of financial markets. It would provide
reserves as necessary through open market operations.
Keep TFFR at 5.25%.
August 16, 2007 conference call: The Federal Reserve Board approved a 50 basis point
reduction in the primary credit rate to 5.75%. The Board also allowed the provision of term
financing for as long as 30 days, renewable by the borrower. In addition, the Federal Reserve
would continue to accept a broad range of collateral for discount window loans, including home
mortgages and related assets, while maintaining existing collateral margins.
Keep TFFR at 5.25%.
September 18, 2007: Short-term financial markets came under pressure over the intermeeting
period amid unease about exposures to subprime mortgages and to structured credit products.
Asset-backed commercial paper outstanding contracted substantially. Yields on Treasury bills
dropped sharply for a period; trading conditions in the market were impaired at times.
Meanwhile, banks took measures to conserve their liquidity and were cautious about
counterparties' exposures to asset-backed commercial paper.
Lower TFFR to 4.75%
October 30-31, 2007: Some improvement in the commercial paper and leveraged loan markets
over the intermeeting period; participants were somewhat less concerned that banks would not
have sufficient balance-sheet capacity to absorb large volumes of assets. Most members saw
substantial downside risks to the economic outlook.
Lower TFFR to 4.5%
December 6, 2007 (Joint session with the Board of Governors): Term Auction Facility
(TAF) established to provide term funding to eligible depository institutions beginning in mid-
December. TAF would not address all of the factors giving rise to stresses in money and credit
markets, notably the ongoing concerns about credit quality and balance sheet pressures. Most
29

participants viewed the TAF as a potentially useful tool to provide additional liquidity. Some
mentioned that a TAF could help alleviate year-end pressures in money markets.
December 11, 2007:
Heightened worries about counterparty credit risk, balance sheet constraints, and liquidity
pressures affected interbank funding markets and commercial paper markets. Spreads over risk-
free rates rose in some cases to levels higher than in August.
Lower TFFR to 4.25%.
2008
January 21, 2008 conference call: Lower TFFR to 3.5%
January 29-30, 2008: Spreads on securities in interbank funding markets over risk-free rates
narrowed somewhat following the announcement of the TAF on December 12 and eased
considerably after year-end, although they remained at somewhat elevated levels. Spreads of
rates on asset-backed commercial paper over risk-free rates also fell, on net, and the level of such
paper outstanding increased in the first two weeks of January for the first time since August.
Lower TFFR to 3%.
March 10, 2008 conference call: Meeting participants discussed the potential usefulness and
risks of instituting a Term Securities Lending Facility (TSLF), under which primary dealers
would be able to borrow Treasury securities for a term of approximately one month against any
eligible collateral. Most participants concluded that this facility was appropriate and could help
alleviate pressures in the financing markets for Treasury and some mortgage-backed securities.
March 18, 2008: Conditions in some short-term funding markets worsened. Spreads in
interbank funding markets and lower-rated commercial paper widened. Obtaining credit through
repurchase agreements backed by agency and private-label mortgage-backed securities became
more difficult amid reports of larger "haircuts" being applied by lenders and news that some
parties had missed margin calls on positions.
Lower TFFR by 75bp to 2.25%
April 29-30, 2008: Lower TFFR to 2%
June 24-25, 2008: Functioning of short-term funding markets showed some improvement;
spreads in interbank funding markets and spreads on lower-rated commercial paper generally
declined. Liquidity in the market for interbank loans at maturities beyond three months remained
thin. Some participants noted that the availability of the recently introduced Federal Reserve
liquidity facilities had probably bolstered the confidence and thus was likely partly responsible
for the improvement in market functioning. Term spreads in interbank funding markets had
declined, but remained elevated by historical standards.
Keep TFFR at 2%
August 5, 2008: On July 30, the Board of Governors and the FOMC announced enhancements
to existing liquidity facilities, including extension of the PDCF and the Term Securities Lending
Facility through January 30, 2009.
Keep TFFR at 2%
September 16, 2008 (right after Freddie, Fannie, BofA & Merill, Lehman):
To address potential liquidity pressures associated with these developments, the Federal Reserve
announced several additional initiatives, including an expansion of collateral eligible for the
Primary Dealer Credit Facility and the Term Securities Lending Facility (TSLF), increases in the
size and frequency of TSLF auctions, and a temporary relaxation of the limitations on broker-
dealers' access to funding from affiliated depository institutions. A consortium of 10 major banks
announced the creation of a liquidity pool from which participants could draw collateralized
30

loans. Despite these enhanced liquidity measures, short-term funding markets remained severely
strained, reflecting investors' heightened concerns about the financial condition of other large
financial firms, including American International Group. To further support market liquidity and
to help keep the federal funds rate near its target, the Federal Reserve conducted very large
reserve-adding open market operations the day before and the morning of the FOMC meeting.
There was agreement that the liquidity facilities established by the Federal Reserve over the past
year had been helpful in ameliorating strains in financial markets, but it was also noted that the
capital of banks and other financial institutions would need to be bolstered to strengthen the
functioning of the financial system and ease constraints on credit.
Keep TFFR at 2%
October 28-29, 2008: Due to the failures or near failures of several large financial institutions,
short-term funding markets came under significant additional pressure over the intermeeting
period, and the Federal Reserve and other central banks took actions to provide liquidity and
improve market functioning. In the overnight federal funds market, financial institutions became
more selective about the counterparties with whom they were willing to trade.

Despite the substantial provision of liquidity by the Federal Reserve and other central banks,
functioning in many credit markets remained very poor, reflecting uncertainty about liquidity
needs and future access to funding as well as concerns about the health of many financial
institutions. To strengthen confidence in U.S. financial institutions, the Treasury, the Federal
Reserve, and the Federal Deposit Insurance Corporation (FDIC) issued a joint statement on
October 14. First, the Treasury announced a voluntary capital purchase plan under which eligible
financial institutions could sell preferred shares to the U.S. government. Second, the FDIC
provided a temporary guarantee of the senior unsecured debt of all FDIC-insured institutions and
their holding companies, as well as all balances in non-interest-bearing transaction deposit
accounts. The statement included notice that nine major financial institutions had agreed to
participate in both the capital purchase program and the FDIC guarantee program. Third, the
Federal Reserve announced details of the CPFF, which was scheduled to begin on October 27.
After this joint statement and the announcements of similar programs in a number of other
countries, financial market pressures appeared to ease somewhat, though conditions remained
strained.
Lower TFFR by 50bp to 1%
December 15-16, 2008: Conditions in short-term funding markets remained strained for most of
the intermeeting period, though some signs of improvement were evident. Meeting participants
discussed how best to employ the Federal Reserve's balance sheet to promote monetary policy
goals. Specifically, participants discussed the merits of purchasing large quantities of longer-
term securities such as agency debt, agency mortgage-backed securities, and Treasury securities.
The available evidence indicated that such purchases would reduce yields on those instruments
and reduce borrowing costs for a range of private borrowers, although there was uncertainty as to
the size of such effects. Participants emphasized that the ultimate objective of special lending
facilities and asset purchases was to support overall market functioning, financial intermediation,
and economic growth.
Lower TFFR to 0-0.25%; purchase up to $100 billion in housing-related GSE debt and up to
$500 billion in agency-guaranteed MBS by the end of the second quarter of 2009

Note: Italics added by the authors
31

References
Adrian, T., and H.S. Shin, 2008, Liquidity and financial contagion, Banque de France: Financial
Stability Review Special issues on liquidity, 11.
Afonso, G., A. Kovner, and A. Schoar, 2009, Stressed, not frozen: The federal funds market in
the financial crisis, MIT and FRB of New York working paper.
Allen, F., and E. Carletti, 2008, The Role of Liquidity in Financial Crises, Working Paper
prepared for the 2008 Jackson Hole Symposium on Maintaining Stability in a Changing
Financial System.
Anderson, R., and C. Gascon, 2009, The Commercial Paper Market, the Fed, and the 2007-2009
Financial Crisis, Federal Reserve Bank of St. Louis Review 91:6, 589-612.
Armantier, O., S. Krieger, and J. McAndrews 2008, The Federal Reserves Term Auction
Facility, Current Issues in Economics and Finance, Federal Reserve Bank of New York 14:5, 1-
11.
Bate, S., S. Bushweller, and E. Rutan, 2003, The fundamentals of asset-backed commercial
paper (Moodys special report), available on Ian Giddys webpage at
http://pages.stern.nyu.edu/~igiddy/ABS/moodysabcp.pdf.
Bernanke, B., 2009, The Crisis and the Policy Response, At the Stamp Lecture, London School of
Economics, London, England, January 13, 2009.
Bruner, R.F., and S.D. Carr, 2007, The Panic of 1907: Lessons Learned from the Markets
Perfect Storm, John Wiley & Sons Inc, Hoboken, NJ.
Cecchetti, S., 2008, Monetary Policy and the Financial Crisis of 2007-2008, NBER Working
Paper, April 3, 2008.
Christensen, J., J. Lopez, and G. Rudebusch, 2009, Do central bank liquidity facilities affect
interbank lending rates? The Federal Reserve Bank of San Francisco Working Paper 2009-13.
Fama, E., (1986), Term premiums and default premiums in money markets, Jopurnal of
Financial Economics, 17. 175-196.
Greenlaw, D., J. Hatzius, A.K. Kashyap and H.S.Shin (2008), Leveraged Losses: Lessons from
the Mortgage Market Meltdown, Report prepared for the U.S. Monetary Policy Forum.
Griffiths, M. and D. Winters, 2005a, The turn of the year in money markets: Test of the risk-
shifting window dressing and preferred habitat hypotheses, Journal of Business, 1337-1364.
Griffiths, M. and D. Winters, 2005b, The year-end price of risk in a market for liquidity, Journal
of Investment Management, 99-109.
32

Kidwell, D., D. Blackwell, D. Whidbee, and R. Peterson, 2006, Financial Institutions, Markets,
and Money, Wiley, 9
th
Edition.
McAndrews, J., A. Sarkar, and Z. Wang, 2008, The effect of the term auction facility on the
London inter-bank offered rate, Staff Report No. 335, Federal Reserve Bank of New York.
Milne, A., 2009, The Fall of the House of Credit: What went wrong in banking and what can be
done to repair the damage? Cambridge University Press.
Mishkin, F. and S. Eakins, 2009, Financial Markets & Institutions, Pearson/Prentice-Hall, 6
th

Edition.
Mizen, P., 2008, The Credit Crunch of 2007-2008: A Discussion of the Background, Market
Reactions, and Policy Responses, Federal Reserve Bank of St. Louis Review, September-
October, 531-567.
Ross, S., 2002, Forensic Finance: ENRON and others, Fourth Angelo Costa Lecture, Rivista Di
Politica Economica, 9-27.
Stigum, M., 1990, The money market, Third edition. Dow Jones-Irwin: Homewood, Ill
Taylor, J., 2009, Getting off track: How government actions and interventions caused, prolonged
and worsened the financial crisis, Hoover Press.
Taylor, J. and J. Williams, 2009, A black swan in the money market, American Economic
Journal: Macroeconomics, 58-83.
Wu, T., 2008, On the Effectiveness of the Federal Reserves New Liquidity Facilities, Federal
Reserve Bank of Dallas.
33

Table 1. CP AA non-financial yields, spreads of other categories over AA non-financial CP
yields, and term spreads.
1-day
yield,%
30-day
yield,%
AA Fin
1d
AA Fin
30d
AA AB
1d
AA AB
30d A2 1d A2 30d
AA Non-
Fin
AA
FIN
AA
AB
A2/
P2
Entire Sample 2.74 2.71 0 8 14 21 31 49 -3 8 6 17
2007 pre-BNP 5.26 5.23 -2 2 3 6 8 12 -3 0 0 1
BNP-Bear 4.24 4.17 -2 11 45 65 47 69 -8 4 12 14
Bear-Lehman 2.10 2.07 -6 32 40 63 61 84 -2 35 21 21
Lehman - 10/31/08 1.50 1.76 -5 106 173 216 327 382 27 137 69 82
Nov.&Dec. '08 0.23 0.41 3 50 70 89 210 475 18 65 36 283
Yr.2009 thru 7/2 0.19 0.21 0 13 29 33 62 95 2 16 7 36
Entire Sample 1.71 1.67 5 19 28 37 52 82 16 26 26 46
2007 pre-BNP 0.03 0.02 2 2 1 2 3 3 3 3 3 4
BNP-Bear 0.78 0.86 4 12 15 41 18 27 16 18 39 30
Bear-Lehman 0.15 0.07 6 14 15 14 11 11 13 16 20 17
Lehman - 10/31/08 0.92 0.39 27 49 64 79 49 91 67 76 135 109
Nov.&Dec. '08 0.22 0.25 8 33 21 35 51 67 25 33 32 90
Yr.2009 thru 7/2 0.06 0.06 2 8 10 14 30 59 6 9 13 35
Spreads of other CP over AA Non-Fin, bp AA Non-Fin Term Spreads, bp
Standard Deviations
Means
Note: The entire sample period is 1/2/2001 through 7/2/2009. Term spreads are spreads between 30-day
and one-day CP yields in each category. Yields are in percent, spreads are in basis points.


34

Table 2. Regression output

PANEL A. Dependent variables are spreads over the target fed funds rate
AA Non-financial AA Financial
Maturity 1-day 30-day 1-day 30-day
Variable Coeff. P-val. Coeff. P-val. Coeff. P-val. Coeff. P-val.
Intercept
-0.014 0.000 0.001 0.457 -0.017 0.000 0.013 0.000
Yr. 07 before BNP
0.006
(2)
0.001 -0.028
(2)
0.000 0.000
(2)
0.977 -0.020
(2)
0.000
BNP-Bear
-0.023
(2)
0.000 -0.018 0.003 0.012 0.382 0.005
(2)
0.168
Bear-Lehman
0.015
(2)
0.001 0.028
(2)
0.000 -0.010 0.163 0.340
(2)
0.000
Lehman-10/31/08
-0.407
(2)
0.011 0.065 0.126 0.132
(2)
0.000 1.070
(2)
0.000
Nov. & Dec. 08
-0.726
(1)
0.000 -0.371
(2)
0.000 -0.156
(2)
0.000 0.304
(2)
0.000
Yr. 09 thru 7/2
-0.047
(2)
0.000 -0.040
(2)
0.000 -0.042
(2)
0.000 0.050
(2)
0.000

AA Asset-backed A2/P2 Non-financial
1-day 30-day 1-day 30-day
Coeff. P-val. Coeff. P-val. Coeff. P-val. Coeff. P-val.
Intercept
0.018 0.000 0.050 0.000 0.061 0.000 0.193 0.000
Yr. 07 before BNP
0.000
(2)
0.828 -0.026
(2)
0.000 0.000
(2)
0.821 -0.098
(2)
0.000
BNP-Bear
0.359
(2)
0.000 0.281
(2)
0.000 0.376
(2)
0.000 0.240
(2)
0.000
Bear-Lehman
0.351 0.000 0.561
(2)
0.000 0.548
(2)
0.000 0.678
(2)
0.000
Lehman-10/31/08
0.832
(1)
0.000 2.223
(2)
0.000 3.121
(2)
0.000 3.724
(2)
0.000
Nov. & Dec. 08
0.073
(2)
0.001 0.276
(2)
0.000 1.791
(2)
0.000 3.777 0.000
Yr. 09 thru 7/2
0.251
(2)
0.000 0.251 0.000 0.246
(2)
0.000 0.229
(2)
0.000

PANEL B. Dependent variables are spreads over AA non-financial CP
AA Asset-backed A2/P2 Non-financial
Maturity 1-day 30-day 1-day 30-day
Variable Coeff. P-val. Coeff. P-val. Coeff. P-val. Coeff. P-val.
Intercept
0.038 0.000 0.057 0.000 0.075 0.000 0.189 0.000
Yr. 07 before BNP
-0.011
(2)
0.000 0.007
(2)
0.000 -0.006
(2)
0.000 -0.051
(2)
0.000
BNP-Bear
0.380
(2)
0.000 0.374
(2)
0.000 0.312
(2)
0.000 0.360
(2)
0.000
Bear-Lehman
0.285
(2)
0.000 0.507
(2)
0.000 0.481
(2)
0.000 0.642
(2)
0.000
Lehman-10/31/08
1.619
(2)
0.000 2.056
(2)
0.000 3.330
(2)
0.000 3.663
(2)
0.000
Nov. & Dec. 08
0.757
(2)
0.000 0.805
(2)
0.000 2.242
(2)
0.000 4.173 0.000
Yr. 09 thru 7/2
0.220
(2)
0.000 0.283
(2)
0.000 0.291
(2)
0.000 0.296
(2)
0.000

AA Financial

1-day 30-day
Coeff. P-val. Coeff. P-val.
Intercept
0.006 0.000 0.017 0.000
Yr. 07 before BNP
-0.015
(2)
0.000 0.010
(2)
0.000
BNP-Bear
-0.021 0.000 0.043
(2)
0.000
Bear-Lehman
-0.057
(2)
0.000 0.298
(2)
0.000
Lehman-10/31/08
0.017 0.758 1.039
(2)
0.000
Nov. & Dec. 08
0.030 0.000 0.430
(2)
0.000
Yr. 09 thru 7/2
-0.005
(2)
0.000 0.102
(2)
0.000

35

Table 2 (Cont.)

PANEL C. Dependent variables are term spreads (30-day yield minus 1-day yield)

Maturity AA Non-financial AA Financial AA Asset-backed A2/P2 Non-financial
Variable Coeff. P-val. Coeff. P-val. Coeff. P-val. Coeff. P-val.
Intercept
0.014 0.000 0.023 0.000 0.030 0.000 0.087 0.000
Yr. 07 before BNP
-0.036
(2)
0.000 -0.010
(2)
0.000 -0.017
(2)
0.000 -0.052
(2)
0.000
BNP-Bear
-0.072
(2)
0.000 0.008 0.454 -0.029 0.002 -0.043 0.000
Bear-Lehman
-0.025
(2)
0.000 0.359
(2)
0.000 0.228
(2)
0.000 0.183
(2)
0.000
Lehman-10/31/08
0.231 0.128 1.521
(2)
0.000 0.868
(2)
0.000 0.849
(2)
0.001
Nov. & Dec. 08
0.144 0.051 0.907
(2)
0.000 0.233
(2)
0.000 2.059
(2)
0.000
Yr. 09 thru 7/2
0.013 0.010 0.142
(2)
0.000 0.047
(2)
0.005 0.148
(2)
0.000

Note: The entire sample period is January 2, 2001 through July 2, 2009. The table reports the first seven
coefficients from Equation (1) for each category of commercial paper in one-day and 30-day maturities. The
coefficients are in percentage points. P-values for the null of a coefficient being equal to zero are next to the
estimated coefficients. For the purposes of estimating the coefficients reported in Panel A, the target federal
funds rate has been set to 0.25% between December 16, 2008 and the end of the sample period.

(1)
and
(2)
mean that a given estimated coefficient is different from the previous subperiods coefficient (in the case
of the first subperiods coefficient from the intercept) at the 5% and 1% level, respectively.
36

Table 3. Regression output with Fed facilities as independent variables

Dependent variables are spreads of commercial paper rates over the target fed funds rate
AA Non-financial AA Financial
Maturity 1-day 30-day 1-day 30-day
Variable Coeff. P-val. Coeff. P-val. Coeff. P-val. Coeff. P-val.
Intercept
-0.014 0.000 0.001 0.472 -0.016 0.000 0.013 0.000
Yr. 07 before BNP
0.007
(2)
0.000 -0.028
(2)
0.000 0.000
(2)
0.975 -0.020
(2)
0.000
BNP to TAF
-0.002 0.859 -0.015
(2)
0.000 0.028 0.314 0.010
(2)
0.012
TAF to TSLF
-0.034
(2)
0.000 -0.102
(2)
0.000 -0.018 0.121 -0.071
(2)
0.000
TSLF to AMLF
-0.006
(2)
0.105 0.027
(2)
0.000 -0.009 0.267 0.340
(2)
0.000
AMLF to CPFF
-0.505
(2)
0.000 0.056 0.118 0.127
(2)
0.000 1.035
(2)
0.000
CPFF to zero TFFR
-0.773
(2)
0.000 -0.467
(2)
0.000 -0.353
(2)
0.000 0.328
(2)
0.000
Zero TFFR to end
-0.049
(2)
0.000 -0.041
(2)
0.000 -0.042
(2)
0.000 0.053
(2)
0.000

AA Asset-backed A2/P2 Non-financial
1-day 30-day 1-day 30-day
Coeff. P-val. Coeff. P-val. Coeff. P-val. Coeff. P-val.
Intercept
0.019 0.000 0.051 0.000 0.061 0.000 0.193 0.000
Yr. 07 before BNP
-0.001
(2)
0.693 -0.017
(2)
0.018 -0.001
(2)
0.733 -0.097
(2)
0.000
BNP to TAF
0.430
(2)
0.000 0.421
(2)
0.000 0.372
(2)
0.000 0.379
(2)
0.000
TAF to TSLF
0.433 0.000 0.201
(2)
0.000 0.406
(2)
0.000 0.238
(2)
0.000
TSLF to AMLF
0.354 0.000 0.576
(2)
0.000 0.499
(2)
0.000 0.656
(2)
0.000
AMLF to CPFF
1.090
(1)
0.001 2.588
(2)
0.000 3.379
(2)
0.000 4.223
(2)
0.000
CPFF to zero TFFR
0.045
(2)
0.066 0.332
(2)
0.000 1.832
(2)
0.000 3.801
(2)
0.000
Zero TFFR to end
0.244
(2)
0.000 0.239
(1)
0.000 0.246
(2)
0.000 0.229
(2)
0.000
Note: The entire sample period is January 2, 2001 through July 2, 2009. The table reports the first seven
coefficients from Equation (3) for each category of commercial paper in one-day and 30-day maturities. The
coefficients are in percentage points. P-values for the null of a coefficient being equal to zero are next to the
estimated coefficients. For the purposes of estimating the coefficients, the target federal funds rate has been set
to 0.25% between December 16, 2008 and the end of the sample period.

(1)
and
(2)
mean that a given estimated coefficient is different from the previous subperiods coefficient (in the case
of the first subperiods coefficient from the intercept) at the 5% and 1% level, respectively.
37

Figure 1. 30-day CP yields

Panel A. AA Financial and AA Non-financial CP, one-month T-bill, TFFR


Panel B. AA Asset-backed and A2/P2 Non-financial CP, one-month T-bill, TFFR

Figure 2. Overnight CP yields
0
1
2
3
4
5
6
1/2/07 4/2/07 7/2/07 10/2/07 1/2/08 4/2/08 7/2/08 10/2/08 1/2/09 4/2/09 7/2/09
Y
i
e
l
d
,

%
YIelds of 30-day AA Fin & AA Non-Fin CP , 1-m T-bill , Target FFR, 1/2/07-7/2/09
AA Fin AA NonFin 1-m T-bill Tgt FFR
BNP Announcement
Bear Failure
Lehman
Bankruptcy
0
1
2
3
4
5
6
1/2/07 4/2/07 7/2/07 10/2/07 1/2/08 4/2/08 7/2/08 10/2/08 1/2/09 4/2/09 7/2/09
Y
i
e
l
d
,

%
Yields of 30-day AA Asset-backed & A2/P2 Non-Fin CP, 1-m T-bill, Target FFR,
1/2/07-7/2/09
AA AB A2/P2 NonFin 1-m T-bill Tgt FFR
BNP Announcement
Bear Failure
Lehman
Bankruptcy
38


Panel A. AA Financial and AA Non-financial CP, one-month T-bill, TFFR


Panel B. AA Asset-backed and A2/P2 Non-financial CP, one-month T-bill, TFFR

0
1
2
3
4
5
6
1/2/07 4/2/07 7/2/07 10/2/07 1/2/08 4/2/08 7/2/08 10/2/08 1/2/09 4/2/09 7/2/09
Y
i
e
l
d
,

%
YIelds of one-day AA Fin & AA Non-Fin CP , 1-m T-bill , Target FFR, 1/2/07-7/2/09
AA Fin AA NonFin 1-m T-bill Tgt FFR
BNP Announcement
Bear Failure
Lehman
Bankruptcy
0
1
2
3
4
5
6
1/2/07 4/2/07 7/2/07 10/2/07 1/2/08 4/2/08 7/2/08 10/2/08 1/2/09 4/2/09 7/2/09
Y
i
e
l
d
,

%
Yields of one-day AA Asset-backed & A2/P2 Non-Fin CP, 1-m T-bill, Target FFR,
1/2/07-7/2/09
AA AB A2/P2 NonFin 1-m T-bill Tgt FFR
BNP Announcement
Bear Failure
Lehman
Bankruptcy
39

Figure 3. Credit Spreads and Term Spreads

Panel A. Credit Spreads (A2/P2 AA NonFin)

Panel B. Term spreads (30-day 1-day) in each CP Category


0
100
200
300
400
500
600
S
p
r
e
a
d
,

b
p
Spreads between A2 Non-financial and AA Non-financial CP yields, 1/2/07-7/2/09
1 day 30 days
BNP
Announcement
Bear Failure
Lehman
Bankruptcy
-100
-50
0
50
100
150
200
250
300
S
p
r
e
a
d
,

b
p
Spreads between 30-day and 1-day CP yields in each category, 1/2/07-7/2/09
AB A2/P2 Fin Non-Fin
BNP Announcement
Bear Failure
Lehman
Bankruptcy

Figure 4. CP Outstanding









0
200
400
600
800
1000
1200
1400
1600
1800
2000
2200
Total
Weekly CP
BNP
announcement
Figure 4. CP Outstanding
Fin NonFin AB
Weekly CP Outstanding, $bln., 1/3/07-7/1/09
Bear failure
Lehman
40

Lehman bankraptcy