The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Working Paper Series
Requiem for a Market Maker:
The Case of Drexel Burnham Lambert
and Below-lnvestment-Grade Bonds
Elijah Brewer III and William E. Jackson III
Working Papers Series
Issues in Financial Regulation
Research Department
Federal Reserve Bank of Chicago
December 1997 (WP-97-25)
FEDERAL RESERVE BANK
OF CHICAGO
Requiem for a Market Maker: The case of
Drexel Burnham Lambert and Below-Investment-Grade Bonds*
by
Elijah Brewer III
Research Department, 11th Floor
230 S. LaSalle Street
Federal Reserve Bank of Chicago
Chicago, Illinois 60604-1413
and
William E. Jackson III
Kenan-Flagler Business School
Campus Box 3490
McColl Building
University ofNorth Carolina
Chapel Hill, North Carolina 27599-3490
JEL Classification Numbers: Gl, G2, G21, G28, L8
[Please Do not Quote without permissionfrom Authors]
December 1997
’We thank Jennifer Conrad, Jason Greene, Gail Greenfield, Anil K. Kashyap, George G.
Kaufinan, Randall S. Kroszner, Richard McEnally, Kathryn Moran, Richard Rendleman, Clifford
Smith, Henri Servaes, Daniel Sullivan, participants at UNC-Chapel Hill’s Finance Area Summer
Seminar Series, Federal Reserve Banks of Atlanta and Chicago Seminar Series, and the Western
Economics Association 1997 meetings, for valuable comments and suggestions. The research
assistance of Jeffrey P. Ballou, Justin L. Brewer, Timothy M. Mumane, Peter Schneider,
Budhiphol Suttiratana, and Nancy E. Waddington is greatly appreciated. The views expressed
here are those of the authors and do not represent the Board of Governors of the Federal Reserve
System or the Federal Reserve Bank of Chicago.
Requiem for a Market Maker-. The Case of Drexel Burnham Lambert and BelowInvestment-Grade Bonds
ABSTRACT
In this article we add to both the financial intermediation and market microstructure literature
by examining the market reactions surrounding the withdrawal of a major financial intermediary
and market maker from a specific securities market. More specifically, we examine the exit of
Drexel Burnham Lambert (Drexel) from the junk bond market. At the time Drexel exited the
market by declaring bankruptcy, it was the dominant market maker and underwriter of junk
bonds. In this article we examine the impact of Drexel’s failure on direct and indirect holders of
junk bonds. That is, we investigate the effect of Drexel’s collapse on junk bond returns, and on
the stock returns of a group of firms that, on average, held significant amounts ofjunk bonds.
We find that the collapse of Drexel had a significant impact on junk bond prices in general,
but a greater impact on the prices of lower quality junk bonds in particular. We interpret this
result to imply that the value of the liquidity services supplied by Drexel was higher for lower-
quality junk bonds. Additionally, we find that junk bonds underwritten by Drexel, as opposed to
other investment banks, experienced a significant decline in prices over the months leading up to
Drexel’s failure announcement. We interpret this result to suggest that the monitoring services
provided by Drexel (for the bonds it underwrote) would not be easily replaced by other financial
intermediaries operating in the junk bond market.
Our results also indicate that the stock returns of life insurance companies with relatively
high junk bond exposure tended to be more negatively affected by Drexel’s financial distress
than the stock returns of life insurance companies with relatively low junk bond exposure.
Requiem for a Market Maker'. The Case of Drexel Burnham Lambert and BelowInvestment-Grade Bonds
On February 13,1990, Drexel Burnham Lambert filed for Chapter 11 bankruptcy protection.
This event led to Drexel’s exit from the below-investment-grade (junk bond) market, adversely
affecting junk bond prices. At the time of its bankruptcy filing, Drexel was the dominant force in
both the primary and secondary markets for junk bonds. The firm underwrote 287 (or 46
percent) of the 618sub-investment-grade debt issues brought to market between 1978 and 1985
and in dollar terms accounted for 57 percent of the $46 billion of new issues brought to market
(Altman and Nammacher, 1987). Drexel’s domination of the market was such that even in the
face of increasing uncertainty regarding the firm’s survival during 1989, its last year of
operations, it was able to maintain a 38.6 percent market share of new issue dollars—
approximately four times the market share of the firm’s closest competitors {Wall Street Journal,
January 2, 1990). The key to Drexel’s domination of the primary market was the extensive
network of repeat investors that it assembled and maintained. To support this investor network,
Drexel established the expertise and reputation necessary to perform credit analysis of its issues
and monitored the post-issuance activities of issuing firms. Such interactions may foster private
information flows over time, which could provide Drexel with a comparative advantage in
monitoring junk bond issuers relative to other investment houses and dispersed debtholders.
This view of Drexel’s services to the junk bond market is consistent with theoretical models of
the asset services view of intermediation, which implies that private information and associated
relationship-specific activities are intrinsic to bank lending.
In models of banking relationships, commercial banks have access to private corporate
{1}
information about their clients and monitor firms’ activities over the course of the loan.1 Banks
are both well informed compared to investors who operate with only public information and have
a comparative advantage in monitoring borrowers relative to dispersed debtholders. Thus, banks
can offer some borrowers lower financing costs relative to the public securities markets. Along
similar lines, in models of investment banking relationships, underwriters can reduce the costs of
asymmetric information by establishing a network of well-informed investors in new issues.
Beatty and Ritter (1986) and Booth and Smith (1986) show that reputational capital is an
important mechanism in the underwriting process, certifying that market prices are consistent
with insider information.
Market prices also carry an immediacy/liquidity discount (Grossman and Miller, 1988; and
Shleifer and Vishny, 1992). In an examination of the transaction costs of liquidity, Amihud and
Mendelson (1986) suggested that an instrument’s liquidity can be inferred from the size of the
bid-ask spread. Less liquid instruments will typically have wider bid-ask spreads. Amihud and
Mendelson find a positive correlation between yields and liquidity. Junk bonds typically have
wide bid-ask spreads, suggesting that their market prices will incorporate a significant liquidity
discount (Amihud and Mendelson, 1989). Thus, the yields on junk bonds will carry an illiquidity
premium to cover transaction costs. Amihud and Mendelson (1986, 1989) argue that liquidity
considerations can have a major effect on market prices of securities. Drexel’s collapse meant
that the dominant market maker was no longer supplying liquidity to the junk bond market.
Drexel’s domination of the secondary market for junk bonds, both of its own issues and
1 Monitoring by bank can be important in mitigating agency conflicts between
shareholders and creditiors. For a discussion of the role of banks in corporation finance, see
Diamond (1984), Hoshi, Kashyap, and Scharfstein (1990a and 1990b), and James (1987).
{2}
others, was widely recognized (see Table 1). The firm was generally considered the primary
source ofjunk bond prices {Wall Street Journal, March 16, 1990), and its willingness to commit
capital to carrying inventory made it an important source of liquidity in the market. At the time
that Drexel sought bankruptcy protection, the firm’s junk bond portfolio was estimated by
analysts to be worth between $1.5 billion and $2 billion (or approximately 1 percent of the entire
market). In addition, Drexel employed aggressive tactics such as the use of the firm’s capital to
buy out preferred customers at the issue price when Flight Transportation Corporation went
bankrupt less than one month after issuance. These tactics served to strengthen investor
confidence in Drexel’s willingness to stand behind both its issuers and investors. Thus, Drexel’s
exit from the junk bond market was expected to affect investors as well as issuing firms.2
According to a Government Accounting Office (GAO) study, insurance companies accounted
for over 30 percent of all junk bond investments in 1987. Drexel’s financial difficulties may
have contributed to solvency problems among those life insurance companies (LICs) holding
junk bonds. For example, two life insurance subsidiaries of First Executive Corporation were
seized by state regulators in April 1991 as a result of junk bond investment losses first disclosed
in January 1990. Executive Life of California had 62.7 percent of its general account assets
invested in junk bonds, while Executive Life of New York held 64 percent of its assets in junk
bonds. Fidelity Bankers of Virginia and First Capital of California, subsidiaries of First Capital
Corporation, were seized in May 1991 due to investment losses in their junk bond portfolios.
2 Slovin, Suuhka, and Polonchek (1993) examine share price effects of firms with lending
relationship with Continental Illinois Bank during its de facto failure and subsequent FDIC
rescue. They find that the bank’s impending failure had negative impact on client firm share
prices. They interpret this result as supporting the asset services models of intermediation that
emphasize the relationship-specific nature of bank lending.
{3}
Fidelity Bankers had 36.9 percent of its assets in junk bonds, while First Capital had about 40
percent of its general account assets invested in junk bonds. Both life insurances companies
were clients of Drexel.
If Drexel maintained a unique capacity for supporting its client investors, then its exit from
the junk bond market is expected to produce negative wealth effects for these junk bondholders.
The assumptions underlying this prediction are that: (1) the large informational production
capacity of Drexel will be lost and not easily replicated in the short run by other underwriters
and, (2) the post-failure junk bond market structure will be less liquid than the pre-failure one.
The remainder of this article is presented in five sections. The first section discusses the
relationship between an investment bank and its client firms. The second section presents the
data and methodology used to examine the impact of Drexel’s collapse on junk bond prices. The
third section provides empirical results on whether Drexel’s collapse negatively affected junk
bond prices. The fourth section investigates whether Drexel’s financial distress affected the
stock market valuation of life insurance companies. The fifth, and last, section offers a brief
summary of our findings.
1. Private information, market makers, and the value of an investment bank
An investment bank that is market maker provides a specialized service of buying and selling
securities in the secondary market. Such an investment bank will face transaction costs, because
buy and sell orders do not arrive simultaneously. An investment bank holds an inventory of
securities until it can arrange placements with investors. Transacting is costly because an
investment bank must raise capital to carry the inventory of securities and faces uncertainty about
the time required to place securities with investors (Demsetz, 1968).
{4}
The more difficult it is for investors to process and evaluate information about a security, the
longer an investment bank will expect to hold that security in its inventory, and the greater will
be the cost of transacting. Because of this investors may demand a large “liquidity premium” on
securities that are difficult to analyze or evaluate. In an examination of the transactions costs
aspect of liquidity, Amihud and Mendelson (1986) find a positive correlation between yields and
liquidity. Thus, if Drexel’s exit from the junk bond market resulted in a market structure that is
not as liquid as the pre-failure one, then junk bond prices should fall, reducing holding period
returns and creating losses to junk bond holders.
In addition to providing liquidity services to market participants, investment banks can
provide monitoring services to reduce agency costs. Jensen and Meckling (1976) developed the
theory of agency costs and suggested that independent auditing firms can reduce these types of
costs. Fama and Jensen (1985) and Smith (1986) suggested that investment banks not only help
find a market for securities, but they play a very key role as monitors. A hypothesis similar to the
monitoring hypothesis is the certification hypothesis, as presented by Baron (1982), Booth and
Smith (1986), Beatty and Ritter (1986), and others.3 Booth and Smith (1986) found evidence that
supports the certification hypothesis, which states that the certification of securities by an
investment bank adds value to issuing firms. The value arises from the ability of issuing firms’
management to communicate to investors through an investment bank that the security price is
consistent with inside information. Consequently, management is willing to pay underwriter fees
in order to communicate or certify the true value of the firm. An underlying assumption of the
3 See Kroszner and Ragan (1994), for an excellent discussion of the certification role of
investment banks before 1933.
{5}
certification hypothesis is that the investment bank has a good reputation. Furthermore, the
certification is more valuable the greater the information asymmetries associated with the firm.
The larger the information asymmetries, the greater is the potential for wealth transfers.
Therefore, the use of an investment bank’s services suggests that the costs of communicating and
the associated potential wealth transfers outweigh the costs of underwriting fees. The
relationship between the investment bank and the issuing firm involves the flow of private
information between the issuer and investment bank and entails relationship-specific
investments. Such setup costs may make it expensive for firms to immediately turn to alternative
funds in response to the failure of their investment bank.
Amihud and Mendelson (1989) suggested that the process of certification can increase a
bond’s liquidity. The liquidity of a bond is reduced when outside investors suspect that insiders
are trading on the basis of privileged information. Market makers will tend to widen the bid-ask
spread in order to protect themselves against better informed traders and to be compensated for
bearing greater liquidity risk. An investment bank’s certification of the firm’s current condition
and future prospects to outside investors reduces the risk of trading against better informed
traders. This is expected to bring about a narrower bid-ask spread and greater liquidity. Greater
liquidity of a firm’s securities may increase its value. Amihud and Mendelson (1989) have
shown that investors require a higher expected return from bonds with lower liquidity to
compensate for the bonds higher trading costs. By increasing the liquidity of the firm’s bonds (as
well as stocks), management can effectively reduce its cost of capital for any given level of
corporate risk. Thus, the impact of Drexel’s collapse could potentially affect firms’ cost of
capital if the surviving junk bond market structure is not as liquid as the pre-failure one. Drexel
{6}
likely provided monitoring and certification, as well as liquidity, services to junk bond market
participants.4 However, Amihud and Mendelson (1989) suggest that eliminating either
certification or monitoring services will have a negative impact on a market’s liquidity.
Therefore, the major focus of this article is on the liquidity services that Drexel provided.
2. Impact of Drexel’s collapse on junk bond prices
We investigate the impact of Drexel’s failure announcement on junk bond prices by
examining the daily abnormal returns associated with several junk bond portfolios. MerrillLynch maintains daily time series data on several high-yield bond indices—a “high quality”
below-investment-grade index, an ’’intermediate quality” index, a “low quality” index, and an
“average quality” index. The high quality index includes about 350 below-investment-grade
bonds that have bond ratings between BB1 and BB3; the intermediate quality index includes
about 390 instruments with bond ratings between BI and B3; the low quality index includes
about 40 below-investment-grade bonds with bond ratings between CCC1 and C3; and the
average quality index is constructed from the bonds included in the high, intermediate, and low
quality indices. We use Merrill-Lynch’s indices to calculate daily returns on the high, low, and
average quality junk bond portfolios. Daily abnormal returns to these portfolios are estimated.
Calculation of abnormal returns for these three portfolios can provide a test of the impact of
Drexel’s financial distress on junk bond prices. The estimation of daily abnormal returns is
based on the multivariate regression model that Cornett and Tehranian (1990) use to examine the
4 See Kroszner (1996), for a discussion ofjunk bonds and the substitution of traded debt
instruments for bank loans over the 1970s and 1980s. Kroszner also noted that junk bonds often
had equity-like characteristics. This suggests that Drexel, as a major underwriter, was likely
involved in the day-to-day management of the some of the junk bond issuing firms, and hence,
monitoring their activities.
{7}
effect of newly introduced banking legislation on depository institutions’ stock returns.5 This
model measures abnormal returns by the coefficients of dummy variables that are included in a
system of market-model equations.
2.1. Methodology
The junk bond price impact of Drexel’s collapse is estimated by employing a single-factor
market model. This approach involves a system of portfolio return equations for each of three
portfolios: (1) high quality junk bonds; (2) low quality junk bonds; and (3) average quality junk
bonds. Thus,
20
Ri,t = «. + Pi A,+
+
E
j = -60
(IA)
20
R2 ,
’
= a2 +
P2 ’ A’ t
+ E
5 =-60
r2
t + e2P
(IB)
20
R3,t = «3 +
P3 A< +
j
E
= -60
+
(IC)
where
Rj t = the return on a portfolio, j (=1, 2, and 3), of different quality of junk bonds on day t (T =
475 daily observations from August 5, 1988, through June 30,1990);
ctj = an intercept coefficient for portfolio j (=1,2, and 3);
5 This approach also has been used by Binder (1985a and 1985b); Thompson (1985);
Malatesta (1986); Karafiath (1988); Karafiath and Glascock (1989); and Karafiath, Mynatt, and
Smith (1991).
{8}
Pj , = risk coefficient for portfolio j (=1,2, and 3);
Rtt = the holding period return on a long-term U.S. Treasury security portfolio;
Tj s = coefficient on the binary variable Dst, for portfolio j (=1, 2, and 3) on day s;
Ds, = a binary variable that is set equal to one on day s in the forecast window and zero
otherwise; and
ej t = an error term for j (=1,2, and 3).
With this specification, the estimated parameters Tj s (j=l, 2, and 3) measure the daily
abnormal returns associated with Drexel’s bankruptcy announcement. We are testing for daily
intercept shifts in the interval day -60 to day 20. Since this interval is “dummied out,” the
observations in the day -60 to day 20 interval do not influence the estimate of the intercept. Only
those observations without dummies (day -379 to day -61 and day +21 to +95) determine the
value of the intercept.
The daily holding period returns on the U.S. Treasury portfolio were calculated as the
percentage changes in the Shearson-Lehman’s long-term Treasury security index, published in
the Wall Street Journal.6
2.2. Testable hypotheses
If Drexel maintained a unique capacity for the production of underwriting, trading, and
monitoring activities, its exit from the market would have implications for prices of below-
investment-grade bonds. The above discussion suggests that Drexel’s financial distress produced
6 Because of the potential shift in the relationship between junk bond returns and the
Treasury return, we estimated the system of equations in (1 A)-(1C) using a lower grade bond
return index as a proxy for the market index. The results were qualitatively similar to those
reported.
{9}
negative price reactions for seasoned junk bonds. Negative junk bond price reactions can be the
result of reduced liquidity in the market for junk bonds. The alternative hypothesis suggests no
reaction from Drexel’s announcement because investors may have already accounted for such an
influence in junk bond prices. We examine this by evaluating the following null hypotheses:
Hl.
The abnormal returnsfor each junk bond portfolio jointly equal zero on, or around,
Drexel’s failure announcement day s.
H2.
The abnormal returns for each junk bond portfolio individually equals zero on, or
around, Drexel's failure announcement day s.
If Hl and H2 are not rejected, then this suggests that no new information concerning the
condition of the junk bond market was conveyed to the market by Drexel’s failure
announcement. Under this scenario, the impact of Drexel’s failure on the junk bond market had
already been discounted by the time its announcement was made on February 13, 1990, and
reflected in junk bond prices. If Drexel’s earlier problems, such as Michael Milken’s dismissal,
had led investors to fully anticipate this deterioration in junk bond prices, then no significant
market response to the failure announcement would occur.
An important related hypothesis is whether the negative price response is the same across junk
bond portfolios. We predict that for any given negative impact of Drexel’s financial distress, the
impact will be larger for lower quality junk bonds than for higher quality junk bonds. The
assumption underlying this prediction is that lower rated firms are more likely to face larger
information asymmetries than other junk bond issuers. The potential for adverse selection is
widely recognized as a source of friction facing firms attempting to raise new capital (see, for
example Myers and Majluf, 1984). The relationships that Drexel maintained with its client firms
mitigated this source of friction and thereby reduced the firms’ cost of capital. Drexel’s “vote of
{10}
confidence” was likely to be more important for those firms with severe information
asymmetries. Thus, we have the following prediction:
H3.
The abnormal returns (or, economic impacts) are the ‘‘same ’’for each junk bond
portfolio on, or around, Drexel’s failure announcement day s.
In the next section, we examine these three hypotheses using daily data for three junk bond
portfolios.
2.3. Empirical results
The daily abnormal returns are generated by estimating equations (1A)-(1C) using seemingly
unrelated regression (SUR) techniques.7 Results from applying the SUR model are presented in
Table 2. Panel A of the table reports the results for high- and low-quality junk bond portfolios.
Panel B of Table 2 shows the results for the average-quality junk bond portfolio.8 Defining
February 13th as day 0, we examine the impact of Drexel’s financial distress using three
windows: Day 0, the three-day window [0,+2], and the five-day window [-2,+2]. On the event
7 The SUR methodology, attributed originally to Zellner (1962), uses joint generalized
least squares as an estimation procedure. When all explanatory variables are identical, as they
are in our system of equations, the parameter estimates and their standard errors are no different
under SUR than those that would result from ordinary least squares. However, tests of
hypotheses across equations are more efficiently performed by using SUR. In addition, when the
disturbances across equations are not independent (i.e., are contemporaneously correlated), SUR
estimation is more efficient than ordinary least squares estimation applied equation-by-equation
(see Johnston [1984]).
8 On December 21,1988, Drexel avoids criminal trial by pleading guilty to six felony
counts and agreeing to dismiss Michael Milken and pay $650 million in fines and restitution. On
March 16, 1989, the SEC increases pressure on Drexel to remove Milken from control of the
firm’s junk-bond operations. On April 13, 1989, Drexel agrees to a settlement of SEC civil
charges that gives regulators unprecedented control over the firm. We tested whether these three
separate announcements had any impact on junk bond prices. The empirical results suggest that
these additional three announcements had little, if any, statistical impact on junk bond prices.
These results are available from the authors upon request.
{11}
date, we find a decline of 1.57 to 4.24 percent in junk bond prices. From the event date to two
days after, the cumulative average decline in junk bond prices ranged between 1.19 and 3.26
percent. From two days before to two days after the event date, the cumulative average decline
in junk bond prices ranged between 1.16 and 3.41 percent. Next, we test whether these negative
junk bond price reactions are statistically significant (Hl and H2).
Test ofHl: The abnormal returns for each junk bond portfolio jointly equal zero on, or around,
Drexel’s failure announcement day s.
The values of the test statistics (F-test) under restrictions implied by the null hypothesis are:
Day 0
t10
= t2,o = t3,o = 0
132.48 (p=0.0001)
Days [0,+2]
29.04 (p=0.0001)
s=0
i-0
j=0
Days [-2,+2]
18.00 (p=0.0001).
j=-2
s=-2
s=-2
These numbers strongly suggest rejection of the null hypothesis for each of the three event
windows. Thus, it appears that there are significant abnormal returns among the three junk bond
portfolios.
Test ofH2: The abnormal returns for each junk bond portfolio individually equals zero on, or
around, Drexel'sfailure announcement day s.
{12}
Table 3 presents the test statistics for each of six event-period windows across the three junk
bond portfolios. Column 1 lists the event windows. Columns 2 through 4 present results for the
three portfolios. Overall, the results suggest that Drexel’s failure had a significant impact on
junk bond prices. The cumulative AR over the interval [-20, -3] is negative and significant.
Thus, this suggests that information (on Drexel Burnham Lambert financial deterioration and
“other” economic news) adversely affecting junk bond prices leaked out in “dribs and drabs”
over days prior to the bankruptcy announcement of February 13Although information
concerning Drexel’s financial condition had leaked out over several weeks prior to February 13,
the failure announcement still provided new information as indicated by the significant negative
reactions over the four event period windows [0], [-2,0], [-2, +2], and [0, +2].10
Test ofH3: The abnormal returns (or, economic impacts) are the "same "for each junk bond
portfolio on, or around, Drexel’s failure announcement day s.
In addition to the identification of the significance of abnormal returns at each of the three
event windows, of particular interest is a test of the hypothesis that the economic impact of
Drexel’s failure was the same for a portfolio of each quality type of junk bond at each of the three
event windows. For day 0, for example, the results indicate an abnormal return of -4.24 percent,
-2.65 percent, and -1.57 percent for the low-quality, average-quality, and high-quality junk bond
9 Cornell and Shapiro (1986) used the term “dribs and drabs” to describe the nature of
negative information flows associated with the Mexican debt crisis of 1982. They postulated that
negative information about Mexico’s financial condition was slowly released throughout 1982
and 1983, and the announcement of Mexico’s default on August 19,1982, provided no new
information. Thus, it is important to examine stock market reaction over several intervals
preceding and following the event announcement.
10 We thank Clifford Smith for suggesting the appropriate intervals for the event windows
used in this analysis.
{13}
portfolios, respectively. We test whether these apparent differences across the three portfolios
are statistically significant. Focusing attention on tests that measure abnormal returns around
Drexel’s failure announcement provides valuable information about the economic impact of the
collapse on prices of junk bonds of different quality.
A joint test of the hypothesis that the economic impact of Drexel’s collapse was the same for
each of the three junk bond portfolios gives the following F-statistics:
Day 0
^1,0
^2,0 '
61.13 (p=0.0001)
l3,0
Day [0,+2]
15.20 (p=0.0001)
EvEv
j=0
s=0
5=0
Day [-2.+2J
j=-2
s=-2
5=-2
10.59 (p=0.0001).
3,5
These results reject the null hypothesis, suggesting that the three junk bond portfolios did not
react in the same way, or to the same magnitude, to Drexel’s collapse. Table 4 provides
additional independent tests of hypothesis 3 across junk bond portfolios. As the second and third
columns of Table 4 indicate, the null hypothesis that the impact on junk bond prices of Drexel’s
{14}
failure is equal can be rejected for the average-quality and high-quality portfolios and the
average-quality and low-quality-portfolios. The results in the fourth column show that the
impact on prices of Drexel’s failure is significantly greater for low-quality junk bonds. Thus, we
conclude that the impact of Drexel’s collapse significantly affected the junk bond prices of lower
quality issues to a greater magnitude.11 This is consistent with the notion that the liquidity
services supplied by an investment banker is more valuable for firms who issues are of lower
quality.
11 To check the robustness of our results, we employed several junk bond indices obtained
from Salomon Brothers. These indices are based on monthly data. The following market model
was utilized:
Rk,t = ak + Pk,A,'
+ TkD + ek,<’
where Rk t = the return on the kth portfolio in month t, R, t = the return on Salomon’s Treasury
portfolio with an average maturity between 7 and 10 years, D is the event binary variable that is
equal to one if month is January or February 1990, otherwise zero, and xk is the coefficient on the
event binary variable. With this specification, the estimated parameter rk measures the monthly
abnormal returns associated with Drexel’s bankruptcy announcement. The results of estimating
[-1,0]
P-value
Composite
-2.6181
0.0150
High-quality
-0.8119
0.1573
Intermediate-quality
-2.6913
0.0127
Low-quality
-5.5394
0.0451
Portfolio
Composite refers to Salomon’s high yield composite index, High-quality refers to Salomon’s
high yield BB-rated index, Intermediate-quality refers to Salomon’s high yield B-rated index, and
Low-quality refers to Salomon’s high yield CCC-rated index. These results also suggest that
lower quality junk bonds responded more than higher quality junk bonds to Drexel’s financial
collapse.
{15}
3. Individual junk bond price reactions to Drexel’s financial distress
In this section, we examine the effects of Drexel’s financial distress on prices of individual
junk bonds. We have already shown that Drexel’s collapse had a more negative impact on the
market valuation of lower quality junk bonds than other quality of junk bonds. Drexel employed
its financial capital to signal to both issuers and investors the firm’s commitment to support the
junk bond market. Thus we ask the question: Is the identity of the underwriter important in
determining the impact of Drexel’s financial distress on junk bond prices? To examine this
question, we estimate the following equation, using weekly data for a sample of junk bonds over
the January 1988 to June 1990 period:
Rit = a0 + a}DREXEL + ^,R,t + £ xsDs + ej t,
(2)
j = -4
where
, is the weekly holding period return on bond i; DREXEL is a binary variable equaling
one for a bond underwritten by Drexel and zero otherwise; Rj, is the holding period return on a
long-term U.S. Treasury security portfolio; Ds is a binary variable that is set equal to one on week
s in the forecast window and zero otherwise; and ej t is an error term.
To identify publicly traded below-investment-grade bonds over the 1988-1990 period, we
utilize prospectuses of several junk bond mutual funds.12 After we obtain the names of the
issuing firms, Moodys’ reports are examined to ascertain the existence of additional junk bonds
12 We obtained data from Franklin’s Age High Income Fund; Dean Witter High Yield
Securities; First Investors High Yield Fund; Lutheran Brotherhood High Yield Fund; T. Rowe
Price High Yield Fund; Value Line Aggressive Income Trust; Keystone B-4 (the discount bond
fund); and Cigna High Yield Income Shares.
{16}
and obtain financial data on the issuing firms. Standard and Poor’s reports are examined to
identify the primary underwriter of each junk bond. Our sample consists of 50 bond issues of 36
firms. In 22 cases, Drexel was the lead underwriter, while in 28 cases other investment banks
were the lead underwriters. Weekly closing bid prices for all 50 bonds over the 132 weeks
beginning December 24, 1987, and ending June 29, 1990, are collected from Barron’s, and used
to calculate holding period returns.
Table 5 lists each junk bond, name of the issuer, rating, and whether the bond was
underwritten by Drexel. Summary statistics on selected financial characteristics of the issuing
firms are provided in Table 6. Before we test for market response differences, it is important to
establish that the risk characteristics are not significantly different between Drexel and non-
Drexel underwritten bonds. Recall that the main result from section 2 of this article is that lower
quality junk bonds on average responded significantly more negatively relative to higher quality
junk bonds to the announcement of Drexel’s failure. The bond ratings from Table 5 and the
accounting measures from Table 6, demonstrate that the riskiness of Drexel and non-Drexel
underwritten bonds are not significantly different.
The coefficient estimates on the Drexel and Ds t variables are provided in panel A of Table 7.
These results indicate that the prices of bonds underwritten by Drexel fell more over this period
than the prices of bonds underwritten by other investment banking firms.'3 To get an idea of the
13 We also estimated the following equation for a portfolio of junk bonds which Drexel
served as the underwriter and a portfolio ofjunk bonds which other investment companies served
as underwriters:
• “o
♦ M,., * £
s = -4
DREXEL e.,.
+
s - -4
{17}
approximate magnitude of this difference over the year leading up to Drexel’s failure we can
simply multiple the coefficient on the Drexel intercept in panel A of Table 7, which represents a
weekly holding period return, by the number of weeks in a year. That is, Drexel junk bonds
declined by [0.2761 percent times 52, or] about 14.4 percent more than junk bonds underwritten
by other investment banks over the year leading up to Drexel’s failure. We believe this result
provides limited support for the monitoring hypothesis.'4
The coefficients on the event variables are consistent with the notion that negative
information about Drexel’s financial condition and ability to maintain its domination of the junk
bond market was slowly released over several weeks prior to the week of the bankruptcy
announcement. From D^, (week ending January 19) to Do (week ending February 16), there
appears to have been a substantial negative impact on junk bond prices from Drexel’s financial
distress. While information concerning Drexel’s financial condition was released over several
weeks prior to the week of February 16, the coefficient on Do suggests that the bankruptcy
The results indicate no significant differences in the reaction (ts) to Drexel’s failure
announcement between junk bonds that were underwritten by Drexel and those that were
underwritten by other investment companies.
14 Anil Kashyap and Randall Kroszner suggested that we need to develop an empirical
test to distinguish between the liquidity hypothesis and the certification hypothesis. Both
Kashyap and Kroszner suggested that we should interact size of junk bond issue with the Drexel
binary variable or size with quality of the junk bond issue. The thinking here is that for some
firms certification would be more important than for other firms. For large firms or firms with
large junk bond issues, certification is less important than for smaller firms. Large firms would
be less dependent on Drexel than smaller firms. So, we need to control for size of the junk bond
issue. Kroszner suggested that we also could use board of director composition as a control
variable to capture the impact of the board on monitoring/controlling the activities of junk bond
issuers, reducing the role of the investment bank. We are in the process of collecting this
information to re-estimate our equation (2). Thus, our results are preliminary.
{18}
announcement still provided new information.
Panel B of Table 7 provides an additional test of whether Drexel’s financial collapse had more
of an impact on lower quality than higher quality junk bonds. Individual firms’ junk bonds were
ranked according to their bond rating. Junk bonds with a quality rating above CCC are classified
as “high-quality” junk bonds, and those with a quality rating CCC or below are classified as
“low-quality” junk bonds. HIGH is a binary variable equaling one for high-quality junk bonds
and zero otherwise. Equation (2) is modified to allow the rss to vary by quality ofjunk bonds.
The results in panel B of Table 7 are consistent with the portfolio results in section 2. In
particular, it appears that news of Drexel’s bankruptcy negatively affected the prices ofjunk
bonds, and that the impact varied with the quality ofjunk bonds. Lower quality junk bonds
experienced a greater price reaction to Drexel’s financial collapse than other junk bonds.
These results indicate that the Drexel Burnham Lambert financial distress had far reaching
ramifications for the junk bond market. Because insurance companies held a substantial amount
of junk bonds in their portfolios, Drexel’s collapse may have led to solvency problems among
those life insurance companies holding junk bonds. The next section of this article examines
the effects of Drexel’s failure on the stock market valuation of life insurance companies.
4. Impact of Drexel’s collapse on the market valuation of life insurance companies (LICs)
4.1. Testable hypotheses
In this section, we investigate the performance of a portfolio of life insurance companies
(LICs) stocks around events leading up to and including Drexel’s failure. The possible LIC
response to Drexel’s failure announcement can be expressed in the form of two pairs of
hypotheses. The first hypothesis is that LIC shareholders returns should reflect quickly and
{19}
without bias events that provide new information (such as Drexel’s failure announcement) about
the value of an LIC’s bond portfolio. Thus, no significant investor response should be found to
Drexel’s failure announcement, because investors had already acted on the deterioration of
Drexel’s financial condition and domination of the junk bond market in response to news and
events in the preceding months. On January 25, 1988, Drexel was informed that the SEC was
about to recommend civil charges of major securities-law violations against the firm and Michael
Milken, the head of its junk-bond operation. On March 29,1989, Michael Milken was indicted
on charges of major securities-law violations. On April 13,1989, Drexel agreed to a settlement
of SEC civil charges that gave, among other things, regulators control of the firm. Thus,
information about Drexel’s financial condition and ability to maintain its domination of the junk
bond market leaked out in “dribs and drabs” over several months. The information-leakage
hypothesis holds that, in light of preceding events and signals, Drexel’s failure announcement
provided no new information. The alternative hypothesis is simply that Drexel’s failure did
indeed provide new information and this information impacted the stock prices of LICs.
The second pair of hypotheses is concerned with the size of the stock prices response for each
LIC. Institutions with large holdings of junk bonds are expected to show more return sensitivity
than LICs with smaller holdings of junk bonds. This rational-pricing hypothesis holds that the
capital market correctly incorporated the junk bond market implications of Drexel’s failure
announcement for each LIC. The opposing hypothesis is that investors did indeed react, but were
unable to discriminate among LICs on the basis of exposure. We call this alternative the
investor-contagion hypothesis. This hypothesis holds that Drexel’s failure announcement was a
“common type of bad signal” initiating a downward revision of an LIC’s market value,
{20}
irrespective of the extent of its exposure to junk bonds. These two pairs of hypotheses can be
summarized as:
Hl.
The event parameter, Tj „ for each LIC (portfolio) jointly equals zero on Drexel’s
failure announcement:
Ti t = r2 t= T3>t...=
H2.
tn,
=0, where N is the number of LICs (portfolios).
The event parameter is equal across all LICs (portfolios) on Drexel’s failure
announcement:
Tl,t = T2,t= T3,t ""= TN,f
4.2. Model
The stock price impact on LICs of Drexel’s failure is estimated by employing a version of the
multivariate model used earlier. This model is expanded to include both a stock market factor
and an interest rate factor. An interest rate factor is appended to the traditional market model to
capture a life insurance company’s sensitivity to unanticipated changes in interest rates.15 Life
insurance companies, like commercial banks and savings and loan associations, are sensitive to
unanticipated interest rate changes, because they typically engage in interest rate intermediation
in which the interest rate sensitivity of their assets differs from that of their liabilities. Therefore,
changes in interest rates will affect the market values of the two sides of their balance sheets
differently and affect both their net worth and their stock value.
The impact of Drexel’s bankruptcy announcement of life insurance companies’ share prices is
estimated by adding a vector of (0,1) binary variables to the right-hand side of the two-factor
15 See Lynge and Zumwalt (1980), Flannery and James (1984), Scott and Peterson (1986),
Kane and Unal (1988), and Kwan (1991) for a discussion of the interest rate sensitivity of
financial intermediaries’ stock returns.
{21}
market model. Defining the event date February 13, 1990 as day zero, the model is estimated
over a 630 day interval. Each day s in the interval day -20 to day +20 is assigned a (0,1) dummy
variable Ds that is equal to one on day s only and is zero otherwise. The estimated regression
coefficient on each dummy variable Ds is a measure of the abnormal return on day s.
The model implies a system of portfolio return equations for each of two portfolios: (1) a high
junk bond exposure portfolio of LICs and (2) a low junk bond exposure portfolio of LICs. Life
insurance companies with a market value of capital to junk bond ratio less than or equal to 75
percent at the end of 1988 are classified as high junk bond institutions. Those with a market
value of capital to junk bond ratio greater than 75 percent are classified as low junk bond
institutions. Thus,
j =20
RH,t = aH +
+
+ 5E
XH,s^s
= -20
+ eH,f>
s = 20
RL,t
+ Pl,
+ E
+
XL,sDs + eL,t’
(3)
j = -20
where
Rj t = the return on a portfolio, j(=H and L), of high-exposure and low-exposure LICs on day t (T
= 630 daily observations from January 4,1988, through June 29,1990);
RMt = the return on a value-weighted portfolio;
Rj t = the return on a long-term U.S. Treasury security portfolio;
{22}
ctj = an intercept coefficient for portfolio j(=H and L);
Pj M = the stock market beta coefficient for portfolio j(=H and L);
PLj = the interest rate beta coefficient for portfolio j(=H and L);
Tj s = coefficient on the binary variable Dst, or the prediction error for portfolio j(=H and L) on
day s;
Ds = a binary variable that is set equal to one on day s in the forecast window and zero otherwise;
and
ej t = is an error term for j(-H and L).
The daily holding period returns on the U.S. Treasury portfolio were calculated as the
percentage change in the Shearson-Lehman’s long-term Treasury security index, published in the
Wall Street Journal.
4.3. Empirical results
Table 8 provides financial data on the sample of 59 life insurance companies. Low junk bond
holders are on average smaller and better capitalized than high junk bond holders. Table 9
provides seemingly unrelated regression estimates of daily abnormal returns for the high and low
junk bond holding life insurance companies for several event intervals.16 The evidence indicates
that Drexel’s failure announcement had a negative and statistically significant impact on the
stock returns of high junk bond life insurance companies. Over the five-day interval [-2,+2]
surrounding Drexel’s failure announcement the high junk bond exposure portfolio shows a
significant negative abnormal return of -13.26 percent (p-value=0.0001). If the analysis is
16 The estimation was conducted in the SUR framework to facilitate hypothesis tests and
pairwise comparisons of abnormal returns. Since, the explanatory variables are identical across
equations, system estimation influences neither the parameter estimates nor their standard errors.
{23}
limited to the three-day interval [0,+2], the high junk bond exposure portfolio shows a negative
abnormal return of-15.65 percent (p-value=0.0001). Drexel’s failure announcement had little, if
any, significant negative impact on the stock returns of the portfolio of low junk bond life
insurance companies. Over the five-day interval [-2,+2] surrounding Drexel’s failure
announcement, the low exposure portfolio showed an abnormal return of -0.87 percent (p-
value=0.3670). Over the three-day event window, the low-exposure portfolio had an abnormal
return of -0.12 percent (p-value=0.8716). Tests of the hypothesis^
h
=
l
f°r
high
and low junk bond portfolios for the two intervals [-2,+2] and [0,+2] yield:
F= 12.85
(=-2
’
z = -2
F = 33.70
t=0
’
(=0
(p = 0.0004)
'
(p = 0.0001).
’
These tests suggest that the economic impact of Drexel’s failure announcement on high junk
bond LICs was significantly different from the impact on low junk bond LICs. Specifically, the
Drexel’s failure announcement resulted in significantly larger negative abnormal returns to
stockholders of the high junk bond exposure LICs than low junk bond exposure LICs. These
results are partially consistent with the rational-pricing hypothesis as the size of investor response
to the failure announcement was related to the degree of junk bond exposure. We attempt to
shed additional light on the degree to which valuation effects realized by LICs reflected junk
bond exposure by examining individual LIC stock returns surrounding Drexel’s failure
{24}
announcement.
4.4. Individual LIC analysis
The portfolio results indicate that more exposed LICs experienced a significant decline in
market value surrounding the failure announcement. Given these findings, we examine whether
the observed decline was uniform across LICs. For each life insurance company, we estimate the
following equation:
20
i,t
“ + PA,,+
+ E
(4)
ei,c
J = -20
This approach yields abnormal returns errors for each LIC on each trading day in the event
interval. For each LIC, cumulative abnormal returns (CAR) may be obtained by summing the t
coefficients:
h
CARj{tx,t^) = 52
’
(5)
*='i
where CAR^ (t,,^) is the cumulative prediction over the interval t, to t2 for the jth portfolio.
4.4.1 LIC-specific empirical results
Table 10 sheds light on the variation in wealth effects within the sample by reporting
individual LIC abnormal returns associated with Drexel’s bankruptcy filing. Panel A of Table 10
presents the results for the high exposure LICs and panel B of Table 10 shows the results for the
low exposure firms. The results for the high exposure group show a significant negative reaction
for four (Amvestors Financial Corporation, First Capital Holding Corporation, First Executive
{25}
Corporation, and Presidential Life Corporation) of the nine LICs over both the five-day window
[-2,+2] and three-day [0,+2]. A fifth insurance company (ICH Corporation) shows a significant
reaction over the three-day window. The results for the low-exposure insurance companies in
panel B of the table show little evidence of a statistically significant reaction to Drexel’s failure
announcement. F tests of the hypothesis:
E Vi = E \2 = - = E
t = -2
t = -2
l=-2
=°
yield the following results:
High junk bond exposure'.
F = 4.3564
(p = 0.0001)
Low junk bond exposure'.
F = 0.7417
(p = 0.9117).
The results from estimating the multivariate regression model for individual LICs are
therefore consistent with those obtained for portfolio returns.17 In particular, over the five-day
17 F tests of the hypothesis
Ev.’Evz
z=0
»=0
t=0
yield the following results:
High junk bond exposure:
F = 7.3465
(p = 0.0001)
Low junk bond exposure:
F = 0.6559
(p = 0.9713).
{26}
event window, we can reject hypothesis Hl that the cumulative abnormal returns for life
insurance companies in the high junk bond exposure group jointly equal zero surrounding
Drexel’s failure announcement. The cross-sectional mean cumulative abnormal returns over the
five-day window is -75.52, suggesting that an equally weighted portfolio of exposed life
insurance companies stocks would have suffered a price decline of 75.52 percent over the five-
day period. The test statistic for the low-junk bond exposure group does not allow rejection of
hypothesis Hl.
Similarly, tests of the hypothesis
E \i= E
l = -2
t = -2
l,2
= •••
=E
t = -2
yield the following results:
High junk bond exposure:
F = 3.8224
(p = 0.0002)
Low junk bond exposure:
F = 0.7360
(p = 0.9152).
The F-statistic testing H2 over the five-day event window allows rejection of the null
hypothesis that the cumulative abnormal returns are equal across different high junk bond life
Thus, over the three-day event window [0,+2] the F-statistic allows rejection of Hl for the high
exposure group of life insurance companies. The insignificant F-statistic does not allow rejection
of Hl for the low exposure group.
{27}
insurance companies.18 Thus, this result indicates that high exposed firms experienced a
significant decline in market valuation over the five days and that the decline was not uniform
across all life insurance companies. Less exposed life insurance companies did not show any
differences in reactions to Drexel’s failure announcement. Given the above findings, we
examine whether the observed differences in response to Drexel’s failure announcement are
proportional to junk bond exposure.
4.4.2 LIC-specific differences in junk bond exposure
Selected cumulative abnormal returns will be used as the dependent variable in the following
model:
HJUNK
MV
C4^(ij,r2) =
LJUNK
MV
j
(6)
where CARj (t,,^) = abnormal return for the jth LIC over the interval t, to t2; [HJUNK/MV]j =
18 F tests of the hypothesis
1=0
Z =0
Z=0
yield the following results:
High junk bond exposure-.
F = 6.4742
(p - 0.0001)
Low junk bond exposure-.
F = 0.6668
(p = 0.9647).
Thus, over the three-day event window [0,+2] the F-statistic allows rejection of H2 for high
exposed firms, while it does not allow rejection of H2 for the low exposed group of life insurance
companies.
{28}
“higher quality” non-investment grade bonds for the jth LIC as a fraction of market value of its
equity; [LJUNK/MV]j = “lower quality” non-investment grade bonds for the jth LIC as a fraction
of market value of its equity; and
is an error term. Estimation of equation (6) allows us to test
whether the market’s ability to distinguish among LICs varies by degree of junk bond exposure
and the quality of their junk bonds.
Ordinary least squares estimation of equation (6) will provide unbiased estimates of the
parameter vector. However, the standard errors of the OLS coefficients will be biased because of
cross-sectional correlation and heteroskedasticity in the abnormal returns. Following Karafaith,
Mynatt, and Smith (1991), equation (6) is estimated using generalized least squares; details are
provided in Appendix A.
Our earlier discussion suggests that the relationship between junk bond exposure and
cumulative ARs will be negative, with LICs that have large exposure to lower quality junk bonds
showing more return sensitivity than those that have smaller exposure to such bonds. This
prediction is based on the notion that an LIC stock price should adjust rapidly to the news
contained in Drexel’s failure announcement and the adjustment should be proportion to its
holdings of lower quality junk bonds. If there is, however, a contagion impact of Drexel’s failure
announcement, we would not expect the lower quality junk bond exposure variable (or the higher
quality junk bond exposure variable) to be significantly related to cumulative ARs. A contagion
effect is by definition universal.
Results of the cross-sectional estimation of equation (6) for selected intervals are presented in
Table 11. These cross-sectional tests confirm our impression from the subgroup analysis and are
consistent with the impact of Drexel’s failure announcement on junk bond prices. In particular,
{29}
our results indicate a significant direct relationship between junk bond exposure and the market
penalty over the five-day window surrounding the bankruptcy filing. The market penalty is more
severe for LICs holding a higher the proportion of lower quality junk bonds relative to market
value of equity. In addition, the coefficient on the higher quality junk bond exposure variable
suggests that Drexel’s failure announcement inflicted less damage on LICs with higher
proportion of higher quality junk bonds relative to equity. If the analysis is limited to the (0,+2)
interval, the negative relationship between lower quality junk bond exposure and the market
penalty is slightly stronger. Both of these findings support the rational-pricing hypothesis as the
magnitude of an LIC stock market reaction varies with its exposure to various quality of junk
bonds.
V. Conclusion
In this article, we examine the implications of Drexel’s financial distress on junk bonds and
their holders. Our results indicate that all types ofjunk bonds were negatively affected by
Drexel’s financial collapse. However, prices of lower quality junk bonds fell more than those of
higher quality junk bonds. We interpret this result to mean that the liquidity services offered by
investment banks are more valuable to lower quality firms than high quality firms. Additionally,
we find that junk bonds underwritten by Drexel, relative to junk bonds underwritten by other
investment banks, experienced a significant decline in price over the year leading up to Drexel’s
failure. We interpret this finding as weak support for the monitoring hypothesis.
Lastly, we find that the market valuation of life insurance companies was negatively affected
by Drexel’s financial collapse. However, it appears that stock market investors reacted rationally
to Drexel’s failure announcement and penalized life insurance companies based on the
{30}
magnitude of their exposure to lower quality junk bonds.
{31}
Appendix A19
We estimate a first-stage regression
R.
- a
j,t
j
$j,MRM,t
s e event window
T.J,s Ds
Zj,'
where Rj t = return to firm j on date t; RM t = return on the stock market portfolio; R,, = return on
the junk bond portfolio;
tjs
= abnormal returns at date s; Ds is equal to 1 if t=s, zero otherwise;
and €j t is an error term with E(€j,) = 0, E(eJJ2)=oJ2, E(eJt ei>t)=Oij Vt, and E(ejt e,J=0 Vi j Vt* r.
For Vs, s e event window,
Rj,' = a. +
+ T
Suppose abnormal returns to firm j at date s is described by
xjs=a + bEXPj
then
,s = « + bEXPj + (a. - a,) + (Py. M - py>JR^ + (p,. 7 - P,
+
- a + bEXPj +
The cumulative abnormal returns between t, and t2 (CARj (t,,^)), can be written as:
19 The approach used in this appendix is a modified version of the techniques developed
in Karafiath, Mynatt, and Smith (1991).
{32}
C^/(„r2) =■ £ V
i = Z,
Then
h.
CAR(t^
S = tt
h
h
Y,Bexpj + Y,Vj,S~a + bEXPj
J = Z,
+ My,
J«Z,
here a^(t2-tl + l)d, &s(f2-/j + l)h, and
+ 52 (Py,/ “ Py,/)-^/,j + 52 ey,j
My = j=/,
52 (ay &y) + 52 (Py.M "
i=Z.
i=Z,
i=Z,
Manipulating the above equations yields
My - ^(“y “ ^y) + ^2^M,s (Py.A/
fy.A/) + ^y./Py.J
$y,y) + 52 ey,j»
3-ti
where T2 = (tj -1, + 1); and averages are taken over the event window [ tl5 tj. We can write Egj2
as:
E]ij = T?var(&) + T2R^ var(fij M) + T2R2
^ varfy ) + T2a2
+ 2T2 Rnf,sCOV^j^J,M^ + 2^2Rjj cov(^y»^y,y) + ^2^M,s^J,sC0V^J,M>^J,j)
=
+ R*>s<fa'X)£ + R^fe'Xfil + yoj
{33}
-1
23
+ 2R M,s
where X^i Ru RJ).
= <>fc2(± * ix'xtf *
r2/x'x£ <■ R2JX’X)^
/v\-l +
* WM * ™r., Wa
2
2
E$j = dj x correction.
Likewise, E$..$.j = a# x correction.
^correction =
Thus,
&gls x correction.
{34}
'/v\-h
References
Altman, Edward I., and Scott A. Nammacher, 1987, Investing in Junk Bonds: Inside the High
Yield Debt Market, Wiley, New York.
Amihud, Yakov, and Haim Mendelson, 1986, “Asset Pricing and the Bid-Ask Spread,” Journal
of Financial Economics 17, 223-249.
Amihud, Yakov and Haim Mendelson, 1989, “Liquidity and Cost of Capital: Implications for
Corporate Management,” Journal ofApplied Corporate Finance 2, 65-73.
Baron, David P., 1982, “A Model of the Demand of Investment Banking Advising and
Distribution Services for New Issues,” Journal of Finance 37, 955-976.
Beatty, Randolph, and Jay Ritter, 1986, “Investment Banking, Reputation, and the Underpricing
of Initial Public Offerings,” Journal ofFinancial Economics 15, 213-232.
Benveniste, Lawrence, Manoj Singh, and William J. Wilhelm, Jr., 1993, “The Failure of Drexel
Burnham Lambert: Evidence on the Implications for Commercial Banks,” Journal of
Financial Intermediation 3, 104-137.
Binder, John J., 1985a, “Measuring the Effects of Regulation with Stock Price Data,” Rand
Journal ofEconomics 16,167-183.
Binder, John J., 1985b, “On the Use of the Multivariate Regression Model in Event Studies,”
Journal ofAccounting Research 23, 370-383.
Binder, John J., 1988, “The Sherman Antitrust Act and the Railroad Cartels,” Journal ofLaw
and Economics 31,443-467.
Booth, James R., and Richard L. Smith, 1986, “Capital Raising, Underwriting and the
Certification Hypothesis,” Journal ofFinancial Economics 15,261-281.
{35}
Cornell, Bradford, and Alan C. Shapiro, 1986, “The Reaction of Bank Stock Prices to the
International Debt Crisis,” Journal of Banking and Finance 10, 55-73.
Demsetz, Harold, 1968, “The Cost of Transacting,” Quarterly Journal ofEconomics 82,44-54.
Diamond, Douglas W., 1984, “Financial Intermediation and Delegated Monitoring,” Review of
Economic Studies 51, 393-414.
Fama, Eugene F., 1976, Foundation ofFinance, Basic Books, New York.
Fama, Eugene F., and Michael C. Jensen, 1985, “Organizatonal Forms and Investment
Decisions,” Journal of Financial Economics 14, 101-119.
Flannery, Mark J., and Christopher James, 1984, “The Effect of Interest Rate Changes on the
Common Stock Returns of Financial Institutions,” Journal ofFinance 39, 1141-1153.
General Accounting Office, 1988, “Financial Markets: Issuers, Purchasers, and Purposes of High
Yield, Non-Investment Grade Bonds,” Washington: Government Printing Office.
Grossman, Sanford J., and Merton Miller, 1988, “Liquidity and Market Structure,” Journal of
Finance 43, 617-637.
Hoshi, Takeo, Anil Kashyap, and David Scharfstein, 1990a, “Bank Monitoring and Investment:
Evidence from Changing Structure of Japanese Corporate Banking Relationship,” in
Asymmetric Information, Corporate Finance and Investment, R. Glenn Hubbard (ed.),
Chicago: University of Chicago Press.
Hoshi, Takeo, Anil Kashyap, and David Scharfstein, 1990b, “The Role of Banks in Reducing
the Cost of Financial Distress in Japan, ''Journal ofFinancial Economics 27, 67-88.
James, Christopher, 1987, “Some Evidence of the Uniqueness of Bank Loans,” Journal of
Financial Economics 19,217-235.
{36}
Jensen, Michael C., and William H. Meckling, 1976, “Theory of the Firm: Managerial Behavior,
Agency Costs and Capital Structure,” Journal ofFinancial Economics 3, 305-360.
Johnston, John, 1984, Econometric Methods, 3rd edition, McGraw-Hill, New York.
Kane, Edward J., and Unal, Haluk, 1990, "Modeling Structural and Temporal Variation in the
Market's Valuation of Banking Firms," Journal ofFinance 45, 113-136.
Karafiath, Imre, 1988, “Using Dummy Variables in the Event Methodology Structure,” Financial
Review 23, 351-358.
Karafiath, Imre, and John Glascock, 1989, “Intra-Industry Effects of a Regulatory Shift: Capital
Market Evidence from Penn Square,” Financial Review 24, 123-134.
Karafiath, Imre, Ross Mynatt, and Kenneth L. Smith, 1991, “The Brazilian Default
Announcement and the Contagion Effect Hypothesis,” Journal ofBanking and Finance
15,699-716.
Kroszner, Randall S., 1995, “The Evolution of Universal Banking and Its Regulations in
Twentieth Century America,” in Universal Banking: Financial System Design
Reconsidered, Anthony Saunders and Ingo Walter, Chicago: Irwin Professional
Publishing, 70-99.
Kroszner, Randall S., and Raguram G. Rajan, 1994, “Is the Glas-Steagall Act Justified? A Study
of the U.S. Experience with Univeral Banking before 1933,” American Economic Review
84,810-832.
Kwan, Simon H, 1991, "Re-examination of Interest Rate Sensitivity of Commercial Bank Stock
Returns Using a Random Coefficient Model," Journal ofFinancial Services Research
5,61-76.
{37}
Lynge, Morgan J., and J. Kenton Zumwalt, 1980, “An Empirical Study of the Interest Rate
Sensitivity of Commercial Bank Returns: A Market Index Approach,” Journal of
Financial and Quantitative Analysis 15, 731-742.
Malatesta, Paul, 1986, “Measuring Abnormal Performance: The Event Parameter Approach
Using Joint Generalized Least Squares,” Journal ofFinancial and Quantitative Analysis
21,27-38.
Millon-Comett, Maria H., and Hassan Tehranian, 1990, “An Examination of the Impact of the
Gam-St German Depository Institutions Act of 1982 on Commercial Banks and Savings
and Loans,” Journal ofFinance 45, 95-111.
Myers, Stewart C., and Nicholas S. Majluf, 1984, “Corporate Investment and Financing
Decisions When Firms have Information That Investors Do Not Have,” Journal of
Financial Economics 13,187-222.
Schipper, Katherine, and Rex Thompson, 1983, “The Impact of Merger-Related Regulations on
the Shareholders of Acquiring Firms,” Journal ofAccounting Research 21,184-221.
Schwert, G. W., 1981, “Using Financial Data to Measure the Effects of Regulation,” Journal of
Law and Economics 24,121-145.
Scott, William L., and Richard L. Peterson, 1986, “Interest Rate Risk and Equity Values of
Hedged and Unhedged Financial Intermediaries,” Journal of Financial Research 9, 325-
329.
Shleifer, Andrei, and Robert W. Vishny, 1992, “Liquidation Values and Debt Capacity: A
Market Equilibrium Approach,” Journal of Finance 47, 1343-1366.
{38}
Slovin, Myron B., Marie E. Suuhka, and John A. Polonchek, 1993, “The Value of Bank
Durability: Borrowers as Bank Stakeholders,” Journal of Finance 48, 247-266.
Smith, Clifford W., 1986, “Investment Banking and the Capital Acquisition Process,"Journal of
Financial Economics 15,3-29.
Thompson, Rex, 1985, “Conditioning the Return Generating Process on Firm Specific Events: a
Discussion of Event Study Methods, "Journal ofFinancial and Quantitative Analysis 20,
151-168.
Zellner, Arnold, 1962, “An Efficient Method of Estimating Seemingly Unrelated Regressions
and Tests for Aggregation Bias,” Journal ofthe American Statistical Association 57, 348368.
White, James A., and Jonathan Clements, 1990, “Drexel Halts All Pricing of Junk—Mutual
Funds Hunt for Backup Data In Battered Market,” Wall Street Journal, 16 March, p. Cl.
Winkler, 1990, “Year-end Review of Markets and Finance—Junk Bonds: Poor Results in ‘89
May Show Profit Erosion for Junk Bonds,” Wall Street Journal, 2 January, p. R26.
{39}
Table 1
Top 10 junk bond underwriters
This table reports the top 10 junk bond underwriters, the dollar amount underwritten in million of
dollars, and the percentages share of the total dollar amount underwritten during 1989.
Information was obtained from Benveniste, Singh, and Wilhelm (1993).
Investment Bank
Amount
Share
Drexel
9748.6
38.6
Shearson
2361.0
9.3
Morgan Stanley
2349.7
9.3
Merrill Lynch
2252.5
8.9
Goldman Sachs
2195.4
8.9
First Boston
2014.9
8.0
Salomon Brothers
1508.3
6.0
Donaldson Lufkin
1308.9
5.2
Bear Steams
547.0
2.2
Kidder Peabody
375.0
1.5
{40}
Table 2
Estimates of daily abnormal returns (AR) for junk bond portfolios
Market model abnormal returns and t-statistics are shown for each day in the interval (day -20 to
day +20) for each portfolio of junk bonds. Panel A reports the results for the high- and lowquality junk bond portfolios; panel B shows those for the average-quality junk bond portfolio.
Panel A
Daily ARs for high- and low- quality junk bond portfolios
t-Statistic
ARLow
Quality Junk
t-Statistic
Day
Date(1990)
-20
Jan. 16
0.0509
0.442
0.0344
0.082
-19
Jan. 17
0.0101
0.088
-0.1424
-0.340
-18
Jan.18
0.0145
0.125
0.6027
1.432
-17
Jan. 19
-0.0470
-0.409
0.2503
0.596
-16
Jan. 22
-0.1249
-1.087
0.2627
0.626
-15
Jan. 23
0.0113
0.098
0.0231
0.055
-14
Jan. 24
-0.0529
-0.460
-0.2372
-0.565
-13
Jan. 25
-0.0478
-0.416
-0.0886
-0.211
-12
Jan. 26
0.0132
0.115
0.1052
0.250
-11
Jan. 29
-1.3035
-11.341®
-2.8633
-6.821®
-10
Jan. 30
-0.0491
-0.428
0.0543
0.130
-9
Jan. 31
0.0627
0.543
-0.6235
-1.480
-8
Feb. 1
-0.2450
-2.134b
0.2612
0.623
-7
Feb. 2
0.0144
0.125
0.1245
0.296
-6
Feb. 5
0.0681
0.593
0.0292
0.070
-5
Feb. 6
0.0126
0.109
0.2169
0.516
-4
Feb. 7
-0.0751
-0.654
-0.3636
-0.867
-3
Feb. 8
-0.2088
-1.816s
0.1786
0.425
-2
Feb. 9
-0.0142
-0.123
-0.3828
-0.908
-1
Feb. 12
0.0492
0.427
0.2209
0.524
ARHigh
Quality Junk
{41}
Table 2 (continued)
Panel A
Day
Daily ARs for high- and low- quality junk bond portfolios
Date(1990)
ARHigh
Quality Junk
t-Statistic
ARLow
Quality Junk
t-Statistic
-4.2400
-10.124°
0
Feb.13
-1.5725
-13.680°
+1
Feb.14
0.0665
0.579
0.7545
1.798
+2
Feb.15
0.3070
2.670°
0.2169
0.517
+3
Feb.16
0.1139
0.992
0.3301
0.787
+4
Feb. 20
0.1851
1.586
0.4661
1.093
+5
Feb. 21
-0.0487
-0.424
0.0192
0.046
+6
Feb. 22
-0.0403
-0.350
+7
Feb. 23
0.0081
0.071
0.0628
0.150
+8
Feb. 26
0.1503
1.306
-0.2778
-0.661
+9
Feb. 27
0.0716
0.623
0.0753
0.179
+10
Feb. 28
-0.2100
- 1.822s
-0.1450
-0.344
+11
Mar. 1
-0.0541
-0.470
-0.8496
-2.022b
+12
Mar. 2
-0.0541
-0.470
-0.3011
-0.717
+13
Mar. 5
0.1607
1.394
-0.1983
-0.471
+14
Mar. 6
-0.0329
-0.286
0.0568
0.135
+15
Mar. 7
0.0104
0.090
0.2734
0.652
+16
Mar. 8
-0.0458
-0.399
0.0289
0.069
+17
Mar. 9
0.0768
0.667
0.0779
0.185
+18
Mar. 12
0.0865
0.753
0.2154
0.514
+19
Mar. 13
0.0124
0.108
0.1957
0.465
+20
Mar. 14
-0.0147
-0.0369
-0.088
-0.128
{42}
-0.4291
-1.021
Table 2 (continued)
Panel B_____ Daily ARs for Average-Quality Junk Bond Portfolio
AR
t-Statistic
Day
Date(1990)
-20
Jan. 16
-0.0033
-0.024
-19
Jan. 17
-0.0572
-0.420
-18
Jan. 18
0.0153
0.112
-17
Jan. 19
-0.1232
-0.903
-16
Jan. 22
-0.0904
-0.664
-15
Jan. 23
-0.0749
-0.549
-14
Jan. 24
-0.1602
-1.175
-13
Jan. 25
-0.1849
-1.356
-12
Jan. 26
0.0264
0.193
-11
Jan. 29
-1.9131
-14.033°
-10
Jan. 30
-0.0069
-0.051
-9
Jan. 31
-0.1014
-0.741
-8
Feb. 1
-0.0852
-0.626
-7
Feb. 2
0.0294
0.215
-6
Feb. 5
0.0484
0.355
-5
Feb. 6
0.0276
0.202
-4
Feb. 7
-0.0073
-0.054
-3
Feb. 8
-0.1465
-1.074
-2
Feb. 9
-0.1417
-1.035
-1
Feb.12
0.0971
0.710
0
Feb. 13
-2.6547
-19.471°
{43}
Table 2 (continued)
Panel B_____ Daily ARs for Average-Quality Junk Bond Portfolio
Day
Date(1990)
AR
t-Statistic
+1
Feb. 14
0.2591
1.901“
+2
Feb.15
0.2501
1.834“
+3
Feb. 16
0.1127
0.828
+4
Feb. 20
0.3019
2.180b
+5
Feb. 21
0.0277
0.203
+6
Feb. 22
-0.0462
-0.339
+7
Feb. 23
0.0148
0.108
+8
Feb. 26
0.1019
0.746
+9
Feb.27
0.0473
0.347
+10
Feb.28
-0.1944
-1.421
+11
Mar. 1
-0.2778
-2.036b
+12
Mar. 2
-0.0443
-0.325
+13
Mar. 5
0.0821
0.601
+14
Mar. 6
0.0038
0.028
+15
Mar. 7
0.0475
0.349
+16
Mar. 8
-0.0324
-0.238
+17
Mar. 9
0.1495
1.096
+18
Mar. 12
0.1384
1.016
+19
Mar. 13
0.0425
0.311
+20
Mar. 14
0.0071
0.052
“a” means significant at the ten percent level, “b” means significant at the five percent level, “c”
means significant at the one percent level.
{44}
Table 3
Cumulative ARs for each of the three junk bond portfolios surrounding Drexel Burnham
Lambert’s failure announement
This table reports the cumulative ARs and p-values for six event-perod windows ([-60, -21],
[-20, -3], [0], [-2, 0], [-2, +2], and [0, +2]) for the three junk bond portfolios. The p-values are
for the null hypothesis that the cumulative ARs for each junk bond portfolio are equal to zero.
High-quality (H)
Average-quality (A)
Low-quality (L)
(-60, -21)
-0.0923
(0.9037)
-0.7377
(0.4150)
-1.8845
(0.4988)
(-20, -3)
-1.8965
(0.0002)
-2.8074
(0.0001)
-2.1757
(0.2327)
(0,0)
-1.5725
(0.0001)
-2.6547
(0.0001)
-4.2500
(0.0001)
(-2, 0)
-1.5375
(0.0001)
-2.6992
(0.0001)
-4.4119
(0.0001)
(-2, +2)
-1.1640
(0.0001)
-2.1900
(0.0001)
-3.4404
(0.0003)
(0,+2)
-1.1990
(0.0001)
-2.1455
(0.0001)
-3.2785
(0.0001)
Interval
{45}
Table 4
Test statistics for the daily abnormal and cumulative abnormal returns for the three junk
bond portfolios
This table provides a test of the hypothesis that the economic impact of Drexel’s failure is the
same for each junk bond portfolio during four event-period windows ([0], [0, +2], [-2,0], and
[-2, +2]). The tests of these pairwise comparisons utilized an F-statistic.
Interval
A vs. H
A vs. L
H vs. L
0
0
118.781 (p=0.0001)
20.649 (p=0.0001)
47.255 (p=0.0001)
0
+2
30.186 (p=0.0001)
3.460 (p=0.0631)
9.470 (p=0.0021)
-2
0
45.446 (p=0.0001)
7.902 (p=0.0050)
18.082 (p=0.0001)
-2
+2
21.181 (p=0.0001)
2.516 (p=0.1129)
6.776 (p=0.0094)
{46}
Table 5
Sampled junk bonds
This table provides a list ofjunk bonds, the name of the issuer, rating at the time of issuance, and whether
Drexel Burnham Lambert was the lead underwriter. Names of the junk bonds were obtained from the
pospectuses of the following high-yield mutual funds: Franklin’s Age High Income Fund, Dean Witter
High Yield Securities, First Investors High Yield Fund, Lutheran Brotherhood High Yield Fund, T. Rowe
Price High Yield Fund, Value Line Aggressive Income Trust, Keystone B-4 (the discount bond fund), and
Cigna High Yield Income Shares. To be included in the sample, each bond held by each High-Yield
mutual fund must be rated less than Baa by Moody’s or BBB by Standard and Poor’s and have weekly
prices listed in Barrons.
Name of Bond
Company Name
Amax 14- % 1994
Amax Inc.
Carter Hawley Hale Stores Inc.
Carter Hawley Hale Stores Inc. 12-16 2002
Carter Hawley Hale Stores Inc. 12-% 1996
Carter Hawley Hale Stores Inc.
Charter Medical Corporation
Charter Medical Corp. 15.85 2008
Coastal Corp. 11-1/b 1998
Coastal Corporation
Coastal Corporation
Coastal Corp. 11-% 2006
Coastal Corp. 8.48 1991
Coastal Corporation
Conseco Inc. 12-% 1997
Conseco Inc.
Jack Eckerd Corp. 1 l-1/e 2001
Jack Eckerd Corporation
Fairfield Communities Inc. 13-% 1992
Fairfield Communites Inc.
Golden Nugget Inc. 13-% 1995
Golden Nugget Inc.
General Homes Corporation 15-16 1995
General Homes Corporation
Griffin Resorts International, Inc. 13-7/b 1998
Resorts International Inc.
Home Shopping Network Inc. 11-% 1996
Home Shopping Network Inc.
ICN Pharmaceuticals Inc.
ICN Pharmaceuticals Inc. 12-7/b 1998
Integrated Resources 10-% 1996
Integrated Resources
Koor Industries, Ltd. 12 1996
Koor Industries, Ltd.
Leisure Technology Inc. 13-5/s 1996
Leisure Technology Inc.
MGM/UA Communications Co. 12-s/e 1993
MGM/UA Communications Co.
MGM/UA Communications Co. 13 1996
MGM/UA Communications Co.
Mark IV Industries Inc. 7 2011
Mark IV Industries Inc.
Maxxam Inc.
Maxxamlnc. 13-5/a 1992
Mesa Limited Partnership 12 1996
Mesa Limited Pamership
NVR, L.P. 10 2002
NVR, L.P.
Occidental Petroleum Corp. 8.95 1994 Occidental Petroleum Corp.
Bal
Public Serv. Co. ofN. Hampshire 17-16 2004
Public Serv. Co. ofN. Hampshire
Public Serv. Co. of N. Hampshire 14-% 1991 Public Serv. Co. ofN. Hampshire
Public Serv. Co. ofN. Hampshire 15 2003
Public Serv. Co. ofN. Hampshire
Public Serv. Co. ofN. Hampshire 14-16 2000
Public Serv. Co. ofN. Hampshire
Pacific Lumber 12 1996
Maxxam Inc.
Phillips Petroleum Co. 14-% 2000
Phillips Petroleum Co.
{47}
Rating Drexel
BB+
B2
B2
B3
Bal
BB+
Ba2
BI
B2
B3
Ba3
Caa
Caa
Ba3
B3
C
CCC
B3
B2
B2
B2
B3
B2
B3
yes
C
C
c
Caa
B+
Ba2
no
no
no
yes
yes
yes
yes
no
no
yes
yes
no
yes
no
no
yes
yes
no
yes
yes
no
yes
yes
no
no
no
no
no
yes
no
Table 5 (continued)
Name of Bond_______________________________ Company Name______________________ Rating Drexel
Resorts International Inc. 11-% 2013
Resorts International Inc. 10 1999
Resorts International Inc. 16-% 2004
Santa Fe Pacific Corporation 16 2003
Stone Container Corporation 13-% 1995
Storage Technology Corporation 13-I/i 1996
Service Merchandise Co. Inc. 11-% 1996
Twentieth Century Fox Film Corp. 10-% 1998
Twentieth Century Fox Film Corp. 13-l/4 2000
Tesoro Petroleum Corporation 12-% 2001
Trump Taj Mahal Funding, Inc. 14 1998
Texas Air Corporation 15-% 1992
Texas Air Corporation 14-% 1993
Texas Air Corporation 14-% 1990
Texas Air Corporation 14.90 1995
UNC Inc. 7-% 2006
Viacom Inc. 15-% 2006
Wickes Companies Inc. 11-% 2001
Wickes Companies Inc. 15 1995
Resorts International Inc.
Resorts International Inc.
Resorts International Inc.
Santa Fe Pacific Corporation
Stone Container Corporation
Storage Technology Corporation
Service Merchandise Co. Inc.
Twentieth Century Fox Film Corp.
Twentieth Century Fox Film Corp.
Tesoro Petroleum Corporation
Trump Taj Mahal Funding, Inc.
Continental Airlines Holdings Inc.
Continental Airlines Holdings Inc.
Continental Airlines Holdings Inc.
Continental Airlines Holdings Inc.
UNC Inc.
Viacom Inc.
Wickes Companies Inc.
Wickes Companies Inc.
{48}
Ca
Ca
Ca
Ba3
Ba3
Ba2
Ba3
B2
B2
B3
B3
Caa
Caa
Caa
Caa
B3
B2
Caa
Caa
no
no
no
no
yes
no
no
no
no
yes
no
no
yes
yes
yes
no
no
yes
yes
Table 6
Firm characteristics for the 22 bonds that were underwritten by Drexel Burnham Lambert
and 28 bonds that were underwritten by other investment bankers
Thirteen companies issued the 22 bonds that were underwritten by Drexel Burnham Lambert and nineteen companies
issued the 28 bonds that were underwritten by other investment bankers. The TIE Ratio is earnings before interest
and taxes divided by interest charges; Cash Flow is net income plus depreciation and other amortization divided by
total assets; Equity Ratio is book value of capital divided by total assets; Sales is net income divided by sales;
Volatility is the standard deviation of ROE over a five year period ending in 1989; and Return is the weekly bond
returns over the January 8, 1988-June 22, 1990 sample period. The fifty junk bonds were divided into two quality
categories. Junk bonds with a quality rating above CCC are classified as “high-quality” junk bonds, and those with a
quality rating CCC or below are classified as “low-quality” junk bonds. Bond returns are reported for both
categories ofjunk bonds by underwriter. The standard deviations are in parentheses below the mean values.
Difference provides the p-value for the null hypothesis that the mean of a variable for Drexel issues minus the mean
of the same variable for non-Drexel issues is zero assuming unequal variances for the two subsamples.
Drexel
Non-Drexel
Mean
$2.50
(2.76)
Difference
Median
$1.66
0.24
1.622
(0.79)
1.43
0.45
0.72
0.983
(1-83)
1.27
0.44
6.35
(30.6)
20.44
10.59
(26.53)
12.15
0.69
ROA
-7.77
(24.44)
-2.40
-2.76
(15.94)
0.91
0.52
ROE
17.92
(51-12)
5.03
-8.00
(91.84)
14.16
0.32
Volatility
38.46
(43.46)
25.03
99.22
(205.11)
23.72
0.22
Mean
$4.53
(5-57)
Median
$2.63
Current Ratio
1.45*
(0.31)
1.51
TIE Ratio
0.53
(1-4)
Equity Ratio
Total Assets
($ Billion)
Return
High-Quality
-0.31
(3.98)
-0.20
(2.42)
—
-0.04
(3.84)
-0.04
(2.37)
—
-0.56
-0.02
(6-14)
(5-64)
1 The number of observations used in computing this variable is 11.
2 The number of observations used in computing this variable is 15.
3 The number of observations used in computing this variable is 17.
Low-Quality
{49}
0.01
0.04
0.06
m
00
ST)
IT)
=?
z—s
*©
-L
<d
>
<D
S3
T
Q
00
s 00
co SO
Os
©
c
o
o
u
<D
CL
U
C
o
L-1
CO
©
M
+
Q
S
O
o
Er 2?
m
os co
rt so
©
©
c
o
1
»r!
+
Q
r- z-s
© 00
-* CN
1)
E
4>
>
JD
©
Q
© ?-x
2
3
>
co
c«
IH
o
x:
S ~
x
m
C
2
u
D.
«
d
LJ ©
•
M
Q'
**!✓
m
X? “■>
O'
• cn
i
i
o
(G
I
c
co
Q
£
0>
>
<D
CL ,
m
£ ®
Q‘
' <*>
00
0)
c
<D
£
QJ
CL
m
t
Q
>3
g
o -5-s
OO
2! A
C
<u
•w
Q
X
1
o
«3
Q
c
.SP
CO
CO
r-* so
23
t s_z
o
£
C3
{5 0 }
<U
DREXEL
Table 7
Av era ge abnorm al firm -spe cific junk bon d retu rns aro und Dre xel Bur nha m La mb ert ’s fail ure ann oun cem ent
+
Q
3
a
g
&
?
o
w
>> T3
3 §
U
o
© 43
S
00 §
13 -S,
CO.
<N
a
§•
£
O
43 O
co.
+
§
1
2
E
§
8 2
U *8 ©
o ■a
s 43
o
‘w
©
T3
3
£
c
E
o g>
_o C .5
"©
43 >
43 a ?
l-l © -2
O 00
S <2
u
u © <L>
& -5
u
a
©
* a
t
00
©
© ©
a 43
o ~
§>1
«
8
a>
V3
§_
s s
C3
Ii
a"
+
P
•5
S3 £
S £
2.129 1
(1.83)"
**»r
o
1
0.90 83
(0. 78 )
£
(-2 .42 )c
□ 5
a
1-2
73
eni
T
(9 80-)
9618 0-
3
+
-2.42 21
cs1
q"
-1.88 11
(-1.61 )
0.1 634
(0. 14 )
0.2 410
(0.25)
2.33 73
(2. 00 )b
1 s-z
S3
4-
£ a
’W'
<u
CM
E ®
SC
o
K
X
Q”’
>
■4-»
c
D
a
u
(X
u
>
IS
u
£
u
<S
I
Vi
vi
§
o
E
a>
>
<D
c
2
a>
CL
c
a>
+■»
-C
x> <u
O
s
S)
e
{51}
p"
(2. 06 )b
1.9514
+
5 Q
t*-~ \o
2.4 734
(2.11) b
§
2.5 757
(2. 20 )b
o
-0.4674
(-0 .40 )
•
a
-1.7284
£
(-1.82)*
cf
-3. 59 39
(-3 .78 ) c
£
73
-3.5 313
(-3.72 ) c
+
+3
-1.3962
(-1-20)
0.3633
(0.3 8)
+4
<u
Va ria ble
+2
Pan el B. Ab norm al bon d retur ns for diff ere nt qua lity of ju nk bonds
T h is p a n e l sh o w s th e a b n o rm a l b o n d re tu rn s a ro u n d D re x e l B u rn h a m L a m b e rt ’ s fa il u re a n n o u n c e m e n t fo r tw o q u a li ty c a te g o rie s o f
ju n k b o n d s. J u n k b o n d s w it h a q u ali ty ra ti n g a b o v e C C C a re c la ssif ie d a s “ h ig h -q u a li ty ” ju n k b o n d s, an d th o se w it h a q u a li ty ra ti n g
a
yzj
a
T3
o
o
e
o
2
CL
o
<D
C
-G
u.
c
60
C
G
QJ
Table 8
Selected financial charateristics of sampled life insurance companies
This table provides some characteristics, at year-end 1989, of nine life insurance companies that have market value
of capital to junk bond ratio less than or equal to 75 percent at the end of 1988 and the 50 life insurance companies
that have market value of capital to junk bond ratio more than 75 percent at the end of 1988. Financial data are from
the National Association of Insurance Commissioners. Difference is the mean of a variable for high junk bond life
insurance companies minus the mean of the same variable for low junk bond life insurance companies; the p-value is
for the null hypothesis that the difference is zero assuming unequal variances for the two subsamples. MVA is the
ratio of market value of equity to general account assets (TA). BVA is the ratio of book value of equity to TA.
Tobin’s Q is the ratio of firm market value (market value of equity plus TA minus book value of equity) to TA.
Equity volatility is obtained by using stock return daily data for the twelve month period ending with the last month
in 1989. JUNK is the ratio of junk bond investments to TA. GIC is the ratio of guaranteed investment contracts to
TA. BMIX, defined as the ratio of income from annunities to total premium income, captures the business mix of a
life insurance company. NOCST is the proportion of life insurance companies’ premium income from states without
premium tax offset policy. In 41 states, life insurance companies are allowed to credit some or all of their ex post
assessments in state-administered guaranty funds against their state premium taxes. In the other states, companies
are permitted to add a premium surcharge but may not credit assessment costs against taxes. In these cases, profits
of surviving LICs would decline if they are unable to pass all of the assessment costs onto existing policyholders.
ROA is net income divided by TA.
Low-Junk
High-Junk
Mean
Median
$5.8
$5.6
Mean
$5.3
Median
$1.4
MVA
0.061
0.051
0.275
0.207
-0.214b
BVA
0.077
0.052
0.132
0.121
-0.055*
Book-to-market
equity ratio
2.113
0.806
0.775
0.598
1.338
Tobin’s Q
0.982
1.009
1.126
1.061
-0.144b
Equity volatility
0.029
0.028
0.022
0.017
0.007*
JUNK
0.150
0.108
0.030
0.018
0.120*
GIC
0.043
0.000
0.046
0.000
BMIX
0.603
0.801
0.196
0.068
0.407*
NOCST
0.120
0.101
0.219
0.180
-0.099b
Total Assets (TA)
($ Billion)
ROA
Difference
$0.5
-0.003
0.020
0.009
0.020
0.017
0.000
“a” means significant at the five percent level, “b” means significant at the one percent level.
{52}
Table 9
Estimates of cumulative abnormal returns for high- and low-junk bond holding life
insurance companies (portfolios are value-weighted by total assets)
This table reports the market model cumulative abnormal returns and p-values for four event-period
windows ([0], [-2, 0], [-2, +2], and [0, +2]) for high- and low-junk bond exposure life insurance
companies. These cumulative abnormal returns are based on daily ARs obtained from seemingly unrelated
regression (SUR) estimates of a two-factor market model using two and one-half years of data (January 4,
1988, through June 29, 1990). The high exposure portfolio includes the nine life insuance companies that
have market value of capital to junk bond ratio less than or equal to 75 percent at the end of 1988; and the
low exposure junk bond portfolio includes the 50 life insurance companies that have market value of
capital to junk bond ratio more than 75 percent at the end of 1988. Portfolio returns are value-weighted
averages of individual stock returns. The p-values are for the null hypothesis that the cumulative
ARs for each junk bond portfolio are equal to zero.
Interval
High-junk
Low-junk
(0, 0)
-3.7124
(0.0144)
-0.2078
(0.6310)
(-2, 0)
-1.3262
(0.6135)
-0.9616
(0.2003)
(-2, +2)
-13.2649
(0.0001)
-0.8751
(0.3670)
(0,+2)
-15.6511
(0.0001)
-0.1213
(0.8716)
{53}
•o
.2
laa
© © <u
u. O
CL
0D
XJ
2 © o
© <3
CA
O
CL ©
X ba
cd C4
©
CA o
*© E c
c
3
o 3
r©
L-»
a\
os
Xi
©
Os
L-i
O
X
©
p
u<
PL
u LZ
00
«s © "E Os
3
o'
cd
©
§ .£
ca
e
3
cd
I
■I
©
J2
3
3
£
d
—
«
J5
S
3
°
>>
§
CL
!
o
©
©
§
3
CA
©
©
<4
h
IW
"S3
3
32
*3
c
(N
4-
©
□
§
C
©
©
■s
2
I
o .S
Cl
<S
$
to
o o in —. so oo oo o m
<N © cs ■'f O f";
o
odddoddoo
^;O> — -tTf— MusO
soooosO'sC'icnC'-xss©
dddddddd
oi
<
d »n ©> d 00 00 cn d
cn v©
cn d d so cn
fe
in d fe cn r-4
d
cn i
■
i
i
v-)sooo©t~-«fs©mgs
o •n —. ■'T so r~; oo © ©
ridoorinso-^
I
lilt
o
©
I
Pl
g
3
d
o
*
□
g
E
o
c
-9
3
w
>
Tl"
V“>
^2
*3
3
B-
3
©
©
o
w
(A
*4
Sh
d
6
52
3
co ©
Cw CL
o >>
xs
CA
©
3
ooosooososciTror'
OOOOf'Sf'scnOOs
ddddddodo
.2
*0
0
a,
S
©
©
©
©
c
2
3
GA
.3
r-oooor^rnOO^t’
— r~-\OCN—;0ss©0
rnOsddoo>ri>n-<0
—■ n --1 "
o
©
.2
w
§
Gw
’g
-3
3
©
X!
3
.3
A
o c 5
&.2 £■
o a s
u go
00 g —
=1 £-3
§
"§0
2
S
•3
cl
a
> .S 3
t*. o
_
3
o 2 O
©
o o
“ 2
flj o
a a > «
36
.3 -b
n. ll
cd X
U pq
3
A
04
3 —R
il
3
©
m o d d d d tt O\
d in \q >n d \O TJ- d
ooooooooo
« +
CA
©
■«
’+2 S
<00'OS’^t^t'~3sTj,K>
o\soc-~soo<NC')"qt'-;
H—— —-d—-do
2
"
© 3
>
<=>
d CL s
» , E
11
©
©
o
©
cT © o e
I CL
X s
o
CA 3
L.
CA
£
>G
eC
o
■a
■© ,©
3 fe
-o
*
oo
©
E
-o .S ■~>'a
’
3
p
i 3
© o o —
_y 4,
•w
P o
c o ■S
<u
> •E .S
u Q
VV l«
3 © ©
£
i2 §
kA
o, is
£ «
CA pm
3!
E •2
3 o Ǥ
£ CA O
fe
S 2 « o
©
© © c
u ■s 1
’ed
oooo^©v©©-’too--d
\o©sdT-indTtdoo
d‘/S'©fe''Ov'O©©
cZ
<
U
6
u
a
>—J
©
© -5
§
oof~Ttm’^so>ncs
— r^r^t^ossqr^r^
dddddddd
o
r4
•=t
£
+
OOOOOOO©
3
©
.o
2
83
"cd
J-A
o
&
o
O i_ bt
2.ff
«S
C"O
J o —
O
E
8
J
o —
5 £
.5 > -2
3
2 e
a .2
g a
GA
«
8
A
c.
S
8
8
3
A
U
3
GA
3
■O
3
8
3
3
Cfl
"3
§
P-
oo cn © oo r* d d OO
CO d d ©
© rt O\ d
Ocn©cn—«dd©*-‘
C
o
■
r^
r*> 00 v
so SO © Os
© so
SO
CM c* oo
Os c* T-.
Os © © , . so
Os m © VI m © Os
m © r* r"i Os r- r- vs Os i—i Tt cm Os 00 00 c- r- cm © Os Os m
©
© © © © © © © © © © © © © © © © © © © © O © © © © © © ©
«
o
CM
&
<
u
V
cn SO
c* © CM so m vs SO so r> 00 c* vs SO 00
Os V> Os © s© SO so
so
00 © Os VS oo vs Tf ©
Os CM V -s- 00 vs
cm c- ©
Os © vj 00 CM v> VS
c- 00
F-"« ©
c*s
© CM
CO
cm
©
cm © so cm so © © © © O cm 00 © ©
© ©
t
1
i
i
«
1
i
(
t
i
1
i
(
t
u
t?
I
o o> r) rr
ic'Ooocicioomioic, wcicfir, Ncocc — oeioo'cioc —
“ 00
" ©
— VO
” OO
“ -- ■ O' Kl
— M —
5> 00 <O O\ W rn Tf OO M rf
co o r-~ ©
©
60666060066 o © © © © © ©©©OOC5OOOOC5OO
co
0$
<
oo oo o\ m C- — cvOvvoOvd'COr’l —« cl Ov v> w> vo — <*1 00 Oc
© CC**
(N O ’t
O\ <f
-d-m-N'OodriocoN-mdricsdoddcirori-'dfNI
I
I
I
I
I
I
I
I
I
I
I
I
Tab le 10 (contin ued )
u
• .S£
$ ™
co +
fc -3
ci
Oi
<
u
SO © Os cn m c- m Os Os cm cn V © SO 00 r- © Os CM r* © V Os c* CM Os SO V) C- ©
vs © OO cm 00 00 □s oo O r- V Os ©
SO cn Os i— m SO Os 00 so so © m 00 Os © vs
©
© © © © © © © © © © ** © © © ©‘ © © © © © © © © © O © © ©
m m vs SO Os V 00 00
r- © SO cn
© CM v> V Tf so
so ^3- t*- •^r Os SO © V vs
cn Os so so
CM © CM
© r* 00
cn so SO © CM VS ’’fr © cn cn SO CO CM CM Os
CM © T 7 rn © © CM CM 00 CM
cm
© *T © © cm cm © © <*s
©
© ©
i
«
i
i
1
i
i
i
i
1
<
1
in
<n
—*
o
to
oooccnoo-;*-r- - rj
©©©©©©"©©©©©
©
lT|OOOO(\10''^OO’!tfO^OO
^t^\0K)'O(N-^O^fnrn
&
Table 10 (continued)
<
u
till
I
II
*1
— r-'00\oe©N^t-'<t^
SO(NOO»^OS'tO»N^
©©©©©©©©©©©
<
u
fO’--'rof\lNMirirMf^f<iro
© = ’T © (N —
co n © ©
ootn©oom©^-©o«nm
OsOro^oOW^O-r^OM
©©©©©’©©©©©©
3 N
co
3
u
CS OOstTt}- — oOfN © O ©
es©o^©rn^^'0 © -* <n
© •’t ci © -* oi O\ Ti-’
vi ©
{56}
cB
to
Table 11
Generalized least squares regression results explaining cumulative abnormal returns of 59 life
insurance companies surrounding Drexel Burnham Lambert’s failure announcement
(cumulative intervals)
This table provides generalized least squares estimates of the association between junk bond exposure and
cumulative abnormal returns using the following equation:
CAR^Q = a, + a2
HJUNK
MV
+ a,
LJUNK
MV
The dependent variable for each interval is the cumulative abnormal return (CAR, (t,,^)) computed from
daily abnormal returns. The daily abnormal returns are estimated from a two-factor market model using
two and one-half years of daily return data (Janaury 4, 1988, through June 29,1990). The variable
[HJUNK/MV] is the ratio of “higher quality” non-investment-grade bonds to market value of equity. The
variable [LJUNK/MV] is the ratio of “lower quality” non-investment-grade bonds to market value of
equity, gj is an error term. The market value of equity is computed at the end of 1988. HJUNK and
LJUNK are from the 1989 Statutory Reports of Condition of sampled insurance companies. The tstatistics are in parentheses and p-values are below the t-statistics.
Interval
Intercept
HJUNK/MV
LJUNK/MV
(-2, 2)
-1.0051
(-0.6547)
0.5152
5.7757
(1.8890)
0.0639
-7.5419
(-3.8239)
0.0003
0.6481
(-2, 0)
-0.5148
(-0.4591)
0.6479
-1.0287
(-0.4605)
0.6469
-0.1244
(-0.0864)
0.9314
0.0203
(0,0)
-0.1426
(-0.3433)
0.7327
0.0266
(0.0322)
0.9745
-1.3763
(-2.5777)
0.0126
0.2748
(0, 2)
-0.6329
(-0.6236)
0.5353
6.8310
(3.3791)
0.0013
-8.7937
(-6.7434)
0.0000
0.7736
{57)
Adjusted R
2
Working Paper Series
A series of research studies on regional economic issues relating to the Seventh Federal
Reserve District, and on financial and economic topics.
REGIONAL ECONOMIC ISSUES
Lean Manufacturing and the Decision to Vertically Integrate
Some Empirical Evidence From the U.S. Automobile Industry
Thomas H. Klier
WP-94-1
Domestic Consumption Patterns and the Midwest Economy
Robert Schnorbus and Paul Ballew
WP-94-4
To Trade or Not to Trade: Who Participates in RECLAIM?
Thomas H. Klier and Richard Mattoon
Restructuring and Worker Displacement in the Midwest Economy
Paul D. Ballew and Robert H. Schnorbus
Financing Elementary and Secondary Education in the 1990s:
A Review of the Issues
Richard H. Mattoon
Community Development-Fiscal Interactions: A Review of the Literature
William H. Oakland and William A. Testa
Community Development-Fiscal Interactions: Theory and Evidence
from the Chicago Area
William H. Oakland and William A. Testa
A Look at the Big Emerging Markets and U.S. Trade
Linda M. Aguilar and Mike A. Singer
Sensitivity of the Chicago Region Econometric Input-Output Model
[CREIM] to Alternative Sources of Interindustry Relationships
Phillip R. Israilevich, Geoffrey Hewings, Graham R. Schindler
and Ramamohan Mahidhara
WP-94-11
WP-94-18
WP-95-2
WP-95-6
WP-95-7
WP-95-9
WP-95-16
1
Working paper series continued
Forecasting Structural Change with a Regional Econometric Input-Output Model
Phillip R. Israilevich, Geoffrey Hewings, Michael Sonis
and Graham R. Schindler
A Mixed Bag: Assessment of Market Performance and Firm Trading
Behavior in the NOx RECLAIM Program
Michael Ari Prager, Thomas H. Klier and Richard H. Mattoon
The Determinants of State Food Manufacturing Growth: 1982-92
Mike Singer
WP-96-2
WP-96-12
WP-97-24
ISSUES IN FINANCIAL REGULATION
Origins of the Modem Exchange Clearinghouse:
A History of Early Clearing and Settlement Methods
at Futures Exchanges
James T. Moser
The Effect of Bank-Held Derivatives on Credit Accessibility
Elijah Brewer III, Bernadette A. Minton and James T. Moser
Small Business Investment Companies:
Financial Characteristics and Investments
Elijah Brewer III and Hesna Genay
WP-94-3
WP-94-5
WP-94-10
Spreads, Information Flows and Transparency Across Trading System
Paul Koftnan and James T. Moser
WP-95-1
The Cultural Affinity Hypothesis and Mortgage Lending Decisions
William C. Hunter and Mary Beth Walker
WP-95-8
Public Benefits and Public Concerns: An Economic Analysis
of Regulatory Standards for Clearing Facilities
William J. Hanley, Karen McCann and James T. Moser
Noisy Trade Disclosure and Liquidity
Subu Venkataraman
Bank Failures, Systemic Risk, and Bank Regulation
George G. Kaufinan
WP-95-12
WP-95-18
WP-96-1
2
Working paper series continued
On Biases in Tests of the Expectations Hypothesis
of the Term Structure of Interest Rates
Geert Bekaert, Robert J. Hodrick and David Marshall
The Equity Premium Puzzle and the Risk-Free Rate Puzzle at Long Horizons
Kent Daniel and David Marshall
Alligators in the Swamp: The Impact of Derivatives on the Financial
Performance of Depository Institutions
Elijah Brewer III, William E. Jackson III, and James T. Moser
WP-96-3
WP-96-4
WP-96-6
Strategic Responses to Bank Regulation: Evidence from HMDA Data
Doug Evanoffand Lewis M. Segal
WP-96-7
Deposit Insurance, Bank Capital Structures and the Demand for Liquidity
Alberto M. Ramos
WP-96-8
Interest-Rate Derivatives and Bank Lending
Elijah Brewer III, Bernadette A. Minton and James T. Moser
WP-96-13
Bank Fragility: Perception and Historical Evidence
George G. Kaufman
WP-96-18
How Should Financial Institutions and Markets Be Structured?
Analysis and Options for Financial System Design
George G. Kaufman and Randall S. Kroszner
WP-96-20
Changes in Trading Activity Following Stock Splits and Their Impact
on Volatility and the Adverse Information Component of the Bid-Ask Spread
A.S. Desai, M. Nimalendran and S. Venkataraman
WP-96-21
Financial Distress and the Role of Capital Contributions by the Owner Manager
S. Venkataraman
The Security Issue Decision: Evidence from Small Business
Investment Companies
Elijah Brewer III, Hesna Genay, William E. Jackson, and Paula R. Worthington
WP-96-22
WP-96-27
3
Working paper series continued
FDICIA After Five Years: A Review and Evaluation
George J. Benston and George G. Kaufinan
WP-97-1
“Peso Problem” Explanations for Term Structure Anomalies
Geert Bekaert, Robert J. Hodrick, and David A. Marshall
WP-97-7
The Role of Credit Market Competition on Lending Strategies
and on Capital Accumulation
Nicola Cetorelli
WP-97-14
The Role of the Financial Services Industry in the Local Economy
Douglas D. Evanoff, Philip R. Israilevich and Graham R. Schindler
WP-97-21
The Trojan Horse or the Golden Fleece? Small Business Investment
Companies and Government Guarantees
Elijah Brewer III, Hesna Genay, William E. Jackson III and Paula R. Worthington
WP-97-22
Requiem for a Market Maker: The Case of Drexel Burnham Lambert
and Below-Investment-Grade Bonds
Elijah Brewer III and William E. Jackson III
WP-97-25
MACROECONOMIC ISSUES
The Effects of Monetary Policy Shocks: Evidence from the Flow of Funds
Lawrence J. Christiano, Martin Eichenbaum and Charles Evans
WP-94-2
Algorithms for Solving Dynamic Models with Occasionally Binding Constraints
Lawrence J. Christiano and Jonas D.M. Fisher
WP-94-6
Identification and the Effects of Monetary Policy Shocks
Lawrence J. Christiano, Martin Eichenbaum and Charles L Evans
WP-94-7
Small Sample Bias in GMM Estimation of Covariance Structures
Joseph G. Altonji and Lewis M. Segal
WP-94-8
Interpreting the Procyclical Productivity of Manufacturing Sectors:
External Effects of Labor Hoarding?
Argia M. Sbordone
WP-94-9
4
Working paper series continued
Small Sample Properties of Generalized Method of Moments Based Wald Tests
Craig Burnside and Martin Eichenbaum
WP-94-12
Evidence on Structural Instability in Macroeconomic Time Series Relations
James H. Stock and Mark W. Watson
WP-94-13
The Post-War U.S. Phillips Curve: A Revisionist Econometric History
Robert G. King and Mark W. Watson
WP-94-14
The Post-War U.S. Phillips Curve: A Comment
Charles L. Evans
WP-94-1S
Identification of Inflation-Unemployment Tradeoffs in the 1970s
Bennett T. McCallum
WP-94-16
The Post-War U.S. Phillips Curve: A Revisionist Econometric History
Response to Evans and McCallum
Robert G. King and Mark W. Watson
Estimating Deterministic Trends in the Presence of Serially Correlated Errors
Eugene Canjels and Mark W. Watson
Solving Nonlinear Rational Expectations Models by Parameterized
Expectations: Convergence to Stationary Solutions
Albert Marcet and David A. Marshall
WP-94-17
WP-94-19
WP-94-20
The Effect of Costly Consumption Adjustment on Asset Price Volatility
David A. Marshall and Nayan G. Parekh
WP-94-21
The Implications of First-Order Risk Aversion for Asset Market Risk Premiums
Geert Bekaert, Robert J. Hodrick and David A. Marshall
WP-94-22
Asset Return Volatility with Extremely Small Costs
of Consumption Adjustment
David A. Marshall
Indicator Properties of the Paper-Bill Spread: Lessons From Recent Experience
Benjamin M. Friedman and Kenneth N. Kuttner
WP-94-23
WP-94-24
5
Working paper series continued
Overtime, Effort and the Propagation of Business Cycle Shocks
George J. Hall
WP-94-25
Monetary policies in the early 1990s—reflections of the early 1930s
Robert D. Laurent
WP-94-26
The Returns from Classroom Training for Displaced Workers
Louis S. Jacobson, Robert J. LaLonde and Daniel G. Sullivan
WP-94-27
Is the Banking and Payments System Fragile?
George J. Benston and George G. Kaufman
WP-94-28
Small Sample Properties of GMM for Business Cycle Analysis
Lawrence J. Christiano and Wouter den Haan
WP-95-3
The Fed Funds Futures Rate as a Predictor of Federal Reserve Policy
Joel T. Krueger and Kenneth N. Kuttner
WP-95-4
Capital Utilization and Returns to Scale
Craig Burnside, Martin Eichenbaum and Sergio Rebelo
WP-95-5
Money, Prices, Interest Rates and the Business Cycle
Robert G. King and Mark W. Watson
WP-95-10
Asset Pricing Lessons for Modeling Business Cycles
Michele Boldrin, Lawrence J. Christiano and Jonas D.M. Fisher
WP-95-11
Inside Money, Outside Money and Short Term Interest Rates
V.V. Chari, Lawrence J. Christiano and Martin Eichenbaum
WP-95-13
Tobin’s q and Asset Returns: Implications for Business Cycle Analysis
Lawrence J. Christiano and Jonas D.M. Fisher
WP-95-14
Sectoral Solow Residuals
Craig Burnside, Martin Eichenbaum and Sergio Rebelo
WP-95-15
The European Unemployment Dilemma
Lars Ljungqvist and Thomas J. Sargent
WP-95-17
6
Working paper series continued
Flexible Employment: Composition and Trends
Lewis M. Segal
WP-95-19
Expectation Traps and Discretion
V. V. Chari, Lawrence J. Christiano and Martin Eichenbaum
WP-96-5
North-South Business Cycles
Michael Kouparitsas
WP-96-9
North-South Financial Integration and Business Cycles
Michael Kouparitsas
WP-96-10
Macroeconomic Effects of Employment Reallocation
Ken Kuttner and Jeffrey Campbell
WP-96-11
A Price Target for U.S. Monetary Policy? Lessons from the Experience
with Money Growth Targets?
Benjamin M. Friedman and Kenneth N. Kuttner
WP-96-14
Catching up with the Keynesians
Lars Ljungqvist and Harald Uhlig
WP-96-15
Chaos, Sunspots, and Automatic Stabilizers
Lawrence J. Christiano and Sharon G. Harrison
WP-96-16
Aggregate Employment Fluctuations with Microeconomic Asymmetries
Jeffrey R. Campbell and Jonas D. M. Fisher
WP-96-17
Using Sibling Data to Estimate the Impact of Neighborhoods
on Children’s Educational Outcomes
Daniel Aaronson
WP-96-19
Credit Market Imperfections and the Heterogeneous Response of Firms
to Monetary Shocks
Jonas D. M. Fisher
WP-96-23
(S, j) Inventory Policies in General Equilibrium
Jonas D.M. Fisher and Andreas Homstein
WP-96-24
7
Working paper series continued
Non-Convex Costs and Capital Utilization: A Study of Production and
Inventories at Automobile Assembly Plants
George J. Hall
WP-96-25
The Growth of Temporary Services Work
Lewis Segal and Daniel Sullivan
WP-96-26
Sticky Price and Limited Participation Models of Money: A Comparison
Lawrence J. Christiano, Martin Eichenbaum and Charles L. Evans
WP-96-28
The Effect of State Fiscal Reform on Population Heterogeneity
Daniel Aaronson
WP-96-29
Money, Sticky Wages, and the Great Depression
Michael D. Bordo, Christopher J. Erceg and Charles L. Evans
WP-97-2
Price Pass-Through and Minimum Wages
Daniel Aaronson
WP-97-3
Habit Persistence and Asset Returns in an Exchange Economy
Michele Boldrin, Lawrence J. Christiano and Jonas D.M. Fisher
WP-97-4
North-South Terms of Trade: An Empirical Investigation
Michael A. Kouparitsas
WP-97-5
Interactions Between the Seasonal and Business Cycles
in Production and Inventories
Steven G. Cecchetti, Anil K. Kashyap and David W. Wilcox
WP-97-6
The Big Problem of Small Change
Thomas J. Sargent, Francois R. Velde
WP-97-8
Monetary Policy and the Term Structure of Nominal Interest Rates:
Evidence and Theory
Charles L. Evans and David A. Marshall
WP-97-10
Employer Learning and Statistical Discrimination
Joseph G. Altonji and Charles R. Pierret
WP-97-11
8
Working paper series continued
A Model of Commodity Money, With Applications to Gresham’s Law
and the Debasement Puzzle
Francois R. Velde, Warren E. Weber and Randall Wright
WP-97-12
The Evolution of Small Change
Thomas J. Sargent and Francois R. Velde
WP-97-13
Algorithms for Solving Dynamic Models with Occasionally
Binding Constraints
Lawrence J. Christiano and Jonas Fisher
WP-97-15
The Return from Community College Schooling for Displaced Workers
Louis S. Jacobson, Robert J. LaLonde and Daniel G. Sullivan
WP-97-16
Modeling Money
Lawrence J. Christiano, Martin Eichenbaum and Charles L Evans
WP-97-17
Monetary Policy Shocks: What Have We Learned and to What End?
Lawrence J. Christiano, Martin Eichenbaum and Charles L Evans
WP-97-18
Volunteer Labor Sorting Across Industries
Lewis M. Segal, Elizabeth Mauser and Burton A. Weisbrod
WP-97-19
Would Freetrade Have Emerged in North America without NAFTA?
Michael A. Kouparitsas
WP-97-20
Temporary Services Employment Durations: Evidence from State UI Data
Lewis M. Segal and Daniel G. Sullivan
WP-97-23
9
@FedFRASER