Can book-to-market, size and momentum be risk factors that predict economic growth?
Jimmy Liew and Maria Vassalou
Journal of Financial Economics 57 (2000) 221-245
Summary
Fama
and French (1992) show that the Capital Asset Pricing Model (CAPM)
cannot explain the cross-section of asset returns in the US. They
propose an alternative model which includes, apart from the market
factor, a factor related to book-to-market (B/M) which they call
High-Minus-Low (HML), and a factor related to size (MV) called
Small-Minus-Big (SMB). In a series of articles, Fama and French (1992,
1993, 1995, 1996) show that their model does a good job explaining
average equity returns. Nevertheless, it still remains unknown whether
their book-to-market and size-related factors have any economic
interpretation. Researchers have found little evidence of a relation
between these three return-based anomalies and intuitive economic risk
factors.
In
this paper, the authors test whether the profitability of HML, SMB, and
WML can be linked to future Gross Domestic Product (GDP) growth. They
construct return-based factors for ten developed markets. In addition to
HML and SMB factors, they also include the WML (winners minus losers)
factors which are returns to long-short portfolios constructed using
momentum information. Using regression analysis, Liew and Vassalou find
that the HML and SMB portfolios are related to future growth in the real
economy. High portfolio returns precede periods of high GDP growth.
Similarly, low portfolio returns are associated with low future GDP
growth. Also, their inclusion of other business cycle variables did not
eliminate the forecasting ability of HML and SMB. Moreover, the authors
showed that a positive relation also exists between the returns of HML
and SMB and future economic growth. Finally, they have also shown that
the HML, SMB, and WML zero investment portfolios provide positive
returns and that the HML and SMB strategies generate significantly lower
turnover than the WML strategy.
Comments
This
paper by Liew and Vassalou was the first to suggest that HML and SMB
contain information useful in predicting future GDP growth. I believe
that a good literature review would have been helpful in order to
introduce the different contributions in this area.
My
first comment concerns the data selection process. The authors did not
specify on which basis they were selecting the foreign countries. Fama
and French (1998) based their selections on the countries with MSCI
accounting ratios (B/M, E/P, C/P and D/P) for at least ten firms in
each. Moreover, MSCI includes only a subset of the firms in any market,
mainly those in Morgan and Stanley’s EAFE index or in the MSCI index
for a country’s market, and this means that most of these firms are
large. A test for size effect should have been performed in order to
control for the significance of the results. In addition, the returns
that are calculated from MSCI data are not without problems.
As
the information about dividends is only available from the dividend
yield (which is the ration of a trailing year of dividends to the
end-of-month stock price), we can calculate the annual return with
accuracy which are enough to estimate expected returns. However, the
test for the asset pricing models will be imprecise unless returns for
shorter intervals are used. The authors calculated monthly total
returns in local currency by spreading evenly the monthly dividends
throughout each year. On the other hand, this approach assumes that the
capital gain component of monthly returns reproduces the volatility and
covariance structure of total monthly returns. In addition, because
their portfolio construction procedure results in the creation of 27
portfolios, some of the portfolios for Italy, the Netherlands, and
Switzerland, in which the number of securities is much smaller than for
the other countries, may contain as few as one to three stocks.
Therefore, the results for these countries cannot be meaningful.
Moreover, their portfolio construction uses three sequential sorts,
which differ from the independent sort used by Fama and French (1993),
and this makes their results specific to this sorting procedure. Also,
the authors did not discuss any possible correlation of returns across
markets. The existence of such correlation might produce biased results
as well.
Other
possible areas that could be covered in this paper in order to make it
different from the Fama and French paper would be to use multivariate
models to help clarifying the fundamental roles played by correlated
macroeconomic factors examined in prior research, as well as assessing
the role of other factors. Furthermore, the authors can identify the
incremental information contained in the market portfolio, after
controlling for the selected macroeconomic fundamentals. The prior
literature focused on only partially overlapping sets of macroeconomic
state variables such as growth expectations, default risk, the term
structure of risk–free interest rates, and inflation. If the state
variables included in different studies are uncorrelated, this will not
cause any problems. However, when the state variables studied are
correlated, the significance of included macroeconomic instruments as
fundamental risk factors will be ambiguous and results will be biased,
because of a correlated omitted variable problem.
Some
innovations can provide possible solutions. For example, since GDP
data is only reported quarterly, one can use instead industrial
production data as a proxy for economic growth as it is reported
monthly. This can be done by using future realized economic growth as a
proxy for innovation in economic growth expectations. Although this
creates an errors–in–variables problem rendering model parameter
estimates unreliable (Petkova and Zhang, 2004; Greene, 2003), one can
avoid this by adopting an approach similar to Vassalou (2003) which
creates a factor–mimicking portfolio to capture the change in industrial
production growth expectations over the next year. The factor mimicking
portfolio is constructed by regressing log changes in realized
industrial production growth over the next year on the excess returns of
a set of base assets and a set control variables capturing information
on expected asset returns and growth (Lamont, 2001; Breeden et al.,
1989). Moreover, it would be interesting to see ho the results are
affected when we use log(GDP) (as it has been always used in economics)
instead of GDP with no transformation.
Finally,
momentum does require frequent trading. Annual winners and losers will
not change that often, but the winning and losing portfolios must still
be turned over at least once per year. Carhart (1997) calculates
transactions costs and concludes that momentum in not exploitable after
those costs are taken into account. Maskowitz and Grinblatt (1999) note
that most of the apparent gains come from short positions in small,
illiquid stocks, and positions that also have high transactions costs.
They also find that a large part of momentum profits come from short
positions taken November, anticipating tax-loss selling in December.
These findings point to the importance of small microstructure effects
for the risk and return in asset market that one should also consider.
Critical Review of
The Debt-Equity Choice
Armen Hovakimian, Tim Opler and Sheridan Titman
Journal of Financial and Quantitative Analysis Vol 36 (1), Mar (2001) 1-24
Summary
A
firm's history plays an important role in determining its capital
structure. Cross-sectional studies of leverage ratios reveal that a
firm's past profitability provides about as much explanatory power as
proxies for firm attributes suggested by theories that trade-off the
various costs and benefits of debt financing. The authors in this paper
employ a two stage regression procedure that allows them to gauge the
relative importance of a firm's history, its attributes and market
conditions in the determination of its capital structure choice. In the
first stage, debt ratios from a sample of Compustat firms during the
1979 – 1997, are regressed cross-sectionally on many of the variables
used in earlier cross- sectional studies. The difference between a
firm's actual debt/assets ratio and its predicted ratio is then included
in the second stage logit regression as a predictor of whether the firm
issues debt or equity. The second stage logit regression also includes
variables that measure the extent to which asymmetric information and
market conditions affect the debt/equity choice. Among their
conclusions, one indicates that the choice of the form of financing
should be examined separately from the choice of the size of financing.
Comments
The
authors examined one debt/asset ratios which is the book value of
short-term plus long term debt divided by the book value of debt plus
the market value of equity. This ratio probably provides the better
specification if managers focus on market value ratios when determining
their capital structure choices. However, since there is at least
anecdotal evidence suggesting that managers often focus on book values
in making these choices the book value of short-term plus long term debt
divided by the book value of total assets should have been studied as
well.
Their two stage approach differs from the empirical strategy of Mackie-Mason (1990) and
others who estimate a similar model, but all in one stage. Mackie-Mason
includes, for example, R&D spending/sales in his logit regression
since there are theories which suggest that high R&D firms tend to
use less debt financing. However, for issues that are relatively small,
this argument implicitly assumes that high R&D firms tend to be
overlevered prior to the issue, and we have no reason to believe ex
ante.
Moreover,
since they are sampling firms that issued debt or equity rather than a
combination, the two step procedure will also be misspecified when a
large amount of new capital is raised. This also means that the size of
the offering needs to be included in their specification and that it
should be interacted with the debt ratio of the firm. This might be done
through inclusion of a dummy variable, equal to one, for those
observations where the new financing increases the assets of the firm by
a certain amount and the product of this dummy variable and the firm's
debt ratio.
The
authors could also get further insights from an analysis of various
subsets of their sample. For example, they would expect that smaller
firms may be more subject to asymmetric information problems than large
firms and explore this possibility by separately estimating their second
stage regression on size stratified subsets of their sample. If
asymmetric information has a large effect on the debt/equity choice,
then the choices of smaller firms should be more influenced by their
past stock performance and by market conditions. They can also split the
sample by other indices of the severity of informational asymmetry
including a dummy for whether a firm is paying a dividend and the extent
to which their stocks are followed by analysts.
Most
of the studies of the firms’ debt-equity issuance (reduction) choice
utilize the standard Probit model. Meng and Schmidt (1985) show that
Bivariate Probit estimation is more efficient than that of independent
Probit equations when the error term of the two decisions are
correlated. On one hand, when unobservable factors that encourage firms
to issue debt also could lead them not to issue equity, the error term
of the interdependent decisions will be highly correlated, which in
return justifies the use of the Bivariate Probit estimation method.
Moreover, using the independent Probit equations model to study the
firms’ debt-equity issuance (reduction) choice leads to a loss of
information due to exclusion of combined issues of debt and equity. On
the other hand, the unobservable factors that encourage firms to issue
debt or equity could also affect the issuance (reduction) size as
proportion of the firms’ financing deficit (surplus). In this case the
Tobit estimation method is justified due to the censoring in the data
and to incorporate these unobservable factors into the estimation. One
can study the factors affecting the firms’ choice of the form of
financing (repurchase) and the size of issue (repurchase) using a two
stage Bivariate Probit – Tobit model. The first stage would examine the
factors affecting the firms’ choice of the form of financing
(repurchases) using Bivariate Probit model. Simultaneously, the second
stage would examine the factors that affect the size of issue
(repurchases) given that the firm decides to use a particular form of
financing (repurchases). The diagnostic tests might prove that using
the independent Probit equations model to study the debt-equity choice
yield inefficient estimators. Using the two stage Bivariate Probit-Tobit
model might rejects Hovakimian, Opler, and Titman's (2001) finding, due
to correlation between the factors affecting the choice of financing
and the size of issue.
Debt-equity
choice analysis, used in Hovakimian et al. (2001), does not allow us to
test how the deviation from the target affects each type of security
issue or repurchase. It is possible, for example, that firms issue
(retire) debt primarily when they want to adjust their leverage toward a
higher (lower) target, but that they issue (repurchase) equity
primarily to take advantage of favorable market conditions. One can
test the target leverage hypothesis separately for equity issues, debt
issues, equity repurchases, and debt reductions. This can be done by
examining the pattern of changes in leverage ratios and their deviations
from target ratios around these corporate financing events.
Finally,
some comments concern the data selection. The period includes the hot
market period in 1983, 1991-1996, where seasoned equity issuance is
high. This means that such a deviation should be interpreted very
cautiously or excluded from the sample. In addition, the measure of
short term debt and long term debt are biased as firms frequently report
revolving bank debt and commercial paper (short tem debt) as long term
debt when they plan to roll it over. Multicollinearity and
error-in-variables problems, due to the two stage regression technique,
and multiple appearances of the same firms in the sample, can be the
source of some results bias in this paper and can cause inconsistent
conclusions with previous literature.
Critical Review of
The Expiration of IPO Share Lockups
Laura Casares Field, Gordon Hanka
Journal of Finance Vol 56 (2), Apr (2001) 471-500
Summary
IPO
lockup agreements prevent insider sale of shares for a specific period
of time (often 180 days). This paper examine the share price reactions
around the expiration of IPO share lockups and try to find several
hypothesis that might explain its existence. Their results indicate
statistically significant negative abnormal return (AR) (-1.5%) in the
event window surrounding the expiration date and a permanent 40 percent
increase in trading volume. These effects were also three times larger
in venture-backed firms compared to non-venture backed firms. When they
looked at insider share sales in the year after the IPO, they found
direct evidence that venture-capital (VC) investors sell more
aggressively than other pre-IPO shareholders. The authors also found
permanent, parallel declines in both the bid and ask price and they
conclude that the AR is not caused by a change in the proportion of
trades at the bid price, temporary price pressure, or increased trading
costs. Their cross-sectional tests showed that the AR may be partly
caused by downward sloping demand curves or by consistently
larger-than-expected insider sales. These results provided only a lower
bound on the frequency of early share sales, as the data did not
include sales by low-level employees, share distributions by VC
partnerships, or effective share sales via hedging techniques such as
forwards, puts, collars, or borrowing against shares. Their sample
consisted of 1,948 IPO lockup agreements in the 10-year period from 1988
through 1997.
Comments
My
first concern in this article is about the data that was used. As it
was mentioned by the authors, the SDC data and those that were present
in the IPO prospectuses differed by a large percentage when they looked
at the number of shares offered, the length of lockup period, and the
number of shares locked up. They also showed that the SDC data are
wrong or missing in 45% of the cases. Thus, even by using the fraction
of shares not sold in the IPO as a proxy for the fraction locked up
(correlation 0.7), the results might still be inaccurate. The fact that
they claim that their results are not different when they use the
noisier measure, might be due to another variable which they don’t know
to affect both cases. Using the data they collected from the IPO
prospectuses would have been safer in this case.
Another
point concerns the measure of AR. It would have been interesting to
see whether or not the results were affected by time variation in betas.
One possible way to do that is to use two estimation approaches for AR
as was performed by Brau et al (2004). The first one would rely on
beta estimates obtained from regressing company daily stock returns
beginning for 90 days prior and ending 11 days prior to lockup date on
the CRSP equally-weighted return index, and the second approach utilizes
a beta estimate obtained from a 200-day window commencing 11 days after
the lockup expiration date.
The
authors looked at many hypotheses to explain the negative abnormal
return but did not find a conclusive one. The lockup expiration date
has institutional characteristics that are similar to those at the IPO
date, and asymmetric information has been suggested as an important
determinant of IPO underpricing. We can try to see whether or not we
should expect the same factors that affect underpricing to have
explanatory power at the lockup expiration date. Ibbotson and Ritter
(1995) state that investors are willing to pay more for a firm with
lockup agreement for two reasons:
- any
negative information being withheld is likely to be divulged before the
lock-up shares can be sold, reducing the benefit of withholding
information
- as long as insiders retain large holdings, their incentives are aligned with outsiders’ incentives.
Moreover,
many empirical works have found that retained ownership by insiders at
the date of the IPO is positively related to firm value. However,
lockup periods can be relatively short and this means that little
mandated information is disclosed between the IPO and the lockup
expiration date. Therefore, the lockup agreement cannot mitigate
informational asymmetries that exist between insiders and outsiders. In
addition, lockup expiration will increase the potential for unaligned
insider and outsider incentives, and this makes the potential agency
costs decrease investor demand for shares, assuming that this lack of
alignment adversely affects general shareholders. Thus, any withheld
information about future prospects is more likely to be negative than
positive and this will in negative AR in days preceding lockup
expiration date. Bank (1999) explains the cause of the negative stock
price reaction to be cause by the expectation of the sale of
newly-released lockup shares in the open market, and Browning (1999)
reports that investors worry about a ‘flood of sales’ and sell stocks in
advance of lockup expiration.
It
is thus interesting to look at evidence for information asymmetry
between insider and outside shareholders at the lockup expiration date.
Some proxies that can be used for asymmetric information are the
underwriter reputation(as underwriter prestige is associated with
marketing of lower risk IPOs (Carter and Manaster 1990, Hanley 1993),
size (as larger firms are expected to have greater information available
to markets and we would expect less uncertainty surrounding the pricing
of these firms (Barry and Brown 1984), growth opportunities (as there
is more uncertainty regarding the valuation of high growth firms
(Garfinkel 1993), and offer price (as offer price data constitute a
measure of risk of the IPOs (Beatty and Ritter 1986, Tinic 1988).
Other
lockup parameters that the authors could have used to increase the
significance of their results and help them assessing their hypotheses
are the number of days in the lockup period as the longer the period the
more information that will be available for the market, and the
percentage of shares in lockup as those with less than 100% in lockup
can signal to the market with their trading of unlocked shares.
Some
other related works to be looked at are the ones by Brav and Gompers
(2000 and 2003), and Tolia and Yip (2003). Brav and Gompers (2000)
provide evidence on the role of analysts around lockup expiration. They
show that analysts, both affiliated and unaffiliated, tend to issues
more optimistic earning forecasts at the time of the lockup expiration.
Brav and Gompers (2003) find support for the notion that lockup serve
as a commitment device to overcome moral hazard problems subsequent to
the IPO. Firms that are unprofitable, with low B/M ratio, and that go
to public with lower-quality underwriters and are not venture
capital-backed have significantly longer lockups. Finally, Tolia and
Yip (2003) try to investigate whether the stock price behavior around
the lockup expiration date is different for “hot” (with first day
returns between 10 and 60%) and “cold” (with zero or negative first day
returns) IPOs. Their methodology; however, need to be further explored
in order to account for other factors that will affect this difference
in the AR.
Critical Review of
Dividends, Share Repurchases, and the Substitution Hypothesis
Gustavo Grullon, Roni Michaely
Journal of Finance Vol LVII (4), Aug (2002) 1649-1684
Summary
The
three goals of this paper are to analyze the recent trend in share
repurchases, to determine if share repurchase is a substitution for
paying out dividends, and to try to explain the reason why firms did not
substitute repurchases for dividends earlier. The authors use a sample
of 15,843 firms over the period 1972 to 2000 from Industrial Compustat
files. When assessing the first objective, the authors find that in the
last 15 years, the majority of the firms used share repurchases instead
of dividends to initiate cash payouts. The larger firms did not cut
the amount of dividends paid, but the growth rate of the dividends was
found to be lower than what it used to be in mid 1980, while the amount
of repurchases was higher. When looking at the second objective of the
paper, the authors find that the marked increase in share repurchase
programs in the United States has been financed with potential
increases in dividends, and that the payout ratio of the firms has
remained relatively stable, which supports their substitution
hypothesis. Using Lintner’s (1956) dividend model to generate expected
future dividends payments, Grullon and Michaely find that dividend
forecast errors are negatively correlated with share repurchase
activities. The authors also report that the market reaction around the
announcement of dividends decreases is significantly less negative for
repurchasing firms than for non repurchasing firms. They also show that
differential taxes between dividends and capital gains seem to matter in
hat the market reaction to repurchases is more positive when the tax
gains from them relative to dividends are larger. Finally, when looking
at the third goal of the paper, the authors conclude that share
repurchase activity experienced a upward shift after the adoption of
Rule 10b-18, which is consistent with the idea that share repurchase
were restricted in the past.
Comments
The
question of how corporations choose their payout policy between the two
has attracted considerable attention from researchers in the last two
decades. One of the most significant trends in corporate finance during
the 1990s was the increasing popularity of open-market stock repurchase
programs. Share repurchases have increased over time, and now constitute
an economically important source of payouts. There are several comments
that the authors of this paper can address to better understand the
interaction between dividends and share repurchases.
First, as
a preliminary step to investigating the substitution hypothesis between
dividends and share repurchase, the authors could have regressed share
repurchases (RP) on dividend changes (ΔD), repurchase payout ratios
(RP/Y) on dividend payout ratios (D/Y), and repurchase payout ratios
(RP/Y) on dividend changes (ΔD). This would help determining whether
these two are real substitutes or complements. Share repurchases can
also complement dividends as a method of returning cash to shareholders.
These two methods are not necessarily close substitutes:
- Dividends
received may be subject to income tax, whereas repurchase receipts may
be subject to capital gains tax. The so-called ‘Dividend Policy Puzzle’
in the US is founded on this difference. Dividends dominate share
repurchase as a method of distribution in spite of the fact that tax on
capital gains is lower. (However, in 2003 President Bush proposed the
abolition of taxation on dividend income. This would serve to remove the
tax disadvantage of dividends.)
- The
signalling properties may differ. For example, variation in the rate of
change in dividends may signal good or bad news. Share repurchases, on
the other hand, are generally interpreted positively.
- Share
repurchases are relatively lumpy and subject to greater managerial
discretion. Dividend payments are regular and generally smoothly
increasing.
- Share repurchases may involve greater ‘out-of-pocket’ expenses such as fees to merchant bankers
- Repurchases may be associated with wealth transfer between shareholders because, unlike dividends, they are non-proportional.
Consistent
with the above, Jagannathan, Stephens and Weisbach (2000) found that,
for US companies, repurchase and dividends are used at different times
from one another by different kinds of firm. Repurchases are very
pro-cyclical, while dividends increase steadily over time. Dividends are
paid by firms with higher ‘permanent’ operating cash flows, while
repurchases are used by firms with higher ‘temporary’ non-operating cash
flows. Repurchasing firms also have much more volatile cash flows and
distributions. Finally, firms repurchase stock following poor stock
market performance and increase dividends following good performance.
The authors comment that ‘these results are consistent with the view
that the flexibility inherent in repurchase programs is one reason why
they are sometimes used instead of dividends’.
Second, there are three ways to look at dividend and share repurchases policies. The
methodology must allow for different decision-making processes. It
should be possible to test three structures, (i) unrelated decisions
(i.e., dividend and share repurchases choice are not related), (ii) simultaneous decisions (i.e., dividend and share repurchases choices
are mutually related), and (iii) sequential decisions (i.e., firms
first decide on whether to pay out or not, and thereafter choose between
dividends and share repurchases). In order to do that, one can use three different logit models to capture these three decision-making processes.
Third,
share repurchases can be simultaneously motivated by various
combinations of many reasons that are not mutually exclusive. Grullon
and Ikenberry (2000) provide five theoretical explanations for stock repurchases. These
are: to distribute funds in excess of investment opportunities (free
cash flow or agency motivation); to signal the market that its shares
are undervalued or simply to take advantage of such undervaluation to
increase the wealth of the remaining shareholders (information asymmetry
motivation); to fend off unwanted takeovers (anti-takeover motivation);
to achieve or maintain target leverage ratios, when such ratios fall
below perceived optimal levels (target-leverage-ratios motivation); and
to avoid the negative impacts of dividends on stock options’ values or
to impact or manipulate reported EPS in one or both of its reported
forms primarily due to the presence of ESOs (stock options motivation).
Grullon
and Michaely report in this paper that, by repurchasing stock rather
than increasing dividends, firms are substituting share repurchases for
cash dividends, although the authors do not attempt to empirically link
such substitution to the presence of ESOs. Another way posits that share
repurchases are motivated by management’s desire to “undo” the dilution
created by stock options. The concept of the undo dilution hypothesis,
as defined by Weisbenner (2000), is that, by buying back shares and thus
decreasing the denominator in EPS calculations, the effects of the
additional shares which are included in the denominator due to stock
options can be offset or undone. Weisbenner (2000) concludes that stock
option grants result in increased share repurchases and increased total
payouts, with the size of the option grants predicting the amount of
repurchases.
In the
past, each hypothesis tended to be addressed separately (or
independently) using a different framework. Thus, it is best to provide a
unified framework and treatment of various hypotheses on share
repurchases so that we obtain a comprehensive evaluation of them. One
can use the VAR framework as a common methodology for examining various
hypotheses regarding the two payout policies. Each hypothesis would be
formulated based on a VAR of relevant variables and characterized as
restrictions on the VAR of the variables.
Fourth,
the authors can investigate payout policy in a time-series VAR context
so that the dynamic and multi-dimensional nature of the two payout
policies is accounted for. Other studies tend to focus on examining
various motives for share repurchases using cross-sectional data, and
they have contributed to understanding of share repurchases. However,
they often provide conflicting evidence. Given that share repurchases
are not a regular event for individual firms, the authors can focus
their analysis on an aggregate time-series analysis between the two
payout policies (see Dittmar and Dittmar (2002) for a recent study on
aggregate patterns in repurchases).
Finally,
if their sample period was larger, this would allow them to estimate
the impact of macro-economic parameters on the distribution of corporate
payout. A possibility would be to spans the years from 1950 through
1997 (the full set of years for which COMPUSTAT data are available on
any of the COMPUSTAT tapes).
Critical Review of
Debtor-in-Possession Financing and Bankruptcy Resolution:
Empirical Evidence
Sandeep Dahiya, Kose John, Manju Puri, Gabriel Ramírez
Journal of Financial Economics Vol 69 (2003) 259-280
Summary
This
paper starts by addressing what distinguishes firms that obtain DIP
financing from firms that do not. Dahiya et al. find that larger firms
are more likely to obtain DIP financing. They also find that firms,
which obtain DIP financing, are more likely to emerge from the Chapter
11 process than firms, which do not. The authors conclude that these
results are consistent with DIP lenders having an information based
role; playing a screening role in which they are able to identify
distressed firms that are strong and likely to emerge quickly, as well
as a monitoring role in which the DIP lenders help firms to emerge
quickly. The authors then explain how is DIP financing related to the
probability and speed of bankruptcy resolution and investigate whether
or not it makes a difference if the DIP financier is an existing
creditor with a prior lending relationship with the firm. They find that
firms that receive DIP financing take a shorter time to resolve their
Chapter 11 filing. They explain that it suggests that DIP lenders
finance positive NPV projects and help the company emerge from
bankruptcy, but if things do not go well with the firm then they are
quick to liquidate to preserve the value of the underlying assets.
Dahiya et al. also find that while many firms receive DIP financing from
an existing lender, a significant number obtain it from a lender with
whom they have no existing lending relationship. The authors use the Bankruptcy DataSource from
New Generation Research Inc. (NGR) to assemble their sample which
consists of 538 firms that filed for Chapter 11 protection between
January 1988 and December 1997 of which 165 received DIP financing.
Comments
The
first point is with respect to the characteristics of firms obtaining
DIP financing. The authors look at leverage, assets level, prepackaged
Chapter 11 and retail as independent variables. One also expects that a
lender will be more inclined to provide DIP financing if the Chapter 11
debtor has the right people who would lead to higher probability of
reorganization given the quality of the business. These two
characteristics should have been included in this analysis to give a
better conclusion.
Several
measures should be examined as well, including income over sales and
the ratios income/assets, and revenue/assets (using log) measured at the
year-end prior to filing for bankruptcy, in order to see if DIP firms
are more profitable than non-DIP firms, as it is expected. Moreover, as
Gilson (1989) observed that bank lenders frequently initiate top
management changes in financially distressed firms. This management
turnover is expected to have a positive impact on the value of the firm.
Moreover, Chatterjee et al. (2001) argue that the evidence of
management turnover in firms with DIP financing is not significantly
different from non-DIP firms. Since the DIP lender is usually a bank
with considerable expertise in the DIP financing market and the
bankruptcy process, one should expect a larger monitoring and
disciplining role in DIP firms, and eventually more management turnover,
with firm value-enhancement (Jensen, 1989). A dummy with a value of 1
in case of changes of CEO during bankruptcy and 0 otherwise can be used.
When
the authors looked at the relationship between DIP financing and
bankruptcy outcomes, they used a Probit model. They could have
separated the sample into two categories: (1) and without DIP financing
(2) and firms who got DIP financing only. The relative loan size should
also be examined as a possible variable affecting the outcome. Other
characteristics might include the DIP loan coming from a bank versus
non-bank as banks invest in information gathering technology and may
have a competitive advantage relative to other non-bank institutions in
evaluating lending opportunities. Moreover, a measure of
solvency/profitability should be included as the more solvent/profitable
the firm, the more likely it is that it will reorganize successfully.
Since equity committees are typically appointed when the firm is
solvent, we should expect their presence to be associated with
successful emergence from bankruptcy. Several measures can be used,
including assets/liabilities, revenue/assets and income/assets, measured
at the year-end prior to filing for bankruptcy. Other good indicators
would be the ratio of total distribution/estimated total debt and a
dummy with a value of 1 in case of equity committee and 0 otherwise.
Furthermore,
since Gilson et al. (1990) argue that financial distress is more likely
to be resolved through a private renegotiation in three situations: 1)
the larger the loss of going concern value if a private renegotiation
failed. 2) the larger the proportion of bank & insurance companies
debt in terms of the total liabilities. 3) the smaller the number of
classes of debt outstanding (see also Betker, 1995), one might argue
that these variables also determine the likelihood of a successful
emergence, following a Chapter 11 filing. The ratio of bank &
insurance companies’ debt/liabilities, the ratio of the market value of
common equity to its liquidation value and the (log) number of creditor
classes, respectively, can be used in order to examine this relation.
One should be careful when using these variables and look at the
possibility of muticollinearity.
Although,
Schwarcz (1985) argues that obtaining new financing can act as a good
signal to trade creditors and encourage them to re-establish the terms
of the trade credit with the company, the positive impact of DIP
financing should be reduced in two circumstances. First, when the new
loan is secured by a lien on already encumbered assets with equal or
senior priority to the existing liens (priming liens). Secondly, when
the DIP lender is a pre-existing creditor and obtains an increase in the
seniority of his prepetition debt. In fact, these situations suggest a
lack of confidence of the DIP lender in a successful reorganization of
the firm. Two dummy variables with a value of 1 in case of DIP financing
(DIP priming liens) and 0 otherwise can be used. Moreover, one
important feature of DIP financing is the time it takes the court to
approve an order. Many times we can have a ‘filing on a pillow’, but
other times the approval might take longer. One can hypothesize that the
more time it takes to obtain DIP financing, the more likely the firm is
to reorganize successfully in Chapter 11, as a careful analysis of the
situation of the bankrupt firm should be required. The variable ‘time to
obtain DIP financing after filing for bankruptcy’ in days should also
be used.
Finally,
trustees are typically appointed in cases when the management is guilty
of fraudulent behavior (under Subsection 1104(a)), but this is usually a
temporary measure, until the company appoints new management. So, the
appointment of trustees is closely related to management turnover during
bankruptcy, thus compounding all the complexity of the process. A dummy
variable with a value of 1 in case of a trustee being appointed and 0
otherwise can be added as well.
Critical Review of
Asset Pricing with Conditioning Information: A New Test
Kevin Wang
Journal of Finance, Vol. LVIII, No. 1, Feb 2003, p161
Summary
This
paper develops a new approach to testing dynamic linear factor models,
which aims at time-variation in Jensen's alphas while using a
nonparametric pricing kernel to incorporate conditioning information. In
application we find that the conditional CAPM performs substantially
better than the static CAPM, but still it is statistically rejected.
Theoretically, it is well known that the conditional CAPM could hold
perfectly – that is, conditional alphas are always zero – but that
time-variation in beta might lead to unconditional pricing errors. In
general, a stock’s unconditional alpha will differ from zero if its beta
covaries with the market risk premium or with market volatility. The
conditional CAPM exhibits time-varying Jensen's alphas that have a
strong size pattern in volatility and a clear book-to-market pattern in
time-series average. These features are well captured by a simple
dynamic version of the Fama and French (1993) three factor model. In
fact, the market portfolio might be conditionally mean-variance
efficient in every period yet, at the same time, not on the
unconditional mean-variance frontier. Several recent papers argue, in
fact, that time variation in beta helps explain the size, B/M, and
momentum effects. Zhang (2002) develops a model in which high-B/M stocks
are riskiest in bad times, and co movement between betas and the risk
premium leads to an unconditional value premium (even though conditional
CAPM alphas are exactly zero).
The
author uses a methodology that avoids specification of time-varying
betas. Instead he uses a nonparametric technique that will enable him
to avoid misspecification about dynamics of conditional betas. He found
that for the conditional CAPM, the nonparametric technique performs
better than the unconditional CAPM. He also finds that the pricing
errors of the conditional CAPM are positively correlated with stock
market. Moreover, the use of labor income risk factor of Jagannathan
and Wang (1996) and the size and book-to-market factors of Fama and
French in this paper did not seem to explain satisfactorily the CAPM
pricing errors. Finally, when he challenged the conditional CAPM and
the conditional Fama and French model with momentum portfolios, Wang
finds that the conditional expected returns of the winners are higher
than those of the losers.
Comments
The
paper focused on modeling the variation in betas using theoretical
framework of the conditional CAPM in a very intelligent way. The
literature overview was thorough and specific. The author was very
explicit in his model, his assumptions, and the data he used. Every
step in his methodology was very sound and supported by arguments. I
can only add few comments to this paper.
Ghysels
(1998) discussed the problem of time variation in betas and risk premia
in detail and stressed the impact of misspecification of beta risk
dynamics on inference and estimation. Also, he argued that betas change
through time very slowly and linear factor models like the conditional
CAPM may have a tendency to overstate the time variation. Further, he
showed that among several well-known time-varying beta models, a serious
misspecification produces time variation in beta that is highly
volatile and leads to large pricing errors. Finally, he concluded that
it is better to use the static CAPM in pricing when we do not have a
proper model to capture time variation in betas correctly.
Large
pricing errors could be due to the linear approach used in a nonlinear
model. Therefore, by treating a non-linear relationship as a linear one
would lead to serious prediction problems in estimation. In order to
overcome these problems, some nonlinear models were suggested in the
literature.
In this
article, Wang explored a nonparametric form of the SDF model as a new
way for dealing with the misspecification errors and conducted a test
based on the nonparametric model. Parametric models for time-varying
betas can be most efficient if the underlying betas are correctly
specified. However, a misspecification may cause serious bias and model
constraints may distort the betas in local area. A nonparametric
modeling is appealing in these situations. One of the advantages for
nonparametric modeling is that no or little restrictive prior
information on betas is needed. Further, it may provide useful insight
for further parametric fitting.
The
conditional CAPM is typically well supported by the data, and thus, it
is important that we test (as it is done in this paper) whether all
coefficient functions are actually varying (namely, if a linear beta
CAPM is adequate); if a parametric model fits the given data such as
testing for structural breaks as in Ghysels (1998) or testing a
threshold model as in Akdeniz, Altay-Salih and Caner (2003) or a
specific parametric form as in Ferson and Harvey (1998, 1999); if there
is no β0(.) at all; and whether some or all coefficient functions are
constant or zero or in a certain parametric form.
In
empirical finance, different models impose different constraints on the
SDF. Particularly, the SDF is usually assumed to be a linear function
of factors in various applications. Further, when the SDF is fully
parameterized such as linear form, the general method of moments (GMM)
of Hansen (1982) can be used to estimate parameters and test the model.
Recently, some more flexible SDF models have been studied by several
authors. Wang pointed out that one of the shortcomings is that it is
difficult to obtain the distribution theory and the effective assessment
of finite sample performance. To overcome this difficulty, instead of
considering the nonlinear pricing kernel, Ghysels (1998) focused on the
nonlinear parametric model and used a set of moment conditions suitable
for GMM estimation of parameters involved. Wang, in this paper, studied
the nonparametric conditional CAPM and gave an explicit expression for
the nonparametric form of conditional CAPM for the excess return. Wang
suggested estimating b(.) by using the Nadaraya-Watson method. Also, he
conducted a simple nonparametric test about the pricing error. He also
extended this setting to multifactor models by allowing b(.) to change
over time; that is, b(Xt) = b(t).
Another
way that might be interesting to study would be to consider the
nonlinear SDF without specifying a particular form of the nonlinear
pricing kernel using the estimation method of GMM.
References
- L. Akdeniz, A. Altay-Salih, M. Caner, "Time Varying Betas Help in Asset Pricing: The Threshold CAPM ," Studies in Nonlinear Dynamics and Econometrics, vol. 6, pp. 1-16 (2003)
- Barry, C.B. and S.J. Brown, 1984, Differential information and the small firm effect, Journal of Financial Economics, 13, 3–29.
- Jim Bau, Dave Carter, Steve Christophe, and Kim Key,"Market Reaction to the Expiration of IPO Lockup Provisions," Managerial Finance, Vol. 30, No. 2, 2004, pp. 87-103.
- Beatty, R.P. and J. R. Ritter, 1986, Investment Banking, Reputation, and the Underpricing of Initial Public Offerings, Journal of Financial Economics 15: 213-232
- Brian
L. Betker, 1995, "Management's Incentives, Equity's Bargaining Power,
and Deviations from Absolute Priority in Chapter 11 Bankruptcies," Journal of Business 68, p. 161
- Brav, A., C. Geczy, and Paul Gompers. "Is the Abnormal Return Following Equity Issuances Anomalous?" Journal of Financial Economics 56, no. 2 (May 2000).
- Brav, Alon and Paul Gompers, 2003, The Role of Lock-ups in Initial Public Offerings, Review of Financial Studies 16, 1-29.
- Breeden,
D. T., M. R. Gibbons, and R. H. Litzenberger (1989). Empirical tests of
the consumption-based CAPM. Journal of Finance 44 (2), 231–263.
- Carhart, M. M. (1997). On persistence in mutual fund performance. Journal of Finance 52 (1), 57–82.
- Carter, R. and S. Manaster, 1990, "Initial Public Offering and Underwriter Reputation," Journal of Finance 45(4), 1045-1067.
- Chatterjee,
Sris, Upinder S. Dhillon and Gabriel G. Ramirez, 2001,
Debtor-inpossession financing, Working paper (Fordham University).
- Dittmar,
Amy, K. and Robert F. Dittmar, 2002, Stock repurchase waves: An
explanation of the trends in aggregate corporate payout policy, Working
paper, Indiana University.
- Fama, Eugene F., and Kenneth R. French, 1992, The Cross-Section of Expected Stock Returns, Journal of Finance 47, 427-465.
- Fama, Eugene F., and Kenneth R. French, 1993, Common Factor Portfolios in the Returns on Stocks and Bonds, Journal of Financial Economics 33, 3-56.
- Fama, E. F. and French, K. R. (1995), ‘Size and book-to-market factors in earnings and returns’, Journal of Finance, 50, 131–155.
- Fama, E. F. and French, K. R. (1996), ‘Multifactor explanation of asset pricing anomalies’, Journal of Finance, 51, 55–84.
- Fama, E. F. and French, K. R. (1998), ‘Value vs growth: the international evidence’, Journal ofFinance, 53, 1975–2000.
- Ferson, W. E. and C. R. Harvey, 1999, “Conditioning Variables and the Cross Section of Stock Returns,” Journal of Finance, 54, 1325-1360.
- Garfinkel,
J., 1993. IPO underpricing, insider selling and subsequent equity
offerings: is underpricing a signal of quality? Financial Management 22,
74-83.
- Ghysels, E., 1998, “On Stable Factor Structures in the Pricing of Risk: Do Time-Varying Betas Help or Hurt,” Journal of Finance, 53, 2, 549-574.
- Gilson, S. C. "Management Turnover and Financial Distress." Journal of Financial Economics 25 (December 1989): 241-262.
- Gilson,
S. C., J. Kose, and L. H. P. Kang. "Troubled Debt Restructurings: An
Empirical Analysis of Private Reorganization of Firms in Default." Journal of Financial Economics 26 (September 1990): 315-353.
- Greene, W. H. Econometric Analysis, 5th Ed., Prentice Hall, 2003
- Grullon, Gustavo, and David Ikenberry, 2000, What do we know about share repurchases?, Journal of Applied Corporate Finance 13, no. 1, 31-51.
- Hanley, K. W. 1993. The Underpricing of Initial Public Offering and the Partial Adjustment Phenomenon. Journal of Financial Economics (October): 231-250.
- Hansen, Lars Peter, 1982. "Large Sample Properties of Generalized Method of Moments Estimators," Econometrica, Econometric Society, vol. 50(4), pages 1029-54.
- Hovakimian, A.; T. Opler; and S Titman, S., 2001. “The debt-equity choice.” Journal of Financial and Quantitative Analysis 36, 1-24.
- Ibbotson, R. G. and Ritter, J. R. (1995) “Initial public offerings” In: Jarrow R.A.,
- Maksimovic V. and Ziemba, W.T. (eds.) Handbooks in Operations Research and
- Management Science, Volume 9: Finance, 1st ed., North Holland: Elsiver Science pp 993-1016.
- Jagannathan, Ravi, and Zhenyu Wang (1996), `The Conditional CAPM and the Cross-Section of Expected Returns,' Journal of Finance, Vol. 51, pp. 3-53.
- Jagannathan,
M., C. P. Stephens, and M. S. Weisbach, 2000, Financial flexibility and
the choice between dividends and stock repurchases, Journal of Financial Economics 57, 355-384.
- Lamont, O. A. (2001). Economic tracking portfolios. Journal of Econometrics 105 (1), 161–185.
- MacKie-Mason, Jeffrey K., 1990, Do taxes affect corporate financing decisions?, Journal of Finance 45, 1471-1495.
- Meng, C., and P. Schmidt, 1985.’’ On the Cost of Partial Observability in the Bivariate Probit Model.” International Economic Review 26, 71-85.
- Moskowitz, Tobias, and Mark Grinblatt, 1999, .Do industries explain momentum?.
- Journal of Finance 54 (No. 4), pp. 1249-1290
- Petkova,
Ralitsa, and Lu Zhang, 2004, Is value riskier than growth? Journal of
Financial Economics Volume 78, issue 1, 2005 pp. 187-202
- Tiniç, Seha M.; "Anatomy of Initial Public Offerings of Common Stock;" Journal of Finance; Vol. 43, Iss. 4; (September 1988); pp. 789-822
- Vassalou,
M. (2003). News related to future GDP growth as a risk factor in equity
returns. Journal of Financial Economics 68, 47–73.
- Weisbenner,
Scott J., 2000, Corporate share repurchases in the 1990s: What role do
stock options play?, Federal Reserve Board, working paper.
- Zhang, X., 2002, “Specification Tests of International Asset Pricing Models,” working paper, Cornell University.
