8/10/2019 How Do Underwriters Trade_CMCRC Research Report
1/28
1
The hidden cost of underwriting
Nicholas Pricha, Sean Foley
*
, Graham Partington and Jiri SvecUniversity of Sydney
30 April 2014
Abstract
We examine agency costs around underwritten seasoned equity offerings (SEOs),
focusing on underwritten dividend reinvestment plans (DRIPs). The underwriters have anincentive to sell stock during the pricing period for the issue. This reduces the price at whichshares are issued and can increase the returns to underwriting. Using data for individual
brokers transactions, we show that underwriting brokers engage in an abnormally high levelof selling during the issue pricing period. Comparison of pricing period returns between stockwith underwritten DRIPs and a matched sample of non-underwritten DRIPs shows thatsignificantly more negative returns accrue to firms that have their issues underwritten.
JEL classification:G14
Keywords: Agency Conflict, Reinvestment plans, DRIP, Underpricing, Underwriting,
SEO
*Email:[email protected]. The authors thank the Securities Industry Research Centre of Asia-Pacific
(SIRCA) for the provision of data and the Capital Markets CRC Limited (CMCRC) for financial support. Theauthors would also like to thank the participants at the JCF Schulich conference on market misconduct as wellas Ryan Davies, Terry Walter, Alex Sacco, Reuben Segara and Angelo Aspris for their thoughtful comments.
mailto:[email protected]:[email protected]:[email protected]:[email protected]8/10/2019 How Do Underwriters Trade_CMCRC Research Report
2/28
2
1 Introduction
Conflicts of interest and the consequent agency costs are regularly encountered in
corporate finance. Such agency conflicts include management manipulating earnings to
maximize the value of funds raised from seasoned equity offerings (SEOs),1
executives
manipulating the timing of information releases to maximize the value of issued stock
options2 or to meet analyst expectations3 and underwriters over-allotting initial public
offering (IPO) stock to profit from the Green Shoe option (Fishe (2002), Aggarwal (2003),
Zhang (2004), Jenkinson and Jones (2007)).
In this paper we examine agency conflicts arising from underwriting in the context of a
seasoned equity offering (SEO). We examine a unique institutional setting where the issue
price is based on an average of the prices at which the stock trades prior to the issue and
where the underwriter has the opportunity to influence this price by trading during the period
over which the price is determined. Furthermore, during the pricing period the underwriter
knows how much stock they will be called upon to take up.
At the time the underwriting agreement is struck the underwriter faces the risk of having
to take up stock if there is a participation shortfall, but typically they are not prohibited from
trading during the pricing period. The underwriters have an incentive to temporarily depress
the stock price during the pricing period. This will lower the issue price and thus provide
extra profit on the underwriters stock allocation assuming the price bounces back from a
temporarily depressed state. Our hypotheses, therefore, are that underwriters engage in
abnormal selling activity over the pricing period and that the stock price is abnormally
depressed during this period.
We focus on new issue dividend reinvestment plans (DRIPs), a subset of SEOs, which
provide the unique institutional setting described above. We utilize a dataset which identifiesthe buying and selling broker for every trade, allowing the buying, and selling behavior of all
brokers to be identified. We test our hypotheses in the Australian market where DRIPs are
invariably new issue DRIPs and are an important source of funds. In 2009, 230 ASX listed
companies raised $11.4 billion using DRIPs, representing 18% of total secondary offerings
1Many firms have documented the subsequent underperformance of SEOs due to accrual management (Rangan,1998, Teoh, Welch and Wong, 1998), real earnings management (Cohen and Zarowin, 2010), and liquidity risk(Lin and Wu, 2013).2
For more on executive options timing see Yermack (1997) and Chauvin and Shenoy (2001).3Marciukaityte and Varma (2013) document executive management of earnings to meet analyst expectations.
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
3/28
3
by large corporations.4 Such is the importance of DRIPs as a source of finance that some
firms choose to have their DRIPs underwritten in order to guarantee the amount of capital to
be raised.
In support of our first hypothesis, we observe aggressive selling by the underwriting
brokers during the pricing period. Abnormal volume is 236% higher than during the
preceding benchmark period. Whilst it is not possible to determine from our data whether the
trades by underwriting brokers are proprietary or client facilitation, we only observe
significant abnormal selling in the pricing period by the underwriting broker. No such selling
is observed for non-underwriting brokers in the underwritten DRIPs, or by brokers in our
matched sample of non-underwritten DRIPs. The results are consistent with manipulation of
the share price by the underwriter during the pricing period in order to generate additional
profit. However, we cannot rule out the alternative motivation of sales to hedge the price risk
of the allocation. If hedging is the motive, the hedging is only partial since on average less
than half of the underwriters allocation is sold during the pricing period. This leaves stock
available for resale and provides the potential to profit from a price rebound at, or subsequent
to, allocation. Whatever the underwriters motive, the consequence is clearly abnormal sales
and, as discussed below, a depressed issue price.
In support of our second hypothesis, we find that underwritten DRIPs have negative
abnormal returns of 4% during the pricing period, which is significantly worse than the
negative abnormal returns of 2.3% experienced by non-underwritten DRIPs. The temporary
decrease in the market price of the stock during the pricing period leads to a reduction in the
issue price of the DRIP shares, resulting in a benefit to all participants in the DRIP
(particularly the underwriter). However the adverse impact on the share price and the lower
issue price is to the detriment of non-participating shareholders.
This paper proceeds as follows. In section 2 we review the DRIP issue process, the
incentives created by the process and how, consistent with their incentives, underwriters can
manipulate prices. Section3 describes the data and method. Section4provides a discussion
of the results while section5 concludes.
4ASX Annual Report, 2010.
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
4/28
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
5/28
5
ex-dividend date. Since the pricing period spans the record date the shortfall will be known
before the pricing period ends.
There is evidence that the management of some companies have concerns over price
pressures caused by UDRIPs. For example, Orica Ltd increased their DRIP pricing period
from 7 to 12 trading days when an underwriter was appointed ...so that the [issue] price was
impacted less by short term variations in the companys share price. Orica Ltd. (ASX:ORI),
23/10/2007.6 More generally, we find that management increase the length of the pricing
period upon the appointment of an underwriter in 46 out of our original sample of 126
UDRIPs (36.51%).
2.2 Agency Conflicts and Underwriter Incentives
This study relates to the broader literature on the link between underwriters incentives
and equity issue underpricing. In IPOs, the existence of the Green Shoe option to oversell
shares in the IPO is shown by Fishe (2002) to create an agency conflict between the
underwriter and the issuing firm, which results in IPO underpricing. The underwriter is able
to oversell the issue, selling a greater number of shares than are actually on offer. Such a
practice necessitates that the underwriter covers this short position. This can be accomplished
either by on-market purchase or through the use of the Green Shoe option. 7 Fishe (2002)
shows that this is analogous to a call option, allowing the underwriter to purchase short-sold
shares at the market price if the price subsequently falls below the issue price, or by using the
Green Shoe option should the price rise. The structure of this call option combined with the
impact of stock-flippers (traders who purchase in the IPO and sell immediately in the
secondary market) results in the underwriter underpricing the issue, to the detriment of the
issuing firm. Empirical support for the model of Fishe (2002) is documented by Aggarwal
(2003) and Ellis, Michaely, and OHara (2000), with Green Shoe options found to be fully
utilized for issues with prices that rise and avoided when post-issue prices fall. This body of
literature on IPO underpricing suggests the presence of an agency conflict as underwriters
maximize their own profit instead of acting in the best interests of the firm. Similarly, in an
underwritten DRIP there is an incentive for the underwriter to manipulate the DRIP issue
price to extract an increased profit, to the detriment of the issuing firms shareholders.
6Firms that have either terminated their underwriting agreement or replaced it with a private placement include
SuncorpMetway Ltd (ASX:SUN), 19/09/2008 and Transpacific Industries Group Ltd. (ASX:TPI), 03/10/2008.7See Aggarwal (2003) for a detailed discussion of the Green Shoe option. On NASDAQ this option is restrictedto 15% overallotment, and the option must be exercised within 30 days.
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
6/28
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
7/28
7
On the basis of the discussion in this section we propose the following hypotheses:
H1: Underwriting brokers will exhibit unusual selling behavior during the pricing period.
H2: Underwriting a DRIP will lead to lower prices and consequent negative abnormal
returns during the pricing period.
3 Data and Method
3.1 Data
The data identifying DRIP announcements is provided by the Securities Industry
Research Centre of Asia-Pacific (SIRCA). We identify 2771 DRIP announcements, 126 of
which are underwritten, between January, 2007 and December, 2011. For each announcementwe collect the date, dividend type, ex-dividend date, record date, and payment date as well as
the ASX stock code and the GICS industry sector classification. The DRIP prospectus and
ASX Appendix 3B documents are used to determine share allotments (both to participating
shareholders and underwriters), along with the corresponding issue prices and the identity of
the underwriting lead manager.8 Details of the pricing period start and end dates are also
obtained from these documents. Stock price and the market index (All Ordinaries) data are
also supplied by SIRCA. Order level data is obtained from SIRCAs Australian equities
database which contains all orders and trade executions submitted in the Australian equity
market. For each order, this data set contains ASX stock codes, times, dates, volume, prices
and the broker identification codes of both the buyer and the seller. There are 93 unique
brokers trading in the UDRIP and DRIP stocks during the sample period. We remove
UDRIPs where we cannot identify the underwriting broker, or which have price sensitive
announcements during the pricing period.9Of our initial sample of 126 UDRIPs, 39 UDRIPs
are removed leaving us with a final sample of 87 UDRIPs.
3.2 Matched Sample Construction
To identify the impact of underwriting a DRIP, UDRIPs are matched to comparable DRIPs.
Matched DRIPs are selected according to Equation (2), which gives a scaled sum of squared
differences between pairs of DRIP and UDRIP firms, across the market capitalization of the
firm and the size of the issue.
8Appendix 3B documents are necessary whenever new shares are issued on the ASX and identify the number,
price and reason for the new issue.9 These include 11 operational results, 9 DRIPs with an unidentifiable underwriting broker, 8 M&Aannouncements, 6 earnings updates, 3 credit rating changes and 2 asset sales.
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
8/28
8
where, and denote the firm market capitalization and issue size for DRIP andUDRIP firms, respectively.In the matching process we ensure that during the pricing period the DRIP does not have
any price sensitive announcements. We select the DRIP and UDRIP pairs with the lowest
matching score within four months of the UDRIP (Time-Match). For robustness testing, we
create a second set of matched firms, based on the lowest matching score within the same
industry, whilst relaxing the contemporaneous time period constraints (Industry-Match). This
generates a new set of matched firms whose fundamental characteristics more closely
resemble their UDRIP counterparts, but which may be drawn from different time periods
within the sample.
The summary statistics for UDRIPs and time-matched and industry-matched DRIPs are
shown in Table 1. Panel A groups UDRIPs by year. The financial crisis of 2008 resulted in a
significant increase in UDRIPs, both by number and dollar value. This reflected the greater
demand for funding certainty during difficult market conditions. As the economic conditionsimproved, the number of UDRIPs declined. While a similar pattern is observed in the
matched DRIPs depicted in Panel B, it is evident that the UDRIPs and the matched DRIP
samples do not have identical numbers of observations by year. This is because the four
month matching period for the time-matched DRIPS spans the year end. Comparing the
equity capital raised across the samples the medians are reasonably similar, but due to large
bank UDRIPs in 2007 and 2008 the means show some substantial differences.
< Insert Table 1 here >
3.3 Broker Trading Behavior
As the underwriting broker is identified in the disclosure documents, we can identify all
trades made under the underwriting brokers ID. Overall volume for underwriting brokers is
higher than that of unaffiliated brokers trading the same UDRIP stock. This is not surprising.
There are fewer small brokers acting as underwriters. Underwriting brokers are generally
larger in size and command greater market share. We account for the size difference by usingeach broker as their own control in constructing trading metrics.
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
9/28
9
Knowing the identity of the broker on the buy and sell side of every trade allows us to
identify the purchasing and selling behavior of all brokers. Trades in UDRIP stocks and the
matched DRIP stocks were analyzed across trading windows, before, during and after the
pricing period. The first day of the pricing period is defined as day 0 and our analysis focuses
on the pricing period window [0, End], where Enddenotes the end of the pricing period. The
returns during the pricing period are then compared to a 5-day and 10-day pre-pricing and
post-pricing period ([-5, -1], [End+1, +5] and [-10, -1], [End+1, +10]). We note that, in the
windows [-5, -1] and [-10, -1], the cause of trading volume should be interpreted with
caution.10Short term trading about the ex-dividend date by both dividend capture traders and
dividend avoidance traders may substantially affect the volumes observed.
Two volume metrics are used to analyze the extent of abnormal trading. The first metric
developed by Chordia, Roll and Subrahmanyam (2002) is used to measure the imbalance
between buying and selling orders that become trades. The second metric is an abnormal
volume metric which is used to measure abnormal volumes separated by whether the broker
acts as the buyer or seller.
Following Chordia et al (2002) the order imbalance metric for each broker is computed as
follows:
() where j indexes the broker and t indexes the trading day, the S and B superscripts represent
seller and buyer respectively. A metric greater (less) than one indicates excess sales
(purchases) made by a broker on a particular day while zero implies that order are in balance.
Order imbalance metrics for each day are then averaged across all brokers for the UDRIP
sample and for the matched DRIP samples, and then further averaged across the trading
window.
Following Henry and Koski (2010), the abnormal volume metric is measured as follows:
where Volumej,tis the total abnormal buying/selling volume of brokerj on each day t during
the event period and Average Volumej is the average buying and selling volume of each10
In 69 out of the 87 UDRIPs the pricing period starts within 10 days of the ex-dividend date.
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
10/28
10
broker j during a clean period measured between 60 and 10 days prior to the start of the
pricing period ([-60, -11]). The abnormal volume metrics are then averaged across all brokers
in each category for each event-day, and then further averaged across the trading window, in
the same way as the order imbalance metric.
3.4
Calculation of abnormal returns
A standard event study method is used to examine the share price response to UDRIPs
during the pricing period. The event windows are the same as those used for the analysis of
volume, [-10, -1], [-5, -1], [0, End], [End+1, 5] and [End+1, 10]. The daily abnormal returns are determined from the market model as follows: [] where is the observed return for security on day and []is the market model return
for security on day , with betas constructed over the period [-180,-11]. Cumulativeabnormal returns are calculated as follows:
where is the CAR for firm over period and mand nare thestarting and ending days of the event window, respectively. These CARs are then averagedfor across firms for each event day, and then averaged again across days in the window of
interest. As a robustness test, and following the Australian DRIP studies of Chan et al. (1993,
1996), we also employ the zero-one market- model, where [] is equal to the marketreturn on day t.
3.5
Regression analysis
To analyze the differences between the returns of UDRIP and DRIP samples in the
pricing period we use the following regression:
whereiis a firm subscript. UDRIPindicates whether the dividend is underwritten and takes a
value of one if the DRIP is underwritten and zero otherwise. We also utilize an alternative
specification for the regression in which an interaction variable U_Sfall is substituted forUDRIP. U_Sfall is the product of the UDRIPdummy and the percentage of shares taken up
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
11/28
11
by the underwriter. The other four variables, measured one month prior to the start of the
pricing period, control for firm-specific factors. is the natural logarithm of themarket capitalization of the firm, is the dividend yield and is a measure of therelative size of the issue.
is the average daily traded value of the stock.
is
the size of the percentage discount applied to the VWAP in order to determine the issue price.
Table 2 reports the descriptive statistics of the DRIP plan structure and the firm
characteristics that are used as control variables. On average, UDRIP plans exhibit a longer
plan pricing period than both time-matched and industry-matched DRIPs. Table 2 also shows
that the participation rate for UDRIPs is lower than for both groups of matched DRIPs. This
is consistent with the literature on rights issues, which shows that rights issues are more likely
to be underwritten when the expected take-up in the offer is low (Bhren, Eckbo and
Michalsen, 1997). The dividend yield is slightly lower for UDRIPs, while the discount
applied to the new shares issued under the UDRIPS and DRIPs is similar. Indeed the median
discounts are identical across all samples at 2.5%.
< Insert Table 2 here >
4 Results
4.1
Broker Trading Behavior
Figure 1 plots Chordia et. als (2002) order imbalance, for the underwriting brokers, the
unaffiliated brokers and brokers in the matched DRIPS. Since the length of the pricing period
varies across firms, we present the order imbalance of each broker group by aligning the
metric by both the start (Panel A) and end (Panel B) of the pricing period. Panel A starts at
day -10 so that it does not overlap with the benchmark period and symmetrically ends at day
+10. Panel B can extend back to day -20 without overlapping the benchmark period and
extends symmetrically to day +20.
Panel A shows that sell orders by underwriting brokers jump substantially during the
pricing period. In contrast, for the unaffiliated brokers and the brokers in the matched DRIP
samples the order imbalance fluctuates around zero during the pricing period. Panel B
demonstrates that after the conclusion of the pricing period the order imbalance for the
underwriting brokers falls sharply towards zero, while no substantive order imbalance
changes are observed in the control samples.
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
12/28
12
< Insert Figure 1 here >
Table 3 provides statistics for the order imbalance. The striking and strongly significant
result is for the underwriting brokers during the pricing period. The pricing period shows a
sharp increase in selling orders by the underwriting brokers with an average sell order
imbalance of 28% of total orders. In contrast, the unaffiliated brokers and the brokers for the
time-matched DRIPs have much smaller, but significant order imbalances on the buy side
during the pricing period and no significant results at other times.
< Insert Table 3 here >
Table 4 provide the results of both a parametric and a non-parametric test of differences
between order imbalance measures of underwriting brokers, unaffiliated brokers and matchedDRIP brokers over the pricing periods. Pairwise comparisons show that the only significant
differences, between the order imbalances for the underwriting brokers and for the other
broker groups, occur in the pricing period. In all cases the underwriting broker is doing
significantly more selling.
< Insert Table 4 here >
4.2
Abnormal Buying and Selling
Figure 2 plots the daily abnormal selling activity for each broker group. We measure the
abnormal volume from both the start (Panel A) and end (Panel B) of the pricing period. From
both panels it is apparent that there is a marked difference between the selling behavior of
underwriting brokers and unaffiliated or DRIP brokers. The non-underwriting brokers do not
exhibit much evidence of unusual selling behavior prior to, during, or post the pricing period.
Underwriting brokers, however, exhibit abnormally high levels of selling during the pricing
period. This abnormal selling jumps to between 300%-400% above the average daily cleanperiod selling volumes at the start of the pricing period, remains elevated for the duration of
the pricing period and then drops markedly about the end of the pricing period. While the
abnormal selling by underwriting brokers is less intense after the end of the pricing period, it
is evident that some abnormal selling is continuing. The rise in abnormal selling in Panel B of
Figure 2, starting about day 15, corresponds to the share allotment dates which typically
occur 15-20 days following the conclusion of the pricing period.
< Insert Figure 2 here >
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
13/28
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
14/28
14
Over the pricing period, Figure 4 shows that both UDRIP stocks and DRIP stocks have
CARs that become negative about the start of the pricing period. However, it is clear that
during the pricing period the UDRIP stocks have a more strongly negative CAR until about
day 10. From about day 10 to day 15 the CAR for the UDRIPs reverses its downward trend
and continues upwards thereafter. Ten days is the median length of the UDRIP pricing period
and there is a batch of UDRIP pricing periods that finish after fifteen days. By the day fifteen
96% of the UDRIP pricing periods are completed. The CAR plot is therefore consistent with
a reversal of price pressure as UDRIP pricing periods conclude.
< Insert Figure 4 here >
Table 6 shows the abnormal CARs over five intervals: [-10, -1], [-5, -1], [0, End],
[End+1, 5] and [End+1, 10], where 0 denotes the start and Enddenotes the end of the pricing
period. The CARs in the windows before and after the pricing period are not significantly
different from zero, neither are they significantly different between the UDRIP and the time-
matched and industry-matched DRIP samples. However, during the pricing period both the
UDRIP and the DRIP samples have significant negative CARs. The UDRIP has a mean
(median) CAR of -4.02% (-2.17%) while the time-matched and industry-matched DRIP have
a mean (median) CAR of -2.26% (-1.96%) and 1.02% (1.34%), respectively. The UDRIPs
mean CAR is significantly more negative than the matched DRIPs mean CARs at the 1%
level. While the median is more negative for the UDRIPs than for the matched DRIPs, the
differences are not significant.
< Insert Table 6 here >
4.4
Cross-sectional regression analysis
Cross-sectional regressions are used to examine the impact of underwriting on prices
during the pricing period, while controlling for various firm-specific variables. The CAR in
the first five days of the pricing period is the dependent variable. We chose five days as all
the CARS in the regression should be measured over the same period and 5 days is the
shortest pricing period present in the sample. The regression results are summarized in Table
7. Columns 1 to 4 measure the market impact of underwritten DRIPs using the UDRIP
dummy. In columns 1 and 2 we report the results for sample that includes UDRIPs and a set
of DRIPs matched by firm size, issue size and the time of the issue (Time). We consider
CARs measured using both the market model (column 1) and the zero-one (market-adjusted)
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
15/28
15
model (column 2) as benchmarks for expected returns. In columns 3 and 4 we repeat the
analysis substituting a set of DRIPs matched by firm size, issue size and the industry of the
firm (Industry). In columns 5 through 8 we delete the UDRIP dummy and instead use an
interaction between the UDRIP dummy and the proportion of the DRIP that was taken up by
the underwriter due to a subscription shortfall. This variable is labeled .The effect of the UDRIP variable is negative across all model specifications and
statistically and economically significant. Underwritten plans experience pricing period
returns which are approximately 2.3% lower than non-underwritten plans after controlling for
differences in firm size, dividend yield, the volume of shares traded during the pricing period
and the discount associated with the DRIP. The results are robust to using DRIPs matched by
time or industry and to using different abnormal return benchmarks.
< Insert Table 7 here >
Columns 5 through 8 in Table 7 show that the effect is also consistently negativeand statistically and economically significant. Given that the mean level of underwriter
participation is 61%, the results imply that UDRIP pricing period returns are, on average,
approximately 2.6% lower than their DRIP counterparts. The evidence from the regression
models clearly indicates that choosing to underwrite a DRIP leads to significant negativeabnormal returns during the pricing period, even after controlling for other potential causes of
price movements.
With respect to the control variables, the effect of , reflecting relatively largerissues, is generally negative and significant, but the effect is more strongly significant in the
regression specifications containing the variable. The effect of, ln(Size), is positiveand significant across all specifications, although the effect weakens in the regression with
the variable. Consistent with increased trading depressing price, the variable, has a coefficient that is negative and significant across the majority ofspecifications, indicating that stocks with high trading in the pricing period experience more
negative returns. The variable , has an insignificant effect across all specifications.5 Conclusion
We hypothesize that there are incentives for underwriter trading that depresses prices over
the pricing period for underwritten DRIPs. The empirical results show both abnormal sellingby the underwriting broker and abnormal price movements during the period in which the
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
16/28
16
pricing of the new shares is determined. Over the pricing period the daily volume of sales
made by the underwriting broker increased by between 200% and 400% relative to trading by
the underwriting broker in the benchmark period. In contrast, there is no significant abnormal
selling by non-underwriting brokers during the pricing period. Furthermore, there is no
abnormal selling in the pricing period for matched samples of DRIPs that are not
underwritten.
Both regression analysis and comparison of CARs between underwritten DRIPs and a
matched sample of DRIPS that were not underwritten, suggests that underwriting results in
significantly more negative returns during the pricing period. On average returns for
underwritten DRIPs are about 2% more negative.
It cannot be conclusively determined whether the observed trading behavior is motivated
by a desire to manipulate the issue price downward, or by a desire to hedge the price risk
arising from the underwriting commitment, or both. Whatever the motivation it serves the
interest of the underwriters, adds to selling pressure and depresses prices during the pricing
period, which consequently depresses the issue price. The result is less capital for the firms
and a wealth transfer to the underwriters. We suggest that firms stem the transfer of wealth
from non-participating shareholders to the underwriter by either selecting a pricing period
that is less susceptible to price manipulation, or by inserting a clause into the underwriting
agreement to restrict the trading activity of the underwriter.
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
17/28
17
References
Aggarwal, R., 2003. Allocation of initial public offerings and flipping activity. Journal of Financial Economics,
68(1), pp.111135.
Aggarwal, R.K. & Wu, G., 2006. Stock Market Manipulations. The Journal of Business, 79(4), pp.19151953.
Bhren, ., Eckbo, B.E. & Michalsen, D., 1997. Why underwrite rights offerings? Some new evidence.Journal of
Financial Economics, 46(2), pp.223261.
Chan, K.K.W., Mccolough, D.W. & Skully, M.T., 1996. Australian dividend reinvestment plans: An event study ondiscount rates.Applied Financial Economics, 6(6), pp.551561.
Chan, K.K.W., McColough, D.W. & Skully, M.T., 1993. Australian Tax Changes and Dividend ReinvestmentAnnouncement Effects: A Pre- and Post-Imputation Study.Australian Journal of Management, 18(1), pp.4162.
Chauvin, K.W. & Shenoy, C., 2001. Stock price decreases prior to executive stock option grants.Journal of
Corporate Finance, 7(1), pp.5376.
Chordia, T., Roll, R. & Subrahmanyam, A., 2002. Order imbalance, liquidity, and market returns.Journal ofFinancial Economics, 65(1), pp.111130.
Cohen, D.A. & Zarowin, P., 2010. Accrual-based and real earnings management activities around seasoned equityofferings.Journal of Accounting and Economics, 50(1), pp.219.
Eades, K. M., P. J. Hess, and E. H. Kim, 1984, On interpreting security returns during the ex-dividend period,Journalof Financial Economics13, 3-34.
Eckbo, B.E. & Masulis, R.W., 1992. Adverse selection and the rights offer paradox.Journal of Financial Economics,32(3), pp.293332.
Ellis, K., Michaely, R. & OHara, M., 2000. When the Underwriter Is the Market Maker: An Examination of Tradingin the IPO Aftermarket. The Journal of Finance, 55(3), pp.10391074.
Fishe, R.P.H., 2002. How Stock Flippers Affect IPO Pricing and Stabilization. Journal of Financial and QuantitativeAnalysis, 37(02), pp.319340.
Henry, T.R. & Koski, J.L., 2010. Short Selling Around Seasoned Equity Offerings.Review of Financial Studies,23(12), pp.43894418.
Jenkinson, T. & Jones, H., 2007. The Economics of IPO Stabilisation, Syndicates and Naked Shorts.EuropeanFinancial Management, 13(4), pp.616642.
Lin, J.-C. & Wu, Y., 2013. SEO timing and liquidity risk. Journal of Corporate Finance, 19, pp.95118.
Marciukaityte, D. & Varma, R., 2008. Consequences of overvalued equity: Evidence from earnings manipulation.Journal of Corporate Finance, 14(4), pp.418430.
Rangan, S., 1998. Earnings management and the performance of seasoned equity offerings.Journal of FinancialEconomics, 50(1), pp.101122.
Teoh, S.H., Welch, I. & Wong, T.J., 1998. Earnings management and the underperformance of seasoned equityofferings.Journal of Financial Economics, 50(1), pp.6399.
Yermack, D., 1997. Good Timing: CEO Stock Option Awards and Company News Announcements. The Journal ofFinance, 52(2), pp.449476.
Zhang, D., 2004. Why Do IPO Underwriters Allocate Extra Shares when They Expect to Buy Them Back?Journal of
Financial and Quantitative Analysis, 39(03), pp.571594.
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
18/28
18
Table 1
Summary of UDRIPs from 2007 to 2011
Panel A provides an overview of the characteristics of UDRIPs from January 2007 to December 2011summarized by year. Panel B and Panel C describe the DRIP sample matched by time and industry,respectively.Market Cap refers to the average market capitalization of companies in the sample measured onemonth prior to the start of the pricing period. Equity Capital Raised is the amount of capital raised by theDRIPs. All percentages are rounded to the nearest percent.
Panel A: UDRIP Plan Distribution by Year
Year FrequencyPercentage
(%)
Market Cap Equity Capital Raised ($ 000s)
($ 000s) Mean Median Total
2007 16 18% 9,349 199,085 29,644 3,185,366
2008 30 34% 9,612 225,759 56,581 6,772,759
2009 18 21% 2,719 49,666 18,912 893,9812010 9 10% 419 8,178 5,779 73,602
2011 14 16% 8,588 140,441 25,788 1,966,180
Sample 87 100% 7,022 148,183 27,936 12,891,888
Panel B: Time Matched DRIP Plan Distribution by Year
Year FrequencyPercentage
(%)
Market Cap Equity Capital Raised ($ 000s)
($ 000s) Mean Median Total
2007 17 20% 5,071 35,939 20,025 610,965
2008 26 30% 7,603 66,210 33,585 1,721,4622009 18 21% 6,752 75,994 17,379 1,367,894
2010 15 17% 8,766 71,541 7,423 1,073,108
2011 11 13% 8,471 62,038 17,160 682,419
Sample 87 100% 7,243 62,711 21,161 5,455,848
Panel C: Industry Matched DRIP Plan Distribution by Year
Year FrequencyPercentage
(%)
Market Cap Equity Capital Raised ($ 000s)
($ 000s) Mean Median Total
2007 13 15% 1,708 14,631 4,122 131,678
2008 17 20% 6,393 47,731 24,508 811,434
2009 18 21% 7,257 80,375 10,572 1,446,750
2010 26 30% 6,519 55,930 16,135 1,454,181
2011 13 15% 14,277 174,609 36,493 1,920,697
Sample 87 100% 7,175 71,170 18,497 5,764,740
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
19/28
19
Table 2
Descriptive Statistics
This table gives descriptive statistics for the UDRIP/DRIP characteristics and the firm characteristics used as control variables, partitioned across the UDRIP sample andmatched DRIP control samples. Each sample includes 87 observations.Pricing Periodis the number of days in the period used to determine the plan issue price. Underwritertake-upis the percentage of the DRIP shares being offered that are subscribed for by the underwriter. Participationis the percentage of shares participating in the DRIP.While in most cases the underwriter take-up plus the participation sums to 1, if the underwriter does not underwrite 100% of the issue the sum could be less than 1. Dividendyieldis the dividend per share divided by the closing price for the stock one month prior to the start of the pricing period. Discountis the size of the discount applied to newshares issued under the DRIP and is applied to the VWAP during the pricing period. Sizemeasures the market capitalization of each firm one month prior to the DRIPannouncement. Ln(Size)is the natural logarithm of the Size variable. Traded valueis the average daily traded value for each stock during the pricing period. Ln(TradedValue)is the natural logarithm of the traded value variable.
UDRIP DRIP Time DRIP Industry
Mean Median Std. Dev. Mean Median Std. Dev. Mean Median Std. Dev.
Pricing period (days) 9.49 10.00 3.96 7.76 8 2.96 8.25 9 3.42
Underwriter takeup(%) 60.98 61.88 17.13 - - - - - -
Participation (%) 32.92 30.15 14.96 40.91 34.20 20.37 39.31 37.11 18.16
Dividend yield (%) 3.06 2.30 2.61 4.00 3.59 2.07 3.70 3.33 1.71
Discount (%) 2.82 2.50 1.51 2.73 2.50 2.07 2.91 2.50 2.52
Size ($m) 7021.61 1298.30 12456.57 7242.63 1405.95 14081.00 7175.20 1411.80 13798.05 7.16 7.17 2.16 7.28 7.25 2.04 7.15 7.25 2.18Traded value ($m) 1.75 0.80 2.21 2.10 0.85 4.22 1.91 1.35 2.01 12.22 13.47 3.83 12.80 13.52 3.17 12.11 13.89 4.53
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
20/28
20
Table 3
Order Imbalances
This table gives the order imbalance metric over the periods [-10, -1], [-5, -1], [0, End], [End +1, +5] and [End+1, +10] for underwriting brokers, unaffiliated brokers and matched DRIP brokers. The daily order imbalancemetric per broker is calculated as the difference between sell volume and buy volume divided by the sum of buyand sell volume. The metric is then averaged across all brokers in each category for each event-day, and thenfurther averaged across the trading window. A measure of 0 indicates that there is no order imbalance. Ameasure greater than 0 indicates abnormal selling whilst a measure less than 0 indicates abnormal buying. ***,** and * represent significance at the 1%, 5% and 10% levels, respectively.
UnderwritingBrokers
Unaffiliated BrokersDRIP Brokers(Time-Match)
DRIP Brokers(Industry-Match)
[-10, -1] 0.012 0.013 -0.005 -0.006
[-5, -1] -0.038 0.006 -0.012 -0.018
[0, End] 0.288*** -0.051** -0.009* -0.008
[End +1, +5] 0.049 0.004 -0.021 0.006
[End +1, +10] 0.011 0.017 -0.009 0.013
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
21/28
21
Table 4
Order Imbalances between Groups Pairwise Comparisons
This table gives the results of pairwise tests of differences between order imbalance measures of underwritingbrokers, unaffiliated brokers and matched DRIP brokers over the pricing periods [-10, 1], [-5, -1], [0,End],[End+1, +5] and [End+1, +10]. ^^^ (###) represents statistical significance at the 1%, ^^ (##) at the 5% and ^
(#) at the 10% level for the paired student tand the Wilcoxon matched pairs signed ranks test.
Underwriting vs.Unaffiliated Brokers
Underwriting vs. DRIPBrokers (Time-Match)
Underwriting vs. DRIPBrokers (Industry-Match)
[-10, -1] -0.001 0.017 0.018
[-5, -1] -0.044 -0.026 -0.02
[0, End] [End+1, +5] 0.045 0.07 0.043
[End+1, +10] -0.006 0.02 -0.002
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
22/28
22
Table 5
Broker Inter-day Trading Activity
This table gives the abnormal volume over the periods [-10, -1], [-5, -1], [0, End], [End+1, 5] and [End+1, 10],where 0 denotes the start and Enddenotes the end of the pricing period for buy and sell volumes across brokergroups. Abnormal volume is measured as the ratio of trades by each broker each day to average daily trades by
the same broker in a benchmark period. One is then subtracted from this ratio and the resulting metric (%abnormal volume) is then averaged across brokers and the event period. Panel A gives abnormal volumes for the
broker underwriting the UDRIP. Panel B gives abnormal volumes for unaffiliated brokers. Panel C and D aretrades in DRIP stocks by all brokers matched by time and industry, respectively. Total is the total abnormalvolume for both buy and sell trades. SalesandPurchases gives abnormal volume conditioned on whether the
broker is selling or buying. A value of 0 implies no abnormal volume. ***, ** and * represent significance at the1%, 5% and 10% levels, respectively.
Panel A: Underwriting Broker Volume during Pricing Periods
Total Sales Purchases
[-10, -1] 39%*** 57%*** 82%***
[-5, -1] 15% 36%** 52%[0, End] 138%*** 236%*** 95%***
[End+1, +5] 45% 66%* 57%**
[End+1, +10] 24% 32%* 49%**
Panel B: Unaffiliated Broker Volume during Pricing Periods
Total Sales Purchases
[-10, -1] 21%** 17%*** 13%*
[-5, -1] 5% 5% -2%
[0, End] 18% 12% 23%*
[End+1, +5] 7% 5% 5%
[End+1, +10] 4% 6% 0%
Panel C: DRIP Broker (matched by time) Volume during Pricing Periods
Total Sales Purchases
[-10, -1] 5% 4% 6%
[-5, -1] 5% 0% 9%
[0, End] 1% -2% 2%
[End+1, +5] -2% 3% -6%
[End+1, +10] -2% 3% -7%
Panel D: DRIP Broker (matched by industry) Volume during Pricing Periods
Total Sales Purchases
[-10, -1] 15%** 15%*** 10%
[-5, -1] 15%* 10% 13%
[0, End] 7% 3% 7%
[End+1, +5] 5% 7% 1%
[End+1, +10] 7% 7% 3%
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
23/28
23
Table 6
Pricing Period CARs
This table gives the CARs from the pricing periods for the UDRIP, time-matched and industry-matched DRIP firms for five event windows [-10, -1], [-5, -1], [0, End],[End+1, 5] and [End+1, 10], where 0 denotes the start and Enddenotes the end of the pricing period. CARs are based on the market model as a benchmark return. The CARsare calculated as the average across all firms of the sum of daily abnormal returns, starting at the beginning of each window. Differences in means and medians depict thedifference between the UDRIP and the respective DRIP matched pairs. ***, ** and * represent statistical significance at the 1%, 5% and 10% levels. Test for differencesbetween samples are based on the paired student t(mean) and Wilcoxon signed rank test (median).
UDRIP Time-matched DRIP Industry-matched DRIP
Mean (%) Median (%) Mean (%) Median (%) Difference inmeans Differencein medians Mean (%) Median (%) Differencein means Differencein medians
[-10, -1] -0.322 -0.509 -0.887 -0.776 0.565 0.267 0.573 -0.232 -0.895 -0.277
[-5, -1] -0.076 -1.037 -0.396 -0.87 0.32 -0.167 0.024 -0.289 -0.1 -0.748
[0, End] -4.024*** -2.174*** -2.255*** -1.959*** -1.769** -0.215 -1.018 -1.336** -3.006** -0.838
[End +1, +5] -0.253 -0.138 -0.559 0.198 0.306 -0.336 -0.095 0.281 -0.158 -0.419
[End +1, +10] -0.259 -0.085 -0.562 0.028 0.303 -0.113 0.369 0.401 -0.628 -0.486
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
24/28
24
Table 7
Pricing Period Regressions
This table reports the cross-sectional regression results for abnormal returns during the pricing period. The Timeheading indicates regressions with UDRIP firms matched with DRIPs within a four month window surroundingthe UDRIP, as well as firm size and issue size. The Industry heading indicates firms matched based on industry,
firm size and issue size. Market and Zero-one headings denote the use of the market model and the zero-one(market-adjusted) model, respectively. The dependent variable is the average CAR for the interval [0, +5];is a dummy variable that equals one if a DRIP is underwritten and zero otherwise (model 1through 4).is the product of the UDRIP dummy and the percentage of shares taken up by the underwriter (model 5through 8). is the natural logarithm of the market capitalization of each firm. is the dividendyield calculated as the dividend per share divided by the share price one month prior to the start of the pricing
period. is the natural logarithm of average daily turnover for each stock during the pricing period. is the discount rate for each plan for firm . Heteroskedasticity consistent standard errors are used.***, **, * represents significance at the 1%, 5% and 10% level, respectively.
Pricing Period CAR with UDRIP Dummy Pricing Period CAR with Underwritten Shortfall
Interaction
Sample Time Industry Time Industry
Model Market Zero-one Market Zero-one Market Zero-one Market Zero-one
(1) (2) (3) (4) (5) (6) (7) (8)
0.036 0.040 0.053 0.048 0.042 0.047 0.062 0.062
-0.023** -0.026** -0.022* -0.023* -0.040** -0.046*** -0.040** -0.045** -0.005* -0.005* -0.004 -0.006* -0.007** -0.007*** -0.006* -0.008**
0.012** 0.012** 0.015** 0.013** 0.009* 0.009* 0.012** 0.010*
-0.009** -0.009** -0.013*** -0.011** -0.007 -0.007 -0.011** -0.009* -0.144 -0.101 0.058 0.027 -0.316 -0.270 -0.166 -0.209
3.49*** 3.70*** 2.83** 2.75** 3.65*** 3.96*** 2.75** 2.93*** 0.089 0.096 0.081 0.077 0.097 0.107 0.080 0.087
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
25/28
25
Figure 1
Broker Order Imbalance
Panel A: Inter-day Order Imbalance aligned by the start of the pricing period
Panel B: Inter-day Order Imbalance aligned by the end of the pricing period
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
26/28
26
Figure 2
Panel A: Inter-day Abnormal Selling aligned by the start of the pricing period
Panel B: Inter-day Abnormal Selling aligned by the end of the pricing period
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
27/28
8/10/2019 How Do Underwriters Trade_CMCRC Research Report
28/28
Figure 4
Pricing Period CARs over [-20, +20]