SSRN-id172633 by blue123

VIEWS: 11 PAGES: 53

									        LONG RUN UNDERPERFORMANCE OF SEASONED EQUITY
                                        OFFERINGS:
                                FACT OR AN ILLUSION?



                          Victor SOUCIK1 and David E. ALLEN2



                                  JEL Classification Code
                                           G12
                        (General Financial Markets: Asset Pricing)




1
  Mr V. Soucik, Edith Cowan University, Joondalup Drive, Perth, Western Australia 6027; Phone: (61-
8) 9400 5470; Fax: (61-8) 9400 5633; E-Mail: v.soucik@cowan.edu.au
2
  Professor D.E. Allen, Edith Cowan University, Joondalup Drive, Perth, Western Australia 6027;
Phone: (61-8) 9400 5471; Fax: (61-8) 9400 5633; E-Mail: d.allen@cowan.edu.au
                                              ABSTRACT



This paper uses Australian data to show that the long run underperformance of seasoned equity

offerings is related to the definition of ‘long-run’. We demonstrate that following the period delimited

by other writers as the long run, issuing firms turn around in their performance and in fact outperform

their corresponding benchmarks, sometimes more than making up for the initial losses. We show that

the initial underperformance affects the issues of companies performing more than one SEO in a

similar fashion. Our results demonstrate that a poor performance following an SEO has, to an extent, a

specific role as the mitigator of costs associated with the issue.
1.     INTRODUCTION

Are new equity issues really a good investment, or is it more rewarding to invest in alternative assets?

The enthusiasm among investors for initial public offerings (IPOs) and many seasoned equity offerings

(SEOs) has been well proven as seen, for example, in a significant oversubscription of almost all initial

company floats in 1997 (Bradley, 1998). But do these investments provide a good return opportunity

beyond the initial gain? In other words, are they a good long-term investment? This is the question we

will attempt to address in some detail inside this paper.



Although the underperformance of initial public offerings has been studied vigorously for a number of

decades, the finding that seasoned equity offerings are also poor “long-run” investments is relatively

new. Early writings can be traced to 1960s in the researches by Stigler (1964) and Friend and

Longstreet (1967), but it wasn’t until mid 1980s that the issue of seasoned equity offerings was

seriously revisited in studies by Masulis and Korwar (1986) and Asquith and Mullins (1986) both

documenting a significant share underperformance of companies who have just conducted a new equity

issue. Masulis and Korwar (1986) in fact observed highly negative returns for 50% of industrial and

32% of public utility stocks in the same time, when the market recorded a significantly positive return.

This was consistent with findings by Mikkleson and Partch (1986) and Schipper and Smith (1986).

However, none of these researchers provided a comprehensive theoretical explanation for their results,

instead explaining the findings as an “empirical phenomenon” (Lim, 1986).



The landmark study into the performance of issuing companies was conducted in 1995 by Loughran

and Ritter (1995) building on foundations laid by Healy and Palepu (1990), Ritter (1991) and

Loughran, Ritter and Rydqvist (1994) into IPOs. They affirmed the original findings by Masulis and

Korwar (1986) observing a 15.7% and 33.4% five-year holding period returns for IPOs and SEOs

during time when the returns on non-issuing firms matched with the issuers by capitalisation were

66.4% and 92.8%, respectively. This finding was also supported by others including Spiess and

Affleck-Graves (1995) who observed the median return for SEO firms to be only 10%, compared with

a 42.3% median return for non-issuers matched by size. Loughran and Ritter (1995) extended this by

reporting that no significant under- performance was found in the first 6 months following an issue, but

a critical period of 18 months ensued during which much of the discrepancy occurred. Jagadeesh,




                                                                                                     Page 1
Weinstein and Welch (1993) also observed that the performance discrepancy in fourth and fifth year

reduces down to only 3% and 1%, respectively. This was consistent with findings by Loughran and

Ritter (1997) who also noted no significant differential during year six and seven. All authors further

confirmed that the observed underperformance is not only significant relative to matched non-issuers,

but also relative to the issuers’ own past performances.



Another startling observation was that of a substantial overperformance by the issuing companies

leading up to the day of issue observed by many writers including Healy and Palepu (1990), Loughran

and Ritter (1995, 1997), Patel, Emery and Lee (1993) and others. Gerard and Nanda (1993) refined this

finding by showing that only 38% of the lead-up performance could be explained by market-wide or

industry-specific factors.



The long-term underperformance of SEOs was not, however, confined just to share-prices but was also

significantly reflected in lower operating performance. Loughran and Ritter (1997), elaborating on their

previous work, observed a 23% and 40% drop in operating income-to-assets and market-to-book ratios,

respectively, and a profit margin which less than halved over the four-year period following an issue.

These results, further confirmed in cross-sectional comparisons with non-issuers, were also consistent

with Spiess and Affleck-Graves (1995) who, in addition, controlled for differences in trading-system,

offer size and firm age. McLaughlin, Safieddine and Vasudevan (1996) re-examined the issue

concentrating on cash flows (found to decline by over 20%) and detected a greater overall performance

drop in companies having larger amounts of free cash.



Collectively these findings have significant economic impact on the investment choices. Loughran and

Ritter (1995:23), for example, concluded “an investor would have had to invest 44 percent more money

in the issuers than in non-issuers of the same size to have the same wealth five years after the offering

date”. On the flip side, however, Ritter (1991) points to the importance of this underperformance in

reducing costs incurred by the firm in making the offering, including the opening return gained from

issue underpricing.




                                                                                                      Page 2
Interestingly, in a recent study Eckbo, Masulis and Norli (1999) proffer that there, in fact, may not be

any underperformance since reduced leverage and increased liquidity of issuing firms that result from

seasoned equity offerings should lower the benchmark against which their post-issue performance is

measured. In a subsequent analysis they found no significant abnormal performance in the adjusted

SEO returns. However, their results could be subject to an alternative interpretation that would question

whether the effects of leverage and liquidity are sufficiently pronounced to warrant the lowering of

benchmark enough to erase the abnormal post-issue returns, thus reverting to the contention of

existence of SEO underperformance.



So the relationship between seasoned equity issues and the subsequent drop in operating, financial and

stock performance thus stands well supported. But a question hotly debated among writers is why

should an issuer experience such underperformance.



One explanation proposed by Loughran and Ritter (1997) asserts that when a firm is substantially

overvalued it is likely to issue equity, taking advantage of the opportune time to augment what Myers

(1984) refers to as financial slack. Thus Myers’ (1984) pecking order hypothesis which ranks funding

preferences as internal equity, external debt and then external equity is not static and during the

“windows of opportunity” the ranking can change to external equity, external debt and then internal

equity, causing preference for SEOs rather than debt. The associated operating performance drop was

interpreted through the continuation of investment in perceived positive NPV projects which, all too

often, turned out to be negative. This is fuelled by over-optimism on part of issuing firms’ managers, as

supported empirically by Healy and Palepu (1990), Brous (1992) and Jain (1992) who find that the

post-issue performance forecasts are systematically too high.



An alternative explanation is that of information asymmetries as summarised by Healy and Palepu

(1990). They assume managers act in the best interests of their existing shareholders, and thus an

announcement to issue new shares will be responded to negatively by the outside markets. The reasons

lie in the application of issue proceeds: (1) If proceeds are used to retire debt, DeAngelo and Masulis

(1980) state that firms with higher business risk and more volatile profits are likely to borrow less

because “extra borrowing increases expected financial distress costs more than the expected benefits”.




                                                                                                      Page 3
Masulis (1983) then argues that investors, having imperfect information, will interpret debt reduction

as increased expected risk by the management, and thus downgrade the firm’s share value. (2) If

proceeds are used to fund capital expenditure, Myers and Majluf (1984) assert that this conveys

information about firm’s existing assets. Acknowledging that SEO creates a trade-off between value

from new investments and wealth transfers from old to new shareholders, and given that managers act

in best interests of existing shareholders, Myers and Majluf argue that investors will interpret SEO as

bad news since “managers are more likely to issue new equity when they expect returns from existing

assets to be lower or riskier than anticipated”. (3) If proceeds are used to fund shortfalls in operating

cash flows, Miller and Rock (1985) argue that, given the identity between cash sources and uses, new

equity offer indicates downward revision by management of the forecasted earnings.



Yet another explanation for long term SEO underperformance is that of “managed earnings” proposed

by Teoh, Welch and Wong (1997) and Rangan (1997) who find that firms with highest levels of

discretionary accruals have the worst post-issue performance. Therefore they conclude “some issuers

consciously attempt to manage earnings, raising stock price to artificially increase the offer price for

the new equity” (Loughran and Ritter, 1997:1847).



The final possibility was advanced by Leland and Pyle (1977) through the “signalling model” and was

empirically supported by Downes and Heinkel (1982). They posited that managers have incentives to

hold large stock holdings only if they expect future cash flows to be larger relative to the current price

and hence decreased fractional stock holding as a result of new equity issue gives a negative signal to

the market about the firm value.



Reporting significant underperformance of firms issuing seasoned equity during the first five years

following the offer our results concur with other studies including Loughran and Ritter (1995, 1997)

and Spiess and Affleck-Graves (1995). We will not, however, conclude that SEO firms thus become

poor long-run performers. Instead we will show that such significant long-run underperformance is a

phenomenon consequential to the wrong definition of the ‘long-run’. The results will demonstrate that

although issuers do underperform non-issuers in the initial years, this trend is later significantly

reversed and by the sixth year issuers actually report significant overperformance. The




                                                                                                       Page 4
overperformance then gradually subsides and by eight year cumulative performance of issuers and non-

issuers becomes approximately equal.



The results will also show that the extent of underperformance, especially over a five-year period, is

related to the initial underpricing, as reflected in the Dilution Yield measure of initial returns. This

concurs with Ritter’s (1991) argument that initial underperformance subsequent to the equity offering

has the effect of reducing costs incurred in making the issue, including the initial return.



The rest of this study is subdivided into five sections. We shall review the research objectives in

Section II. Section III describes the methodology and data sources used in the research. We will report

our results in Section IV, followed by Section V in which we conclude and pose some final remarks

and suggestions.



2.     RESEARCH OBJECTIVES

In this paper we intend to re-examine the above findings and attempt to control for a number of

methodological biases which we believe were present in the previous research.



We begin by re-examining whether the issuers in our sample actually do underperform with respect to a

number of benchmarks including non-issuers and the market:

Hypothesis 1a: Firms issuing seasoned equity do not underperform with respect to corresponding

                   non-issuers.

Hypothesis 1b: Firms issuing seasoned equity do not underperform with respect to the market



Next we will conjecture that the observed long-run underperformance is a phenomenon conditioned by

the definition of the ‘long run’, based on Loughran and Ritter’s (1997) observation that the SEO firms

continue to invest significantly more in capital projects than non-issuers in every of the five post-issue

years examined. We will thus hypothesise that it takes more than five years for these capital projects to

come to fruit, after which the issuers will significantly outperform non-issuers, perhaps more than

making up for the initial shortfall. Some preliminary support comes from Loughran and Ritter (1997)

who found that critical underperformance occurs in seventh to twenty-fourth month following an SEO,




                                                                                                      Page 5
after which the performance gap significantly narrows. McLaughlin, Safieddine and Vasudevan (1996)

in fact detected no significant difference in the sixth year following an issue:

Hypothesis 2a: SEO firms do not crossover from period of underperformance to period of

                  overperformance relative to non-issuers.

Hypothesis 2b: SEO firms do not underperform non-issuers, in aggregate, over the real ‘long-run’.



In the third stage we will examine the performance of companies which had more than one seasoned

equity offering. We will hypothesise that if underperformance is, in fact, a consequence of issue, then it

should accompany each offering. Moreover, controlling for changes in company characteristics that

resulted from the previous issue, each underperformance should be qualitatively equal and

quantitatively similar to the previous one:

Hypothesis 3:     Performance characteristics of firms having more than one issue of seasoned equity

                  are not economically and statistically similar for each offering.



The next stage will analyse the impact various extraneous factors have on SEO performance in a belief

that some of these factors may systematically differentiate issuers from non-issuers, hence questioning

whether the return differential is actually a result of issuing. In particular we will control for:

i. Firm Age – A more established firm generally implies greater stability of its costs and revenues, and

       hence lesser ex-ante uncertainty. Consequently a seasoned issue should impact more strongly on

       younger firms raising an expectation of a negative relationship between age and the extent of

       underpricing, implying a positive relationship with abnormal returns. Welch (1989) and Spiess

       and Affleck Graves (1995) observed weak support for this proposition.

ii. Company Beta – As a proxy for company risk, the higher the beta the greater would be the expected

       return as a compensation for the risk, and hence the smaller should be the post-issue

       underperformance. Hence a positive relationship is anticipated.

iii. Market Capitalisation – Just like age, size may also reflect on the ex-ante uncertainty implying a

       positive relationship between firm’s capitalisation and the abnormal returns. However, Allen

       and Patrick (1996) – studying the performance of IPOs – find this factor statistically

       insignificant.




                                                                                                      Page 6
iv. Year of issue – This factor is included to account for the possibility that differences in the economic

       environment surrounding the issue play a role in the amount of underpricing. Specifically, year

       of issue will look at chronological dependence between the underperformance and the time of

       issue. This loosely reflects on the potential existence of a ‘learning curve’ with respect to post-

       issue performance.

v. Volume of SEOs in the issue year – While year of issue is taken more as a ‘raw’ reflection on the

       conditions affecting the offering, volume of SEOs proxies specifically for a particular aspect of

       these conditions – the level of market optimism. “The annual volume of listings should be

       greater in rising markets, reflecting the firm’s desire to list when the market is buoyant” (Allen

       and Patrick, 1996:142). This notion was somewhat supported by empirical studies of Ibbotson

       and Jaffe (1975) and Ritter (1984) who document the existence of a hot issue phenomenon; and

       Loughran, Ritter and Rydqvist (1993) who evidence the impact of business cycles.

These arguments can be summed in following hypotheses:

Hypothesis 4a: The extent of SEO underperformance is not a function of age.

Hypothesis 4b: The extent of SEO underperformance is not a function of beta.

Hypothesis 4c:    The extent of SEO underperformance is not a function of market capitalisation.

Hypothesis 4d: The extent of SEO underperformance is not a function of the chronological attribute

                  of the issue.

Hypothesis 4e:    The extent of SEO underperformance is not a function of volume of seasoned equity

                  offerings in the year of issue.



Finally we shall inspect the relationship between the initial underpricing and subsequent

underperformance of issuers. This is done in an attempt to prove or disprove the argument posited by

Ritter (1991) that immediate decline in performance following an offering has the effect of reducing

the cost to the firm incurred in making the issue. More precisely, we shall first re-confirm whether

there is, in fact, a significantly positive opening return for issuers, and then we will check for the

presence of a negative relationship between initial and subsequent returns:

Hypothesis 5a: SEO firms do not record significant opening returns.

Hypothesis 5b: The extent of SEO underperformance is not a function of the size of initial returns.




                                                                                                       Page 7
3.     RESEARCH METHOD

3.1.   DATA

The raw sample consists of 137 seasoned equity offerings made between January 1984 and October

1993. The period was chosen so as to permit at least five years of price data for each SEO company in

the sample (leading up to 1998). To be included, the identified SEOs must have met following criteria

(which parallel Loughran and Ritter, 1997): (1) the company is listed on the Australian Stock Exchange

and recorded in the DataStream Database at the time of the issue, (2) the offer must be a cash offer for

common stock, (3) the book value of assets at the end of the fiscal year of issuing must be at least $5

million in terms of the 1990 purchasing power and (4) the company undertaking the SEO is not a

financial company or a regulated utility.



Furthermore, during the stages of our study where we examine no more than five years of issuers’

performance we follow Healy and Palepu’s (1990) procedure and remove all issues by the same

company made within five years after the SEO to avoid a period overlap bias. Thus, when a company

made a seasoned equity issue, that company cannot re-enter the sample until five years after the offer

date. This causes a deletion of 35 SEOs from the sample, leaving a total of 102 issues made by 94

companies. Following an argument by Loughran and Ritter (1997:1826) it should also be noted that

because whether or not a firm had previously issued is observable at the time of the issue, we do not

introduce an ex-post sample selection bias. Therefore, if a firm issues seasoned equity without having

done so in the past five years it is included in the sample irrespective of whether or not it has issued

subsequently.



On the other hand, where we study the performance of multiple SEOs we have deleted all companies

from the raw sample that have had only one seasoned equity issue. This lead to the deletion of 60 firms

leaving 77 SEOs in the sample. These represent 32 firms of which 24 made two equity issues, 6 made

three issues and 2 made five issues.



Finally, for the analysis of the long run performance it was necessary to narrow the time frame to

October 1986 instead of 1993. This reduced the sample to 26 SEOs. During this period 5 of the

companies had multiple issues, so further deletions produced a clean sample of 21 firms. It is also




                                                                                                     Page 8
important to note that although each of these companies did have continuous data for the ensuing 12 –

year period, this was not a precondition in the selection process, and as such survivorship bias has not

been introduced.



The original list of seasoned equity offerings was sourced from the DataStream Database and

crosschecked with the listing from Securities Data Company (SDC) database. The offer prices, post-

issue price series, company betas, industry codes and the market values were also obtained from the

DataStream Database. The date of incorporation and the date of listing were obtained from the 1998

Australian Stock Exchange Yearbook and the 1998 Australian Stock Exchange Investor Handbook.

Where these data were unavailable we have examined companies’ prospectuses in an effort to fill the

remaining gaps. Even after this exercise, however, incorporation and listing dates were missing for 7

(5) and 18 (15) issues (firms), respectively causing their deletion from this part of cross-sectional

regressions.



3.2.   METHODOLOGY

3.2.1. Choice of Performance Benchmarks

At first, a benchmark needs to be established against which SEO performance can be measured. Results

are sensitive to the choice of benchmark as demonstrated by Ritter (1991). We have therefore chosen

three such benchmarks – (1) a size-matched sample of non-issuing firms, (2) an industry-and-size-

matched sample of non-issuing firms and (3) a market index benchmark



Our size-matched benchmark is created in several steps. First, in the middle of each issue year (defined

as 30 June), all common stocks listed on the ASX that have not made an issue in the last five years are

ranked according to their market capitalisation. Next, for each issuing firm in the sample a non-issuer is

selected from the list that has capitalisation closest to, but higher than the issuer. If the sample firm has

already the largest capitalisation, then a match with next highest market value is chosen. This then

becomes a size – matched non-issuer benchmark. If the non-issuer becomes delisted before the end date

for the corresponding issuer, a second (and if necessary third, fourth, etc.) matching firm is spliced in

after the delisting date of the first matching firm (based on Loughran and Ritter, 1995). Splicing

procedure follows the matching methodology for the first non-issuing match. If a chosen non-issuer




                                                                                                         Page 9
subsequently issues, we treat it as being delisted at the date of the offer and replace returns from the

issue day onwards with the next match. These re-matching procedures have the benefit of eliminating

any survivorship or look-ahead biases from the benchmark selection process. It should be noted that

unlike many previous studies, which have performed the non-issuer matching on 31st of December, we

have selected 30th June to remove the bias towards greater size-match accuracy for firms issuing close

to the year-end.



The methodology for industry-and-size-matching parallels the above. Non-issuers are first grouped

according to their industry codes assigned by the Australian Stock Exchange and then ranked by their

market capitalisation. For each issuer, a non-issuer with the same industry code and capitalisation

closest to, but higher than the sample firm, was chosen as the matching benchmark. If a match in an

identical industry could not be chosen, the issuer was matched by capitalisation only. Re-matching

following a delisting or an equity issue was again performed as above.



Finally, the market index benchmark was established by pairing the performance of each issuer with

the performance of the All Ordinaries Index over the corresponding time. This enables the evaluation

of SEO’s performance with respect to the market as a whole, as proxied by the index.



3.2.2. Time Series Methodology

3.2.2.1. Time Definitions

For the time series examination of SEO performance, we follow Spiess and Affleck-Graves (1995) and

define each year as consisting of 12 months, each month comprising 21 trading days. Hence annual

anniversaries of SEOs using this trading-day definition may slightly precede their respective calendar

anniversaries.



Initial (or opening) return is calculated over the first trading day on which the seasoned equity was

issued. Post-issue returns are computed during the period following the offer date, ie excluding the first

day. Three separate time frames are defined:




                                                                                                     Page 10
i.      Short term – Defined as 3 years following the offer date. This period is selected to permit

        examination of the “critical underperformance period”, proposed by Loughran and Ritter (1997)

        to occur during months 7 to 24 following the issue; and at the same time allow for comparison

        with other studies that looked at SEO performance over a three-year period.

ii.     Medium term – Defined as 5 years following the offer date. The five-year time frame should

        enable comparison of underpricing with other studies, which took five years to already

        constitute long term.

iii.    Long term – Defined as 12 years following the offer date. This was selected to correct for the

        potentially wrong definition of the long term chosen by previous writers. A twelve year period

        was chosen so as to be long enough for many of SEO’s R&D and Capital projects to come to

        fruition thereby permitting testing of Hypotheses 2a and 2b on one hand; and short enough to

        allow for a relatively robust sample of issues on the other.



3.2.2.2. Performance Measurement

We use Cumulative Abnormal Returns (CAR) method to measure the performance of firms issuing

seasoned equity. The description of this methodology parallels Allen and Patrick (1996) and Spiess and

Affleck-Graves (1995).



Raw daily returns for issuers and non-issuers are first calculated as

                            PISS ,t                          PBM ,t                           AOI t
               rISS ,t =              −1        rBM ,t =              −1         rAOI ,t =             −1
                           PISS ,t −1                       PBM ,t −1                         AOI t −1

where              PISS,t = closing price of the SEO firm on day t

                   PBM,t = closing price of the benchmark non-issuing firm on day t

                   AOIt = closing value of the All Ordinaries Index on day t



The abnormal return is then calculated as the raw return from the issuing firm minus the return on the

corresponding non-issuer or the All Ordinaries Index. Hence

                                ari ,t = rISS ,t − rBM ,t        ari ,t = rISS ,t − rAOI ,t

where              rISS,t = Raw return for SEO on day t

                   rBM,t = Raw return for non-issuer benchmark firm on day t



                                                                                                            Page 11
                  rAOI,t = Raw return for All Ordinaries Index on day t



Next, the average abnormal return for the day t across all SEOs is calculated as the equally weighted

arithmetic average of the individual abnormal returns:

                                                 1 n
                                           ARt =  ∑ ari ,t
                                                  n  i =1
where             n = number of SEOs in the sample



Finally, the CAR from the first day after the offering until day t is calculated as the sum of the daily

average abnormal returns until t. Hence

                                                          t
                                           CARt = ∑ ARd
                                                         d =1




To test for the significance of the resulting cumulative abnormal return we use a modified t-statistic

that also accounts for the autocovariance that may exist in the time series:

                                                      CARt ⋅ n
                                t (CARt ) =
                                                t ⋅ var + 2 ⋅ (t − 1) ⋅ cov

where             var = average cross-sectional variance over the measurement period

                  cov = first-order autocovariance of the ARt series



The var and cov values for our sample are shown in Table 1.

                                                <Table 1>

Because Conrad and Kaul (1993) have documented a potential upward or downward bias introduced by

cumulating short term abnormal returns over long periods, we also define holding-period return as an

alternative returns measure:

                                                 b            
                                    HPRi ,a:b = ∏ (1 + Ri ,t ) − 1
                                                 t =a         

where             Ri,t = Raw return of firm i on day t

                  a = Beginning of the holding period

                  b = End of the holding period




                                                                                                   Page 12
The above formula will be used to measure “the total returns from a buy and hold strategy in which a

stock is purchased at the first closing market price after listing” (a=1) and held for the subsequent

short-term (b=3×252=756), medium term (b=5×252=1260) and long term (b=12×252=3024) period.



3.2.3. Cross Sectional Methodology

In the ‘cross-sectional analysis’ stage of our study we regress the returns of SEOs (dependent variable)

on a number of controlling factors (independent variables):

                                  CARi = α i + β i Ω i + ε i (univariate)

                CARi = α i + β1,i Ω1,i + β 2 ,i Ω 2 ,i + ... + β n ,i Ω n ,i + ε i (multivariate)



where             CARi = Cumulative abnormal return of SEO i for a five year period

                  Ωi = Control variable whose effect on SEO performance is being measured

                  αi, βi = Regression coefficients

                  εi = Regression error terms



The rationale for the choice of control variables was detailed in the Research Objectives section, above.



i.      Age (2 VARIABLES)

        c   INAGE           Number of years from the time of SEO firm’s incorporation in Australia.

                            This variable will look at the impact of issuing firm’s maturity (as proxied

                            by its effective life) on the extent of underperformance.

        c   PUBAGE          Number of years from the time of SEO firm’s listing on an organised stock

                            exchange in Australia. Through this variable we examine the impact of time

                            during which the SEO is ‘in the public eye’, on the post-issue performance.



ii.     Company Beta (1 VARIABLE)




                                                                                                     Page 13
       c    BETA           This variable will be used to test the adequacy of our benchmark in

                           controlling for the effect of firm’s risk as measured by the variability of its

                           returns relative to the market.



iii.   Market Capitalisation (1 VARIABLE)

       c    EQUITY         This variable proxies for the firm size and is calculated as the market value

                           of the firm expressed in 1990 dollars:

                                                          EQUITY = ln( MVadj )



iv.    Year of Issue (1 VARIABLE)

       c    ISSYR          The year in which each issue is made. This roughly proxies for the possible

                           chronological impact that issue years may have on the extent of

                           underperformance and loosely reflects on the existence of a learning curve

                           for the seasoned equity issuers.



v.     Volume of SEOs in the issue year (2 VARIABLES)

       To account for the possibility of either the overall climate, or the climate specific to sample

       firms impacting on their post-issue performance, we have used two variables for this factor:

       c    TOTVOL         Measure of the effect of total annual volume of SEOs on the issuing firm’s

                           performance. Calculated as

                                                      TOTVOL = ln(1 + ΨTOT )

       c    SAMPVOL Measure of the effect of sample annual volume of SEOs on the issuing

                           firm’s performance. Calculated as

                                                    SAMPVOL = ln(1 + ΨSAMP )



The final element is to investigate the impact of initial underpricing on the subsequent performance of

the issuer. The initial underpricing will be defined as   Ri = ℜi − R AOI , with raw return (ℜi) estimated

using four methods:




                                                                                                      Page 14
       c     CORERT        It calculates how deeply was each new share in the offer discounted with

                           respect to the closing price on the day of the issue.

                                                                  P 
                                                         CORERT =  0  − 1
                                                                   IP 
                           where     P0 =     Closing price on the day of the issue (t=0)

                                     IP =     Subscription price for each new share in the SEO



       c     ABSRT         Compares the closing price at the offer date with the closing price on the day

                           just prior to the issue, thus incorporating everything from the effect of equity

                           addition to the reaction by the market.

                                                                  P 
                                                          ABSRT =  0  − 1
                                                                  P 
                                                                   −1 


       c     DILRT         Similar to CORERT, but also takes into account the proportion of new

                           equity issued with respect to the equity in place prior to the offer thus

                           reflecting on the dilution effect of new shares.

                                                              (η + 1) × P0 
                                                              η × P + IP  − 1
                                                     DILRT =               
                                                                   o       

                           where     η=       Ratio at which new equity is issued.



       c     TOTRT         A holding period return for an investor who acquires the necessary number

                           of shares (η) on the last day before the SEO, exercises the right to buy the

                           extra equity, and sells it at the close of the day of the issue.

                                                             (η + 1) × P0 
                                                             η × P + IP  − 1
                                                    TOTRT =               
                                                                  −1      


As before, each of the market-adjusted definitions of the initial returns will be regressed on the three –

year and five – year CARs of the issuers. For all analyses, t-statistic and p-value will be used to assess

the significance of regression results.




                                                                                                      Page 15
4.     RESULTS

4.1.   TIME – SERIES PATTERNS

4.1.1. Short – Term and Medium – Term SEO performance

To compare the findings in our sample with previous results, we commence by examining the

performance of SEOs over period of up to five years. The results, summarised in Table 2 and Figure 1,

are quite striking.

                                               <Table 2>

                                               <Figure 1>

For the size-match-adjusted returns, the CAR reaches -26.07% significant at 5% level at the end of

third year. This is very similar to findings of Spiess and Affleck-Graves (1995) and Loughran and

Ritter (1995) who recorded three-year CARs at -23.15% and -27.2%, thus supporting the contention of

some correspondence between Australian and the US markets. Furthermore, this figure is also

consistent with Allen and Patrick (1996) and Ritter (1991) who studied the underperformance of initial

public offerings (IPOs) observing -25.38% and -29.13% CARs, respectively. Such finding would

therefore suggest that the causality of underperformance could be the act of issuing, rather than the type

of equity offering. Extending the view to the medium term, underperformance of SEOs largely subsides

resulting in CAR of just -15.03% at the end of fifth year. Although not as dramatic, Loughran and

Ritter (1995) also noted a drop-off in the level of underperformance after the third year. In fact,

McLaughlin, Safiedinne and Vasudevan (1996) observed insignificant abnormal returns in the fifth

year following an SEO.



When taking industry-and-size matched non-issuers as the benchmark, the underperformance of issuers

becomes more marked. The CAR is significantly negative for each of the five post-issue years reaching

-47.71%, which is significant at 1% level. Similar to the size-match-adjusted returns, underperformance

is not evident during years four and five and the CAR reaches -39.46% significant at 5% by the end of

medium term. This result parallels Spiess and Affleck-Graves (1995) who found five-year industry-

and-size matched CAR to be -31.24%, although they did not observe a peak in year three. Two further

observations can be made:

i.     The underperformance using this benchmark is significantly propelled by the first two years

       accounting collectively for a -39.22% CAR.




                                                                                                     Page 16
ii.    But for the extremity of these two years, underperformance from both size-matched and

       industry-and-size matched benchmarks is exceptionally similar. Both record significantly

       negative CAR in the first year, which peaks during year three and then becomes insignificantly

       positive. In fact, both benchmarks show economically significant overperformance in the last six

       months of 5.1% and 4.2% for size and industry-and-size benchmarks respectively, although

       neither is statistically significant.

Finally it should be noted that the five year CAR is significantly more negative (differential t-statistic

of -3.87) for the industry-and-size benchmark than it is for size-only benchmark. It can be therefore

concluded that the size matching hides some of the underperformance, which is revealed when industry

characteristics are accounted for.



The last benchmark – All Ordinaries Index – is quite conspicuous recording an overperformance of

38.85% (3.27) in the medium term in contrast to a -78.4% five-year CAR observed by Allen and

Patrick (1996) on IPOs. Although we could attribute this disparity to the difference of samples (SEOs

v. IPOs), this may not be reasonable given the similarity of results with previous benchmarks. Noting

that Allen and Patrick’s (1996) sample covers only the period from 1974 to 1984, we decided to

investigate the possible cause for this phenomenon by looking separately at the firms issuing before

and after the October crash of 1987. To do so, we have assigned all firms issuing at least one year prior

to the crash into the ‘pre-crash’ sub-sample, and all SEOs that occurred at least one year after the crash

into the ‘post-crash’ sub-sample, deleting all observations immediately surrounding the crash itself. To

mitigate bias, we have kept the number of firms in the post-crash sample the same as the one in pre-

crash sample.

                                               <Figure 2>



While pre-crash sample recorded a -18.2% (-2.01) CAR three years after the issue, post-crash SEOs

have overperformed at +22.1% (2.31) during the same period. The underperformance of pre-crash

SEOs then subsided (perhaps reflecting on the fact that for many of these SEOs this is already a period

after the crash), however the post-crash firms continued to overperform reaching 34.3% at the end of

fifth year. T-test also shows that the difference of the two CARs is statistically significant at 1% level.

A possible explanation for this phenomenon is that it was mainly the firms with good future prospects




                                                                                                      Page 17
that dared to issue seasoned equity following the crash, and as such they have outperformed an

otherwise sluggish market. However, when compared to their matched non-issuing counterparts, the

perception of good performance dissipates.



In sum, therefore, results reject Hypothesis 1a – firms issuing seasoned equity do underperform with

respect to corresponding non-issuers. Hypothesis 1b requires some qualification. Firms issuing

seasoned equity underperform the all ordinaries index prior to the stock market crash, but outperform

the market after the crash.



4.1.2. Long – Term SEO performance

A twelve-year SEO performance is observed to establish whether the above underperformance is a

persistent phenomenon.



The results from Table 3 and Figure 3 are quite remarkable despite the expected drop-off in

significance levels resulting from the reduced SEOs in the sample.

                                                <Table 3>

                                               <Figure 3>



When size-matched non-issuers are used as the benchmark, the performance of SEOs over the first five

years parallels the full sample reported in Table 2, suggesting that such underperformance is a

phenomenon attributable to the act of issuing rather than selection of firms in the sample. CAR is

significantly negative in both cases during the first year, peaking in year three at -22.52% significant at

5% level, which is consistent with -26.07% for the short-term sample, also significant at 5%.

Underperformance then subsides in both cases during fourth and fifth year following the issue.

Strikingly, however, the performance continues to improve further during years six and seven. The

CAR crosses into positive territory in the first quarter of year 6 recording significantly positive 15.29%

by the end of the year. It then proceeds to peak at +28.9% in the middle of seventh year before starting

its decline, reaching zero once again in the last quarter of year eight. From this point onwards the CAR

hovers around zero until year twelve, not recording any statistically significant values along the way.




                                                                                                      Page 18
The results identified above closely resemble those obtained using industry-and-size matched non-

issuers as the benchmark. Underperformance is again evident during the first three years, but also

disappears in years four and five. In a manner similar to the size-matched benchmark, significant

overperformance is present toward the end of year five and in years six and seven. Over these three

years the cumulative abnormal return amounts to +33.06% (1.78) which is significant at 5% and

consistent with size-matched CAR of +37.12% (1.66) over the same period. In both cases year eight

registers a substantial downturn of -12.98% and -19.84% for industry-and-size and size-matched

benchmarks, respectively, both significant at 10% level. By the end of year twelve, the CAR is

economically significant at -40.59%, although such underperformance is now statistically insignificant

with t-statistic at just -0.71. A notable difference may be observed in the fact that the CAR computed

using industry-and-size benchmark does not cross over to the positive territory. This can be attributed

to the more pronounced underperformance in years two and three (totalling over -36%) not present for

the size-matched benchmark. A possible influence could be credited to the fact, that over 50% of SEOs

in this sample were in mining and oil exploration industries. Consequently, if these industries

performed quite well compared to the market, then even a moderate underperformance of SEOs would

be largely amplified when compared to the industry-and-size matched counterparts.



Finally, looking at the performance against the all ordinaries index, a similar picture arises: significant

underperformance (-16.41%; -1.37) in the first year, followed by a cross over into the positive territory,

then a significant overperformance in years six (+19.74%; 1.64) and seven (+19.11%; 1.60) and a

subsequent drop-off in year eight. Two differences can be noted: First, the crossover from negative to

positive CAR occurred much earlier. Second, the drop-off in year eight was not as significant, and was

in fact followed by slight overperformance thereafter. An explanation could, once again, stem from the

effect of stock market crash examined earlier, noting also that the crossover was roughly concomitant

with the actual event of the crash.



Overall, the results reject Hypothesis 2a – firms issuing seasoned equity do cross over from the period

of underperformance to the period of overperformance relative to the non-issuer benchmarks and the

market. Except for the AOI-adjusted returns, results also reject Hypothesis 2b – SEOs do not

underperform, in aggregate, over the real ‘long-run’. These two conclusions thus support the argument




                                                                                                      Page 19
that the firms issuing seasoned equity initially underperform perhaps due to the correction of run-up

overperformance. During this period, however, firms continue to invest more than their counterparts in

Research and Development and Capital Infrastructure. When these expenditures eventually come to

fruition around year six, these issuers turn around their poor performance into significantly positive

abnormal returns. As the rest of the market eventually catches up with these developments, issuers

loose their competitive advantage and their performance eventually reverts to closely match their non-

issuing counterparts.



4.1.3. Performance of Multiple Issuers

The industry composition of multiple issuers is quite diversified, with no industry constituting more

than 10% of the overall sample. Consequently, we will only be looking at the performance of these

firms using size-matched non-issuing counterparts as the benchmark. Given there are no firms issuing

four SEOs and only two firms offering five, we will not examine the effect of fourth and fifth issuing as

this may not be representative of the population. Furthermore, since the average time between issues is

around 2.6 years, we shall look at the performance over the first two years only to mitigate bias from

overlap. An added benefit of this time frame is that it permits the observation of the critical

underperformance period defined by Loughran and Ritter (1995) to be from 7th to 24th month

following the issue. Table 4 summarises the results on a six-monthly basis. Figure 4 graphs the

findings and Figure 5 then presents the corresponding trend lines for clarity.

                                               <Table 4>

                                               <Figure 4>

                                               <Figure 5>



As expected, the underperformance of first SEO issue is quite evident, reaching -21.38% by the end of

second year, significant at 5% level, propelled mainly by the first six-month period. In a similar

fashion, second and third issues also recorded a significant two-year cumulative abnormal return of

-32.78% and -34.75%, respectively, both significant at 5% level, and in a the same fashion these were

also fuelled primarily during the first six months. Moreover, months seven to twelve showed a decrease

in underperformance to -8.95% and -6.32%, which is consistent with -6.27% CAR for the first issue.




                                                                                                    Page 20
To investigate these similarities further we have conducted a series of paired t-tests on the abnormal

returns for the multiple issues and found the probabilities that the two samples in the pair have come

from the same underlying population. The (unreported) results show that there is over 70% chance that

first issue has the same average abnormal returns over time as the second and third issue. This is

despite the fact that the subsequent issues, made on average over two years later, necessarily face

different economic climates than their predecessors. Consequently, these probabilities show

specifically the similarities that arise from the act of issuing, rather than the associated market

environment. Moreover, the correspondence between second and third issue is even stronger showing

almost 95% probability that the underlying population samples are the same.



In sum, therefore, the above results reject Hypothesis 3 – performance characteristics of firms having

more than one issue of seasoned equity are, in fact, economically and statistically similar for each

offering. This statement further provides some support to the contention that the underperformance is a

consequence of the act of issuing.



4.2.   CROSS – SECTIONAL PATTERNS

4.2.1. Univariate Cross-Sectional Regressions

Tables 5, 6 and 7 present results derived from regressing the three-year and five-year cumulative

abnormal returns using each benchmark on the seven control variables detailed in section 3.2.3 above.

                                              <Table 5>

                                              <Table 6>

                                              <Table 7>



The resulting coefficients concur with our predictions. The Company Age factors, INAGE and

PUBAGE are both positive implying a negative relationship between firm’s maturity and the extent of

underpricing. While time since listing (PUBAGE) does not seem to play a major role in the abnormal

returns, the amount of time since company’s incorporation (INAGE) shows statistical significance for

all benchmarks. These results are consistent with Spiess and Affleck-Graves (1995) who also observed

this relationship using different methodology. To put this into perspective, we have calculated three-




                                                                                                  Page 21
and five-year HPRs for each firm and sorted the results by deciles and halves according to INAGE

(Table 8).

                                                <Table 8>



The BETA coefficient was found to be positive as expected and highly significant using all benchmarks

(1% level for non-issuer benchmarks and 5% level for all ordinaries index benchmark). This provides a

strong support to the contention that the risk of the firm, as proxied by beta, plays an important role in

the determination of firm’s CAR. The significance of this result prompted further research, the subject

of a second paper. The Market Capitalisation variable, EQUITY, has also shown negative coefficient as

expected. Although the result was statistically insignificant using all benchmarks, it is consistent with

findings of Allen and Patrick (1996) suggesting similarity between post-issue performance of IPOs and

SEOs. ISSYR, the variable representing the year of SEO issue, was found to be statistically and

economically insignificant indicating that the chronological distribution of SEOs does not impact on

the relative performance of issuers and that the importance of a ‘learning curve’ is thus likely to be

minimal. Finally, the SEO volume variables – TOTVOL and SAMPVOL – were found to be positive as

anticipated. While they were insignificant against non-issuers’ benchmark, TOTVOL was found to be

highly significant against the AOI benchmark. Although not as pronounced, these results are

paralleling those of Allen and Patrick (1996) who also found volume of IPOs in a given year to impact

on the post-issue abnormal returns computed using All Ordinaries Index.



A second basic observation is the high consistency across time. Although statistical significance is

somewhat reduced in the three-year data, the economic significance of five-year and three-year

regressions is generally very similar.



Finally, the results are also quite consistent across benchmarks. Perhaps with the exception of TOTVOL

variable, the statistical significance of control factors is consistent between the size-matched, industry-

and-size matched non-issuers and the All Ordinaries Index as the benchmark.



In sum the results reject Hypotheses 4a and 4b – the extent of underperformance is a function of age,

when age is expressed as the amount of time since incorporation and is related to company beta. They




                                                                                                      Page 22
fail to reject Hypotheses 4c and 4d – the extent of underperformance is not a function of market

capitalisation nor the chronological year of issuance. They also reject Hypothesis 4e – the extent of

underperformance is related to the volume of SEOs issued in the corresponding year, but only when the

market performance is used as a benchmark. Hypothesis 4e is not rejected for other benchmarks.



4.2.2. Univariate Cross-Sectional Regressions: Initial Underpricing Tests

As Table 9 depicts, a highly significant and positive opening return can be noted under each method of

calculation. CORERT provides a conspicuously large initial return, although such could be expected

given its ignorance of the ratio in which the discounted new shares are distributed relative to the

existing equity.

                                               <Table 9>



Table 10, 11 and 12 show the results of regressing the CARs on the corresponding initial gains under

different benchmarks.

                                              <Table 10>

                                              <Table 11>

                                              <Table 12>



With the exception of DILRT variable, none of the regressions provide consistently significant results

and can be discounted as the potential factors affecting the extent of underperformance. DILRT,

however, records significantly negative coefficients under each benchmark. This lends some credence

to the proposition that the post-issue underperformance mitigates some of the costs associated with the

issue. Moreover, given the definition of DILRT, the observed relationship covers only the extent to

which new shares are underpriced relative to the existing equity, but excludes the effect of any other

information released on the opening day. The relationship thus seems to be directly related to the cost

of underpricing an SEO.



These results therefore reject Hypothesis 5a finding that firms issuing seasoned equity do record

significantly positive initial returns. They also reject Hypothesis 5b showing that the extent of initial




                                                                                                    Page 23
returns and subsequent underperformance are related, conditional on a correct definition of the initial

gain.



4.2.3. Multivariate Cross-Sectional Regressions

To account for possible interdependence in the above results we have performed a multivariate

regression analysis of the five-year cumulative abnormal returns on the four most ‘significant’ control

variables. The correlation matrix for these variables was is presented in Table 13, while the results of

the multivariate regressions are shown in Table 14.

                                              <Table 13>



Although low in absolute values, the INAGE – DILRT, BETA – INAGE and BETA – TOTVOL

correlations are quite noticeable.

                                              <Table 14>



It can be immediately noticed that, with the exception of TOTVOL for non-issuer benchmarks, all

results are significant at least at a 10% level. This is not unexpected, however, given that TOTVOL was

also insignificant in univariate regressions for the same benchmarks. Even more striking is the

observation that almost all variables retained their significance at levels concomitant with their

corresponding univariate regressions concluding that the relationships so established were actual, rather

than a phenomenon of high correlation. Consequently, this result also lends credence to the rejection of

Hypotheses 4a, 4b, 4e and 5b concerning the effect of age, beta, volume of SEOs and initial

underpricing, respectively. Moreover, the Adjusted R-squares from the multivariate regressions

(ranging from 10.1% to 18.1%) demonstrate the relatively strong explanatory power of these variables.



5.      CONCLUSION

Our results have shown that the general conclusion of many writers accusing firms issuing seasoned

equity of long run underperformance is dependent on the definition of the ‘long run’. When long run

was defined as twelve years instead of the usual five years, SEOs can be clearly found to turn around

their performance particularly during years six and seven. Depending on the selection of benchmark, by




                                                                                                    Page 24
the end of year twelve the cumulative performance ranges from positive to negative CAR, with a strong

average tendency towards zero.



This research has also demonstrated that the immediate post-issue underperformance pattern is

qualitatively equal and quantitatively similar for each SEO when a single firm makes more than one

issue. This finding lends credence to the assertion that such initial poor performance is a phenomenon

directly attributable to the act of issuing.



A series of regression results have pointed to a number of factors that bear influence on the extent of

the initial underperformance. Decreased ex-ante uncertainty associated with older firms causes a

negative relationship between the age and the extent of underpricing. Moreover, the greater is the SEO

cost specifically associated with underpricing of the new equity, the greater will be the

underperformance that follows the issue. Finally, the greater is the risk of the issuer as proxied by its

beta, the more compensation will investors require for taking on new shares, and hence the smaller will

be the subsequent loss of performance.



Collectively, these findings add plausibility to the following possible scenario. Following an equity

issue, firms deploy their newfound wealth into capital investments and research and development,

which were found by Loughran and Ritter (1997) to exceed significantly their non-issuing counterparts.

Facing such expenditures as well as the high costs that accompanied the issue, the outflows

immediately impact on firms through loss of profitability and hence performance. As the capital

expenditures and the R&D projects come later to fruition, these firms suddenly find themselves

substantially outperforming the non-issuers. The competition, however, eventually catches up and

gradually erodes any abnormal profits that the issuer may have been enjoying. Subsequently, the

performance of issuers and non-issuers will match each other, until such time when another equity

issue is made.



The implication of our findings also takes on more practical application. For example, an active trader

may consider buying shares just prior to SEO so as to reap the benefits of initial underpricing, but sell

soon after to avoid the initial post-issue underperformance. Such trader may again repurchase shares in




                                                                                                    Page 25
this firm around year five to six following the SEO, and take advantage of the recovery phase selling

before year eight before underperformance will, again, step in. For a passive investor with a long term

horizon our findings provide very good news – there is no need to panic and sell before an SEO since

the long term returns will not be eroded by the act of issuing. It should be reiterated, however, that all

results are based on sample averages, and as such these strategies may not work on an individual-firm

basis.




                                                                                                     Page 26
5.1.    TABLE 1
             VAR and COV Values for the Three Time Frames and Three Benchmarks
The VAR and COV parameters calculated from the data are input into the formula (viz below) used to determine
the significance of the calculated cumulative abnormal returns. VAR and COV are the variance and the first-order
autocovariance of excess returns, respectively. The excess returns are measured against the size-matched and
industry-and-size matched non-issuers benchmarks and the market index benchmark. An added benefit of this
formula is its ability to account for possible autocorrelation that may exist in the observed CARs.
                                                                   CARt ⋅ n
                                               t (CARt ) =
                                                             t ⋅ var + 2 ⋅ (t − 1) ⋅ cov


                                                                                           Benchmark
                                                  MV                                        Ind + MV        AOI
Short               Var                      0.00166442                                    0.00258872    0.00104262
                    Cov                     -1.1237E-06                                    -1.2684E-06   5.2958E-08
Medium              Var                      0.00203780                                    0.00243249    0.00114219
                    Cov                     -1.0696E-06                                    -4.7655E-07   -5.3733E-07
Long                Var                      0.00154412                                    0.00226587    0.00119837
                    Cov                     -2.2031E-06                                    -5.4077E-06   -2.2159E-06




                                                                                                                       Page 27
5.2.    TABLE 2
        Short and Medium Term Performance of SEOs measured against the i) size-matched
    non-issuers, ii) industry-and-size matched non-issuers and the iii) market-index benchmarks
                                        over the January 1984 to October 1993 period.
The table reports the excess returns of issuers calculated under the three benchmarks for a period of five years
since issuing. The data cover 102 seasoned equity offerings from January 1984 to October 1993. The y(CARt)
column shows the cumulative abnormal return of daily abnormal returns specifically for the one year shown in the
Years Since Seasoning column. The CARt shows the cumulative abnormal return of all daily abnormal returns
since the date of seasoned equity issue. The corresponding calculations have been defined by equations:
                                  1 n                                                                    1                                          
                      CARt = ∑    ∑ { ISS (i ,d ) − rBM ( i ,d ) }                            ∑   n  ∑ {r                         − rBM ( i ,d ) }
                               t                                                                    t               n
                                               r                                   y (CARt ) =                          ISS ( i ,d )
                             d =1   n  i =1                                                      
                                                                                                 d = t −1          i =1                                  
It should be noted that, for clarity, subscript t in the above equations refers to the year in the ‘Years Since
Seasoning’ column. Mathematically, however, each such year comprises of 252 trading days whose returns (ri,day)
are actually used in the computations. The significance levels (presented in brackets) have been calculated based
on the following t-statistic formula (where t refers to the specific time frame covered by CARt or yCARt):
                                                                                CARt ⋅ n
                                                         t (CARt ) =
                                                                         t ⋅ var + 2 ⋅ (t − 1) ⋅ cov

The added benefit of this t-statistic definition is its ability to account for series autocorrelation when reporting
significance level.
                                   Size matched                              Inds+Size matched                                                    AOI matched
Years        since       y(CARt)                 CARt                    y(CARt)              CARt                             y(CARt)                       CARt
seasoning                [%]                     [%]                     [%]                  [%]                              [%]                           [%]
                                    *                       *                     ***                        ***
1                        -10.02                  -10.02                  -25.21               -25.21                           0.36                          0.36
                         (-1.41)                 (-1.41)                 (-3.25)              (-3.25)                          (0.07)                        (0.07)
                                                            *                                                ***
2                        -4.25                   -14.27                  -14.01               -39.22                           6.05                          6.41
                         (-0.42)                 (-1.42)                 (-1.28)              (-3.58)                          (0.81)                        (0.85)
                                                            **                                               ***
3                        -11.80                  -26.07                  -8.49                -47.71                           6.34                          12.75*
                         (-0.96)                 (-2.12)                 (-0.63)              (-3.55)                          (0.70)                        (1.39)
                                                            *                                                ***
4                        6.20                    -19.88                  3.24                 -44.47                           10.65                         23.40**
                         (0.44)                  (-1.40)                 (0.21)               (-2.87)                          (1.00)                        (2.20)
5                        4.85                    -15.03                  5.01                 -39.46**                         15.45*                        38.85***
                         (0.31)                  (-0.95)                 (0.29)               (-2.28)                          (1.30)                        (3.27)
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                                                                                                                                                                        Page 28
                                                                                                      FIGURE 1
                                          Five-Year Performance of Firms Issuing Seasoned Equity
This figure maps the cumulative daily abnormal returns for the 102 SEOs detailed in Table 2 over the five-year
period (ie 5years x 12 months x 21days = 1260 trading days). The CARs under different benchmarks are shown.
With the exception of AOI-based excess returns, the performance characteristics are largely parallel. AOI
benchmark was subject to further investigation based on pre- and post- October 1987 crash data (viz Figure 2).

    0.5

                                                                                                                                                                                                               AOI -
    0.4
                                                                                                                                                                                                               MATCH


    0.3


    0.2


    0.1


     0




                                                                                                                                                                                          1021
                                                                                                                                                                                                 1055
                                                                                                                                                                                                        1089
                                                                                                                                                                                                               1123
                                                                                                                                                                                                                      1157
                                                                                                                                                                                                                             1191
                                                                                                                                                                                                                                    1225
                                                                                                                                                                                                                                           1259
                        103
                              137
                                    171
                                          205
                                                239
                                                      273
                                                            307
                                                                  341
                                                                        375
                                                                              409
                                                                                    443
                                                                                          477
                                                                                                511
                                                                                                      545
                                                                                                            579
                                                                                                                  613
                                                                                                                        647
                                                                                                                              681
                                                                                                                                    715
                                                                                                                                          749
                                                                                                                                                783
                                                                                                                                                      817
                                                                                                                                                            851
                                                                                                                                                                  885
                                                                                                                                                                        919
                                                                                                                                                                              953
                                                                                                                                                                                    987
              35
                   69
          1




   -0.1


   -0.2
                                                                                                                                                                                                                  SIZE -MATCH

   -0.3


   -0.4


                                                                                                                                                                                                  INDS-AND-SIZE
   -0.5
                                                                                                                                                                                                  MATCH

   -0.6




                                                                                                                                                                                                                                           Page 29
                                                           FIGURE 2
                         Comparison of Pre-Crash and Post-Crash Performance of SEOs
The excess returns of issuing companies is further examined when All Ordinaries Index is taken as a benchmark to
investigate the reasons for positive CAR exhibited in Figure 1. An equal number of relative returns was taken up to
one year before the crash (constituting the pre-crash sample) and at least one year after the crash (constituting the
post-crash sample). It can be observed that while pre-crash sample parallels in its characteristics excess returns
measured against other benchmarks, after the crash positive abnormal returns can be observed. A possible
explanation may stem from the requirement of a robust issuer that dares to issue new equity following the 1987
crash.

    0.4


                                                                                                         POST-CRASH

    0.3




    0.2




    0.1




     0
          1   48   95   142 189 236 283 330 377 424 471 518 565 612 659 706 753 800 847 894 941 988 1035 1082 1129 1176 1223



   -0.1
                                                                                                             PRE-CRASH


   -0.2




   -0.3




                                                                                                                               Page 30
5.3.     TABLE 3
              Long-Term Performance of SEOs measured against the i) size-matched
    non-issuers, ii) industry-and-size matched non-issuers and the iii) market-index benchmarks
                                  over the January 1984 to October 1996 period.
The table reports the excess returns of issuers calculated under the three benchmarks for a period of twelve years
since issuing. The data cover 21 seasoned equity offerings from January 1984 to October 1986. The y(CARt)
column shows the cumulative abnormal return of daily abnormal returns specifically for the one year shown in
the ‘Years Since Seasoning’ column. The CARt shows the cumulative abnormal return of all daily abnormal
returns since the date of seasoned equity issue. The mathematical expressions for each of these measures, as
well as the corresponding t-statistic, are identical to those defined in Table 2, above.
                      Size matched                     Inds+Size matched                   AOI matched
Years        since    y(CARt)           CARt           y(CARt)         CARt                y(CARt)       CARt
seasoning             [%]               [%]            [%]             [%]                 [%]           [%]
                               *                 *              *                                   *
1                     -20.96            -20.96         -24.05          -24.05*             -16.41        -16.41*
                      (-1.54)           (-1.54)        (-1.46)         (-1.46)             (-1.37)       (-1.37)
                                                                *               **
2                     3.02              -17.95         -22.43          -46.48              14.65         -1.76
                      (0.22)            (-0.93)        (-1.36)         (-2.00)             (1.22)        (-0.10)
                                                 *                              **
3                     -4.58             -22.52         -13.62          -60.10              4.74          2.98
                      (-0.34)           (-1.29)        (-0.83)         (-2.11)             (0.40)        (0.14)
                                                 *                              **
4                     -1.03             -23.56         -4.72           -64.83              7.16          10.14
                      (-0.08)           (-1.31)        (-0.29)         (-1.97)             (0.60)        (0.42)
5                     22.84**           -0.72          12.08*          -52.75*             15.20         25.34
                      (1.68)            (-0.02)        (1.22)          (-1.43)             (1.27)        (0.95)
                              *                 *                               *                  *
6                     16.01             15.29          1.16            -51.59              19.74         45.08*
                      (1.28)            (1.26)         (0.07)          (-1.28)             (1.64)        (1.54)
                                                *              *                                   *
7                     -1.73             13.55          19.82           -31.77              19.11         64.19**
                      (-0.13)           (1.21)         (1.46)          (-0.73)             (1.60)        (2.03)
                               *                                *
8                     -19.84            -6.29          -12.98          -44.75              -3.39         60.80**
                      (-1.46)           (-0.16)        (-1.22)         (-0.96)             (-0.28)       (1.80)
9                     -4.69             -10.98         -0.50           -45.24              7.96          68.76**
                      (-0.35)           (-0.27)        (-0.03)         (-0.92)             (0.66)        (1.91)
10                    5.93              -5.05          -1.44           -46.68              3.32          72.08**
                      (0.44)            (-0.12)        (-0.09)         (-0.90)             (0.28)        (1.90)
11                    0.29              -4.76          -1.56           -48.24              6.19          78.27**
                      (0.02)            (-0.11)        (-0.10)         (-0.88)             (0.52)        (1.97)
12                    8.05              3.29           7.65            -40.59              6.61          84.88**
                      (0.59)            (0.07)         (0.47)          (-0.71)             (0.55)        (2.05)




                                                                                                                   Page 31
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                              Page 32
                                                               FIGURE 3
                              Long Run Performance of Firms Issuing Seasoned Equity
This figure maps the cumulative daily abnormal returns for the 21 SEOs detailed in Table 3 over the twelve-year
period (ie 12 years x 12 months x 21days = 3024 trading days). The CARs under different benchmarks are shown.
Astoundingly the chart of excess returns against size-matched benchmarks shows a reversal in performance of
issuing companies gradually more than making up for the initial losses. As the competition eventually catches up,
however, this performance edge is then eroded with issuers’ cumulative returns at the end of the period matching
those of non-issuers. These characteristics are mirrored in data measured against the other benchmarks, albeit
finishing with either positive or negative CARs. Overall data have a strong average tendency towards matched
performance over the real ‘long-run’.

     1
                                                                                                                   AOI-MATCH

   0.8



   0.6



   0.4



   0.2
                                                                                                                         SIZE-MATCH

     0
          1   113 225 337 449 561 673 785 897 1009 1121 1233 1345 1457 1569 1681 1793 1905 2017 2129 2241 2353 2465 2577 2689 2801 2913

   -0.2



   -0.4



   -0.6
                                                                                                                      INDS-AND-SIZE
                                                                                                                      MATCH


   -0.8




                                                                                                                                          Page 33
5.4.     TABLE 4
                                  Two-Year Performance of Multiple SEOs
This table summarises the abnormal returns of companies performing multiple seasoned equity offerings,
measured against the three benchmarks. No companies have performed four seasoned equity offerings, and only
two companies performed five. As these would be unrepresentative of the population we only look at one, two
and three consecutive issues. The average time between consecutive SEOs is 2.6 years. We shall therefore look
at a period of only 2 years to mitigate bias on one hand, but enable the examination of the critical 7th to 24th
month, on the other. Data cover 77 SEOs over the period from January 1984 through to October 1993. The
p(CARt) column shows the cumulative abnormal return of daily abnormal returns only for the 6 months period
shown in the ‘Time Since Seasoning’ column. The CARt shows the cumulative abnormal return of all daily
abnormal returns since the date of SEO.
                                 First issue                    Second issue                   Third issue
Time        since    p(CARt)           CARt             p(CARt)        CARt          p(CARt)         CARt
seasoning            [%]               [%]              [%]            [%]           [%]             [%]
                             *                 *                 **             **            *
6 mths               -9.23             -9.23            -14.36         -14.36        -13.30          -13.30*
                     (-1.34)           (-1.34)          (-1.78)        (-1.78)       (-1.64)         (-1.64)
                                                   *                            **
12 mths              -6.27             -15.51           -8.95          -23.31        -6.32           -19.62**
                     (-0.78)           (-1.36)          (-1.11)        (-2.04)       (-0.78)         (-1.72)
                                                                                **
18 mths              2.49              -14.17           -5.00          -28.47        -5.96           -25.37**
                     (0.31)            (-1.01)          (-0.62)        (-2.03)       (-0.74)         (-1.81)
                                                   **                           **
24 mths              -8.36             -21.38           -4.47          -32.78        -9.17           -34.75**
                     (-1.03)           (-1.62)          (-0.55)        (-2.03)       (-1.13)         (-2.15)
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                                                                                                                Page 34
                                                                FIGURE 4
                                Cumulative Abnormal Performance of Multiple Issuers
This figure maps the cumulative abnormal returns measured against the size-matched non-issuers’ benchmarks and
summarised in Table 4, for companies issuing more than one seasoned equity offering. In this raw CAR chart it
can already be noticed that each offering produces qualitatively same and quantitatively similar post-issue
performance pattern. This suggests that the underperformance generally observed to accompany SEOs in the short
term is a consequence of the act of issuing. To clarify this proposition further, trend lines have been obtained for
each of the CAR patterns, presented in Figure 5 below.

   0.05


     0
          1   16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241 256 271 286 301 316 331 346 361 376 391 406 421 436 451 466 481 496

  -0.05


   -0.1


  -0.15


   -0.2


  -0.25


   -0.3


  -0.35


   -0.4


  -0.45

                                                     FIRST ISSUE      SECOND ISSUE       THIRD ISSUE




                                                                                                                                        Page 35
                                                                 FIGURE 5
                                 Trend lines of Multiple SEOs’ post-issue Performance
The trend lines of cumulative abnormal returns following each of the three issues by the same issuer are presented.
If it is true that the post issue underperformance is, in fact, a consequence of the act of issuing, then a qualitatively
identical and quantitatively similar pattern should be present for each trend line. This is what we observe.
Following each SEO the issuer immediately enters a period of downturn that progresses into a plateau towards the
end of first / beginning of second year, and then proceeds onto a further period of underperformance. It should be
noted that to avoid a period overlap we only examine the first two years following each issue. This, however,
prevent us from following through to find out if each issue by the same firm eventually turns around on to the
period of overperformance as hinted in the medium term and clearly demonstrated in the long run.

    0.05



      0
           1   16 31 46 61 76 91 106 121 136 151 166 181 196 211 226 241 256 271 286 301 316 331 346 361 376 391 406 421 436 451 466 481 496

   -0.05



    -0.1



   -0.15
                                                                                                                          FIRST ISSUE


    -0.2



   -0.25
                                                                                                                  THIRD ISSUE
                                                                                    SECOND ISSUE
    -0.3



   -0.35



    -0.4



   -0.45




                                                                                                                                          Page 36
5.5.    TABLE 5
               Cross-Sectional Univariate Regression of CARs (measured against the
                       Size-Matched benchmark) on Characteristics Variables
The table shows the constants, coefficients and their respective t-statistics for univariate regression of
cumulative abnormal returns as measured relative to size-matched non-issuers benchmark, against the seven
variables suggested to have an impact on the extent of post-issue underperformance. This regression was
defined by CARiS − BM = α i + β i Ω i + ε i (where CARS-BM refers to the excess returns measured relative to size-

matched benchmark; α and β are the intercept and coefficient, respectively; Ω is the independent factor and ε
represents the error term). Both the short-term and the medium-term time frames have been examined. R-Square
statistics are also provided. It can be observed that with the notable exception of beta, none of the other factors
turns out to be highly significant. Moreover, the beta factor also records a relatively high R-square statistic at
7.1% and 8.4% for three and five year returns, respectively. These findings appear to be relatively consistent
across time. We have also re-run these regressions (unreported) whilst controlling for the effects of
heteroscedasticity, including computations based on ARCH/GARCH models. Similar conclusions have been
reached.
                                Constant                       Coefficient                     R-Square
Control              3-Year          5-Year          3-Year          5-Year          3-Year          5-Year
Variable
INAGE                -0.392**        -0.209          0.001           0.003*          0.001201        0.000879
                     (-2.43)         (-1.02)         (0.33)          (1.28)
PUBAGE               -0.217          -0.077          -0.004          0.008           0.004665         0.007998
                     (-1.33)         (-0.32)         (-0.57)         (0.75)
BETA                 -1.009***       -1.225***       0.847***        1.222***        0.070828        0.083900
                     (-3.55)         (-3.26)         (2.76)          (3.01)
EQUITY               -0.466          0.162           0.037           -0.060          0.005937         0.008730
                     (-1.68)         (0.45)          (0.77)          (-0.94)
ISSYR                35.50           85.23           -0.018          -0.043          0.002641         0.008496
                     (0.51)          (0.93)          (-0.51)         (-0.93)
TOTVOL               -0.427          -0.242          0.048           0.026           0.000270         0.000429
                     (-0.44)         (-0.19)         (0.16)          (0.07)
SAMPVOL              0.780           0.988           0.442           0.483           0.021670         0.014687
                     (1.09)          (1.05)          (1.49)          (1.22)
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                                                                                                               Page 37
5.6.    TABLE 6
              Cross-Sectional Univariate Regression of CARs (measured against the
               Industry-and-Size-Matched benchmark) on Characteristics Variables
This table shows the constants, coefficients and the corresponding t-statistics resulting from a univariate
regression of cumulative abnormal returns (measured relative to industry-and-size-matched non-issuers
benchmark) against the seven variables posited as having an impact on the extent of post-issue
underperformance. This regression was defined by CARiIS − BM = α i + β i Ω i + ε i (where CARIS-BM refers to the

excess returns measured relative to industry-and-size-matched benchmark; α and β are the intercept and
coefficient, respectively; Ω is the independent factor and ε represents the error term). Both the short-term and
the medium-term time frames have been examined. R-Square statistics are also provided. It can be instantly
noted that beta factor is a prominent characteristic in these regressions, recording a significance at a 1% level.
An R-square of 3.7% (5.9%) for the three-year (five-year) period is also high in relative terms. We can further
observe a significant impact (at 5% level) of the INAGE variable that measures the time since incorporation.
Some influence is also observed from TOTVOL, a variable indicating the amount of seasoned equity issues in
the corresponding issue years. These findings appear to be relatively consistent across time. We have also re-run
these regressions (unreported) whilst controlling for the effects of heteroscedasticity, including computations
based on ARCH/GARCH models. Similar conclusions have been reached.
                               Constant                        Coefficient                    R-Square
Control              3-Year          5-Year          3-Year          5-Year          3-Year          5-Year
Variable
INAGE                -0.146*         0.096           0.014           0.013**         0.006784        0.049633
                     (-1.48)         (0.36)          (0.42)          (2.18)
PUBAGE               -0.248          -0.086          0.009           0.014           0.009221        0.012129
                     (-1.25)         (-0.25)         (0.66)          (0.92)
BETA                 -1.152***       -1.418***       0.836**         1.235***        0.036774        0.058722
                     (-2.89)         (-3.08)         (1.94)          (2.49)
EQUITY               -0.262          0.308           -0.030          -0.122          0.002010        0.024509
                     (-0.69)         (0.70)          (-0.45)         (-1.58)
ISSYR                -10.46          0.488           0.005           -0.001          0.000107        0.000529
                     (-0.11)         (0.01)          (0.10)          (0.02)
                                              *
TOTVOL               -1.384          -2.955          0.289           0.785*          0.005128        0.027713
                     (-1.02)         (-1.88)         (0.71)          (1.68)
SAMPVOL              -0.651          -1.338          0.097           0.421           0.000558        0.007642
                     (-0.66)         (-1.16)         (0.23)          (0.87)
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                                                                                                              Page 38
5.7.    TABLE 7
               Cross-Sectional Univariate Regression of CARs (measured against the
                       Market-Index benchmark) on Characteristics Variables
The table shows the constants, coefficients and the corresponding t-statistics for univariate regression of
cumulative abnormal returns as measured relative to All Ordinaries Index – matched benchmark, against the
seven variables proposed to have an impact on the extent of post-issue underperformance. This regression was
defined by CARiAOI − BM = α i + β i Ω i + ε i (where CARAOI-BM refers to the excess returns measured relative to

market index (AOI) benchmark; α and β are the intercept and coefficient, respectively; Ω is the independent
factor and ε represents the error term). Both the short-term and the medium-term time frames have been
examined. R-Square statistics are also provided. As with previous two benchmarks, beta factor again stands out
as having a significant impact on the post-issue performance recording also relatively high R-square of 3.0%
and 4.6% for short and medium term regressions, respectively. Importantly, INAGE representing the time since
incorporation and the TOTVOL indicating the volume of SEOs in a given year both gather on prominence with
5% and 1% significance levels, respectively. These findings appear to be relatively consistent across time. We
have also re-run these regressions (unreported) whilst controlling for the effects of heteroscedasticity, including
computations based on ARCH/GARCH models. Similar conclusions have been reached.
                                  Constant                     Coefficient                     R-Square
Control              3-Year            5-Year        3-Year          5-Year          3-Year          5-Year
Variable
INAGE                0.449**           0.545***      0.009           0.006**         0.008795        0.022341
                     (2.31)            (3.56)        (1.05)          (1.76)
                              *                ***
PUBAGE               0.674             0.574         0.007           0.012           0.010003        0.022499
                     (1.59)            (3.37)        (0.97)          (1.62)
                                                             **
BETA                 -0.217            -0.225        0.387           0.697**         0.030371        0.046127
                     (-1.07)           (-0.762)      (1.76)          (2.19)
EQUITY               0.301             -0.462        -0.05           -0.254          0.035649        0.007146
                     (0.44)            (-0.46)       (-0.26)         (0.844)
ISSYR                -38.5             17.30         0.019           -0.008          0.006284        0.000565
                     (-0.79)           (0.24)        (0.79)          (-0.24)
TOTVOL               0.461**           1.506***      0.064**         0.211***        0.000686        0.182521
                     (2.43)            (5.90)        (-1.91)         (4.70)
                              *                *
SAMPVOL              0.436             1.221         -0.132          0.069           0.003956        0.005021
                     (1.26)            (1.30)        (-0.63)         (0.22)
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                                                                                                               Page 39
                                                    TABLE 8
                    Comparison of Holding Period Returns by Age of the Issuer
A comparison of holding period returns sorted by age is presented for both the short term (3 years) and the
medium term (5 years). The HPR was defined by the equation
                                                         b            
                                            HPRi ,a:b = ∏ (1 + Ri ,t ) − 1
                                                         t =a         

The age comparison is based on deciles as well as halves. The results point to a substantial difference between
short and medium term holding period returns when oldest half (average age of 54.9 years) and youngest half
(average age 14.5 years) is compared. This differential in returns is even more pronounced when deciles are
compared. These results suggest greater vulnerability of younger firms to post-issue drop in performance as
opposed to the more robust older firms.
                                     Avg Age                        3 Years                 5 Years
                                     (years)            HPR %              Diff %   HPR %       Diff %
Oldest Decile                        88.6               -20.2                       -29.4
Youngest Decile                      4.5                -53.9              33.7     -57.6       28.2

Oldest Half                          54.9               -27.6                       -37.2
Youngest Half                        14.5               -44.7              17.1     -50.0       12.8




                                                                                                            Page 40
                                                     TABLE 9
          Summary of Issuers’ Initial Returns based on Four Measurement Definitions
This table summarises the average total opening gains accompanying a seasoned equity offering as defined by
the four measures of initial returns:
            P                      P                     (η + 1) × P0               (η + 1) × P0 
   CORERT =  0  − 1        ABSRT =  0  − 1
                                     P            DILRT = 
                                                             η × P + IP  − 1
                                                                                TOTRT = 
                                                                                          η × P + IP  − 1
                                                                                                        
             IP                     −1                        o                          −1      

It is immediately noticeable that each measure is significant at least at a 5% level. A conspicuous result in
these figures is the initial return of 163.4% measured by the CORERT variable. This is not unexpected,
however, since this is the most crude indicator that computes opening gains simply as a ratio of post-issue
share-value to the sale-price of new equity, whilst ignoring all other aspects including the proportion in which
the new equity was offered.
Measurement Method                               Initial Return                     t-statistic
CORERT                                           163.4%                             6.43***
ABSRT                                            10.9%                              2.68***
DILRT                                            4.9%                               1.66**
TOTRT                                            21.6%                              3.66***
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                                                                                                              Page 41
                                                                TABLE 10
                    Results of Cross-Sectional Regression of Initial Returns Regression on
                               Size-Matched Benchmark-Adjusted Excess Returns
Presented are the results from regressing the cumulative abnormal returns (computed relative to the size-
matched          non-issuers   benchmark)          on     the    four     measures   of    opening   gains,   according    to
       S − BM                                      S-BM
CAR   i         = α i + β i Ω i + ε i (where CAR          refers to the excess returns measured relative to size-matched

benchmark; α and β are the intercept and coefficient, respectively; Ω is the independent opening return measure
and ε represents the error term). It may be observed that DILRT is the only measure that is consistently
significant across time at minimum 5% level. In relative terms it also exhibits a high R-square value of 4.1%
and 7.1% for short and medium terms, respectively. This finding supports the contention that the post-issue
underperformance acts as a mitigator of issue costs, when those costs are proxied by the initial returns that
account for offer underpricing as well as the dilution ratio (viz DILRT definition). It should be noted that in
(unreported) results we have re-examined these findings whilst controlling for the effects of heteroscedasticity,
including computations based on ARCH/GARCH models. Similar conclusions have been reached.
                                        Constant                           Coefficient                   R-Square
Control                   3-Year             5-Year              3-Year          5-Year         3-Year         5-Year
Variable
CORERT                    -0.250**           -0.139              -0.005          -0.005         0.009123       0.00542
                          (-2.54)            (-1.07)             (-0.96)         (-0.74)
                                   ***
ABSRT                     -0.263             -0.134              -0.037          -0.206         0.001151       0.01979
                          (-2.69)            (-1.05)             (-0.33)         (-1.42)
                                   **                                     **
DILRT                     -0.218             -0.071              -0.992          -1.729***      0.040755       0.07059
                          (-2.23)            (-0.55)             (-2.06)         (-2.74)
TOTRT                     -0.260***          -0.127              -0.032          -0.134*        0.002832       0.02730
                          (-2.66)            (-0.98)             (-0.53)         (-1.67)
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                                                                                                                          Page 42
                                                  TABLE 11
              Results of Cross-Sectional Regression of Initial Returns Regression on
                 Industry-And-Size-Matched Benchmark-Adjusted Excess Returns
This table summarises the results from regressing cumulative abnormal returns (computed relative to the
industry-and-size-matched non-issuers benchmark) on the four indicators of opening gains, defined by
                                            IS-BM
CARiIS − BM = α i + β i Ω i + ε i (where CAR      refers to the excess returns measured relative to industry-and-

size-matched benchmark; α and β are the intercept and coefficient, respectively; Ω is the independent opening
return measure and ε represents the error term). Two coefficients have been detected to be significant.
Consistent with other benchmarks, DILRT is again significantly related to post-issue performance, such
relationship being consistent across time. The second measure, TOTRT, is also significant at a 5% level in the
medium term, however this is not consistent over time. It should be noted that the significance of specifically
these two measures is not coincidental as their definitions parallel but for the inclusion of market reaction in
TOTRT. It should be noted that in (unreported) results we have re-examined these findings whilst controlling
for the effects of heteroscedasticity, including computations based on ARCH/GARCH models. Similar
conclusions have been reached.
                                Constant                      Coefficient                    R-Square
Control              3-Year          5-Year         3-Year          5-Year          3-Year          5-Year
Variable
CORERT               -0.398***       -0.295*        -0.006          -0.011          0.007318        0.020750
                     (-2.93)         (-1.87)        (-0.85)         (-1.45)
                              ***             *
ABSRT                -0.401          -0.292         -0.168          -0.407          0.002418        0.053198
                     (-2.98)         (-1.90)        (-1.11)         (-1.22)
                              ***             *              *
DILRT                -0.409          -0.290         -0.225          -0.957**        0.011119        0.019877
                     (-2.96)         (-1.81)        (-1.33)         (-2.36)
                              ***             *
TOTRT                -0.401          -0.288         -0.084          -0.224**        0.010101        0.052813
                     (-2.98)         (-1.87)        (-1.00)         (-2.35)
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                                                                                                             Page 43
                                                     TABLE 12
              Results of Cross-Sectional Regression of Initial Returns Regression on
                         Market-Index Benchmark-Adjusted Excess Returns
Presented are the constants, coefficients and the corresponding t-statistics from the regression of cumulative
abnormal returns (computed relative to the AOI-matched benchmark) against the four measures of opening
returns, according to CARiAOI − BM = α i + β i Ω i + ε i (where CARAOI-BM refers to the excess returns measured

relative to market index (AOI) benchmark; α and β are the intercept and coefficient, respectively; Ω is the
independent opening return measure and ε represents the error term). As with the other benchmarks, DILRT is
once again the only measure that is consistently significant across time. TOTRT exhibits marginal significance
(at 10% level) for the short term, but this disappears in the medium term. These findings add further credence to
the proposition that the post-issue underperformance mitigates some of the issue costs, when the measure of
these costs accounts for offer underpricing as well as the dilution ratio (viz DILRT definition). It should be
noted that in (unreported) results we have re-examined these findings whilst controlling for the effects of
heteroscedasticity, including computations based on ARCH/GARCH models. Similar conclusions have been
reached.
                                  Constant                      Coefficient                  R-Square
Control              3-Year            5-Year         3-Year          5-Year        3-Year          5-Year
Variable
CORERT               0.0127*           0.390***       -0.001          -0.0015       0.001844        0.000862
                     (1.83)            (3.84)         (-0.43)         (-0.29)
                                               ***
ABSRT                0.107             0.384          -0.082          -0.005        0.000571        0.000182
                     (1.58)            (3.82)         (-0.24)         (-0.04)
                              *                ***             **
DILRT                0.126             0.407          -0.133          -0.448*       0.029892        0.018230
                     (1.78)            (3.97)         (-1.75)         (-1.49)
                                               ***             *
TOTRT                0.107             0.386          -0.068          -0.006        0.025214        0.000855
                     (1.57)            (3.83)         (-1.60)         (-0.09)
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                                                                                                             Page 44
                                                TABLE 13
           Correlation Matrix of Four Major Factors influencing post-Issue Returns
This table compiles the correlation matrix between the four variables shown to bear the most impact on the
post-issue performance of SEOs. Although quite small in absolute values, several correlations are quite
substantial in relative values. The most prominent is the correlation between the INAGE and DILRT variables
standing at -0.28049, a fairly strong negative relationship. The correlations between BETA and TOTVOL and
between BETA and INAGE also show substantial negative association between these two pairs. For this
reason a multivariate regression has also been performed and is reported in Table 14, below.
                      DILRT                 TOTVOL                 BETA                   INAGE
DILRT                 1
TOTVOL                -0.09909              1
BETA                  -0.02044              -0.11563               1
INAGE                 -0.28049               -0.05715              -0.14971               1




                                                                                                         Page 45
                                                       TABLE 14
       Results from Cross-Sectional Multivariate Regression Analysis of the Four Major
  Factors on post-Issue Excess Returns measured against the Three Benchmark Definitions.
Shown below are the results of a multivariate regression of CAR (measured against the three benchmarks) on
the four variables shown in univariate regressions to bear the most impact on the post-issue performance, but
also shown to exhibit some cross-correlation. This regression was defined by the equation

CARi = α i + β1,i Ω1,i + β 2,i Ω 2,i + ... + β n ,i Ω n ,i + ε i (where α and βx’s are the intercept and factor coefficients,

respectively; Ωx’s are the independent control variables and ε represents the error term). With the exception of
TOTVOL under size-matched and industry-and-size-matched benchmarks, all coefficients continue to exhibit
significance at a minimum 10% level. A strong significance of Adjusted R-square at a 1% level across all
benchmarks also points to a good selection of the chosen factors. It should further be noted that again BETA
has the strongest and most robust influence as measured by the significance of the corresponding coefficient
against each benchmark. This multivariate regression was also re-examined controlling for the effects of
heteroscedasticity, including computations based on ARCH/GARCH models. The (unreported) results draw
identical conclusions.
                                                                    Benchmark
                       Size-Match                         Ind-and-Size Match                 AOI
DILRT                  -1.767***                          -1.441**                           -0.656**
                       (-3.08)                            (-1.99)                            (-1.83)
TOTVOL                 -0.017                             0.802                              0.285**
                       (-0.05)                            (1.03)                             (1.66)
                               ***                                **
BETA                   1.319                              1.091                              0.780***
                       (3.59)                             (2.12)                             (2.74)
                                *                                  **
INAGE                  -0.003                             -0.014                             -0.006*
                       (-1.63)                            (-2.26)                            (-1.67)
              2
Adjusted R             0.181                              0.117                              0.101
                               ***                                ***
F-Statistic            6.123                              4.084                              3.612***
* Significant at 10% level
** Significant at 5% level
*** Significant at 1% level




                                                                                                                          Page 46
                                         REFERENCES



Allen D.E. and Patrick M., Some Further Australian Evidence on the Long-Run Performance of Initial

Public Offerings: 1974-1984, Pacific Basin Financial Markets, Volume 2A, JAI Press Inc., 1996:133-

155



Ali A., Bias in analysts’ earnings forecasts as an explanation for the long-run underperformance of

stock following equity offerings, Working Paper, University of Arizona



Asquith P. and Mullins D., Equity Issues and Offering Dilution, Journal of Financial Economics,

1986:61-89



ASX, The Australian Stock Exchange Yearbook 1998, Australian Stock Exchange, Sydney, 1998



ASX, The Australian Stock Exchange Investor Handbook, Australian Stock Exchange, 1998 Ed.,

Brisbane, 1997



Beatty R.P. and Ritter J.R., Investment Banking, Reputation and the Underpricing of Initial Public

Offerings, Journal of Financial Economics, 1986:213-232



Brous P.A., Common Stock Offerings and Earnings Expectations: A test of the Release of

Unfavourable Information, Journal of Fiance, 1992: 1517-1536



Cheng L.L., The motives, timing and subsequent performance of seasoned equity issues, PhD

Dissertation, University of Massachusetts, 1995



Cheung Y.W. and Lai K.S., Lag order and critical values of the Augmented Dickey-Fuller test, Journal

of Business and Economic Statistics, 1995:277-280




                                                                                               Page 47
Conrad J. and Kaul G., Long term market overreaction or biases in computed returns?, Journal of

Finance, 1993:39-63



Cunningham S., Predictability of British Stock Market prices, Journal of Royal Statistical Society

Series C, 1973



DeAngelo H. and Masulis R., Optimal Capital Structure under Corporate and Personal Taxation,

Journal of Financial Economics, 1980:3-29



Eckbo E.B., Masulis R.W. and Norli ∅., Seasoned Public Offerings: Resolution of the ‘New Issues

Puzzle’, http://mba.tuck.dartmouth.edu/pages/faculty/espen.eckbo/docs/resear, 1999



Enders W., Applied Econometric Time Series, John Wiley and Sons, New York, 1995



Friend I. and Longstreet J.R., Price Experience and Return on new stock issues, Investment Banking

and the New Issues Market, World Publishing Co., Cleveland, 1967



Gerard B. and Nanda V., Trading and Manipulation around Seasoned Equity Offerings, Journal of

Finance, 1993:213-245



Hansen R.S. and Crutchley C., Corporate Earnings and Financing: An Empirical Analysis, Journal of

Business, 1990:347-371



Healy P.M. and Palepu K.G., Earnings and Risk Changes Surrounding Primary Stock Offers, Journal

of Accounting Research, 1990:25-48



Ibbotson R.G. and Jaffe J.F., Hot Issue Markets, The Journal of Finance, 1975:1027-1042



Jagadeesh N., Weinstein M. and Welch I., Initial Public Offerings and Subsequent Equity Offerings,

Journal of Financial Economics, 1993:153-175




                                                                                              Page 48
Jensen M., The Modern Industrial Revolution, Exit, and the Failure of Internal Control Systems,

Journal of Finance, 1980:831-880



Korajzcyk R.A., Lucas D. and McDonald R., Understanding Stock Price behaviour around the time of

Equity Issues, University of Chicago Press, Chicago, 1991



Leland H. and Pyle D., Informational asymmetries, financial structure and financial intermediaries,

Journal of Finance, 1977:371-388



Lim K.S., The Performance of Issuing Firms, Journal of Business, February 1986:775-801



Loughran T., Ritter J.R. and Rydqvist K., Initial Public Offerings: International Insights, Pacific-Basin

Finance Journal, 1994:165-199



Loughran T. and Ritter J.R., The New Issues Puzzle, Journal of Finance, 1995:23-51



Loughran T. and Ritter J.R., The Operating Performance of Firms conducting Seasoned Equity

Offerings, Journal of Finance, December 1997:1823-1850



Madala G.S., Introduction to Econometrics, 2nd Ed., MacMillan Publishing Company, New York,

1992



Masulis R., The Impact of Capital Structure Change of Firm Value: Some Estimates, Journal of

Finance, 1983:107-126



Masulis R. and Korwar A., Seasoned Equity Offerings: An empirical investigation, Journal of Financial

Economics, 1986:91-118




                                                                                                    Page 49
McLaughlin R., Safieddine A. and Vasudevan G.K., The Operating Performance of Seasoned Equity

Issuers: Free Cash Flow and Post-Issue Performance, Financial Management, 1996:41-53



Mikkleson W. and Partch M., Valuation Effects of Security Offerings and the Issuance process, Journal

of Financial Economics, 1986:31-60



Miller M. and Rock K., Dividend Policy under Asymmetric Information, Journal of Finance,

1985:1031-1051



Myers S. and Majluf N., Corporate Financing and Investment Decisions When Firms Have

Information that Investors Do Not Have, Journal of Financial Economics, 1984:187-221



Mustow D., The long run performance of Initial Public Offerings and some factors affecting that

performance, Honours Dissertation, University of Melbourne.



Patel A., Emery D.R. and Lee Y.W., Firm performance and Security Type in Seasoned Offerings,

Journal of Financial Research, 1993:181-193



Rangan S., Earnings Management and the Performance of Seasoned Equity Offerings, Journal of

Financial Economics, 1997



Ritter J.R., The Hot Issue Market of 1980, Journal of Business, 1984:215-240



Ritter J.R., The long-run performance of initial public offerings, Journal of Finance, 1991:3-27



Rock K., Why new issues are underpriced, Journal of Financial Economics, 1986:187-212



Teoh S.H., Welch I. and Wong T.J., Earnings Management and the underperformance of Seasoned

Equity Offerings, Journal of Financial Economics, 1997




                                                                                                   Page 50
Schipper K. and Smith A., Comparison of Equity Caveats and Seasoned Equity Offerings: Share Price

Effects and Corporate Restructuring, Journal of Financial Economics, 1986:153-186



Spiess D.K. and Affleck-Graves J., Underperformance in long-run stock returns following seasoned

equity offerings, Journal of Financial Economics, 1995:243-267



Stigler G., Public regulation of security market, Journal of Business, 1964:117-142



Welch I., Seasoned offerings, imitation costs, and the underpricing of initial public offerings, Journal

of Finance, 1989:421-449




                                                                                                   Page 51

								
To top