Rights Offerings, Renounceability and Underwritten Status

Transcription

1 0 Rights Offerings, Renounceability and Underwritten Status Balasingham Balachandran - Monash University Robert Faff - Monash University Michael Theobald - University of Birmingham Acknowledgments: The authors gratefully acknowledge the helpful comments of Sugato Chakravarty, Charles Corrado Alistair Marsden, seminar participants at Monash University and particularly Keryn Chalmers; Alan Ramsay and Michael Skully. We are also pleased to acknowledge the research assistance provided by Tuan Anh Nguyen, Eswaran Velayutham, Hanghang Tu, Berty Vidanapathirana, Vineet Tawani and Ke Xu and funding provided by an ARC Discovery grant (DP ). Address for correspondence: Dr Balasingham Balachandran, Department of Accounting & Finance, Faculty of Business & Economics, P.O. Box 197, Caulfield East, Melbourne, Victoria, 3145, Australia. FAX No. (61) Bala.Balachandran@BusEco.Monash.edu.au Abstract Rights offerings in Australia can be renounceable or non-renounceable. Exploring the articulation between this renounceability dimension and underwriting provides new insights into the literature on rights issues. Rights offerings are analysed from two perspectives market reactions to rights announcements and the factors driving the choice of issue type. Rights offerings are partitioned into a range of alternative classifications based upon renounceability, underwritten status and expected takeup, and a number of testable hypotheses are developed. Evidence is provided to support the relationship between quality signals and we find that fully underwritten provides a quality signal. We have also developed an enhanced version of the expected takeup variable of Eckbo and Masulis (1992) and Bohren et al (1997), and find support for our adapted signalling model of Heinkel and and Schwartz (1986). Managers expecting (i) lower takeup use non-renounceable with fully underwriting, and (ii) higher takeup select renounceable with fully underwriting to signal their superior quality. JEL Classification code: G14, G32

2 1 Rights Offerings, Renounceability and Underwritten Status 1. Introduction High quality firms, seeking to raise new capital via rights offerings, signal their quality by choosing underwritten equity issues (Heinkel and Schwartz, 1986). The ability to choose whether rights issues are renounceable or non-renounceable 1 provides management of Australian firms with the opportunity to considerably enhance the signal of quality in the context of the asymmetric information models developed by Heinkel and Schwartz (1986) and Eckbo and Masulis (1992). 2 Accordingly, our study investigates the unique interaction of these two characterizing features of Australian rights offerings; namely, their renounceability and their underwritten status. By focusing upon the renounceability and full/partial underwriting dimensions, we are able to provide new empirical evidence and to develop further insights into existing theories regarding the choice of rights offerings. We focus upon the way in which the market reacts to these differing dimensions and we also investigate the determining factors for particular types of rights issue. Heinkel and Schwartz (1986) predict that high quality firms which select underwritten rights issues (involving underwriter investigations) should generate the least unfavourable market reaction on the rights announcement. Low quality firms would choose firm commitment offerings, which should generate the most unfavourable market reaction, and intermediate quality firms would select non-underwritten (uninsured) rights issues generating moderately 1 Renounceable rights issues allow existing shareholders who do not wish to take up their entitlement of new shares, to sell part or all of their rights to the issue. In contrast a non-renounceable rights issue does not permit shareholders to sell the right any unused entitlement is forfeited. New Zealand companies also issue both renounceable and non -renounceable rights. 2 Non-renounceable issues in Australia are different from open offers in the UK. In the case of the latter, an arranging bank or broker negotiates privately with investing institutions over several days before the public announcement so that come the announcement day the arranger has a list of placees who have verbally agreed to buy blocks of shares (in the event that they are not taken up by existing shareholders). Potential placees agree to become insiders and are obliged not to trade in the issuer s shares until after the announcement. The new shares are offered pro rata to existing shareholders, as in the rights issue, and if taken up they are said to be clawed back from placees (Armitage, 2002, p. 1247).

3 2 unfavourable price reaction. There is, as yet, no empirical support for the prediction that superior quality firms would be more likely to use underwritten rights issues than non-underwritten issues. Our study adapts and uses the Heinkel and Schwartz (1986) framework to analyse the impact of rights issue announcements in Australia, where rights issues are very popular, and provides good support for the predictions deriving from this enhanced version of their model. The differences in impact between underwriters fully certifying and incompletely certifying issues are developed by Booth and Smith (1986). In this modelling context, there should be no certification related abnormal returns with full certification, while in the case of incomplete certification there will be a negative reaction. Eckbo and Masulis (1992) argue that managers and shareholders have asymmetric information with regard to firm value which influences expectations about the willingness of existing shareholders to participate in equity offerings, thereby determining the flotation method. They argue that managers expecting low shareholder participation, choose firm commitment offerings and retain underwriters to certify firm quality. If shareholders are effective monitors and they represent a cost efficient source of financing, managers of undervalued firms issue non-underwritten rights, since they expect the offered shares to be fully taken up. Prior studies on rights issues have produced conflicting results. Most notably, Eckbo and Masulis (1992) examined a clean sample of 135 underwritten rights issues and 57 nonunderwritten rights issues of US firms and find significantly negative price reactions for underwritten rights issues and insignificant price reactions for non-underwritten rights issues for the two day announcement period. Bohren et al (1997) report that Norwegian firms (using a sample of 200 issues) experienced positive price reactions to non-underwritten rights issues and insignificant reactions to underwritten issues for the two day announcement period, while Cronqvist and Nilsson (2005) show that Swedish firms (using a sample of 160 issues) had insignificant price reactions for both underwritten and non-underwritten rights issues for the three day announcement period. Slovin, Sushka and Lai (2000) find significantly negative price

4 3 reactions for both underwritten and non-underwritten rights issues made by British firms for the two days announcement period. Marsden (2000) finds a significantly negative price reaction for a clean sample of 62 renounceable underwritten rights issues and a significantly positive price reaction for a clean sample of 26 uninsured renounceable rights issues in New Zealand. 3 Australian evidence, while now somewhat dated, reveals a similar conflict. Ball et al. (1977) found a significant positive announcement period abnormal return in the announcement month, while in a later unpublished Australian study, Dehnert (1993) reported a significantly negative abnormal return for the two-day announcement period. The mixed empirical evidence regarding rights issues may reflect variations in institutional characteristics across different countries. These differing institutional characteristics and features could, when combined with robust theories and well-specified hypotheses, provide strong insights into the rights issue type choice. In countries like Australia, major changes in regulations and taxation systems, for example, are likely factors inducing time-series effects in addition to these cross-sectional impacts. Accordingly, further updated research is warranted, with particular emphasis on the renounceability and underwriting dimension. Heinkel and Schwartz (1986) argue that if the rights issue fails, the terminal stock price will be less than the subscription price at the rights issue expiration date. Thus, a variable of interest in the context of rights issue quality is the market reaction for the period from the announcement date to rights offer expiry date. No previous studies have examined the market reaction for the full period from the announcement date to offer expiry date to evaluate the impact of the quality of the rights issue. However, a few studies document the price reactions for various windows within the period from the announcement date to the rights offer expiry date (see for example, Eckbo and Masulis, 1992; Bohren et al, 1997). As such, based on information provided in these papers, an estimate can be made of the magnitude of the average price reaction 3 Marsden (2000) focused on a clean sample of 88 renounceable issues in New Zealand, taken from a bigger group of 570 issues over the period 1976 to His exclusions include 105 non-renounceable issues. This exclusion is notable since we retain this type of rights issue as an important focus in our study.

5 4 for the period from the announcement date to the offer expiry date. From this we deduce that the reaction in the Eckbo and Masulis (1992) study is stronger for underwritten rights issues than for non-underwritten rights issues and thereby provides some support for the Heinkel and Schwartz (1986) model. However, further analysis is warranted to confirm the statistical robustness of these reactions in a rigorous experimental setting. Prior to the 1960s, rights issues were the dominant method of seasoned equity offerings in the US (Stevenson, 1957). They have now become rare and replaced by firm commitment offers (particularly public offers). Indeed, Eckbo and Masulis (1992) document that firm commitment offers comprised 99% of seasoned equity offers made by US industrial firms by As such, it might seem that an understanding of the financial lessons drawn from rights offerings is of little interest particularly given such evidence must, of necessity, involve non- US settings. However, the recent work of Bohren et al (1997); Slovin et al (2000) and more recently Cronqvist and Nilsson (2005) suggest otherwise. For example, Slovin et al (2000) argue: Given the similarity of American and British corporate finance, differences in the practices used in each country to price and market seasoned equity provide a setting that facilitates an examination of the effects of equity issuance and adverse selection on firm value. In an analogous fashion to the Norwegian, UK, Swedish and Hong Kong settings, rights issues have remained relatively popular in Australia over recent periods. The critical question is: what new and incremental insights will an Australian study deliver? At one level our answer echoes the sentiment of Slovin et al (2000), above. That is, as in the UK, Australia has a strong similarity to the US in terms of corporate finance, but differences in practices and institutional features create an opportunity to glean empirical and theoretical insights, thereby increasing the general understanding of financial phenomena. As such, an important part of the answer to this question stems from the existence of two very distinct types of equity rights offerings in Australia: renounceable and non-renounceable rights. Renounceability gives shareholders a valuable option in a world of transactions costs and other market frictions.

6 5 Moreover, our sample also gives rise to a dimension relating to their underwritten status by interacting with renounceability we have a two by three combination of cases: (a) fully underwritten/renounceable; (b) partially underwritten/renounceable; (c) non underwritten/ renounceable; (d) fully underwritten/non-renounceable; (e) partially underwritten/nonrenounceable; and (f) non-underwritten/non-renounceable. Collectively, a comparative analysis across these different scenarios provides deeper insights for this literature since the differing combinations will provide varying signal potential. Several basic questions are of concern here, namely: (a) do Australian shareholders also suffer a loss in their wealth when companies announce rights issues; (b) what are the underlying factors that drive investors behaviour when they react to these announcements?; (c) do market reactions vary between non-renounceable and renounceable issues and among fully underwritten, non-underwritten issues and partially underwritten rights issues and, if so, for what reason(s)?; (d) what are the underlying factors that drive managers choice between non-renounceable and renounceable rights issues?; (e) what are the underlying factors that induce managers to fully underwrite rights issues?; and, finally, (f) what are the factors that determine the rights offer price discount? The key motivations for undertaking research on the impact of rights issue announcements on shareholders wealth in Australia are threefold. First, since the time of the introduction of the imputation tax system in 1987, 4 the incentive for firms to raise equity (via rights issues, placements, and dividend reinvestment plans) in preference to debt and retained earnings, has increased. An important reason for this equity focus is that resident investors are better off with imputation credits and, hence, higher dividend payments (see Twite, 2001). Therefore, conducting a study based on Australian data will contribute to the existing literature by either confirming or opposing the current state of financial knowledge. Second, a relatively large number of rights issues in Australia have been announced in the non-renounceable form 4 Under the imputation tax system, Australian resident shareholders are eligible to obtain imputation credits for income tax paid by companies to reduce their personal income tax obligation.

7 6 and this makes them very different from rights issues made in other countries. Therefore, Australian evidence will provide insights into the wealth impact of non-renounceable rights issue announcements and whether this impact is different to that of renounceable issues. Finally, the Australian market is one characterized by companies that typically have a closely held shareholding that is, it is a market in which there is a high level of ownership concentration. 5 Conducting a study in the Australian context will, therefore, extend our knowledge as to whether or not a company s level of ownership concentration (Claessens, Djankov, Fan and Lang (2002), Shleifer and Vishny (1986) plays any role in determining the effect of rights issue announcements upon share prices. We modify the Heinkel and Schwartz (1986) model by partitioning the firms with underwritten rights issues into fully underwritten rights issues (high quality firms) and partially underwritten rights issues (low quality firms). We argue that high quality firms are willing to pay the investigation costs to the underwriter and thus secure a fully underwritten rights issue. In contrast, low quality firms are unwilling to pay the total investigation costs (to facilitate a fully underwritten rights issue), as investigation costs may be high and/or the underwriter may not be willing to insure all the shares in the issue (due to high risk). This results in partial underwriting. Thus, we expect the most (least) unfavourable market reaction to occur for partially (fully) underwritten rights issues, with a moderately unfavourable reaction relating to non-underwritten rights issues. 6 The empirical results documented in this paper provide support for these predictions. Eckbo and Masulis (1992) and Bohren et al (1997) analyse the impact upon issue decisions of the expected take-up of the issue. We argue that the take-up measure employed in Bohren et al (1997) is misspecified and not appropriate in Australian setting and, hence, we generate a robust and more appropriate measure of the expected take-up variable. In the context 5 The top twenty shareholders own more than 50% of the outstanding shares in 70% of the Australian companies examined in this study. 6 We fully develop and explain this argument in following section of our paper.

8 7 of Australian rights issues, we find results that are somewhat different to those reported in Bohren et al (1997) but which are consistent with the predictions deriving from our quality signal perspective. The plan of this paper is as follows. Section 2 discusses the institutional background in Australia. Section 3 discusses the underlying theory and development of hypotheses. Section 4 discusses the research design and methodology used to analyse the impact of rights issue announcements in Australia. Section 5 presents the empirical results using a standard event study methodology and cross-sectional regression analysis. In Section 6 we analyse the factors that drive the choice of the rights issue type using probit analysis. Section 7 discusses the relationship between subscription price discount, risk (idiosyncratic risk) and ownership concentration. Finally, the conclusions are presented in Section Institutional Background The Australian equities market is the 8 th largest national market in the world by market capitalisation, representing 2.08% of the global market in 2003 in contrast to the US which accounts for about half (see, ASX, Exposure Draft, 2003). The percentage of issued capital that may be issued without shareholders approval is 20%, except for public offers for cash or any bona fide private financing at the NYSE. There is no element of pro rata participation under the NYSE framework. 7 London stock exchange (LSE) listing rules require that unless shareholders otherwise permit, entities proposing to issue equity securities for cash must first offer those securities to existing shareholders. Prior approval is required for transactions involving an increase exceeding 25% to any of a company s assets, profits, turnover, market capitalisation or gross capital at LSE. That is, large rights issues have to be approved at extra ordinary general meeting or annual general meeting. In Sweden, rights offerings have to be approved by simple majority vote at a shareholders meeting whereas private placements have to be approved by 7 In the US there was originally a presumption of pre-emptive rights. US firms began removing pre-emptive rights clauses from corporate charters following the Great Depression (Ursel and Trepanier, 2001).

9 8 shareholders representing two-thirds of votes and capital (Cronqvist and Nilson, 2005). In Norway, the seasoned equity offerings process starts with management requesting shareholder authorisation for the issue and approval may be given for a specific amount to be raised within the following year or until the next general meeting (Bohren et al, 1997). In Australia, companies are not allowed to issue more than 15% of capital without the prior approval of shareholders. 8 However, prior approval is not required for an issue to holders of ordinary securities made under a pro rata issue and to holders of other equities to the extent that the terms of the issue of the equity securities permit participation in the pro rata issue. Rights issues in Australia are made on pro rata basis and as such shareholders approval is not required. Australia is one of few countries worldwide that allows firms to make a choice regarding whether a rights issue is renounceable or non-renounceable. 9 A renounceable rights issue allows existing shareholders who do not wish to take up their entitlement of new shares, to sell part or all of their rights to the issue. In contrast, a non-renounceable rights issue does not permit shareholders to sell the right any unused entitlement is forfeited. This renounceability feature is an important institutional difference from the US setting a difference that we seek to exploit in the current paper. In undertaking a rights issue in Australia, a company and, if the rights issue is underwritten, the underwriter will, in practice, usually decide on structural matters, such as the price at which new shares are offered, the number of shares offered and whether the issue will be renounceable. If the rights issue is underwritten, the role of the underwriter is to guarantee that the funds sought by the company will be raised. The agreement between the underwriter and the company is set out in a formal underwriting agreement. Typical terms of an underwriting arrangement require the underwriter to subscribe for any shares offered but not taken up by shareholders. The underwriting agreement will normally enable the underwriter to terminate its 8 Placements do not require shareholder approval if the amount to be raised is less than 15% of the company issued shares. 9 New Zealand companies also have this option.

10 9 obligations in defined circumstances. Some companies use principal underwriters and principal underwriters use sub-underwriters to certify the rights issues. Principal underwriters pass a major part of the risk to the sub-underwriters. Underwriters and sub-underwriters can be financial institutions, stock-brokers, major shareholders of the company or other related or unrelated parties. Companies adopt a number of strategies to deal with a shortfall in a rights issue rather than solely requiring the underwriter to subscribe for the initial shortfall. Major examples of such strategies are shortfall facilities and back end book builds. A shortfall facility allows other shareholders to subscribe for any shares not taken up by shareholders under the rights issue. Where the rights issue is underwritten, this participation will usually be resolved in advance of determining the shortfall available to the underwriter. A back-end bookbuild is a bookbuild conducted in respect of the shortfall in a rights issue (a bookbuild being an offer of securities to investors - typically institutions - for which bids are sought from the investors and the allotments and issue price are determined based on those bids). 10 A back-end bookbuild is another method of dispersing widely the rights not taken up under a rights issue and minimising the chance of the issue having a control effect. A shareholder who subscribes for shares under the rights issue may increase its voting power if the issue is not fully subscribed and it is not fully underwritten, or where that shareholder is also an underwriter or sub-underwriter (see, Takeover Panel Guidance Note 17: Rights Issues). The ability to sell renounceable rights makes it easier for shareholders to access the benefits of a rights issue and assist in ensuring equality and reasonableness of opportunity per Section 602(c) of the of the Corporations Act A company considering a non-renounceable rights issue has to consider structuring a rights issue in such a way to avoid other factors which would tend to increase the risk of the rights issue having an effect on control or potential control 10 In this case, companies must make the issue within three months after the close of the offer, and the directors must have stated as part of the offer that they reserve the right to issue the shortfall at their discretion. The issue price must not be less than the subscription price.

11 10 of the company. 11 If the Takeover Panel finds (from direct or circumstantial evidence, including the structure of the offer) that the rights issue has been structured for the purpose of affecting control of the company, it will be less ready to accept the company's judgement that a rights issue was an appropriate method of raising the required funds or that aspects of the structure of the issue such as price, timing and renounceability were in fact reasonably necessary for fundraising purposes. The Australian Stock Exchange (ASX) requires all public companies that make rights issue announcements to lodge their prospectuses with the Australian Securities and Investments Commission (ASIC) and the ASX. Notably, the subscription price for new shares offered under a rights issue by Australian companies is made public at the time of the announcement. This contrasts with the US where the offer price is typically only finalized just prior to the subscription period. Prospectuses must contain the following information: type of rights issue; whether renounceable or non-renounceable; entitlement ratio for new shares offered to existing shareholders on a pro-rata basis; total amount payable for the new shares, subscription price for new shares, identity of the underwriter, broker (if any); underwriters fees, ex-right date, trading period for the renounceable rights issues, as well as the record date and payment date. This is similar to the UK (see Slovin et al, 2000, p. 162). According to the ASX listing rules, companies should (a) announce renounceable and non renounceable issues at least 6 business days before the record date, (b) send notices to shareholders containing information regarding details of the time table of the issue, and a statement that the prospectus and offer information statement has been lodged with the ASX and is available on the company s website at least 5 business days before the record date, (c) set the 11 For example, Bains Harding Ltd announced a 1 for 1 non-renounceable rights issue at 15 cents on 9 April 2001 (10 days before the announcement the market price was 15 cents versus 22 cents three moths before the announcement), which is underwritten by its largest shareholder KAEFER Isoliertechnik GmbH who has a 21.65% holding at the balance sheet date immediately prior to the announcement of the rights offer. Ultimately, KAEFER took 96% of this issue as an underwriter and shareholder, thereby controlling 48.62% shares immediately after the rights issue.

12 11 Ex date 4 business days before the record date, (d) send prospectuses, offer information statement and acceptance forms to persons entitled no more than 4 business days after the record date, (e) set closing of acceptances at least 10 business days after the company announces that the despatch of the prospectus and offer information statement has been completed, (e) the end of rights trading in the case of renounceable issues, and (f) disclose the extent of under subscription to the ASX within three business days of acceptances closing. Shareholder takeup disclosed under this rule includes shares taken up by existing shareholders (i.e. shares taken up by existing shareholders on their pro rata basis and those shares subscribed by existing shareholders using the shortfall facility) and in the case of renounceable issues shares subscribed by the person who bought the rights. 3. Theory and Hypothesis Development 3.1 Asymmetric Information Theories Myers and Majluf (1984) developed an equilibrium model in which stock issues cause a decline in price. Their model assumes that the firm s investment requirements are fixed and known by all investors, and that the firm raises either that specific amount of equity or none at all. Krasker (1986) developed a two period model that generalised the Myers and Majluf (1984) structure by modelling the issue size as a continuous choice variable. Krasker s (1986) model predicts that the stock price will be negatively correlated with issue size and argues that investors must interpret stock issues unfavourably and indeed, must interpret larger issues more unfavourably than smaller ones. Krasker s model strengthens the Myers and Majluf s conclusion that there is an adverse selection problem associated with the issuance of risky securities, whether debt or equity. Booth and Smith (1986) develop a model for rationalising the issue process within an asymmetric information setting in terms of underwriter certification, distinguishing between full and incomplete certification and the differential price reaction effects. Lucas and McDonald (1990) developed a dynamic, infinite horizon model of the firm s equity issue decision under adverse selection.

13 12 Heinkel and Schwartz (1986) developed a signalling model in which the method of external financing conveys information regarding the quality of the firm. They argue that there is information content in the choice of financing method made by the firm and the choice reveals, either partially or totally, the quality of the firm. Specifically, their rational expectations, partially revealing information equilibrium model predicts that (a) high quality firms signal through rights issues with standby agreements i.e. such firms are willing to pay the fixed investigation costs associated with standby agreements; (b) intermediate quality firms signal through the subscription price of an uninsured rights issue; and (c) lower quality firms remain indistinguishable to investors by selling their shares to an uninformed underwriter at a fixed price (firm commitment offer); the underwriter then resells the shares to investors at that price plus the underwriter s fixed underwriting expense. According to the Heinkel and Schwartz (1986) signalling model, the highest quality firms are willing to pay the exogenous, fixed investigation costs associated with a standby rights offer and will find it less disadvantageous to incur such investigation costs. For these firms, the correctly priced insurance provided by the standby agreement allows them to be indifferent to the choice of subscription price. The intermediate quality firms are willing to pay the endogenously determined signalling costs (which are unobservable when the financing choice is made and therefore difficult to measure empirically) incurred with an uninsured rights offer by setting their subscription price at a level such that there is a significant probability of failure of the offer thereby revealing the quality of the firm. That is, upper bound intermediate quality firms distinguish themselves from lower bound intermediate quality firms in choosing a relatively higher subscription price relative to the current share price. Eckbo and Masulis (1992) develop an analytical framework to explain a firm s choice of equity flotation method. They suggest that the choice between uninsured rights, insured rights and firm commitment underwritten offers depends on information asymmetries, shareholder characteristics and direct flotation costs. Further, they suggest that the asymmetric information

14 13 between managers and shareholders influences expectations about the willingness of existing shareholders to participate in equity offerings which, in turn, determine the flotation method. Managers expecting low shareholder participation choose firm commitment offerings and retain underwriters to certify firm quality. If shareholders are effective monitors and represent a cost efficient source of financing, managers of undervalued firms issue uninsured rights, since they expect the offered shares to be fully taken up - underwriting by the arranger provides imperfect certification that the issuer is not overvalued, which enables the issuer to sell shares at a higher price. The value of this certification benefit is negatively related to the proportion of the issue expected to be taken up by existing shareholders, so that a higher take-up implies a lower probability that the issue will be underwritten. The predictions regarding market reaction depend on how reliable certification is and the relationship between the positive effect of underwriting on market price and the negative effect of a low takeup in underwritten offers. Conditional on the expected current shareholder takeup, they argue that under- and over-valued issuers of stock choose between uninsured rights, insured rights and firm commitment offers, and the market responds to these choices by a zero reaction to uninsured rights, most negatively to firm commitment offers and less negatively to insured rights. They suggest that US firms with dispersion in ownership choose firm commitment offers while firms in other markets with high ownership concentration use the rights issue method. 3.2 Hypothesis Development General The models that were discussed in Section 3.1, above, suggest that firms will attempt to signal their quality via the issuance route selected where asymmetric information exists. One strong implication of these models was that high quality firms would signal this quality in the presence of such asymmetries by ensuring that the issue were fully underwritten. This is consistent with the modelling and implications developed in Heinkel and Schwartz (1986), wherein such high quality firms are willing to incur the total investigation and other related costs associated with

15 14 full underwriting, given the associated and significant signalling benefits that accrue. Similarly, equating full certification and full underwriting leads to a corresponding prediction in the context of the Booth and Smith (1986) model. Effectively, then, we have the prediction that high (non-high) quality firms will employ full (less than full) underwriting arrangements. However, richer hypotheses are possible by analysing further the scenario wherein the issue is not fully underwritten. As in Heinkel and Schwartz (1986), this scenario may be partitioned into two classifications namely, partially underwritten and non-underwritten. Heinkel and Schwartz (1986) demonstrate that, within a partially revealing equilibrium, low quality firms will attempt to disguise their quality by utilising partially underwritten issues. Furthermore, intermediate quality firms would signal via the subscription price with the implication that, if this is the primary signalling device and with their lesser need for disguise, they would more likely opt for uninsured rights issues. By utilising the Heinkel and Schwartz (1986) structure, then, there will be three hypotheses available for investigation and testing. That is, based on this firm quality perspective, we expect the most unfavourable market reaction to partially underwritten rights issues, the least unfavourable reaction to fully underwritten rights issues and an intermediate level unfavourable reaction to non-underwritten rights issues. The resultant hypotheses are, then: Hypothesis 1: High quality firms will use fully underwritten rights offerings (either insured or standby) to signal their quality and will experience the least unfavourable price reactions around rights offering announcement dates. Hypothesis 2: Intermediate quality firms will use non-underwritten rights issues and will experience intermediate degrees of unfavourable price reaction around rights offering announcement dates. Hypothesis 3: Low quality firms will use partially underwritten rights issues and will experience the most unfavourable price reaction around rights offering announcement dates.

16 15 Note that within hypotheses 1-3 there are two aspects the first regarding the drivers of issue type choice and the second relating to market reactions. Rights offer failure is costly and the rights issuer who privately expects a fall in stock price over the rights offer period selects a low offer price relative to the current (uninformed) market price in order to prevent the offer from failing. The underwriter will incur the direct costs of certification of the rights issues only to the point where the marginal cost of certification equals the marginal benefit so that shareholder take up is maximized subject to the constraint that the issue is not over-priced. Booth and Smith (1986) argue that if the underwriter certifies a lower price to protect or help establish his reputation or uses discounting as a substitute for full certification, a drop in value followed by either normal or positive abnormal returns after the last day of the acceptance is expected. We argue that if the underwriter who agrees to certify the rights issue expects a lower takeup from existing shareholders, that underwriter will demand higher investigation costs and/or impose a higher issue price discount to protect them from offer failure. Marsh (1980) argues that as an alternative to underwriting, companies can reduce the risk of a failed issue equally effectively by setting the issue price sufficiently low. Bohren et al (1997) suggest that in the absence of information asymmetries, an underwriter guarantee and a deep rights issue discount are substitute mechanisms for ensuring a full subscription to the rights offer. Thus, we argue that issuers of rights offerings in Australia will use two mechanisms to increase shareholder takeup: underwriting the rights issue and/or issue price discount. The following table summarises the quality of the firms versus underwriting. Quality High Intermediate Low Issue Type Fully Underwritten Non-Underwritten (Uninsured) Partially Underwritten Fully underwritten issues with high discounts would be associated with lesser quality firms, while lower discounts on uninsured or partially underwritten issues would be associated with firms at the higher ends of the quality ranges normally using these issue types. That is, as

17 16 suggested in Heinkel and Schwartz (1986), an intermediate quality firm using an uninsured issue could signal quality via a higher subscription price. The corresponding hypothesis is, then: Hypothesis 4: The quality of firms using the fully underwritten/partially underwritten/non-underwritten rights offer will be negatively correlated with the subscription price discount. Thus we expect a negative relationship between subscription price discount and announcement period abnormal return Expected Takeup, Quality Certification and Renounceability Myers and Majluf (1984) argue that managers, who are assumed to maximise the intrinsic value of the firm s shares, elect to issue and invest if and only if the net issue benefit is nonnegative. That is, when b-(c+f) 0, where, b is the value of the project, c is cost or benefit of selling undervalued or overvalued stock, and f is the flotation costs. Eckbo and Masulis (1992) argue that (i) managers select the floatation method that maximises the net issue benefit (b-(c+f)) given the expected current shareholder takeup and recognising that the equilibrium value of c itself depends on both the value of current shareholder takeup and the flotation method choice; (ii) managers treat takeup as an exogenous factor as this is determined by factors largely beyond managerial control, such as personal consumption and wealth constraints, diversification benefits and portfolio rebalancing costs, benefits from maintaining a shareholder s current voting and dividend rights, and individual tax brackets; (iii) managers have better information than the market about current shareholders expected takeup; and (iv) underwriter certification narrows, but does not eliminate the information asymmetry between the firm and market. They conclude that the equilibrium value of c depends on the market s estimate of current shareholders expected takeup, conditional on the information in the offer announcement; and therefore, (i) undervalued firms with higher current shareholders expected takeup will choose nonunderwritten rights issues or opt for underwritten issues if the sum of the expected certification

18 17 benefit and the net project value exceeds the underwriting fee, and (ii) overvalued firms with higher current shareholders expected takeup will choose non-underwritten issues informing the market of the higher expected takeup or conceal the information on expected takeup and select a firm commitment offer to exploit the imperfect certification process. According to Heinkel and Schwartz (1986), there is information content in the choice of a financing method and the choice reveals, either partially or totally, the quality of the firm. If the rights issue fails, the terminal stock price will be less than the subscription price at the rights issue expiration date. A critical assumption of their model is that sufficient time passes between the setting of the rights offer subscription price and expiration of the right for there to be significant stock price movement (Heinkel and Schwartz, 1986, p. 17). Thus, the variable of interest is the price reaction for the period from the date of the announcement of the equity issue to the equity offer expiry date. Our calculations reveal an average price reaction in the US, from the day before the announcement date to day of the offer expiration, of % for nonunderwritten rights issues of industrial firms compared to -3.52% for underwritten rights issues of industrial firms and -2.67% for the firm commitment offers of industrial firms (based on information reported by Eckbo and Masulis, 1992). However, Eckbo and Masulis (1992) report the average direct flotation costs as a percentage of market value of common equity is 1.05% for firm commitment offers, 0.93% for underwritten rights issues and 0.80% for non-underwritten rights issues. It raises an interesting question as to which issue is more costly considering the price decline and flotation costs. Its seems that non-underwritten rights issues are very costly for US industrial firms as the market s negative reaction to non-underwritten rights issue is 350% of the reaction to fully underwritten rights issues and 462% of the reaction to firm commitment offers in Eckbo and Masulis s (1992) study. This casts considerable doubt over the significance of the Eckbo and Masulis s (1992) current shareholders expected takeup variable in the determination of the flotation method.

19 18 We argue that the variables of interest in the choice of financing method may be the expected takeup of current as well as potential new shareholders and the market reaction to the choice of financing method during the equity offer period. For example, if the expected current as well as new shareholders takeup is low, then the market reaction will be strongly negative to a firm commitment offer during the offer period as the underwriter will struggle to sell the shares (firms may have to set the issue price much lower than the market price prior to the issue). In the case of underwritten rights issues, if the expected takeup of current shareholders is low the underwriter will certify the issue only if the underwriter can sell the shares to potential shareholders at least at the subscription price. However, ownership concentration will also play a role in the financing choice as that may have an impact on control dilution. Eckbo and Masulis (1992, p. 330) suggest that US, UK and Japanese firms have experienced substantial dispersion of ownership and thus go for firm commitment method whereas other markets, which are relatively small and closely held, continue to rely on rights issues (underwritten or nonunderwritten). Thus, the best strategy for firms with high ownership concentration may be to make rights issues in such a way to protect from control dilution and use other mechanisms to increase the takeup of current as well as potential new shareholders in the case of non participation of current shareholders. One strategy is to use a lower subscription price or deeper subscription price discount as suggested by Marsh (1980) and Bohren et al, (1997) to increase the takeup of current shareholders. However, this mechanism may provide a potentially negative signal of higher degree of overvaluation of the stock, and hence high quality firms are reluctant to use this strategy. Australian firms use a number of strategies to increase shareholder takeup. Underwriter certification as discussed by Heinkel and Schwartz (1986) and expanded in our Hypothesis 1 can be used to increase shareholder takeup. We argue that full underwriter certification will increase

20 19 the confidence of shareholders and increase their participation in takeup. 12 Australian companies also use shortfall facilities and back end book builds to deal with a shortfall in a rights issue rather than solely requiring the underwriter to subscribe for the initial shortfall. Australian firms have an opportunity to make two distinct types of rights issues: renounceable and nonrenounceable. If companies believe that they have a market for rights and expect some shareholders will not participate in the takeup, they can use renounceable issues rather than nonrenounceable issues to provide an opportunity for non-participating shareholders to sell the rights. 13 Thus, renounceable issues will give an opportunity to participate in the takeup by other existing shareholders or new shareholders who buy the rights. The market for rights will also guarantee that the terminal stock price will not be lower than the subscription price. Taking on board all of these considerations, we argue that the relevant takeup variable in the determination of financing choice in the Australian setting is the shareholders takeup reported by the companies to ASX under the disclosure rule of under subscription. This takeup variable includes shares taken by existing shareholders on a pro rata basis and shortfall facility and also the shares taken by those who bought rights in the case of renounceable issues. We argue that the higher the expected takeup, the greater the certification of the quality. We expect that firms will use renounceable rather than non-renounceable issues; and fully underwritten issues rather than non-underwritten or partially underwritten issues to increase the shareholders takeup. Thus, we propose the following hypotheses: Hypothesis (5a): Firms with renounceable rights issues will have higher expected takeup than firms with non-renounceable issues. Hypothesis (5b): Firms with fully underwritten rights issues will have higher expected takeup than firms with non- underwritten or partially underwritten issues. 12 We note that firms with very high ownership concentration do not need to underwrite the issue to certify the quality. That is, there are high quality firms that may not underwrite the rights issues. 13 Renounceable issues will give an opportunity for all shareholders to participate in the issue and enjoy the benefits of the renounceable rights, and will satisfy the equality and reasonableness of opportunity per Section 602(c) of the Corporations Act 2001.

21 20 High quality firms are prepared to provide costly signals to benefit from investors participation and support. There will be some mimickers. However, they will face information risk and will get hurt during the subscription period or immediately after that period. We argue that high quality/undervalued firms not necessarily will have higher takeup. Takeup will depend on a range factors as discussed earlier. Takeover panel code guidance notes on rights issues states that A rights issue which is priced more closely to the market price of the securities (or even at a premium) provides less incentive for the rights issue to be taken up by all shareholders and, therefore, may increase the likelihood of control becoming concentrated with an underwriter or other participating major shareholder. Thus, some firms may choose fully underwriting with non-renounceability and a lower subscription price discount to provide a very strong signalling mechanism. Fully underwriting with a lower discount conveys additional signal that the investment opportunity is important (i.e. not to be missed) and certify the quality/undervaluation of the firm; and non-renounceable with fully underwritten provides a further signal at the cost of control dilution. Who will choose this very costly signal? Potentially very high quality firms with lower expected takeup and/or lower ownership concentration. Considering the Hypotheses 1-4, and their potential interaction with the takeup, we further propose the following hypotheses: Hypothesis 6(a): Upper bound high quality firms with non-high expected takeup are more likely to make non-renounceable/underwritten rights issues with smaller subscription price discount to certify their quality/undervaluation and the significance of the investment opportunity. This may also provide a strong signal to the market and existing shareholders as it has costly signal underwriting fees and the possibility of control dilution. 14 These firms will experience the least unfavourable price reaction or positive price reaction. However, lower bound high quality firms may use a deeper subscription 14 Some firms with high expected takeup may also use this signal as a strong signalling mechanism.

22 21 price discount and the market will react negatively as they perceive it as an overvaluation signal with the significance of the investment opportunity. Hypothesis 6(b): Upper bound high quality firms with higher expected takeup are more likely to make renounceable/underwritten rights issues with smaller subscription price discount to signal their quality/undervaluation and the significance of the investment opportunity. These firms will experience the least unfavourable price reaction or positive price reaction. However, lower bound high quality firms may use a deeper subscription price discount and the market will react negatively as they perceive it as an overvaluation signal with the significance of the investment opportunity. Hypothesis 6(c): Upper bound intermediate quality firms with non-high expected takeup are more likely to make non-renounceable/non-underwritten rights issues with smaller subscription price discount. These firms will experience smaller unfavourable price reaction or no price reaction. However, lower bound high quality firms may use a deeper subscription price discount and the market will react negatively as they perceive it as an overvaluation signal with the significance of the investment opportunity. Hypothesis 6(d): Upper bound intermediate quality firms with higher expected takeup are more likely to make renounceable/non-underwritten rights issues with smaller subscription price discount. These firms will experience smaller unfavourable price reaction or no price reaction. However, lower bound high quality firms may use a deeper subscription price discount and the market will react negatively as they perceive it as an overvaluation signal with the significance of the investment opportunity.