The long- and short-run financial impacts of cross listing on Australian firms

Abstract

This paper investigates the short-run and long-run performance of Australian cross-listed firms relative to their industry rivals. The role of share trading liquidity and firm visibility in explaining abnormal returns is also investigated. In the short run, an abnormal return of 1.91% for cross-listed firms is found at announcement, while no significant abnormal returns is found for rivals on the event day. For the long-run analysis, only rival firms (especially for rivals of the non-market leaders) experience significant abnormal returns, which are negative. Cross listing into NEW ZEALAND and other countries induces a more negative impact on rivals than the UK. Lastly, liquidity is found to be a determinant of the short- and long-run abnormal returns.

Keywords

Australian firms cross listing financial performance internalisation rivals

1. Introduction

With the advent of globalisation and deregulation of the financial landscape in the past decade, there has been a surge in cross-border listings. In 1997, nearly 4700 firms cross listed on overseas exchanges globally, while there were around 1000 new foreign listings for that year (World Federation of Exchanges, 2011). Popular locations for foreign listing included the UK, the US and Japan. A decade later, the number of cross-listed firms had declined to 2837 firms in 2006, while the number of new foreign listings fell to 299, less than a third of the 1997 levels. By 2010, the number of new foreign listings remained relatively low at 394 (World Federation of Exchanges, 2011).

What motivates firms to go overseas to raise capital? Researchers have debated this question since the early 1990s, when international equity listing or ‘cross listing’ was gaining popularity. Among the argued benefits that cross listings create are a reduced cost of capital, broadening of the shareholder base, increased liquidity and the bonding of firms to a stronger legal framework (Karolyi, 2006; King and Mittoo, 2007). However, international equity raising attracts costs as well. These include costs associated with adherence to the overseas exchange’s regulatory and accounting framework, additional reporting costs and underwriting fees.

If there are net positive benefits of cross listing that accrue to these firms, the number of international equity listings should be increasing over the years. The declining trend of cross listing highlighted above raises the question of whether the benefits of internationalisation are enduring in the long term or are transitory in nature. In other words, are there permanent gains to cross list overseas?

Another perspective on cross listing is the potential effects that it has on the other firms in the industry. While it is clear that seeking shareholder interest overseas could possibly bring benefits to the cross-listed firm, there could be spillover effects on the other domestic firms in the industry of cross-listed firms. According to Melvin and Valero (2008), the act of going overseas could possibly alter the competitive landscape of the industry, because cross-listed firms are perceived to be at an advantage relative to non-cross-listed rivals in the home market. It is therefore reasonable to conjecture that the remaining firms in the industry would be affected to a certain extent, as prior studies have shown.¹ Accordingly, the aim of this paper is to examine the impact of cross listing on the short- and long-run performance of cross-listed firms in comparison to their rival firms in the Australian context. In addition, this paper also investigates the impact of cross listing on liquidity and visibility. Given that Australia is a small domestic capital market, any changes in liquidity and shareholder visibility are expected to be advantageous for cross-listed Australian firms.

This study will contribute to the literature in several ways. Firstly, the Australian market is an interesting research setting because, despite being a developed economy, gains are still expected from cross listing due to the shallower nature of its capital markets compared to the US and UK markets.² Secondly, this study will contribute to the growing body of literature on the long-run impact of cross listing. This allows a comparison of both short- and long-run benefits of cross listing and sheds some light on the possible reasons behind the recent trend reversal in cross listing. Thirdly, studies on cross listing emphasise the effects of cross listing on the listing firm and often do not seek to compare the effect from the perspective of rival firms. To date, only a few studies (Bradford et al., 2002; Melvin and Valero, 2008) document the spillover impact of cross listing on rival firms and are limited to the US context. Our study provides a more complete picture of the net benefits of cross listing by explicitly comparing the impact of cross listing on cross-listed firms versus their rival firms both in the short-run and long-run in the Australian context.

The findings of this study will be relevant to the management of both cross-listed and rival firms. If cross-listing gains are found to be transitory in nature, Australian firms seeking to raise funds overseas should reconsider cross-listing motives. Managers of firms intending to cross list would have to weigh up the cost and benefits of such a strategy. Domestic rival firms need to consider if the competitive landscape in the industry would change due to the cross listing of their competitor and whether it is beneficial for their firm to follow suit to cross list overseas. From the investors’ perspective, they could benefit from better understanding of the effect of cross listing. For example, if it is found that cross-listing gains are temporary in nature, investors should not overreact to this event.

This study finds an abnormal return of 1.91% in response to the cross-listing announcement, while rival firms experience significantly negative abnormal returns in the long run. In particular, cross listing in New Zealand and other countries bring a more negative impact to the rivals in the long run, relative to cross listing in the UK. Liquidity is found to be significant in explaining abnormal returns, highlighting the importance of liquidity gain as the main motivation for Australian firms to cross list.

The remainder of this paper is organised as follows. Section 2 provides a literature review, while Section 3 outlines the data and research method. Section 4 presents and discusses the empirical results. Finally, Section 5 concludes.

2. Literature review

2.1. Performance of cross-listed firms

Prior literature on cross listing focuses on the short-run performance of the listing firms. Foerster and Karolyi (1999) utilise a sample of 183 American Depository Receipts (ADRs) and ordinary listings in the US and find a listing week abnormal return of 1%. Roosenboom and Van Dijk (2009) examine cross listings from 44 different countries on eight major stock exchanges between 1982 and 2002 and found that the destination market plays an important role in value creation. They found the highest announcement return of 1.3% for cross listings on US exchanges, followed by 1.1% on the London Stock Exchange, 0.6% in continental Europe and an insignificant 0.5% return on the Tokyo Stock Exchange. The highest announcement returns for cross listing on the US exchange can be attributed to improved disclosure and bonding with the greater regulatory requirements of the US market.

In the long run, however, the performance of cross-listed firms tells a different story. Foreign firms listing in the US are found to underperform the local market benchmarks by 8–15% in the following three years after cross listing (Foerster and Karolyi, 2000). A similar result is evident in the study of Canadian firms by Mittoo (2003). In a similar vein, Sarkissian and Schill (2009) fail to find any permanent valuation gains for a global sample of firms 10 years pre- and post-cross listing.³ King and Segal (2009), utilising a sample of cross-listed Canadian firms between 1988 and 2005, find mixed evidence for permanent valuation gains in terms of ‘visibility’. They argue that increased visibility upon cross listing is not permanent unless the shareholder base increment is maintained over time. Collectively, these findings pose serious questions as to whether cross-listing benefits are enduring.

2.2. Theories on the benefits of cross listing

2.2.1. Market segmentation

One of the theories developed to explain the abnormal performance of cross-listed firms is the market segmentation theory. Firms internationalise to overcome investment barriers that they face in domestic markets and to diversify risk (Bancel and Mittoo, 2001; Mittoo, 1992). The presence of investment barriers in domestic markets hinders access to overseas capital, thereby limiting growth of the firms. By listing in an overseas market, firms are able to access foreign capital and increase exposure to global market factors. The ultimate result is diversification through risk sharing, thereby reducing the cost of raising capital.

2.2.2. Liquidity and multi-market trading

Amihud and Mendelson (1986) develop an asset pricing model which shows that returns of securities are an increasing concave function of liquidity. Consequently, increasing liquidity results in higher valuation and returns. By listing in multiple and larger markets, firms are able to enjoy more liquidity due to increased trading volume, exposure and reduced trading costs (Domowitz et al., 1998; Hargis, 2000). In fact, managers have cited increased liquidity as one of the motivations to list in foreign markets (Bancel and Mittoo, 2001; Mittoo, 1992). Foerster and Karolyi (1998) find a 30% increase in trading volume for 52 Canadian firms listed in the US markets between 1981 and 1990. Mittoo (2003) finds a reduction in trading costs by 1.46% for Canadian firms in the US between 1990 and 1998. Increased liquidity can be an advantage for firms coming from small domestic markets.

2.2.3. Investor recognition

Merton (1987) proposes an equilibrium model of incomplete information. A shadow cost exists due to incomplete information, leading to higher expected returns for securities due to the higher premium attributed to incomplete information. Cross listing in multiple markets can widen the shareholder base and increase the ‘visibility’ of firms. As investors become aware of these firms, the premium or shadow cost is reduced, leading to higher valuations. Foerster and Karolyi (1999) and Baker et al. (2002) document results consistent with this investor recognition theory. A wider shareholder base and increased profile enhances liquidity and price discovery in markets.

2.2.4. Bonding and corporate governance

Bonding theory postulates that cross listing can enhance corporate governance and better protect the rights of minority shareholders (Coffee, 2002; Doidge et al., 2004). Firms list in markets covered by a tougher legal framework and disclosure rules, thereby ‘bonding’ themselves to more effective legal institutions. This attracts more investors, especially those concerned with tunnelling and disclosure issues. According to Doidge et al. (2004), investors in the US are well protected relative to other countries globally. Reduced expropriation of minority shareholders by the dominant shareholders frees up resources for growth funding, thereby leading to higher firm valuation.

2.3. Home market rivals and spillover effects

Cross listing is argued to confer positive effects on the listing firm and is perceived to affect the competitive landscape of industries. Since these proposed advantages only accrue to firms that internationalise, non-cross-listing firms in the same industry are perceived to be disadvantaged. Stulz (1999) argues that firms cross list to signal to investors and distinguish themselves from ‘losers’. Prior studies document support for the existence of adverse spillover effects brought about by cross listing. Levine and Schmukler (2007) find a negative spillover effect of liquidity on remaining non-cross-listed home market firms. Melvin and Valero (2008) analyse the spillover effects of cross listing on the home market rivals using a sample of 14 US cross-listed firms between 1986 and 2002 and find that rival firms in the home markets declined in performance.

2.4. Australian firms and cross listing

Faff et al. (2002) analyse the performance one year pre- and post-cross listing for 22 Australian firms cross listed overseas as at 1996. They utilise a multivariate Generalized Autoregressive Conditional Heteroskedasticity (GARCH) model in computing abnormal returns. For the 20-day period post-listing, they find significant negative abnormal returns and no significant return in the one year post-listing. Mixed results for cost of capital reduction are also documented. They offer market timing and insider knowledge as explanations. Ahmed et al. (2006), utilising bootstrapping methods, study Australian firms cross listing overseas from 1980 to 2000; they find results that are consistent with Faff et al. (2002).

Durand et al. (2006) investigate the post-listing benefits of Australian firms cross listed between 1990 and 1999. They find a negative long-run performance for this sample, suggesting that rather than being a signal of bonding and quality, cross listing may be a sign of desperation. Frijins et al. (2010) examine the dynamics of price discovery for a sample of seven firms with bilateral cross listings in Australia and New Zealand between 2002 and 2007. They find that the home market is generally dominant in terms of price discovery; however, the role of the Australian market, being the larger of the two markets, increased over time.

3. Data and method

3.1. Data

A search is performed on Datastream to identify all possible listings of Australian firms on various exchanges. Then, the host exchange websites are searched for cross-listed firms. However, not all exchange websites provide foreign firm statistics.⁴ For these exchanges, the research department of the exchange is contacted directly to obtain the required data.

To be included in the initial sample, a firm must have (1) the Australia Securities Exchange (ASX) as the home exchange and (2) a foreign market listing that is exchange traded. These requirements deliver an initial sample of 125 firms. Consistent with prior research (e.g. Foerster and Karolyi, 1999), over-the-counter listings, level 1 ADRs⁵ and other non-exchange traded listings are excluded. The primary reason for exclusion is that non-exchange traded foreign listings do not have disclosure and reporting obligations that are as high as the main or second board exchange listings. The sample of cross-listed firms is also filtered for investment funds and preference shares due to differing operating activities, leaving 110 firms.

Based on this group of cross-listed firms, an extensive search of the announcement dates is performed up to three years prior to the cross-listing dates. If announcement dates are unavailable from the host exchange website, the dates are identified from Aspect Huntley’s DatAnalysis database, annual reports, online news articles or company web pages. Due to the lack of information surrounding the announcement dates of cross listing, 13 firms were excluded from the sample.

The next stage involves matching each cross-listing firm with a rival firm. The criteria for matching are that the rival firm must be a domestic firm listed on the ASX⁶ and is not cross listed during the three years post-cross listing of the listing firm. Each firm is matched with a domestic rival that has the closest market capitalisation at the cross-listing date. To identify the domestic rival, the Datastream Level 4 Industry/Sector Classification is used.

The final stage in constructing the sample of matched pairs of cross-listed firms and rivals involves screening for data availability. Daily closing share prices for the sample firms, risk-free rate, local equity market and global equity indices, price-to-book value ratios and total assets data are sourced from Datastream. Daily closing bid-ask prices are obtained from Bloomberg. Other company level accounting data, such as foreign sales, are sourced from annual company reports available online via Aspect Huntley FinAnalysis and Connect 4 Databases. Each cross-listed and rival firm is then filtered for one year pre- and three years post-cross listing for share price and market capitalisation availability on Datastream. After these remaining filters, the final sample consists of 80 matched pairs of cross-listed and rival firms covering the period from September 1989 to August 2005 (noting that up to three years of data are needed subsequent to the cross-listing announcement). Table 1 summarises the sample selection and filtering criteria, explaining the transition from the initial to the final samples.

Table 1.

Sample selection criteria.

Sample selection criteria	No. of firms
Initial firms with: Australia Securities Exchange (ASX) as home exchange and has foreign market listing that is exchange-traded	125
Exclusions
Investment funds and preference shares	15
Cross-listing announcement dates cannot be reliably identified	13
Cannot be matched to a rival	6
No reliable data for share prices for the 1 year pre- and the 3 year post-cross listing (either missing data or due to trading suspension)	11
Final sample	80

Table 2, Panel A, provides the distribution of the cross-listed firms across years and host exchanges. New Zealand and the UK are the most popular cross-listing destination for Australian firms, representing about 30% each of our sample. The popularity of New Zealand as a cross-listing destination for Australian firms may be due to proximity preference. Sarkissian and Schill (2004) suggested that such preference is due to familiarity, because there is additional information flow between countries with a similar culture and close geographical proximity. The table also shows that cross listing into the UK has increased in popularity post-year 2000. Table 2 Panel B groups the cross-listed firms in the sample according to Datastream industry sectors. Mining firms represent about a third of the cross-listing firms. This is likely due to the high capital requirements for mining ventures, which drive these firms to raise capital overseas. Table 3 presents the descriptive statistics for the sample cross-listed and rival firms. The rivals are matched with the cross-listed firms based on industry and size. The size of the rival firms is on average smaller than the cross-listing firms.

Table 2.

Sample description.

Panel A: Frequency of cross listings across various exchanges
Year\host country	US	UK	Luxembourg	Canada	Singapore	New Zealand	South Africa	Germany	Switzerland	Totals
1989						2				2
1990	1								1	2
1991						1				1
1992										0
1993	1									1
1994	1					1				2
1995						1				1
1996				1		1				2
1997	1		1	2		6				10
1998						5				5
1999	1							2		3
2000		2		1	1					4
2001	1	4			1		1			7
2002	1	3				2				6
2003	1	2								3
2004		6		3		5	1	1		16
2005	4	6		2				3		15
Totals	12	23	1	9	2	24	2	6	1	80

Panel B: Industry sector distribution

Industry sector					Frequency	Industry sector				Frequency

Banks					0	General retailers				2
Beverages					1	Industrial engineering				0
Chemicals					1	Industrial metals & mining				6
Construction & materials					4	Industrial transportation				1
Electronic & electrical equipment					1	Media				2
Financial services (sector)					2	Mining				26
Fixed line telecommunications					1	Oil & gas producers				15
Food & drug retailers					1	Pharmaceuticals & biotechnology				9
Food producers					2	Software & computer services				2
General industrials					1	Support services				2

Table 3.

Descriptive statistics of cross-listed and rival firms.

Variable (mean)	Cross-listed firms	Rival firms
Total assets	AUS $705 million	AUS $227 million
Market capitalisation	AUS $650 million	AUS $227 million
Return on assets	13%	8%
Financial leverage	1.66	1.68
Price-to-book ratio	3.91	3.38

3.2. Method

3.2.1. Short-run analysis

Following Foerster and Karolyi (1999) and Melvin and Valero (2008), an event study approach is employed to examine the share price reaction of firms surrounding the cross-listing event. Cross-listing announcement dates are used as the event dates. An event window of (−15, +15) is used with a 100-day estimation period. To estimate abnormal returns, a domestic market model is employed. The market model is given in Equation (1):

R_{i} = α_{i} + β_{i} R_{m} + έ_{i}

(1)

where R_i is the return of the firm, and R_m is the market return proxied by the returns on the All Ordinaries Index.⁷

To test the significance of the Cumulative Abnormal Returns (CARs), a Z-test is employed. However, there is potential for the Z-test statistic to be misspecified if returns are not normally distributed. Thus, the non-parametric Cowan sign test is also employed to complement the analysis. The event study analysis is conducted for both listing and rival firms.

3.2.2. Long-run analysis: calendar time approach

According to Mitchell and Stafford (2000), using Buy Hold Abnormal Returns (BHARs) or CARs in measuring long-term performance is flawed because these methods assume independence of cross-sectional returns. Therefore, a calendar time approach is employed in this study.

The calendar time approach involves the creation and rebalancing of a portfolio of the firms relevant to the given research question. The period covered is from the earliest cross-listing event until the latest within the sample. The portfolio of cross-listed firms is rebalanced every month to include firms that have just experienced an event (cross listing) while firms that have been in the portfolio for 36 months (three years) are dropped. The monthly returns on the portfolio are calculated on a value-weighted basis. Returns are defined as returns in excess of the risk-free rate. To test the post-36 month cross-listing performance, the monthly returns over the period of analysis (earliest cross listing to the latest) is then regressed on a constant (alpha) and the domestic market’s excess return (All Ordinaries Index), as shown in Equation (2). The alpha is interpreted as the long-term post-event abnormal return. The analysis is conducted for both listing and for rival firms:

R_{i} - R_{f} = α_{i} + β_{1} (R_{m} - R_{f}) + έ_{i}

(2)

3.2.3. Two-factor international asset pricing model

Foerster and Karolyi (1999) argued that by cross listing overseas, a firm is exposed to global market factors that could alter their risk behaviour, cost of capital and ultimately their share prices, in addition to domestic market factors. Following Foerster and Karolyi (1999) and Mittoo (2003), a firm’s pre-listing, listing and post-listing abnormal returns (captured by the separate alpha terms) are estimated from the following equation:

R_{i t} - R_{f t} = α_{i}^{P r e} + β_{1}^{P r e} (R_{A u s t} - R_{f t}) + β_{2}^{P r e} (R_{W o r l d t} - R_{f w t}) + α_{i}^{L i s t} D_{i}^{L i s t} + α_{i}^{Post} D_{i}^{Post} + β_{3}^{P o s t} (R_{A u s t} - R_{f t}) D_{t}^{P o s t} + β_{4}^{P o s t} (R_{W o r l d t} - R_{f w t}) D_{t}^{P o s t} + έ_{i t}

(3)

where

R_it – R_ft is stock i returns in excess of the Australian risk-free-rate for month t;

R_{Aus t} – R_ft is the home market excess return (Australia);

R_Worldt – R_fwt is the world excess return;

D_i^List=1 if observations are from listing month, 0 otherwise;

D_i^Post=1 if observations are from post-listing months (+1, +36), 0 otherwise.

All returns are denominated in Australian dollars and are in excess of the risk-free rate. The risk-free rate of return is proxied by the yield on the JP Morgan Australian Government Bond Index sourced from Datastream. The global excess return is the excess returns on the Datastream International World Index. For the global risk-free rate, the yield on the JP Morgan global government bond index (excluding Australia) is used.

The estimation of the pooled cross section and time series of returns is performed using the Schipper and Thompson (1983) method. The abnormal returns are the alphas, which are split into the listing and post-listing period. Using this method, both short-run and long-run abnormal performances are observed in a single regression. This analysis is repeated for the rival firms. Newey–West heteroskedasticity and autocorrelation consistent standard errors and covariance adjustment is applied in the regression.

3.2.4. Cross-sectional analysis on the abnormal returns of cross-listed firms

The cross-section model is defined in Equation (4). Equation (3) is regressed for each individual firm to obtain the dependent variables for the cross-section analysis. For the short-run analysis, the dependent variable employed in Equation (4) is the excess abnormal returns for the listing month (α_i^Pre +α_i^List from the two-factor international asset pricing model (IAPM) regression). For the long-run analysis, the dependent variable is the excess abnormal returns for the post-listing period (α_i^Pre +α_i^Post from the two-factor IAPM regression). The model is given by

α_{i} = c_{i} + β_{1} Δ S P R E A D + β_{2} Δ λ + β_{3} G R O W T H + β_{4} S I Z E + β_{5} F O R E I G N + δ_{1} M N G D U M + δ_{2} N Z D U M + δ_{3} U K D U M + δ_{4} M L E A D E R D U M + έ_{i}

(4)

where ΔSPREAD represents the change in spread percentage (SP%) before and after cross listing; Δλ is the change in shadow cost around the cross-listing announcements; GROWTH is the price-to-book ratio; SIZE is the natural log of total assets; FOREIGN is the ratio of foreign sales to total sales for the fiscal year prior to cross listing; MNGDUM is a mining dummy variable taking a value of 1 for mining stocks and 0 otherwise; NZDUM (UKDUM) is a dummy variable taking a value of 1 if the cross listing occurs in New Zealand (the UK) and 0 otherwise; and MLEADERDUM is a dummy variable taking a value of 1 if the cross lister is a market leader (if its market capitalisation is greater than 5% of the industry market capitalisation). The cross-sectional regression is adjusted for Newey–West heteroskedasticity and autocorrelation consistent standard errors and covariance.

The spread percentage is calculated as

SP % = (Ask - Bid) / [(Ask + Bid) / 2]

(5)

where Bid and Ask is the daily closing bid and ask prices of a firm’s stock. To construct ΔSPREAD for the short run, the average spread percentage for the month before the listing month is subtracted from the average spread percentage of the month following the listing month, as in Kadlec and McConnell (1994). For the long-run cross-sectional analysis, the ΔSPREAD is estimated as the 12-month average spread percentage before listing (months −12 to −1) subtracted from the average spread percentage for the third year (months +25 to +36) following the cross-listing month. The more liquid a security is, the higher the price (Amihud and Mendelson, 1986). As a particular stock experiences increased liquidity, the bid-ask spread would tighten to reflect the decline in trading costs, leading to higher valuation. Thus, a negative relation is expected between ΔSPREAD and abnormal returns, because a negative ΔSPREAD would be interpreted as higher liquidity.

To capture the change in visibility resulting from cross listing, a ‘shadow cost’ proxy of incomplete information for each firm is estimated following Kadlec and McConnell (1994) and Foerster and Karolyi (1999). The change in shadow cost (Δλ) around cross-listing announcements is estimated as

Δ λ = σ_{\in_{i}}^{2} M C A P_{i} (\frac{1}{S H R_{t + 1}} - \frac{1}{S H R_{t}})

(6)

where $σ_{\in_{i}}^{2}$ is the residual variance from the two-factor IAPM regression (Equation (3)), MCAP_i is the normalised market capitalisation of the firm in the listing week and SHR_t is the number of shareholders recorded in the year before (t) and after (t+1) listing. If a cross-listed firm experiences an increase in shareholder base, the Δλ will be negative, indicating a drop in shadow cost. Thus, when regressed against the abnormal returns, the coefficient of Δλ is expected to be negative in the cross-sectional regression.

4. Results and discussions

4.1. Short-run analysis

Table 4 presents the event study results for cross-listed and rival firms for various windows within the (−15, +15) event period. Panel A presents the full sample results, while Panel B displays results for the sub-sample of cross-listed market leaders and their rivals. For the full sample of cross-listed firms, the mean abnormal return on the announcement date is 1.91% (significant at the 1% level). For the three days surrounding the cross-listing date (−1, +1), the mean CAR of 2.82% is also significant at the 1% level. This suggests that investors react positively to the announcement of cross listing. Previous foreign listing studies, such as Foerster and Karolyi (1999), similarly documented a small significant abnormal performance of 1% within the cross-listing month or one week around the listing date. However, a significant CAR (at the 1% level) is found in the pre-event window (−15, −2), which might suggest information leakage or anticipation of the cross-listing announcement. Interestingly, there seems to be a short-term reversal in abnormal returns beyond day 1, where the mean CAR for (+2, +15) is −4.22% (significant at the 1% level) – indeed, this negative reaction is twice the magnitude of the initial day-zero positive impact of the cross-listing event, but less than the cumulated CAR from (−15, +1) of +5.8%.

Table 4.

Event study results for cross-listed and rival firms.

Panel A: Full sample
Period	Cross-listed firms				Industry rivals
	Mean CAR	Z-stat	Positive fraction	Cowan’s sign test	Mean CAR	Z-stat	Positive fraction	Cowan’s sign test
(−15, −2)	0.0298**	3.7711	0.6589	0.4431	0.0095	0.2015	0.6857	0.7667
(−1, −1)	0.0011	0.0855	0.5875	1.2234	0.0146**	3.4640	0.8000*	2.0179
(0, 0)	0.0191**	4.0416	0.7500	1.8292	−0.0047	−1.7053	0.6625	0.6564
(+1, +1)	0.0080	1.5950	0.7375	1.5944	0.0009	0.1549	0.7125	0.3161
(−1, +1)	0.0282**	5.7222	0.6917	1.2703	0.0108	1.9136	0.7250	0.9685
(+2, +15)	−0.0422**	−10.4492	0.6295	1.6279	0.0110*	−2.1833	0.6821	1.0266
(−15, +15)	−0.0158	−0.9560	0.7157**	6.5998	0.0313	0.8820	0.7581**	6.6938

Panel B : Market leaders only
Period	Cross-listed firms that are market leaders				Industry rivals of market leader
	Mean CAR	Z-stat	Positive fraction	Cowan’s sign test	Mean CAR	Z-stat	Positive fraction	Cowan’s sign test

(−15, −2)	0.0444*	2.2944	0.6345	0.6629	0.0144	−1.2907	0.6555	0.3796
(−1, −1)	−0.0037	−0.9032	0.5294	0.7123	0.0185	1.8256	0.7647	0.8543
(0, 0)	0.0162	1.3390	0.7647	1.2801	−0.0020	−0.6410	0.5882	0.6896
(+1, +1)	0.0037	0.5219	0.7059	0.7820	0.0069	0.0578	0.7647	0.8543
(−1, +1)	0.0162	0.9577	0.6667	0.6667	0.0234	1.2424	0.7059	0.5883
(+2, +15)	0.0166	−0.8532	0.6303	0.6303	0.0304	0.0923	0.6050*	2.0302
(−15, +15)	0.0822*	2.3989	0.7002**	4.0855	0.0682	0.0439	0.7059	1.8912

Note: This table presents the mean Cumulative Abnormal Returns (CARs) around the listing date for 80 matched pairs of cross-listed firms and rivals in the sample. The results in Panel A are for the full sample, while Panel B reports the results for the sub-sample of cross-listed market leaders and their rivals as a robustness check. The positive fraction column indicates the fraction of the firms in the sample that had positive CARs in the window. ** and * indicate statistical significance at the 1% level and 5% level, respectively.

Due to potential stock illiquidity and non-normality of the sample return distribution, which are highlighted in the event study methodology literature (MacKinlay, 1997), the non-parametric Cowan’s sign test is also reported in Table 4. Similar to the Z-test, the Cowan sign tests indicate that on day zero, the CAR is positive and statistically significant at the 10% level.

Regarding the sample of rival firms, a negative abnormal return of −0.47% is found on the event day, albeit only significant at the 10% level. In addition, a positive abnormal return of 1.46% (CAR of 1.10%) is found on the day prior to the event day (post-event period +2, +15), significant at the 1% (5%) level, respectively. This suggests that the negative reaction for the rivals is very short term in nature. The Cowan sign test similarly displays a significant positive fraction of CARs for day −1. The results for the rival firms are somewhat consistent with the findings of Melvin and Valero (2008), where they show that home market rivals of the cross-listed firms in their US study experience a negative impact when firms cross list. One possible explanation for the Australian results is that while investors react positively to cross-listing news, they only view non-cross-listed rivals firms at a marginal disadvantage relative to the cross-listing firms in the short run.

As a robustness test, the reaction of cross-listed firms that are market leaders within their industry and their rivals is also investigated. To this end, a cross-listed firm is defined as a market leader if its market capitalisation is greater than 5% of the industry. The results of the robustness test on the short-run analysis are shown in Table 4, Panel B. For the cross-listed firms, the mean CAR is not statistically significant. However, we detected a significant pre-event window (−15, −2) CAR, suggesting that information leakage or anticipation of the cross-listing news associated with market dealers is more probable. In contrast to the full sample, there is no significant short-term reversal in returns post-event day for the market leader sub-sample. For the rival firms, the robustness test results again indicate no significant mean CARs on the event day. However, positive fractions of mean CARs are found for the post-event window (+2, +15) at the 5% level.

4.2. Long-run analysis

Table 5, Panel A, reports the outcome from estimating the market model applied to the 36-month calendar time approach. The coefficient of interest is the alpha coefficient, which is interpreted as the excess abnormal returns in the long run. The alpha is not statistically significant for the cross-listed firms. The short-run positive abnormal returns have disappeared when the longer time horizon is investigated. This finding is consistent with the long-run hypothesis that gains are transitory. Sarkissian and Schill (2009), in their long-term focus study of cross-listed firms, document similar results for a global sample 10 years post-cross listing using CARs from a market model. This finding is also consistent with Durand et al. (2006), who use cross-listing dates as the event date with a sample from 1990–1999.

Table 5.

Long-run performance analysis using the calendar time approach from September 1989 to July 2008.

Panel A: Market model (Months +1, +36)
	Coefficient				Adjusted R²	Chi sq-test (p-value)
	alpha	R_m–R_f	SMB	HML	Adjusted R²	Chi sq-test (p-value)
Cross-listedfirms	0.0015(0.2251)	0.9619**(11.0316)			0.5078	11.2630**
Rivals	−0.0239**(−3.0402)	0.7496**(7.5159)			0.2467

Panel B: Fama and French three-factor model (months +1, +36)

Cross-listedfirms	0.0024(0.3639)	0.9767**(11.1053)	−0.6102(−0.2022)	3.9562(1.2573)	0.5074	13.3305**
Rivals	−0.0251**(−3.3942)	0.7254**(7.9471)	6.5227*(2.2832)	−1.5771(−0.2695)	0.2576

Panel C: Market model (months +1, +36), sample with market leaders only

Cross-listedfirms	0.0146(1.4747)	1.0907**(9.0238)			0.3968	3.8724*
Rivals	−0.0073(−0.6295)	0.9466**(6.9046)			0.2019

Panel D: Fama and French three-factor model (months +1, +36), sample with market leaders only

Cross-listedfirms	0.0152(1.5426)	1.0999**(9.1523)	−0.2322(−0.0684)	2.5700(0.7478)	0.3924	4.2835*
Rivals	−0.0081(−0.6880)	0.9311**(7.0525)	4.8296(1.5673)	−0.4233(−0.0645)	0.1997

Note: This table presents the estimated coefficients and t-statistics (in parentheses) of the market model (Equation (2)) and Fama–French model in Panel A and Panel B, respectively. The analysis is repeated for market leaders in Panel C and Panel D. ** and * indicate statistical significance at the 1% level and 5% level, respectively.

Also in Panel A of Table 5, the long-run market model analysis indicates that the domestic rival firms suffer a negative abnormal return of 2.39% post-cross listing 36 months, which is significant at the 1% level. One possible explanation is that increased visibility or profile of the company, as a result of cross listing, allows it to avoid a negative market effect occurring in the relevant industry over this period. Rival firms that are not cross listed are seen as missing out on the increased visibility and broader shareholder base in the long run. Kadlec and McConnell (1994) find strong support for the investor recognition hypothesis proposed by Merton (1987). This suggests that investors’ recognition could be a reason for the negative abnormal returns for rivals in the long run.

As a robustness check on the long-run analysis, the Fama and French model is used in the calendar time portfolio regression (Equation (2)). The outcome is also reported in Table 5, in Panel B. Most notably, we see that the results are robust and consistent with the main market model regression. With regard to the main cross-listing sample, the calendar time portfolio alpha is still insignificant at the 5% level. For the rival firms, the negative alpha remains quite similar (now −2.51%) and significant at the 1% level.⁸

The long-run analysis is repeated for the sub-sample of cross-listed market leader firms and their rivals (Panels C and D). Interestingly, the alpha is no longer significant, although the Chi-square test indicates a significant difference between the alphas of the cross-listed firms and their rivals. This finding suggests that the results of the full sample are driven by the non-market leaders, rather than the market leaders. That is, our sample evidence suggests that the rivals lose out more when the non-market leaders cross list overseas, possibly because non-market leaders have more to gain from cross listing (e.g. due to increased exposure) relative to their rivals, than do market leaders relative to their rivals.

4.3. Two-factor international asset pricing model

Table 6 reports pre-listing, listing month and post-listing month performance estimates by employing the two-factor IAPM (Equation (3)). We show results for the full sample (Panel A) and for the following sub-samples (Panel B): New Zealand; the UK; other countries; and market leaders. For the cross-listed firms in the full sample, the one-year pre-listing performance (alpha pre) is 2.37%, suggesting that cross-listed firms have better performance one year prior to cross listing. In addition, the cross-listing firms have a listing month abnormal performance of 3.18% (alpha pre of 2.37% plus alpha list of 0.81%), which is significant at the 10% level. However, for rival firms, the post-listing month performance of −0.98% (alpha pre of 1.75% plus alpha post of −2.73%) is significant at the 10% level. Consistent with the long-run analyses discussed earlier, this demonstrates the negative impact to the rival firms who chose not to cross list in the long run. As an aside, the global beta of the rivals has increased. Although not being cross listed, rival firms are exposed to international market factors indirectly through competition.

Table 6.

Risk-adjusted return performance analysis with two-factor international asset pricing model.

	Pre-cross-listing period(Months −12, −1)			Listing month	Post-cross-listing period(+1,+36)			Adjusted R²
	α_ɩ^Pre	β₁ ^Pre (Aus)	β₁ ^Pre (World)	α_i^List	α_i^P^ost	β₃ ^P ^ost (Aus)	β₄ ^P ^ost (World)	Adjusted R²
Panel A: Full sample
Cross listed	0.0237*	1.0535**	0.0359	0.0081	−0.0222	0.5573*	−0.2919	0.0794
	(2.3272)	(5.3060)	(0.2564)	(0.4654)	(−1.9353)	(2.4466)	(−1.7896)
Rivals	0.0175	1.2545**	−0.2594	−0.0012	−0.0273*	−0.1958	0.4135*	0.0521
	(1.6389)	(6.0356)	(−1.7718)	(−0.0668)	(−2.2687)	(−0.8209)	(2.4214)
Panel B: Various sub-samples
New Zealand
Cross listed	0.0192	1.1521**	0.1817	0.0130	−0.0179	0.5163	−0.5022	0.1017
	(1.0847)	(3.5606)	(0.7456)	(0.4567)	(−0.9080)	(1.4070)	(−1.8440)
Rivals	−0.0128	0.6771	−0.2163	0.0162	−0.0092	0.0140	0.5216	0.0284
	(−0.5729)	(1.6629)	(−0.7054)	(0.4527)	(−0.3710)	(0.0303)	(1.5217)
United Kingdom
Cross listed	0.0241	1.1044*	0.2307	0.0426	−0.0199	0.6868	−0.5424	0.0750
	(1.1623)	(2.4592)	(0.8901)	(1.2648)	(−0.8550)	(1.3564)	(−1.7238)
Rivals	0.0605**	2.0656**	−0.4170	0.0031	−0.0469*	−0.6302	0.5709*	0.0922
	(3.2113)	(5.0585)	(−1.7692)	(0.0998)	(−2.2210)	(−1.3688)	(1.9955)
Other countries
Cross listed	0.0277	0.9715**	−0.1855	−0.0205	−0.0268	0.5011	0.0263	0.0730
	(1.7632)	(3.2097)	(−0.8278)	(−0.7237)	(−1.4950)	(1.4265)	(0.0997)
Rivals	0.0099	1.2313**	−0.2138	−0.0115	−0.0268	−0.1263	0.2303	0.0455
	(0.6190)	(3.9731)	(−0.9320)	(−0.3978)	(−1.4617)	(−0.3512)	(0.8518)
Market leaders
Cross listed	0.0677**	1.5168**	−0.0509	−0.0080	−0.0562**	−0.1501	0.0656	0.1531
	(4.1552)	(5.5985)	(−0.2471)	(−0.2692)	(−2.9949)	(−0.4675)	(0.2716)
Rivals	0.0327	1.3611**	−0.0091	0.0149	−0.0450*	−0.4250	0.1803	0.0835
	(1.8185)	(4.5566)	(−0.0399)	(0.4572)	(−2.1749)	(−1.2007)	(0.6767)

Note: This table presents the estimated coefficients and t-statistics (in parentheses) of the two-factor international asset pricing model with dummy list and dummy post: R_it −R_ft = α_i^Pre + β₁^Pre(R_Aust–R_ft) + β₂^Pre(R_Worldt–R_fwt) + α_i^ListD_i^List + α_i^PostD_i^Post + β₃^Post(R_Aust–R_ft) D_t^Post + β₄^Post(R_Worldt–R_fwt) D_t^Post + ϵ́_it (Equation (4)).

and * indicate statistical significance at the 1% level and 5% level, respectively.

As a robustness test, the full sample was divided into sub-samples. Given that firms which cross listed in New Zealand and the UK make up a large proportion of the sample, the analysis is repeated for these two groups separately. For the New Zealand sub-sample, the alpha pre and alpha list are not statistically significant. However, a negative and significant post listing performance of −2.2% (alpha pre of −1.28% plus alpha post of −0.92%) for the rival firms of New Zealand cross listings is found at the 5% level. One of the possible explanations is that New Zealand and Australia have close proximity of market factors in terms of culture, geography and legal framework efficiency, as found by La Porta et al. (1998) and Sarkissian and Schill (2004). Thus, cross listing in New Zealand does not seem to bring abnormal returns, but rival firms that do not follow suit suffer negative consequences.

For the UK sub-sample, there is a significant listing month performance of 6.67% (6.36%) abnormal returns for the cross-listed (rival) firms, both significant at the 10% level. This shows that firms that choose to cross list in the UK are only enjoying marginal benefits from cross listing relative to their rivals from the investor’s point of view. There is also a pre-listing run up of 6.05% for the rival firms. The results for the ‘other countries’ sub-sample are similar to the New Zealand sub-sample.

The market leader sub-sample shows that the pre-listing run up for the cross-listed firms is positive and significant of 6.77% at the 1% level, while the run up is only 3.27% for the rivals (significant at the 10% level). Only the market leaders present a significant listing month performance of 5.97% (at the 10% level). While the negative effect only accrued to the rivals in the long run in the full sample, this is not confirmed in the market leader-only sample. These findings corroborate the analyses presented in the long-run analysis (Section 4.2), where rivals tend to lose out more when non-market leaders cross list overseas.

4.4. Cross-sectional analysis

Table 7 presents the cross-sectional regression analysis. The results are divided into two panels: Panel A (Panel B) presents the cross-sectional regression on the short-run (long-run) abnormal performance alpha.

Table 7.

Cross-sectional regression of risk-adjusted abnormal returns on firm-specific variables.

Variable	Panel A: Listing month abnormal returns	Panel B: Post-listing months abnormal returns (months +1, +36)
Constant	0.4138*(2.1612)	0.0326(0.4718)
ΔSPREAD	−2.8100*(−2.0050)	−0.6750**(−5.8987)
Δλ (× 10⁸)	3.4300(0.6126)	−1.1900(−0.7514)
GROWTH	−0.0035(−1.2324)	0.0005(0.4135)
SIZE	−0.0164(−1.7271)	−0.0015(−0.4098)
FOREIGN	−0.0836(−1.9212)	−0.00042(−0.0185)
MNGDUM	−0.0420(−0.7834)	−0.0123(−0.7797)
NZDUM	−0.0706(−1.5835)	−0.0052(−0.3304)
UKDUM	0.0209(0.4711)	0.0070(0.4512)
MLEADER DUM	0.0291(0.6745)	0.0104(0.5411)
Adjusted R²	0.0307	0.1216

Note: This table reports the estimated coefficients and t-statistics (in parentheses) from the cross-sectional regression α_i = c_i + β₁ΔSPREAD + β₂ Δλ + β₃GROWTH + β₄SIZE + β₅FOREIGN + δ₁MNGDUM + δ₂NZDUM + δ₃UKDUM+ δ₄MLEADERDUM + ϵ́_i, where ΔSPREAD and Δλ are the test variables, while growth, size, foreign sales (FOREIGN), mining dummy (MNGDUM), New Zealand dummy (NZDUM), United Kingdom (UKDUM) dummy and market leader dummy (MLEARDERDUM) are included as control variables. Variable definitions are provided in the test. ** denotes significance at the 1% level and * denotes significance at the 5% level.

With the short-run abnormal performance regression, we find that ΔSPREAD is negative and significant at the 5% level. This indicates that the decline in trading costs (i.e. liquidity gain) is a positive determinant of short-run abnormal returns, which is consistent with the expectation that domestic liquidity should improve due to additional foreign order flow generated by cross listing. The change in visibility (shadow cost) proxy, Δλ is not statistically significant, indicating that visibility is not a determinant of cross listing abnormal returns for the Australian sample.

Size is found to be negatively associated with short-run abnormal returns (at the 10% level), indicating that smaller firms enjoys greater abnormal returns upon cross listing, perhaps due to the magnitude of liquidity gain. There is a negative relationship between the level of foreign sales and the abnormal performance of the cross-listed firms in the short run (at the 10% level). While the extant literature show that firms with higher foreign sales are more likely to go overseas to cross list (Bailey et al., 2006; Doidge et al., 2004; King and Segal, 2009), this study shows that once cross listed, the level of foreign sales is negatively associated with abnormal performance. Investors seem to value foreign sales as a source of revenue diversification, which is viewed differently to the diversification of shareholder base achieved with cross listing.

In the long-run multivariate analysis (36 months post-listing), the results are similar to the short run, where ΔSPREAD is negative and statistically significant at the 1% level, while Δλ is not statistically significant. This suggests that liquidity gain is the primary positive determinant of abnormal return of Australian firms, which is consistent with the conjecture that the key motivation for firms to cross list overseas is to improve share trading liquidity.

5. Conclusions

This paper examines the impact of cross listing on Australian firms in the short and long run and spillover effects on domestic rivals. The role of liquidity and visibility as determinants of abnormal performance of cross listing is also investigated. The short-run event study analysis indicates that there is a small listing gain of 1.91% during the event window. For the long-run analysis, cross-listed firms are found to have no significant abnormal returns. This is in line with the current literature on long-term performance that finds no permanent gains arising from cross listing. Rival firms (especially rivals of the non-market leaders), on the other hand, experience negative abnormal returns in the long run. The analysis based on an IAPM also finds that there are significant positive abnormal returns for the cross-listing firms during the listing month, but negative abnormal returns for industry rival firms in the long run.

A cross-sectional analysis is also conducted to investigate whether liquidity and visibility changes are determinants of cross-listing performance. Results from the cross-sectional regressions suggest that liquidity gain is a significant positive determinant in explaining abnormal returns for cross listing.

Overall, the results suggest that Australian firms seeking to raise funds overseas should reconsider their cross-listing motives by weighing up the costs and benefits of cross listing. Since the cross-listing gains are temporary in nature, investors should not overreact and bid up the stock prices beyond the fair value upon cross listing of a firm. In addition, the findings of our study indicate that domestic rival managers might need to consider changes in the competitive landscape within the industry, as there is evidence that non-cross-listed home market rivals are negatively affected in the long run.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Date of acceptance of final transcript: 20 December 2011.

Accepted by Associate Editor, Tom Smith (Finance).

Notes

References

Ahmed

Kim

Henry

(2006) International cross-listings by Australian firms: a stochastic dominance analysis of equity returns. Journal of Multinational Financial Management 16: 494–508.

Amihud

Mendelson

(1986) Asset prices and the bid ask spread. Journal of Financial Economics 17: 223–249.

Bailey

Andrew Karolyi

Salva

(2006) The economic consequences of increased disclosure: evidence from international cross-listings. Journal of Financial Economics 81: 175–213.

Baker

Nofsinger

Weaver

(2002) International cross-listing and visibility. Journal of Financial and Quantitative Analysis 37: 495–521.

Bancel

Mittoo

(2001) European managerial perceptions of the net benefits of foreign stock listings. European Financial Management 7: 213–236.

Bradford

Martin

Whyte

(2002) Competitive and information effects of cross-border stock listings. Journal of Financial Research 25: 399–413.

Coffee

Jr (2002) Racing towards the top? The impact of cross-listings and stock market competition on international corporate governance. Columbia Law Review 102: 1757–1831.

Dimson

(1979) Risk measurement when shares are subject to infrequent trading. Journal of Financial Economics 7: 197–226.

Doidge

Karolyi

Stulz

(2004) Why are foreign firms listed in the US worth more? Journal of Financial Economics 71: 205–238.

10.

Domowitz

Glen

Madhavan

(1998) International cross-listing and order flow migration: evidence from an emerging market. Journal of Finance 53: 2001–2027.

11.

Durand

Gunawan

Tarca

(2006) Does cross-listing signal quality? Journal of Contemporary Accounting and Economics 2: 48–67.

12.

Faff

Hodgson

Saudagaran

(2002) International cross-listings towards more liquid markets: the impact on domestic firms. Journal of Multinational Financial Management 12: 365–390.

13.

Foerster

Karolyi

(1998) Multimarket trading and liquidity: a transaction data analysis of Canada-US interlistings. Journal of International Financial Markets, Institutions and Money 8: 393–412.

14.

Foerster

Karolyi

(1999) The effects of market segmentation and investor recognition on asset prices: evidence from foreign stocks listing in the United States. Journal of Finance 54: 981–1013.

15.

Foerster

Karolyi

(2000) The long-run performance of global equity offerings. Journal of Financial and Quantitative Analysis 35: 499–528.

16.

Frijins

Gilbert

Tourani-Rad

(2010) The dynamics of price discovery for cross-listed shares: evidence from Australia and New Zealand. Journal of Banking and Finance 34: 298–308.

17.

Hargis

(2000) International cross-listing and stock market development in emerging economies. International Review of Economics and Finance 9: 101–122.

18.

Kadlec

McConnell

(1994) The effect of market segmentation and illiquidity on asset prices: evidence from exchange listings. Journal of Finance 49: 611–636.

19.

Karolyi

(1998) Why do companies list shares abroad? A survey of the evidence and its managerial implications. Financial Markets, Institutions and Instruments 7: 1–60.

20.

Karolyi

(2006) The world of cross-listings and cross-listings of the world: challenging conventional wisdom. Review of Finance 10: 99–152.

21.

King

Mittoo

(2007) What companies need to know about international cross-listing. Journal of Applied Corporate Finance 19: 60–74.

22.

King

Segal

(2009) The long-term effects of cross-listing, investor recognition, and ownership structure on valuation. Review of Financial Studies 22: 2393–2421.

23.

La Porta

Lopez-de-Silanes

Shleifer

. (1998) Law and finance. Journal of Political Economy 106: 1113–1155.

24.

Levine

Schmukler

(2007) Migration, spillovers, and trade diversion: the impact of internationalization on domestic stock market activity. Journal of Banking and Finance 31: 1595–1612.

25.

MacKinlay

(1997) Event studies in economics and finance. Journal of Economic Literature 35: 13–39.

26.

Maynes

Rumsey

(1993) Conducting event studies with thinly traded stocks. Journal of Banking and Finance 17: 145–157.

27.

Melvin

Valero

(2008) The dark side of international cross-listing: effects on rival firms at home. European Financial Management 15: 66–91.

28.

Merton

(1987) A simple model of capital market equilibrium with incomplete information. Journal of Finance 42: 483–510.

29.

Mitchell

Stafford

(2000) Managerial decisions and long-term stock price performance. Journal of Business 73: 287–329.

30.

Mittoo

(1992) Managerial perceptions of the net benefits of foreign listing: Canadian evidence. Journal of International Financial Management and Accounting 4: 40–62.

31.

Mittoo

(2003) Globalization and the value of US listing: revisiting Canadian evidence. Journal of Banking and Finance 27: 1629–1661.

32.

Roosenboom

Van Dijk

(2009) The market reaction to cross-listings: does the destination market matter? Journal of Banking and Finance 33: 1898–1908.

33.

Sarkissian

Schill

(2004) The overseas listing decision: new evidence of proximity preference. Review of Financial Studies 17: 769–809.

34.

Sarkissian

Schill

(2009) Are there permanent valuation gains to overseas listing? Review of Financial Studies 22: 371–412.

35.

Schipper

Thompson

REX

(1983) The impact of merger-related regulations on the shareholders of acquiring firms. Journal of Accounting Research 21: 184–221.

36.

Scholes

Williams

(1977) Estimating betas from nonsynchronous data. Journal of Financial Economics 5: 309–327.

37.

Stulz

(1999) Globalization, corporate finance, and the cost of capital. Journal of Applied Corporate Finance 12: 8–25.

38.

World Federation of Exchanges (2011) Statistics. Available at: http://www.world-exchanges.org/WFE/home.asp?action=document&;menu=10