Does a liability of foreignness in liquidity apply to US IPOs?

Abstract

We provide evidence regarding two unanswered and consequential questions regarding share trading liquidity, a primary motive for US listings, for the prominent listing cohort of foreign-firm US initial public offerings (FIPOs). First, we test whether FIPOs exhibit a ‘liability of foreignness (Bell et al. 2012) in liquidity’ (LFL) compared with matched domestic-firm IPOs (DIPOs), despite listing requirements that are more stringent than for the mature cross-listed foreign firms studied previously. Second, we test whether US IPO LFL is moderated by FIPO home country institutional attributes that promote liquidity. Our findings for 327 FIPOs from 36 countries between 1990 and 2012 reveal that US IPO LFL is moderated, but not eliminated, by FIPO home country attributes, thus indicating incomplete bonding with US institutions. These findings extend prior research and serve to inform foreign firms considering US IPOs, exchanges competing for them, listing facilitators, regulators, and investors regarding a salient listing consideration.

Keywords:

• foreign listings
• initial public offerings
• liquidity
• liability of foreignness

JEL:

• G01
• M4
• M49

Acknowledgements

This study developed from an earlier working paper entitled, ‘Liability of Foreignness in Global Stock Markets: Liquidity Dynamics of Foreign IPOs in the US’ coauthored with Igor Filatotchev of King’s College. We are most grateful to Igor Filatotchev for his contributions. We would like to thank our editors and the two anonymous reviewers. We also thank Sonia Falconieri, Veliko Fotak, Matthew Pinnuck, Marianna Russo, Naomi Soderstrom and participants in seminars at Australian National University, Brunel University, Central Bank of Chile, UniSinos, University of Adelaide, University of Bristol, University of Essex, University of Melbourne, University of Western Australia, the AIB/JIBS Workshop, Auckland Finance Meeting, INFINITI Conference in Ljubljana, SFA Conference in Captiva Island, World Finance Conference in Buenos Aires, and YSF Sussex Conference for their helpful suggestions and comments. We extend special thanks to Jay Ritter for allowing Jonathan Jona access to hardcopies of prospectuses of foreign IPOs.

Notes

1 Zaheer (1995, p. 343) defines liability of foreignness as ‘all additional costs a firm operating in a market overseas incurs that a local firm would not incur’ that include transportation and coordination costs, reduced perceived legitimacy, and costs arising from information asymmetries regarding products, brands, management practices, home country institutions and cultural attributes (for a comprehensive review see Rugman et al. 2011).

2 In the New York Stock Exchange (NYSE) ‘International Listings’ website, the first reason listed under ‘Why Companies List in the US’ is ‘More Liquidity’ (https://www.nyse.com/listings/international-listings). Following Kyle (1985) and Amihud and Mendelson (1986), we interpret liquidity as the ability to trade a desired quantity of a specific financial asset quickly and at low cost, with liquidity measures detailed below. Liquidity as a listing benefit reflects the expression ‘liquidity begets liquidity’ (Brown 2011, p. 278), describing the attraction of investors to invest and companies to list in more liquid markets resulting in even higher liquidity in those markets. Liquidity is also a primary determinant of IFRS adoptions (Brown 2011) in combination with strong country institutions (Daske et al. 2008, Wysocki 2011). Saudagaran (1988), Saudagaran and Biddle (1995), and Blass and Yafeh (2001) empirically document determinants of foreign stock exchange listing choices, Fanto and Karmel (1997) and Bancel and Mittoo (2001) present corroborative managerial perceptions, Doidge et al. (2004) considers cross-exchanges valuation determinants including liquidity, and Karolyi (2006) reviews findings on liquidity and other exchange choice influences. Complementary evidence indicates that accounting information plays a key role in mitigating information asymmetry in US capital markets (e.g. Brown 2011, Sadka 2011, Zimmerman 2015) via enhanced transparency (Lang and Maffett 2011) and information quality (Ng 2011).

3 We study simultaneous IPOs on more than one foreign exchange in section 5.3 with qualitatively similar findings.

4 In comparison to the present study that examines US IPO LFL, finding it moderated by FIPO home country institutions that promote liquidity, Filatotchev et al. (2020) examine US IPO earnings management, finding it is enhanced by FIPO home country institutions that promote liquidity by the reasoning that they lower SEC enforcement and US private litigation threats. Thus, both studies find incomplete bonding for FIPOs and moderating effects for FIPO home country institutions promoting liquidity, with opposite directionality for LFL and earnings management.

5 Our study period reflects that before 1990, FIPOs comprised less than 5% of US IPOs and is comparable with liquidity study periods used in prior studies of mature cross-listed foreign firms.

6 Our results are consistent with a domestic investor information disadvantage vis-à-vis home-country investors, irrespective of foreign home locale (Van Nieuwerburgh and Veldkamp 2009) and irrespective of additional liquidity costs that might arise from imperfectly connected trading venues (Domowitz et al. 1998) per our global trading test.

7 Zaheer (1995), Domowitz et al. (1998), Bacidore and Sofianos (2002) and Bacidore et al. (2005) examine US listings by non-US firms that are by nature primarily mature cross-listers, and they do not separately examine FIPOs.

8 Blass and Yafeh (2001) argue that Israeli firms conduct IPOs in the US to access a larger and more sophisticated investor pool relative to their domestic market but do not consider the influence of home country institutions.

9 Baker et al. (2018) observe that ‘whereas Chinese firm U.S. listings via reverse merger (CRMs) have dominated prior media, regulator and research attention regarding financial reporting quality, CRMs have effectively ceased, leaving listings via initial public offering (CIPOs) the relevant remaining class of Chinese firms on the U.S. exchanges.’

10 This includes publishing preliminary and subsequent amendments to the registration statements. For more details see SEC website https://www.sec.gov/divisions/corpfin/internatl/foreign-private-issuers-overview.shtml.

11 See Domowitz et al. (1998), Bacidore and Sofianos (2002), Shleifer and Wolfenzon (2002), Bacidore et al. (2005), Siegel (2005), Chung (2006), Eleswarapu and Venkataraman (2006), Fernandes and Ferreira (2008), Stulz (2009), Doidge et al. (2010) cited above.

12 As noted above, our sample period reflects that there were relatively few FIPOs before 1990 (Figure 1) and our interest in comparability with prior study findings regarding liquidity for mature foreign firms cross-listed on US exchanges. For example, Filatotchev et al. (2020) examine FIPOs between 1990 and 2009 inclusive. Our sample includes both direct stock listings and Level-III ADRs subject to like listing requirements.

13 As robustness checks, we include in untabulated regressions indicator variables for FIPOs from Israel, China, UK and Canada. We find no change in our main results with results available from the authors on request.

14 The additional analysis section below also presents robustness test results using a propensity score matching procedure with qualitatively similar findings.

15 Spamann (2010) shows that his revised index markedly differs from both the La Porta et al. (1998) original index as well as its later revision that is provided in Djankov et al. (2008).

16 Retrieved from http://www.prsgroup.com/icrg.aspx. Since both Spamann (2010) and La Porta et al. (2006) do not provide measures for China, for the first and second measures we use the values for China from Ding et al. (2010).

17 An alternative approach is to calculate the median product score by year and set HOME equal to one if above the year’s product score and zero otherwise. However, country product scores are very stable and thus there is little difference between the approaches.

18 See additional analysis section for robustness results using alternative home country institution measures.

19 As discussed later, to account for different trading protocols across US stock exchanges, we control for NASDAQ listings.

20 Given the inconclusive evidence in the literature, we excluded underpricing from all tests in an untabulated robustness exercise with qualitatively similar findings available from the authors on request.

21 In robustness tests we use the log of total assets as an alternative measure for firm size (LSALES) with qualitatively similar results that are available on request from the authors.

22 We statistically test the significance of these liquidity differences below.

23 This declining illiquidity pattern for FIPOs is consistent with prior findings regarding IPO trading volumes (Ellis et al. 2002, Corwin et al. 2004, Ellis 2006, Ellul and Pagano 2006).

24 As our focus is on IPOs, we restrict the reported results to the first 12 months of trading. LFL persists when we extend the univariate analysis to five years post IPO (untabulated), with results available from the authors on request.

25 Given the multifaceted nature of liquidity and to assess the robustness of our findings, we also examine the Amihud (2002) ratio as an alternative proxy for illiquidity, with qualitatively similar findings. In comparison to the bid-ask spread that reflects transaction costs and the price for immediacy provided by dealers, the Amihud ratio reflects the price impact of transactions. We do not report these results for brevity, but they are available from the authors on request.

26 We compare the ratio of coefficient of the constant (0.058) with the median bid-ask spreads of FIPOs in the first year of listing (0.046).

27 We estimate model 1 without the 1999 and 2000 year fixed effects as these are captured in the indicator variable DOTCOM.

28 We estimate model 1 without the 2008 and 2009 year fixed effects as these are captured in the indicator variable GFC.

29 This finding is consistent with the theoretical model of Van Nieuwerburgh and Veldkamp (2009) predicting that a foreign listing will not completely overcome the information disadvantage of domestic investors in the host country, resulting in a home bias irrespective of the home country of the foreign firm.

30 We use a home country enforcement index as measured and presented in Brown et al. (2014).

31 These indexes are available for both 2007 and 2008. To account for regulatory changes, we consider the indexes for the periods before and after the adoption in November 2007 of the Markets in Financial Instruments Directive (MiFID). We report the results for 2008. The results for 2007 are qualitatively similar and we do not report them for brevity, but we make them available upon request.

32 We confirm these differences for our sample in untabulated tests available from the authors on request.

33 We believe this to be implausible in our case since there are no significant differences in the two sample of firms before the matching, aside from marginally larger IPO proceeds (PROC) and larger (median) ROA for FIPOs than for DIPOs. We nonetheless follow this literature and test the robustness of our results using PSM techniques. For completeness, we do not find significant differences in the two sample of firms before the matching procedure. Tests are not reported for brevity but available from the authors upon request.