Investor Diversity and Liquidity in The Secondary Loan Market

For example, Chiang and Venkatesh (1988), Sarin et al. (2000) and Dennis and Weston (2001) study the impact of informed investors on stocks’ bid-ask spreads while Kedia and Zhou (2009) investigate whether informed investors affect bond liquidity.

Relatedly, Gopal and Schnabl (2021) documents that the upsurge of small business lending by finance companies and FinTech lenders made up for the decrease in bank lending after the 2008 financial crisis.

Investor diversity may matter for other reasons. For example, Hong and Page (2004) show that diverse problem-solving teams outperform homogeneous teams because they benefit from different backgrounds, ideas, and experiences of their members. Levine et al. (2014) find that, in a diverse market, traders are more likely to scrutinize the other traders’ behavior instead of placing greater confidence in the actions of the other traders and mimicking them, leading to more accurate pricing.

For example, information on the lead bank share is missing for more than half of DealScan credits and information on the shares of syndicate participants is even more sparse.

See, for example, Demsetz (1968), Tinic (1972), Stoll (1978 and 1989), Copeland and Galai (1983), Glosten and Milgrom (1985), Kyle (1985), Easley and O’Hara (1987).

Copeland and Galai (1983) and Glosten and Milgrom (1985) suggest that the presence of informed traders imposes adverse selection costs on other traders, leading to a reduction in liquidity. Admati and Pfleiderer (1988) and Holden and Subrahmanyam (1992), in contrast, show that when multiple informed investors of a security compete with each other, the speed of the price discovery increases, leading to higher liquidity. Manconi and Massa (2009) and Han and Zhou (2014) empirically study the role of informed investors on bond liquidity, a number of other papers examine the impact of informed investors on stock liquidity (e.g. Chiang and Venkatesh 1988; Sarin et al. 2000; Dennis and Weston 2001; Easley and O’Hara 1987; Glosten and Harris 1988; Lee et al. 1993; Kavajecz 1999; Liu 2013; Blume and Keim 2012.

Theories of the bid-ask spread posit that the quoted spread accounts for three forms of transaction costs faced by dealers (order processing costs, inventory costs, and information costs) which vary inversely with the liquidity of security. It is believed that the size of the order flow and frequent trading activities reduce the fixed costs of trading (Demsetz 1968). Similarly, the ease with which a dealer can either acquire a security or dispose of it will lower its inventory costs. See Tinic (1972), Stoll (1978 and 1989), Copeland and Galai (1983), Glosten and Milgrom (1985), Kyle (1985) and Easley and O’Hara (1987) for theories of bid-ask spreads.

While we are able to identify the presence of major dealers among the loan investors, we are unable to determine whether each one of them acts as a market maker for most of our loans.

Before 1998, there was no mark-to-market service available to the loan buyers (Marsh 2017).

We have performed a similar comparison to that reported in Table 1, but this time using investor concentration – our second measure of investor diversity, HHI – and found, consistent with Fig. 1, that loans with less concentrated syndicates have a lower average bid-ask spread. Results available upon request.

In a separate analysis, indeed we find that the negative relationship between bid-ask spreads and the loan share of lead arranger is driven by the arranger’s role as a dealer for the loan. When the arranger does not play this role, that relationship becomes positive, consistent with the idea that the presence of an informed investor adversely affects market liquidity. These findings are only suggestive because we have information on whether the arranger acts as a dealer for only 22% of our loan-year observations. Also, whether an arranger acts as a dealer is endogenous.

We also try to address the concern of reverse causality by putting our measures of investor diversity on the left hand side and the lagged loan bid-ask spreads on the right hand side with the same set of control variables. We find that the lagged loan liquidity does not explain the number of investor types in the syndicate, TY PES, while it does explain the concentration in the loan syndicate, HHI, but in a way inconsistent with reverse causality. These tests, however, suffer from an important limitation – they rely on annual data.

Nonetheless, we also run the tests on term loans B and get equally strong results when we proxy diversity by the number of investor types but weaker results when we do it by the HHI of investor types’ loan shares.

For example, Peristiani and Santos (2019) document that CLOs are active traders in the secondary loan market.

In another test, we control for TRADING, the total number of trading days over the year as proxied by the number of days in which the price of the loan changes. This variable comes out negative and statistically significant, but adding it to our model does not affect our measures of investor diversity.