The Informational Efficiency of the Equity Market As Compared to the Syndicated Bank Loan Market

1. The Informational Efficiency of the Equity MarketAs Compared to the Syndicated Bank Loan Market Linda Allen, Baruch College, CUNY and Aron A. Gottesman, Pace University

2. Objective • Objective: To investigate the integration of the secondary market for syndicated bank loans with the secondary market for public equity. • Key factors that can influence cross-market integration: • Differential access to private information. • Liquidity differences across markets. • Syndicated loans may be less sensitive than equity to certain firm-specific information because of debt’s limited potential for upside gain.

3. Information characteristics of the secondary market for syndicated loans • The assembling and setting of the terms of loan syndication are primary market transactions. • After the loan syndication is closed, banks can sell their loan syndication shares in the secondary market. • Both screening and monitoring intermediation services are provided in syndicated bank loans. • The lead arranger is typically a relationship bank that has access to private information about the borrower and therefore can effectively screen the loan.

4. Information characteristics of the secondary market for syndicated loans • Ongoing monitoring is induced by the lead bank’s relatively large stake in the loan. • Moreover, the structure of the syndicated loan mandates ongoing monitoring through a series of financial and non-financial covenants that require the borrower to make regular disclosures of private information to all members of the syndicate. • Trading in the syndicated bank loan market is limited to financial institutions and sophisticated investors as a result of the designation of these instruments as Rule 144a securities. • There is no direct participation by individual (retail) investors. Exclusion of retail investors enhances the market’s informational efficiency as uninformed noise traders do not contribute to price volatility in the syndicated loan market.

5. Information characteristics of the secondary market for syndicated loans • Syndicated loans lie between illiquid relationship bank loans and arms-length public equity. • Dennis and Mullineaux (2000) document that the volume of loan syndications “has increased at well over a 20% rate annually over the past decade” and “topped $1 trillion in 1997.” • The Wall Street Journal (see Zuckerman and Sapsford, (2001)) referred to the syndicated loan market as a “multi-trillion dollar debt bazaar that has become the nation’s largest capital market during the last decade.” • Thomas and Wang (2004) note that the increased liquidity in the bank loan market after 1993 approximates conditions in the high yield bond market.

6. Liquidity characteristics of the secondary market for syndicated loans • The dramatic growth in secondary loan market trading volume in the 1990s was fueled, in part, by the adoption of the Basel Capital Accords that induced banks to seek ways to remove capital-intensive loans from their books. • The 1995/1997 standardization of settlement procedures via the Loan Syndications and Trading Association’s (LSTA) development of standardized trading documentation and T+10 (now T+7) settlement procedures for par/near par loans reduced the incidence of trade disputes in the syndicated loan market. • The growth of mutual funds of senior bank loans fueled market demand.

7. Liquidity constraints in the secondary market for syndicated loans • Order processing costs are considerable, particularly for the sale of assignments, which require borrower consent and legal documentation of the transfer of the lender’s share of the loan. • A single informed lender (the lead bank) trades with other less informed banks (syndicate members) and non-banks (mutual funds and investment banks). • Thus, the risk shifting concern that the lead bank will syndicate the “bad loans” and keep the “good loans” contributes to an information asymmetry that is implicit in the structure of the market. • In contrast, information is symmetric in equity markets, particularly in the wake of the SEC’s adoption of Regulation FD in Oct. 2000 that mandated fair disclosure of any material and forward-looking information to the entire market, rather than to a favored institution.

8. Four Hypotheses • Private Information Hypothesis: Loan markets have access to private information available to syndicate members. Predicts loan market will lead. • Liquidity Hypothesis: Uninformed traders migrate to the more liquid equity markets. Predicts equity market will lead. • Asymmetric Price Reaction Hypothesis: Loan markets are more sensitive to negative information, whereas equity markets react to both negative and positive information. Predicts loan markets will lead by negative information, and equity market will lead by positive information. • Integrated Market Hypothesis: Prior hypotheses may hold at different points in time; the appropriate market depends on the relative liquidity costs and gains from trade. Eventually, information is reflected in all markets. Predicts that both markets will lead each other.

9. Previous evidence: Integration tests Kwan (1996): • Employs market integration tests of whether contemporaneous and lagged stock returns explain public bond returns. • Finds evidence of significant coefficients on lagged stock returns for both investment grade and below investment grade bonds. • Suggests that stock markets lead the publicly-held bond market. Suggests support for the liquidity hypothesis.

10. Previous evidence: Integration tests Hotchkiss and Ronen (2002): • Refute Kwan, using intraday data on individual high yield bonds that trade on NASD’s Fixed Income Pricing System (FIPS). • Although Hotchkiss and Ronen (2002) do not test it directly, their results also offer support for the liquidity hypothesis, since they find informational efficiency for the liquid corporate bonds in their sample. • Neither Kwan (1996) nor Hotchkiss and Ronen (2002) reverse their model specification in order to explicitly test whether lagged bond returns can explain stock returns.

11. Previous evidence: Event studies Allen, Guo and Weintrop (2004): • Compare abnormal returns in the syndicated bank loan market to equity markets in response to information about earnings. • Find that negative earnings announcements are reflected approximately one month earlier in bank loan prices than in stock prices. • The abnormal returns coincide with the private release of earnings information to syndicate members as mandated by bank loan covenants. • These results suggest that negative information about firm-specific events is incorporated in the bank loan market and transmitted to the stock market. • Although the results are consistent with the asymmetric price reaction hypothesis, they are not tests of integration across markets.

12. Previous evidence: Event studies Altman, Gande and Saunders (2004): • Compare abnormal returns around default announcement dates in syndicated bank loan, public debt and equity markets. • Find that syndicated bank loan markets lead all public markets in reacting to default announcements. • Syndicated bank loan secondary market prices fall significantly more during the preannouncement period than do bond prices or stock prices. • Moreover, the announcement effect is smaller in the syndicated bank loan market than in either the bond or equity markets.

13. Previous evidence: Event studies • By focusing on single events, these papers bias their results in favor of the private information hypothesis and the asymmetric price reaction hypothesis to the detriment of testing the liquidity hypothesis. • After all,a market-moving event such as an impending default will most likely overcome illiquidity constraints in loan markets, thereby biasing results in favor of the private information hypothesis. • For example, Green (2004) examines transaction data in the US Treasury market and finds that the more substantial the information release (e.g., the greater the surprise or precision of the macroeconomic announcement), the less important the liquidity considerations (in terms of order flow data) in impacting prices.

14. Previous evidence: Event studies • Some support for the integrated market hypothesis comes from Chen, Lung, and Tay (2005), who show that informed participants trade in both the equity and options markets, first accessing the greater liquidity of the equity markets (so that stock returns lead options trading), but preferring the options markets in the presence of information asymmetries. • Thus, traders use all available financial markets to opportunistically benefit from information. Prices will adjust rapidly to information revealed in other companion markets. • This process of informational integration and efficiency is consistent with the integrated markets hypothesis.

15. Key Variables Measures of return and year • RBt= loan return • RSt= equity return • RLt= loan index return: S&P/LSTA Loan Index- weekly returns on a portfolio of almost 500 loan facilities • RMt= equity index return: S&P 500 Composite Index from CRSP • RDt= T-Bill return • Yt= year of the observation

16. Key Variables Measures of loan market liquidity • NBAt= sum of bids and asks for the loan • SPRDt= relative loan spread • TERM = 1 if the loan is a term loan • REVOLVER = 1 if the loan is a revolver loan Measures of equity market liquidity • Vt= volume of trades on the equity market • ESPRDt = relative equity spread • ENBAt = sum of equity bid and asks / 1000

17. Key Variables Measures of Private Information • COV = 1 if financial covenants present • DISTRESSt = 1 if loan price <= 70 • PARt = 1 if loan price >= 90 • INTANGIBLE= 1 if borrower has 2-digit SIC codes 28, 73, 37, 35, 36 and 38 • DUALMMt = 1 if at least one lender associated with the given syndicated loan is also a market maker for the equity of the borrower on the day of the observation, and zero otherwise.

18. Sample selection methodology • We obtain a sample of secondary market data from the Loan Pricing Corporation (LPC) that consists of the average bid and average ask quotations on all syndicated bank loans that had at least 2 quotes on a given date for any week during the January 1999 through May 2003 period. • Initial sample size: • Secondary market observations: 129,172 • Loan facilities: 1,621 • Borrower: 763 • After filtering for key variables and lag return availability: • Secondary market observations: 43,578 • Loan facilities: 719 • Borrower: 334

19. Descriptive statistics, key variables

20. Primary market and price characteristics

21. Primary market and price characteristics

22. Financial covenants

23. General covenants

24. Liquidity and price characteristics

25. Empirical Methodology: Overview Three types of empirical tests: 1. SUR (Tables 5, 8, 9, 12, 14, 15, 18) (1) (2)

26. Empirical Methodology: Overview 2. Granger causality tests (Tables 6, 10, 11, 13, 16, 17, 19) (3) (4) • Each of the above equations are OLS estimated twice: with/without the lagged value of the other market. • The sum of square residuals are compared using F-tests and asymptotically equivalent tests of whether β2 = 0.

27. Empirical Methodology: Overview 2. Abnormal returns tests (Table 7) • For every week in the time period, we separately form equally weighted portfolios consisting of loan return observations for which the lagged equity return in excess of the T-bill return is positive or negative. • We simulate a portfolio consisting of long positions in loans with positive lagged equity returns and short positions in loans: Loan return, positive lagged equity returns portfolio – Loan return, negative lagged equity returns portfolio • We also test equity returns using lagged loan return portfolios: Equity return, positive lagged loan returns portfolio – Equity return, negative lagged loan returns portfolio

28. Empirical Methodology: Overview • Samples tested: • Full sample integration tests • (Tables 5, 6, 7) • Direct tests of the private information hypothesis • (Panel A of Tables 8, 9, 10, 11) • Direct tests of the liquidity hypothesis • (Panels B and C of Tables 8, 9, 10, 11) • Direct tests of the asymmetric price reaction hypothesis • (Tables 12, 13, 14, 15, 16, 17) • Direct tests of the integrated market hypothesis • (Tables 18 and 19)

29. Full sample integration tests • Table 5: regression tests • Lagged values of both loan and equity returns have explanatory power with regards to each other. • Consistent with the integrated market hypothesis.

30. Full sample integration tests • Table 6: Bivariate Granger Causality Tests • Results appear to reject null hypothesis that lagged values of returns do not Granger cause the returns of each other. • Consistent with the integrated market hypothesis.

31. Full sample integration tests • Table 7: Abnormal Returns on Long/Short Portfolios • Alpha is estimated in a single factor regression of the return on the long/short portfolio on the market index. • Results appear to support the contention that equity markets lead the loan market. • No evidence that loan markets lead equity markets. • We will not present tables for the remaining results.

32. Direct tests of the private information hypothesis • (Panel A of Tables 8, 9, 10, 11) • Greater private information: • Loans with financial covenants • Distressed loans, which are intensely monitored. • Loans to intangible firms, which are intensely monitored due to information asymmetries. • Less private information: • loans without financial covenants • Par loans • Results generally do not suggest that private information results in the loan market leading the equity market.

33. Direct tests of the liquidity hypothesis • (Panels B and C of Tables 8, 9, 10, 11) • Greater loan liquidity: High number of loan quotes. Term loans. • Less loan liquidity: Low number of loan quotes. Revolver loans. • Unclear loan liquidity: • High loan spreads: May be less liquid, or may be associated with liquid distressed loans • Low loan spreads: May be more liquid, or may be associated with less liquid par loans. • Greater equity liquidity: High equity volume. Low equity spreads. High number of equity quotes • Less equity liquidity: Low equity volume. High equity spreads. Low number of equity quotes

34. Direct tests of the liquidity hypothesis • Results suggest that equity market returns Granger cause loan market returns regardless of equity or loan liquidity. • Results provide some support for liquidity hypothesis: Loan market returns more likely to Granger cause equity market returns when loan market is more liquid.

35. Direct tests of the asymmetric price reaction hypothesis • “Good news” subsample: loan return - loan index > 0 • “Bad news” subsample: loan return - loan index <= 0 • If accurate, loan markets will lead by the “bad news” subsample, and equity market will lead by the “good news” subsample. • Results do not support the asymmetric price reaction hypothesis: Loan (equity) returns Granger cause equity (loan) returns for both good and bad news groups. • Further, Granger subgroup tests (Tables 16 and 17) do not identify a consistent pattern across “good news” and “bad news” with regard to either information-advantaged subsegments or liquid-market subsegments.

36. Direct tests of the integrated market hypothesis • “Dual Market Maker” subsample: at least one lender associated with the given syndicated loan facility is also a market maker for the equity of the borrower on the day of the observation. • “No Dual Market Maker” subsample: No lender associated with the given syndicated loan facility is also a market maker for the equity of the borrower on the day of the observation. • We should expect superior market integration for the dual market maker subsample. • Loan returns only Granger cause equity returns when the market maker is also a lender. • Suggests that information flows freely from equity markets to loan markets, but that access to private information (available to syndicate members) is required to enhance the flow of information from loan markets to equity markets.

37. Conclusions • We find evidence consistent with considerable market cointegration across equity and syndicated bank loan markets. • Information flows even more freely when the same financial intermediary acts as an equity market maker and a member of a loan syndicate. • Extremely rich database for future work: • How are the liquidity characteristics of the equity market and structural characteristics of loans affected by the presence of equity market maker/syndicate members?