Equity investor sentiment and bond market reaction: Test of overinvestment and capital flow hypotheses,☆☆

https://doi.org/10.1016/j.finmar.2020.100589Get rights and content

Highlights

  • Equity investor sentiment is negatively correlated with contemporaneous bond returns.

  • The negative relation between equity sentiment and contemporaneous bond returns is stronger for overinvestment firms.

  • Equity investor sentiment is positively correlated with subsequent bond returns.

  • The positive relation between equity sentiment and subsequent bond returns is mitigated for overinvestment firms.

  • Bond liquidity moderates the relation between equity sentiment and bond returns via the capital flow across markets.

  • Equity investor sentiment leads to a delayed bond rating downward revision for overinvestment firms.

Abstract

In this paper, I examine the effect of equity investor sentiment on the bond market. Sentiment can affect bond returns via two channels. First, in times of high investor sentiment, overvalued equity can lead to firm overinvestment, resulting in a negative impact on bond pricing due to an increase in default risk. Second, overvalued equity attracts capital flow to the equity market from the bond market, which can create a downward pressure on bond pricing. Consistent with these channels, I find that equity investor sentiment exhibits a significant negative relation with contemporaneous bond returns. This effect is stronger for firms that overinvest. Furthermore, I find a positive relation between sentiment and subsequent bond returns, consistent with a return reversal predicted by the capital flow channel (due to the backflow of capital); however, there is no return reversal observed for overinvestment firms. Additionally, I find a negative (but delayed) impact of equity investor sentiment on bond ratings for overinvestment firms, consistent with an increase in default risk of these firms. Overall, my study highlights that the bond market reacts to equity investor sentiment in a rational way, consistent with bond investors’ payoff function.

Introduction

Investor sentiment refers to “a belief about future cash flows and investment risks that is not justified by the facts at hand” (Baker and Wurgler, 2007, p. 1). De Long et al. (1990, 1991) propose models showing how sentiment can move asset prices in the financial markets in the presence of noise traders with erroneous stochastic beliefs. The findings of several empirical studies suggest that during periods of high market-wide sentiment, equity securities are overvalued as evidenced by reversals of subsequent returns (e.g., Brown and Cliff, 2005; Baker and Wurgler, 2006, 2007; Ben-Rephael et al., 2012; Stambaugh et al., 2012). While the impact of investor sentiment on equity pricing has been studied extensively, there is limited research examining whether equity investors’ sentiment has any effect on debt market pricing, especially corporate bond pricing. The relation between the stock market and the bond market is unclear theoretically1 and mixed empirically (e.g., Shiller, 1982; Baele et al., 2010; Chordia et al., 2017) given they differ in many respects. Previous studies show that the correlation between stock returns and bond returns is low and stock anomalies may or may not extend to the bond market (e.g., Chordia et al., 2017). Therefore, it is not straightforward to predict the relation between equity market sentiment and bond pricing, given the findings in the equity market. In this paper, I develop and test hypotheses that relate equity market sentiment to bond market pricing. An investigation on the relation and possible mechanisms between equity investor sentiment and bond market movement could help to advance our knowledge about the bond market and the stock-bond relation.

The finance literature documents that firms overinvest during high sentiment periods (e.g., Arif and Lee, 2014), either due to manager empire building or managers catering to noise traders, among other reasons (e.g., Titman et al., 2004). Furthermore, empirical evidence shows that corporate overinvestment results in greater earnings disappointments in subsequent periods (e.g., Titman et al., 2004; Arif and Lee, 2014) and leads to an increase in firm default risk (e.g., Altman, 1968; Zmijewski, 1984; Callen et al., 2009). Given the concave payoff structure in bonds, bondholders are more sensitive to downside risk (Bai et al., 2019) and are expected to view overinvestment negatively because it is associated with higher firm default risk and reduction in bond value (e.g., Jensen and Meckling, 1976; Hong and Sraer, 2013). Thus, I hypothesize that, through the overinvestment channel, high equity market sentiment will have a negative impact on bond pricing.2

The negative impact of high equity market sentiment on the bond market could also arise through another channel that is related to the flow of capital across the bond and equity markets. Prior studies argue that returns influence investors’ trading such that positive (negative) returns lead to more (fewer) trades (e.g., Ben-Rephael et al., 2011). Following this argument, several studies find that contemporaneous equity returns are strongly associated with re-balancing decisions between the equity and bond markets (e.g., Warther, 1995; Edwards and Zhang, 1998; Fant, 1999). Ben-Rephael et al. (2012) argue that shifts between bond funds and equity funds are indicative of investor sentiment. They show that the net exchange of equity funds is positively correlated with contemporaneous aggregate equity market returns and that this relation reverses in subsequent months. Consistent with these studies, I expect greater capital inflow to the bond market during low equity sentiment periods and greater capital outflow from the bond market during high equity sentiment periods. The capital inflow (outflow) generates positive (negative) price pressure on the traded securities and can lead to overvaluation (undervaluation) (e.g., Coval and Stafford, 2007; Khan et al., 2012). As long as individual bonds have downward-sloping demand curves, the capital flow channel predicts a negative relation between equity investor sentiment and contemporaneous bond returns, on average. Further, subsequent to a high sentiment period, a reversal of bond returns is expected due to the capital backflow from the equity market to the bond market; thus, I predict a positive relation between equity investor sentiment and subsequent bond returns, on average.

I also examine a competing hypothesis, the contagion effect, which predicts a positive rather than a negative contemporaneous bond market reaction to equity market sentiment. Stocks and bonds issued by the same firms represent claims on the same underlying assets. Any change in the expectation of future cash flows conceptually applies to both stocks and bonds. This suggests a contagion effect of sentiment from the stock market to the bond market. Opposite to the effects predicted by the capital flow channel, the contagion hypothesis predicts that equity investor sentiment is positively related to contemporaneous bond returns and negatively related to subsequent bond returns, on average.

I measure equity investor sentiment and corporate bond returns on a quarterly basis. Baker and Wurgler (2006) develop a measure of equity investor sentiment based on the first principal component of six standardized sentiment proxies (adjusted for a common business cycle component). The sentiment proxies include the closed-end fund discount, NYSE share turnover, the number of IPOs, the average IPO first-day returns, the equity share in total new issues, and the dividend premium. To test the effect of the overinvestment channel, I use the residual from the capital expenditure investment model in Polk and Sapienza (2009) as a proxy for the deviation from the optimal investment level (CAPXRES). Following Biddle et al. (2009), I assign firm-quarter observations to the overinvestment (no-overinvestment) group if the residual falls in the top tercile (bottom two terciles) of the sample. I test the differential relation between equity investor sentiment and contemporaneous bond returns for the overinvestment and the no-overinvestment groups. I control for Fama-French risk factors, default risk, term risk, and bond characteristics associated with bond returns as documented by previous studies. Both time and industry fixed effects are included in the regressions.

Using the U.S. bond transaction data from 2003 to 2016, I first verify that equity investor sentiment is positively associated with firm overinvestment.3 Since the overinvestment and the capital flow channels have the same directional prediction with respect to the relation between equity investor sentiment and contemporaneous bond returns, it is difficult to attribute the result to one or the other of the channels. I attempt to resolve this issue by contrasting the effects on contemporaneous versus subsequent bond returns. If across-market capital flows drive the observed contemporaneous relation, I expect a reversal in subsequent bond returns due to the backflow of capital when the high equity investor expectations decline; such reversal is less likely if overinvestment drives the relation because the effects of overinvestment are likely to persist over a longer time period.

Overall, I find a significant negative relation between equity investor sentiment and contemporaneous bond returns. Using the proxy of overinvestment, CAPXRES, I find that the observed negative effect is significant for both the overinvestment and the no-overinvestment groups, consistent with the predicted effects of both channels, overinvestment and capital flow. Moreover, this finding rules out the contagion effect hypothesis, which predicts a positive relation. In addition, the relation is significantly more negative for the overinvestment group, supporting the effect of the firm overinvestment channel. To further distinguish the effects of the overinvestment and the capital flow channels, I test the relation between equity investor sentiment and the subsequent 12-month’s bond returns. If the effect is mainly due to the capital flow channel, I expect a positive relation for both groups to reflect the reversal of returns in the subsequent period due to the backflow of capital. I find a significant bond return reversal subsequent to the high sentiment period for the no-overinvestment group; however, the overinvestment group exhibits an insignificant relation between sentiment and subsequent bond returns. Thus, my results suggest that, when there is no firm overinvestment, the contemporaneous negative bond market reaction to equity investor sentiment can be attributed to the flow of capital from the bond market. On the other hand, given that I do not observe a subsequent return reversal, the capital flow channel is less likely to be the primary explanation for the overinvestment group; the bond investors’ negative reaction is at least partially due to the effect of overinvestment resulting in a higher risk exposure for bondholders.

Next, I test the cross-sectional variation in the effect of the capital flow channel based on the level of bond liquidity. Prior literature finds that the price pressure generated by capital overflow has less influence on liquid bonds than on illiquid bonds (e.g., Brandt and Kavajecz, 2004). Consistent with this evidence, I find that the relations between equity investor sentiment and both contemporaneous and subsequent bond returns concentrate in the subsample of illiquid bonds within the no-overinvestment group. However, as expected, I do not observe a differential relation for the overinvestment group, confirming overinvestment and capital flow as two distinct channels through which equity investor sentiment influences the bond market.

Finally, I examine the impact of equity investor sentiment on bond rating changes. Since bond ratings reflect the default risk of bond issues, a change in the expectation of the issuer’s ability to fulfill its financial obligations should lead to a bond rating revision. Consistent with overinvestment increasing a firm’s default risk, I find a positive relation between equity investor sentiment and downward revisions of bond ratings for the overinvestment group, but only in the subsequent period. This is consistent with the argument that bond rating agencies update bond ratings with a delay (e.g., Pinches and Singleton, 1978; Beaver et al., 2006). In contrast, I find an insignificant effect on bond ratings for the no-overinvestment group, consistent with the fact that the negative effect on bond returns due to the capital flow channel does not imply a change in bond default risk.

In addition, I conduct a battery of robustness tests. Specifically, I show that the main results and references are not driven by the financial crisis of 2008 and are not sensitive to the alternative proxies of firm overinvestment and equity investor sentiment.

My paper contributes to the literature on equity market investor sentiment by examining its effects on the bond market. Prior empirical studies on investor sentiment mainly focus on the effect of sentiment on the behavior of equity market participants.4 I focus on the bond market and provide empirical evidence on the negative relation between sentiment and contemporaneous bond returns, consistent with the effects of both the flow of capital and overinvestment channels. Similar to bond investors, rating agencies respond negatively to sentiment (although with a delay), but only for the overinvestment group due to their (potentially) higher default risk. Overall, my study highlights that behavioral biases in the equity market do not automatically get transmitted to the bond market. In fact, the bond market reacts negatively to sentiment-induced overinvestment in a rational way, consistent with bond investors’ payoff functions.

The remainder of this paper unfolds as follows. I develop my hypotheses in Section 2. In Section 3, I describe the data, sample selection procedure, and definitions of the main variables. In Section 4, I present the details of the research design and discuss the results. Additional tests are presented in Section 5. I conclude in Section 6.

Section snippets

Hypothesis development

In this section, I discuss different channels through which equity investor sentiment could affect bond pricing and develop hypotheses accordingly.

Data and sample selection

The data come from a variety of sources. Bond price and trade data are from the Trade Reporting and Compliance Engine (TRACE) transactions database. Ratings and bond characteristics data are from the Mergent Fixed Income Securities Database (FISD). Accounting variables are from Compustat and stock return data are from CRSP. The equity investor sentiment index is obtained from Jeffrey Wurgler’s website.7 The stock market (MKT) and SMB and HML factors are

Descriptive statistics

Table 2 provides descriptive statistics of the primary variables of interest for the final sample. The mean quarterly (annual) bond returns adjusted for the T-bill return is 1.2% (2.3%) consistent with prior studies (e.g., Bessembinder et al., 2009; Easton et al., 2009). The average bond rating score is around 9.0, which corresponds to Standard & Poor’s (S&P) BBB group (Moody’s Baa2 group).14 The average coupon

Relation between equity investor sentiment and bond ratings

In this section, I examine bond rating revisions to provide more evidence on the effect of equity investor sentiment on the bond market. This analysis helps to differentiate the effects of the overinvestment and capital flow channels. Credit ratings for corporate bonds represent a summary of forward-looking opinions about the credit risk of bond issues and issuers’ ability to meet the financial obligations.19 Changes in the

Conclusion

Equity investor sentiment has been shown to have a significant impact on the equity market. However, limited empirical evidence has been offered regarding its effect on the bond market. In this study, I examine whether equity investor sentiment affects the bond market. I show that equity investor sentiment affects bond returns through (i) the price pressure generated by money reallocation across the equity and bond markets as a result of investors chasing high returns (capital flow channel),

References (55)

  • M.C. Jensen et al.

    Theory of the firm: managerial behavior, agency costs and ownership structure

    J. Financ. Econ.

    (1976)
  • S.H. Kwan

    Firm-specific information and the correlation between individual stocks and bonds

    J. Financ. Econ.

    (1996)
  • H. Lin et al.

    Liquidity risk and expected corporate bond returns

    J. Financ. Econ.

    (2011)
  • J.J. McConnell et al.

    Corporate capital expenditure decisions and the market value of the firm

    J. Financ. Econ.

    (1985)
  • C. Shi

    On the trade-off between the future benefits and riskiness of R&D: a bondholders’ perspective

    J. Account. Econ.

    (2003)
  • R. Shiller

    Consumption, asset markets, and macroeconomic fluctuations

    Carnegie-Rochester Conf. Ser. Public Policy

    (1982)
  • R. Shiller et al.

    Stock prices and bond yields: can their comovements be explained in terms of present value models?

    J. Monetary Econ.

    (1992)
  • R.F. Stambaugh et al.

    The short of it: investor sentiment and anomalies

    J. Financ. Econ.

    (2012)
  • J.C. Stein

    Agency, information and corporate investment

    Handb. Econ. Finance

    (2003)
  • V.A. Warther

    Aggregate mutual fund flows and security returns

    J. Financ. Econ.

    (1995)
  • E.I. Altman

    Financial ratios, discriminant analysis and the prediction of corporate bankruptcy

    J. Finance

    (1968)
  • S. Arif et al.

    Aggregate investment and investor sentiment

    Rev. Financ. Stud.

    (2014)
  • L. Baele et al.

    The determinants of stock and bond return comovements

    Rev. Financ. Stud.

    (2010)
  • M. Baker et al.

    When does the market matter? Stock prices and the investment of equity-dependent firms

    Q. J. Econ.

    (2003)
  • M. Baker et al.

    Investor sentiment and the cross-section of stock returns

    J. Finance

    (2006)
  • M. Baker et al.

    Investor sentiment in the stock market

    J. Econ. Perspect.

    (2007)
  • J. Bao et al.

    The illiquidity of corporate bonds

    J. Finance

    (2011)
  • Cited by (8)

    View all citing articles on Scopus

    I am deeply indebted to my adviser, Pervin Shroff, for her invaluable guidance and support. I thank Ivy Xiying Zhang, Jianfeng Yu, and Paul Glewwe. I appreciate helpful comments from Tarun Chordia (the editor), Daniel Collins, Yu Gao, Andrew Winton, Haibin Wu, Liandong Zhang, the anonymous reviewer, and seminar participants at the University of Minnesota, National University of Singapore, Singapore Management University, Hong Kong University of Science and Technology, Hong Kong Polytechnic University, the Chinese University of Hong Kong, New York University, Baruch College, City University of Hong Kong, and Peking University.

    ☆☆

    The financial support from City University of Hong Kong is greatly acknowledged.

    View full text