Investor Sentiment and Market-Wide Liquidity Pricing

Abstract

The recent asset pricing evidence on the return-liquidity risk relationship is mixed and somewhat ambiguous. We reevaluate the importance of market-wide liquidity and liquidity risk for equity pricing by taking the role of investor sentiment into account. Regarding the market-wide liquidity level as a systematic factor, we find that high market sentiment tends to weaken the effect of market-wide illiquidity - both expected and unexpected - on stocks returns. With respect to systematic liquidity exposure as a priced risk factor, the results show that the effect of exposure to shocks in aggregate liquidity on expected returns is significantly positive when sentiment is low, while it is significantly negative when sentiment is high suggesting that “rational” asset pricing is only valid in the low sentiment regime without too much turbulence caused by sentiment traders.

Keywords:

• Investor sentiment
• asset pricing
• aggregate liquidity
• liquidity risk

JEL Classification:

• G11
• G12
• G14
• G17

Notes

1 For an extensive review of this literature, see Amihud, Mendelson and Pedersen (2005, 2013).

2 Numerous studies argue that the beliefs of many stock market investors share a common time-varying sentiment component with the potential to influence stock prices in the same direction at the same time (e.g. Brown and Cliff, 2004, 2005; Yuan 2005, Baker and Wurgler, 2006, 2007; Yu and Yuan 2011; Gao, Yu, and Yuan 2012; Antoniou, Doukas, and Subrahmanyam 2013; Yu, 2013).

3 Many studies document that there exist short-sale impediments in the stock market. These impediments arise from, but not limited to, high shorting costs (D'Avolio, 2002), institutional constraints, risks in arbitrage (Pontiff, 1996; Shleifer and Vishny, 1997; and Wurgler and Zhuravskaya, 2002), and behavioral biases of traders (Barber and Odean, 2008). Empirical evidence shows also that individual traders, who are natural candidates for sentiment traders, seldom short (e.g. Barber and Odean, 2008).

4 Securities that are not common stock are excluded from our sample.

5 Tests are limited to NYSE-traded stocks to avoid some problems that can result from differences in market microstructures.

6 Ince and Porter (2006) show that after a careful screening, which mainly consists in excluding securities that are not common shares, inferences drawn from Thomson Reuters Datastream data are similar to those drawn from CRSP (Center for Research in Security Prices) data.

7 The AAII sentiment index is the result of a weekly survey conducted by the American Association of Individual Investors. The IIA sentiment index reflects the outlook of over 130 independent financial market newsletter writers. Both sentiment indices are computed as the spread between the percentage of bullish investors and the percentage of bearish investors (Bull–Bear).

8 Examples are Acharya and Pedersen (2005), Liu (2015) and Avramov, Cheng and Hameed (2016).

9 Amihud (2002) demonstrates that $\mathit{\text{ILLIQ}}$ is positively related to measures of price impact and fixed trading costs over the time period in which he has the microstructure data. Similarly, Hasbrouck (2009) finds a correlation of 0.668 between $\mathit{\text{ILLIQ}}$ and Kyle’s λ, which he derives from intraday data, and a correlation between $\mathit{\text{ILLIQ}}$ and effective costs, also estimated from intraday data, of 0.612. Goyenko, Holden, and Trzcinka (2009) also compare the 5-minute price impact estimated from intraday data and conclude that $\mathit{\text{ILLIQ}}$ is well suited as a measure of price impact.

10 $\mathit{\text{Adjusted illiquidity}}\left(t\right)=\mathit{\text{raw illiquidity}}\left(t\right)*\left(m_t/m_1\right),$ where $m_t$ is the total dollar value at the end of month t-1 of the stocks included in the average in month t, and month 1 corresponds to July 2001. The ratio $\left(m_t/m_1\right)$ serves as a common detrending factor that controls for the time trend in the illiquidity measure.

11 Using the logarithmic form of the average illiquidity across stocks eliminate the effect of outliers that are commonly observed during periods of low trading activity. An alternative method is used by Acharya and Pedersen (2005) who truncate the Amihud illiquidity measure to a level between 0.25 and 30 percent and normalize it before using it.

12 b is expected to be positive since higher illiquidity in a month raises expected illiquidity for the following month.

13 To investigate the relationship between the market excess return and the market illiquidity, one should not regress the observed market excess returns solely on the expected market illiquidity since Amihud (2002) shows that such an approach would generate biased coefficients. This bias can be eliminated by including the residual term of Equation (1). Fortunately, this residual can be interpreted as the unexpected innovation in illiquidity.

14 A similar methodology has been used by French, Schwert, and Stambaugh (1987) to examine the relationship between risk and stock excess returns.

15 Expected illiquidity is said to have a positive effect on the ex ante stock excess return (stock return in excess of the Treasury bill rate), since rational investors expect higher returns if they anticipate higher market illiquidity.

16 Unexpected illiquidity is considered as having a negative effect on current returns, because a sudden persistent increase in illiquidity raises stock expected returns and lowers stock prices.

17 Brown and Cliff (2004, 2005) provide evidence that, due to sentiment’s high persistence, the predictability of returns over long horizons is stronger than the predictability of near-term returns.

18 The term yield premium is the monthly change in the term spread, i.e., the difference between the 10-year U.S. Treasury bond and three month Treasury bill yields. The default yield premium is the monthly change in the default spread, i.e., the difference between the yields on Moody’s Baa or better corporate bond yield index and the yield on a 10-year constant maturity Treasury bond

19 See for example Fama and French (1989), Fama (1990), Chordia, Roll and Subrahmanyam (2001) and Amihud (2002).

20 The equal- (value-) weighted Amihud illiquidity ratio has a first-order serial correlation of 0.669 (0.701).

21 An AR(1) process is sufficient to remove autocorrelations in the aggregate Amihud illiquidity series.

22 We use the logarithmic transformation of market illiquidity following Amihud (2002).

23 Multiplying the fitted residual by negative one is done to convert the aggregate Amihud illiquidity ratio from a measure of illiquidity into a measure of liquidity.

24 We keep the number of independent variables to a minimum because running a regression where the number of regressors is large may add a lot of noise. The same estimation procedure is used by Lou and Sadka (2011).

25 Value-weighted portfolios are formed by weighting each stock in the quintile by its relative market value within the quintile the month prior to portfolio formation.

26 Fama-French Factors (MKT, SMB and HML) and the momentum factor (MOM) are collected from Kenneth French’s homepage < www.dartmouth.edu/∼kfrench/>.

27 The intuition for such a measure is that risk-averse market makers accommodate order flow from liquidity motivated traders or non-informational traders (Campbell, Grossman, and Wang 1993) In providing liquidity, they require compensation in the form of higher expected returns. The larger the order flow is, the higher compensation for the market markers, leading to greater volume-return reversals when current volume is high.