## SHRIHARI SANTOSH

*Associate Professor of Finance*

**Research Interests**: empirical and theoretical asset pricing, behavioral economics, machine learning

# PUBLISHED and FORTHCOMING

**Journal of Finance**,* *June 2018, 73(3), 1183-1223 (with Serhiy Kozak and Stefan Nagel)

**ABSTRACT: **We argue that empirical tests of reduced-form factor models do not shed light on competing theories of investor beliefs. Since asset returns have substantial commonality, absence of near-arbitrage opportunities implies a stochastic discount factor (SDF) that is a function of a few dominant sources of return variation. Consistent with this view, we show: an SDF based on the first few principal components explains many recently studied anomalies; if this was not true, near-arbitrage opportunities with extremely high Sharpe Ratios would exist; in-sample near-arbitrages vanish out-of-sample. However, a reduced-form factor SDF of this kind is perfectly consistent with an economy in which all cross-sectional variation in expected returns is caused by sentiment. Components of sentiment-investor demand that line up with common factor loadings affect asset prices because it is risky for arbitrageurs to take the opposite position, while components orthogonal to these factor loadings are neutralized. If investor sentiment is time-varying, the SDF can take the form of an ICAPM. For these reasons, tests of reduced-form factor models, horse races between “characteristics” and “covariances,” and firm investment-based models that take as given an arbitrary or a reduced-form factor SDF cannot discriminate between alternative models of investor beliefs.

**Journal of Financial Economics***, *February 2020, 135(2) 271-292, **Lead Article** (with Serhiy Kozak and Stefan Nagel)

2020 1st Place JFE Fama-DFA Prize

**ABSTRACT: **We construct a robust stochastic discount factor (SDF) that summarizes the joint explanatory power of a large number of cross-sectional stock return predictors. Our method achieves robust out-of-sample performance in this high-dimensional setting by imposing an economically motivated prior on SDF coefficients that shrinks the contributions of low-variance principal components of the candidate factors. While empirical asset pricing research has focused on SDFs with a small number of characteristics-based factors—e.g., the four- or five-factor models discussed in the recent literature—we find that such a characteristics-sparse SDF cannot adequately summarize the cross-section of expected stock returns. However, a relatively small number of principal components of the universe of potential characteristics-based factors can approximate the SDF quite well.

**Journal of Financial Economics**, September 2020, 137(3), 740-751 (with Serhiy Kozak)

**ABSTRACT: **Controlling for changes in wealth, the price of "discount-rate" risk reveals whether increases in equity risk premia represent "good" or "bad" news to rational investors. We employ a new empirical methodology and find that the price is negative. This finding suggests that discount rates are high during times of high marginal utility of wealth. Our approach relies on using future realized market returns to consistently estimate covariances of asset returns with the market risk premium. Covariances drive observed patterns in a broad cross-section of stock and bond expected returns.

**Review of Financial Studies**, May 2020, 33(5), 1980-2018 (with Valentin Haddad and Serhiy Kozak)

2018 Q-Group Jack Treynor Prize

**ABSTRACT: **An optimal factor timing portfolio is equivalent to a conditional SDF. We use economic restrictions to determine and characterize both empirically. With these restrictions, we find that long-short equity factors are strongly and robustly predictable. A number of these portfolios have small or zero average price of risk, suggesting that the economic risks investors worry about conditionally are often very different from those they worry about on average. This manifests in the very different compositions of the conditional and unconditional SDFs. For bonds and foreign exchange strategies,long-short portfolios sorted on maturity or interest rate differential are also predictable.

**Review of Financial Studies**, October 2022 (with Christian L. Goulding and Xingtan Zhang)

**ABSTRACT:** We study the interaction of noisy demand and skewed asset payoffs. In our model, price as a function of quantities is convex in a neighborhood around zero if and only if skewness is positive. The combination of convexity and noise produces the idiosyncratic skewness effect--a documented negative relationship between an asset's idiosyncratic skewness and its expected return. We further offer an explanation for the idiosyncratic volatility puzzle. Finally, our theory predicts that higher idiosyncratic skewness strengthens the idiosyncratic volatility effect (and vice versa). We find support for this prediction in the cross section of stock returns.

**Critical Care Explorations**, May 2022, 4(5) (with John Roddy, Daniel Wells, Katharine Schenck, and Sadashiv Santosh)

**ABSTRACT: **Patients were divided into tocilizumab and baricitinib cohorts based on actual medication received. The primary outcome was hospital discharge alive and free from mechanical ventilation within 60 days, assessed by logistic regression. Three hundred eighty-two patients were included: 194 in the tocilizumab cohort and 188 in the baricitinib cohort. Most baseline characteristics in the two cohorts were similar. All patients received dexamethasone. Two patients were lost to follow-up. In the remaining 380 patients, probability of successful discharge in the two cohorts was quantitatively similar in unadjusted, multivariate-adjusted, and propensity score-matched analyses. Hospital length of stay, rates of thromboembolic events, and rates of hospital-acquired infections were all similar in the two cohorts.

**Critical Care Explorations**, April 2020, 2(4) (with Robert Sbertoli, Zeyu Hu, Jonathan Henke, Eric Wu, Stephen Osmon, Edward Charbek, Zafar Jamkhana, and Sadashiv Santosh)

**ABSTRACT: **Patients were divided into low-dose hydrocortisone (75–150 mg/d) and standard-dose hydrocortisone (200–400 mg/d) cohorts based on initial prescribed hydrocortisone dose. Rates of shock reversal and adverse events in the two cohorts were compared. Two-hundred thirteen patients were included—41 in low-dose and 172 in standard-dose cohorts. Baseline characteristics including initial vasopressor requirement and Sequential Organ Failure Assessment scores were similar. Average rates of change in vasopressor needs, conditional hazard rate for vasopressor withdrawal, and cumulative probability for vasopressor withdrawal were all quantitatively similar for low-dose and standard-dose hydrocortisone. Insulin requirement (particularly in those with diabetes mellitus), blood glucose in those with diabetes mellitus, and frequency of secondary infections seemed to be lower in the low-dose hydrocortisone cohort. Mortality and other secondary outcomes were similar.

**Southern Medical Journal**, December 2016, 109(12), 785-791 (with Conor McCartney, Setu Patolia, and Sadashiv Santosh)

**ABSTRACT: **Ninety-nine hospital admissions from 76 patients met the inclusion criteria. Patients received mean 2.9 ± 2.6 red blood cell (RBC) units. In the multivariate analysis, increased prothrombin time, low initial hemoglobin, admission to intensive care unit, congestive heart failure, white race (P = 0.08), and syncope (P = 0.09) were independent predictors of RBC transfusion. A total of 28% received fresh frozen plasma and 8% received platelets. Prolonged prothrombin time was the only independent predictor of fresh frozen plasma transfusion (P < 0.001). Platelet transfusion was predicted by thrombocytopenia at platelet counts <100,000/mm3 (P < 0.001) and white race. Coronary artery disease was associated with reduced platelet transfusion. Other baseline comorbidities, bleeding source, and active/recent hemorrhage on endoscopy did not predict transfusion. Desmopressin use was not associated with reduced RBC needs, even after adjusting for other covariates. Ninety-seven percent of patients survived to discharge.

# WORKING PAPERS

## Smart Money and Investor Sentiment *NEW*

*NEW*

(with Brian Waters, Xingtan Zhang, and Wei Zhou)

**ABSTRACT:** We study a dynamic asset pricing model with smart money and sentiment traders. Since sentiment spreads across traders from period to period, trading due to sentiment is persistent. Smart money's superior fundamental information also confers a superior ability to predict future sentiment-based trades. We find that the horizon at which sentiment can be predicted is a key determinant of smart money's trading behavior. When predictability is short-term, smart money trades against sentiment traders which mitigates mispricing. However, when smart money can predict longer-term sentiment, it can be profitable for smart money to trade in the same direction as sentiment traders in order to manipulate market participants' beliefs about future sentiment-based trades. In doing so, smart money increases the opportunity to exploit this non-fundamental information in the future. The model also sheds light on conflicting evidence that relates sentiment-based trading to past, present, and future asset prices.

## Disagreement, Skewness, and Asset Prices *NEW*

*NEW*

(with Christian L. Goulding and Xingtan Zhang)

**ABSTRACT:** We present a frictionless model which bridges two seemingly unrelated empirical anomalies: (1) the negative relationship between dispersion in financial analysts’ forecasts and expected returns and (2) the negative relationship between idiosyncratic skewness and expected returns. The results obtain because (1) empirically, most stocks have positive expected skewness, (2) positive skewness implies that investors' demand schedules are convex in the relevant price range, and hence, (3) trades due to disagreement do not ``cancel out"; asset prices are inflated even without short-selling constraints. Our theory further predicts that skewness and disagreement have an interactive pricing impact. We find support for this prediction in the cross section of stock returns.

**ABSTRACT:** Investors' return on their portfolios, as proxied by the market, is a theoretically appealing but empirically unsuccessful asset pricing factor. In practice, many institutional investors choose to deviate substantially from the market portfolio. We propose a simple model in the spirit of Merton (1987) which implies that an asset's expected return is linear in its average idiosyncratic beta with respect to each active investor's portfolio return. We estimate this relation using 13F holdings data for active investors and find that a unit increase in investor betas commands 5-10\% greater annual expected return. The results are robust to alternative ways of estimating betas and using daily or monthly returns. In sum, investors appear to be compensated for holding stocks that have a high covariance with the idiosyncratic component of their portfolio.

**ABSTRACT:** We present a portable model of distorted learning which embodies Tversky and Kahneman’s “belief in the law of small numbers.” When adjusting beliefs in response to new information the decision maker overweights the sample, updating as if the sample size were inflated. The degree of distortion is embodied in a single parameter specific to the agent and not to the particular stochastic setting. We show that the beliefs of such an agent preserve many dynamic properties of fully rational Bayesian beliefs. Though exaggerated likelihoods delivers similar predictions to diagnostic expectations in a static setting, the models imply dramatically different belief dynamics. We present examples of distorted Kalman filtering in a Gaussian environment as well as a non-linear setting with stochastic volatility.

**ABSTRACT:** I develop a dimensionless measure of the path of price discovery using public information shocks as an instrument for changes in firm value. I find that price discovery occurs largely through trading. Using high-frequency data, I find substantial cross-sectional variation in the speed of price of price discovery when measured in “real time.” This heterogeneity, however, essentially disappears when measured in “trade time.” Consistent with previous work work, I find that trade time is well approximated by the accumulation of transactions, not share volume, dollar volume, or turnover. These findings support the hypothesis of Kyle and Obizhaeva (2016b) that “information flows take place in the same business time as trading.”

## Behavioral Biases, Order Imbalance, and Expected Returns (with Yuan Hou)

**ABSTRACT: **We test the hypothesis that anomalous expected returns are caused by biased investor demand coupled with imperfect competition among rational investors. A unique data set on short-horizon digital options, which allows us to track prices as well as the imbalance between "buy" and "sell" orders, enables us to trace the path from behavioral biases to order imbalance to distorted prices and expected returns. The conditional information ratio (Sharpe ratio) from exploiting these behavioral biases is on average 0.18 per event net of transaction costs