Knowledge that Transforms

We use slow and fast time-series momentum to characterize four stock market cycles—Bull, Correction, Bear, and Rebound. The steep market declines of Bears concentrate in high-risk states, yet predict negative expected returns, which is difficult to rationalize by most models of time-varying risk premia. Using a model to analyze slow and fast momentum strategies, we estimate both relatively high mean persistence and realization noise in U.S. stock market returns. Intermediate-speed momentum portfolios, formed by blending slow and fast momentum strategies, translate predictive information in market cycles into positive unconditional alpha, for which we propose a novel decomposition.

We hypothesize that the well-documented negativity bias, the psychological tendency to asymmetrically emphasize negative over positive aspects, can help explain several financial market phenomena: why most individuals hold strongly bearish views of both short- and long-term equity return distributions, why individuals exhibit heterogeneous beliefs, and the stock market participation puzzle. Using variation in the perceived risk of mortality from the swine flu pandemic as our primary proxy for an individual's negativity bias, we find strong support for our hypothesis even when controlling for alternative mechanisms including optimism, risk aversion, ambiguity aversion, and anxiety.

Using the COVID-19 pandemic as a laboratory, we show that asset markets assign a time-varying price to firms’ disaster risk exposure. The cross-section of stock returns reflected firms’ different exposure to the pandemic, as measured by their vulnerability to social distancing. As predicted by theory, realized and expected return differentials moved in opposite directions, initially widening and then narrowing. When inferred from market outcomes, firm resilience correlates mainly with exposure to social distancing: vulnerability to social distancing is priced in changes of firms’ expected returns, while measures of financial and environmental resilience are not.

We use variation in the access to a government credit certification program to estimate the financial and real effects of supporting small firms. This program was first implemented during the global financial crisis, but has remained active ever since, allowing us to analyze its effects both during recessions and recoveries. Eligible firms have access to government loan guarantees and a credit quality certification. We estimate real effects using a multidimensional regression discontinuity design. We find that eligible firms borrow more and at lower rates than non-eligible firms, allowing them to increase investment and employment during crises. Industry-level analysis shows reduced productivity heterogeneity in more exposed industries, which is consistent with improved credit allocation. However, when the economy is recovering the effects of the program are less pronounced and centered on the certification component. The cost-per-job in the recovery period is half of the one estimated for the crisis period (5784€ and 11,788€, respectively).

The component of the volatility of total factor productivity (TFP) that is orthogonal to the dividend price ratio is shown to have long-run predictive ability for excess market returns. This finding implies that TFP volatility should also predict real cash flows and/or real interest rates: it is found to mainly predict real cash flows through inflation. A model with endogenous growth, Epstein-Zin preferences and price rigidities reconciles both TFP volatility-driven long-run predictability and its real implications. Within the model, we justify the similar (to that of TFP volatility) predictive ability of a low-frequency notion of market volatility as well as the cross-sectional pricing of TFP volatility risk in alternative asset classes.

I examine the asset pricing implications of technological innovations that allow capital to displace labor: automation. I develop a theory in which firms with displaceable labor are negatively exposed to such technology shocks. In the model, firms optimally adopt technology to gain competitive advantage but in equilibrium competition erodes profits and decreases firm value. Empirically, I find that firms with high share of displaceable labor have negative exposure to technology shocks. A long-short portfolio sorted on this variable mimics macroeconomic measures of technology shocks. Negatively exposed firms earn a 4% annual return premium consistent with displacement risk from technological progress.

We document that banks that cut lending more during the Great Recession were lending to riskier firms ex-ante. To understand the aggregate implications of this sorting pattern, we build an assignment model in which banks have heterogeneous costs to take on risky loans and firms have different credit risks. In the model, aggregate loan volume depends on the entire distribution of bank holding costs and firm credit risks. We then use our model to recover the change in the distribution of bank holding costs during the Great Recession and show that it explains two-thirds of the decline of aggregate loan volume during this period.

Open banking facilitates data sharing consented to by customers who generate the data, with the regulatory goal of promoting competition between traditional banks and challenger fintech entrants. We study lending market competition when sharing banks’ customer transaction data enables better borrower screening for fintechs. Open banking promotes competition if it helps level the playing field for all lenders in screening borrowers; however, if it over-empowers fintechs, it can also hinder competition and leave all borrowers worse off. Due to the credit quality inference from borrowers’ sign-up decisions, this remains true even if borrowers have the control of whether to share their banking data. We also study extensions with fintech affinities and data sharing on borrower preferences.

We show that exchange rate correlations tend to be explained by the global trade network while consumption correlations tend to be explained by productivity correlations. Sharing common trade linkages with other countries increases exchange rate correlations beyond bilateral linkages. We explain these findings using a model of the global trade network with market segmentation. Interdependent global production generates international comovements, while market segmentation disconnects the drivers of exchange rate correlations from the drivers of consumption correlations. Moreover, we show that the trade network generates common factors found in exchange rates. Our findings offer a trade-based account of the origins of international comovements and shed light on important frictions in international markets.

We construct a measure of systematic default defined as the probability that many firms default at the same time. We account for correlations in defaults between firms through exposures to common shocks. Systematic default spikes during recessions, is correlated with macroeconomic indicators, and predicts future realized defaults. More importantly, it predicts future equity and corporate bond index returns both in- and out-of-sample. Finally, we find that the cross-section of average stock returns is related to firm-level exposures to systematic default risk.