Session 18: New Research in Asset Pricing

We examine a perplexing phenomenon wherein technological innovations induce short-term contractions, using a two-sector New-Keynesian model. Pivotal to explaining the evidence is the joint effect of sticky prices, in conjunction with the separation of production into investment and consumption sectors. This setup alters the cyclicality of capital good prices, which modulates production during innovative phases. The model addresses key asset-pricing puzzles: Why is there a negative link between investment returns and stock returns? Why do valuations surge after adverse labor-market events? Why do both high book-to-market and high gross profits forecast future returns positively, despite their divergent ties to technology? Why is the slope of the equity yield term structure procyclical? The mechanism of innovation-led contractions serves as a unifying thread, weaving together previously isolated questions while offering a novel perspective.

We collect a novel dataset encompassing the vast majority of the portfolio holders of the U.S. Treasury market and estimate granular demand functions for the Federal Reserve and preferred habitat investors such as commercial banks, mutual funds, insurance companies, and foreign investors. We embed these estimated demand functions in an equilibrium model of the U.S. Treasury market where risk-averse strategic arbitrageurs interact with preferred habitat investors and the Federal Reserve. Our tractable model accommodates cross-elasticities across different maturities, macroeconomic dynamics, and both conventional and unconventional monetary policies. We quantify arbitrageurs’ risk aversion using data on dealers’ and hedge funds’ Treasury positions. We find (1) that the Treasury market is elastic because of low estimated arbitrageur risk aversion that significantly weakens demand impact; (2) a positive term premium response to monetary policy tightening due to high estimated cross-elasticities, rationalizing excess sensitivity of long rates; (3) a weak effect of Fed purchases on bond yields unless the Fed credibly commits to a persistent expansion of its balance sheet.

Can environmentally-minded investors impact the cost of capital of green firms even when they invest through financial intermediaries? To answer this and related questions, I build an equilibrium intermediary asset pricing model with three investors, two risky assets, and a riskless bond. Specifically, two heterogeneous retail investors invest via a financial intermediary who decides on the portfolio allocation that she offers between a green and a brown equity. Both retail investors and the financial intermediary can tilt towards the green asset, beyond pure financial considerations. Perhaps surprisingly, the green retail investor can have significant impact on the pricing of green assets, even when she invests via an intermediary who does not tilt: a sizable green premium –that is, a lower cost of capital– can emerge on the equity of the green firm. This good news comes with important qualifications, however: the green retail investor has to take large leveraged positions in the portfolio offered by the intermediary, her strategy must be inherently state-dependent, and economic conditions or the specification of preferences can overturn or limit the result. When the financial intermediary decides (or is made) to tilt instead, the impact on the green premium is substantially larger, although it is largest when preference are aligned with retail investors. I also study what happens when the green retail investor does not know the weights in the portfolio offered by the intermediary, the potential impact of greenwashing, and the effect of portfolio constraints. Taken together, these findings highlight the central role that financial intermediaries can play in channeling financing (or not) towards the green transition.

A common concern is that ambitious climate policy is—at least in parts—obstructed by corporate lobbying activities. We quantify corporate anti- and pro-climate lobbying expenses, identify the largest corporate lobbyists and their motives, establish how climate lobbying relates to corporate business models, and document whether and how climate lobbying is priced in financial markets. Firms spend on average $277k per year on anti-climate lobbying ($185k on pro-climate lobbying). Recently, firms have tried to camouflage their climate lobbying activities. Large anti-climate lobbyists have more carbon-intensive business models and face more climate-related incidents in the future. Firms that spend more on anti-climate lobbying earn higher returns, probably because of a risk premium. Their stock prices went up when the Waxman-Markey Cap-and-Trade Bill failed, and down when the Inflation Reduction Act was announced.

How do strategic interactions between financial institutions influence adaptation behavior to emergent risks where regulatory oversight is limited? We join detailed bank supervisory data with high resolution climate data to identify adaptation behavior of banks to climate change. We exploit heterogeneous in bank learning about climate shocks following Hurricane Harvey to examine the influence of the adaptation choices of competitors. First, banks that learn about climate risks reduce lending to riskier markets while uninformed banks increase market share. Second, despite learning, informed banks are less likely to adapt when facing greater competitive pressures. Finally, banks are less likely to adapt in markets where competitors are also less likely to do so, suggesting strategic complementarities in adaptation behavior. More broadly, our paper sheds light on the role of competitive forces in how banks manage emergent risks.

We analyze the intergenerational consequences of rare disasters in a calibrated overlapping generations model featuring realistic household portfolios and equilibrium asset prices. House-holds own houses and additionally trade in bonds and equity. In a disaster, young households suffer from reduced labor income and tightened borrowing constraints. Older households lose a large portion of their savings invested in risky assets. The relative winners are households shortly before retirement, who have a comparatively stable labor income, are not borrowing constrained, and are young enough to benefit from large returns of assets purchased during the disaster at depressed prices. In order to solve the model, we advance contemporary deep learning based solution methods along two complementary dimensions. First, we introduce an economics-inspired neural network architecture that, by construction, ensures that market clearing conditions are always satisfied. Second, we illustrate how to solve models with multiple assets by introducing them step-wise into the economy. These two innovations enable us to reduce the number of equilibrium conditions, that are not fulfilled exactly, and to substantially improve the stability of the training algorithm.

Cross-sectional decompositions using professional forecasts show high price-earnings ratios are accounted for by both low expected returns and overly high expected earnings growth. The magnitudes and timing of the comovements between prices, earnings growth, and returns are consistent with gradual learning rather than expectations being highly sensitive to recent realizations. Earnings growth surprises do not translate 1-1 into one-period returns, but instead are gradually reflected in returns over time. A structural model incorporating constant-gain learning about mean earnings growth, coupled with risk premia linked to cash flow timing, replicates our findings and generates realistic dispersion and persistence in price-earnings ratios.

We use option prices to derive bounds on the probability of a crash in an individual stock, and argue that the lower bound should be close to the truth. Empirically, the lower bound is highly statistically and economically significant; on its own, it outperforms 15 stock characteristics proposed by the prior literature combined. In a multivariate regression, a one standard deviation increase in the bound raises the predicted crash probability by 3 percentage points, whereas a one standard deviation increase in the next most important predictor (a measure of short interest) raises the predicted probability by only 0.3 percentage points.

Large economic, ecological, natural, and health-related disasters have the potential to set off a sequence of secondary calamities, initiating cascading effects that impose substantial additional economic costs. For instance, the 2008 world economic crisis triggered a chain of financial catastrophes that rippled across global markets, highlighting the severe impact of contagion effects. This paper delves into the repercussions of such contagion effects for optimal public policy. We compare the optimality of precautionary measures taken ahead of time with a ”reactive” approach to disaster management, i.e., disaster-mitigation efforts adopted after the gravity of the first shock has been established. The paper develops a novel dynamic stochastic framework, where disaster arrivals are modeled via the Hawkes process, which allows for statistical modeling of such events through its self-excitation mechanism, in contrast to processes with independent increments, like Wiener or Poisson processes. This modeling choice captures the clustering of events and the subsequent triggering of additional disasters, mirroring real-world scenarios such as the aftermath of the 2008 crisis or the cascading impacts seen during the COVID-19 pandemic. We derive analytical solutions and show that the optimal policy consists of devoting a stochastic fraction of output to disaster-mitigation. The mitigation propensity is an increasing function of the Hawkes intensity and essentially tracks disaster arrivals. This implies that the policy is indeed reactive. This result contrasts with the existing literature, which does not account for the possibility of contagion and therefore finds a constant mitigation propensity to be optimal. By incorporating the Hawkes process, we provide a more accurate representation of disaster dynamics and offer improved policy recommendations for managing the complex interdependencies of catastrophic events.

I derive a novel production-based asset pricing result that relates exchange rates to the intertemporal marginal rates of transformation of capital (i.e., returns on investment) within multinational firms with headquarters at home and foreign affiliates abroad. This q-theory of multinationals is analogous to a standard international consumption-based model, but uses multinational producers and production functions rather than consumers and stochastic discount factors. Its structural estimation infers exchange rates from US multinational aggregates on real investment, output, and capital stocks. Tested across 44 countries since the 1980s, it successfully prices cross-sectional carry trade spreads through the fundamentals of US foreign affiliates.

I develop a real options model to explain average returns and return volatilities of stock portfolios sorted on the book-to-market ratio. While average returns increase monotonically across portfolios, return volatilities are U-shaped. My model combines business cycle variations with countercyclical economic uncertainty. Operating leverage and procyclical growth options make both value stocks and growth stocks risky, generating U-shaped return volatilities. Growth stocks additionally load on the negative variance risk premium which reduces their expected return. Using structural estimation, my model jointly fits average returns and return volatilities, thereby solving a long-standing problem in investment-based asset pricing. Further reduced-form evidence supports the model channels.

This paper proposes latent factor models for multidimensional panels called 3D-PCA. Factor weights are constructed from a small set of dimension-specific building blocks, which give rise to proportionality restrictions of factor weights. While the set of feasible factors is restricted, factors with long/short structures often found in pricing factors are admissible. I estimate the model using a 3-dimensional data set of double-sorted portfolios of 11 characteristics. Factors estimated by 3D- PCA have higher Sharpe ratios and smaller cross-sectional pricing errors than models with PCA or Fama-French factors. Since factor weights are subject to restrictions, the number of free parameters is small. Consequently, the model produces robust results in short time series and performs well in recursive out-of-sample estimations.

We leverage state-of-the-art large language models (LLMs) such as ChatGPT and LLaMA to extract contextualized representations of news text for predicting stock returns. Our results show that prices respond slowly to news reports indicative of market inefficiencies and limits- to-arbitrage. Predictions from LLM embeddings significantly improve over leading technical signals (such as past returns) or simpler NLP methods by understanding news text in light of the broader article context. For example, the benefits of LLM-based predictions are especially pronounced in articles where negation or complex narratives are more prominent. We present comprehensive evidence of the predictive power of news on market movements in 16 global equity markets and news articles in 13 languages.

The fight against climate change exposes businesses to major transition risks. To successfully navigate this challenge, businesses need to fundamentally rethink their carbon footprint, energy supplies and, in many cases, entire production chain. In a sign that they are taking this challenge seriously, many publicly traded companies in the U.S. have announced ambitious commitments to reduce future carbon emissions. In this paper we identify green pledges—decarbonization commitments—from newswires articles using AI (GPT-4). About 9% of U.S. firms have made green pledges, and these companies tend to be larger and browner than those without pledges. The announcement of a green pledge significantly raises the company’s stock price, and this stock market impact is stronger for larger and browner firms. Firms that make green pledges subsequently reduce their CO2 emissions by more than other firms, a result that alleviates concerns about “greenwashing”. Our evidence suggests that green pledges have material new information for investors, can reduce perceived climate transition risk, and may be the result of financial incentives for companies to reduce their carbon footprint.