Qixiang Gao - Research

Working Paper

Greening the Production (Job Market Paper)

Abstract: Using global firm level environmental expenditure data from 2002 to 2024, I study how firms accumulate clean capital and increase its substitutability with dirty capital. The clean share of physical capital rises from below 1% to nearly 10%. Under a balanced productivity CES benchmark, the two capital stocks are gross substitutes, with substitutability varying across industries and rising over time. Allowing for unbalanced productivity, clean investment and the installed clean capital share predict jumps in relative clean marginal productivity. I develop and estimate a dynamic model that jointly endogenizes clean productivity growth and the elasticity of substitution. The two channels reinforce one another, implying that the green transition reshapes not only capital allocation but also the production technology itself. Policies like carbon pricing and clean financing become substantially more effective when the elasticity of substitution is endogenous.

Asset Pricing with a Latent Network with Qiyang Yu

Abstract: We develop a network augmented characteristic pricing framework in which expected returns reflect both firm specific information and signals transmitted through cross firm network exposures. The estimator imposes a parsimonious latent group structure that do not reply on any pre-specified network assignment. The resulting latent peer index, LPI, admits an economically transparent decomposition into own characteristic and network components. Applied to U.S. equities, LPI contains cross sectional return information beyond both the conventional industry peer index and firms’ own characteristics, with the network component carrying a distinct price of risk. LPI sorted portfolios earn positive returns and factor adjusted alphas across weighting schemes, subsamples, and holding horizons. The associated factor remains unspanned by established factor models and continues to add pricing information after conditioning on the industry peer index and the market. A trading strategy based on LPI also outperforms the benchmark signals and the market. These findings show that network transmission provides an economically important and empirically distinct dimension of characteristic based asset pricing.

Herding as a Price Multiplier

Abstract: I study the price impact of investors' flow of fund and mutual funds' herding behavior in a double-layered demand system where money flows from household to institutions and eventually to capital markets. The model features a flow-performance-herding feedback loop that imposes both a fundamental and a behavioral multiplier effect upon steady state asset price, thus predicts that even a transitory shock to asset fundamentals can have a permanent influence on asset price through the behavioral channel. Empirically, the above price dynamics framework provides explanation to the post earnings announcement drift anomaly and illustrates why post-crisis government intervention policies often over-tune the market.

Work in Progress

A Spectral Solution to the Multi-horizon Pricing Problem with Hui-Jhong Choi [Draft upon Request]

Abstract: Based on previous empirical indications, this paper formally shows that transitory and persistent shocks to firm specific characteristics across frequencies carry different prices of risk. Although there exists a pricing kernel spanned by the long-short portfolio of observable characteristics aggregated across frequencies that correctly prices the one period return, it is no longer a mean variance efficient return when rolled over to multiple horizons, in the sense that it generates a non zero pricing error proportional to the pricing horizon. Using the frequency principal component constructed pricing kernel alleviates this problem.

The Term Structure of Corporate Required Rates with Hui-Jhong Choi [Draft upon Request]

Abstract: This paper proposes a systematic approach, combining recent advances in characteristics based asset pricing with dimension reduction techniques, to estimate the term structure of required equity returns for U.S. stocks. The estimated premium sensitivities of firm characteristics change substantially with the investment horizon, indicating that the composition of priced risk shifts as the holding period lengthens. In an application to corporate investment, we find that longer horizon required rates are more strongly associated with firm level investment than short horizon premia, and that firms with longer asset maturity are especially sensitive to these long term required rates, consistent with the idea that their investments face longer cash flow duration.

Volatility Risk Sharing under Uncertainty [Draft upon Request]

Abstract: Volatility risk is priced whenever a volatility-of-volatility sctucture exists, which can be isolated by the difference between implied and realized volatility, i.e., the volatility risk premium. At the equilibrium of futures market, the volatility risk premium is determined jointly by uncertainty in the economy, risk attitudes of hedgers and speculators, as well as the accuracy of their a priori belief of future spot price distribution. Empirically, I find out that unforeseeable fluctuations in macroeconomy and financial markets cause volatility risk in the WTI crude oil futures market to be underpriced. The market participants’ risk attitudes are strongly correlated with uncertainty and their cross effects exacerbate hedging constraints. The cost of hedging against futures price variance will increase and risk sharing will decrease if uncertainty in the macroeconomic environment is high. Furthermore, the predictive power of some futures market pricing factors like open interest momentum and hedging demand on volatility risk premium could be enhanced when taking uncertainty into consideration.

Page updated

Google Sites

Report abuse