Working Papers

Revise and Resubmit

Active

13.) "The Fed Information Effect and Firm Investment." With Indrajit Mitra, Yu Xu, and Linghang Zeng, 2025.

We provide firm-level evidence that the Federal Reserve’s economic outlook expressed during FOMC announcements has real effects. Using a high frequency measure of Fed information, we show that more cyclical firms see a greater upward (downward) revision in equity analyst forecasts of sales and earnings per share following a positive (negative) Fed information shock. We construct a heterogeneous firm New Keynesian model incorporating this finding to analyze Fed information’s effect on investment. Our model predicts greater sensitivity of firm profitability and investment to Fed information for more cyclical firms. We find evidence for both predictions.

14.) "Public Debt and Private Investment: The Role of Firm Leverage." With Altan Pazarbasi, and Andrea Tamoni, 2026.

This paper provides novel empirical evidence that financial heterogeneity matters in firms' investment responses to rising public debt. After government debt increases, low-leverage firms cut investment more and have weaker cash flows than high-leverage firms. We rationalize empirical findings using a heterogeneous-firm model with financial frictions. In the model, low-leverage firms are more responsive to rising public debt because they rely more on cash flows for investment and experience a larger shift in the marginal benefit of investment. Our results highlight a cash flow channel of fiscal policy distinct from the standard discount rate mechanism.

15.) "Fatal Attraction? An Early Investigation of the AI Bubble." With Xindi He, Andrea Tamoni, and Yiyang Zhang, 2026.

The equity market has rewarded AI-relevant firms with outsized growth and elevated valuations. However, if AI fails to deliver the expected earnings growth and cost savings, today's financial and physical investments may be misallocated. To assess whether the equity market is currently experiencing an AI bubble, we construct a firm-level AI Rank Score using asset embeddings. Asset pricing tests show that firms with high AI Rank Scores have higher expected returns, higher valuation ratios, and are more profitable. Institutional investors, hedge funds in particular, hold portfolios with high delta exposure and position tilts toward the AI stocks. While we cannot rule out the AI-bubble hypothesis, the market is presently not as detached from fundamentals as a bubble would imply.

16.) "Asset Pricing with Hand-to-Mouth Households." 2017.

I develop a heterogeneous agent model in a general equilibrium production economy to analyze the asset pricing implications when the marginal pricer can potentially lose the ability to save and invest. Building on the saver-spender dichotomy of Mankiw (2000), an optimizing household (saver) is faced with the possibility of being switched to a hand-to-mouth household (spender), and vice versa, with time-varying probabilities. The baseline model is capable of generating large risk premia for both stocks and Treasury bonds while matching a set of macroeconomic moments. The potential to encounter a household-level rare-disaster-like switching shock results in a pricing kernel, purely driven by the saver's consumption process, that is the weighted average of the marginal utility of staying as a saver and the marginal utility of getting switched to a spender. The switching mechanism amplifies the covariances between the real pricing kernel and the return on dividend claims of firms as well as inflation. The former generates large equity risk premium while the latter produces large inflation risk premium on nominal bonds.

Google Sites

Report abuse