Research

Accepted and Published Papers

Collateral Competition: Evidence from Central Counterparties (with Magda Grothe and Stathis Tompaidis) ESRB Working Paper No 131

Journal of Financial Economics, Volume 149, Issue 3, September 2023

We analyze competition and risk management at central counterparties (CCPs) using a granular transaction-level dataset, and find that CCPs decrease collateral in response to lower collateral at their competitors, an effect that becomes stronger as the correlation between positions increases. To interpret our findings, we derive a model in which collateral is driven by risk and CCP competition. Our results are consistent with the model and suggest that a single monopolistic CCP would require more collateral. We also show that amid the substantial increase in collateral during the Covid-19 pandemic, the probability of a margin breach did not significantly change.

The Effects of Capital Requirements on Good and Bad Risk-Taking (with Roberto Robatto)

Review of Financial Studies, Volume 36, Issue 2, February 2023

We study optimal capital requirement regulation in a dynamic quantitative model in which nonfinancial firms, as well as households, hold deposits. A novel general equilibrium channel that operates through firms' deposits mitigates the cost of increasing capital requirements. In the calibrated model, (i) the optimal capital requirement is 7.3 percentage points higher than in a comparable model in which all deposits are held by households, and (ii) setting the capital requirement higher than the true optimum is not as costly as one would gauge from the comparable model. We also provide some independent evidence that supports our novel channel.

Special Repo Rates and the Cross-Section of Bond Prices (with Stefania D'Amico) 

Review of Finance, Volume 26, Issue 1, February 2022

We price the risky component of specialness spreads---identified by their deviations from the expected auction cycle---within a dynamic term structure model estimated using daily prices of all outstanding Treasury securities and corresponding special collateral (SC) repo rates. This allows us to derive a time-varying SC risk premium that we quantitatively link to various price anomalies, such as the on-the-run premium. The SC risk premium explains about 80% of the on-the-run premium and a substantial share of other Treasury price anomalies, suggesting that unexpected fluctuations in the specialness spreads of recently-issued nominal Treasury securities are a common risk factor.

Zero-Coupon Yields and the Cross-Section of Bond Prices 

Review of Asset Pricing Studies, Volume 11, Issue 2, June 2021

I estimate a dynamic term-structure model on an unbalanced panel of Treasury coupon bonds, without relying on an interpolated zero-coupon yield curve. A linearity-generating (LG) model, which separates the parameters that govern the cross-sectional and time-series moments of the model, takes about eight minutes to estimate on a sample of over 1 million bond prices. The traditional exponential affine model takes about two hours, due to a convexity term in coupon-bond prices that cannot be concentrated out of the cross-sectional likelihood. I quantify the on-the-run premium and a ``notes vs. bonds'' premium from 1990--2017 in a single, easy-to-estimate no-arbitrage model.

Working Papers

Measuring Measurement Error (with Garrett Schaller)

Revision requested by the Journal of Financial Economics

We highlight the importance of measurement error in applied empirical work using a sample of 2,185 instrumental variables regressions from 326 papers published in top economics and finance journals. If published instruments are valid for measurement error, then our estimates imply that only 20–40% of the variance of the average regressor is attributable to the underlying variable of interest. Publication bias does not quantitatively explain our results, although we cannot rule out the influence of heterogeneous treatment effects or instrument invalidity. Our estimator can also bolster identification arguments when IV estimates are unexpectedly large. 

The Benchmark Greenium (with Stefania D'Amico and Johannes Klausmann)

Exploiting the unique "twin" structure of German government green and conventional securities, we use a dynamic term structure model to estimate a sovereign risk-free greenium, distinct from the yield spread between the green security and its conventional twin (i.e., the green spread). The model purifies the green spread from pecuniary factors unrelated to environmental preferences. While the model-implied greenium exhibits a significant relation with proxies of shocks to climate concerns---and the green spread does not---the green spread correlates with stock market prices and measures of flight-to-quality. We also estimate the full greenium term structure and expected green returns.

Bailing out (Firms’) Uninsured Deposits: A Quantitative Analysis (with Roberto Robatto)

We analyze the effects of (not) bailing out uninsured deposits in a quantitative, general equilibrium model in which firms’ deposits are valued for their safety and uninsured deposits might be bailed out by the government. We calibrate our model around two novel facts: first, firms and other non-household depositors hold more uninsured deposits than households, and second, uninsured depositors have been bailed out in 94% of bank failures. Our results suggest negligible real effects of not bailing out uninsured deposits in 2023: although riskier deposits reduce firms’ ability to engage in production, they also induce a substitution away from deposits and towards investment in productive capital, increasing employment. We also find roughly zero real effects of moving to a 100% deposit insurance regime.

Testing for Instrument Validity with Higher-Order Cumulants (with Garrett Schaller)

Instrumental variables estimators are commonly used in economics and finance to establish causal relationships. Although instruments that fail the exclusion restriction do not reliably estimate parameters of interest, testing the exclusion restriction is uncommon, due to the difficulty of finding multiple valid instruments. We derive closed-form instrumental variable estimators that allow for tests of over-identifying restrictions even for the case of a single valid instrument. We also derive estimators that are consistent when instruments and regressors are mis-measured with correlated errors. Monte Carlo simulations suggest that our estimators have power to reject even in relatively small samples. We also apply our estimators to the IV regressions of Mian & Sufi (2014) and cannot reject the null hypothesis that the exclusion restriction holds.

The Impact of Lending Shocks Across the Firm Size and Age Distribution (with Nuno Paixao)

Financial frictions are often mentioned as an important factor affecting firms' growth, especially for small firms, yet few studies include the full size distribution of firms. Taking advantage of the universe of firms in the LBD, we analyze the causal impact of lending shocks on firms of different size or age. We find that in response to local aggregate lending shocks, on average, very small firms (1--4 employees) actually shrink, medium-sized firms grow faster, while very large firms (at least 500 employees) are unaffected. Firm size, rather than firm age, best describes the cross-sectional effects of credit shocks, although we do find that older (small and medium-sized) firms are less likely to exit following an inflow of credit, while young firms appear to be unaffected.  

Do Financial Factors Drive Aggregate Productivity? Evidence from Indian Manufacturing Establishments (online appendix)

Work in Progress

Financial frictions in the U.S., Europe and India (with Chen Yeh)