Yi Ding
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yi Ding has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Xinghua Zheng
Department of ISOM Hong Kong University of Science and Technology Clear Water Bay Kowloon Hong Kong
4 shared publications - Viktor Todorov
Northwestern University
3 shared publications Torben G. Andersen
2 shared publicationsYingying Li
2 shared publicationsGuoli Liu
Hong Kong University of Science & Technology (HKUST) - Department of Information Systems, Business Statistics and Operations Management
1 shared publication- Seunghyeon Yu
College of Business Korea Advanced Institute of Science and Technology (KAIST) Seoul Korea
1 shared publication
Latest Publications
- Open Access
- 26 June 2026
- Published by Elsevier in SSRN Electronic Journal
A Fine Lens on Common Trading Flows
We study systematic trading activities at ultra-high frequencies and their implications for price formation. A continuous-time, high-dimensional factor model is proposed for stock trading volume. This model features a Common Trading Intensity (CTI) factor that captures common trading flow. We develop estimation and inference theories for studying the fine dynamics of CTI. Using trading volume data from U.S. equities, we find strong negative temporal dependence in short-horizon CTI movements, pointing to the presence of intraday systematic liquidity frictions. Moreover, large negative CTI shocks are followed by brief intraday price adjustments that unwind rapidly, suggesting efficient intraday systematic liquidity rebalancing.
- 26 June 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessA Fine Lens on Common Trading Flows
We study systematic trading activities at ultra-high frequencies and their implications for price formation. A continuous-time, high-dimensional factor model is proposed for stock trading volume. This model features a Common Trading Intensity (CTI) factor that captures common trading flow. We develop estimation and inference theories for studying the fine dynamics of CTI. Using trading volume data from U.S. equities, we find strong negative temporal dependence in short-horizon CTI movements, pointing to the presence of intraday systematic liquidity frictions. Moreover, large negative CTI shocks are followed by brief intraday price adjustments that unwind rapidly, suggesting efficient intraday systematic liquidity rebalancing.
- Open Access
- 3 April 2026
- Published by Elsevier in SSRN Electronic Journal
Predictive Factor Model for Jump Intensities
We propose a parsimonious factor model for jump intensities for a large number of stocks, which we call the Jump Factor Intensity Thinning (J-FIT) Model. The model utilizes a jump factor process to capture the complete grid of potential jumps, and a thinning procedure to capture individual stock jumps. The jump factor process features temporal dependence, which leads to predictability. We develop estimators and inference theories for the model parameters, enabling applications such as jump intensity prediction and constructing prediction intervals for the jump risk exposures of individual stocks and sectors. Empirically, we provide solid evidence that jumps are predictable with the J-FIT model, whose predictive performance substantially sur- passes various benchmark models. We also demonstrate that J-FIT is useful for tail risk management.
- 3 April 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessPredictive Factor Model for Jump Intensities
We propose a parsimonious factor model for jump intensities for a large number of stocks, which we call the Jump Factor Intensity Thinning (J-FIT) Model. The model utilizes a jump factor process to capture the complete grid of potential jumps, and a thinning procedure to capture individual stock jumps. The jump factor process features temporal dependence, which leads to predictability. We develop estimators and inference theories for the model parameters, enabling applications such as jump intensity prediction and constructing prediction intervals for the jump risk exposures of individual stocks and sectors. Empirically, we provide solid evidence that jumps are predictable with the J-FIT model, whose predictive performance substantially sur- passes various benchmark models. We also demonstrate that J-FIT is useful for tail risk management.
- Open Access
- 10 March 2026
- Published by Elsevier in SSRN Electronic Journal
The Granular Origin of Tail Dispersion Risk
We study the asset pricing implications of shifts in the investment opportunities set characterized by the tail shape of the cross-sectional return distribution. The associated risk depends on its origin: whether the returns are driven by a systematic event or not. In the former case, it stems from the assets' exposure to systematic tail risk. Alternatively, it arises from idiosyncratic tail risk in the individual asset returns. We show empirically that shocks to the cross-sectional tail shape are a source of priced risk: fat idiosyncratic tails are favored by investors, while fat-tailed exposures to systematic jumps are disliked.
- 10 March 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessThe Granular Origin of Tail Dispersion Risk
We study the asset pricing implications of shifts in the investment opportunities set characterized by the tail shape of the cross-sectional return distribution. The associated risk depends on its origin: whether the returns are driven by a systematic event or not. In the former case, it stems from the assets' exposure to systematic tail risk. Alternatively, it arises from idiosyncratic tail risk in the individual asset returns. We show empirically that shocks to the cross-sectional tail shape are a source of priced risk: fat idiosyncratic tails are favored by investors, while fat-tailed exposures to systematic jumps are disliked.
- 19 February 2026
- Published by Elsevier in Journal of Econometrics
The factor structure of jump risk
- Open Access
- 18 February 2026
- Published by Elsevier in SSRN Electronic Journal
Rate-optimal Estimation of the Scatter Matrix in Ultra-High Dimensional Elliptical Factor Models with $2+\varepsilon$th Moment
We study the estimation of scatter matrices in elliptical factor models with 2+εth moment. We develop an idiosyncratic-projected self-normalization method to remove the effect of the heavy-tailed scalar component, and propose a robust estimator of the scatter matrix that achieves the same convergence rate as in the sub-Gaussian case under an ultra-high dimensional setting. The convergence rate is optimal and leads to superior performance in estimating high-dimensional global minimum variance portfolios.
- 18 February 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessRate-optimal Estimation of the Scatter Matrix in Ultra-High Dimensional Elliptical Factor Models with $2+\varepsilon$th Moment
We study the estimation of scatter matrices in elliptical factor models with 2+εth moment. We develop an idiosyncratic-projected self-normalization method to remove the effect of the heavy-tailed scalar component, and propose a robust estimator of the scatter matrix that achieves the same convergence rate as in the sub-Gaussian case under an ultra-high dimensional setting. The convergence rate is optimal and leads to superior performance in estimating high-dimensional global minimum variance portfolios.
- Open Access
- 17 February 2026
- Published by Elsevier in SSRN Electronic Journal
Tails of Cross-Sectional Return Distributions at High Frequencies1
We develop nonparametric estimates for tail risk in the cross-section of asset prices at high frequencies. We show that the tail behavior of the crosssectional return distribution depends on whether the time interval contains a systematic jump event. If so, the cross-sectional return tail is governed by the assets’ exposures to the systematic event while, otherwise, it is determined by the idiosyncratic jump tails of the stocks. We develop an estimator for the tail shape of the cross-sectional return distribution that display distinct properties with and without systematic jumps. Empirically, we provide evidence for symmetric cross-sectional return tails at high-frequency that exhibit nontrivial and persistent time series variation. A hypothesis of equal cross-sectional return tail shapes during periods with and without systematic jump events is strongly rejected by the data.
- 17 February 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessTails of Cross-Sectional Return Distributions at High Frequencies1
We develop nonparametric estimates for tail risk in the cross-section of asset prices at high frequencies. We show that the tail behavior of the crosssectional return distribution depends on whether the time interval contains a systematic jump event. If so, the cross-sectional return tail is governed by the assets’ exposures to the systematic event while, otherwise, it is determined by the idiosyncratic jump tails of the stocks. We develop an estimator for the tail shape of the cross-sectional return distribution that display distinct properties with and without systematic jumps. Empirically, we provide evidence for symmetric cross-sectional return tails at high-frequency that exhibit nontrivial and persistent time series variation. A hypothesis of equal cross-sectional return tail shapes during periods with and without systematic jump events is strongly rejected by the data.
- Open Access
- 1 January 2024
- Published by Elsevier in SSRN Electronic Journal
Sub-Gaussian High-Dimensional Covariance Matrix Estimation under Elliptical Factor Model with 2 εth Moment
We study the estimation of high-dimensional scatter matrices under elliptical factor models with 2 + εth moment. For such heavy-tailed data, robust estimators like the Huber-type estimator in Fan et al. (2018) can not achieve a sub-Gaussian convergence rate. In this paper, we develop an idiosyncratic-projected self-normalization method to remove the effect of the heavy-tailed scalar component and propose a robust estimator of the scatter matrix that achieves the sub-Gaussian rate. The estimator demonstrates superior performance in estimating high-dimensional global minimum variance portfolios.
- 1 January 2024
- Published by Elsevier in SSRN Electronic Journal
Open AccessSub-Gaussian High-Dimensional Covariance Matrix Estimation under Elliptical Factor Model with 2 εth Moment
We study the estimation of high-dimensional scatter matrices under elliptical factor models with 2 + εth moment. For such heavy-tailed data, robust estimators like the Huber-type estimator in Fan et al. (2018) can not achieve a sub-Gaussian convergence rate. In this paper, we develop an idiosyncratic-projected self-normalization method to remove the effect of the heavy-tailed scalar component and propose a robust estimator of the scatter matrix that achieves the sub-Gaussian rate. The estimator demonstrates superior performance in estimating high-dimensional global minimum variance portfolios.
- Open Access
- 10 February 2022
- Published by Elsevier in SSRN Electronic Journal
Stock Co-Jump Networks
We propose a network model with communities to study the stock co-jump dependency. To estimate the community structure, we extend the SCORE algorithm in Jin (2015) and develop a Spectral Clustering On Ratios-of-Eigenvectors for networks with Dependent Multivariate Poisson edges (SCORE-DMP) algorithm. We prove that SCORE-DMP enjoys strong consistency in community detection for the proposed co-jump network with dependent edges. Empirically, using high-frequency data of S&P 500 constituents, we identity two co-jump networks according to whether the market jumps or not and show that the identified community structure helps in predicting stock returns.
- 10 February 2022
- Published by Elsevier in SSRN Electronic Journal
Open AccessStock Co-Jump Networks
We propose a network model with communities to study the stock co-jump dependency. To estimate the community structure, we extend the SCORE algorithm in Jin (2015) and develop a Spectral Clustering On Ratios-of-Eigenvectors for networks with Dependent Multivariate Poisson edges (SCORE-DMP) algorithm. We prove that SCORE-DMP enjoys strong consistency in community detection for the proposed co-jump network with dependent edges. Empirically, using high-frequency data of S&P 500 constituents, we identity two co-jump networks according to whether the market jumps or not and show that the identified community structure helps in predicting stock returns.