Embedded Files
Research Papers
"Statistical Identification and Estimation of Non-Gaussian Structural Vector Autoregressions with Nonstationary Regressors" (August 2025).
Abstract:
This paper propose an identification and estimation scheme for non-Gaussian SVAR models in the presence of nonstationary regressor such as near unit roots and possible cointegration. When structural shocks are non-Gaussian and are assumed serially independent then observationally equivalent processes can be obtained regardless of the nonstationary dynamics. We develop an estimation method and establish the asymptotic properties of the proposed estimator and test statistic.
Keywords: non-Gaussian SVAR; nonstationary regressors
Photo Credit: © Christis Katsouris (2011)
"Estimation and Inference in Systems of Predictive Quantile Regressions with Persistent Predictors" (July 2025).
Abstract:
This paper develops estimation and inference procedures for systems of predictive quantile regressions with nonstationary predictors and generated regressor to capture tail network-based interdependencies. We study the asymptotic properties of the proposed estimation and inference procedures and show that system-based and pooled-based estimators are robust to equation-specific nuisance parameters of persistence using the IVX instrumentation. The efficacy of our method is studied with simulated and real-life data under the presence of network dependence.
Keywords: Predictive Quantile Regression, Local-to-Unity; SUR
Photo Credit: © Christis Katsouris (2011)
"Estimating Conditional Value-at-Risk with Nonstationary Quantile Predictive Regression Models" (July 2024).
Abstract:
This paper develops an asymptotic distribution theory for an endogenous instrumentation approach in quantile predictive regressions when both generated covariates and persistent predictors are used. The generated covariates are obtained from an auxiliary quantile predictive regression model and the statistical problem of interest is the robust estimation and inference of the parameters that correspond to the primary quantile predictive regression in which this generated covariate is added to the set of nonstationary regressors. The proposed doubly IVX corrected estimator is robust to the unknown persistence regardless of the presence of generated regressor obtained from the first stage procedure. The asymptotic properties of the two-stage IVX estimator (mixed Gaussianity) are established while the asymptotic covariance matrix is adjusted to account for the first-step estimation error.
Link: Preprint arXiv:2311.08218.
Keywords: CoVaR; Predictive Quantile Regression, Local-to-Unity; Subvector Inference
Photo Credit: © Christis Katsouris (2012)
"Predictability Tests Robust against Parameter Instability" (July 2023).
Abstract:
This paper proposes a set of Wald type statistics for structural break testing in predictive regression models with regressors generated as local unit root processes. We focus on an environment with a single structural break at an unknown location when predictors exhibit mildly integratedness or high persistence. We derive the limiting distributions of the sup OLS-Wald and sup IVX-Wald statistic following the instrumental variable methodology proposed by Phillips & Magdalinos (2009). The asymptotic theory of the paper is developed from first principles, following limit results of seminal studies in the literature. The finite-sample size performance of the Wald statistics based on these estimators is assessed via Monte Carlo experiments with critical values computed using bootstrap techniques.
Links: [Manuscript: Preprint arXiv] [Appendix: Preprint arXiv]
Keywords: Structural Break Testing; Predictive Regression; Local-to-Unity
Photo Credit: © Christis Katsouris (2011)
Research in Progress
"Inferring Predictive Accuracy in Nested Predictive Regressions Robust against Parameter Instability" (July 2024).
Abstract:
This paper investigates predictive accuracy testing robust against parameter instability, extending the novel approach to predictive accuracy testing in nested environments of Pitarakis (2023, Econometric Theory). Our econometric framework compares forecasts based on a benchmark predictive regression against an unconstrained predictive regression with predictors generated using the local-to-unity parametrization. We propose suitable self-normalizations for our test statistics based on both HAC robust covariance matrices as well as suitable transforamations that by-pass the incidental parameter problem that appear due to data-driven dependent structures. Discussion on the recent developments can be found in Katsouris (2023, arXiv:2308.01418). Moreover, we study the limiting behaviour of the test statistics under both the null as well under the alternative. We examine the finite-sample properties of the tests via extensive Monte Carlo experiments. A local power analysis also establishes the ability of the test to detect departures from the null of equal predictive accuracy and absence of parameter instability for nearly nonstationary regressors.
Keywords: Forecast evaluation; Predictive accuracy testing; Nested model comparisons
Photo Credit: © Christis Katsouris (2012)
Academic and Research Trajectory:
September 2020 to September 2025
(a). Participation to Econ Job Market Search and Matching Procedures
Academic year 2020/2021: Doctoral Researcher at the Department of Economics, University of Southampton.
Academic year 2021/2022: Doctoral Researcher at the Department of Economics, University of Southampton. I first entered the Economics Job Market during the academic year 2021/2022; with interviews at the University of Bristol, Aarhus University and University of Exeter.
Academic year 2022/2023: Worked as a Visiting Lecturer in Economics at the University of Exeter Business School.
Academic year 2023/2024: Worked as a Postdoctoral Researcher in Econometrics at the Faculty of Social Sciences, University of Helsinki.
Academic year 2024/2025: Joined the Christis G. Katsouris Institute of Econometrics and Data Science as a Research Lab Team Leader.
(b). Research Projects, Research Grants and Research Outcomes
(see below)
Research Projects
Research Project A
Estimation and Inference for Predictive Regression Models
Research Objectives:
The main research objectives of Research Project A is the development of robust estimation and inference methodologies for predictive regression models under various econometric assumptions. A first line of research within this theme is the development of asymptotic theory for detecting parameter instability in predictive regressions under the assumption of nonstationarity when the functional form corresponds to a conditional mean function as well as a conditional quantile function. Consequently, we can investigate the presence of mean or quantile predictability based on nonstationary time series data. A second line of research within this theme, is the development of uniform inference with general autoregressive roots which is robust to the whole spectrum of persistence properties. In particular, we utilize current instrumental variable methodology from the literature (IVX) and focus on establishing the related asymptotic theory for the quantile autoregressive and quantile predictive regressions. Further computational aspects include robust estimation and bootstrap methods forpersistenty data in the presence of heteroscedasticity.
Research Outputs:
Estimating Conditional Value-at-Risk with Nonstationary Quantile Predictive Regressions
Christis Katsouris, 2023
DOI: 10.48550/arXiv.2311.08218
Unified Inference for Dynamic Quantile Predictive Regression
Christis Katsouris, 2023
DOI: 10.48550/arXiv.2309.14160
Asymptotic Theory for Unit Root Moderate Deviations in Quantile Autoregressions and Predictive Regressions
Christis Katsouris, 2023
DOI: 10.48550/arXiv.2204.02073
Estimation in Threshold Predictive Regression Models with Locally Nearly Unstable Regressors
Christis Katsouris, 2022
DOI: 10.48550/arXiv.2305.00860
Bootstrapping Nonstationary Autoregressive Processes with Predictive Regression Models
Christis Katsouris, 2022
DOI: 10.48550/arXiv.2307.14463
Research Outcomes:
During the 'Research in Training' stage I made various innovative suggestions for implementing modifications and extensions to existing econometric methods (such as in relation to the IVX estimator proposed by Phillips and Magdalinos). These suggestions which reiterate the main properties of the IVX instrumentation method for robust-to-persistence estimation and inference techniques in various settings, have resulted to novel approaches implemented by strong research teams. A related 'novel thinking', take for example the case of long-horizon predictability; a rather 'strange' way for transforming regression variables in order to test for the presence of long-horizon predictability. Take also the case of a framework for return-predictability regressions with overlapping observations. Then possible statistical efficiency gains can be realized by combining these two seemingly unrelated techniques.
The research conducted in the above manuscripts, discuss some important points regarding related open problems in the literature, which were examined in various follow-up papers by other authors. For example, I raised the question on the implementation of an IVX bootstrap-based estimator in [Preprint arXiv], as a potential extension to the original IVX method of Phillips & Magdalinos (JoE, 2007), which is then addressed in the paper of Demetrescu, Rodrigues & Taylor (JoE, 2023). Another issue of consideration that I raise in [Preprint arXiv], is the correct estimation of the density function term f(e) in quantile predictive regression models with persistent predictors, as originally implemented in the paper of Lee (JoE, 2023). This issue is further discussed in the paper of Maynard, Shimotsu & Kuriyama (JoE, 2024), who propose a fully modified bias corrected quantile-based estimator as well as by Liu, Xiao, Liu & Long (Preprint, 2024), who propose a novel simulation-based estimator for the density function of the innovations around a fixed quantile. Moreover, the issue of parameter instability in quantile predictive regressions studied in my paper [Preprint arXiv], as well as the issue of stochastic equicontinuity for persistent processes discussed in my paper [Preprint arXiv], are both further explored in the paper of Hoga (Preprint, 2024). Specifically, motivated from the literature of modelling the risk measures of VaR and CoVaR ('Tenet' specifications), this paper [Preprint arXiv] proposes a 'two stage procedure' for estimating these risk measures using quantile predictive regression models. A corresponding 2SLS testing approach, in a closely related framework, is then proposed by Demetrescu, Rodrigues & Taylor (JoE, 2024). Lastly, the issue of quantile-based estimators for uniform inference in autoregressions and predictive regression models is discussed and currently under further development in my paper [Preprint arXiv].
# Non-Stationary Time Series Models, # Robust-Persistence Methods, # Uniform Inference
© Institute of Econometrics and Data Science Christis G. Katsouris
Photo Credit: © Christis Katsouris (2013)
References:
Ex-Post Related Literature:
Demetrescu, M., Rodrigues, P.M., and Taylor, A.M. (2024). "Predictive Quantile Regressions with Persistent and Heteroskedastic Predictors: A Powerful 2SLS Testing Approach". Available at SSRN 4767275.
Maynard, A., Shimotsu, K., and Kuriyama, N. (2024). "Inference in Predictive Quantile Regressions". Journal of Econometrics, 245(1-2), 105875.
Demetrescu, M., Georgiev, I., Rodrigues, P.M., and Taylor, A.R. (2023). "Extensions to IVX Methods of Inference for Return Predictability". Journal of Econometrics, 237(2), 105271.
Demetrescu, M., Rodrigues, P.M., and Taylor, A.R. (2023). "Transformed Regression-based Long-Horizon Predictability Tests". Journal of Econometrics, 237(2), 105316.
Liu, Y., and Phillips, P.C.B. (2023). "Robust Inference with Stochastic Local Unit Root Regressors in Predictive Regressions". Journal of Econometrics, 235(2), 563-591.
Phillips, P.C.B. (2023). "Estimation and Inference with Near Unit Roots". Econometric Theory, 39(2), 221-263.
Demetrescu, M., and Rodrigues, P. M. (2022). "Residual-Augmented IVX Predictive Regression". Journal of Econometrics, 227(2), 429-460.
Related Literature:
Demetrescu, M., and Hillmann, B. (2025). "Gaussian Inference in Predictive Regressions for Stock Returns". Journal of Financial Econometrics, 23(2), nbaf004.
Yang, B., Long, W., Liu, X., and Peng, L. (2025). "A Unified Predictability Test Using Weighted Inference and Random Weighted Bootstrap". Journal of Financial Econometrics, 23(2), nbaf003.
Liu, X., Liu, Y., Long, W., and Xiao, P. (2025). "Testing Predictability of Stock Returns under Quantile Regression: A Bootstrapping Double-Weighted Approach". Econometric Reviews, 1-22.
Fei, Y., Lui, Y. L., and Yu, J. (2024). "Testing Predictability in the Presence of Persistent Errors". Working Paper 202401. Faculty of Business Administration, University of Macau.
Liao, X., Li, X., and Fan, Q. (2024). "Robust Inference for Multiple Predictive Regressions with an Application on Bond Risk Premia". Preprint arXiv:2401.01064.
Liao, X., Li, X., and Fan, Q. (2024). "Robust Bond Risk Premia Predictability Test in the Quantiles". Preprint arXiv:2410.03557.
Liu, X., Xiao, P., Liu, Y., and Long, W. (2024). "Bootstrapping the Double-Weighted Predictability Test for Predictive Quantile Regression". Available at SSRN 4740614.
Liu, B., and Pang, T. (2024). "Weighted Composite Quantile Inference for Nearly Nonstationary Autoregressive Models". Statistical Methods & Applications, 1-43.
Xu, K. L., and Guo, J. (2024). "A New Test for Multiple Predictive Regression". Journal of Financial Econometrics, 22(1), 119-156.
Hong, S., Henderson, D.J., Jiang, J., and Ni, Q. (2023). "Unifying Estimation and Inference for Linear Regression with Stationary and Integrated or Near-Integrated Variables". Journal of Financial Econometrics.
Cai, Z., Chen, H., and Liao, X. (2023). "A New Robust Inference for Predictive Quantile Regression". Journal of Econometrics, 234(1), 227-250.
Yang, B., Long, W., and Yang, Z. (2022). "Testing Predictability of Stock Returns under Possible Bubbles". Journal of Empirical Finance, 68, 246-260.
Research Project B
Break-Point Detection Methods: Estimation, Inference and Simulations
Research Objectives:
The main research objective of Research Project B is the development of econometric estimation and inference techniques suitable for when economic relations are assumed to have structural breaks at unknown locations within the full sample. Under the presence of possible multiple structural breaks in time series regressions, consistently estimating these break-point estimates using joint and sequential procedures allows for the detection of structural instabilities (such as structural breaks and threshold effects). Moreover, within the context of nearly unstable autoregressive models and predictive regression models with nonstationary regressors, test statistics can be also employed in settings where underline bubble dynamics are present (such as house price bubbles), although appropriate modifications of limiting distributions are required when constructing robust-persistent estimators. Generally, in comparison to change-point testing procedures (such as the sup-Wald statistic), change-point estimation is considered to be a more challenging task both theoretically and methodologically, regardless of whether we employ a retrospective or a sequential perspective. These issues are also of practical relevance to the case of predictive regression models with nonstationary regressors. Further computational aspects include the implementation of fast and efficient algorithmic procedures for dating the break-points (such as via the use of information criteria), the implementation of sequential monitoring schemes as well as out-of-sample forecasting schemes under the presence of parameter instability.
Research Outputs:
Break-Point Date Estimation for Nonstationary Autoregressive and Predictive Regression Models
Christis Katsouris, 2022
DOI: 10.48550/arXiv.2308.13915
Structural Break Detection in Quantile Predictive Regression Models with Persistent Covariates
Christis Katsouris, 2021(Preprint arXiv)
DOI: 10.48550/arXiv.2302.05193
Predictability Tests Robust to Parameter Instability
Christis Katsouris, 2021
DOI: 10.48550/arXiv.2307.15151
Partial Sum Processes of Residual-Based and Wald-type Break-Point Statistics in Time Series Regression Models
Christis Katsouris, 2020
DOI: 10.48550/arXiv.2202.00141
# Nonstationary Econometrics, # Structural Break Testing, # Forecast Evaluation with Breaks
© Institute of Econometrics and Data Science Christis G. Katsouris
Photo Credit: © Christis Katsouris (2010)
References:
Ex-Post Related Literature:
Schweikert, K. (2025). "Detecting Multiple Structural Breaks in Systems of Linear Regression Equations with Integrated and Stationary Regressors". Oxford Bulletin of Economics and Statistics.
Cai, Z., and Chang, S.Y. (2024). "A New Test on Asset Return Predictability with Structural Breaks". Journal of Financial Econometrics, 22(4), 1042-1074.
Andersen, T.G., and Varneskov, R.T. (2022). "Testing for Parameter Instability and Structural Change in Persistent Predictive Regressions". Journal of Econometrics, 231(2), 361-386.
Andersen, T.G., and Varneskov, R.T. (2022). "Consistent Local Spectrum Inference for Predictive Return Regressions". Econometric Theory, 38(6), 1253-1307.
Zhu, F., Liu, M., Ling, S., and Cai, Z. (2022). "Testing for Structural Change of Predictive Regression Model to Threshold Predictive Regression Model". Journal of Business & Economic Statistics, 41(1), 228-240.
Research Project C
Modelling Financial Networks: Theory, Estimation and Applications
Research Objectives:
The main research objective of Research Project C is the development of robust econometric identification, estimation and inference methods under the presence of network dependence. Overall, network connectivity has important implications in the risk management of economic and financial events which motivates the use of statistical procedures such as testing for financial contagion in stock markets, modelling the transmission of risk spillovers and market exuberance as well as monitoring the macroeconomic effects of climate change events. A modelling approach which allows to capture such interdependencies is the Vector Autoregression Model with both temporal and contemporaneous effects. Recently, several studies proposed the use of network-driven models which capture information based on the underline network dependence of time series. Complex networks can be modelled by such time series regressions (such as the NVAR model), which allow to incorporate features that capture both the underline network structure (such as via an adjacency matrix) as well as node-specific characteristics. Firstly, we focus on econometric identification and estimation methods as well as on asymptotic theory for network-driven model specifications which capture time series dynamics under general dependence structures. Secondly, we relax the network time-invariant assumption which implies that the causal structure remains unchanged over the sample period, since causal links may change during the sample period at unknown locations. Further computational aspects include linear and nonlinear optimisation methods for estimating VAR models with macroeconomic data.
Research Outputs:
Robust Estimation in Network Vector Autoregression with Nonstationary Regressors
Christis Katsouris, 2024
DOI: 10.48550/arXiv.2401.04050
Statistical Estimation for Covariance Structures with Tail Estimates using Nodewise Quantile Predictive Regression Models
Christis Katsouris, 2022
DOI: 10.48550/arXiv.2305.11282
Optimal Portfolio Choice and Stock Centrality for Tail Risk Events
Christis Katsouris, 2021
DOI: 10.48550/arXiv.2112.12031
# Large Covariance Matrices, # VAR Models, # Quantile Predictive Regression
© Institute of Econometrics and Data Science Christis G. Katsouris
Photo Credit: © Christis Katsouris (2023)
References:
Ex-Post Related Literature:
Uematsu, Y., and Yamagata, T. (2025). "Discovering the Network Granger Causality in Large Vector Autoregressive Models". Journal of the American Statistical Association, 1-23.
Chen, S., and Schienle, M. (2024). "Large Spillover Networks of Nonstationary Systems". Journal of Business & Economic Statistics, 42(2), 422-436.
Xu, X., Wang, W., Shin, Y., and Zheng, C. (2024). "Dynamic Network Quantile Regression Model". Journal of Business & Economic Statistics, 42(2), 407-421.
Ando, T., Greenwood-Nimmo, M., and Shin, Y. (2022). "Quantile Connectedness: Modelling Tail Behavior in the Topology of Financial Networks". Management Science, 68(4), 2401-2431.
Research Project D
Structural Vector Autoregressive Models: Identification and Estimation
Research Objectives:
The main research objective of Research Project D is the development of identification and estimation methods for multivariate time series when our interest is the structural analysis of macroeconomic data. While identification techniques for SVAR models can be economic theory driven which requires to impose variable restrictions (such as using sign or short and long-run restrictions) along with suitable distributional assumptions, the statistical identification of SVAR models require a valid identification scheme based on the non-Gaussianity property. Despite the large literature covering theoretical and methodological aspects for both approaches, less attention has been paid to the aspect of individually identified structural shocks versus the case of simultaneously identified structural (multiple) shocks. In this research theme we focus on developing robust identification and estimation methods for SVAR models using the non-Gaussianity property of structural shocks. Within the macroeconometrics literature, under the assumption of time series stationarity (stability) SVAR models have a Wold representation which permits identification and estimation. In addition, we focus on the identification of cointegrating vectors (such as the case of possibly cointegrated data) when SVAR models are identified by exploiting the non-Gaussianity of structural shocks. Similarly, the case where the SVAR models includes nonstationary regressors which are modelled using the local to unity parametrization, is of interest to the identification and estimation perspective we follow which exploits the information induced from deviations from the non-Gaussianity of structural shocks. Firstly, we focus on establishing that the set of observationally equivalent parameters satisfy the conditions imposed for the validity of our identification scheme under the presence of possibly nonstationary regressors. Secondly, we focus on persistent-robust estimation which ensure asymptotically valid inferences. Further identification approaches include the instrumental variable approach which is robust against the degree of the instrument's identification strength under the presence of persistence of unknown form. Further computational aspects include likelihood-based estimation and independent component analysis using optimisation algorithms.
Research in Progress:
Structural Analysis of Non-Gaussian and Non-Stationary Vector Autoregressions
Christis Katsouris, 2025
# Non-Gaussian SVAR Models, # Identification Methods, # Structural Inference
© Institute of Econometrics and Data Science Christis G. Katsouris
Photo Credit: © Christis Katsouris (2013)
References:
Related Literature:
Holberg, C., and Ditlevsen, S. (2025). "Uniform Inference for Cointegrated Vector Autoregressive Processes". Journal of Econometrics, 247, 105944.
Hoesch, L., Lee, A., and Mesters, G. (2024). "Locally Robust Inference for Non‐Gaussian SVAR Models". Quantitative Economics, 15(2), 523-570.
Katsouris, C. (2023). "Structural Analysis of Vector Autoregressive Models". Preprint arXiv:2312.06402.
Cheng, X., Han, X., and Inoue, A. (2022). "Instrumental Variable Estimation of Structural VAR Models Robust to Possible Nonstationarity". Econometric Theory, 38(5), 845-874.
Chevillon, G., Mavroeidis, S., and Zhan, Z. (2020). "Robust Inference in Structural Vector Autoregressions with Long-Run Restrictions". Econometric Theory, 36(1), 86-121.
Lanne, M., Meitz, M., and Saikkonen, P. (2017). "Identification and Estimation of Non-Gaussian SVARs". Journal of Econometrics, 196(2), 288-304.
Research Project E
Estimating Causal Effects using Machine Learning Methods and High Dimensional Data
Research Objectives:
Estimating treatment effects in economics are commonly used for program evaluation and belongs to the family of causal inference problems. However, in contrast to randomized experiments, observational studies are prone to the presence of confounding variables which are not balanced among treated and control individuals. In particular, Research Project D focuses on developing identification and estimation methods for treatment effect heterogeneity across various data structures. We study the statistical properties of treatment effect estimators for learning causal effects in settings with possibly high dimensional covariates. We focus on machine learning methods for large p and small n problems using nonparametric and semiparametric estimation approaches. Further computational aspects include shrinkage optimisation techniques and doubly robust estimation methods.
Research in Progress:
Estimating Optimal Treatment Regimes for Analyzing Treatment Heterogeneity in Quantile Regressions
Christis Katsouris, 2025
# Panel Data Models, # Heterogeneous Treatment Effects, # Doubly Robust Estimation
© Institute of Econometrics and Data Science Christis G. Katsouris
Photo Credit: © Christis Katsouris (2017)
References:
* Research Proposal prepared August 2023. Revised January 2025.
Related Presentations:
Katsouris C. (May 2024). "Presentation in Econometrics with an Application to Estimating Heterogeneous Treatment Effects at the Quantiles". Interview for the position of Lecturer in Econometrics, Department of Economics, University of Manchester, (28th of May 2024).
Katsouris C. (December 2023). "Some Aspects on Identification and Estimation of Heterogeneous Treatment Effects with Many Covariates at the Quantiles". Presentation at the Time Series and Machine Learning Study Group, Faculty of Social Sciences, University of Southampton, (8th of December 2023).
Keywords: Heterogeneous Treatment Effects, Machine Learning Methods, Program Evaluation, Panel Data Regression Models, High-Dimensional Statistics
1. Scope of Research Methods
Econometric methods and statistical theory allow to study many real-life problems found in economics, finance and macroeconomics. The econometrician addresses issues in relation to model identification, econometric estimation and inference as well as specification analysis. Moreover, while cross-sectional data are increasingly used in applied work, and empirical researchers account for certain data-driven features when estimating econometric models, the asymptotic theory for high-dimensional and data-rich environments based on machine learning techniques is of paramount importance. Recent advances allow researchers to implement empirical analyses which address questions for problems of economic relevance using state-of-the-art econometric techniques as well as making policy recommendations using big data. In general, Research Project E focuses on econometric methodologies for program evaluation as well as on treatment effect estimation and causal inference in microeconometrics.
1.1 Applied Micro-Econometric Research Questions
We consider econometric frameworks for program evaluation. In particular, we are interested in developing estimation and inference procedures for causal learning about the presence of heterogeneous treatment effects (HTEs). The main intuition of a statistical framework that is robust to heterogeneity of treatment is that it allows to assess the impact effects of interventions with respect to a target variable. Such target variables could be, for example, a measure of the cost effectiveness of an intervention scheme, the potential health improvements from a prescribed treatment or the economic outcomes of participants to an intervention program, after receiving the treatment of interest or participating to the intervention program under study (e.g., participating in a leadership development workshop). In these settings, we consider the case of selection on observable set of covariates, which includes both baseline covariates (e.g., years of education, income, area of residence) as well as control variables (e.g., health measurements) and permits to construct treatment effect estimators with machine learning methods. Moreover, estimation and inference on relevant metrics such as risk group classifications of study participants can be obtained. An excellent review of current developments in the literature related to causal inference techniques with observational data, can be found in Gunsilius (arXiv:2503.07811, 2025). Consider patients who are diagnosed with hypertropic cardiomyopathy. A false classification of a patient into a low risk group which would imply not receiving the treatment when is needed, has more serious implications with respect to health outcomes than doing the opposite type of error (see Müller et al. (JRSS B, 2024)).
From the economic perspective, labour market dynamics are directly linked to sustainable wellbeing and the mental health of the workforce, which directly impacts the productivity frontier of the economy. Social planners are interested to minimise any potential aggregate productivity losses due to mental health problems, which implies that the role of the econometrician and these techniques are to provide economically and statistically meaningful insights regarding the suitability of certain treatments or interventions. Econometric frameworks which are robust to such features (such as covariate shifts), imply that when survey data are used to obtain estimates of treatment effects for a particular sub-population, can provide insights on ways to reduce the factors that could otherwise exacerbate the impact of these conditions on individual-specific health measures and on related measures of inequalities and productivity dynamics.
2. Causal Inference for Impact/Program Evaluation or Policy Intervention
Related applications in economics include the study of the effectiveness of economic programs (program evaluation) which aim to reduce socioeconomic inequalities. Specifically, within this context the main objective is to identify whether social interventions have different effects for different people. These differences in the effects of interventions on different groups of individuals (subgroup analysis), is what we call heterogeneous treatment effects (HTEs). Although, the econometric analysis of such policy-relevant questions, which are commonly based on quasi-experimental studies, requires a different modelling approach than the statistical techniques used for data from observational studies, the underline objectives are similar. Overall, in econometrics there is now a growing literature which focuses on program evaluation using panel data regression models that goes beyond the measurement of linear impact effects. Further applications include the study of dyadic, network and panel data regressions when considering a counterfactual analysis of such data with respect to production functions.
Firstly, the study of team incentives and firm performance with respect to creation of innovation and productivity spillovers. According to Gaynor, Rebitzer, & Taylor (JPE, 2004) "such incentives can improve efficiency in the allocation of resources when the allocation process is based on the professional judgment of multiple agents". These authors study the transmission mechanism of team incentives with respect to measurable outcomes. Moreover, Mas & Moretti (AER, 2009), examine positive spillover effects of productivity which are relevant in the context of social pressure and prosocial behaviour. The authors find that when more productive workers arrive at shifts, they include a productivity increase only in workers who are in their line of vision. Another related application within this line of literature, is the measurement of team productivity with respect to the quality of research publications, using a counterfactual framework which allows to compare potential outputs based on a random allocation of team members versus potential outcomes from a self-sorting team arrangement. Specifically, Bonhomme et al. (2024) show that the estimates of average outputs are lower in the counterfactual scenario where workers are randomly allocated across teams. Within the same context of learning spillovers, cooperation and team formation the research findings of Zajkowski et. al (Nature, 2024) from the RIKEN Center for Brain Science, provide fresh perspectives on how humans build trust and interact within groups. The authors observe that allowing people to freely form and adjust connections, rather than adhering to rigid group structures, could lead to better collaboration. Importantly, team dynamics can be adjusted via alternative team formations as a mechanism for maintaining effective cooperation at the desirable level of group cohesion and social learning.
From the decision-making perspective, think for example the scenario where instead of evaluating the research output of individuals with respect to some predefined collective level of outputs from research groups in organisations, evaluating the research outputs from research groups regardless of individual performances. Would this provide incentives to team members to work towards cohesive collective research outputs, thereby increasing their cooperation and inclusion strategies towards the production of research outputs? Certainty these are interesting research questions which could be examined in appropriately designed studies in order to evaluate the potential impact on research outputs and on team members' collaboration levels without having to change a management practice (such as changing the evaluation procedure). Therefore, based on the examples discussed above, the presence of heterogeneous treatment effects serve as an identification mechanism to the effectiveness of intervention schemes (such as a re-design of research output evaluation schemes). Consequently, being able to obtain robust estimates of such treatment effects are also linked to the enhancement of socioeconomic impacts, such as the creation of positive knowledge spillover effects across research organisations and firms - contributing to firm productivity and economic growth as well as towards more tangible societal benefits which justify investments to R&D activities. For example, in management studies, several authors investigate the optimal formation of teams which ensures operational efficiency is maximised while frictions are reduced (such as the exploration-exploitation trade-off in multi-armed bandit problems, see Komiyama & Noda (Management Science, 2024)).
Secondly, firm research and innovation capabilities can facilitate the creation of knowledge spillover across firms which is often leveraged to impact aggregate productivity growth and output (under the presence of firm heterogeneity). Within this stream of literature, the measurement of distributional treatment effects of certain interventions which enhance the innovation and productivity frontier of firms can be evaluated using as an outcome (target) variable a proxy covariate for firm productivity. From the econometrics point of view, suitable hypothesis testing can be implemented using test statistics for the stochastic dominance relations between the distributions of the potential outcomes with respect to the intervention strategy. Although, statistically determining stochastic dominance relations is used for social welfare comparisons, within our setting this methodology allows to identify the economically and statistically significant pathways of a 'knowledge flow' management practice, for example, on a treated versus an untreated group with respect to a suitable firm output metric or a firm productivity metric (such as the degree of innovation). In the aggregate economy context these research questions are related to the identification of R&D spillovers due to structural change, such as during the 'Great Recession', or technological transformation such as during the 'Great Resignation'. Undoubtedly, R&D spillovers across heterogeneous firms reinforce the role of knowledge diffusion and innovation via market structures such as firms, businesses and corporations, which enables the implementation of transformational technologies that lead to aggregate productivity growth. From the social planner's perspective, these issues are relevant to the study of optimal design of corporate taxation and R&D spillovers (e.g., viewed as a dynamic mechanism design problem with spillovers; see Akcigit et al. (Ecta, 2022)).
Thirdly, program evaluation methods are used when evaluating the distributional impact of extreme weather events, such as in the case of a large-scale environmental disaster with respect to economic outcomes (e.g., impact of the weather shock on the local wage distribution). Machine learning methods allows to identify the causal effects of a cash-program intervention via an optimal treatment regime approach. In the case of extreme climate event shocks, a program intervention can take the form of cash-based anticipatory action to forecasted extreme climate events or the ex-post evaluation of the impact of such extreme event shocks on economic outcomes of local communities. Regarding suitable econometric methods, Botosaru, Giacomini & Weidner (Preprint arXiv, 2024) propose a framework for constructing forecasted treatment effects which is relevant to the aforementioned example. Additional econometric aspects include: (i) estimation of optimal combination of counterfactuals, (ii) uncertainty quantification under dataset shift; which allow to evaluate the impact of policy interventions and to determine the optimal combination of counterfactuals.
3. Main Research Objectives
We focus on econometric identification methods as well as estimation and inference procedures in the presence of heterogeneous treatment effects. We are interested in machine learning methods for estimation and inference, which are robust in the presence of heterogeneous effects on the economic outcomes of interest. Examining these robustness properties within a potential outcome framework worth further study. For example, Adusumilli (2023) propose asymptotics for sequential experiments, which facilitates statistical inference.
3.1 Literature Review and Recent Advances
Developing machine learning methods tailored to the case of quantile treatment effects and conditional quantile distributional effects, involves various challenges worth investigating further. Specifically, one approach is to set up the statistical problem from the social planner's perspective (e.g., see Spini (Preprint arXiv, 2021)), where a statistical distance metric is constructed in order to represent the decision of the planner based on the presence of covariate shifts. Moreover, Alvarez & Orestes (Preprint, 2024), propose a framework where the statistical distance metric corresponds to quantile-based measures. However a major challenge in the case of quantile treatment effects, is that the estimand of the conditional quantile function versus the estimand of the unconditional quantile function are not asymptotically equivalent (see Firbo et al. (Ecta, 2009)), since the law of iterated expectations does not hold in exact form in the case of quantiles. Asymptotic approximation results with respect to the conditional quantile operator are proposed by de Castro et al. (Bernoulli, 2023), who establish theory and representations for optimisation problems. Robust estimation of HTEs requires computational feasible techniques for estimating these distributional and/or conditional treatment effects at either fixed quantile levels or a continuum of quantiles. For example, extending the approach proposed by Firbo et al. (Ecta, 2009) to a framework with multivariate quantiles is examined by Merlo, Petrella, Salvati & Tzavidis (Preprint arXiv, 2023).
In the context of heterogeneous treatment effects, Wang & Zhu (Metrica, 2017) propose a conditional empirical likelihood approach in quantile regression models. Towards this direction, using quantile mixture density function techniques allows to obtain suitably balanced estimates for propensity score weights, which can then facilitate the construction of sample estimators for HTEs (e.g., L-moments approach which is used in quantile optimisation settings; see Fallahgoul, Mancini & Stoyanov (JFE, 2024)). When the statistical objective of interest is the measurement and estimation of outcome heterogeneity from discrete or continuous treatments, in the presence of covariate shifts, then one would be interested to use copula invariance principles to construct statistical comparisons under conditional quantile independence which is more challenging than conventional conditional mean independence (e.g., see Chernozhukov, Fernández-Val, Han & Wüthrich, (Preprint arXiv, 2024)). Lastly, evaluating the model adequacy using machine learning methods, is an important task. For example, the martingale difference approximation can be used to establish asymptotic theory for estimators and tests; since a conditional martingale transform approach permits to obtain asymptotically distribution-free test statistics.
3.2 Econometric Aspects of Research Interest
Methodologies for Identification and Estimation of Heterogeneous Treatment Effects for Panel Data.
Methodologies for Dynamic Treatment Effects and Counterfactual Analyses for Panel Data.
Double Debiased Machine Learning Methods in the context of Heterogeneous Treatment Effects.
High-Dimensional Estimation and Inference for Average/Quantile Treatment Effect under Treatment Heterogeneity. In particular, the development of causal inference machine learning techniques using conditional quantile functional forms, robust to the presence of covariate shifts due to treatment effects heterogeneity. From the social planner's statistical problem perspective, such a setting corresponds to the largest distribution shift before the decision (worst case distributional shift).
Econometric Inference for Unobserved Heterogeneity in Quantile Treatment Effects. Statistical testing within these settings requires caution with respect to the properties of covariates (such as a binary covariate versus a continuous covariate). These statistics are linked to the concept of testing for separability in structural relations which can be approximated via tests for conditional independence (see, Lu & White (JoE, 2014)).
Empirical applications of interest in relation to econometric methods for identification and estimation of HTEs are: dynamic causal effects of shocks to the macroeconomy and cross-country convergence, which are related to financial connectedness and economic linkages.
3.3 Empirical Examples of Quasi-Experimental Studies
Oregon Health Insurance Experiment:
The OHI experiment allows to implement the proposed econometric methods for studying the relation between insurance coverage and health utilization. Specifically, the OHI experiment considers a limited expansion of its Medicaid program for uninsured low-income adults by offering insurance coverage to the lottery winners from a waiting list of 90,000 people. These data collections allow to study the impact of insurance by a means of a large-scale randomized controlled trial across a low-income population. The outcome variable is a count variable which measures the number of outpatient visits in the last six months, which was elicited via a large-scale mail survey (sample of about 25,000 observations, after excluding individuals with missing information in any of the variables). Control variables include education level, gender, employment status, blood measurements and household size, indicators for the survey wave. Other measures include blood-pressure, cholesterol and glycated hemoglobin levels, screening for depression, medication inventories and self-reported diagnoses, health statues, health care utilization, and out-of-sample spending for such services. Specifically, this randomized controlled study showed that Medicaid coverage generated no significant improvement in measured physical health outcomes in the first 2 years, but it did increase use of health care services, rates of diabetes detection and management, lower rates of depression, and reduce financial strain. Another expected result is that participation to treatment is found to be a contributing factor to the frequency of doctor visits. For example, Baicker et al. (2013), found that about one year after enrolment those selected by the lottery have substantial and statistically significantly higher health care utilization, lower out-of-pocket medical expenditures and better self-reported health than the control group that was not given the opportunity to apply for Medicaid. Furthermore, a relevant outcome of interest is the impact of health insurance on increased health care utilization. From the policy planner perspective, the econometrician is interested to study the impact of health insurance on self-reported health, financial strain and overall well-being.
Moving to Opportunity Experiment:
The MTO housing experiment is based on the social theory of neighborhood effects (see, Sampson (AJS, 2008)). According to Sampson (AJS, 2008) the MTO experiment: "used randomization to solve the selection problem it has been said to offer the clearest answer so far to the threshold questions of whether important neighborhood effects exist". Within the counterfactual analysis paradigm, the question of interest is whether the same individual, residing in a poor neighborhood, would follow a different course if he or she in fact resided in a non-poor neighborhood (Does the offer of a housing voucher to move to a non-poor neighborhood affect the later outcomes of the extremely poor). Randomly assigning individuals to neighborhood treatment is the scientifically proposed way to equate otherwise dissimilar people, permitting estimation of an average causal effect. The experimental design is summarized as follows: families below the poverty line and living in concentrated poverty in five cities in the mid-1990s were deemed eligible to apply for housing vouchers. Those that did so were randomly assigned to groups such as experimental versus control. The experimental group was offered a housing voucher that, if used, had to be applied toward residence in a neighborhood with less than 10% poverty. In particular, the existence of the self-selection phenomenon especially within the experimental group, poses problems for valid causal inference. Specifically, the "intent to treat" effect will significantly underestimate the "treatment effect on the treated", or the effect of actually moving. A possible proposed solution to the presence of such biases is to ensure that randomization of being offered a voucher to be used as an instrumental variable to identify the impact of actually using a voucher. Thus, valid causal inference depends on the exclusion restriction that the offer of a voucher will affect outcomes only if participants use the voucher. Overall, the five main outcomes that have been studied are adult economic self-sufficiency, mental health, physical health, education, and risky behaviour. Significant positive effects of the MTO intervention have been reported for adult mental health, young female education, as well as physical and mental health of female adolescents. Our aim is to apply the proposed econometric methods in order to examine the conditional mean and quantile heterogeneous treatment effects, thereby evaluating the validity of the quasi-experimental results.
Extreme Weather Events Shocks 'Experiment':
Gros et al. (2019), consider the household-level effects of providing cash in anticipation of extreme weather events. In particular, the authors present the results of a mixed-methods quasi-experimental study, based on a post-disaster household survey. Specifically, their research assess the effectiveness of forecast-based cash distribution in helping beneficiaries to take preparatory early actions and reduce the negative impacts of the flood on their health, well-being, assets and livelihoods. The assessment shows that the cash grants contributed to improving households' access to food, a reduction in high-interest debt accrual of vulnerable households, and reduced psychological stress during and after the flood period, compared to a control group of similarly vulnerable and flood-affected communities that did not receive the forecast-based cash assistance. However, there is a need for further research to assess the longer-term effects of forecast-based cash on the socioeconomic development and well-being of the most vulnerable, which motivates further the usefulness of our econometric methods.
4. Literature and Bibliography
I. Econometrics and Statistics Literature
a. Statistical Theory and Methods
> Statistical Learning Theory
Bai, R., Zhang, Y., Yang, H., and Zhu, Z. (2024). "Transfer Learning for High-dimensional Quantile Regression with Distribution Shift". Preprint arXiv:2411.19933.
Beare, B.K., and Kaji, T. (2025). "Necessary and Sufficient Conditions for Convergence in Distribution of Quantile and PP Processes in L1(0, 1)". Preprint arXiv:2502.01254.
Li, T., Shi, C., Lu, Z., Li, Y., and Zhu, H. (2024). "Evaluating Dynamic Conditional Quantile Treatment Effects with Applications in Ridesharing". Journal of the American Statistical Association, 1-15.
Merlo, L., Petrella, L., Salvati, N., and Tzavidis, N. (2024). "Unified Unconditional Regression for Multivariate Quantiles, M-Quantiles, and Expectiles". Journal of the American Statistical Association, 119(547), 2154-2165.
Müller, M.M., Reeve, H.W., Cannings, T.I., and Samworth, R.J. (2024). "Isotonic Subgroup Selection". Journal of the Royal Statistical Society Series B.
Adusumilli, K. (2023). "Optimal Tests following Sequential Experiments". Preprint arXiv:2305.00403.
de Castro, L., Costa, B.N., Galvao, A.F., and Zubelli, J.P. (2023). "Conditional Quantiles: An Operator-Theoretical Approach". Bernoulli, 29(3), 2392-2416.
Duchi, J., Hashimoto, T., and Namkoong, H. (2023). "Distributionally Robust Losses for Latent Covariate Mixtures". Operations Research, 71(2), 649-664. Preprint arxiv:2007.13982.
Kallus, N., and Oprescu, M. (2023). "Robust and Agnostic Learning of Conditional Distributional Treatment Effects". In ICAIS (pp. 6037-6060).
Ma, S., Zhu, L., Zhang, Z., Tsai, C.L., and Carroll, R.J. (2019). "A Robust and Efficient Approach to Causal Inference Based on Sparse Sufficient Dimension Reduction". Annals of Statistics, 47(3), 1505.
Zhang, Y. (2018). "Extremal Quantile Treatment Effects". Annals of Statistics, 46(6B), 3707-3740.
> Policy Learning and Heterogeneous Treatment Effects
Mo, W., Tang, W., Xue, S., Liu, Y., and Zhu, J. (2024). "Minimax Regret Learning for Data with Heterogeneous Subgroups". Preprint arXiv:2405.01709.
Xu, R., Gao, J., Oka, T., and Whang, Y. J. (2024). "Quantile Random-Coefficient Regression with Interactive Fixed Effects: Heterogeneous Group-Level Policy Evaluation". Econometric Reviews, 1-19.
Giessing, A., and Wang, J. (2023). "Debiased Inference on Heterogeneous Quantile Treatment Effects with Regression Rank Scores". Journal of the Royal Statistical Society Series B, 85(5), 1561-1588.
Guo, X., Wei, L., Wu, C., and Wang, J. (2021). "Sharp Inference on Selected Subgroups in Observational Studies". Preprint arXiv:2102.11338.
Wang, L., Zhou, Y., Song, R., and Sherwood, B. (2018). "Quantile-Optimal Treatment Regimes". Journal of the American Statistical Association, 113(523), 1243-1254.
Wager, S., and Athey, S. (2018). "Estimation and Inference of Heterogeneous Treatment Effects using Random Forests". Journal of the American Statistical Association, 113(523), 1228-1242.
Ma, S., and Huang, J. (2017). "A Concave Pairwise Fusion Approach to Subgroup Analysis". Journal of the American Statistical Association, 112(517), 410-423.
Zhou, X., Mayer-Hamblett, N., Khan, U., and Kosorok, M.R. (2017). "Residual Weighted Learning for Estimating Individualized Treatment Rules". Journal of the American Statistical Association, 112(517), 169-187.
b. Econometric Methods and Applications
Baybutt, A., and Navjeevan, M. (2026). "Doubly-Robust Inference for Conditional Average Treatment Effects with High-Dimensional Controls". Journal of Econometrics, 253, 106180.
Chen, X., Chen, Z., and Gao, W. Y. (2025). "Inference on Welfare and Value Functionals under Optimal Treatment Assignment". Preprint arXiv:2510.25607.
Chen, X., Feng, X., Galvao, A.F., and Ge, Y. (2025). "Treatment Effects Inference with High-Dimensional Instruments and Control Variables". Preprint arXiv:2503.20149.
Chernozhukov, V. et al. (2025). "Policy Learning with Confidence". Preprint arxiv:2502.10653.
Jin, Z., and Sun, J. (2025). "Neyman-Orthogonal Moment for Instrumental Variable Quantile Regression Model with High Dimensional Data". Economics Letters, 112369.
Alvarez, L., and Orestes, V. (2024). "Quantile Mixture Models: Estimation and Inference". Working paper, MIT.
Bonhomme, S., Jochmans, K., and Weidner, M. (2024). "A Neyman-Orthogonalization Approach to the Incidental Parameter Problem". Preprint arXiv:2412.10304.
Botosaru, I., Giacomini, R., and Weidner, M. (2024). "Forecasted Treatment Effects with Short Panels". Available at SSRN 4815865.
Chernozhukov, V., Fernández-Val, I., Han, S., and Wüthrich, K. (2024). "Estimating Causal Effects of Discrete and Continuous Treatments with Binary Instruments". Preprint arXiv:2403.05850.
Ellul, S., Carlin, J. B., Vansteelandt, S., and Moreno-Betancur, M. (2024). "Causal Machine Learning Methods and Use of Sample Splitting in Settings with High-Dimensional Confounding". Preprint arXiv:2405.15242.
Fallahgoul, H., Mancini, L., and Stoyanov, S. (2024). "An L-Moment Approach for Portfolio Choice under Non-Expected Utility". Journal of Financial Econometrics, nbae027.
Ghanem, D., Sant'Anna, P. H., and Wüthrich, K. (2024). "Selection and Parallel Trends". Preprint arXiv:2203.09001.
Guggenberger, P., Mehta, N., and Pavlov, N. (2024). "Minimax Regret Treatment Rules with Finite Samples when a Quantile is the Object of Interest". Working paper, Department of Economics, Pennsylvania State University.
Martinez-Iriarte, J., Montes-Rojas, G., and Sun, Y. (2024). "Unconditional Effects of General Policy Interventions". Journal of Econometrics, 238(2), 105570.
Park, G. (2024). "Debiased Machine Learning when Nuisance Parameters Appear in Indicator Functions". Preprint arXiv:2403.15934.
Chernozhukov, V., Newey, M., Newey, W. K., Singh, R., and Srygkanis, V. (2023). "Automatic Debiased Machine Learning for Covariate Shifts". Preprint arXiv:2307.04527.
Hsu, Y.C., Huang, T.C., and Xu, H. (2023). "Testing for Unobserved Heterogeneous Treatment Effects with Observational Data". Econometric Theory, 39(3), 582-622.
Hsu, Y.C., Huber, M., and Yen, Y.M. (2023). "Doubly Robust Estimation of Direct and Indirect Quantile Treatment Effects with Machine Learning". Preprint arXiv:2307.01049.
Lei, L., and Ross, B. (2023). "Estimating Counterfactual Matrix Means with Short Panel Data". Preprint arXiv:2312.07520.
Lu, X., Miao, K., and Su, L. (2023). "Estimation of Heterogeneous Panel Data Models with an Application to Program Evaluation". Available at SSRN 4758814.
Chernozhukov, V., Escanciano, J.C., Ichimura, H., Newey, W.K., and Robins, J.M. (2022). "Locally Robust Semiparametric Estimation". Econometrica, 90(4), 1501-1535.
Chiang, H. D., Ma, Y., Rodrigue, J., and Sasaki, Y. (2022). "Dyadic Double/Debiased Machine Learning for Analyzing Determinants of Free Trade Agreements". Preprint arXiv:2110.04365.
Callaway, B. (2021). "Bounds on Distributional Treatment Effect Parameters using Panel Data with an Application on Job Displacement". Journal of Econometrics, 222(2), 861-881.
Sun, L., and Abraham, S. (2021). "Estimating Dynamic Treatment Effects in Event Studies with Heterogeneous Treatment Effects". Journal of Econometrics, 225(2), 175-199.
Spini, P. E. (2021). "Robustness, Heterogeneous Treatment Effects and Covariate Shifts". Preprint arXiv:2112.09259.
Callaway, B., and Li, T. (2019). "Quantile Treatment Effects in Difference in Differences Models with Panel Data". Quantitative Economics, 10(4), 1579-1618.
Ura, T. (2018). "Heterogeneous Treatment Effects with Mismeasured Endogenous Treatment". Quantitative Economics, 9(3), 1335-1370.
Heckman, J. J., Humphries, J. E., and Veramendi, G. (2016). "Dynamic Treatment Effects". Journal of Econometrics, 191(2), 276-292.
Qu, Z., and Yoon, J. (2015). "Nonparametric Estimation and Inference on Conditional Quantile Processes". Journal of Econometrics, 185(1), 1-19.
Donald, S.G., and Hsu, Y.C. (2014). "Estimation and Inference for Distribution Functions and Quantile Functions in Treatment Effect Models". Journal of Econometrics, 178, 383-397.
Tetenov, A. (2012). "Statistical Treatment Choice based on Asymmetric Minimax Regret Criteria". Journal of Econometrics, 166(1), 157-165.
Firpo, S., Fortin, N.M., and Lemieux, T. (2009). "Unconditional Quantile Regressions". Econometrica, 77(3), 953-973.
c. Synthetic Control Method and Applications
Gunsilius, F.F. (2025). "A Primer on Optimal Transport for Causal Inference with Observational Data". Preprint arXiv:2503.07811.
Chen, J. (2023). "Synthetic Control as Online Linear Regression". Econometrica, 91(2), 465-491.
Gunsilius, F.F. (2023). "Distributional Synthetic Controls". Econometrica, 91(3), 1105-1117.
Abadie, A., and L’hour, J. (2021). "A Penalized Synthetic Control Estimator for Disaggregated Data". Journal of the American Statistical Association, 116(536), 1817-1834.
Chernozhukov, V., Wüthrich, K., and Zhu, Y. (2021). "An Exact and Robust Conformal Inference Method for Counterfactual and Synthetic Controls". Journal of the American Statistical Association, 116(536), 1849-1864.
d. Sample Splitting for Time Series Analysis
Francq, C., Trapani, L., and Zakoian, J.M. (2025). "Inference on Breaks in Weak Location Time Series Models with Quasi-Fisher Scores". Working paper. Available at MPRA 123741.
Davis, R. A., and Fernandes, L. (2024). "Sample Splitting and Assessing Goodness-of-fit of Time Series". Preprint arXiv:2403.07158.
Lee, Y., and Wang, Y. (2023). "Threshold Regression with Nonparametric Sample Splitting". Journal of Econometrics, 235(2), 816-842.
McElroy, T. S., and Roy, A. (2022). "Model Identification via Total Frobenius Norm of Multivariate Spectra". Journal of the Royal Statistical Society Series B, 84(2), 473-495.
Strieder, D., and Drton, M. (2022). "On the Choice of the Splitting Ratio for the Split Likelihood Ratio Test". Electronic Journal of Statistics, 16(2), 6631-6650.
Zhang, X., Lu, Z., Wang, Y., and Zhang, R. (2020). "Adjusted Jackknife Empirical Likelihood for Stationary ARMA and ARFIMA Models". Statistics & Probability Letters, 165, 108830.
Lunde, R. (2019). "Sample Splitting and Weak Assumption Inference for Time Series". Preprint arXiv:1902.07425.
II. Applied Econometrics and Network Econometrics Literature
a. Applied Econometrics
(Causal Inference and Program Evaluation)
Elias, M., Gillitzer, C., Kaplan, G., La Cava, G., and Prasad, N. V. (2025). "The Mortgage Debt Channel of Monetary Policy when Mortgages are Liquid". NBER Working Paper (No. w34461). Available at nber/w34461.
Cronin, C.J., Forsstrom, M.P., and Papageorge, N.W. (2024). "What Good are Treatment Effects without Treatment? Mental Health and the Reluctance to use Talk Therapy". Review of Economic Studies, rdae061.
Adjaho, C., and Christensen, T. (2022). "Externally Valid Policy Choice". Preprint arXiv:2205.05561.
Bessone, P., Rao, G., Schilbach, F., Schofield, H., and Toma, M. (2021). "The Economic Consequences of Increasing Sleep Among the Urban Poor". Quarterly Journal of Economics, 136(3), 1887-1941.
Currie, J. M., and MacLeod, W. B. (2020). "Understanding Doctor Decision Making: The Case of Depression Treatment". Econometrica, 88(3), 847-878.
Gros, C., et al. (2019). "Household-Level Effects of providing Forecast-based Cash in Anticipation of Extreme Weather Events: Quasi-experimental Evidence from Humanitarian Interventions in the 2017 Floods in Bangladesh". International Journal of Disaster Risk Reduction, 41, 101275.
Kitagawa, T., and Tetenov, A. (2018). "Who Should be Treated? Empirical Welfare Maximization Methods for Treatment Choice". Econometrica, 86(2), 591-616.
b. Network Econometrics
(Causal Inference and Counterfactuals under Network Interference)
Chodorow-Reich, G., Gabaix, X., and Viviano, D. (2025). "Propagation of Shocks in Networks: Identification and Applications". Available at SSRN 4885984.
Chen, L., and Sasaki, Y. (2025). "Heterogeneous Effects of Endogenous Treatments with Interference and Spillovers in a Large Network". Preprint arXiv:2512.14515.
Jetsupphasuk, M., Li, D., and Hudgens, M. G. (2025). "Estimating Causal Effects using Difference-in-Differences under Network Dependency and Interference". Preprint arXiv:2502.03414.
Sun, K., and Xiao, Z. (2025). "Difference-in-Differences Under Network Interference". Preprint arXiv:2509.24259.
Shirani, S., Luo, Y., Overman, W., Xiong, R., and Bayati, M. (2025). "Can We Validate Counterfactual Estimations in the Presence of General Network Interference?". Preprint arXiv:2502.01106.
Wang, J., and Yang, S. (2025). "Estimation of Heterogeneous Treatment Effects in Network-Based Quasi-Experiments". Available at SSRN 5220294.
Wang, Y., and Jetsupphasuk, M. (2025). "Causal Inference in Longitudinal Data under Unknown Interference". Preprint arXiv:2106.15074.
Wu, H. (2025). "Unobserved Heterogeneous Spillover Effects in Instrumental Variable Models". Preprint arXiv:2511.22643.
Bong, H., Fogarty, C. B., Levina, E., and Zhu, J. (2024). "Heterogeneous Treatment Effects under Network Interference: A Nonparametric Approach Based on Node Connectivity". Preprint arXiv:2410.11797.
Gao, M. (2024). "Endogenous Interference in Randomized Experiments". Preprint arXiv:2412.02183.
Owusu, J. (2024). "A Nonparametric Test of Heterogeneous Treatment Effects under Interference". Preprint arXiv:2410.00733.
Viviano, D. (2024). "Policy Targeting under Network Interference". Review of Economic Studies, rdae041.
Leung, M. P. (2023). "Network Cluster‐Robust Inference". Econometrica, 91(2), 641-667.
Owusu, J. (2023). "Randomization Inference of Heterogeneous Treatment Effects under Network Interference". Preprint arXiv:2308.00202.
Viviano, D., Lei, L., Imbens, G., Karrer, B., Schrijvers, O., and Shi, L. (2023). "Causal Clustering: Design of Cluster Experiments under Network Interference". Preprint arXiv:2310.14983.
Auerbach, E. (2022). "Testing for Differences in Stochastic Network Structure". Econometrica, 90(3), 1205-1223.
Leung, M. P. (2022). "Causal Inference under Approximate Neighborhood Interference". Econometrica, 90(1), 267-293.
Leung, M. P. (2020). "Treatment and Spillover Effects under Network Interference". Review of Economics and Statistics, 102(2), 368-380.
Bargagli-Stoffi, F.J., Tortù, C., and Forastiere, L. (2020). "Heterogeneous Treatment and Spillover Effects under Clustered Network Interference". Preprint arXiv:2008.00707.
Zacchia, P. (2020). "Knowledge Spillovers through Networks of Scientists". Review of Economic Studies, 87(4), 1989-2018.
c. Social Learning and Team Formation
Akker, R. V. D., Werker, B. J., and Zhou, B. (2025). "Local Asymptotic Normality for Multi-Armed Bandits". Preprint arXiv:2512.12192.
Xu, Y., and Zhou, B. (2025). "Batched Adaptive Network Formation". Preprint arXiv:2507.18961.
Douglas, C., Provost, F., and Sundararajan, A. (2024). "Naive Algorithmic Collusion: When Do Bandit Learners Cooperate and When Do They Compete?". Preprint arXiv:2411.16574.
Bonhomme, S. (2021). "Teams: Heterogeneity, Sorting, and Complementarity". Preprint arXiv:2102.01802.
Weidmann, B., and Deming, D. J. (2021). "Team Players: How Social Skills Improve Team Performance". Econometrica, 89(6), 2637-2657.
III. Modelling Mental Health Trajectories and Risk Factors Literature
a. Time Series and Panel Data Methods for Mental Health
Andersen, A. L., Iyer, R., Johannesen, N., Jørgensen, M., and Peydró, J. L. (2025). "Household Leverage and Mental Health Fragility". CEPR Discussion Paper (No. 17711). Available at cepr/dp17711.
Simons, J. R., Chen, Y., Brunner, E., and French, E. (2025). "Forecasting Dementia Incidence". Preprint arXiv:2509.07874.
b. Machine Learning Methods for Mental Health
Brantner, C. L., Nguyen, T. Q., Parikh, H., Zhao, C., Hong, H., and Stuart, E. A. (2025). "Precision Mental Health: Predicting Heterogeneous Treatment Effects for Depression through Data Integration". Journal of the Royal Statistical Society Series C, qlaf068.
Kim, J. H., et al. (2025). "Improving Accuracy in the Estimation of Probable Dementia in Racially and Ethnically Diverse Groups with Penalized Regression and Transfer Learning". American Journal of Epidemiology, kwaf001.
Wen, S., Li, Y., Kong, D., and Lin, H. (2025). "Prediction of Cognitive Function via Brain Region Volumes with Applications to Alzheimer’s Disease Based on Space-Factor-Guided Functional Principal Component Analysis". Journal of the American Statistical Association, 1-13.
Guo, X., Zeng, D., and Wang, Y. (2024). "A Semiparametric Inverse Reinforcement Learning Approach to Characterize Decision Making for Mental Disorders". Journal of the American Statistical Association, 119(545), 27-38.
Zhang, C., et al. (2024). "Dynamic Risk Score Modelling for Multiple Longitudinal Risk Factors and Survival". Computational Statistics & Data Analysis, 189, 107837.
Avagyan, V., and Vansteelandt, S. (2022). "High-Dimensional Inference for the Average Treatment Effect under Model Misspecification using Penalized Bias-Reduced Double-Robust Estimation". Biostatistics & Epidemiology, 6(2), 221-238.
Kim, G.S., Paik, M.C., and Kim, H. (2017). "Causal Inference with Observational Data under Cluster-Specific Non-Ignorable Assignment Mechanism". Computational Statistics & Data Analysis, 113, 88-99.
Lin, W., Feng, R., and Li, H. (2015). "Regularization Methods for High-Dimensional Instrumental Variables Regression with an Application to Genetical Genomics". Journal of the American Statistical Association, 110(509), 270-288.
c. Psychological Dimensions of Mental Health
> Unobserved Heterogeneity and Health Outcomes
Cuddy, E., and Currie, J. (2026). "Rules versus Discretion: Treatment of Mental Illness in US Adolescents". Journal of Political Economy, 134(1), 000-000.
Costantini, S. (2025). "How Do Mental Health Treatment Delays Impact Long-Term Mortality?". American Economic Review, 115(5), 1672-1707.
Janys, L., and Siflinger, B. (2024). "Mental Health and Abortions among Young Women: Time-Varying Unobserved Heterogeneity, Health Behaviors, and Risky Decisions". Journal of Econometrics, 238(1), 105580.
Mazzonna, F., and Peracchi, F. (2024). "Are Older People Aware of their Cognitive Decline? Misperception and Financial Decision-Making". Journal of Political Economy, 132(6), 1793-1830.
> Environment Risk Factors and Health Outcomes
Chen, Y., Sun, R., Chen, X., and Qin, X. (2023). "Does Extreme Temperature Exposure take a toll on Mental Health? Evidence from the China Health and Retirement Longitudinal Study". Environment and Development Economics, 28(5), 486-510.
Chen, S., and Zhang, B. (2023). "Estimating and Improving Dynamic Treatment Regimes with a Time-Varying Instrumental Variable". Journal of the Royal Statistical Society Series B, 85(2), 427-453.
> Personality Risk Factors and Health Outcomes
Gilbert, J. B. et al. (2024). "Item-level Heterogeneous Treatment Effects of SSRIs on Depression: Implications for Inference, Generalizability, and Identification". Epidemiologic Methods, 13(1), 20240006.
Chen, Y. C., Cheng, M. Y., and Wu, H. T. (2014). "Non-parametric and Adaptive Modelling of Dynamic Periodicity and Trend with Heteroscedastic and Dependent Errors". Journal of the Royal Statistical Society Series B, 76(3), 651-682.
d. Neurobiological Dimensions of Mental Health
Liu, et al. (2025). "Subsecond Dopamine Fluctuations Do Not Specify the Vigor of Ongoing Actions". Nature Neuroscience, 1-7.
Lila, E., Zhang, W., and Rane Levendovszky, S. (2024). "Interpretable Discriminant Analysis for Functional Data supported on Random Nonlinear Domains with an Application to Alzheimer’s Disease". Journal of the Royal Statistical Society Series B, 86(4), 1013-1044.
Zhao, Q., Small, D. S., and Ertefaie, A. (2022). "Selective Inference for Effect Modification via the Lasso". Journal of the Royal Statistical Society Series B, 84(2), 382-413.
Bibliography:
Tsiatis, Anastasios A., et al. (2019). Dynamic Treatment Regimes: Statistical Methods for Precision Medicine. Chapman and Hall/CRC.
van der Laan, M.J., and Rose, S. (2018). Targeted Learning in Data Science. Springer International Publishing.
Chakraborty, B., and Moodie, E.E. (2013). Statistical Methods for Dynamic Treatment Regimes. Springer-Verlag.
Ziegler, A. (2011). Generalized Estimating Equations. Springer Science & Business Media.
Ruszczynski, A. (2011). Nonlinear Optimization. Princeton University Press.
Horowitz, J.L. (2009). Semiparametric and Nonparametric Methods in Econometrics. Springer Press.
Liese, F., and Miescke, K.J. (2008). Statistical Decision Theory. Estimation, Testing, and Selection. New York: Springer Press.
Kosorok, M. R. (2008). Introduction to Empirical Processes and Semiparametric Inference. New York: Springer Press.
Heckman, J.J., and Vytlacil, E.J. (2007). Econometric Evaluation of Social Programs, Part I: Causal Models, Structural Models and Econometric Policy Evaluation. Handbook of Econometrics, 6, 4779-4874.
Heckman, J.J., and Vytlacil, E.J. (2007). Econometric Evaluation of Social Programs, Part II: Using the Marginal Treatment Effect to Organize Alternative Econometric Estimators to Evaluate Social Programs, and to Forecast their Effects in New Environments. Handbook of Econometrics, 6, 4875-5143.
Tsiatis, Anastasios A. (2006). Semiparametric Theory and Missing Data. New York: Springer Press.
Boyd, S. (2004). Convex Optimization. Cambridge University Press.
Google Sites
Report abuse