Kansas Econometrics Workshop

Date: April 27 (Saturday), 2024

Venue: Gathering Room 2, The Oread Hotel

Hotel: Oread Hotel, located at 1200 Oread Avenue, Lawrence, KS 66044 (877) 263-6347 or  (785) 843-1200


07:30-08:20 Breakfast (Gathering Room 2, Oread Hotel)


08:20-08:30 Opening Remark: Richard Yi, Chair of Economics Department, KU

Session I: Chair, Zongwu Cai, University of Kansas


[1] 08:30 - 09:00 Bill Barnett, University of Kansas

“Nonseparability of Credit Card Services within Divisia Monetary Aggregates”

[2] 09:00 - 09:30 Xuming He, Washington University

“Some Recent Developments in Expected Shortfall Regression”

[3] 09:30-10:00 Eric Zivot, University of Washington

“Improving Price Leadership Share for Measuring Price Discovery”


               10:00-10:30 Coffee Break (Gathering Room 2, Oread Hotel)


                Session II: Chair, Zongwu Cai, University of Kansas


[4] 10:30-11:00 Jeffrey S. Racine, McMaster University

“Locally Adaptive Online Functional Data Analysis”

[5] 11:00-11:30 Runze Li, Pennsylvania State University

“Feature-Splitting Algorithms for Ultrahigh Dimensional Quantile Regression”

[6] 11:30-12:00 Mathew Harding, University of California at Irvine

“Combining Instrumental Variable Estimators for a Panel Model with Factors”


                12:00-14:00 Lunch (Gathering Room 2, Oread Hotel)


              Session III: Chair, John Keating, University of Kansas


[7] 14:00-14:30 Le Wang, Virginia Tech University

“Generalized Intergenerational Mobility Regressions”

[8] 14:30-15:00 Daniel Henderson, University of Alabama

“Estimation and Inference for Varying Coefficient Multidimensional Fixed-Effects Panel Data Models”

[9] 15:00-15:30 Wei Long, Tulane University

“Estimation of Intergenerational Mobility: A Time-Varying Mixed Copula Method”


            15:30-16:00 Coffee Break (Gathering Room 2, Oread Hotel)


                Session IV:  John Keating, University of Kansas

[10] 16:00-16:30 Huixin Bi, Federal Reserve Bank of Kansas City

“Dynamics of Job Search Effort and Vacancies: Evidence from Classified Advertisements”

[11] 16:30-17:00 Shu Shen, University of California at Davis

“Dynamics Panel Instrumental Variable Regression with Varying-intensity Repeated Treatments: Theory and the China Syndrome Application”

[12] 17:00-17:30 Molin Zhong, Board of Governors of the Federal Reserve System

“Risk and Monetary Policy in a Data-Rich Model”

[13] 17:30-18:00 John Keating, University of Kansas

“Structurally Interpreting Mis-specified Long-run Recursive Vector Autoregressions”


          18:15-21:00 Reception and Dinner at the Hotel (Gathering Room 2)



"Nonseparability of Credit Card Services within Divisia Monetary Aggregates"

Bill Barnett, University of Kansas

We use the New-Keynesian DSGE framework and VAR to investigate the usefulness and relevancy of monetary services, augmented to include credit card transaction services. We use the new credit-card-augmented Divisia monetary aggregates in the models to further the existing research on their usefulness and relevancy. In this research, we compare three different monetary aggregates within the New-Keynesian framework: (1) the aggregation theoretical “true” monetary aggregate, (2) the credit-card-augmented Divisia monetary aggregate, and (3) the simple sum monetary aggregate. We acquire the following primary results. (1) The credit-card-augmented Divisia monetary aggregate tracks the theoretical (true) monetary aggregate, while simple-sum does not. Although this result would be expected from the theory in classical economic models, the result is not an immediate implication of the theory in New-Keynesian models and therefore needs empirical confirmation. (2) Under the recursive VAR framework, the credit-card-augmented Divisia monetary aggregate serves as a preferable monetary policy indicator compared to the traditional federal funds rate. (3) On theoretical grounds, we find that the separability condition for existence of a monetary aggregator function would fail, if credit card deferred payment services were excluded from the monetary services block, unless all markets are perfect. Co-author: Hyun Park.

"Some Recent Developments in Expected Shortfall Regression"

Xuming He, Washington University

Expected shortfall, measuring the average outcome (e.g., portfolio loss) above a given quantile of its probability distribution, is a common financial risk measure. The same measure can be used to characterize treatment effects in the tail of an outcome distribution, with applications ranging from policy evaluation in economics and public health to biomedical investigations. Expected shortfall regression is a natural approach of modeling covariate-adjusted expected shortfalls. Because the expected shortfall cannot be written as a solution of an expected loss function at the population level, computational as well as statistical challenges around expected shortfall regression have led to stimulating research. We discuss some recent developments in this area, with a focus on a new optimization-based semiparametric approach to estimation of conditional expected shortfall that adapts well to data heterogeneity with minimal model assumptions. Co-author: Yuanzhi Li. 

"Improving Price Leadership Share for Measuring Price Discovery"

Eric Zivot, University of Washington

We propose an improvement to the information leadership (IL) measure of price discovery as introduced in Yan and Zivot (2010), and the information leadership share (ILS) measure proposed by Putnins (2013). Our improved IL and ILS measures integrate the price discovery share (PDS) from Sultan and Zivot (2015) with the component share (CS) from Gonzalo and Granger (1995). In contrast to the IL metric by Yan and Zivot (2010), which combines the information share (IS) measure from Hasbrouck (1995) with CS, we demonstrate that our improved IL measure accurately reflects the ratio of initial responses of competing markets to a permanent shock, even when the residuals of reduced-form vector error correction models are correlated. Simulation evidence strongly supports the superiority of our measures over a wide spectrum of existing price discovery metrics (Lien and Shrestha, 2009; Putnis, 2013; Sultan and Zivot, 2015; Patel et al., 2020). We demonstrate the effectiveness of our improved measures by examining price discovery for various Chinese stocks cross-listed in Shanghai and Hong Kong (SH-HK) both before and after the initiation of the Shanghai-Hong Kong Stock Connect.  Co-authors: Shulin Shen and Yixuan Zhang.

"Locally Adaptive Online Functional Data Analysis"

Jeffrey S. Racine, McMaster University

One drawback with classical smoothing methods (kernels, splines, wavelets etc.) is their reliance on assuming the degree of smoothness (and thereby assuming continuous differentiability up to some order) for the underlying object being estimated. However, the underlying object may in fact be irregular (i.e., non-smooth and even perhaps nowhere differentiable) and, as well, the (ir)regularity of the underlying function may vary across its support. Elaborate adaptive methods for curve estimation have been proposed, however, their intrinsic complexity presents a formidable and perhaps even insurmountable barrier to their widespread adoption by practitioners. We contribute to the functional data literature by providing a pointwise MSE-optimal, data-driven, iterative plug-in estimator of “local regularity” and a computationally attractive, recursive, online updating method. In so doing we are able to separate measurement error “noise” from “irregularity” thanks to “replication”, a hallmark of functional data. Our results open the door for the construction of minimax optimal rates, “honest” confidence intervals, and the like, for various quantities of interest.

"Feature-Splitting Algorithms for Ultrahigh Dimensional Quantile Regression"

Runze Li, Pennsylvania State University

This paper is concerned with computational issues related to penalized quantile regression (PQR) with ultrahigh dimensional predictors. Various algorithms have been developed for PQR, but they become ineffective and/or infeasible in the presence of ultrahigh dimensional predictors due to the storage and scalability limitations. The variable updating schema of the feature-splitting algorithm that directly applies the ordinary alternating direction method of multiplier (ADMM) to ultrahigh dimensional PQR may make the algorithm fail to converge. To tackle this hurdle, we propose an efficient and parallelizable algorithm for ultrahigh dimensional PQR based on the three-block ADMM. The compatibility of the proposed algorithm with parallel computing alleviates the storage and scalability limitations of a single machine in the large-scale data processing. We establish the rate of convergence of the newly proposed algorithm. In addition, Monte Carlo simulations are conducted to compare the finite sample performance of the proposed algorithm with that of other existing algorithms. The numerical comparison implies that the proposed algorithm significantly outperforms the existing ones. We further illustrate the proposed algorithm via an empirical analysis of a real-world data set.

"Combining Instrumental Variable Estimators for a Panel Model with Factors"

Mathew Harding, University of California at Irvine

We address the estimation of factor-augmented panel data models using observed measurements to proxy for unobserved factors or loadings and explore the use of internal instruments to address the resulting endogeneity. The main challenge consists in that economic theory rarely provides insights into which measurements to choose as proxies when several are available. To overcome this problem, we propose a new class of estimators that are linear combinations of instrumental variable estimators and establish large sample results. We also show that an optimal weighting scheme exists, leading to efficiency gains relative to an instrumental variable estimator. Simulations show that the proposed approach performs better than existing methods. We illustrate the new method using data on test scores across US school districts. Co-authors: Carlos Lamarche and Chris Muris.

"Generalized Intergenerational Mobility Regressions"

Le Wang, Virginia Tech University

Current research on intergenerational mobility (IGM) is informed by statistical approaches based on log-level and rank regressions, whose economic interpretations remain largely unknown. We reveal the subjective value-judgements in them: they are represented by weighted-sums (or aggregators) over heterogeneous groups, with controversial economic properties. The rank-regression weights are improper and inconsistent aggregators. We propose a general construction of IGM measures that can incorporate any transparent economic preferences. They are interpreted as the marginal effect of parental normalized social welfare on children's normalized welfare. Conventional regressions are special cases with implicit economic preferences that fail inequality-aversion and the Pigou-Dalton principle of transfers. Empirically, we show that the disparate findings in the literature may reflect differences in subjective valuation of different groups, rather than "mobility". A variety of economic preferences, with varying inequality aversion, demonstrate a nuanced view of mobility, and perspectives on geographic-differences and dynamics of it.

"Estimation and Inference for Varying Coefficient Multidimensional Fixed-Effects Panel Data Models"

Daniel Henderson, University of Alabama

This paper presents an estimation method and an array of hypothesis tests for varying coefficient multidimensional panel data regression models. We derive the asymptotic distribution of the proposed nonparametric estimator, and to construct valid tests, we develop the necessary central limit theory to conduct inference. The presence of multiple effects over differing dimensions requires nontrivial changes to the well-known central limit theory for U-statistics. The types of inference we can conduct offer a diverse array of hypotheses for applied work and we explicitly present test statistics for some of the most important hypothesis tests. To illustrate the usefulness of our proposed suite of tests, we provide an empirical application focusing on the gravity model of international trade. We first reject the random effects specification and then find that the result from the cross-sectional setting that the linear in parameters model cannot be rejected versus a semiparametric alternative extends to the multidimensional panel data setting. Lastly, an appendix contains a detailed set of simulations that supports our estimator's asymptotic developments and reveals that our testing infrastructure possesses correct asymptotic size and high power. Co-authors: Christopher F. Parmeter and Alexandra Soberon.

"Estimation of Intergenerational Mobility: A Time-Varying Mixed Copula Method"

Wei Long, Tulane University

This study examines the trajectory of intergenerational mobility in the United States over the past three decades. Employing a time-varying mixed copula model as outlined by Yang et al. (2022), our findings reveal a gradual decline in intergenerational mobility, particularly in terms of permanent incomes, between parents and children over time. This heightened income inequality persists across various racial and regional demographics, although subgroup analyses unveil some heterogeneous patterns in the paths of intergenerational mobility. Co-authors: Zongwu Cai, Qi Li, Guannan Liu and Xuehong Luo.

"Dynamics of Job Search Effort and Vacancies: Evidence from Classified Advertisements"

Huixin Bi, Federal Reserve Bank of Kansas City

We introduce a new dataset of job search effort and vacancies from the early twentieth century, providing new insights into the response of search effort to labor market conditions. Combining scanned images of U.S. newspapers with morphological and textual analysis, we construct measures of job posts placed by both firms and job seekers in the classified advertisements. To be the best of our knowledge, we are the first to systematically document and use the “situation-wanted” advertisements placed by job seekers. We use moments from national measures of the help- and situation-wanted ads to derive a range of plausible estimates for the elasticity of search effort costs and discipline a model of equilibrium search effort by the unemployed to examine its implications for labor market dynamics. Co-authors: Nora Traum and Nicolas Petrosky-Nadeau.

"Dynamics Panel Instrumental Variable Regression with Varying-intensity Repeated Treatments: Theory and the China Syndrome Application"

Shu Shen, University of California at Davis

Instrumental variable models with repeated endogenous treatments are popular in empirical research. This paper shows that if treatment effect dynamics is present and external instruments are serially correlated, the current empirical approaches adopted in the literature is invalid. Using the proposed new model and semi-parametric approach, we find strong evidence of path-dependency in the contemporaneous impact of increased Chinese import competition on U.S.\ manufacturing employment. Specifically, the trade shock an industry received over 1991-1999 monotonically magnifies the negative impact of the 1999-2011 trade shock. The magnifying effect is mild initially but soars after the 1991-1999 import exposure passes a certain threshold.

"Risk and Monetary Policy in a Data-Rich Model"

Molin Zhong, Board of Governors of the Federal Reserve System

In this paper, we quantify the role of financial conditions and U.S. monetary policy in shaping risk measures associated with a large set of economic indicators. Specifically, we estimate a factor augmented VAR model with endogenous stochastic volatility and isolate U.S. financial and monetary policy shocks. We find substantial heterogeneity in how risk evolves over the business cycle both across economic indicators and across sectors of the economy. Furthermore, preliminary findings reveal that monetary policy can help reduce downside risks. Co-authors: Dario Caldara and Haroon Mumtaz.

"Structurally Interpreting Mis-specified Long-run Recursive Vector Autoregressions"

John Keating, University of Kansas

Recursive empirical models are easy to estimate but are often and for good reason considered mis-specified. Consequently, these models may seem incapable of yielding useful information about economic structure. However, this paper proves long-run recursive models used to identify vector autoregressions that are indeed mis-specified may still be informative about features of a structure. This is possible under weaker assumptions than would be necessary to consistently estimate a structure. This paper shows that approaches taken in some existing work generalize to reduced-form time series model of arbitrary size. Specifically, parameters from a general linear structure with an arbitrary number of endogenous variables are mapped into coefficients of a long-run recursive VAR model. The long-run and dynamic effects of shocks in the long-run recursive model each are specific functions of particular structural parameters. Given assumptions about signs of some of these structural parameters, a long-run recursive model's estimates may in fact be used to infer the sign of some other structural effect. Specific examples illustrate ways the general results can be used.