nep-ecm New Economics Papers
on Econometrics
Issue of 2018‒08‒27
nineteen papers chosen by
Sune Karlsson
Örebro universitet

  1. Reducing Dimensions in a Large TVP-VAR By Eric Eisenstat; Joshua C.C. Chan; Rodney W. Strachan
  2. Change Point Estimation in Panel Data with Time-Varying Individual Effects By Otilia Boldea; Bettina Drepper; Zhuojiong Gan
  3. Trade Liberalization, Absorptive Capacity and the Protection of Intellectual Property Rights By YAMAMOTO, Yohei
  4. Semiparametric Identification in Panel Data Discrete Response Models By Eleni (E.) Aristodemou
  5. Change Point Detection in the Conditional Correlation Structure of Multivariate Volatility Models By Barassi, Marco; Horvath, Lajos; Zhao, Yuqian
  6. The Spatial Efficiency Multiplier and Common Correlated Effects in a Spatial Autoregressive Stochastic Frontier Model By Glass, Anthony J.; Kenjegalieva, Karligash; Sickles, Robin C.; Weyman-Jones, Thomas
  7. Asymptotic results under multiway clustering By Laurent Davezies; Xavier D'Haultfoeuille; Yannick Guyonvarch
  8. Intergenerational Income Elasticities, Instrumental Variable Estimation, and Bracketing Strategies By Pablo Mitnik
  9. Two-Step Estimation and Inference with Possibly Many Included Covariates By Matias D. Cattaneo; Michael Jansson; Xinwei Ma
  10. On Using Predictive-ability Tests in the Selection of Time-series Prediction Models: A Monte Carlo Evaluation By Costantini, Mauro; Kunst, Robert M.
  11. Probabilisitic Forecasts in Hierarchical Time Series By Puwasala Gamakumara; Anastasios Panagiotelis; George Athanasopoulos; Rob J Hyndman
  12. A Composite Likelihood Approach for Dynamic Structural Models By Canova, Fabio; Matthes, Christian
  13. Estimating Models with Dynamic Network Interactions and Unobserved Heterogeneity By Luisa Corrado; Salvatore Di Novo
  14. The Bigger Picture: Combining Econometrics with Analytics Improve Forecasts of Movie Success By Steven F. Lehrer; Tian Xie
  15. Model invariance when estimating random parameters with categorical variables By Burton, Michael
  16. Identifying Noise Shocks By Joshua Chan; Luca Benati; Eric Eisenstat; Gary Koop
  17. Predicting fixed effects in panel probit models By Kunz, J.S.;; Staub, K.E.;; Winkelmann, R.;
  18. The Simple Empirics of Optimal Online Auctions By Dominic Coey; Bradley Larsen; Kane Sweeney; Caio Waisman
  19. Big Data & Macroeconomic Nowcasting: Methodological Review By George Kapetanios; Fotis Papailias

  1. By: Eric Eisenstat (University of Queensland); Joshua C.C. Chan (University of Technology Sydney); Rodney W. Strachan (University of Queensland)
    Abstract: This paper proposes a new approach to estimating high dimensional time varying parameter structural vector autoregressive models (TVP-SVARs) by taking advantage of an empirical feature of TVP-(S)VARs. TVP-(S)VAR models are rarely used with more than 4-5 variables. However recent work has shown the advantages of modelling VARs with large numbers of variables and interest has naturally increased in modelling large dimensional TVP-VARs. A feature that has not yet been utilized is that the covariance matrix for the state equation, when estimated freely, is often near singular. We propose a speci?cation that uses this singularity to develop a factor-like structure to estimate a TVP-SVAR for 15 variables. Using a generalization of the re-centering approach, a rank reduced state covariance matrix and judicious parameter expansions, we obtain e¢ cient and simple computation of a high dimensional TVP-SVAR. An advantage of our approach is that we retain a formal inferential framework such that we can propose formal inference on impulse responses, variance decompositions and, important for our model, the rank of the state equation covariance matrix. We show clear empirical evidence in favour of our model and improvements in estimates of impulse responses.
    Keywords: Large VAR; time varying parameter; reduced rank covariance matrix
    JEL: C11 C22 E31
    Date: 2018–03–16
  2. By: Otilia Boldea; Bettina Drepper; Zhuojiong Gan
    Abstract: This paper proposes a method for estimating multiple change points in panel data models with unobserved individual effects via ordinary least-squares (OLS). Typically, in this setting, the OLS slope estimators are inconsistent due to the unobserved individual effects bias. As a consequence, existing methods remove the individual effects before change point estimation through data transformations such as first-differencing. We prove that under reasonable assumptions, the unobserved individual effects bias has no impact on the consistent estimation of change points. Our simulations show that since our method does not remove any variation in the dataset before change point estimation, it performs better in small samples compared to first-differencing methods. We focus on short panels because they are commonly used in practice, and allow for the unobserved individual effects to vary over time. Our method is illustrated via two applications: the environmental Kuznets curve and the U.S. house price expectations after the financial crisis.
    Date: 2018–08
  3. By: YAMAMOTO, Yohei
    Abstract: We propose a new method for the structural identification of a dynamic causal relationship in factor-augmented vector autoregression models based on changes in the unconditional shock variances that occur on a historical date. The proposed method can incorporate both observed and unobserved factors in the structural vector autoregression system and it allows the contemporaneous matrix to be fully unrestricted. We derive the asymptotic distribution of the impulse response estimator and consider a bootstrap inference method. Monte Carlo experiments show that the proposed method is robust to the misspecification of the contemporaneous matrix unlike the existing methods. Both the asymptotic and bootstrap methods obtain a satisfactory coverage rate when the shock of an observed factor is studied, although the latter is more accurate when the shock of an unobserved factor is considered. An empirical example based on the same data employed by Bernanke et al. (2005) provides similar point estimates and somewhat wider confidence intervals, thereby supporting their identification strategy.
    Keywords: dynamic casual effect, factor-augmented vector autoregression, identification through heteroskedasticity, impulse response
    JEL: C14 C22
    Date: 2018–08
  4. By: Eleni (E.) Aristodemou (University of Amsterdam, The Netherlands)
    Abstract: This paper studies semiparametric identification in linear index discrete response panel data models with fixed effects. Departing from the classic binary response static panel data model, this paper examines identification in the binary response dynamic panel data model and the ordered response static panel data model. It is shown that under mild distributional assumptions on the fixed effect and the time-varying unobservables, point-identification fails but informative bounds on the regression coefficients can still be derived. Partial identification is achieved by eliminating the fixed effect and discovering features of the distribution of the unobservable time-varying components that do not depend on the unobserved heterogeneity. Numerical analyses illustrate how the identified set changes as the support of the explanatory variables varies.
    Keywords: Static and Dynamic Panel Data; Binary Response Models; Ordered Response Models; Semiparametric Identification; Partial Identification
    JEL: C01 C33 C35
    Date: 2018–08–17
  5. By: Barassi, Marco; Horvath, Lajos; Zhao, Yuqian
    Abstract: We propose semi-parametric CUSUM tests to detect a change point in the correlation structures of non--linear multivariate models with dynamically evolving volatilities. The asymptotic distributions of the proposed statistics are derived under mild conditions. We discuss the applicability of our method to the most often used models, including constant conditional correlation (CCC), dynamic conditional correlation (DCC), BEKK, corrected DCC and factor models. Our simulations show that, our tests have good size and power properties. Also, even though the near--unit root property distorts the size and power of tests, de--volatizing the data by means of appropriate multivariate volatility models can correct such distortions. We apply the semi--parametric CUSUM tests in the attempt to date the occurrence of financial contagion from the U.S. to emerging markets worldwide during the great recession.
    Keywords: Change point detection, Time varying correlation structure, Volatility processes, Monte Carlo simulation, Contagion effect
    JEL: C12 C14 C32 G10 G15
    Date: 2018–07–11
  6. By: Glass, Anthony J. (Loughborough U); Kenjegalieva, Karligash (Loughborough U); Sickles, Robin C. (Rice U and Loughborough U); Weyman-Jones, Thomas (Loughborough U)
    Abstract: We extend the emerging literature on spatial frontier methods in a number of respects. One contribution includes accounting for unobserved heterogeneity. This involves developing a random effects spatial autoregressive stochastic frontier model which we generalize to a common correlated effects specification to account for correlation between the regressors and the unit specific effects. Another contribution is the introduction of the concept of a spatial efficiency multiplier to show that the efficiency frontiers from the structural and reduced forms of a spatial frontier model differ. To demonstrate various features of the estimators we develop we carry out a Monte Carlo simulation analysis and provide an empirical application. The application is to a state level cost frontier for U.S. agriculture which is a popular case in the efficiency literature and is thus well-suited to highlighting the features of the estimators we propose.
    JEL: C23 C51 D24 Q10
    Date: 2018
  7. By: Laurent Davezies; Xavier D'Haultfoeuille; Yannick Guyonvarch
    Abstract: If multiway cluster-robust standard errors are used routinely in applied economics, surprisingly few theoretical results justify this practice. This paper aims to fill this gap, by developing a general asymptotic theory for multiway clustering. We first prove, under nearly the same conditions as with i.i.d. data, the weak convergence of empirical processes under multiway clustering, as the number of clusters tends to infinity in each dimension. This result, which is new even under one-way clustering, implies simple central limit theorems for sample averages. It is also key for showing the asymptotic normality of nonlinear estimators such as GMM and smooth functionals of the empirical cumulative distribution function. We then establish consistency of various asymptotic variance estimators, including that of Cameron et al. (2011) but also a new estimator that is positive by construction. Next, we show the general consistency, for linear and nonlinear estimators, of the pigeonhole bootstrap, a resampling scheme adapted to multiway clustering. Monte Carlo simulations suggest that inference based on either asymptotic normality and our variance estimator or the pigeonhole bootstrap may be accurate even with a very small number of clusters in each dimension.
    Date: 2018–07
  8. By: Pablo Mitnik (Stanford University)
    Abstract: Although the intergenerational income elasticity (IGE) has long been the workhorse measure of economic mobility, this elasticity has been widely misinterpreted as pertaining to the conditional expectation of children’s income when it actually pertains to its conditional geometric mean. This has led to a call to replace it by the IGE of the expectation, which requires developing the methodological knowledge necessary to estimate the latter with short-run measures of income. This paper contributes to this aim. It advances a “bracketing strategy” for the estimation of the IGE of the expectation that is equivalent to that used to bracket the conventional IGE with estimates obtained with the Ordinary Least Squares and Instrumental Variable (IV) estimators. The proposed bracketing strategy couples estimates generated with the Poisson Pseudo Maximum Likelihood estimator and a Generalized Method of Moments IV estimator of the Poisson or exponential regression model. To achieve its goal the paper develops two generalized error-in-variables models for the IV estimation of the IGE of the expectation, and compares them to the corresponding model underlying the IV estimation of the conventional IGE. By considering the bracketing strategies from the perspective of the partial-identification approach to inference, the paper also specifies how to construct confidence intervals for the IGEs from the bounds estimated with those strategies. Lastly, using data from the Panel Study of Income Dynamics, the paper shows that the bracketing strategies work as expected, and assesses the information they generate and how this information varies across instruments and short-run measures of parental income.
    Keywords: intergenerational income elasticity, PSID, Panel Study of Income Dynamics, GMM, Generalized Method of Moments
    JEL: J62 C02
    Date: 2018–07
  9. By: Matias D. Cattaneo; Michael Jansson; Xinwei Ma
    Abstract: We study the implications of including many covariates in a first-step estimate entering a two-step estimation procedure. We find that a first order bias emerges when the number of \textit{included} covariates is "large" relative to the square-root of sample size, rendering standard inference procedures invalid. We show that the jackknife is able to estimate this "many covariates" bias consistently, thereby delivering a new automatic bias-corrected two-step point estimator. The jackknife also consistently estimates the standard error of the original two-step point estimator. For inference, we develop a valid post-bias-correction bootstrap approximation that accounts for the additional variability introduced by the jackknife bias-correction. We find that the jackknife bias-corrected point estimator and the bootstrap post-bias-correction inference perform excellent in simulations, offering important improvements over conventional two-step point estimators and inference procedures, which are not robust to including many covariates. We apply our results to an array of distinct treatment effect, policy evaluation, and other applied microeconomics settings. In particular, we discuss production function and marginal treatment effect estimation in detail.
    Date: 2018–07
  10. By: Costantini, Mauro (Department of Economics and Finance, Brunel University, London); Kunst, Robert M. (Institute for Advanced Studies, Vienna, and University of Vienna)
    Abstract: Comparative ex-ante prediction experiments over expanding subsamples are a popular tool for the task of selecting the best forecasting model class in finite samples of practical relevance. Flanking such a horse race by predictive-accuracy tests,such as the test by Diebold and Mariano (1995), tends to increase support for the simpler structure. We are concerned with the question whether such simplicity boosting actually benefits predictive accuracy in finite samples. We consider two variants of the DM test, one with naive normal critical values and one with bootstrapped critical values, the predictive-ability test by Giacomini and White (2006), which continues to be valid in nested problems, the F test by Clark and McCracken (2001), and also model selection via the AIC as a benchmark strategy. Our Monte Carlo simulations focus on basic univariate time-series specifications, such as linear (ARMA) and nonlinear (SETAR) generating processes.
    Keywords: Forecasting, time series, predictive accuracy, model selection
    JEL: C22 C52 C53
    Date: 2018–07
  11. By: Puwasala Gamakumara; Anastasios Panagiotelis; George Athanasopoulos; Rob J Hyndman
    Abstract: Forecast reconciliation involves adjusting forecasts to ensure coherence with aggregation constraints. We extend this concept from point forecasts to probabilistic forecasts by redefining forecast reconciliation in terms of linear functions in general, and projections more specifically. New theorems establish that the true predictive distribution can be recovered in the elliptical case by linear reconciliation, and general conditions are derived for when this is a projection. A geometric interpretation is also used to prove two new theoretical results for point forecasting; that reconciliation via projection both preserves unbiasedness and dominates unreconciled forecasts in a mean squared error sense. Strategies for forecast evaluation based on scoring rules are discussed, and it is shown that the popular log score is an improper scoring rule with respect to the class of unreconciled forecasts when the true predictive distribution coheres with aggregation constraints. Finally, evidence from a simulation study shows that reconciliation based on an oblique projection, derived from the MinT method of Wickramasuriya, Athanasopoulos and Hyndman (2018) for point forecasting, outperforms both reconciled and unreconciled alternatives.
    Date: 2018
  12. By: Canova, Fabio (BI Norwegian Business School, CAMP, and CEPR); Matthes, Christian (Federal Reserve Bank of Richmond)
    Abstract: We describe how to use the composite likelihood to ameliorate estimation, computational, and inferential problems in dynamic stochastic general equilibrium models. We present a number of situations where the methodology has the potential to resolve well-known problems. In each case we consider, we provide an example to illustrate how the approach works and its properties in practice.
    Keywords: dynamic structural models; composite likelihood; identification; singularity; large scale models; panel data
    JEL: C10 E27 E32
    Date: 2018–07–23
  13. By: Luisa Corrado (DEF & CEIS,University of Rome "Tor Vergata"); Salvatore Di Novo (University of Rome "Tor Vergata")
    Abstract: In this paper, we propose an approach to estimate models with network interactions in the presence of individual unobserved heterogeneity. The latter may impact the formation of ties and/or exogenous effects, thereby undermining identification of the associated parameters. In a panel setting, we devise a way to cope with these sources of endogeneity by relying on observable variations. When exogenous effect are involved, one can control for unobserved heterogeneity by including time-averages of the endogenous variables. When unobserved individual traits affect the process of network formation, it is possible to explore the role of network statistics. We derive a 2SLS estimator in order to address simultaneity bias, relying on sources of variation provided by the product between successive powers of the network matrix and the matrix of exogenous covariates; we assess the performances of the method via a Monte Carlo exercise, considering various combination of models and different ranges of parameters for both network interactions and the social multiplier. We also separately assess the cases in which unobserved sources hit the network structure only or act on exogenous effects as well. Focusing on the former case, our approach may be also applied when a simple cross-section is available. More generally, it does not require full knowledge of the spectrum of agents' interactions.
    Keywords: Networks,Individual Unobserved Heterogeneity,Dynamic Network Formation,network Statistics.
    JEL: C31 C36
    Date: 2018–08–09
  14. By: Steven F. Lehrer; Tian Xie
    Abstract: There exists significant hype regarding how much machine learning and incorporating social media data can improve forecast accuracy in commercial applications. To assess if the hype is warranted, we use data from the film industry in simulation experiments that contrast econometric approaches with tools from the predictive analytics literature. Further, we propose new strategies that combine elements from each literature in a bid to capture richer patterns of heterogeneity in the underlying relationship governing revenue. Our results demonstrate the importance of social media data and value from hybrid strategies that combine econometrics and machine learning when conducting forecasts with new big data sources. Specifically, while recursive partitioning strategies greatly outperform dimension reduction strategies and traditional econometric approaches in forecast accuracy, there are further significant gains from using hybrid approaches. Further, Monte Carlo experiments demonstrate that these benefits arise from the significant heterogeneity in how social media measures and other film characteristics influence box office outcomes.
    JEL: C52 C53
    Date: 2018–06
  15. By: Burton, Michael
    Abstract: This paper shows that econometric models that include categorical variables are not invariant to choice of ‘base’ category when random parameters are estimated, unless they are allowed to be correlated. We show that the invariance can lead to significant increases in Type I errors, and distortions in the implied behaviour of respondents. We hypothesise that these biases may influence the economic policy implications of published models that contain this error, but it’s impossible to be sure without re-estimating the model correctly.
    Keywords: Research Methods/ Statistical Methods
    Date: 2018–05–12
  16. By: Joshua Chan (University of Technology Sydney); Luca Benati (University of Bern); Eric Eisenstat (University of Queensland); Gary Koop (University of Strathclyde)
    Abstract: We make four contributions to the ‘news versus noise’ literature: (I) We provide a new identification scheme which, in population, exactly recovers news and noise shocks. (II) We show that our scheme is not vulnerable to Chahrour and Jurado’s (2018) criticism about the observational equivalence of news and noise shocks, which uniquely holds if the econometrician only observes a fundamental, and agents’ expectations about it. By contrast, we show that observational equivalence breaks down when the econometrician observes macroeconomic variables encoding information about the signal (and therefore about news and noise shocks), because they are chosen by agents conditional on all information, including the signal itself. (III) We propose a new econometric methodology for implementing our identification scheme, and we show, via a Monte Carlo study, that it has an excellent performance. (IV) We provide several empirical applications of our identification scheme and econometric methodology. Our results uniformly suggest that, contrary to previous findings in the literature, noise shocks play a minor role in macroeconomic fluctuations.
    Date: 2018–01–01
  17. By: Kunz, J.S.;; Staub, K.E.;; Winkelmann, R.;
    Abstract: We present a method to estimate and predict fixed effects in a panel probit model when N is large and T is small, and when there is a high proportion of individual units without variation in the binary response. Our approach builds on a bias-reduction method originally developed by Kosmidis and Firth (2009) for cross-section data. In contrast to other estimators, our approach ensures that predicted fixed effects are finite in all cases. Results from a simulation study document favorable properties in terms of bias and mean squared error. The estimator is applied to predict period-specific fixed effects for the extensive margin of health care utilization (any visit to a doctor during the previous three months), using German data for 2000-2014. We find a negative correlation between fixed effects and observed characteristics. Although there is some within-individual variation in fixed effects over sub-periods, the between-variation is four times as large.
    Keywords: Perfect prediction; Bias reduction; modified score function;
    JEL: I11 I18 C23 C25
    Date: 2018–08
  18. By: Dominic Coey; Bradley Larsen; Kane Sweeney; Caio Waisman
    Abstract: We study reserve prices computed to maximize the expected profit of the seller based on historical observations of incomplete bid data typically available to the auction designer in online auctions for advertising or e-commerce. This direct approach to computing reserve prices circumvents the need to fully recover distributions of bidder valuations. We derive asymptotic results and also provide a new bound, based on the empirical Rademacher complexity, for the number of historical auction observations needed in order for revenue under the estimated reserve price to approximate revenue under the optimal reserve arbitrarily closely. This simple approach to estimating reserves may be particularly useful for auction design in Big Data settings, where traditional empirical auctions methods may be costly to implement. We illustrate the approach with e-commerce auction data from eBay. We also demonstrate how this idea can be extended to estimate all objects necessary to implement the Myerson (1981) optimal auction.
    JEL: C10 D44 L10
    Date: 2018–06
  19. By: George Kapetanios; Fotis Papailias
    Abstract: This paper is concerned with an introduction to big data which can be potentially used in nowcasting the UK GDP and other key macroeconomic variables. We discuss various big data classifications and review some indicative studies in the big data and macroeconomic nowcasting literature. A detailed discussion of big data methodologies is also provided. In particular, we focus on sparse regressions, heuristic optimisation of information criteria, factor methods and textual-data methods.
    Keywords: Big Data, Machine Learning, Sparse Regressions, Factor Models
    JEL: C32 C53
    Date: 2018–07

This nep-ecm issue is ©2018 by Sune Karlsson. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.