|
on Econometrics |
| By: | Zongwu Cai (Department of Economics, The University of Kansas, Lawrence, KS 66045, USA); Ying Fang (The Wang Yanan Institute for Studies in Economics, Xiamen University, Xiamen, Fujian 361005, China); Qiuhua Xu (School of Finance, Southwestern University of Finance and Economics, Chengdu, Sichuan 611130, China) |
| Abstract: | This paper proposes a functional-coecient panel data model with cross-sectional dependence motivated by re-examining the empirical performance of conditional capital asset pricing model. In order to characterize the time-varying property of assets' betas and alpha, our proposed model allows the betas to be unknown functions of some macroeconomic and financial instruments. Moreover, a common factor structure is introduced to characterize cross-sectional dependence which is an attractive feature under a panel data regression setting as different assets or portfolios may be affected by same unobserved shocks. Compared to the existing studies, such as the classic Fama-MacBeth two-step procedure, our model can achieve substantial eciency gains for inference by adopting a one-step procedure using the entire sample rather than a single cross-sectional regression at each time point. We propose a local linear common correlated effects estimator for estimating time-varying betas by pooling the data. The consistency and asymptotic normality of the proposed estimators are established. Another methodological and empirical challenge in asset pricing is how to test the constancy of conditional betas and the significance of pricing errors, we echo this challenge by constructing an L2-norm statistic for functional-coefficient panel data models allowing for cross-sectional dependence. We show that the new test statistic has a limiting standard normal distribution under the null hypothesis. Finally, the method is applied to test the model in Fama and French (1993) using Fama-French 25 and 100 portfolios, sorted by size and book-to-market ratio, respectively, dated from July 1963 to July 2018. |
| Keywords: | Cross-sectional dependence; Functional coefficients; Local linear estimation; Nonlinear panel data models; Nonparametric test |
| JEL: | C12 C13 C14 C23 |
| Date: | 2020–09 |
| URL: | http://d.repec.org/n?u=RePEc:kan:wpaper:202009&r=all |
| By: | Eric Hillebrand; Manuel Lukas; Wei Wei |
| Abstract: | Relations between economic variables are often not exploited for forecasting, suggesting that predictors are weak in the sense that the estimation uncertainty is larger than the bias from ignoring the relation. In this paper, we propose a novel bagging estimator designed for such predictors. Based on a test for finite-sample predictive ability, our estimator shrinks the OLS estimate not to zero, but towards the null of the test which equates squared bias with estimation variance, and we apply bagging to further reduce the estimation variance. We derive the asymptotic distribution and show that our estimator can substantially lower the MSE compared to the standard ttest bagging. An asymptotic shrinkage representation for the estimator that simplifies computation is provided. Monte Carlo simulations show that the predictor works well in small samples. In an empirical application, we find that our proposed estimators works well for inflation forecasting using unemployment or industrial production as predictors. |
| Keywords: | inflation forecasting, bootstrap aggregation, estimation uncertainty, weak predictors, shrinkage methods. |
| JEL: | C13 C15 C18 |
| Date: | 2020 |
| URL: | http://d.repec.org/n?u=RePEc:msh:ebswps:2020-16&r=all |
| By: | Kim, Jihyun; Meddahi, Nour |
| Abstract: | Nowadays, a common method to forecast integrated variance is to use the fitted value of a simple OLS autoregression of the realized variance. However, non-parametric estimates of the tail index of this realized variance process reveal that its second moment is possibly unbounded. In this case, the behavior of the OLS estimators and the corresponding statistics are unclear. We prove that when the second moment of the spot variance is unbounded, the slope of the spot variance’s autoregression converges to a random variable as the sample size diverges. The same result holds when one uses the integrated or realized variance instead of the spot variance. We then consider the class of diffusion variance models with an affine drift, a class which includes GARCH and CEV processes, and we prove that IV estimation with adequate instruments provide consistent estimators of the drift parameters as long as the variance process has a finite first moment regardless of the existence of the second moment. In particular, for the GARCH diffusion model with fat tails, an IV estimation where the instrument equals the sign of the centered lagged value of the variable of interest provides consistent estimators. Simulation results corroborate the theoretical findings of the paper. |
| Keywords: | volatility; autoregression; fat tails; random limits. |
| Date: | 2020–05 |
| URL: | http://d.repec.org/n?u=RePEc:tse:wpaper:124237&r=all |
| By: | Svetlana Litvinova; Mervyn J. Silvapulle |
| Abstract: | We show that the full-sample bootstrap is asymptotically valid for constructing confidence intervals for high-quantiles, tail probabilities, and other tail parameters of a univariate distribution. This resolves the doubts that have been raised about the validity of such bootstrap methods. In our extensive simulation study, the overall performance of the bootstrap method was better than that of the standard asymptotic method, indicating that the bootstrap method is at least as good, if not better than, the asymptotic method for inference. This paper also lays the foundation for developing bootstrap methods for inference about tail events in multivariate statistics; this is particularly important because some of the non-bootstrap methods are complex. |
| Keywords: | full-sample bootstrap, intermediate order statistic, extreme value index, Hill estimator; tail probability, tail quantile. |
| JEL: | C13 C15 C18 |
| Date: | 2020 |
| URL: | http://d.repec.org/n?u=RePEc:msh:ebswps:2020-15&r=all |
| By: | Piero Mazzarisi; Silvia Zaoli; Carlo Campajola; Fabrizio Lillo |
| Abstract: | Identifying risk spillovers in financial markets is of great importance for assessing systemic risk and portfolio management. Granger causality in tail (or in risk) tests whether past extreme events of a time series help predicting future extreme events of another time series. The topology and connectedness of networks built with Granger causality in tail can be used to measure systemic risk and to identify risk transmitters. Here we introduce a novel test of Granger causality in tail which adopts the likelihood ratio statistic and is based on the multivariate generalization of a discrete autoregressive process for binary time series describing the sequence of extreme events of the underlying price dynamics. The proposed test has very good size and power in finite samples, especially for large sample size, allows inferring the correct time scale at which the causal interaction takes place, and it is flexible enough for multivariate extension when more than two time series are considered in order to decrease false detections as spurious effect of neglected variables. An extensive simulation study shows the performances of the proposed method with a large variety of data generating processes and it introduces also the comparison with the test of Granger causality in tail by [Hong et al., 2009]. We report both advantages and drawbacks of the different approaches, pointing out some crucial aspects related to the false detections of Granger causality for tail events. An empirical application to high frequency data of a portfolio of US stocks highlights the merits of our novel approach. |
| Date: | 2020–05 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2005.01160&r=all |
| By: | Poncela, Pilar; Ruiz, Esther |
| Abstract: | In this paper, the authors comment on the Monte Carlo results of the paper by Lucchetti and Veneti (A replication of "A quasi-maximum likelihood approach for large, approximate dynamic factor models" (Review of Economics and Statistics), 2020)) that studies and compares the performance of the Kalman Filter and Smoothing (KFS) and Principal Components (PC) factor extraction procedures in the context of Dynamic Factor Models (DFMs). The new Monte Carlo results of Lucchetti and Veneti (2020) refer to a DFM in which the relation between the factors and the variables in the system is not only contemporaneous but also lagged. The authors´ main point is that, in this context, the model specification, which is assumed to be known in Lucchetti and Veneti (2020), is important for the properties of the estimated factors. Furthermore, estimation of the parameters is also problematic in some cases. |
| Keywords: | Dynamic factor models,EM algorithm,Kalman filter,principal components |
| JEL: | C15 C32 C55 C87 |
| Date: | 2020 |
| URL: | http://d.repec.org/n?u=RePEc:zbw:ifwedp:20207&r=all |
| By: | Masayuki Sawada; Kohei Kawaguchi |
| Abstract: | We propose an estimation procedure for discrete choice models of differentiated products with possibly high-dimensional product attributes. In our model, high-dimensional attributes can be determinants of both mean and variance of the indirect utility of a product. The key restriction in our model is that the high-dimensional attributes affect the variance of indirect utilities only through finitely many indices. In a framework of the random-coefficients logit model, we show a bound on the error rate of a $l_1$-regularized minimum distance estimator and prove the asymptotic linearity of the de-biased estimator. |
| Date: | 2020–04 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2004.08791&r=all |
| By: | Chen, J.;; Gu, Y.;; Jones, A.M.;; Peng, B.; |
| Abstract: | The empirical and methodological efforts in using the generalised linear model to model healthcare costs have been mostly concentrated on selecting the correct link and variance functions. Another type of misspecification - misspecification of functional form of the key covariates - has been largely neglected. In many cases, continuous variables enter the model in linear form. This means that the relationship between the covariates and the response variable is entirely determined by the link function chosen which can lead to biased results when the true relationship is more complicated. To address this problem, we propose a hybrid model incorporating the extended estimating equations (EEE) model and partially linear additive functions. More specifically, we partition the index function in the EEE model into a number of additive components including a linear combination of some covariates and unknown functions of the remaining covariates which are believed to enter the index non-linearly. The estimator for the new model is developed within the EEE framework and based on the method of sieves. Essentially, the unknown functions are approximated using basis functions which enter the model just like the other predictors. This minimises the need for programming as the estimation itself can be completed using existing EEE software programs. The new model and its estimation procedure are illustrated through an empirical example focused on how children’s Body Mass Index (BMI) z-score measured at 4-5 years old relates to their accumulated healthcare costs over a 5-year period. Results suggest our new model can reveal complex relationships between covariates and the response variable. |
| Keywords: | body mass index; extended estimating equations; generalised linear model; healthcare cost; sieve estimation; |
| JEL: | C14 I10 P46 |
| Date: | 2020–02 |
| URL: | http://d.repec.org/n?u=RePEc:yor:hectdg:20/03&r=all |
| By: | Dimitris Korobilis; Davide Pettenuzzo |
| Abstract: | As the amount of economic and other data generated worldwide increases vastly, a challenge for future generations of econometricians will be to master efficient algorithms for inference in empirical models with large information sets. This Chapter provides a review of popular estimation algorithms for Bayesian inference in econometrics and surveys alternative algorithms developed in machine learning and computing science that allow for efficient computation in high-dimensional settings. The focus is on scalability and parallelizability of each algorithm, as well as their ability to be adopted in various empirical settings in economics and finance. |
| Date: | 2020–04 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2004.11486&r=all |
| By: | Karlsson, Sune (Örebro University School of Business); Mazur, Stepan (Örebro University School of Business) |
| Abstract: | We propose a general class of multivariate fat-tailed distributions which includes the normal, t and Laplace distributions as special cases as well as their mixture. Full conditional posterior distributions for the Bayesian VAR-model are derived and used to construct a MCMC-sampler for the joint posterior distribution. The framework allows for selection of a specific special case as the distribution for the error terms in the VAR if the evidence in the data is strong while at the same time allowing for considerable flexibility and more general distributions than offered by any of the special cases. As fat tails can also be a sign of conditional heteroskedasticity we also extend the model to allow for stochastic volatility. The performance is evaluated using simulated data and the utility of the general model specification is demonstrated in applications to macroeconomics. |
| Keywords: | Scale mixture of normals; Elliptically contoured distribution; Mixture distributions; Stochastic volatility; Markov Chain Monte Carlo |
| JEL: | C11 C15 C16 C32 C52 |
| Date: | 2020–04–27 |
| URL: | http://d.repec.org/n?u=RePEc:hhs:oruesi:2020_005&r=all |
| By: | Yeonwoo Rho; Yun Liu; Hie Joo Ahn |
| Abstract: | This paper proposes a new nonparametric mixed data sampling (MIDAS) model and develops a framework to infer clusters in a panel dataset of mixed sampling frequencies. The nonparametric MIDAS estimation method is more flexible but substantially less costly to estimate than existing approaches. The proposed clustering algorithm successfully recovers true membership in the cross-section both in theory and in simulations without requiring prior knowledge such as the number of clusters. This methodology is applied to estimate a mixed-frequency Okun's law model for the state-level data in the U.S. and uncovers four clusters based on the dynamic features of labor markets. |
| Date: | 2020–04 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2004.09770&r=all |
| By: | Mengya Liu; Fukan Zhu; Ke Zhu |
| Abstract: | This paper proposes a new family of multi-frequency-band (MFB) tests for the white noise hypothesis by using the maximum overlap discrete wavelet packet transform (MODWPT). The MODWPT allows the variance of a process to be decomposed into the variance of its components on different equal-length frequency sub-bands, and the MFB tests then measure the distance between the MODWPT-based variance ratio and its theoretical null value jointly over several frequency sub-bands. The resulting MFB tests have the chi-squared asymptotic null distributions under mild conditions, which allow the data to be heteroskedastic. The MFB tests are shown to have the desirable size and power performance by simulation studies, and their usefulness is further illustrated by two applications. |
| Date: | 2020–04 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2004.09161&r=all |
| By: | Giuseppe Cavaliere (Department of Economics, University of Bologna, Italy); Heino Bohn Nielsen (Department of Economics, University of Copenhagen, Denmark); Anders Rahbek (Department of Economics, University of Copenhagen, Denmark) |
| Abstract: | This article provides an introduction to methods and challenges underlying application of the bootstrap in econometric modelling of economic and financial time series. Validity, or asymptotic validity, of the bootstrap is discussed as this is a key element in deciding whether the bootstrap is applicable in empirical contexts. That is, as detailed here, bootstrap validity relies on regularity conditions, which need to be verified on a case-by-case basis. To fix ideas, asymptotic validity is discussed in terms of the leading example of bootstrap-based hypothesis testing in the well-known first order autoregressive model. In particular, bootstrap versions of classic convergence in probability and distribution, and hence of laws of large numbers and central limit theorems, are discussed as crucial ingredients to establish bootstrap validity. Regularity conditions and their implications for possible improvements in terms of (empirical) size and power for bootstrap-based testing, when compared to asymptotic testing, are illustrated by simulations. Following this, an overview of selected recent advances in the application of bootstrap methods in econometrics is also given. |
| Keywords: | Bootstrap theory; Bootstrap implementation; Econometric time series analysis; Testing; Asymptotic theory; Autoregressive models |
| JEL: | C12 C13 C15 C22 C32 C50 |
| Date: | 2020–12–17 |
| URL: | http://d.repec.org/n?u=RePEc:kud:kuiedp:2002&r=all |
| By: | Andreas Tryphonides |
| Abstract: | The paper demonstrates that identification in heterogeneous agent economies can be robust to alternative assumptions regarding the underlying structure using identifying restrictions that are consistent with a variety of mechanisms. These restrictions generate moment inequalities and therefore set identification, which can lead to wide confidence sets for the structural parameters. The paper shows that employing aggregated survey data that are informative about the extensive margin of adjustment i.e. the proportion of agents whose behavior is distorted over time due to financial frictions can provide additional information that can tighten the corresponding bounds. The paper provides identification analysis both in an extended partial equilibrium analytical example and in a general equilibrium setting. I apply this approach to the Spanish economy, where the proportion of constrained consumers is identified by combining information from different surveys. The results suggest that the extensive margin is empirically informative. |
| Date: | 2020–05 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2005.02010&r=all |
| By: | Kim, Jihyun; Park, Joon; Wang, Bin |
| Abstract: | In the paper, we introduce and analyze a new methodology to estimate the volatility functions of jump diffusion models. Our methodology relies on the standard kernel estimation technique using truncated bipower increments. The relevant asymptotics are fully developed, which allow for the time span to increase as well as the sampling interval to decrease and accommodate both stationary and nonstationary recurrent processes. We evaluate the performance of our estimators by simulation and provide some illustrative empirical analyses. |
| Keywords: | nonparametric estimation; jump diffusion;aymptotics; diffusive and jump; volatility functions; Lévy measure; optimal bandwidth; bipower increment; threshold truncation. |
| JEL: | C14 C22 C58 |
| Date: | 2020–05 |
| URL: | http://d.repec.org/n?u=RePEc:tse:wpaper:124234&r=all |
| By: | Jean-Jacques Forneron; Serena Ng |
| Abstract: | Assessing sampling uncertainty in extremum estimation can be challenging when the asymptotic variance is not analytically tractable. Bootstrap inference offers a feasible solution but can be computationally costly especially when the model is complex. This paper uses iterates of a specially designed stochastic optimization algorithm as draws from which both point estimates and bootstrap standard errors can be computed in a single run. The draws are generated by the gradient and Hessian computed from batches of data that are resampled at each iteration. We show that these draws yield consistent estimates and asymptotically valid frequentist inference for a large class of regular problems. The algorithm provides accurate standard errors in simulation examples and empirical applications at low computational costs. The draws from the algorithm also provide a convenient way to detect data irregularities. |
| Date: | 2020–04 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2004.09627&r=all |
| By: | Thomas-Agnan, Christine; Laurent, Thibault; Ruiz-Gazen, Anne; Nguyen, T.H.A; Chakir, Raja; Lungarska, Anna |
| Abstract: | Econometric land use models study determinants of land-use-shares of different classes: ``agriculture'', ``forest'', ``urban'' and ``other'' for example. Land-use-shares have a compositional nature as well as an important spatial dimension. We compare two compositional regression models with a spatial autoregressive nature in the framework of land use. We study the impact of the choice of coordinate space. We discuss parameters interpretation taking into account the non linear structure as well as the spatial dimension. We compute and interpret the semi-elasticities of the shares with respect to the explanatory variables and the spatial impact summary measures. |
| Keywords: | compositional regression model; marginal effects; simplicial derivative; elasticity; semi-elasticity. |
| JEL: | C10 C39 C46 C65 M31 Q15 |
| Date: | 2020–05 |
| URL: | http://d.repec.org/n?u=RePEc:tse:wpaper:124242&r=all |
| By: | Gael M. Martin; David T. Frazier; Christian P. Robert |
| Abstract: | The Bayesian statistical paradigm uses the language of probability to express uncertainty about the phenomena that generate observed data. Probability distributions thus characterize Bayesian inference, with the rules of probability used to transform prior probability distributions for all unknowns - models, parameters, latent variables - into posterior distributions, subsequent to the observation of data. Conducting Bayesian inference requires the evaluation of integrals in which these probability distributions appear. Bayesian computation is all about evaluating such integrals in the typical case where no analytical solution exists. This paper takes the reader on a chronological tour of Bayesian computation over the past two and a half centuries. Beginning with the one-dimensional integral first confronted by Bayes in 1763, through to recent problems in which the unknowns number in the millions, we place all computational problems into a common framework, and describe all computational methods using a common notation. The aim is to help new researchers in particular - and more generally those interested in adopting a Bayesian approach to empirical work - make sense of the plethora of computational techniques that are now on offer; understand when and why different methods are useful; and see the links that do exist, between them all. |
| Keywords: | history of Bayesian computation, Laplace approximation, Markov chain Monte Carlo, importance sampling, approximate Bayesian computation, Bayesian synthetic likelihood, variational Bayes, integrated nested Laplace approximation. |
| JEL: | C11 C15 C52 |
| Date: | 2020 |
| URL: | http://d.repec.org/n?u=RePEc:msh:ebswps:2020-14&r=all |
| By: | Skrobotov, Anton (Скроботов, Антон) (The Russian Presidential Academy of National Economy and Public Administration) |
| Abstract: | This paper propose the review of the model selection approaches based on information criteria. Different types of non-stationary modelas with unit root, cointegration and non-stationary volatility are considered. |
| Keywords: | unit root, information criteria, non-stationary volatility, seasonality |
| Date: | 2020–03 |
| URL: | http://d.repec.org/n?u=RePEc:rnp:wpaper:032015&r=all |
| By: | Ursula Laa; Dianne Cook; Andreas Buja; German Valencia |
| Abstract: | Multivariate data is often visualized using linear projections, produced by techniques such as principal component analysis, linear discriminant analysis, and projection pursuit. A problem with projections is that they obscure low and high density regions near the center of the distribution. Sections, or slices, can help to reveal them. This paper develops a section pursuit method, building on the extensive work in projection pursuit, to search for interesting slices of the data. Linear projections are used to define sections of the parameter space, and to calculate interestingness by comparing the distribution of observations, inside and outside a section. By optimizing this index, it is possible to reveal features such as holes (low density) or grains (high density). The optimization is incorporated into a guided tour so that the search for structure can be dynamic. The approach can be useful for problems when data distributions depart from uniform or normal, as in visually exploring nonlinear manifolds, and functions in multivariate space. Two applications of section pursuit are shown: exploring decision boundaries from classification models, and exploring subspaces induced by complex inequality conditions from multiple parameter model. The new methods are available in R, in the tourr package. |
| Keywords: | multivariate data, dimension reduction, projection pursuit, statistical graphics, data visualization, exploratory data analysis, data science. |
| Date: | 2020 |
| URL: | http://d.repec.org/n?u=RePEc:msh:ebswps:2020-17&r=all |
| By: | Ruda Zhang; Patrick Wingo; Rodrigo Duran; Kelly Rose; Jennifer Bauer; Roger Ghanem |
| Abstract: | Economic assessment in environmental science concerns the measurement or valuation of environmental impacts, adaptation, and vulnerability. Integrated assessment modeling is a unifying framework of environmental economics, which attempts to combine key elements of physical, ecological, and socioeconomic systems. Uncertainty characterization in integrated assessment varies by component models: uncertainties associated with mechanistic physical models are often assessed with an ensemble of simulations or Monte Carlo sampling, while uncertainties associated with impact models are evaluated by conjecture or econometric analysis. Manifold sampling is a machine learning technique that constructs a joint probability model of all relevant variables which may be concentrated on a low-dimensional geometric structure. Compared with traditional density estimation methods, manifold sampling is more efficient especially when the data is generated by a few latent variables. The manifold-constrained joint probability model helps answer policy-making questions from prediction, to response, and prevention. Manifold sampling is applied to assess risk of offshore drilling in the Gulf of Mexico. |
| Date: | 2020–04 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2004.11780&r=all |
| By: | Chan, Mark K.; Kwok, Simon |
| Abstract: | We develop a class of regression-based estimators, called Principal Components Difference-in-Differences estimators (PCDID), for treatment effect estimation. Analogous to a control function approach, PCDID uses factor proxies constructed from control units to control for unobserved trends, assuming that the unobservables follow an interactive effects structure. We clarify the conditions under which the estimands in this regression-based approach represent useful causal parameters of interest. We establish consistency and asymptotic normality results of PCDID estimators under minimal assumptions on the specification of time trends. We show how PCDID can be extended to micro/group-level data and be used for testing parallel trends under the interactive effects structure. The PCDID approach is illustrated in an empirical exercise that examines the effects of welfare waiver programs on welfare caseloads in the US. |
| Keywords: | principal components difference-in-differences, interactive fixed effects, factor- augmented regressions, treatment effects, parallel trends |
| Date: | 2020–03 |
| URL: | http://d.repec.org/n?u=RePEc:syd:wpaper:2020-03&r=all |
| By: | Francesca Molinari |
| Abstract: | This chapter reviews the microeconometrics literature on partial identification, focusing on the developments of the last thirty years. The topics presented illustrate that the available data combined with credible maintained assumptions may yield much information about a parameter of interest, even if they do not reveal it exactly. Special attention is devoted to discussing the challenges associated with, and some of the solutions put forward to, (1) obtain a tractable characterization of the values for the parameters of interest which are observationally equivalent, given the available data and maintained assumptions; (2) estimate this set of values; (3) conduct test of hypotheses and make confidence statements. The chapter reviews advances in partial identification analysis both as applied to learning (functionals of) probability distributions that are well-defined in the absence of models, as well as to learning parameters that are well-defined only in the context of particular models. A simple organizing principle is highlighted: the source of the identification problem can often be traced to a collection of random variables that are consistent with the available data and maintained assumptions. This collection may be part of the observed data or be a model implication. In either case, it can be formalized as a random set. Random set theory is then used as a mathematical framework to unify a number of special results and produce a general methodology to carry out partial identification analysis. |
| Date: | 2020–04 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2004.11751&r=all |
| By: | Michael Roberts; Indranil SenGupta |
| Abstract: | In this paper we present a sequential hypothesis test for the detection of general jump size distrubution. Infinitesimal generators for the corresponding log-likelihood ratios are presented and analyzed. Bounds for infinitesimal generators in terms of super-solutions and sub-solutions are computed. This is shown to be implementable in relation to various classification problems for a crude oil price data set. Machine and deep learning algorithms are implemented to extract a specific deterministic component from the crude oil data set, and the deterministic component is implemented to improve the Barndorff-Nielsen and Shephard model, a commonly used stochastic model for derivative and commodity market analysis. |
| Date: | 2020–04 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2004.08889&r=all |
| By: | Mattia Guerini; Patrick Musso; Lionel Nesta |
| Abstract: | We develop a new method to estimate the parameters of threshold distributions for market participation based upon an agent-specific attribute and its decision outcome. This method requires few behavioral assumptions, is not data demanding, and can adapt to various parametric distributions. Monte Carlo simulations show that the algorithm successfully recovers three different parametric distributions and is resilient to assumption violations. An application to export decisions by French firms shows that threshold distributions are generally right-skewed. We then reveal the asymmetric effects of past policies over different quantiles of the threshold distributions. |
| Keywords: | Parametric Distributions of Thresholds; Maximum Likelihood Estimation; Fixed Costs; Export Decision. |
| Date: | 2020–05–05 |
| URL: | http://d.repec.org/n?u=RePEc:ssa:lemwps:2020/09&r=all |
| By: | Marc-Oliver Pohle |
| Abstract: | I provide a unifying perspective on forecast evaluation, characterizing accurate forecasts of all types, from simple point to complete probabilistic forecasts, in terms of two fundamental underlying properties, autocalibration and resolution, which can be interpreted as describing a lack of systematic mistakes and a high information content. This "calibration-resolution principle" gives a new insight into the nature of forecasting and generalizes the famous sharpness principle by Gneiting et al. (2007) from probabilistic to all types of forecasts. It amongst others exposes the shortcomings of several widely used forecast evaluation methods. The principle is based on a fully general version of the Murphy decomposition of loss functions, which I provide. Special cases of this decomposition are well-known and widely used in meteorology. Besides using the decomposition in this new theoretical way, after having introduced it and the underlying properties in a proper theoretical framework, accompanied by an illustrative example, I also employ it in its classical sense as a forecast evaluation method as the meteorologists do: As such, it unveils the driving forces behind forecast errors and complements classical forecast evaluation methods. I discuss estimation of the decomposition via kernel regression and then apply it to popular economic forecasts. Analysis of mean forecasts from the US Survey of Professional Forecasters and quantile forecasts derived from Bank of England fan charts indeed yield interesting new insights and highlight the potential of the method. |
| Date: | 2020–05 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2005.01835&r=all |
| By: | Gilad Francis; Nick James; Max Menzies; Arjun Prakash |
| Abstract: | We develop a new method to find the number of volatility regimes in a non-stationary financial time series. We use change point detection to partition a time series into locally stationary segments, then estimate the distributions of each piece. The distributions are clustered into a learned number of discrete volatility regimes via an optimisation routine. Using this method, we investigate and determine a clustering structure for indices, large cap equities and exchange-traded funds. Finally, we create and validate a dynamic portfolio allocation strategy that learns the optimal match between the current distribution of a time series with its past regimes, thereby making online risk-avoidance decisions in the present. |
| Date: | 2020–04 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2004.09963&r=all |
| By: | Jesper R.-V. Soerensen (Department of Economics, University of Copenhagen, Denmark); Mogens Fosgerau (Department of Economics, University of Copenhagen, Denmark) |
| Abstract: | We study the additive random utility model of discrete choice under minimal assumptions. We make no assumptions regarding the distribution of random utility components or the functional form of systematic utility components. Exploiting the power of convex analysis, we are nevertheless able to generalize a range of important results. We characterize demand with a generalized Williams-Daly-Zachary theorem. A similarly generalized version of Hotz-Miller inversion yields constructive partial identification of systematic utilities. Estimators based on our partial identification result remain well defined in the presence of zeros in demand. We also provide necessary and sufficient conditions for point identification. |
| Keywords: | Additive random utility model; Discrete choice; Convex duality; Demand inversion; Partial identification |
| JEL: | C25 C6 D11 |
| Date: | 2020–12–17 |
| URL: | http://d.repec.org/n?u=RePEc:kud:kuiedp:2001&r=all |
| By: | Alexander Arimond; Damian Borth; Andreas Hoepner; Michael Klawunn; Stefan Weisheit |
| Abstract: | Utilizing a generative regime switching framework, we perform Monte-Carlo simulations of asset returns for Value at Risk threshold estimation. Using equity markets and long term bonds as test assets in the global, US, Euro area and UK setting over an up to 1,250 weeks sample horizon ending in August 2018, we investigate neural networks along three design steps relating (i) to the initialization of the neural network, (ii) its incentive function according to which it has been trained and (iii) the amount of data we feed. First, we compare neural networks with random seeding with networks that are initialized via estimations from the best-established model (i.e. the Hidden Markov). We find latter to outperform in terms of the frequency of VaR breaches (i.e. the realized return falling short of the estimated VaR threshold). Second, we balance the incentive structure of the loss function of our networks by adding a second objective to the training instructions so that the neural networks optimize for accuracy while also aiming to stay in empirically realistic regime distributions (i.e. bull vs. bear market frequencies). In particular this design feature enables the balanced incentive recurrent neural network (RNN) to outperform the single incentive RNN as well as any other neural network or established approach by statistically and economically significant levels. Third, we half our training data set of 2,000 days. We find our networks when fed with substantially less data (i.e. 1,000 days) to perform significantly worse which highlights a crucial weakness of neural networks in their dependence on very large data sets ... |
| Date: | 2020–05 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2005.01686&r=all |