|
on Econometrics |
| By: | Fabrizio Durante; Ostap Okhrin; ; |
| Abstract: | Complex phenomena in environmental sciences can be conveniently represented by several inter-dependent random variables. In order to describe such situations, copula-based models have been studied during the last year. In this paper, we consider a novel family of bivariate copulas, called exchangeable Marshall copulas. Such copulas describe both positive and (upper) tail association between random variables. Specically, inference procedures for the family of exchangeable Marshall copulas are introduced, based on the estimation of their (univariate) generator. Moreover, the performance of the proposed methodologies is shown in a simulation study. Finally, an illustration describes how the proposed procedures can be useful in a hydrological application. |
| Keywords: | Copula, Kendall distribution, Marshall-Olkin distribution, Non-parametric Estimation, Risk Management |
| JEL: | C13 C14 |
| Date: | 2014–01 |
| URL: | http://d.repec.org/n?u=RePEc:hum:wpaper:sfb649dp2014-014&r=ecm |
| By: | Robertson, Donald; Sarafidis, Vasilis; Westerlund, Joakim |
| Abstract: | This paper proposes a new panel unit root test based on the generalized method of moments approach for panels with a small number of time periods and a large number of cross-section units, N. In the model that we consider the deterministic trend function is essentially unrestricted and the errors are cross-sectionally correlated in a very general fashion. In spite of these allowances, the GMM-statistic is shown to be asymptotically unbiased, square root N-consistent and asymptotically normal for all values of the autoregressive (AR) coefficient, ρ, including unity, making it an ideal candidate for unit root inference. Results from both simulated and real data are provided to suggest that the asymptotic properties are borne out well in small samples. |
| Keywords: | Panel data, unit root test, cross-section dependence, common factors, GMM. |
| JEL: | C12 C13 C33 C36 |
| Date: | 2014–02–05 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:53419&r=ecm |
| By: | Elena Di Bernardino (CEDRIC - Centre d'Etude et De Recherche en Informatique du Cnam - Conservatoire National des Arts et Métiers (CNAM)); Didier Rullière (SAF - Laboratoire de Sciences Actuarielle et Financière - Université Claude Bernard - Lyon I : EA2429) |
| Abstract: | Calculating return periods and critical layers (i.e., multivariate quantile curves) in a multivariate environment is a di cult problem. A possible consistent theoretical framework for the calculation of the return period, in a multi-dimensional environment, is essentially based on the notion of copula and level sets of the multivariate probability distribution. In this paper we propose a fast and parametric methodology to estimate the multivariate critical layers of a distribution and its associated return periods. The model is based on transformations of the marginal distributions and transformations of the dependence structure within the class of Archimedean copulas. The model has a tunable number of parameters, and we show that it is possible to get a competitive estimation without any global optimum research. We also get parametric expressions for the critical layers and return periods. The methodology is illustrated on hydrological 5-dimensional real data. On this real data-set we obtain a good quality of estimation and we compare the obtained results with some classical parametric competitors |
| Keywords: | Multivariate probability transformations; level sets estimation; copulas; hyperbolic conversion functions; risk assessment; multivariate return periods. |
| Date: | 2014–01–31 |
| URL: | http://d.repec.org/n?u=RePEc:hal:wpaper:hal-00940089&r=ecm |
| By: | Christian Aßmann; Jens Boysen-Hogrefe; Markus Pape |
| Abstract: | Due to their indeterminacies, static and dynamic factor models require identifying assumptions to guarantee uniqueness of the parameter estimates. The indeterminacy of the parameter estimates with respect to orthogonal transformations is known as the rotation problem. The typical strategy in Bayesian factor analysis to solve the rotation problem is to introduce ex-ante constraints on certain model parameters via degenerate and truncated prior distributions. This strategy, however, results in posterior distributions whose shapes depend on the ordering of the variables in the data set. We propose an alternative approach where the rotation problem is solved ex-post using Procrustean postprocessing. The resulting order invariance of the posterior estimates is illustrated in a simulation study and an empirical application using a well-known data set containing 120 macroeconomic time series. Favorable properties of the ex-post approach with respect to convergence, statistical and numerical accuracy are revealed |
| Keywords: | Bayesian Estimation; Factor Models; Multimodality; Rotation Problem; Ordering Problem; Orthogonal Transformation |
| JEL: | C11 C31 C38 C51 C52 |
| Date: | 2014–01 |
| URL: | http://d.repec.org/n?u=RePEc:kie:kieliw:1902&r=ecm |
| By: | BAUWENS, Luc; STORTI, Giuseppe |
| URL: | http://d.repec.org/n?u=RePEc:cor:louvrp:-2469&r=ecm |
| By: | Silvia Figini (Department of Political and Social Sciences, University of Pavia); Mario Maggi (Department of Economics and Management, University of Pavia) |
| Abstract: | The performance of predictions models can be assessed using a variety of methods and metrics. Several new measures have recently been proposed that can be seen as refinements of discrimination measures, including variants of the AUC (Area Under the ROC curve), such as the H index. It is widely recognized that AUC suffers from lack of coherency especially when ROC curves cross. On the other hand, the H index requires subjective choices. In our opinion the problem of model comparison should be more adequately handled using a different approach. The main contribution of this paper is to evaluate the performance of prediction models using proper loss function. In order to compare how our approach works with respect to classical measures employed in model comparison, we propose a simulation studies, as well as a real application on credit risk data. |
| Keywords: | Model Comparison, AUC, H index, Loss Function, Proper Scoring Rules, Credit Risk |
| Date: | 2014–01 |
| URL: | http://d.repec.org/n?u=RePEc:pav:demwpp:demwp0064&r=ecm |
| By: | Paul Eckerstorfer; Johannes Halak (Government of the Federal State of Upper Austria); Jakob Kapeller (Department of Philosophy and Theory of Science, University of Linz, Linz, Austria); Bernhard Schütz (Department of Economics, University of Linz, Linz, Austria); Florian Springholz (Department of Philosophy and Theory of Science, University of Linz, Linz, Austria); Rafael Wildauer (Department of Economics, Kingston University London) |
| Abstract: | It is a well-known criticism that due to its exponential distribution, survey data on wealth is hardly reliable when it comes to analyzing the richest parts of society. This paper addresses this criticism using Austrian data from the Household Finance and Consumption Survey (HFCS). In doing so we apply the assumption of a Pareto distribution to obtain estimates for the number of households possessing a net wealth greater than four million Euros as well as their aggregate wealth holdings. Thereby, we identify suitable parameter combinations by using a series of maximum-likelihood estimates and appropriate goodness-of-fit tests to avoid arbitrariness with respect to the fitting of the Pareto-Distribution. Our results suggest that the alleged non-observation bias is considerable, accounting for about one quarter of total net wealth. The method developed in this paper can easily be applied to other countries where survey data on wealth are available. |
| Keywords: | wealth distribution, non-observation bias, Pareto distribution |
| JEL: | C46 C81 D31 |
| Date: | 2014–01 |
| URL: | http://d.repec.org/n?u=RePEc:jku:econwp:2014_01&r=ecm |
| By: | Yves S. Schüler (Department of Economics, University of Konstanz, Germany) |
| Abstract: | This paper proposes to relate conditional quantiles of stationary macroeconomic time series to the different phases of the business cycle. Based on this idea, I introduce a Bayesian Quantile Structural Vector Autoregressive framework for the analysis of the effects of uncertainty on the US real economy. For this purpose, I define a novel representation of the multivariate Laplace distribution that allows for the joint treatment of multiple equation regression quantiles. I find significant evidence for asymmetric effects of uncertainty over the US business cycle. The strongest negative effects are revealed during recession periods. During boom phases uncertainty shocks improve the soundness of the economy. Moreover, the phase of the financial sector matters when the real economy is at recession but not if the economy is at boom. When the financial system is in a bad state, an uncertainty shock leads to a deeper recession than in times when the financial system is in a good state. |
| Keywords: | Uncertainty, Economic Cycles, Quantile SVAR, Multivariate Laplace |
| JEL: | C32 E44 G01 |
| Date: | 2014–01–27 |
| URL: | http://d.repec.org/n?u=RePEc:knz:dpteco:1402&r=ecm |
| By: | Wojciech Charemza; Carlos Diaz; Svetlana Makarova |
| Abstract: | Empirical evaluation of macroeconomic uncertainties and their use for probabilistic forecasting are investigated. A new weighted skew normal distribution which parameters are interpretable in relation to monetary policy outcomes and actions is proposed. This distribution is fitted to recursively obtained forecast errors of monthly and annual inflation for 38 countries. It is found that this distribution fits inflation forecasts errors better than the two-piece normal distribution, which is often used for inflation forecasting. The new type of ‘fan charts’ net of the epistemic (potentially predictable) element is proposed and applied for UK and Poland. |
| Keywords: | macroeconomic forecasting, inflation, uncertainty, monetary policy, non-normality, density forecasting |
| JEL: | C54 E37 E52 |
| Date: | 2014–01 |
| URL: | http://d.repec.org/n?u=RePEc:lec:leecon:14/01&r=ecm |
| By: | Anton Velinov |
| Abstract: | This paper discusses the type of trajectory a country's public debt path follows. In particular, a Markov switching ADF model is used to assess the sustainability of public debt by testing whether a government's present value borrowing constraint holds. Building on the work of Raybaudi et al. (2004) and Chen (2011), the model in this paper generalizes their methodology. The number of lags and states are in principle unrestricted and all of the parameters can be switching. Debt trajectories of 16 countries are investigated using long time series on debt/GDP obtained from Reinhart and Rogoff (2011). Two different criteria are used to test the null hypothesis of a unit root in each state. The countries with a sustainable debt path are found to be Finland, Norway, Sweden, Switzerland and the UK, while Greece and Japan are found to have unsustainable debt trajectories. The debt paths of the remaining countries are mainly characterized as being in a unit root state and are therefore labeled as uncertain. Robustness tests indicate that the model is robust to the sample size and the number of states used. Further, it is shown that the models used in this paper offer an improvement to existing models investigating this subject. |
| Keywords: | Markov switching, debt trajectory, debt sustainability, unit root |
| JEL: | C22 C24 H60 |
| Date: | 2014 |
| URL: | http://d.repec.org/n?u=RePEc:diw:diwwpp:dp1359&r=ecm |
| By: | Pedro Saramago (Centre for Health Economics, University of York, UK); Ling-Hsiang Chuang (Pharmerit International, Rotterdam, the Netherlands); Marta Soares (Centre for Health Economics, University of York, UK) |
| Abstract: | Objectives: Network meta-analysis (NMA) methods extend the standard pair-wise framework to allow simultaneous comparison of multiple interventions in a single statistical model. Despite published work on NMA mainly focussing on the synthesis of aggregate data (AD), methods have been developed that allow the use of individual patient-level data (IPD) specifically when outcomes are dichotomous or continuous. This paper focuses on the synthesis of IPD and AD time to event data, motivated by a real data example looking at the effectiveness of high compression treatments on the healing of venous leg ulcers. Methods: This paper introduces a novel NMA modelling approach that allows IPD (time to event with censoring) and AD (event count for a given follow-up time) to be synthesised jointly by assuming an underlying, common, distribution of time to healing. Alternative model assumptions were tested within the motivating example. Model fit and adequacy measures were used to compare and select models. Results: Due to the availability of IPD in our example we were able to use a Weibull distribution to describe time to healing; otherwise, we would have been limited to specifying a uniparametric distribution. Absolute effectiveness estimates were more sensitive than relative effectiveness estimates to a range of alternative specifications for the model. Conclusions: The synthesis of time to event data considering IPD provides modelling flexibility, and can be particularly important when absolute effectiveness estimates, and not just relative effect estimates, are of interest. |
| Date: | 2014–01 |
| URL: | http://d.repec.org/n?u=RePEc:chy:respap:95cherp&r=ecm |