Sonderforschungsbereich (SFB) 823

Permanent URI for this collection

http://hdl.handle.net/2003/26250

Statistical modelling of nonlinear dynamic processes

Im Zentrum des SFB stehen zeitvariable dynamische Prozesse in den Wirtschafts- und Ingenieurwissenschaften. Die statistische Modellbildung in diesen Bereichen sieht sich mit vielfältigen intervenierenden Variablen und komplexen Prozessen mit zum Teil unübersichtlichen Abhängigkeiten konfrontiert, die sich mit konventionellen Modellen nicht beschreiben lassen. Ein Beispiel: In der aktuellen Finanzkrise haben fast alle ökonomischen Modelle bei Diagnose und Prognose versagt. Während 2007 in ruhigeren Börsenzeiten die Aktienmärkte unterschiedliche Entwicklungen und Trends zeigten, riss 2008 die Krise nahezu alle ins Minus, mit nahezu prozentual gleichen Verlusten. Wieso nehmen internationale Kapitalmarktabhängigkeiten in wirtschaftlichen Abschwungphasen drastisch zu? Und wie ist zu erklären, dass die jeweiligen Märkte in Aufschwungphasen nicht diese simultane Kursausschläge zeigen? Die abrupten und/oder graduellen Änderungen - die so genannten Strukturbrüche - zu finden und zu quantifizieren, ist das wichtigste Ziel der Wissenschaftler im neuen SFB.

Und diese Probleme beschränken sich keineswegs auf die Wirtschaft. ähnliche Probleme existieren auch in den Ingenieurwissenschaften. So ist beispielsweise bei der Blechumformung oder Betonverarbeitung nicht davon auszugehen, das Variablen im Prozess immer konstant ihren Einfluss ausüben, sondern es auch hier zu Strukturbrüchen kommt, die in die statistische Modellbildung einfließen müssen.

Browse

Now showing 1 - 20 of 581

Online monitoring of dynamic networks using flexible multivariate control charts
(2022) Flossdorf, Jonathan; Fried, Roland; Jentsch, Carsten
The identification of differences in dynamic networks between various time points is an important task and involves statistical procedures like two-sample tests or changepoint detection. Due to the rather complex nature of temporal graphs, the analysis is challenging which is why the complexity is typically reduced to a metric or some sort of a model. This is not only likely to result in a loss of relevant information, but common approaches also use restrictive assumptions and are therefore heavily limited in their usability. We propose an online monitoring approach usable for flexible network structures and able to handle various types of changes. It is based on a sound choice of a set of network characteristics under consideration of their mathematical properties which is crucial in order to cover the relevant information. Subsequently, those metrics are jointly monitored in a suitable multivariate control chart scheme which performs superior to a univariate analysis and enables both parametric and non-parametric usage. The user also benefits from a handy interpretation of the structural reasons for the detected changes which is a crucial advantage in the rather complex field of dynamic networks. Our findings are supported by an extensive simulation study.
The influence of different diamond spacings in diamond impregnated tools on the wear behavior and material removal
(2022) Dreier, Julia; Ferreira, Manuel; Malcherczyk, Dennis; Biermann, Dirk; Müller, Christine H.; Tillmann, Wolfgang
The influence of the spacing of four diamonds on the breakout time and material removal is investigated for a diamond impregnated tool for machining concrete workpieces. A statistical analysis using the Cox-model shows a positive effect of larger spacings on the lifetime of the diamonds where no effect on the material removal can be found. Moreover, a relationship between the position of the diamond and its lifetime is observed.
Model checks and simultaneous prediction bands for load sharing models in prestressed concrete beams
(2022) Leckey, Kevin; Heinrich, Jens; Müller, Christine H.; Maurer, Reinhard
This article presents a new method to test on whether a parametric model is capable of describing data properly. It also introduces a simple procedure to generate simultaneous prediction bands based on independent copies of a process. The performance of these prediction bands, e.g. in a leave-one-out cross-validation, will also be used as another indication of whether data is modeled properly. Both methods are applied to data from fatigue experiments on prestressed concrete beam girders. These experiments highlight a couple of different influences on the fatigue of such girders, namely the so-called cable factor and the deflection force. Both effects are incorporated into different load sharing models for component failures which then are compared and used for predicting these failure times.
Pro-environmental behavior as a means of self-signaling: Theory and evidence
(2022) Flörchinger, Daniela; Frondel, Manuel; Sommer, Stephan; Andor, Mark A.
Recent research indicates that pro-environmental behavior may be driven by concerns about one’s moral identity. Using identiﬁcation with the environmentalist movement Fridays for Future, this paper develops and empirically tests a straightforward model of self-signaling. We assume that pro-environmental behavior, here taking the train rather than the plane for a journey, serves as a means of self-signaling. On the basis of a large-scale survey experiment with revealed preferences, we ﬁnd evidence that respondents who receive an identity prime in the form of a reminder of their previously stated attitude towards Fridays for Future are more likely to behave in line with the movement’s moral principles in that they take the train. Our explanation of this outcome is that individuals attempt to avoid cognitive dissonance by choosing the more environmentally benign alternative. Our results suggest that pro-environmental behavior may be enhanced by appealing to an individual’s self-image so that costly interventions that are designed to convince subjects of new moral principles may be unnecessary.
Fairness and the support of redistributive environmental policies
(2021) Andor, Mark A.; Lange, Andreas; Sommer, Stephan
Exemptions from costly policy measures are frequently applied to alleviate financial burdens to specific market participants. Using a stated-choice experiment with around 6,000 German household heads, we test how exemptions for lowincome households and energy-intensive companies influence the political acceptability of additional cost for the promotion of renewable energies. We find that the support for the policy is substantially higher when low-income households are exempt rather than the industry. Introducing exemptions for low-income households on top of existing exemptions for the industry increases the acceptability of the policy. We show that the support for exemptions as one example of distributional policy design is associated with individual behavioral measures like inequality aversion and fairness perceptions.
Körperschallanalyse der Ermüdung von Spannbetonbauteilen
(2021) Dreier, Julia; Hafer, Marlies; Heinrich, Jens; Leckey, Kevin; Malcherczyk, Dennis; Maurer, Reinhard; Müller, Christine H.
Das Ermüdungsversagen von Spannbetonbauteilen mit Spanngliedern aus mehrdrähtigen Litzenbündeln erfolgt i. d. R. drahtweise. Das Brechen solcher einzelner Spanndrähte kann unter Laborbedingungen akustisch sehr gut exakt erfasst werden. Hier wird untersucht, ob das Brechen der Spanndrähte schon kurz vorher akustisch messbar ist. Dies erfolgte über eine Körperschallanalyse mit 512 Frequenzen im Frequenzbereich 0.003 MHz – 1.562 MHz. Wir beschreiben hier eine Analyse-Möglichkeit, identifizieren Probleme dabei und machen Vorschläge für verbesserte zukünftige Analysen. Insbesondere war es von Nachteil, dass nur in einem zweistündigen Takt zweimal pro Stunde nur für ca. Sekunden gemessen werden konnte. Damit konnte kein Zusammenhang zu den Drahtbrüchen festgestellt werden. Allerdings ergab sich ein Zusammenhang zu einem Steifigkeitsparameter.
Fiscal policy, international spillovers, and endogenous productivity
(2021) Klein, Mathias; Linnemann, Ludger
The paper presents empirical evidence on the international effects of US fiscal policy from structural vector autoregressions identified through external instruments in a panel setting for the G7 countries. An exogenous increase in US government spending is estimated to produce sizeable positive responses of output and consumption in the rest of the G7 countries, both about half as large as their domestic US counterparts, while strongly depreciating the US terms of trade and lowering short-run real interest rates. Moreover, fiscal shocks are estimated to have a strongly positive impact on hourly labor productivity in the private sector. We solve a two-country New Keynesian model in closed form and show that a low cost elasticity of varying technology utilization can simultaneously explain the positive productivity, consumption and international spillover effects as well as the real depreciation resulting from expansionary US government spending shocks.
Akzeptanz der CO2-Bepreisung in Deutschland: Die hohe Bedeutung der Rückverteilung der Einnahmen
(2021) Frondel, Manuel; Helmers, Viola; Mattauch, Linus; Pahle, Michael; Sommer, Stephan; Schmidt, Christoph M.; Edenhofer, Ottmar
Im Jahr 2021 wurde in Deutschland die sogenannte CO2-Bepreisung fossiler Kraft- und Brennstoffe eingeführt, um deren Verbrauch zum Zwecke des Klimaschutzes zu reduzieren. Dieser Preisaufschlag auf fossile Energieträger wird in den kommenden Jahren sukzessive erhöht. Dieser Beitrag untersucht die Akzeptanz der CO2-Bepreisung für die Zeit vor Einführung des CO2-Preises im Jahr 2019. Eine Erhebung unter mehr als 6.000 Haushalten zeigt, dass eine leichte absolute Mehrheit von 53,7 % der Befragten grundsätzlich bereit ist, zu Klimaschutzzwecken höhere Kosten in Kauf zu nehmen. Die Zustimmung zu einer CO2-Bepreisung nimmt jedoch mit sinkendem Einkommen deutlich ab: Bei Befragten der untersten Einkommensgruppe liegt die Zustimmungsrate bei knapp unter 40 %. Erwartungsgemäß verringert sich die Zustimmung auch mit der Höhe des CO2-Preises. So wurde ein CO2-Preis von 50 Euro von einer Mehrheit der Befragten von 50,6 % abgelehnt. Um bei bis zum Jahr 2025 auf 55 Euro steigenden CO2-Preisen die mehrheitliche Akzeptanz der Bürger zu gewinnen, wird hier für einen breit angelegten Ausgleichsmechanismus durch die Reduzierung verzerrender und sozial ungerechter Steuern und Abgaben auf den Strompreis plädiert, die insbesondere Gering- und Durchschnittsverdienern zugutekommt. Andernfalls könnten die über die Zeit steigenden CO2-Preise eine hohe soziale Sprengkraft entfalten.
Efficiency gains in structural vector autoregressions by selecting informative higher-order moment conditions
(2021) Keweloh, Sascha Alexander; Hetzenecker, Stephan
This study combines block-recursive restrictions with non-Gaussian and mean independent shocks to derive identifying and overidentifying higher-order moment conditions for structural vector autoregressions. We show that overidentifying higher-order moments can contain additional information and increase the efficiency of the estimation. In particular, we prove that in the non-Gaussian recursive SVAR higher-order moment conditions are relevant and therefore, the frequently applied estimator based on the Cholesky decomposition is inefficient. Even though incorporating information in valid higher-order moments is asymptotically efficient, including many redundant and potentially even invalid moment conditions renders standard SVAR GMM estimators unreliable in finite samples. We apply a LASSO-type GMM estimator to select the relevant and valid higher-order moment conditions, increasing finite sample precision. A Monte Carlo experiment and an application to quarterly U.S. data illustrate the improved performance of the proposed estimator.
Approximation and error analysis of forward-backward SDEs driven by general Lévy processes using shot noise series representations
(2021) Massing, Till
We consider the simulation of a system of decoupled forward-backward stochastic differential equations (FBSDEs) driven by a pure jump Lévy process L and an independent Brownian motion B. We allow the Lévy process L to have an infinite jump activity. Therefore, it is necessary for the simulation to employ a finite approximation of its Lévy measure. We use the generalized shot noise series representation method by Rosinski (2001) to approximate the driving Lévy process L. We compute the Lp error, p > 2, between the true and the approximated FBSDEs which arises from the finite truncation of the shot noise series (given sufficient conditions for existence and uniqueness of the FBSDE). We also derive the Lp error between the true solution and the discretization of the approximated FBSDE using an appropriate backward Euler scheme.
K-depth tests for testing simultaneously independence and other model assumptions in time series
(2021) Dohme, Hendrik; Malcherczyk, Dennis; Leckey, Kevin; Müller, Christine
We consider the recently developed K-depth tests for testing simultaneously independence and other model assumptions for univariate time series with a potentially related d-dimensional process of explanatory variables. Since these tests are based only on signs of residuals, they are easy to comprehend. They can be used in a full version and in a simplified version. While former investigations already showed that the full version is appropriate for testing model assumptions, we concentrate here on either testing the independence assumption on its own or on simultaneously testing independence- and model assumptions with both types of tests. In an extensive simulation study, we compare these tests with several known independence test such as the runs test, the Durbin-Watson test, and the Von-Neumann-Rank-Ratio test. Finally, we demonstrate how the K-depth tests can be used for improved modelling of crack width time series depending on temperature measurements in a bridge monitoring.
Block-recursive non-Gaussian structural vector autoregressions
(2021) Keweloh, Sascha Alexander; Hetzenecker, Stephan; Seepe, Andre
This study combines block-recursive restrictions with higher-order moment conditions to identify and estimate non-Gaussian structural vector autoregressions. The estimator allows to impose a block-recursive structure on the SVAR and for a given block-recursive structure we derive a conservative set of assumptions on the dependence and Gaussianity of the shocks to ensure identification. We use a Monte Carlo simulation to illustrate the advantages of the proposed blockrecursive estimator compared to unrestricted, purely data driven estimators in small samples. The block-recursive estimator is used to analyze the interdependence of monetary policy and the stock market. We find that a positive stock market shock contemporaneously increases the nominal interest rate, while contractionary monetary policy shocks lead to lower stock returns on impact.
A feasible approach to incorporate information in higher moments in structural vector autoregressions
(2021) Keweloh, Sascha Alexander
Generalized method of moments and continuous updating estimators based on second- to fourth-order moment conditions can be used to solve the identification problem and estimate non-Gaussian structural vectorautoregressions. However, estimating the asymptotically optimal weighting matrix and the asymptotic variance of the estimators is challenging in small samples. I show that this can lead to a severe bias, large variance, and inaccurate inference in small samples. I propose to use the assumption of independent structural shocks not only to derive moment conditions but also to derive alternative estimators for the asymptotically optimal weighting matrix and the asymptotic variance of the estimator. I demonstrate that these estimators greatly improve the performance of the generalized method of moments and continuous updating estimators in terms of bias, variance, and inference.
Das Klimaschutz-Sofortprogramm von Bündnis90/Die Grünen: Mögliche Auswirkungen auf Emissionen und Gesellschaft
(2021) Frondel, Manuel
In diesem Beitrag werden die Auswirkungen des Klimaschutz-Sofortprogramms der Partei Bündnis90/Die Grünen im Hinblick auf die gesellschaftlichen Verteilungswirkungen und die Potentiale zur Emissionsminderung bewertet. Aufgrund von Unklarheiten in der Ausgestaltung zahlreicher Maßnahmen ist es prinzipiell unmöglich, die damit in Summe einhergehenden Emissionsminderungen zu quantifizieren. Stattdessen wird sich hier auf diejenigen der großen Mannigfaltigkeit an Maßnahmen kapriziert, die im Programm hinreichend klar formuliert sind, um eine Bewertung zu erlauben, zumindest in qualitativer Hinsicht. Baerbock und Habeck (2021:2) kündigen unter anderem an, die erneuerbaren Stromerzeugungskapazitäten schneller ausbauen und den Kohleausstieg auf das Jahr 2030 vorziehen zu wollen. Diese nationalen Maßnahmen verursachen unnötig hohe Kosten. Es wäre kostengünstiger, diese dem Markt bzw. den steigenden Preisen für Emissionszertifikate zu überlassen. Lobenswert ist hingegen das Versprechen von Baerbock und Habeck (2021: 7), dass sie eine transatlantische Klimapartnerschaft zwischen der EU und den USA auf den Weg bringen möchten, da für eine effektive weltweite Klimapolitik internationale Kooperation unabdingbar ist. Für effektive Minderungen der globalen Emissionen ist ein solches bilaterales Bündnis allerdings zu wenig. Ein Bündnis zum Zwecke der effektiven und effizienten Senkung der Treibhausgasemissionen sollte deutlich umfassender sein und zumindest auf Ebene der G20-Staaten initiiert werden sowie ein Abkommen über die Etablierung eines einheitlichen CO2-Preises in diesen Ländern beinhalten. Allein einem möglichst umfassenden Klimaschutzabkommen über einen einheitlichen CO2-Preis trauen Experten die effektive Senkung der globalen Treibhausgasemissionen zu.
Statistical inference for function-on-function linear regression
(2021) Dette, Holger; Tang, Jiajun
We propose a reproducing kernel Hilbert space approach to estimate the slope in a function-on-function linear regression via penalised least squares, regularized by the thin-plate spline smoothness penalty. In contrast to most of the work on functional linear regression, our main focus is on statistical inference with respect to the sup-norm. This point of view is motivated by the fact that slope (surfaces) with rather different shapes may still be identified as similar when the difference is measured by an L2-type norm. However, in applications it is often desirable to use metrics reflecting the visualization of the objects in the statistical analysis. We prove the weak convergence of the slope surface estimator as a process in the space of all continuous functions. This allows us the construction of simultaneous confidence regions for the slope surface and simultaneous prediction bands. As a further consequence, we derive new tests for the hypothesis that the maximum deviation between the “true” slope surface and a given surface is less or equal than a given threshold. In other words: we are not trying to test for exact equality (because in many applications this hypothesis is hard to justify), but rather for pre-specified deviations under the null hypothesis. To ensure practicability, non-standard bootstrap procedures are developed addressing particular features that arise in these testing problems. As a by-product, we also derive several new results and statistical inference tools for the function-on-function linear regression model, such as minimax optimal convergence rates and likelihood-ratio tests. We also demonstrate that the new methods have good finite sample properties by means of a simulation study and illustrate their practicability by analyzing a data example.
Confidence surfaces for the mean of locally stationary functional time series
(2021) Dette, Holger; Wu, Weichi
The problem of constructing a simultaneous confidence band for the mean function of a locally stationary functional time series {Xi,n(t)}i=1,...n is challenging as these bands can not be built on classical limit theory. On the one hand, for a fixed argument t of the functions Xi,n, the maximum absolute deviation between an estimate and the time dependent regression function exhibits (after appropriate standardization) an extreme value behaviour with a Gumbel distribution in the limit. On the other hand, for stationary functional data, simultaneous confidence bands can be built on classical central theorems for Banach space valued random variables and the limit distribution of the maximum absolute deviation is given by the sup-norm of a Gaussian process. As both limit theorems have different rates of convergence, they are not compatible, and a weak convergence result, which could be used for the construction of a confidence surface in the locally stationary case, does not exist. In this paper we propose new bootstrap methodology to construct a simultaneous confidence band for the mean function of a locally stationary functional time series, which is motivated by a Gaussian approximation for the maximum absolute deviation. We prove the validity of our approach by asymptotic theory, demonstrate good finite sample properties by means of a simulation study and illustrate its applicability analyzing a data example.
Some practical aspects of sequential change point detection
(2021) Sivanesan, Sivanja; Dette, Holger; Ziggel, Daniel
In this report we investigate the finite sample properties of a new online monitoring scheme which was recently introduced by Gösmann et al. (2020) by means of a simulation study and a real data example. We also develop an algorithm which can be used in an active risk management. We start with an introduction in the basic notation and an explanation of the monitoring procedure, continue with an extensive simulation study to provide recommendations for the choice of several tuning parameters. Finally we present some illustration analyzing the Standard & Poor’s 500, MSCI World and MSCI Emerging Markets indices.
Dose response signal detection by parametric and least squares bootstrap
(2021) Bastian, Patrick; Dette, Holger; Kokot, Kevin; Bornkamp, Björn; Bretz, Frank
Wasserverbrauch privater Haushalte in Deutschland: Eine empirische Mikroanalyse
(2021) Frondel, Manuel; Niehues, Delia A.; Sommer, Stephan
Deutschland ist ein eher wasserreiches Land. Dennoch könnten es klimatische Veränderungen notwendig machen, künftig sorgsam mit der Ressource Wasser umzugehen, vor allem in Zeiten von Trockenheit. Vor diesem Hintergrund schätzt dieser Beitrag die Preiselastizität des Wasserverbrauchs privater Haushalte und differenziert dabei zwischen Haushalten, die eine grobe Kenntnis der Wasserpreise haben, und Haushalten ohne Preiskenntnis. Auf Basis von ca. 1.100 Beobachtungen für Haushalte, die in Einfamilienhäusern wohnen, und unter Verwendung der Summe der Kubikmeter- Preise für Wasser und Abwasser findet sich eine moderate, aber statistisch signifikant von Null verschiedene Preiselastizität von -0,102. Haushalte, die über die Kenntnis der Wasserpreise verfügen, weisen tendenziell eine höhere Elastizität auf, während Haushalte ohne Preiskenntnis keine statistisch signifikante Reaktion in ihrem Wasserverbrauch zeigen. Preise können demnach nur in begrenztem Umfang als Mittel zur Steuerung des Wasserverbrauchs eingesetzt werden.
Carbon pricing in Germany’s road transport and housing sector: Options for reimbursing carbon revenues
(2021) Frondel, Manuel; Schubert, Stefanie
In 2021, Germany launched a national emissions trading system (ETS) in its road transport and housing sectors that increases the cost burden of consumers of fossil fuels, the major source of carbon dioxide (CO2) emissions. A promising approach to secure public acceptance for such a carbon pricing would be to entirely reallocate the resulting “carbon” revenues to consumers. This article discusses three alternatives that were discussed in the political arena prior to the introduction of the national carbon pricing: a) a per-capita reallocation to private households, b) the reduction of electricity prices by, e.g., decreasing the electricity tax, as well as c) targeted financial aid for vulnerable consumers, such as increasing housing benefits. To estimate both the revenues originating from carbon pricing and the resulting emission savings, we employ a partial equilibrium approach that is based on price elasticity estimates on individual fossil fuel consumption from the empirical literature. Most effective with respect to alleviating the burden of poor households would be increasing housing benefits. While this measure would not require large monetary resources, we argue that the remaining revenues should be preferably employed to reduce Germany’s electricity tax, which becomes more and more obsolete given the steadily increasing amount of electricity generated by renewable energy technologies.

Browse

Recent Submissions