C22 Time-Series Models; Dynamic Quantile Regressions (Updated!)
Refine
Year of publication
Document Type
- Working Paper (18)
- Article (1)
Has Fulltext
- yes (19)
Is part of the Bibliography
- no (19)
Keywords
- Exponential smoothing (3)
- GARCH-Prozess (3)
- Zeitreihenanalyse (3)
- ARCH-Prozess (2)
- Börsenkurs (2)
- Deutschland (2)
- Schätzfunktion (2)
- Volatilität (2)
- joint inference (2)
- loss function (2)
Wir verwenden eine neue, auf der Burr-Verteilung basierende Spezifikation aus der Familie der Autoregressive Conditional Duration (ACD) Modelle zur ökonometrischen Analyse der Transaktionsintensitäten während der Börseneinführung (IPO) der Deutsche Telekom Aktie. In diesem Fallbeispiel wird die Leistungsfähigkeit des neu entwickelten Burr-ACD-Modells mit den Standardmodellen von Engle und Russell verglichen, die im Burr-ACD Modell als Spezialfälle enthalten sind. Wir diskutieren außerdem alternative Möglichkeiten, Intra- Tagessaisonalitäten der Handelsintensität in ACD Modellen zu berücksichtigen.
Several recent studies have expressed concern that the Haar prior typically imposed in estimating sign-identi.ed VAR models may be unintentionally informative about the implied prior for the structural impulse responses. This question is indeed important, but we show that the tools that have been used in the literature to illustrate this potential problem are invalid. Speci.cally, we show that it does not make sense from a Bayesian point of view to characterize the impulse response prior based on the distribution of the impulse responses conditional on the maximum likelihood estimator of the reduced-form parameters, since the the prior does not, in general, depend on the data. We illustrate that this approach tends to produce highly misleading estimates of the impulse response priors. We formally derive the correct impulse response prior distribution and show that there is no evidence that typical sign-identi.ed VAR models estimated using conventional priors tend to imply unintentionally informative priors for the impulse response vector or that the corre- sponding posterior is dominated by the prior. Our evidence suggests that concerns about the Haar prior for the rotation matrix have been greatly overstated and that alternative estimation methods are not required in typical applications. Finally, we demonstrate that the alternative Bayesian approach to estimating sign-identi.ed VAR models proposed by Baumeister and Hamilton (2015) su¤ers from exactly the same conceptual shortcoming as the conventional approach. We illustrate that this alternative approach may imply highly economically implausible impulse response priors.
We propose a new framework for modelling time dependence in duration processes on financial markets. The well known autoregressive conditional duration (ACD) approach introduced by Engle and Russell (1998) will be extended in a way that allows the conditional expectation of the duration process to depend on an unobservable stochastic process, which is modelled via a Markov chain. The Markov switching ACD model (MSACD) is a very flexible tool for description and forecasting of financial duration processes. In addition the introduction of an unobservable, discrete valued regime variable can be justified in the light of recent market microstructure theories. In an empirical application we show, that the MSACD approach is able to capture several specific characteristics of inter trade durations while alternative ACD models fail. Furthermore, we use the MSACD to test implications of a sequential trade model.
We examine both the degree and the structural stability of inflation persistence at different quantiles of the conditional inflation distribution. Previous research focused exclusively on persistence at the conditional mean of the inflation rate. As economic theory provides reasons for inflation persistence to differ across conditional quantiles, this is a potentially severe constraint. Conventional studies of inflation persistence cannot identify changes in persistence at selected quantiles that leave persistence at the median of the distribution unchanged. Based on post-war US data we indeed find robust evidence for a structural break in persistence at all quantiles of the inflation process in the early 1980s. While prior to the 1980s inflation was not mean reverting, quantile autoregression based unit root tests suggest that since the end of the Volcker disinflation the unit root can be rejected at every quantile of the conditional inflation distribution.
This paper examines the relationship between oil movements and systemic risk of financial institution in major petroleum-based economies. We estimate ΔCoVaR for those institutions and observe the presence of elevated increases in its levels corresponding to the subprime and global financial crises. The results provide evidence in favor of risk measurement improvements by accounting for oil returns in the risk functions. The spread between the standard CoVaR and the CoVaR that includes oil absorbs in a time range longer than the duration of the oil shock. This indicates that the drop in the oil price has a longer effect on risk and requires more time to be discounted by the financial institutions. To support the analysis, we consider also the other major market-based systemic risk measures.
The use of GARCH models with stable Paretian innovations in financial modeling has been recently suggested in the literature. This class of processes is attractive because it allows for conditional skewness and leptokurtosis of financial returns without ruling out normality. This contribution illustrates their usefulness in predicting the downside risk of financial assets in the context of modeling foreign exchange-rates and demonstrates their superiority over use of normal or Student´s t GARCH models.
This study uses Markov-switching models to evaluate the informational content of the term structure as a predictor of recessions in eight OECD countries. The empirical results suggest that for all countries the term spread is sensibly modelled as a two-state regime-switching process. Moreover, our simple univariate model turns out to be a filter that transforms accurately term spread changes into turning point predictions. The term structure is confirmed to be a reliable recession indicator. However, the results of probit estimations show that the markov-switching filter does not significantly improve the forecasting ability of the spread.
This paper provides theory as well as empirical results for pre-averaging estimators of the daily quadratic variation of asset prices. We derive jump robust inference for pre-averaging estimators, corresponding feasible central limit theorems and an explicit test on serial dependence in microstructure noise. Using transaction data of different stocks traded at the NYSE, we analyze the estimators’ sensitivity to the choice of the pre-averaging bandwidth and suggest an optimal interval length. Moreover, we investigate the dependence of pre-averaging based inference on the sampling scheme, the sampling frequency, microstructure noise properties as well as the occurrence of jumps. As a result of a detailed empirical study we provide guidance for optimal implementation of pre-averaging estimators and discuss potential pitfalls in practice. Quadratic Variation , MarketMicrostructure Noise , Pre-averaging , Sampling Schemes , Jumps
This paper proposes tests for out-of-sample comparisons of interval forecasts based on parametric conditional quantile models. The tests rank the distance between actual and nominal conditional coverage with respect to the set of conditioning variables from all models, for a given loss function. We propose a pairwise test to compare two models for a single predictive interval. The set-up is then extended to a comparison across multiple models and/or intervals. The limiting distribution varies depending on whether models are strictly non-nested or overlapping. In the latter case, degeneracy may occur. We establish the asymptotic validity of wild bootstrap based critical values across all cases. An empirical application to Growth-at-Risk (GaR) uncovers situations in which a richer set of financial indicators are found to outperform a commonly-used benchmark model when predicting downside risk to economic activity.
Recently, the Bank of Japan outlined a “two perspectives” approach to the conduct of monetary policy that focuses on risks to price stability over different time horizons. Interpreting this as pertaining to different frequency bands, we use band spectrum regression to study the determination of inflation in Japan. We find that inflation is related to money growth and real output growth at low frequencies and the output gap at higher frequencies. Moreover, this relationship reflects Granger causality from money growth and the output gap to inflation in the relevant frequency bands. Keywords: spectral regression, frequency domain, Phillips curve, quantity theory. JEL Numbers: C22, E3, E5