An Evaluation of Early Warning Models for Systemic Banking Crises: Does Machine Learning Improve Predictions?
Johannes Beutel, Sophia List, Gregor von Schweinitz
Abstract
This paper compares the out-of-sample predictive performance of different early warning models for systemic banking crises using a sample of advanced economies covering the past 45 years. We compare a benchmark logit approach to several machine learning approaches recently proposed in the literature. We find that while machine learning methods often attain a very high in-sample fit, they are outperformed by the logit approach in recursive out-of-sample evaluations. This result is robust to the choice of performance measure, crisis definition, preference parameter, and sample length, as well as to using different sets of variables and data transformations. Thus, our paper suggests that further enhancements to machine learning early warning models are needed before they are able to offer a substantial value-added for predicting systemic banking crises. Conventional logit models appear to use the available information already fairly effciently, and would for instance have been able to predict the 2007/2008 financial crisis out-of-sample for many countries. In line with economic intuition, these models identify credit expansions, asset price booms and external imbalances as key predictors of systemic banking crises.
Read article
Inference in Structural Vector Autoregressions when the Identifying Assumptions are not Fully Believed: Re-evaluating the Role of Monetary Policy in Economic Fluctuations
Christiane Baumeister, James D. Hamilton
Journal of Monetary Economics,
2018
Abstract
Point estimates and error bands for SVARs that are set identified are only justified if the researcher is persuaded that some parameter values are a priori more plausible than others. When such prior information exists, traditional approaches can be generalized to allow for doubts about the identifying assumptions. We use information about both structural coefficients and impacts of shocks and propose a new asymmetric t-distribution for incorporating information about signs in a nondogmatic way. We apply these methods to a three-variable macroeconomic model and conclude that monetary policy shocks are not the major driver of output, inflation, or interest rates.
Read article
Who Benefits from GRW? Heterogeneous Employment Effects of Investment Subsidies in Saxony Anhalt
Eva Dettmann, Mirko Titze, Antje Weyh
IWH Discussion Papers,
No. 27,
2017
Abstract
The paper estimates the plant level employment effects of investment subsidies in one of the most strongly subsidized German Federal States. We analyze the treated plants as a whole, as well as the influence of heterogeneity in plant characteristics and the economic environment. Modifying the standard matching and difference-in-difference approach, we develop a new procedure that is particularly useful for the evaluation of funding programs with individual treatment phases within the funding period. Our data base combines treatment, employment and regional information from different sources. So, we can relate the absolute effects to the amount of the subsidy paid. The results suggest that investment subsidies have a positive influence on the employment development in absolute and standardized figures – with considerable effect heterogeneity.
Read article
College Choice and the Selection of Mechanisms: A Structural Empirical Analysis
J.-R. Carvalho, T. Magnac, Qizhou Xiong
Abstract
We use rich microeconomic data on performance and choices of students at college entry to study the interaction between the revelation of college preferences through exams and the selection of allocation mechanisms. We propose a method in which preferences and expectations of students are identified from data on choices and multiple exam grades. Counterfactuals we consider balance costs arising from congestion and exam organization. Moving to deferred acceptance or inverting the timing of choices and exams are shown to increase welfare. Redistribution among students or schools is sizeable in all counterfactual experiments.
Read article
Bottom-up or Direct? Forecasting German GDP in a Data-rich Environment
Katja Drechsel, Rolf Scheufele
Abstract
This paper presents a method to conduct early estimates of GDP growth in Germany. We employ MIDAS regressions to circumvent the mixed frequency problem and use pooling techniques to summarize efficiently the information content of the various indicators. More specifically, we investigate whether it is better to disaggregate GDP (either via total value added of each sector or by the expenditure side) or whether a direct approach is more appropriate when it comes to forecasting GDP growth. Our approach combines a large set of monthly and quarterly coincident and leading indicators and takes into account the respective publication delay. In a simulated out-of-sample experiment we evaluate the different modelling strategies conditional on the given state of information and depending on the model averaging technique. The proposed approach is computationally simple and can be easily implemented as a nowcasting tool. Finally, this method also allows retracing the driving forces of the forecast and hence enables the interpretability of the forecast outcome.
Read article
Qual VAR Revisited: Good Forecast, Bad Story
Makram El-Shagi, Gregor von Schweinitz
Abstract
Due to the recent financial crisis, the interest in econometric models that allow to incorporate binary variables (such as the occurrence of a crisis) experienced a huge surge. This paper evaluates the performance of the Qual VAR, i.e. a VAR model including a latent variable that governs the behavior of an observable binary variable. While we find that the Qual VAR performs reasonably well in forecasting (outperforming a probit benchmark), there are substantial identification problems. Therefore, when the economic interpretation of the dynamic behavior of the latent variable and the chain of causality matter, the Qual VAR is inadvisable.
Read article
Bottom-up or Direct? Forecasting German GDP in a Data-rich Environment
Katja Drechsel, Rolf Scheufele
Abstract
This paper presents a method to conduct early estimates of GDP growth in Germany. We employ MIDAS regressions to circumvent the mixed frequency problem and use pooling techniques to summarize efficiently the information content of the various indicators. More specifically, we investigate whether it is better to disaggregate GDP (either via total value added of each sector or by the expenditure side) or whether a direct approach is more appropriate when it comes to forecasting GDP growth. Our approach combines a large set of monthly and quarterly coincident and leading indicators and takes into account the respective publication delay.
Read article
The Performance of Short-term Forecasts of the German Economy before and during the 2008/2009 Recession
Katja Drechsel, Rolf Scheufele
International Journal of Forecasting,
No. 2,
2012
Abstract
The paper analyzes the forecasting performance of leading indicators for industrial production in Germany. We focus on single and pooled leading indicator models both before and during the financial crisis. Pairwise and joint significant tests are used to evaluate single indicator models as well as forecast combination methods. In addition, we investigate the stability of forecasting models during the most recent financial crisis.
Read article
Auswirkungen der aus dem Konjunkturpaket II für das Zentrale Innovationsprogramm Mittelstand (ZIM) bereitgestellten Mittel auf die konjunkturelle Entwicklung. Gutachten im Auftrag des Bundesministerium für Wirtschaft und Technologie (BMWi)
Jutta Günther, Udo Ludwig, Hans-Ulrich Brautzsch, Brigitte Loose, Nicole Nulsch
One-off Publications,
2011
Abstract
The ZIM program (Zentrales Innovationsprogramm Mittelstand) is a technologically open program of the Federal Ministry of Economics and Technology to support small and medium enterprises and Science organizations in their research and innovation activities. It became operative July 1, 2008 and offers three program lines: individual projects, cooperative projects, and networks. In reaction to the global economic crisis the ZIM program was increased for the years 2009 and 2010 – in addition to the regulary scheduled 626 Million – by 900 Million Euro through the Konjunkturpaket II (KP II).
In this study, the analysis of the macroeconomic effects of the ZIM program in Germany has been carried out – first time in the evaluation of federal support programs for research and innovation – by the use of the input output method.
The pdf file includes an english summary with details about the study's results.
Read article
Should We Trust in Leading Indicators? Evidence from the Recent Recession
Katja Drechsel, Rolf Scheufele
Abstract
The paper analyzes leading indicators for GDP and industrial production in Germany. We focus on the performance of single and pooled leading indicators during the pre-crisis and crisis period using various weighting schemes. Pairwise and joint significant tests are used to evaluate single indicator as well as forecast combination methods. In addition, we use an end-of-sample instability test to investigate the stability of forecasting models during the recent financial crisis. We find in general that only a small number of single indicator models were performing well before the crisis. Pooling can substantially increase the reliability of leading indicator forecasts. During the crisis the relative performance of many leading indicator models increased. At short horizons, survey indicators perform best, while at longer horizons financial indicators, such as term spreads and risk spreads, improve relative to the benchmark.
Read article