Testing for Structural Breaks at Unknown Time: A Steeplechase
Makram El-Shagi, Sebastian Giesen
Computational Economics,
No. 1,
2013
Abstract
This paper analyzes the role of common data problems when identifying structural breaks in small samples. Most notably, we survey small sample properties of the most commonly applied endogenous break tests developed by Brown et al. (J R Stat Soc B 37:149–163, 1975) and Zeileis (Stat Pap 45(1):123–131, 2004), Nyblom (J Am Stat Assoc 84(405):223–230, 1989) and Hansen (J Policy Model 14(4):517–533, 1992), and Andrews et al. (J Econ 70(1):9–38, 1996). Power and size properties are derived using Monte Carlo simulations. We find that the Nyblom test is on par with the commonly used F type tests in a small sample in terms of power. While the Nyblom test’s power decreases if the structural break occurs close to the margin of the sample, it proves far more robust to nonnormal distributions of the error term that are found to matter strongly in small samples although being irrelevant asymptotically for all tests that are analyzed in this paper.
Read article
Distance Functions for Matching in Small Samples
Eva Dettmann, Christian Schmeißer, Claudia Becker
Computational Statistics & Data Analysis,
No. 5,
2011
Abstract
The development of ‘standards’ for the application of matching algorithms in empirical evaluation studies is still an outstanding goal. The first step of the matching procedure is the choice of an appropriate distance function. In empirical evaluation situations often the sample sizes are small. Moreover, they consist of variables with different scale levels which have to be considered explicitly in the matching process. A simulation is performed which is directed towards these empirical challenges and supplements former studies in this respect. The choice of the analysed distance functions is determined by the results of former theoretical studies and recommendations in the empirical literature. Thus, two balancing scores (the propensity score and the index score) and the Mahalanobis distance are considered. Additionally, aggregated statistical distance functions not yet used for empirical evaluation are included. The matching outcomes are compared using non-parametric scale-specific tests for identical distributions of the characteristics in the treatment and the control groups. The simulation results show that, in small samples, aggregated statistical distance functions are the better choice for summarising similarities in differently scaled variables compared to the commonly used measures.
Read article
Has the Euro Increased International Price Elasticities?
Oliver Holtemöller, Götz Zeddies
IWH Discussion Papers,
No. 18,
2010
published in: Empirica
Abstract
This paper analyzes the role of common data problems when identifying structural breaks in small samples. Most notably, we survey small sample properties of the most commonly applied endogenous break tests developed by Brown, Durbin, and Evans (1975) and Zeileis (2004), Nyblom (1989) and Hansen (1992), and Andrews, Lee, and Ploberger (1996). Power and size properties are derived using Monte Carlo simulations. Results emphasize that mostly the CUSUM type tests are affected by the presence of heteroscedasticity, whereas the individual parameter Nyblom test and AvgLM test are proved to be highly robust. However, each test is significantly affected by leptokurtosis. Contrarily to other tests, where skewness is far more problematic than kurtosis, it has no additional effect for any of the endogenous break tests we analyze. Concerning overall robustness the Nyblom test performs best, while being almost on par to more recently developed tests in terms of power.
Read article
Testing for Structural Breaks at Unknown Time: A Steeplechase
Makram El-Shagi, Sebastian Giesen
Abstract
This paper analyzes the role of common data problems when identifying structural breaks in small samples. Most notably, we survey small sample properties of the most commonly applied endogenous break tests developed by Brown, Durbin, and Evans (1975) and Zeileis (2004), Nyblom (1989) and Hansen (1992), and Andrews, Lee, and Ploberger (1996). Power and size properties are derived using Monte Carlo simulations. Results emphasize that mostly the CUSUM type tests are affected by the presence of heteroscedasticity, whereas the individual parameter Nyblom test and AvgLM test are proved to be highly robust. However, each test is significantly affected by leptokurtosis. Contrarily to other tests, where skewness is far more problematic than kurtosis, it has no additional effect for any of the endogenous break tests we analyze. Concerning overall robustness the Nyblom test performs best, while being almost on par to more recently developed tests in terms of power.
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
A First Look on the New Halle Economic Projection Model
Sebastian Giesen, Oliver Holtemöller, Juliane Scharff, Rolf Scheufele
Abstract
In this paper we develop a small open economy model explaining the joint determination of output, inflation, interest rates, unemployment and the exchange rate in a multi-country framework. Our model – the Halle Economic Projection Model (HEPM) – is closely related to studies recently published by the International
Monetary Fund (global projection model). Our main contribution is that we model the Euro area countries separately. In this version we consider Germany and France, which represent together about 50 percent of Euro area GDP. The model allows for country specific heterogeneity in the sense that we capture different adjustment patterns to economic shocks. The model is estimated using Bayesian techniques. Out-of-sample and pseudo out-of-sample forecasts are presented.
Read article
Is there a Superior Distance Function for Matching in Small Samples?
Eva Dettmann, Claudia Becker, Christian Schmeißer
Abstract
The study contributes to the development of ’standards’ for the application of matching algorithms in empirical evaluation studies. The focus is on the first step of the matching procedure, the choice of an appropriate distance function. Supplementary o most former studies, the simulation is strongly based on empirical evaluation ituations. This reality orientation induces the focus on small samples. Furthermore, ariables with different scale levels must be considered explicitly in the matching rocess. The choice of the analysed distance functions is determined by the results of former theoretical studies and recommendations in the empirical literature. Thus, in the simulation, two balancing scores (the propensity score and the index score) and the Mahalanobis distance are considered. Additionally, aggregated statistical distance functions not yet used for empirical evaluation are included. The matching outcomes are compared using non-parametrical scale-specific tests for identical distributions of the characteristics in the treatment and the control groups. The simulation results show that, in small samples, aggregated statistical distance functions are the better
choice for summarising similarities in differently scaled variables compared to the
commonly used measures.
Read article
Threshold for employment and unemployment. A spatial analysis of German RLM's 1992-2000
Christian Dreger, Reinhold Kosfeld
External Publications,
2006
Abstract
Changes in production and employment are closely related over the course of the business cycle. However, as exemplified by the laws of Verdoorn (1949, 1993) and Okun (1962, 1970), thresholds seem to be present in the relationship. Due to capacity reserves of the firms, output growth must exceed certain levels for the creation of new jobs or a fall in the unemployment rate. While Verdoorn's law focuses on the growth rate of output sufficient for an increase in employment, in Okun's law, the fall in the unemployment rate becomes the focus of attention. In order to assess the future development of employment and unemployment, these thresholds have to be taken into account. They serve as important guidelines for policymakers. In contrast to previous studies, we present joint estimates for both the employment and unemployment threshold. Due to demographic patterns and institutional settings on the labour market, the two thresholds can differ, implying that minimum output growth needed for a rise in employment may not be sufficient for a simultaneous drop in the unemployment rate. Second, regional information is considered to a large extent. In particular, the analysis is carried out using a sample of 180 German regional labour markets, see Eckey (2001). Since the cross-sections are separated by the flows of job commuters, they correspond to travel-to-work areas. Labour mobility is high within a market, but low among the entities. As the sectoral decomposition of economic activities varies across the regions, the thresholds are founded on a heterogeneous experience, leading to more reliable estimates.The contribution to the literature is twofold. First, to the best of our knowledge, no previous paper has investigated a similar broad regional dataset for the German economy as a whole before. By using a panel dataset, information on the regional distributions around the regression lines as well as theirs positional changes is provided for each year. Second, the methods applied are of new type. They involve a mixture of pooled and spatial econometric techniques. Dependencies across the regions may result from common or idiosyncratic (region specific) shocks. In particular, the eigenfunction decomposition approach suggested by Griffith (1996, 2000) is used to identify spatial and non-spatial components in regression analysis. As the spatial pattern may vary over time, inference is conducted on the base of a spatial SUR model. Due to this setting, efficient estimates of the thresholds are obtained. With the aid of a geographic information system (GIS) variation of the spatial components can be made transparent. With Verdoorn’s and Okun’s law the figures show some significant patterns become obvious over time. In respect to Verdoorn’s law, for instance, a stripe of high values in the north-western part from Schleswig-Holstein via Lower Saxony and North Rhine Westfalia to Rhineland Palatinate is striking in all years but 1994 and 1995. In most periods the spatial component is likewise concentrated in Saxony. Clusters of low values can be found in northern Bavaria and, in some periods, in Thüringen and Mecklenburg-Vorpommern. Other parts of Germany appear to be more fragmented consisting of relative small clusters of low, medium and high values of the spatial component. With Okun’s law some changing spatial patterns arise. In all, spatially filtering provides valuable insights into the spatial dimensions of the laws of Verdoorn and Okun.
Read article
Mikroökonometrische Evaluation der ökonomischen Wirkungen betrieblicher Mitbestimmung - Möglichkeiten und Grenzen des Matching-Ansatzes
Birgit Schultz
Jahrbücher für Nationalökonomie und Statistik,
No. 5,
2006
Abstract
The evaluation of economic effects of workers participation is not simple from the methodical point of view because of specific characteristics of establishments with works councils. Especially recent studies show contradictory results. In this study problems are pointed out, discussed, and options for solution are presented on the example of workers participation in East German establishments of industry and construction by the IAB-Establishment Panel.
An optimal matching-algorithm which supplies good matching-results for small samples to assign 'statistical establishment-twins' is applied. But by reason of very short primarily spells it can only calculate short time effects. Therefore, the matching method is additionally used to construct longer observation periods. By this new application establishments with recently founded works councils are matched with so called 'proxy establishments' with existing works councils. As a result short observation periods are prolonged and information about long-term effects can be given.
The effects on productivity, profitability and qualification level of employees show neither in short-term nor in long-term a significant impact on workers' participation.
Read article
Options and Limits of the Matching Approach – An Application to Workers Participation
Birgit Schultz
IWH Discussion Papers,
No. 15,
2006
Abstract
The evaluation of economic effects of workers participation is not simple from the methodical point of view because of specific characteristics of establishments with works councils. Especially recent studies show contradictory results. In this study problems are pointed out, discussed, and options for solution are presented on the example of workers participation in East German establishments of industry and construction by the IABEstablishment Panel. An optimal matching-algorithm which supplies good matching-results for small samples to assign ‘statistical establishment-twins’ is applied. But by reason of very short primarily spells it can only calculate short time effects. Therefore, the matching method is additionally used to construct longer observation periods. By this new application establishments with recently founded works councils are matched with so called ‘proxy establishments’ with existing works councils. As a result short observation periods are prolonged and information about long-term effects can be given. The effects on productivity, profitability and qualification level of employees show neither in short-term nor in long-term a significant impact on workers’ participation.
Read article