Volume 224, Issue 5

This paper explores the existence of the credit channel in the transmission of monetary policy in Germany on the basis of a structural analysis of aggregate bank loan data. The empirical analysis is carried out in a vector error correction model (VECM), which allows to identify long-run cointegration relationships that can be interpreted as loan supply and loan demand equations. In this way, the fundamental identification problem inherent in reduced form approaches based on aggregate data is explicitly adressed. The short-run dynamics of the VECM is investigated by means of impulse response analysis, which sets out the impact of a monetary policy shock on the variables in the system. Empirical evidence consistent with the existence of a credit channel operating in Germany alongside the interest rate channel can be reported. Zusammenfassung Der Beitrag untersucht die Existenz des Kreditkanals in der geldpolitischen Transmission in Deutschland auf Basis einer strukturellen Analyse aggregierter Bankkreditdaten. Die empirische Untersuchung wird im Rahmen eines Vektorfehlerkorrekturmodells (VECM) durchgeführt, in dem die Identifizierung von langfristigen Angebots- und Nachfragebeziehungen im Markt für Bankkredite durch die Wahl von Restriktionen auf Kointegrationsvektoren erfolgt. Auf diese Weise wird das fundamentale Identifikationsproblem, das bei Schätzung einer reduzierten Form auftritt, explizit angegangen. Die Analyse der kurzfristigen Dynamik des VECM wird anhand von Impuls-Antwort-Funktionen durchgeführt, die den Einfluss geldpolitischer Schocks auf die Variablen des Systems darstellen. Die empirischen Ergebnisse deuten auf die Relevanz eines Kreditkanals in Deutschland, der zusätzlich zum Zinskanal wirkt, hin.

In this paper aggregate supply and demand functions for Germany’s imports and exports are estimated for the period 1975:1–1999:4. Previous studies are extended in two regards: First, the multivariate estimation technique of Johansen (1988, 1991) is employed, which helps to avoid the potential bias of the predominately used single-equation approaches. Second, structural error-correction models are estimated for the import and export market. This allows for conclusions about the short-run response of buyers and sellers in both markets to changes in their forcing variables. Zusammenfassung In dem vorliegenden Beitrag werden aggregierte Angebots- und Nachfragefunktionen für deutsche Im- und Exporte im Zeitraum 1975:1–1999:4 geschätzt. Vorgängerstudien werden in zwei Richtungen erweitert: Zum einen kommt das multivariate Schätzverfahren von Johansen (1988, 1991) zur Anwendung, wodurch der potenzielle Bias der häufig verwendeten Eingleichungsansätze vermieden wird. Zum anderen werden für die Im- und Exportmärkte strukturelle Fehlerkorrekturmodelle geschätzt. Dies ermöglicht eine Analyse der kurzfristigen Reaktionen von Anbietern und Nachfragern in beiden Märkten auf Änderungen ihrer Bestimmungsvariablen.

More and more investors react sceptically to stock picking recommendations from banks and newspapers. Often they try to carry out active portfolio management strategies which are based on forecasts of whole stock markets or stock indices. Those strategies are known as index picking or index market timing. So it is an amazing fact that periodically published stock index forecasts are a very rare thing. In Germany only the Centre for European Economic Research (ZEW) publishes stock index forecasts of the most important stock markets in Europe, the United States and Japan every month. This treatise brings into focus if ZEW stock index forecasts are useful for realizing active portfolio management strategies. The forecast data material analysis includes the Theil’s inequality coefficient (U 2 ), the topical orientated trend adjustment coefficient (GOVA coefficient) and the forecast quality matrix. The data analysis shows that ZEW stock market forecasts from October 1994 till December 2002 were not able to anticipate the future stock market development. Active portfolio management strategies get no support from ZEW stock market forecasts.

As a follow up of a previous paper (Strecker 2004) the vertical aggregation of individual random errors to totals for a population with Ν units is suggested by means of the Variate Difference Method. Because this method requires at least three (re-)enumerations which can hardly be realized in practice, the statistic response variability is introduced as a surrogate to account for the influence of random errors on survey totals. Using this statistic the confidence interval ot the random errors in survey totals can be calculated with data from two (re-)enumerations. The estimations of the response variability and of the confidence interval are illustrated with data from an official Belgian survey.

Adequacy (formerly: correspondence) generally concerns the correspondence of statement and facts. In economics we specially look for the statistical adequacy of variables and data (by means of a minimization of the so-called adequation discrepancy). Hereby we aim at empirical validity of hypotheses. This is extremely difficult in the case of a stationary statistical system on the one hand (see the System of National Accounts) and an evolving system of macroeconomic definitions on the other hand. The definitions of macro-economic variables with some importance within the growth and business cycle models are sometimes contradictory and insofar without any value for orientation.