Greg Tkacz

Neural network forecasting of Canadian GDP growth

Abstract The objective of this paper is to improve the accuracy of financial and monetary forecasts of Canadian output growth by using leading indicator neural network models. We find that neural networks yield statistically lower forecast errors for the year-over-year growth rate of real GDP relative to linear and univariate models. However, such forecast improvements are less notable when forecasting quarterly real GDP growth. Neural networks are unable to outperform a naive no-change model. More pronounced non-linearities at the longer horizon is consistent with the possible asymmetric effects of monetary policy on the real economy.

Estimating the Fractional Order of Integration of Interest Rates Using a Wavelet OLS Estimator

The debate on the order of integration of interest rates has long focused on the I(1) versus I(0) distinction. In this paper we instead use the wavelet OLS estimator of Jensen (1999) to estimate the fractional integration parameters of several interest rates for the United States and Canada from 1948 to 1999. We find that most rates are mean-reverting in the very long run, with the fractional order of integration increasing with the term to maturity. The speeds of mean reversion are lower in Canada, likely because of a positive country-specific risk premium. We also demonstrate that interest rate yield spreads involve noticeable persistence, indicating that these are also not strict I(0) processes. One consequence of these findings is that shocks to most interest rates and their spreads are very long lasting, yet not necessarily infinite.

TESTING FOR ASYMMETRY IN THE LINK BETWEEN THE YIELD SPREAD AND OUTPUT IN THE G-7 COUNTRIES

The difference in yields between long-term and short-term securities has been used both as a business cycle leading indicator and as an indicator of the current impact of monetary policy. This paper tests for an asymmetry, in the form of a threshold effect, such that the impact of the yield spread on output is greater on one side of the threshold than the other. The test allows for an unknown threshold, and the asymptotic distribution of the resulting statistic is obtained by the method of Hansen (1996). We test using data from each of the G-7 countries, and find that, while the yield spread does generally show a signifcant link with output, only in the U.S. and Canada is there strong evidence of an asymmetry of this type. The evidence of asymmetry that we find suggests a high value of the threshold in both the U.S. and Canada.

Endogenous thresholds and tests for asymmetry in US prime rate movements

Abstract This paper tests the hypothesis that movements in the prime rate are dependent upon the interest rate regime. We find that the responsiveness of prime to changes in the Federal Funds rate is relatively symmetric, but that the speed of adjustment of prime is faster when there is a substantial deviation from the equilibrium relationship linking these rates. These two results would confirm that lending markets are competitive in the US, and that monetary policy actions are readily transmitted to lending rates.

Forecast Content and Content Horizons for Some Important Macroeconomic Time Series

The pattern of decay of forecast content (or skill) with increasing horizon is well known for many types of meteorological forecasts; by contrast, little generally accepted information about these patterns or content horizons is available for economic variables. In this paper we estimate content horizons for a variety of macroeconomic quantities; more generally, we characterize the pattern of decay of forecast content as we project farther into the future. We find a wide variety of results for the different macroeconomic quantities, with models for some quantities providing useful content several years into the future, for other quantities providing negligible content beyond one or two months or quarters.

Inflation changes, yield spreads, and threshold effects

Using interest rate yield spreads to explain changes in inflation, we investigate whether such relationships can be modelled using two-regime threshold models.

Analyzing Economic Effects of September 11 and Other Extreme Events Using Debit and Payments System Data

Les décideurs politiques font parfois face à des événements extrêmes et imprévus qui peuvent influencer l’économie d’un pays et avoir des impacts négatifs par exemple sur l’emploi, les recettes fiscales et les cibles d’inflation. Dans certaines de ces situations, ils tentent de neutraliser ces effets en utilisant des outils monétaires ou budgétaires. Toutefois, pour établir la solution appropriée à adopter, il faut mesurer en temps réel la gravité des événements en question. Or, dans le cas d’événements momentanés, les données trimestrielles des bureaux de statistiques gouvernementaux sont peu utiles, puisqu’elles ne sont disponibles qu’après un décalage dans le temps par rapport aux événements, et que les effets de ceux-ci, de courte durée, peuvent être masqués par la nature cumulative des données dans le temps. Cependant, grâce à des progrès technologiques des dernières années, on a maintenant accès à des sources de données sur l’activité économique qui sont beaucoup plus précises et à jour. Ainsi, dans cet article, nous montrons qu’il aurait été possible d’utiliser les données portant sur le volume de transactions par carte de débit et sur le volume et la valeur des opérations de débit et par chèques effectuées les jours ouvrables pour analyser l’impact sur les dépenses des consommateurs canadiens des attaques terroristes du 11 septembre 2001, de l’épidémie de SRAS du printemps 2003 et de la grande panne d’électricité d’août 2003. Nos résultats montrent que, contrairement à ce qu’on a pu croire sur le moment, les effets de ces événements ont été de courte durée et peu importants.

Combining Forecasts with Nonparametric Kernel Regressions

We introduce a flexible nonparametric technique that can be used to select weights in a forecast-combining regression. We perform a Monte Carlo study that evaluates the performance of the proposed technique along with other linear and nonlinear forecast-combining procedures. The simulation results show that when forecast errors are correlated across models, the nonparametric weighting scheme dominates. As a general rule, our simulation results suggest that the practice of combining forecasts, no matter the technique employed in selecting the combination weights, can yield lower forecast errors on average. An application to inflation forecasting is also presented to demonstrate the use of all forecast-combining techniques.

Nowcasting GDP: Electronic Payments, Data Vintages and the Timing of Data Releases

We describe and assess the usefulness of a newly-constructed database of electronic payments, comprised of debit and credit card transactions as well cheques that clear through the banking system, as indicators of current GDP growth. Apart from capturing a broad range of spending activity, these variables are available on a very timely basis, thereby making them suitable candidate indicators. Controlling both for the release dates of various variables and the vintage of GDP available to analysts at the time a nowcast is produced, we generate nowcasts of GDP growth for a given quarter over a span of five months, which is the period over which interest in nowcasts would exist. We find that nowcast errors fall by about 60 per cent between the first and final nowcast. Evidence on the value of the additional payments variables is mixed, however; the point estimates suggest reductions in forecast loss at some nowcast horizons, but with considerable variability.

Using dynamic factor models to forecast Canadian inflation: the role of US variables-super-1

This article evaluates the forecasting performance of dynamic factor models for Canadian inflation. We find that factor models are as good as more conventional forecasting models, while a model estimated using only US data, is at least as useful as a model that incorporates Canadian data. This suggests that the United States is a source of data that could be beneficial to its trading partners for forecasting purposes.