Bagging or Combining (or Both)? An Analysis Based on Forecasting U.S. Employment Growth

Forecasting a macroeconomic variable is challenging in an environment with many potential predictors whose predictive ability can vary over time. We compare two approaches to forecasting U.S. employment growth in this type of environment. The first approach applies bootstrap aggregating (bagging) to a general-to-specific procedure based on a general dynamic linear regression model with 30 potential predictors. The second approach considers several methods for combining forecasts from 30 individual autoregressive distributed lag (ARDL) models, where each individual ARDL model contains a potential predictor. We analyze bagging and combination forecasts at multiple horizons over four different out-of-sample periods using a mean square forecast error (MSFE) criterion and forecast encompassing tests. We find that bagging forecasts often deliver the lowest MSFE. Interestingly, we also find that incorporating information from both bagging and combination forecasts based on principal components often leads to further gains in forecast accuracy.

[1]  C. Granger,et al.  Handbook of Economic Forecasting , 2006 .

[2]  Norman R. Swanson,et al.  Chapter 5 Predictive Density Evaluation , 2006 .

[3]  A. Timmermann.  Chapter 4 Forecast Combinations , 2006 .

[4]  Stephen Figlewski,et al.  Optimal Price Forecasting Using Survey Data , 1983 .

[5]  J. Stock,et al.  Macroeconomic Forecasting Using Diffusion Indexes , 2002 .

[6]  Paul Newbold,et al.  Some comments on the evaluation of economic forecasts , 1973 .

[7]  F. Diebold,et al.  Comparing Predictive Accuracy , 1994, Business Cycles.

[8]  Norman R. Swanson,et al.  Predictive Density Evaluation , 2005 .

[9]  David E. Rapach,et al.  Forecasting US state-level employment growth: An amalgamation approach , 2012 .

[10]  D. Hendry,et al.  Econometric Evaluation of Linear Macro-Economic Models , 1986 .

[11]  David E. Rapach,et al.  Forecasting US employment growth using forecast combining methods , 2008 .

[12]  J. M. Bates,et al.  The Combination of Forecasts , 1969 .

[13]  Allan Timmermann,et al.  Persistence in forecasting performance and conditional combination strategies , 2006 .

[14]  Tae-Hwy Lee,et al.  To Combine Forecasts or to Combine Information? , 2010 .

[15]  L. Kilian,et al.  How Useful Is Bagging in Forecasting Economic Time Series? A Case Study of U.S. Consumer Price Inflation , 2008 .

[16]  J. Stock,et al.  Forecasting Output and Inflation: The Role of Asset Prices , 2001 .

[17]  J. Stock,et al.  A dynamic factor model framework for forecast combination , 1999 .

[18]  N. Swanson,et al.  Seeing Inside the Black Box: Using Diffusion Index Methodology to Construct Factor Proxies in Largescale Macroeconomic Time Series Environments , 2008 .

[19]  Yang Yang,et al.  Bagging binary and quantile predictors for time series , 2006 .

[20]  Serena Ng,et al.  Evaluating latent and observed factors in macroeconomics and finance , 2006 .

[21]  J. Stock,et al.  Combination forecasts of output growth in a seven-country data set , 2004 .

[22]  J. Bai,et al.  Determining the Number of Factors in Approximate Factor Models , 2000 .

[23]  Leo Breiman,et al.  Bagging Predictors , 1996, Machine Learning.

[24]  W. Newey,et al.  A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelationconsistent Covariance Matrix , 1986 .

[25]  Stephen Taylor,et al.  Forecasting Economic Time Series , 1979 .

[26]  Jonathan H. Wright,et al.  Forecasting Inflation , 2011 .

[27]  P. Newbold,et al.  Tests for Forecast Encompassing , 1998 .