Search results

Recently, it has been suggested that macroeconomic forecasts from estimated dynamic stochastic general equilibrium (DSGE) models tend to be more accurate out-of-sample than random walk forecasts or Bayesian vector autoregression (VAR) forecasts. Del Negro and Schorfheide (2013) in particular suggest that the DSGE model forecast should become the benchmark for forecasting horse-races. We compare the real-time forecasting accuracy of the Smets and Wouters (2007) DSGE model with that of several reduced-form time series models. We first demonstrate that none of the forecasting models is efficient. Our second finding is that there is no single best forecasting method. For example, typically simple AR models are most accurate at short horizons and DSGE models are most accurate at long horizons when forecasting output growth, while for inflation forecasts the results are reversed. Moreover, the relative accuracy of all models tends to evolve over time. Third, we show that there is no support to the common practice of using large-scale Bayesian VAR models as the forecast benchmark when evaluating DSGE models. Indeed, low-dimensional unrestricted AR and VAR forecasts may forecast more accurately.

Small-scale VARs are widely used in macroeconomics for forecasting US output, prices, and interest rates. However, recent work suggests these models may exhibit instabilities. As such, a variety of estimation or forecasting methods might be used to improve their forecast accuracy. These include using different observation windows for estimation, intercept correction, time-varying parameters, break dating, Bayesian shrinkage, model averaging, etc. This paper compares the effectiveness of such methods in real-time forecasting. We use forecasts from univariate time series models, the Survey of Professional Forecasters, and the Federal Reserve Board's Greenbook as benchmarks.

This article surveys recent developments in the evaluation of point and density forecasts in the context of forecasts made by vector autoregressions. Specific emphasis is placed on highlighting those parts of the existing literature that are applicable to direct multistep forecasts and those parts that are applicable to iterated multistep forecasts. This literature includes advancements in the evaluation of forecasts in population (based on true, unknown model coefficients) and the evaluation of forecasts in the finite sample (based on estimated model coefficients). The article then examines in Monte Carlo experiments the finite-sample properties of some tests of equal forecast accuracy, focusing on the comparison of VAR forecasts to AR forecasts. These experiments show the tests to behave as should be expected given the theory. For example, using critical values obtained by bootstrap methods, tests of equal accuracy in population have empirical size about equal to nominal size.

This chapter investigates the performance of different dimension reduction approaches for large vector autoregressions in multi-step ahead forecasts. The authors consider factor augmented VAR models using principal components and partial least squares, random subset regression, random projection, random compression, and estimation via LASSO and Bayesian VAR. The authors compare the accuracy of iterated and direct multi-step point and density forecasts. The comparison is based on macroeconomic and financial variables from the FRED-MD data base. Our findings suggest that random subspace methods and LASSO estimation deliver the most precise forecasts.

We develop a Bayesian approach for modeling brain activation and connectivity from functional magnetic resonance image (fMRI) data. Our approach simultaneously estimates local hemodynamic response functions (HRFs) and activation parameters, as well as global effective and functional connectivity parameters. Existing methods assume identical HRFs across brain regions, which may lead to erroneous conclusions in inferring activation and connectivity patterns. Our approach addresses this limitation by estimating region-specific HRFs. Additionally, it enables neuroscientists to compare effective connectivity networks for different experimental conditions. Furthermore, the use of spike and slab priors on the connectivity parameters allows us to directly select significant effective connectivities in a given network.

We include a simulation study that demonstrates that, compared to the standard generalized linear model (GLM) approach, our model generally has higher power and lower type I error and bias than the GLM approach, and it also has the ability to capture condition-specific connectivities. We applied our approach to a dataset from a stroke study and found different effective connectivity patterns for task and rest conditions in certain brain regions of interest (ROIs).

This study first calculates a profit rate for China’s economy over the period 1952–2014; the rate shows a downward trend in the long term but also exhibits cyclical fluctuations. Then, structural vector autoregressive models are used to examine the Chinese economic structure and, thanks to impulse response functions, the role of the profit rate in investment, capital accumulation, and GDP growth rates. Then, based on a priori constraints relative to this structure, the study tests whether these assumptions are verified over the period studied in the context of the transformations of China. The impulse response functions are further examined by using Bayesian analysis. Finally, the authors conclude that the period from 1952 to 2014 should be divided into several sub-periods with distinct structural characteristics.

After briefly reviewing the past history of Bayesian econometrics and Alan Greenspan's (2004) recent description of his use of Bayesian methods in managing policy-making risk, some of the issues and needs that he mentions are discussed and linked to past and present Bayesian econometric research. Then a review of some recent Bayesian econometric research and needs is presented. Finally, some thoughts are presented that relate to the future of Bayesian econometrics.

Over the past 25 years as a marketing academic, I have been fortunate to have collaborated with various researchers and firms and have contributed to the advancement of the marketing field. This is a review article that tracks my progress through these years that has led me to explore different areas of marketing, thereby shaping me as a researcher and an academic. As I see now, all of my research work can be viewed from a decision-making point of view – decisions that marketers can make either at the market, brand/firm/store, or the customer level. These decisions have in turn been transformed into strategies or tactics leading up to successful implementations and improved bottom-line results. The development of strategies/tactics and successful implementations can be seen in nearly 10 areas of research that I have involved myself in. This article also highlights how my research studies have contributed and advanced the creation of knowledge in each of these research areas.

There is a paradox in the normative foundations for chronic and intertemporal poverty measurement. Measures that reflect particular aversion to chronicity of poverty cannot also reflect particular aversion to fluctuations in the level of poverty when poverty is intense, yet good arguments are made in favour of each of these properties. I argue that the paradox may be explained if the poverty analyst implicitly predicts that an individual observed to experience persistent poverty will continue to experience poverty when unobserved. The paradox may then be resolved by separating the normative exercise of evaluation, applying a measure that reflects particular aversion to fluctuations, from a positive exercise of modelling and prediction. This proposal is illustrated by application to panel data from rural Ethiopia, covering the period 1994–2004. Several dynamic models are estimated, and a simple model with household-specific trends is found to give the best predictions of future wellbeing levels. Appropriately normalised measures of intertemporal poverty are applied to the predicted and observed trajectories of wellbeing, and results are found to differ substantially from naïve application of the measures to observed periods only. While similar results are obtained by naïve application of the measures that embody particular aversion to chronicity, separation of the normative and positive exercises maintains conceptual clarity.