Mr. Jean-Francois Dauphin, Mr. Kamil Dybczak, Morgan Maneely, Marzie Taheri Sanjani, Mrs. Nujin Suphaphiphat, Yifei Wang, and Hanqi Zhang
This paper describes recent work to strengthen nowcasting capacity at the IMF’s European department. It motivates and compiles datasets of standard and nontraditional variables, such as Google search and air quality. It applies standard dynamic factor models (DFMs) and several machine learning (ML) algorithms to nowcast GDP growth across a heterogenous group of European economies during normal and crisis times. Most of our methods significantly outperform the AR(1) benchmark model. Our DFMs tend to perform better during normal times while many of the ML methods we used performed strongly at identifying turning points. Our approach is easily applicable to other countries, subject to data availability.
We develop a framework to nowcast (and forecast) economic variables with machine learning techniques. We explain how machine learning methods can address common shortcomings of traditional OLS-based models and use several machine learning models to predict real output growth with lower forecast errors than traditional models. By combining multiple machine learning models into ensembles, we lower forecast errors even further. We also identify measures of variable importance to help improve the transparency of machine learning-based forecasts. Applying the framework to Turkey reduces forecast errors by at least 30 percent relative to traditional models. The framework also better predicts economic volatility, suggesting that machine learning techniques could be an important part of the macro forecasting toolkit of many countries.
This paper introduces methods that allow analysts to (i) decompose the estimates of unobserved quantities into observed data, (ii) to better understand revision properties of the model, and (iii) to impose subjective prior constraints on path estimates of unobserved shocks in structural economic models. For instance, a decomposition of the flexible-price output gap, or a technology shock, into contributions of output, inflation, interest rates, and other observed variables' contribution is feasible. The intuitive nature and analytical clarity of the suggested procedures are appealing for policy-related and forecasting models.
Jonas Dovern, Mr. Ulrich Fritsche, Mr. Prakash Loungani, and Ms. Natalia T. Tamirisa
We examine the behavior of forecasts for real GDP growth using a large panel of individual forecasts from 30 advanced and emerging economies during 1989–2010. Our main findings are as follows. First, our evidence does not support the validity of the sticky information model (Mankiw and Reis, 2002) for describing the dynamics of professional growth forecasts. Instead, the empirical evidence is more in line with implications of "noisy" information models (Woodford, 2002; Sims, 2003). Second, we find that information rigidities are more pronounced in emerging economies than advanced economies. Third, there is evidence of nonlinearities in forecast smoothing. It is less pronounced in the tails of the distribution of individual forecast revisions than in the central part of the distribution.
This paper surveys dynamic stochastic general equilibrium models with financial frictions in use by central banks and discusses priorities for future development of such models for the purpose of monetary and financial stability analysis. It highlights the need to develop macrofinancial models which allow analysis of the macroeconomic effects of macroprudential policy tools and to evaluate elements of the Basel III reforms as a priority. The paper also reviews the main approaches to introducing financial frictions into general equilibrium models.
The driving force of U.S. economic growth is expected to rotate from the fiscal stimulus and inventory rebuilding in 2009 to private demand in 2010, with consumption and particularly investment expected to be important contributors to growth. The strength of U.S. investment will hence be a crucial issue for the U.S. and global recovery. On the basis of several traditional models of investment, we forecast that the U.S. investment in equipment and software will grow by about 10 percent on average over the 2010-12 period. The contribution of investment to real GDP growth will be 0.8 percentage points on average over the same period.
This paper simulates out-of-sample inflation forecasting for Germany, the UK, and the US. In contrast to other studies, we use output gaps estimated with unrevised real-time GDP data. This exercise assumes an information set similar to that available to a policymaker at a given point in time since GDP data is subject to sometimes substantial revisions. In addition to using real-time datasets for the UK and the US, we employ a dataset for real-time German GDP data not used before. We find that Phillips curves based on ex post output gaps generally improve the accuracy of inflation forecasts compared to an AR(1) forecast but that real-time output gaps often do not help forecasting inflation. This raises the question how operationally useful certain output gap estimates are for forecasting inflation.
This paper attempts to explain short- and long-term dynamics of-and forecast-inflation in Tajikistan using the Vector Error Correction Model (VECM) and Autoregressive Moving Average Model (ARMA). By analyzing different transmission channels through the VECM, we were able to evaluate their relative dominance, magnitude, and speed of transition to the equilibrium price level, with the view of identifying those policy tools that will enhance the effectiveness of monetary policy. We found that excess supply of broad money is inflationary in both the short and long term. The dynamic analysis also demonstrates that the exchange rate and international inflation have a strong impact on local prices. Available monetary instruments, such as the refinancing rate, have proven to be ineffective. Therefore, the Tajik monetary authority could greatly benefit from enhancing its monetary instruments toolkit, including by developing the interest rate channel, to improve its monetary policy execution and to achieve stable inflationary conditions.
This paper contributes to the debate on the role of money in monetary policy by analyzing the information content of money in forecasting euro-area inflation. We compare the predictive performance within and among various classes of structural and empirical models in a consistent framework using Bayesian and other estimation techniques. We find that money contains relevant information for inflation in some model classes. Money-based New Keynesian DSGE models and VARs incorporating money perform better than their cashless counterparts. But there are also indications that the contribution of money has its limits. The marginal contribution of money to forecasting accuracy is often small, money adds little to dynamic factor models, and it worsens forecasting accuracy of partial equilibrium models. Finally, non-monetary models dominate monetary models in an all-out horserace.
The paper proposes an algorithm that uses forecast encompassing tests for combining forecasts. The algorithm excludes a forecast from the combination if it is encompassed by another forecast. To assess the usefulness of this approach, an extensive empirical analysis is undertaken using a U.S. macroecoomic data set. The results are encouraging as the algorithm forecasts outperform benchmark model forecasts, in a mean square error (MSE) sense, in a majority of cases.