Massimiliano Giuseppe Marcellino

Macroeconomic forecasting in the Euro area: Country specific versus area-wide information

Abstract This paper compares several time series methods for short-run forecasting of Euro-wide inflation and real activity using data from 1982 to 1997. Forecasts are constructed from univariate autoregressions, vector autoregressions, single equation models that include Euro-wide and US aggregates, and large-model methods in which forecasts are based on estimates of common dynamic factors. Aggregate Euro-wide forecasts are constructed from models that utilize only aggregate Euro-wide variables and by aggregating country-specific models. The results suggest that forecasts constructed by aggregating the country-specific models are more accurate than forecasts constructed using the aggregate data.

Some Cautions on the Use of Panel Methods for Integrated Series of Macroeconomic Data

We show how the use of panel data methods such as those proposed in single equations by Kao (1999) and Pedroni (1999) or in systems by Larsson and Lyhagen (1999) to investigate economic hypotheses such as purchasing power parity or the term structure of interest rates may be affected by the existence of cross-unit cointegrating relations. The existing literature assumes that such relations, that tie the units of the panel together, are not present. Using empirical examples from a panel of OECD countries we show that this assumption is very likely to be violated. Simulations of the properties of panel cointegration tests in the presence of cross-unit relations are then presented to demonstrate the serious cost of assuming away such relations. Some fixes are proposed as a way of dealing with these more general scenarios.

Testing for PPP: Should we use panel methods?

A common finding in the empirical literature on the validity of purchasing power parity (PPP) is that it holds when tested for in panel data, but not in univariate (i.e. country specific) analysis. The usual explanation for this mis-match is that panel tests for unit roots and cointegration are more powerful than their univariate counterparts. In this paper we suggest an alternative ex-planation for the mismatch. More generally, we warn against the use of panel methods for testing for unit roots in macroeconomic time series. Existing panel methods assume that cross-unit cointegrating or long-run relationships, that tie the units of the panel together, are not present. However, using empirical examples on PPP for a panel of OECD countries, we show that this assumption is very likely to be violated. Simulations of the properties of panel unit root tests in the presence of long-run cross-unit relationships are then presented to demonstrate the serious cost of assuming away such relationships. The empirical size of the tests is substantially higher than the nominal level, so that the null hypothesis of a unit root is rejected very often, even if correct.

Some Consequences of Temporal Aggregation in Empirical Analysis

I derive the generating mechanism of a temporally aggregated process when the disaggregated one belongs to the vector autoregressive integrated moving average class. I then study the effects of temporal aggregation on a set of characteristics of usual interest such as exogeneity, causality, cointegration, and common features. An empirical example with Canadian interest rates illustrates the main issues.

Dating the Euro Area Business Cycle

In this Paper we compare alternative approaches for dating the euro area business cycle and analysing its characteristics. First, we extend a commonly used dating procedure to allow for length, size and amplitude restrictions, and to compute the probability of a phase change. Second, we apply the modified algorithm for dating both the classical euro area cycle and the deviation cycle, where the latter is obtained by a variety of methods, including a modified HP filter that reproduces the features of the BK filter but avoids end-point problems, and a production function based approach. Third, we repeat the dating exercise for the main euro area countries, evaluate the degree of synchronization, and compare the results with the UK and the US. Fourth, we construct indices of business cycle diffusion, and assess how widespread are cyclical movements throughout the economy. Finally we repeat the dating exercise using monthly industrial production data, to evaluate whether the higher sampling frequency can compensate the higher variability of the series and produce a more accurate dating.

Leading Indicators for Euro-area Inflation and GDP Growth

In this Paper we evaluate the role of a set of variables as leading indicators for Euro-area inflation and GDP growth. Our evaluation is based on using the variables in the ECB euro area model database, plus a set of similar variables for the US. We compare the forecasting performance of each indicator with that of purely autoregressive models, using an evaluation procedure that is particularly relevant for policy-making. The evaluation is conducted both ex-post and in a pseudo real time context, for several forecast horizons, and using both recursive and rolling estimation. We also analyse three different approaches to combining the information from several indicators. First, we discuss the use as indicators of the estimated factors from a dynamic factor model for all the indicators. Second, an automated model selection procedure is applied to models with a large set of indicators. Third, we consider pooling the single indicator forecasts. The results indicate that single indicator forecasts are on average better than those derived from more complicated methods, but for them to beat the autoregression a different indicator has to be used in each period. A simple real-time procedure for indicator-selection produces good results.

Are there any reliable leading indicators for US Inflation and GDP Growth

In this paper we evaluate the relative merits of three approaches to information extraction from a large data set for forecasting, namely, the use of an automated model selection procedure, the adoption of a factor model, and single-indicator-based forecast pooling. The comparison is conducted using a large set of indicators for forecasting US inflation and GDP growth. We also compare our large set of leading indicators with purely autoregressive models, using an evaluation procedure that is particularly relevant for policy making. The evaluation is conducted both ex-post and in a pseudo real time context, for several forecast horizons, and using both recursive and rolling estimation. The results indicate a preference for simple forecasting tools, with a good relative performance of pure autoregressive models, and substantial instability in the leading characteristics of the indicators.

Fiscal Forecasting: the Track Record of the IMF, OECD and EC

We analyse the relative performance of the IMF, OECD and EC in forecasting the government deficit, as a ratio to GDP, for the G7 countries. Interesting differences across countries emerge, sometimes supporting the hypothesis of an asymmetric loss function (i.e. of a preference for underprediction or overprediction), and potential benefits from forecast pooling.

Factor-GMM estimation with large sets of possibly weak instruments

The use of factor analysis for instrumental variable estimation when the number of instruments tends to infinity is analysed. In particular, the focus is on situations where many weak instruments exist and/or the factor structure is weak. Theoretical results, simulation experiments and empirical applications highlight the relevance of Factor-GMM estimation, which is also easily implemented.

A Parametric Estimation Method for Dynamic Factor Models of Large Dimensions

The estimation of dynamic factor models for large sets of variables has attracted considerable attention recently, due to the increased availability of large datasets. In this paper we propose a new parametric methodology for estimating factors from large datasets based on state space models and discuss its theoretical properties. In particular, we show that it is possible to estimate consistently the factor space. We also develop a consistent information criterion for the determination of the number of factors to be included in the model. Finally, we conduct a set of simulation experiments that show that our approach compares well with existing alternatives.