Introduction

Economic forecasting is a well-recognized field of study in economists. However, the ability of economists to make forecast has often been questioned. Recently, Lahiry (2011) summarizes this idea, saying that ‘There is skepticism not only by laymen but by most academic economists regarding the true capability of macroeconomic forecasters. The conventional wisdom is that economic forecasters are mere charlatans.’ Despite such negative opinions residing in part of the professional and academic domains of economics, businesses and policymakers need predictions, especially when the world economy is experiencing the Great Recession. In this sense, Kennedy (2003, p. 364) asserted that ‘even if forecasts are poor, there are none better, and perhaps a poor forecast is better than none at all.’ The accuracy of economic forecasting depends on many factors, such as the quality of the data, the adequacy of the chosen method, and computational limitations. When it comes to predictions for (small) geographical/georeferenced entities, the scarcity of disaggregated data has limited the development of methods and the consequent interest in this particular empirical exercise for a long time. In contrast, forecasts at the regional/local level are a necessary complement to forecasts at the national level, allowing, for example, policymakers to allocate public expenditure between regions more efficiently. This idea reinforces the need for econometric techniques for obtaining local predictions, and in particular those where the information on each spatial unit is considered together with the one on its neighbouring units, as the latter ones are most likely to influence the former’s socioeconomic outcomes with their policies due to spatial proximity (e.g., think of interregional commuting) (see, e.g., Baltagi et al., 2012, 2014). Moreover, emerging empirical evidence suggests that, depending on the data set-up available (e.g., in terms of the cross-sectional and time dimensions) different methodological approaches may be desirable (Mayor and Patuelli, 2012, 2014). The aim of this special issue is to provide an outlet for some recent advances in regional forecasting, showing, at the same time, future research directions. Both methodological and applied contributions have been selected, varying methodologically from traditional, timeseries-based panel econometrics to spatial econometrics. The special issue contains nine