... Forecaster herding has been extensively studied using theoretical models (see for example, Scharfstein and Stein, 1990;Bikhchandani and Hirshleifer, 1992;Ehrbeck and Waldmann, 1996;Teraji, 2003 , among others; for related research on preferences for conformism, see, e.g., Klick and Parisi, 2008 ), in experimental research ( Anderson and Holt, 1997;Drehmann and Oechssler, 2005;Morone and Sandri, 2009 ), and also in empirical research ( Chang and Cheng, 20 0 0;Clement and Tse, 2005;Bernhardt and Campello, 2006 among others; for related research on biases in how experts update beliefs, see Sinkey, 2015 , among others). 2 While theories of herding behavior have been extensively used in the field of financial economics to study forecasts of stock prices ( Chang and Cheng, 20 0 0;, metal prices ( Pierdzioch and Stadtmann, 2013 ), and exchange rates ( Pierdzioch and Stadtmann, 2012;Fritsche et al., 2015 ), herding behavior has also been studied in a variety of other fields including, for example, macroeconomic modeling ( Gaffeo and Canzian, 2011 ) and the modeling of the foreign direct investment decisions of firms ( Pinheiro-Alves, 2011 ). Forecaster herding arises if, for example, weak forecasters and forecasters in an early stage of their career remain close to the consensus forecast ( Lamont, 2002 ;see, however, Ashiya and Doi, 2001 ). ...