Asymmetric Loss and the Rationality of Inflation Forecasts: Evidence from South Korea

Abstract

We examine asymmetry in the loss functions of South Korean consumers' and the Bank of Korea's (BOK) inflation forecasts, and test the rationality of these forecasts under the assumption of a possible asymmetric loss function. Under an asymmetric loss function, we find evidence of asymmetry and support for rationality. We also examine whether the BOK's forecasts incorporate respective forecasts and consensus forecasts efficiently. They broadly use available information efficiently, and their results are robust to inflation-targeting measures and the recent global financial crisis. However, our results suggest that the information efficiency of the BOK's forecasts for consumers was affected during the period 2007-2008.

Ontario's Auction Market for Financial Transmission Rights: An Analysis of its Efficiency

“…Following , for example Keane and Runkle (1990), testing the informational efficiency hypothesis amounts to testing whether the additional variables do not explain any of the variation of the spot price and are therefore orthogonal to the error term in equation ( 7); that is, testing the null hypothesis:…”

The Rationality of USDA Forecasts under Multivariate Asymmetric Loss

“…Krüger and LeCrone (2019) show that this method has a high power and is robust to fat tails, serial correlation, and outliers. The method has been used to evaluate forecasts of a number of economic variables by professional forecasters (Aretz, Bartram, and Pope 2011;Pierdzioch, Rülke, and Stadtmann 2013; Mamatzakis and Koutsomanoli‐Filippaki 2014; Fritsche et al 2015; Pierdzioch, Reid, and Gupta 2016; Tsuchiya 2016a, 2016b; Christodoulakis 2020), government agencies (Auffhammer 2007; Krol 2013; Tsuchiya 2016a; Giovannelli and Pericoli 2020), international organizations (Christodoulakis and Mamatzakis 2008; Tsuchiya 2016a; Giovannelli and Pericoli 2020), and central banks (Capistrán 2008; Baghestani 2013; Pierdzioch, Rülke, and Stadtmann 2015; Ahn and Tsuchiya 2019; Caunedo et al 2020). These studies overwhelmingly suggest that forecasts that are biased or inefficient under MSE loss are rational under asymmetric loss.…”

Central bank forecasting: A survey

“…2 Simulations of the FRB/US model are used to construct "fan charts" that are published in the Tealbook as an indicator of the uncertainty surrounding the model-based forecasts Brayton et al (2014); Reifschneider and Tulip (2019), but these reflect model uncertainty rather than subjective uncertainty and are released with a long delay. 3 Similarly, recent evidence for central banks in Latin America and Asia have also pointed to asymmetric preferences, again finding a higher cost associated with positive inflation forecast errors (inflation above target) than negative ones (inflation below target) (Ahn & Tsuchiya, 2019;Pierdzioch et al, 2015). 4 While the FOMC does not provide such fan charts, the SEP includes qualitative assessments of uncertainty, discussed in Reifschneider and Tulip (2019).…”

“…33Similarly, recent evidence for central banks in Latin America and Asia have also pointed to asymmetric preferences, again finding a higher cost associated with positive inflation forecast errors (inflation above target) than negative ones (inflation below target) (Ahn & Tsuchiya, 2019; Pierdzioch et al., 2015). …”