Test
Can Awareness Reduce (and Reverse) Identity-driven Bias in Judgement? Evidence from International Cricket
Speaker – Prof Subhasish M. Chowdhury, Department of Economics, University of Sheffield, UK.
Date & Time – Wednesday, 23rd Aug 2023, 4pm, Venue – ERU seminar room, 6th Floor, Library Building, ISI Kolkata
Abstract: Competition is often judged by official decision makers, such as judges, juries, and referees. Systematic bias in those judgements, frequently related to social identities, may have undesirable effects. We investigate whether raising awareness can correct or even reverse such bias. We use a natural experiment from international Test cricket to focus on the match umpires and their decisions. Previous research has found evidence of biased judgements favouring the home team when the umpires shared the same nationality. Policy makers solved this by employing neutral country umpires. From June 2020, home umpires temporarily returned, sometimes in empty stadiums, because of the COVID-19 pandemic. We argue that these umpires were then under substantial scrutiny, due to the previous bias being well-known and highlighted in the media, alongside a technology-driven decision review system. Through a behavioural model, we show that such circumstances may result in the in-group judgement bias being eliminated or reversed. We find no evidence of the historical bias in umpire judgements returning during the pandemic. Instead, we find over-compensating behaviour, with a pre-pandemic home team advantage of 26% in the frequency of subjective and difficult ‘leg before wicket’ decisions being eliminated by the return of home umpires. Tight decisions tended to go against the home team more frequently when home umpires were officiating. We conclude that awareness not only has a long-term effect on eliminating identity-driven judgement bias but also may reverse it against the in-group.
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Title - Institutional Origin of Market Power and International Trade
Speaker – Dr. Joy Das, Lecturer, Cornell University.
Date & Time – Wednesday, 16-Aug-2023, 4pm, Venue – ERU seminar room, 6th Floor, Library Building, ISI Kolkata
Abstract: There are twofold purposes of this study: first, to propose a trade general equilibrium model in which the institution is modeled within the framework of international market concentration, and second, to empirically test our theoretical findings on the optimizing integration strategies by firms given the institution of a country. The study delves into the scope to which the institutional quality of a country affects the bilateral industry-level trade flows of manufacturing goods and services. Based on the interactive general equilibrium trade model of country-specific and industry-specific institutionally intensive variables, including the traditional control variables in bilateral trade, we analyze the six-digit NAICS classified industry-level bilateral trade flows from 220 countries and 389 industries for the year 1997. Corresponding to the Dixit-Stiglitz differentiated product solution, the theoretical portion of the study confirms that the new differentiated product solution is also a functional form of the institutions. The causal relationship is established by addressing the econometric challenges and endogeneity problems first with the Ordinary Least Squares (OLS) method of estimation and then with the Tobit model and Poisson Pseudo Maximum Likelihood (PPML) method of estimation with several robustness checks. Results indicate that countries with stronger institutions shift industries from highly concentrated markets with lower trade shares to lower concentrated markets with higher trade shares. Moreover, the institutions of the exporting countries are causal factors that manipulate the market structures for bilateral trade.
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Title: Revisiting Anselin et al. (1996): the Last Word on Spatial Testing.
Speaker: Prof. Anil K Bera, Department of Economics, University of Illinois at Champaign-Urbana.
Date & Venue: 2nd August 2023, Wednesday, ERU Lecture Theatre
Abstract: In spatial regression the dependence may arise from three kinds of interaction effects: (i) endogenous interaction effect; (ii) dependence through errors and (iii) exogenous interaction effect. Thus, a natural question would be that which of these three effects should be included in the model. To answer this question, we start with the most general model that encompasses all the above three types of dependencies, called the general nesting spatial (GNS) model. Anselin et al. (1996), Regional Science and Urban Economics, devised specification tests for a spatial model with only the first two kinds of interaction effects, i.e., (i) and (ii). Their tests had been revolutionary in the literature with more than 2,500 citations. However, in recent times, empirical researchers are in favor of adding the exogenous interaction effect (iii). Therefore, the need of the hour is to update Anselin et al. The GNS model, although appealing, remains largely untouched due to some possible identification problems.Thus, the aim of this paper is two-fold. First, to analytically explore the identification problem of the GNS model. Second, to develop simple specification tests for all the three types of spatial interactions. These and the Cox-type non-nested tests, that we construct subsequently, will lead to selecting the most appropriate spatial econometric model. The usefulness and practical applicability of our suggested tests will be demonstrated using extensive Monte Carlo studies and empirical illustrations.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
