Behavioural Economics & Experimental Methods Dr Anwar Shah School of Economics, QAU Introduction of Experimental Economics Main Books: 1.Markets, Games, and Strategic Behavior By Charles A. Holt 2.Nudge: Improving decisions about health, wealth and Happiness by Richard H. Thaler and Cass R. Sunstein (Published by Yale University Press) 3.Experimental Methods: A Primer for Economists by Daniel Friedman and Shyam Sunder (1994) Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 2 What is “Experimental Economics” • Experimental Economics is a branch of Economics that uses controlled experiments – To evaluate Theories and Behavioral assumptions – To test policies and their implementation. • Controlled experiments are a method of empirical investigation Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 3 Why use experiments? • 1. To make inferences about theory, which otherwise is hard to examine. – Most theories cannot be tested directly with field (happenstance) data such as whether people are rationale, whether invisible hand theory holds, whether people free ride etc., • 2. To examine and study topics for which no information is available. – Do people prefer equality over efficiency? – Do people compare themselves with others? Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 4 Experimental Methodology • What is experimental methodology and how controlled experiments are used in economics • How controlled experiments offer an alternative way of analyzing economic theories and problems Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 5 Why Data Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 6 Types of Data Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 7 Types of Data • Happenstance data: the by-product of uncontrolled naturally occurring economic activity • Experimental data: Generated for scientific purposes in a controlled environment Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 8 Types of Data • Solely on the basis of field data, it is difficult to decide whether and when a theory fails, and to pinpoint the aspects responsible for this failure”. - Bank of Sweden (2002) • Nevertheless, laboratory experiments have recently been accepted by mainstream economics… Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 9 Types of Data and Experiments Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 10 Types of Data and Experiments Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 11 Data and Experiments • Experimental economics uses controlled experiments (primarily in the laboratory) • Laboratory experiments is a (“new”) method for studying human behavior, theories, institutions in a simplified environment that mimics situations that are encountered in markets and other economic interactions. – New since Chamberlin (1948) – Simplified to ensure control and causality • Experiments are also conducted in the field, over the internet and in brain scanners. Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 12 Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 13 Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 14 Advantages and Disadvantages • Advantages: – Easy to interpret due to control – Focus on the topics of interest – Sometimes can be a way to save money! (e.g. policy experiments, auction experiments) • Disadvantages – Cost – External Validity • Happenstance field data has advantages and disadvantages, reversed Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 15 Advantages: Control • Experimenter can distinguish between exogenous and endogenous variables • Experimenter controls information conditions • Experimenter controls (to some extent) incentives by paying money • There are few unobservable variables – Causality is not a problem as experimenter can implement ceteris paribus condition – As a result, sophisticated econometrics are not necessary and fewer observations are required • Experimenter controls the conditions under which data is generated Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 16 Advantages: Control • A case of how lack of control can lead to misleading results • Question: • Do employment and training programs increase (mean) annual earnings of participants? • Randomized field experiment: Participants and nonparticipants chosen randomly • This ensures experimental control by avoiding selection bias (assuming sample is large enough) • By using simple statistics and comparing the two groups La Londe (1986, AER) showed that the course increased earnings by $900 on average. Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 17 Advantages: Control • Then, he ignored control group and assumed that he had happenstance field data. • He used various econometric specifications to estimate impact of course on earnings • The estimates varied considerably and in some cases had the opposite sign! • See also Cox and Oaxaca (1991) Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 18 Purpose of Experiments • Scientific purposes – – – – Study theoretical predictions Study underlying assumptions Compare competing theories Offer advice to theory (e.g. which of the different equilibria predicted by theory will occur) – Show the way forward to theory (e.g. does gender matter as all economic theories are gender neutral? (highlight the importance of institutions such as markets and auctions?) – Measure individual characteristics (e.g. risk attitudes, fairness attitudes) Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 19 Purpose of Experiments • Other purposes – To influence decision/policy making – To understand behavior of consumers, voters and managers. – Educational purposes (learning objectives) Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 20 Difference from Psychology • There is partial overlap of research interest as both are social sciences, but both have some different interests: • economists care for economically relevant situations. • Methodological differences: – – – – – – economists despise deception (lying) economists dislike hypothetical questions economists reward each decision (salience) economists use written instructions economists use repeated trials and less one-shot trials different relation to existing theory • But experimental economists are learning from psychologists and vice versa. Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 21 Some Objections • Experiments are unrealistic – So are economic models! • Simplicity is often a virtue • Whether realism is important depends on the purpose of the experiment: • If the purpose of study is to test a theory • Then the evidence is important for theory building but not for a direct understanding of reality. Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 22 Some objections • Experiments are artificial, because of – subject pool bias (students) – low stakes – Small number of participants – Inexperienced subjects Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 23 Answers of the Objections • Most criticisms can actually be answered by conducting more experiments! • Use other subjects (e.g. Fehr et al., JLE 1998: soldiers; Cooper et al. AER 1999: managers; Fehr/List, 2006, JEEA: CEOs) • Increase the stake level (e.g. Holt/Laury, AER 2002; Cameron, JRU 1999) • Increase the number of participants, (e.g . Isaac and Walker J Pub E 1990; and Walker, J.Pub.E 1990; Bellemare/Kröger2003) • Recruit experienced participants (e.g. Kagel/Levin, AER 1986) Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 24 Summary • Experimental economics is a relatively new field in economics, which use controlled laboratory experiments for the purpose of scientific knowledge. • Experiments economics provides new tool for the collection of data in social sciences, which otherwise was difficult to collect. Such collection of data help us to test theories, assumptions of theories and suggest polices. Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 25 Thanks & Allah Hafiz Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 26 Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 27 Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 28 Appendix Behavioural Economics & Experimental Methods (Dr Anwar Shah, QAU) 29