Econ 440: Lecture 7 - The Endowment Effect PDF

Econ 440: Lecture 7 Prof. Ragan Petrie Texas A&M University September 17, 2024 The Endowment Effect The Endowment Effect  The Endowment Effect Motivating Example ▶ Professor A likes to collect wine ▶ He will not pay more than $35 for a bottle ▶ Once he gets a...

Econ 440: Lecture 7 Prof. Ragan Petrie Texas A&M University September 17, 2024 The Endowment Effect The Endowment Effect  The Endowment Effect Motivating Example ▶ Professor A likes to collect wine ▶ He will not pay more than $35 for a bottle ▶ Once he gets a bottle, he will not sell it for less than $100 ▶ So, for a price between $35 and $100, will neither sell nor buy ▶ What is going on? ▶ Once he obtains the bottle, its value goes up ▶ This phenomenon is the endowment effect: owning an item changes its apparent value to the owner ▶ Explanation: prospect theory ▶ Before owning the bottle, reference point is not having it ▶ Once he owns the bottle, having it becomes the reference point ▶ Because of loss aversion, losing a bottle hurts more than gaining one feels good  The Endowment Effect The Classic Experiment ▶ Undergraduates participate sequentially in 11 markets for goods like pens and mugs ▶ Half of subjects randomly endowed with item ▶ Value of item for buyer and sellers gathered through price lists ▶ Buyers express willingness to pay (WTP) for items ▶ Sellers express willingness to accept (WTA) to give up items ▶ Items transacted at market prices Source: Kahneman, Knetsch, and Thaler (1991)  The Endowment Effect Results ▶ What relationship should we expect between WTP and WTA? ▶ Results ▶ Median WTA (seller stated price) 2-3 times higher than median WTP (buyer stated price) ▶ Interpretation: Sellers’ stated price is higher because they are endowed with the items, and this increases their value ▶ We call this the willingness to pay vs willingness to accept gap Source: Kahneman, Knetsch, and Thaler (1991)  The Endowment Effect Loss aversion ▶ Loss aversion is one of the most well-supported theories from the field ▶ We already saw how it could explain behavior of cab drivers ▶ We will see two more famous examples today ▶ One additional real-world example, this time of professional golfers ▶ A laboratory experiment where reference points come from expectations  The Endowment Effect Is Tiger Woods Loss Averse? Is Tiger Woods Loss Averse?  The Endowment Effect Is Tiger Woods Loss Averse? Why Golf? ▶ Pope and Schweitzer (2011) examine observational data from professional golfers playing on the PGA tour ▶ Golf suggests a natural reference point for score on each hole: par ▶ Par is the number of strokes a professional golfer typically takes to complete the hole ▶ Because object of game is to minimize number of strokes, going below par is the gain domain and above par is the loss domain ▶ Terminology for going above or below par: ▶ Eagle: two shots below par ▶ Birdie: one shot below par ▶ Bogey: one shot above par ▶ Double bogey: two shots above par ▶ Stakes are very high: typical tournament pays out $5 million in prizes to the top finishers  The Endowment Effect Is Tiger Woods Loss Averse? Connection to Prospect Theory ▶ Let ∆x indicate the score relative to par ▶ Prospect theory value function (with no diminishing sensitivity): ( ∆x ∆x ≥ 0 v (∆x) = λ∆x ∆x < 0 ▶ Focus on putting (usually last 1-2 shots of the hole) ▶ Miss the putt for score ∆x − 1, or ▶ Make the putt for score ∆x ▶ Prediction: Putts attempted for par, bogey, and double-bogey will be more accurate than putts attempted for birdie and eagle  The Endowment Effect Is Tiger Woods Loss Averse? Value Function Applied to Golf  The Endowment Effect Is Tiger Woods Loss Averse? Results  The Endowment Effect Reference Points and Effort Provision Reference Points and Effort Provision  The Endowment Effect Reference Points and Effort Provision Expectations as Reference Point ▶ So far, we have mostly thought of a reference point as a fixed number, independent of the choice at hand ▶ However, it is possible that a reference point is based on the expected outcome ▶ For example, equal chances of getting $10, $40, or $50 ▶ Here, reference point of $0 makes little sense ▶ Getting $10 feels like a loss, while getting $40 or $50 feels like a gain ▶ So a more reasonable reference point might be the expected value: $10 + $40 + $50 = $33.33 3  The Endowment Effect Reference Points and Effort Provision Setup: Effort Task ▶ Abeler et al (AER, 2011) perform lab experiment where they manipulate expected payment for a task ▶ Task: count the number of zeros in an array of 150 randomly ordered ones and zeros ▶ This is really annoying Example ▶ Subjects performed as many of these tasks as they wanted, for up to 60 minutes ▶ One of two possible payment schemes ▶ With 50% probability, paid 10 cents per correct answer (piece rate) ▶ With 50% probability, paid a fixed amount: either 3 Euros (LOW treatment) or 7 Euros (HIGH treatment) ▶ Do not know which payment scheme will be used until after they have decided to stop working  The Endowment Effect Reference Points and Effort Provision Expected Results ▶ Assume that reference point for earnings is average of what subject will earn if paid piece rate and what subject will earn if paid fixed amount ▶ Which treatment would have higher effort?  The Endowment Effect Reference Points and Effort Provision Results Research Proposal Example Research Proposal Example  Research Proposal Example Submission I think an interesting research topic would be budgeting money. Being that we’re all in college, we all have to stick to our own budgets. The experiment would include giving a sample size of students all the same sum of money and having them buy groceries for themselves. It would have to deal with how people budget money around foods they want to buy vs. foods they should buy. It would be interesting to set a certain type of time limit to this being: how long can you make that food last to survive yourself. Like how much money are you allocating for essentials and how much is going to waste. I think the kicker of it would have to be that the participants aren’t aware of the time limit they’ll be given to live off their choices til after. It would be cool to measure ”waste” as items that aren’t needed (who bought beer or something).  Research Proposal Example What is the research question? ▶ As proposed, this is not an experiment. It is a measurement exercise ▶ i.e. everyone is given the same amount (no treatments) ▶ the study is descriptive, asking how or what, not why ▶ What theories would inform us about how people vary their consumption? ▶ Think of the basic consumer problem, maxxt t βδ t Ut (xt ; θ) s.t. a P budget constraint ▶ How many things can be changing? preferences, budget constraints, heterogeneous preferences, how sophisticated you are (β, δ), etc. ▶ Which of these features would we like to study and keep the rest constant? ▶ Are you measuring whether people waste or they are wasteful?  Research Proposal Example What is the research question? ▶ Things that can be tested (research questions) ▶ Q1: Do people display projection bias, i.e. they buy things they think they will consume but they do not? ▶ Q2: Are decisions costly, so that individuals make “mistakes” when under time pressure?  Research Proposal Example Why is this interesting? ▶ Lots of waste, i.e. energy, packaging, time loss, landfill, because of these mistakes ▶ Net benefit to society is not clear – as people get more efficient, they demand less in the short run, i.e. those whose jobs depend on your “waste” would be most impacted  Research Proposal Example What are hypotheses? ▶ H0: time pressure makes people more wasteful  Research Proposal Example What is an experimental design to test this? ▶ subjects are given an amount of money, i.e. $50, to do grocery shopping ▶ T1: they are told to make a grocery list to spend the $50 and have 24 hours to do this ▶ T2: they are told to make a grocery list to spend the $50 and have to do this in 5 minutes ▶ T1’: they are told to make a grocery list to spend the $50 and have 5 min to do this in a quiet space ▶ T2’: they are told to make a grocery list to spend the $50 and have 5 min to do this while remembering 7 digit number  Research Proposal Example What is an experimental design to test this? ▶ Outcomes: Items on the shopping list ▶ Observe what items are consumed and what are thrown away during the next week (or next month)  Research Proposal Example Your turn ▶ You will answer the following ▶ What is the research question? ▶ Why is this interesting? ▶ What are the hypotheses? ▶ Broadly, what would be the experimental design to test your hypotheses?  Research Proposal Example Idea to workshop Idea 1: I would like to further research the ideas of risk-averse, risk-neutral, and risk-seeking in individuals and why they develop the way they do in different groups of people. We have looked into the way these ideas vary in individuals in class, but I want to know why certain people develop more profound instances of either risk neutrality, risk aversion, or risk loving characteristics. My research question would be ”Are children who grow up on single parent households more risking seeking than those who grow up in a two parent household?” This study may involve some pretty lengthy and detailed panel data that may take a while to collect with a sample size in the thousands. Although, (with some further research) I may be able to collect data that is available to the public. The study would collect data in certain cities in the U.S. and then go on to collect data from certain major world cities and compare results. This is an ambitious task but could possibly have some ”game changing” results. I hypothesize that children who live in single parent households are more risk seeking than those that live in two parent households.  Research Proposal Example Example of Counting Zeros Task Back