8.24 - Experience & Endowment effect

ST vs Endowment effect

ST - Subject should have same choice always if feasible set the same

Endowment effect - People more attached to things they already ‘have’

Knetsch (1989) recap

3 groups

• 2 Groups given mug or chocolate bar and allowed to swap

• Other group had choice of either (no endowment)

ST predicts proportion who end with mug should be consistent in all groups as feasible set the same in all groups

• apart from random sampling variation

Evidence shows Endowment effect (90% of initial E vs ~50% in choice group)

Why introduce a market to test Endowment effect

Knetsch (1989) results could be from weak lure of endowment that ‘tie-breaks’ if nearly indifferent

• Only if many are indifferent between M / C

Choices in real life often money vs good (not good vs good)

• EE might not survive if “keeping the endowment” is “giving up money”

  • i.e. not selling

Kahneman et al (1990) - OV

Look at their experiment 1

• 11 rounds of market exchange involving different goods - repeated market context

  • Induced value tokens + mugs + pens

• Can test for endowment effect for relevant good in each round

• Can test for survival of endowment effect in later rounds after earlier experience

Design

• Each round 50% of subjects endowed with 1 unit of good

• These subjects asked for a range of prices if they would sell at that price

• Other subjects asked if would buy at these prices

• Given responses experimenter selects price to match S & D - market price

  • Subjects then trade if they said they would

Innovative method at the time but not best practice now

• small samples + experiment in class

Kahneman et al (1990) - Theory

ST

• Call the 50% who like the good most ‘likers’ + others ‘haters’

  • If subject a Liker or Hater shouldn’t be affected by endowment

• Regardless of endowment, after trading those holding the good should he those who want the good - Coasian insight

• Random allocation of subjects means Endowed group comprises 50%

  Likers and 50% Haters

• Therefore endowed haters will trade to non likers but endowmed likers wont

Volume of trade should be 50% (those endowed but haters)

Endowment effect

• Fewer that 50% of units traded

Kahneman et al (1990) - Induced value tokens

Tokens have no intrinsic value

• value induced by willingness of experimenter to buy them for cash at end of experiment - only value from disposal

Induced values can be varied by subject

• Experimenters can induce distributions of willingness to sell/buy at different prices, creating Likers and Haters

For Mugs / pens value is from subjects own preferences

• more chance good kept

Kahneman et al (1990) - RESULTS

No EE for induced value tokens - ST holds

Volume of trade much lower that ST for M/P

• EE persists over all rounds

EE only for goods people expect to keep

List (2003) - OV

Sports trading cards field experiment - 2 different goods (small market for both goods)

• test for EE in setting that:

  • Feels natural to participants - Sports memorabilia marketplace

  • Offers much of control that lab offers

  • Has measurable variation in participants’ prior market experience

Wide variation in trading experience - dealers make more trades + 80% male + avg age = 36

List (2003) - Design

On entering show (or, for dealers, at their stalls), subjects asked if would fill out survey about participation in similar events

• Information on prior trading experience

In return for agreeing,, subject immediately given either good A or good B

• Pseudo-random allocation, based on clock

After completing survey, the other good is shown and subject given chance to swap

List (2003) - Theory

ST

• same feasible set so preference should vary with which card given 1st

• swap rate should be ½ (indpendent of proportion preferring either)

  • prefer A = p prefer B = 1- p → avg swap rate = (1-p+p)/2 = ½

List (2003) - RESULTS

Swap rates higher for dealers - not sig different from 50%

Non-dealers show EE (BUT experience non-dealers also have 50% swap rate)
Experience shown to push preferences towards ST

• Increasing Trades per month → less EE

EE found outside lab with non-students

List (2003) - Caveats

Resale - dealers + high-volume traders may never intend to keep good

• EE not expected to hold

Good-specificity - Participants have specialised interest in unusual goods

• does this trading experience erode EE in other contexts or just sports trading

Reverse causality - Does trading experience / dealership erode endowment effect or do people with weak endowment effects select into dealership / trade a lot

List (2004) - OV & results

Uses Mugs & candy in same trading market-place

• addresses resale + good-specificity issues

Results

• No EE for dealers or high-volume traders

• Strong EE for non dealers (aggregate + low-volume traders)

List (2011) - OV & results

Focuses on group who will commit to come back again to sports trading place

• reduces observations

Treated group get additional gift of multiple pieces of memorabilia

• may increase their trading between events

Test EE between treated & untreated groups

RESULTS

• Some evidence that gifts (induced extra trading experience) weaken subsequent EE

• Causality of trading experience causing reduced EE

• Although sample size very small

Critique of List’s approach

The market consumers have an extremely specialised interest - does external validity improve from normal set of students?

market experience is of a particular kind - dealers trade on both sides of the market

• being a dealer is uncommon

• Experience highlighted role (dealer vs non-dealer) + extent of experience (intensity - volume of trades)