Salience in Experimental Tests of the Endowment Effect by Pedro Bordalo, Nicola Gennaioli, Andrei Shleifer Presented by Maria Weber 1
Agenda I. Introduction II. Research Question III. Salience IV. Of Mugs and Pens V. Of Mugs and Bucks VI. Conclusion VII. Critique 2
I. Introduction The endowment effect: 1st stage: endowment with a good (e.g. a mug) 2nd stage: opportunity to trade the endowed good for another good (e.g. a pen) just around 10% chose to trade gap between WTA and WTP for one and the same good common explanation: loss aversion (prospect theory of Kahnemann and Tversky) new approach for explanation: impact of contextual factors such as nature of the goods information about market prices 3
II. Research Question New modeling of the endowment effect: context dependency based on the salience mechanism decision making: focusing on / overweighing of the features whereby options are salient ("stand out") evaluation in the two stages happens in different contexts: 1st stage: comparison to status quo of having nothing overweighing of good s best attributes "warm glow" of receiving gift / becoming owner 2nd stage: contrast of endowed and new good salience of the new good s disadvantages (underevaluation of good) the overweighed valuation of the endowed good persist 4
III. Salience Two-attribute good: good k is a vector of qualities (q 1k, q 2k ) intrinsic utility: v(q 1k, q 2k ) = w 1 q 1k + w 2 q 2k perceived value differs to due overweighing of salient attribute evaluation of good k if attribute i is salient and j not: with δ ϵ (0, 1] = degree of neglection of nonsalient features which attribute is salient for good k depends on Ʃ weights = 1 1. consideration set C (all goods considered for decision making) 2. salience function σ (measures the extent of salience of an attribute relative to the average) assumptions for σ: 1. ordering 2. diminishing sensitivity 3. reflection 4. symmetric 5. homogenous of degree zero 5
IV. Of Mugs and Pens q 1 = "quality for drinking and q 2 = "quality for writing equal weights for both attributes w 1 = w 2 = 1/2 mug M is a good (q M, 0), a pen P is a good (0, q P ) intrinsic value of both goods: q/2 endowment stage (endowed with mug): consideration set C e =, average good (q/2, 0) the quality of the mug is salient: σ(q,q/2) > σ(0,0) perceived value higher than intrinsic value the mug s salience ranking persists to the trading stage with probability ϒ 6
IV. Of Mugs and Pens trading stage (trade or not trade the mug for a pen): consideration set C t = pen s quality for writing is not salient: σ(0,q/3) > σ(q, q/3) pen s lack of quality for drinking is more salient than it s high quality for writing perceived value lower than intrinsic value, average good (q/3, q/3) mug s average valuation in the trading stage: (3) v LT ( = q * as long as ϒ > 0 the decision maker keeps the mug (endowment effect) 7
IV. Of Mugs and Pens Predictons due to the changed perspective on the endowment effect: (i) No endowment effect for identical objects (overweighing of good s best attributes for both). No endowment effect in upgrading. The endowment effect requires a trade off between the endowed good and the new good. (ii) There is a reverse endowment effect for bads. Focus on the bad s downside instead of the good s upside. "cold glow" of ownership, eagerness to trade (iii) No endowment effect for comprehensive endowments. Endowment with both goods, warm glow of ownership for both goods 8
V. Of Mugs and Bucks Gap between willingness to pay (WTP) and willingness to accept (WTA): q 1 = quality q of the mug and q 2 = price disutility utility from the mug (q, -p) is q/2 - p/2 endowment stage: - consideration set C e = - salience of mug s upside trading stage: - consideration set C WTA = - including option of getting his WTA - downside of all options is salient (as before) - again perceived value lower than intrinsic value (4) v LT ( = WTA * < WTA/2 9
V. Of Mugs and Bucks monetary gain is undervalued, focus lies on the loss of the mug value of mug as before (see (3)) from equating (4) and (3) we get the value of the WTA (5) WTA = q*(1+ ϒ* ) taking into account the WTP: not endowed with the mug (missing warm glow) consideration set C WTP = correct valuation: (6) WTP = q (price and quality are equally salient) comparing (5) and (6): positive WTA WTP gap = q *ϒ * endowment effect 10
V. Of Mugs and Bucks Dependency of the WTA WTP gap and the market price p M now information about the market price is given to decision makers consideration set: C t = new: option of selling the mug in the market again downsides of options are salient α = probability of seeling, (1-α) = probability of keeping, 1/δ = boost of mug s quality due to warm glow of endowment (7) WTA = αp M + (1 - α)q * 1/δ (decision maker s selling price) stating the WTP for the decision maker we get C WTP = new option: trading the mug at the market price 11
V. Of Mugs and Bucks downside of each option salient (given p M >> q), in particular the price WTP paid when buying the mug (8) WTP = (p M + (1-α)q)* (decision maker s buying price) with α = probability of reselling the mug equations (7) and (8) give the WTA WTP gap (1) the gap arises whenever there is uncertainty about trading (α < 1) (2) the selling price is more (less) sensitive than the buying price to the market price when market price is high (low) the WTA WTP gap increases with changing market price (in both directions) 12
VI. Conclusion the given model does feature salience and context dependence instead of prospect theory s loss aversion the difference in the evaluation in the two stages (absolute vs. comparative) is highlighted focus on the endowment s most attractive attributes vs. focusing on the alternative s downside endowment effect the focus on the most attractive attributes of the owned good makes ownership special whereas there is no warm glow if there is a better option connection between the endowment effect and risk behavior same mechanism of salience can help understand the attitude towards risk 13
VII. Critique the paper gave me a quite interesting insight in the understanding of the endowment effect from a new point of view can hardly find any critique Possible extensions: complex goods (positive values for both qualities): authors write in a foodnote that main results go through for complex goods uncertainty in the information about the market price 14
Thank you for your attention! 15