Temptations and Dynamic Consistency

1
Temptations and Dynamic Consistency

• Enrica Carbone
• Università di Bari

2
Recent development of the literature on
intertemporal choice

• Quasi-Hyperbolic discounting, Laibson(1997)
• Consumers have an higher discount rate between
  the present and the next period than between any
  two of the subsequent period.
• This discount function implies time inconsistency
  

3
Discounting models

• Exponential discounting
• u(ct) d u(ct1) d2 u(ct2) ... dT-t1
  u(cT)
• (quasi-)Hyperbolic discounting
• u(ct) ß d u(ct1) d2 u(ct2) ... dT-t1
  u(cT)

4
Investment goods and temptation goods

• An important distinction was made by Read,
  Loewenstein and Kalyanaraman (1999) between goods
  which are virtues and vices
• Stefano Della Vigna and Ulrike Malmendier, (2004)
  made the same distinction between investment
  goods and leisure goods
• Investment goods have current costs and future
  benefits relative to the best alternative
  activity
• Leisure (temptation) goods current benefits and
  future costs relative to the best alternative
  activity

5
Previous empirical work

• I follow up two experiments by Read et al. (1998
  and 1999).
• In one experiment they ask people to chose
  between instant-win and prize-draw lottery
  tickets.
• In the other they ask people to chose between
  healthy and unhealty snacks.

6
Hypothesis

• Naive hyperbolic people should plan according to
  the discount factor d, but when they actually
  choose they should apply the discount factor ßd.

7
Experiment

• To test this hypothesis we created a spot market
  and a future market for goods and activities
  (some of the goods were temptation goods and some
  were investment goods)
• Naive quasi-hyperbolic people when they decide
  what to buy they discount at a factor ßdltd and
  they should choose more temptation goods with
  respect to the exponential people, because
  comparatively they give less weight to the future
  cost and more weight to the present benefit.

8
Experiment forward market

• 15 minutes to read the instructions and choose
  the activities and the goods.
• 3 hours and 45 minuts to consume the goods and
  to play the activities previously bought.
• The participants could not talk to other
  participants during the experiments.

9
Experiment spot market

• 15 minutes to read the instructions.
• 3 hours e 45 minutes to consume the goods and to
  play the activities they wanted.
• The participants could not talk to other
  participants during the experiments.

10
Goods sold both in the spot market and in the
future market

• Espressino
• Espresso coffee
• Cioccolat Snack
• Focaccia
• Sandwich
• Oransoda
• Lemonsoda
• Chinotto
• Crostini
• Heqalthy salad Insalata
• Fruit
• Yoghurt skimmed
• Crisps
• Cucciolone (icecream)
• Magnum Nuts (icecream)
• Cornetto (icecream)
• Sweets

11
Activities sold both in the spot market and in
the future market

• Textbook reading
• Research on the electronic catalogue of the
  department
• Pectoral muscles exercise device
• Reading Magazines
• Videogames
• Listening music

12
Hypotesis tested

• If people are hyperbolic in the spot market, the
  temptations goods and activities consumed are
  significantly more than in the forward market.

13
Steps for analysing the data

• Combine multiple purchases of the same
  good/service by the same person
• Add quantities together, multiply by cost/unit
• Add missing data
• If someone didnt buy a certain good, add an
  observation that has an expenditure equal to 0
• Normalize expenditure with the following formula
• where e is the expenditure on good j.
• Get mean ratings for short term and long term for
  each good

14
Table1
15
Conclusion

• People consume more temptation goods both in the
  spot market and in the future market (there is
  only a general discount effect).