Modeling self-control problems II: Dual-self models Lectures in Behavioral economics Fall 2011, Part 4 (Based on notes by Kjell Arne Brekke) 13 September 2011 G.B. Asheim, ECON4260, #4 1 Dual self models of impulse control Thaler and Sherfin (1981) ● Consider each individual as an organization ● Consist of a planner and doers ● The planner is concerned with lifetime utility ● The doers exists only one period and is completely selfish Current interpretation ● Planner is prefrontal cortex ● Doers are the rest of the brain (reptilian brain) Fudenberg and Levine (2006) ● Equivalent formulation: Planner control doers at a cost. 13 September 2011 G.B. Asheim, ECON4260, #4 2 1 Using new insight from neurology—Neuroeconomics The importance for economics discussed “Economics needs brains” “The case for a mindless economics” Gul and Pesendorfer (2008) New knowledge about the brain has made the dual self model much more plausible Some argue as if such evidence is almost conclusive Others insist that econ econ. th. th should be tested on econ. econ data Camerer,, Loewenstein and Prelec (2004, ( , 2005)) Consumer theory makes no claims about the brain Nor about underlying psychology Neurology can inspire new theory building To be tested on economic data 13 September 2011 G.B. Asheim, ECON4260, #4 3 The prefrontal cortex Phineas Gage survived a pole through frontal lobe After the damage, the doctor described him: “.. devising many plans for future operation which are no sooner arranged than they are abandoned. … Ladies were advised not to stay in his presence. The hallmark of Gage’s new condition was his complete inability to direct or control himself” 13 September 2011 G.B. Asheim, ECON4260, #4 4 2 Choices under cognitive load Shiv and Fedorikhin (1999) S bj t were instructed Subjects i t t d to t memorize i 2 (7) digits di it inducing low (high) cognitive load Then walk from one room to another, in-between choosing between a cake and a fruit salad Choice Memorize 37 (5274657) 13 September 2011 Recall G.B. Asheim, ECON4260, #4 5 Choices under cognitive load (cont.) Results High load: 59 percent chose cake Low load: 37 percent chose cake Interpretation Prefrontal cortex cares about the long run (health) Prefer the fruit salad Sugar (the cake) is tempting to reptilian brain (the crocodile within) High cognitive load reduces the prefrontal cortex’s ability to control the impulses from the reptilian brain Cognitive load; short term memory has 7 slots 13 September 2011 G.B. Asheim, ECON4260, #4 6 3 The dual self model Thaler and Shefrin (1981) do not talk about prefrontal f l cortex andd reptilian ili brain, b i b but this hi h how the model is often interpreted today. (And this is this interpretation that we will use here). Their motivation: “The idea of self-control is paradoxical unless it is assumed that the psyche contains more than one energy system, and that these energy systems have some degree of independence from each other” (McIntosh, 1969). 13 September 2011 G.B. Asheim, ECON4260, #4 7 The dual self model (cont.) A planner l (th (the prefrontal f t l cortex; t frontal f t l lobe) l b ) A sequence of doers (the reptilian brain) Setting Income stream: y (y1, y2, … , yT) Consumption plan: c (c1, c2, … , cT) The utility of the doer at time t: Zt(ct) The utility of the planner: V(Z1, Z2, … , ZT) 13 September 2011 G.B. Asheim, ECON4260, #4 8 4 Alter incentives Utility Unaltered pref. The planner can alter the Cost of impulse control doer' s incentive s through t . Doer' s utility : Z t (ct ; t ). Altered preferences Such altering of the doer' s preferences induces costs. If t is chosen such that arg max Z t (ct ; t ) cˆ, then Z t (cˆ; t ) Z t (cˆ;0). 13 September 2011 ĉ G.B. Asheim, ECON4260, #4 c Consumption 9 Equivalent formulation: Fudenberg and Levine (2006) Doer' s utilityy u (ct ) is strictlyy increasing. g Maximum feasible consumption ct . Cost of self control : C (c, c*) (u (c*) u (c)). The more the doer sacrifices the more effort required. Total utility : Z t u (ct ) C (ct , ct *). Planner maximizes V ( Z1 , Z 2 ,..., ZT ). 13 September 2011 G.B. Asheim, ECON4260, #4 10 5 Application: Immediate money First choice problem (in 1000 $): Second choice problem (in 1000 $): A: 1 in year 1, 1 or B: 1.5 1 5 in year 2 C: 1 in year 2, or D: 1.5 in year 3 Basic income stream: 10 each year. Doer’s utility: u(ct) lnct May save, but no borrowing. Cost of self control: 0.5ln ct ln ct Planner’s utility: V ( Z1 , Z 2 , Z 3 ) Z1 Z 2 Z 3 13 September 2011 11 G.B. Asheim, ECON4260, #4 No saving due to control costs—B as an example Income stream: 10, 11.5, 10 In year 2, c2 11.5 dV 1.5 1 dc2 c2 21.5 c2 dV dc Saving some of the additional income 2 c2 11.5 1.5 1 would induce cost of self control: 0 11.5 10 For c2 c2 11.5, utility equals ln10 ln c2 ln(10 11.5 c2 ) 0.5(ln11.5 ln c2 ) 1.5 ln c2 ln(10 11.5 c2 ) ln10 0.5 ln11.5 Optimal consumption stream: 10, 11.5, 10 13 September 2011 G.B. Asheim, ECON4260, #4 12 6 Choosing A or B B : ln10 ln11.5 ln10 0.5ln11 ln10 ln 1150 ln1096 1.1 Additional income in A or B will be cons. immediately. y Without problems of self control, B is best: A: ln11 + ln10 + ln10 ln1100 B: ln10 + ln11.5 + ln10 But with A as an option, c1 11 , while choosing B i li c1 10 . implies The cost of self control exceed additional utility. The planner will choose A. 13 September 2011 G.B. Asheim, ECON4260, #4 13 Choosing C or D Additional income in C or D will be cons. immediately. With no cost of self-control, D is best: C: ln10 + ln11 + ln10 ln1100 D: ln10 + ln10 + ln11.5 ln1150 With cost of self control Onlyy 10 available to the doer in year y 1 in both C and D. Choice made before doer in year 2 or 3 may influence it. No cost of self control. The planner will choose D. 13 September 2011 G.B. Asheim, ECON4260, #4 14 7 Compared to Gul and Pesendorfer (2001, 2004) With onlyy A available,, utilityy is ln11 + ln10 + ln10 With only B available, utility is ln10 + ln11.5 + ln10 With a choice between A and B, the planner will choose A dduee to the cost of self control. control Thus: Th s { A} ~ { A, B} {B} ln 1100 ln 1100 ln 1150 13 September 2011 G.B. Asheim, ECON4260, #4 15 . Compared to Gul and Pesendorfer (2001, 2004) (cont.) B: 2 in year 2 With onlyy A available,, utilityy is ln11 + ln10 + ln10 With only B available, utility is ln10 + ln12 + ln10 With a choice between A and B, the planner will choose h s B in spite spit off self s lf control ntr l costs. sts Th Thus: s { A} { A, B} {B} ln 1100 ln 1144 ln 1200 13 September 2011 G.B. Asheim, ECON4260, #4 16 8 Control methods Alter preferences Alter opportunities: Rules Thi corresponds This d to the h approach h studied di d Precommitment, (e.g. non-liquid assets) Limiting choices (avoid Las Vegas) Rules of thumb (ban on borrowing) Why would rules of thumb work? Public announcement (“I’ve stopped smoking”) Other? 13 September 2011 G.B. Asheim, ECON4260, #4 17 Applications in Thaler and Shefrin (1981) Rich in ideas; weak in formal proofs. The effect of mandatory pension saving A person saves a share s of income Saving p < s becomes mandatory Does total saving change? Bonus-payments and saving Two person with annual income 84 (thousand), one get 7 per month the other gets 6 per month + a bonus of 12 once a year. Who saves most? 13 September 2011 G.B. Asheim, ECON4260, #4 18 9 Bonus payment B gets the bonus, A does not A saves 0.4 per month. Cost of impulse control: 12(ln7 – ln6.6) 0.71 B saves 4.8 4 8 in one month. month A earns 7 p per month B earns 6 except one month with 12 Both have doer utility lnc Both save 4.8, what are their cost of impulse control? Cost of impulse control : (ln18 – ln13.2) = 0.31 Least cost of self control for B, since marginal utility the one month is lower. 13 September 2011 G.B. Asheim, ECON4260, #4 19 Conclusion Dual self model “Planner” (p (prefrontal cortex)) controllingg “Doer” (reptilian brain) Equivalent to a cost of self control if deviating from short run optimum. Explains Choosing 1 now over 11.55 in one year year, while also choosing 1.5 in two years over 1 in one year. Demand for commitment. Consumption correlating with income. 13 September 2011 G.B. Asheim, ECON4260, #4 20 10 Additional references Colin Camerer, George Loewenstein and Drazen Prelec (2004): Neuroeconomics: Why Economics Needs Brains. Scandinavian Journal of Economics, 106, 555-579 C li Camerer, Colin C G George L Loewenstein t i andd Drazen D P l Prelec (2005):Neuroeconomics: How Neuroscience Can Inform Economics, Journal of Economic Literature, Vol XLIII, pp. 9-64 Faruk Gul and Wolfgang Pesendorfer (2008): The Case for Mindless Economics, in The Foundations of Positive and Normative Economics, by A. Caplin and A. Shotter (eds.). Oxford Univ Press http://www.princeton.edu/~pesendor/mindless.pdf p // p / p / p Baba Shiv and Alexander Fedorikhin (1999): Heart and Mind in Conflict: The interplay of Affect and Cognition in Consumer Decision Making, Journal of Consumer Research, 26, 127-90. Rita Carter (1998): Mapping the Mind. University of California Press 13 September 2011 G.B. Asheim, ECON4260, #4 21 11