Question

Ebenezer Scrooge, in his spare time, happens to be an intelligent and data-driven economist. He tells you he doesn’t believe in the value of behavioral economic concepts because he’s seen neither data to support them nor real-life, high-stakes examples of their consequences. For each of the following concepts, explain to Ebenezer what sort of experimental or field data could support the need for a behavioral model. (These data can be the result of a hypothetical experiment, the existence of a real-world phenomenon, or a reference to a study covered in the course.) Then, provide Scrooge with an example of how the predictions of such a behavioral model can have implications in an important economic context or policy debate.

(a) Social preferences

(b) Effects of social information nudges

(c) Time inconsistency

Answer 1

1) Traditional economic model is of homo economicus: rational and self-interested agent. Behavioral economics seeks a richer description of economic agents by relaxing these two assumptions:

1. Bounded rationality: limited cognitive abilities, viz., people are not always expected discounted utility maximizers 2. Social preferences: assumption of self-interest must sometimes be appended to account for interdependent preferences such as fairness and envy.

*Focus of this presentation: social preferences branch of behavioral economics.

Types of Social Preferences Emerging consensus among behavioral economists:

Distributive preferences: preferences over the final distribution, e.g., equity, efficiency, altruism; related to consequences or outcomes. Reciprocal preferences: desire to reward or punish others beyond mere consequences, e.g., being “more than fair” to someone who has been fair to you; related to intentions or types of agents.

Field Evidence of Social Preferences

Kahneman, Knetsch and Thaler (AER 1986)

People care about fairness. Ex: It is sometimes unfair to raise prices to eliminate a shortage, e.g., on snow shovels after a snowstorm . This concern for fairness constrains profit seeking.

Blinder and Choi (QJE 1990)

Wage setters at firms confirmed the importance of fairness considerations in contributing to wage rigidity.

Babcock, Wang and Loewenstein (QJE 1996)

Survey teachers unions and school boards. Strike activity is positively related to differences between teachers unions and school boards over the salaries they consider fair and comparable.

Andreoni, Erard and Feinstein (JEL 1998)

Amount of tax evasion is affected by the perceived fairness of the tax system.

Zajac (Political Economy of Fairness 1995)

Public support for the regulation of private industries depends on perceived fairness of firms’ policies.

Experimental Evidence of Social Preferences

Güth, Schmittberger and Schwarze (JEBO 1982)

First experimental test of the “ultimatum game,” a simple two player, two stage bargaining game. Proposer offers a division of a fixed sum to an anonymous Responder, who then accepts or rejects the offer ¾ If the Responder accepts, the sum is divided as agreed ¾ If the Responder rejects, both receive nothing. The subgame perfect equilibrium: the Proposer offers the minimum amount and the Responder accepts.

Camerer (Behavioral Game Theory 2003)

Hundreds of replications of the ultimatum game produce similar results. Average Proposer offers are about 40% Responders reject offers below 20% about one-half of the time.

Forsythe, Horowitz, Savin and Sefton (GEB 1994).

Ultimatum game contains a possible confound: even self-interested Proposers might make generous offers if enough Responders are expected to reject unfair offers. “Dictator game” removes this confound: Proposers can transfer any amount of a fixed sum, or nothing, and Responders have no recourse. Results: Proposers offer less in the Dictator game than in the Ultimatum game but still give, on average, about 20-25% of the stakes.

Fehr, Kirchsteiger and Riedl (QJE 1993)

First experimental “gift-exchange” game: Experimental labor market in two stages: 1. Employers first offer a wage to workers 2. Workers then choose a costly effort level that increases the employers’ earnings. Results: 1) worker effort is directly related to employer wage offers, and 2) average wages exceed the market clearing wage.

Charness (JLaborE 2004)

Gift exchange experiment found that: 1) worker effort varies more when the wage offered by an employer is intentional than when it is random, and 2) both distributive and reciprocal preferences matter.

2) A three-fold change to the design and delivery of public services has been taking place over the past decade. Expectations of user choice or personalisation, emergent localism and most particularly the implications of the economic crisis each increase tensions within the public service framework. One key factor underpins all of them: they require fundamental change in the expectations of individuals, communities and service providers if the most is to be made of diminishing resources and whilst securing public well-being. Many experts have said that the critical public service challenge of the decade is to encourage behaviour that benefits both the individual and the state, whilst preventing long term expense. They want to discourage behaviour which creates user dependency and attracts further costs. Behaviour change is vitally important, they say, because we can no longer provide the services we have always done, in the way we have always provided them. Various approaches to altering the behaviour of citizens have been outlined in a growing body of evidence including Nudge (Thaler and Sustein) ‘Think’ (John et al) and MINDSPACE (Dolan et al).

However, our belief is that behaviour change alone is helpful but will not be sufficient because it focuses too heavily on individuals and not on the system and community as a whole. There is too much reliance on service users choosing to do something different when actually the need is for the individual and the community to  think differently. We believe that this requires an attitudinal or cultural change and not simply behavioural change. INLOGOV’s new model for public services provides a useful distinction between individual co-production, community co-production and self-help activities which this chapter will draw upon.

Why changing behaviours and attitudes is important

We know from what Barnet Council has called its ‘Graph of Doom’, at Figure 1: Barnet's Graph of Doom, 2012 Councils are navigating within a ‘perfect storm’ of reducing funding and increasing demands from demographic change, public expectations and the rising cost of delivering services. So somehow the level of demand on services from citizens needs to be contained, and reduced. Just changing the way in which existing services are delivered will not save enough money. For example if the current trend of people needing care continues and the use of personal budgets in their current form is extended, there is a clear risk of double pressures on the public purse, as current services such as day care continue to be provided rather than decommissioned. We believe that following simultaneous outcomes will be required in the future, some of which will be the responsibility of public services:

Reduced dependence/reliance on state to pick up the pieces
Improved individual well being and resilience
New and improved community/social networks
Sustainability – both in terms of the environment and also the future of public services.
A better understanding in the community of the cost of public service and its relation to taxation
A shift in the underlying expectations of individual citizens and communities of the deal that they have with the state as to the provision of public services.

A new contract between the citizen and the state

The underlying implication of the above is that there is a need to change the contract between the individual and the state. There has been a range of reports and statements from think tanks and central government departments extolling this approach.  

The RSA 2020 Public Services Need to shift current approach in culture, power and finance. Together, these shifts will open up space for new, bottom-up approaches to solving public problems. This requires a different, but still active role for the state: a state that is active in stimulating social productivity, building citizen capabilities and fostering social resilience. Putting People First HM Government 2007 and subsequent All social care users should have access to a personal budget, with the intention that they can use it to exercise choice and control to meet their agreed social care outcomes. A new social contract between the citizen and the local state Commission on the Future of Local Government Leeds City Council July 2012 Local government and partners should forge a new social contract between the citizen and the state in which services can be delivered with rather than to individuals and communities. The RSA 2020 Public Services final report 4 provides a good summary of many of the more detailed points. It calls for a new ‘social citizenship’ approach where as citizens we should have a duty to contribute as well as a right to receive support. Much of this work has made use of behaviour change research and techniques, such as Nudge and other disciplines including work under way in fields as diverse as deep academic social psychology and the Volkswagen fun theory piano staircase5 . However there are problems with a simply behavioural approach because of the lack of a lasting underlying change: “The deepest problem with nudge is that it is not transformative. Indeed, darkly, this may be why it is so popular. Nudge changes the environment in such a way that people change their 4 From social security to social productivity: a vision for 2020 Public Services , the final report of the Commission on 2020 Public Services 5 Thefuntheory.com, Volkswagen 2009 behaviour, but it doesn’t change people at any deeper level in terms of attitudes, values, motivations etc. In this respect, nudge creates what psychologist Paul Watzlawick calls ‘firstorder change’ rather than second order change” sets out an adapted version of the choice architecture as suggested in Nudge, which demonstrates the various approaches open to public sector agencies to use to try and change behaviour. Most public sector agencies, including local and central government, and health providers, have experimented with some or all of these approaches.

3) Behavioral economics explains phenomena like time-inconsistent intertemporal choice and procrastination as the product of “present-biased” preferences and possible naivete about the choices of one’s own future selves.

For a time-consistent planner, other things equal, if it’s beneficial to do something next week/month/year/etc., it’s (even more) beneficial to do it now. But many people don’t act that way. Familiar quotations you would never hear from a time-consistent planner:
? Next month, I’ll quit smoking.
? Next week, I’ll catch up on the required reading.
? Tomorrow morning, I’ll wake up early and exercise.
? After Christmas, I’ll start eating better.

Not only do these quotations all reflect time-inconsistency, they are all inconsistent in the same direction: All excessively favor gratification (or avoiding nongratification) now, at the expense of future gratification (or avoidance).

Excessively favoring gratification now at the expense of future gratification can usefully be modeled via “present-biased” preferences, on which more below. Future bias would yield quotations you seldom hear:
? I plan to watch more TV next year.
? I plan to eat more cookies and doughnuts next year.
? I plan to smoke more cigarettes next year.
? I plan to borrow more on my credit card next year.
? I plan to exercise less next year.
? I plan to wake up later next year.

Not only do these quotations all reflect time-inconsistency, they are all inconsistent in the same direction: All excessively favor gratification (or avoiding nongratification) now, at the expense of future gratification (or avoidance).

Excessively favoring gratification now at the expense of future gratification can usefully be modeled via “present-biased” preferences, on which more below. Future bias would yield quotations you seldom hear:
? I plan to watch more TV next year.
? I plan to eat more cookies and doughnuts next year.
? I plan to smoke more cigarettes next year.
? I plan to borrow more on my credit card next year.
? I plan to exercise less next year.
? I plan to wake up later next year.

One possible exception:
? John Maynard Keynes’s last words: “I wish I had drunk more champagne”. (But he was a workaholic, so perhaps even this was present bias.) 10 “Excessively” relative to what? The standard exponential discounting benchmark. 11 As the familiar quotations also illustrate, a common symptom of present-biased preferences is procrastination, whereby the current self defers nongratification in pursuit of a long-term goal while planning for some future self to pursue it. 12 Procrastination is a prominent problem (numbers from 2011; I’ll update soon…):
? With 14,800,000 Google hits, 33% more prominent than ennui with 11,100,000, but 47% less important than plagiarism with 27,800,000.
? Far less prominent than pride with 418,000,000; lust 216,000,000; wrath 148,000,000; greed 114,000,000; or envy 110,000,000.
? But close behind sloth with 15,900,000 (not counting those with toes), and easily tops gluttony with 5,450,000.
? (For comparison: Obama 723,000,000 Google hits; Lady Gaga 479,000,000; The Simpsons 159,000,000; Emma Watson 92,300,000; Marx (not counting Groucho) 90,900,000; Keynes 51,000,800.) Future bias would make people accelerate unpleasant tasks or defer gratification. English actually has a word for this:
? “Preproperation” (456 Google hits and counting). 13 Less casual evidence With exponential discounting, plausible discounting for a year from now implies implausibly low (virtually zero) discounting for a day or a week from now. Thaler (1981 Economics Letters) asked subjects to choose between money now versus more money later, hypothetically (but works for real choices too):
? What amount makes you indifferent between $15 today and $X in 1 month? Typical response: X = 20. Implied discount rate: 345% per year.
? What amount makes you indifferent between $15 today and $X in ten years? Typical response: X = 100. Implied discount rate ? 19% per year. 14 Shapiro (2005 J. Public Economics) examined consumption patterns among food-stamp recipients, and found that over the month between food-stamp deliveries, a family’s caloric intake declines by about 10-15 percent. Survey evidence revealed rising desperation over the month, suggesting that high elasticity of intertemporal substitution is not a likely explanation. Households with more short-run impatience (estimated from hypothetical intertemporal choices) were more likely to run out of food during the month. To explain these patterns with a standard exponential-discounting model would require calibration with an extreme annual discount factor of 0.23. Models like those discussed below explain the patterns more gracefully. 15 Read, Loewenstein, and Kalyanaraman’s (1999 J. Behavioral Decision Making) and others found similar results for real choices in kind, avoiding some confounds. Subjects chose among 24 movie videos.
? Some were “lowbrow”: e.g Four Weddings and a Funeral
? Some were “highbrow”: e.g. Schindler’s List 16 (The canonical highbrow movie: A man seeks answers about life, death, and the existence of God as he plays chess against the Grim Reaper during the Black Death.)

Results:
? Choosing for tonight: 66% of subjects chose lowbrow.
? Choosing for next Tuesday: 37% chose lowbrow.
? Choosing for second Tuesday: 29% chose lowbrow. 18 Badger et al. (2007 J. Health Econ.)

on recovering heroin addicts’ monetary equivalent of an extra dose of Bupronorphine (like Methadone, reduces craving). Elicited willingness to pay for second dose (still attractive to addicts) from 13 long-term heroin addicts regularly receiving a single dose: $10? $20?...$100? Subjects told (truthfully) that one of their choices, randomly selected, would be implemented: Thus had incentive to choose according to true preferences. Half of the subjects were asked when more deprived (2 hours before scheduled dose), half when less deprived (right after scheduled dose). Half were asked about a dose today, half about a dose on their next visit. (Subjects were always experienced; and were always asked about a second dose, to be delivered in the satiated state.)