Lecture 11 Utility

  • Utility
    • Conditions to define utility function:
      • mapping/function: \(\{ \text{Alternatives} \} \to \mathbf{R}\)
      • assign larger number to more preferred alternatives
    • Historically: "utility" means pleasure, happiness, satisfaction etc of receiving, owning or consuming
    • Utility = INDEX of measure of preference
    • If Rational preference relation, we can find a utility function to represent it if \(\{ \text{Alternatives} \}\) is finite.
    • For infinite, complication. Also, if a utility function represents preference relation then the preference relation is rational!
      • However, reverse implication is not true.
    • Given utility function \(u(\cdot)\), the following conditions hold
      • \(x \succ y \iff u(x) \gt u(y)\)
      • \(x \sim y \iff u(x) = u(y)\)
    • Utility functions are not unique. Here, we use ordinal utility: that just allows us to order things
      • The absolute numbers don't say anything. It is only the relationship that matters.
        • e.g. Lataji > Arijit Singh
      • Don't ascribe any significance to the ratios either
        • e.g. if \(u(\text{Lataji}) = 2u(\text{Arijit})\) it doesn't mean that I like Lataji "twice as much" as Arijit Singh (the scale is not scaling)
  • How do utilities relate to indifference curves?
    • \(u(\cdot)\) assigns one number to each IC
    • Two bundles on the same IC \(\implies\) same utility
    • \(u_{1} \lt u_{2} \lt u_{3}\)
  • How does utility relate to behavior?
    • Choosing the most preferred item corresponds to choosing the item with the highest utility.
    • You choose rationally insofar as you maximize utility. We thus, behave rationally.
  • Historical evolution of utility
    • Underlies the theory of consumer choice. We concern ourselves with riskless choice.
    • Jeremy Bentham, 1789: assumption that objective is EU maximization. (A fundamental concept of NM)
      • utility 'points out what we ought to do, as well as what we shall do'
      • Utility has a hedonic2 characteristic (noted by Kahneman (2000) as experienced utility)
    • Meaning has changed in favor of the concept: decision utility
      • Utility = "Weight assigned to the outcome of a decision", revealed by people's choices
      • Call it the "Revealed preference meaning" of utility, generally used on NM
      • Two obvious advantages
        • Easier to measure: can be inferred from choices and actions
        • No commitment to a hedonistic philosophy
      • Sen (1987): "maximization of expected utility is not always what people are trying to achieve"
  • Decision Utility
    • Easiest to measure in terms of revealed preference (usually discussed by economists)
    • Doesn't reflect attitude or judgements
    • Study by Tversky and Griffin (2000)
      • 66 undergrads
      • For the same role and experience
      • Mag A: 35000 (others make 38000)
      • Mag B: 33000 (others make 30000)
      • Half were asked: which they would choose
        • Decision utility
        • When asked to make a choice, they make arguments and justify their choice
      • Other half: which would make them happier
        • Hedonic or experienced utility
        • In this case, expected in future, people try to imagine how they would feel
      • Results:
        • 84% chose (A)
        • 62% thought (B) would make them happier
    • More recent Study: Comerford & Ubel (2013): Preference Choice for Effortful jobs
      • Express a preference for more effort
      • But choose jobs with less effort (lol)
      • In lab experiment
        • They set their own wages
        • Those whose wage demands got them effortless jobs: got lower enjoyment
      • Observation: Discrepancy between CHOICE and PREFERENCE
    • So, there is a discrepancy between
      • preference & choice
      • decision utility & experienced utility
      • choice & attitude (consideration of car may involve attributes of safety, fuel economy etc)
    • SDM approach involves two steps:
      • values for each attribute to some scale
      • weights for each attribute to compare them
      • e.g. loss of one mile to the gallon \(\sim\) two cubic feet of boot space
      • This procedure is more relevant in determining attitudes or judgements
    • For preference: most important attribute is weighed more heavily (maybe cause its a more convenient rationale for choice)
  • Prominence Effect
    • Explained by Somatic-Marker Hypothesis (SMH) proposed by neuroscientist, Damasio (1994)
    • Somatic-Marker is an unpleasant gut feeling, when a bad outcome connected to a response comes to mind
    • Automated alarm signal that forces attention to the negative outcome of a given actions "Behave: Danger ahead! If you choose this option".
  • The SMH
    • This auto signal makes you reject negative course of action and protects against future losses.... choose from fewer alternatives (better)
    • In context of prominence effect, Damasio observed... gut feeling can still lead to good and quick decisions
    • Visceral factors1... proposing that they can lead to better rather than worse decisions.
  • Cardinal & Ordinal utility
    • Early Economists:
      • "Utility can be quantitatively measured using utils", using a ratio scale with a zero point.
      • They can say, "Basket B yielded twice as much as utility as basket A"
      • Can even be considered interpersonally.
    • Disfavor cardinal utility for Ordinal (ranking system)
    • Interpersonal additions of utility is INVALID
    • Ordinal > Cardinal because:
      • fewer assumptions
      • Equilibrium conditions can be expressed in terms of ordinal utility
      • DMU Law can also be expressed
    • However NM uses cardinal, as it is more tractable mathematically.

  1. Powerful internal states that can affect our attention, motivation and behavior (order matters)... sometimes causing us to act against our rational, long-term interests. 

  2. Hedonistic simply means there is a variance in the way it is interpreted/experienced by different people.