ECO331: Behavioural & Experimental Economics - Mental Accounting and Myopic Loss Aversion

Overview

• Topic: Mental Accounting and Myopic Loss Aversion in Behavioural & Experimental Economics (ECO331)

• Main plan: summarize key ideas from Gazzale’s slides, aligned with Thaler’s view that mental accounting matters for decisions under risk, income measurement, and real-world behavior

• Key links:

  • Mental accounting explains sunk cost fallacy

  • Hedonic editing interacts with prospect theory

  • Mental accounting combines with loss aversion to produce myopic loss aversion

  • Relevance to labour economics and tax/transfer settings

• Core takeaway: Decisions depend on how people partition wealth into separate accounts, not just on total wealth; this partitioning interacts with framing, time, and perceived gains/losses to shape risk-taking and consumption choices

What mental accounting is and why it matters

• Mental accounting is a set of cognitive operations people use to organize, evaluate, and keep track of financial activities

• Key questions about accounts:

  • What counts as being in the same account?

  • Items in the same category (e.g., clothing) or same size/frequency may be treated as twins in the same account

  • When are accounts opened and closed?

  • A "closing" of an account registers utility and can affect next decisions

  • How do accounts affect attitudes toward risk?

• Formal framing comment: homo economicus would treat wealth in a single net-benefit account, but actual behavior uses multiple mental accounts

The mental accounting framework and big ideas

• Mental accounting helps explain sunk cost fallacy: continuing a project due to past costs already sunk

• It also relates to hedonic editing (and prospect theory): how people segment gains/losses to maximize utility

• Core equation (conceptual): \text{MLA} = \text{Mental Accounting} + \text{Loss Aversion}

  • Myopic Loss Aversion (MLA) combines mental accounting with loss aversion to explain risk-averse behavior in many real-world settings

• Labour economics: mental accounting can shape work-leisure and wage-taking decisions under different framing and benefit structures

Mental accounting and risk aversion

• Classical prediction (homo economicus): risk attitudes are derived from expected utility; small gambles should be near risk-neutral

• Reality: risk aversion is pervasive and often not explained by EUT alone, pointing to narrow framing and mental accounting effects

• Myopic loss aversion: the combination of mental accounting and loss aversion helps explain why people avoid small risks repeatedly over time

Anomalies and example illustrations

• Example 1: wealth and consumption vs small wealth changes

  • If you leave an apartment with $50 in wealth and face a small wealth decrement or a ticket cost, your willingness to pay and your decision can depend on whether the loss is seen as coming from a particular account or from overall wealth

• Example 2: willingness to walk for money illustrates how context and framing alter perceived value

  • Willing to walk 5 blocks for $5? contrasts with other small incentives (e.g., a $10 discount for buying at current store vs a $5 discount 5 blocks away)

• In both examples, the same total payoff can be valued differently depending on the mental account in which it is embedded

Real-world anomalies from the U.S. context (real-world evidence)

• I. Tax rebates vs withholding during stimulus periods

  • Prediction under single account (homo economicus): rebates and tax withholding changes are equally stimulating

  • Reality with credit constraints: rebates are often saved more than spent

  • Early evidence: Shapiro & Slemrod (2009): MPC about 0.33; low-income households use rebates to pay down debt

  • Parker, Souleles, Johnson, McClelland (2011): MPC about 0.7 with durable purchases; withholding changes can shift consumption more strongly (MPC ≈ 1)

  • 2009 American Recovery and Reinvestment Act: primary tax cut operated through reduced withholding, affecting savings/spending behavior

• II. SNAP (food stamps) benefits

  • SNAP benefits are restricted to groceries; unlike cash, some predict effects should mimic cash transfers

  • Evidence (Hastings & Shapiro 2018; Hoynes, McGranahan, Schanzenbach 2015): ~84% of recipients use benefits for groceries; restricting use can alter perceived value and utilization

• Implication: framing and restriction of benefits alter consumption responses beyond what purely cash-equivalence would predict

Narrow categories and relativity in costs and incomes

• Narrow accounts: empirical evidence from Kooreman (2000)

  • Two income accounts: (i) Dutch child-benefit system, (ii) other income

  • MPC for children’s clothing with child-benefit: ~0.12; for adult clothing: ~−0.02

  • MPC for children’s clothing from other income: ~0.01; for adult clothing: ~0.04

• Mental accounting and relativity: how costs are described affects choices

  • Framing: costs described as “for only the cost of a cup of coffee a day” or “Dollar-a-Day club” tend to be perceived differently

  • Gourville (1998): costs are compared to expenditures of similar size/frequency (different accounts) rather than to large, infrequent costs

  • Pennies a Day: ongoing small costs are perceived differently than large, one-time expenditures (vacations, expensive outfits) when compared within different mental accounts

• Implication: aggregate one-time costs vs. ongoing small costs can lead to different decisions depending on comparator sets and accounts

Temporal framing and donation behavior (Figurative illustration)

• Gourville (1998) study: effects of temporal frame and request amount on donation likelihood

  • Donor decisions vary by whether the ask is framed as a small daily amount or larger one-time ask

  • Example figures show different donation likelihoods for various frames and amounts, highlighting the role of temporal framing in charitable giving

• Takeaway: small, frequent charges are perceived and evaluated differently than large, infrequent charges, due to separate mental accounts

Income side and permanence of income

• Permanent-income hypothesis: people smooth consumption over time, implying one MPC across income channels

  • Conceptual basis: consumers allocate spending based on expected lifetime resources, not just current income

• Landsberger (1966): German restitution payments to Israelis in the late 1950s; observed MPC heterogeneity across payment sizes

  • Israeli MPC was approximately
    \text{MPC}_{\text{Israeli}} \approx 0.75

  • Small payments: MPC ≈ 2; Medium payments: MPC ≈ 0.5–0.6; Largest payments: MPC ≈ 0.2

  • This illustrates strong variability in MPC by the size of transfer and the perceived significance of each payment within separate accounts

• Eco331 takeaway: income is partitioned into multiple accounts with different apparent MPCs; the way income is allocated matters for consumption choices

Mental accounting of income: windfalls and how to spend them

• Classic eco-331 example: how to spend a $500 windfall when it arrives via different channels (work bonus vs government refund)

  • Predicted choices: debt repayment, savings, or consumption (splurges)

  • The same dollar windfall can be allocated differently depending on its source due to mental accounting

• Implication: the source of funds can influence how they are allocated, even if the economic impact on wealth is identical

Is mental accounting good or bad?

• Costs of narrow bracketing

  • Narrow bracketing can lead to suboptimal decisions if important interdependencies across accounts are ignored

• Motivated bracketing

  • People may deliberately frame or bracket decisions in a way that supports a desired outcome or reduces cognitive effort, potentially improving or impairing welfare depending on context

Mental accounting across time: opening and closing accounts

• Hedonic editing (creative accounting): how outcomes are classified within accounts to maximize utility

• Prospect theory and the breakdown of gains/losses when organizing them into accounts

• Central ideas:

  • Break up gains into smaller gains to feel better about them

  • Offset small losses with larger gains

  • Segregate big losses from small gains to avoid cross-account negative feelings

  • Aggregate losses to reduce perceived impact

• Practical insight: how and when to open/close accounts can shape risk-taking and happiness through hedonic editing

Opening/closing accounts and the reluctance to realize losses

• If we do not aggregate losses, we should avoid realizing them; people delay closing loss-making accounts to avoid locking in a loss

• Odean (1998): investors tend to sell winning stocks more readily than losing stocks

  • They realize gains by selling winners but hold on to losers, which can hurt overall performance

• Consequence: investment behavior is influenced by how gains and losses are partitioned into separate accounts rather than by the total net wealth change

Sunk costs and mental accounting

• Mental accounting as an explanation for sunk cost bias: people prefer to realize future benefits within the same account as past costs to avoid wasting sunk costs

• Hypothetical example: concert during a snowstorm

  • Would you go to the concert if you had bought the ticket? What if you received the ticket for free?

• Gourville & Soman (1998): empirical work on gym attendance and the timing of semiannual dues supports the link between sunk costs and current decisions

• Implication: sunk costs lead to continued investment even when marginal benefits are negative if those costs are embedded in the same mental account as expected future gains

Sunk costs: broad vs narrow framing

• Broad frame: lifetime earnings scenarios (e.g., $4,000,010 for certain vs a mixed lottery-like scenario)

• Narrow frame: small, incremental outcomes (e.g., $10 certain vs $0 with 0.5 probability or $21 with 0.5 probability)

• Mental accounting predicts people may react differently under broad vs narrow framing due to how each frame maps onto distinct accounts and reference points

Mental accounting and risk aversion: revisiting homo economicus

• The simple EUT explanation cannot capture most risk-averse behavior observed in real life

• Narrow framing suggests that people evaluate small gambles in isolation, rather than in aggregate with overall wealth

• Arrow (1971): EUT predicts near risk neutrality for small gambles; real-world deviations imply non-trivial risk preferences under mental accounting

Rabin (2000) – risk aversion and diminishing marginal utility

• Setup: consider a payoff where a person prefers certain $10 over a 50% chance of $21, and this preference extends as wealth grows

  • The plan rejection (choice) informs how MU declines with wealth changes

• The key idea: as wealth W rises, the marginal utility of an additional lump sum declines; the incremental gain of $21 becomes less appealing, reinforcing risk aversion

• Iterated rejections show that MU declines rapidly as wealth increases, implying extremely high risk aversion for larger gains

• Result: If MU(W) = 1000, then MU(W+1000) ≈ 10 and MU(W+2000) ≈ 0.1, illustrating very steep declines in marginal utility with wealth

• Takeaway: This helps explain why people exhibit strong risk aversion in real-world decisions and supports the need for non-EUT explanations like prospect theory and mental accounting

Myopic loss aversion and the integration with prospect theory

• Myopic loss aversion combines mental accounting with loss aversion from prospect theory to explain a wide range of observed risk-averse behaviors

• Practical implication: decision-makers evaluate prospects after considering both the chance of gains/losses and how outcomes are partitioned into separate mental accounts

• This framework reconciles why people may reject a certain loss in one context but accept a larger risk in another when framed differently

Mental accounting and prospect theory: Samuelson’s bet (illustrative example)

• Classic bet: a 50-50 gamble where you can lose $100 or gain $200; if you reject this, you face a similar but different gamble

• EUT would reject both, but real people may reject the sure loss while accepting larger risky gambles when considered as separate outcomes

• The challenge arises if you assess each outcome in isolation rather than as part of a single portfolio

• Overall takeaway: prospect theory combined with mental accounting explains why people exhibit seemingly inconsistent risk preferences depending on framing and account structure

Practical and ethical implications

• Framing effects and account design can influence consumer welfare and policy effectiveness

• Tax policy, rebates, and transfer programs can be made more or less effective by considering mental accounting (e.g., restricting the use of rebates, timing withholding changes, or packaging stimulus messages differently)

• Financial education should address how people partition wealth to help them make more globally optimal decisions rather than suboptimal, account-specific choices

• Ethical considerations: policymakers and firms should be transparent about how framing and account structures influence choices to avoid manipulation and to promote welfare

Summary of key concepts and references

• Mental accounting: organizing, evaluating, and keeping track of financial activities across multiple accounts

• Sunk costs: bias arising when past costs influence current decisions within the same account

• Hedonic editing: reclassifying outcomes to maximize perceived utility within frames

• Prospective theory: non-linear value function and loss aversion shaping risk preferences

• Myopic loss aversion: mental accounting + loss aversion leading to heightened aversion to risk when decisions are evaluated over time

• Real-world evidence: rebates, withholding, SNAP benefits, and windfall allocations differ from simple cash-equivalence predictions

• Narrow framing vs broad framing: decision outcomes depend on how problems are framed across time and accounts

• Important empirical figures and references:

  • Shapiro & Slemrod (2009): MPC ≈ 0.33; rebates used to pay debts; low-income groups

  • Parker, Souleles, Johnson, McClelland (2011): MPC ≈ 0.7 for durable purchases; changes in withholding magnify spending response; ARRA 2009 effect through withholding

  • Hastings & Shapiro (2018); Hoynes, McGranahan, Schanzenbach (2015): SNAP benefits usage; 84% use for groceries

  • Kooreman (2000): MPC by categories of income and clothing in Dutch data

  • Gourville (1998): Pennies-a-Day; temporal framing in donations

  • Landsberger (1966): restitution payments and MPC heterogeneity

  • Odean (1998): disposition effect – selling winners vs losers

  • Rabin (2000): diminishing marginal utility and risk aversion dynamics

  • Samuelson’s bet: illustration of risk preferences under isolation vs aggregation

Equations and key formulas (LaTeX)

• Myopic loss aversion concept
  ext{MLA} = ext{Mental Accounting} + ext{Loss Aversion}

• Permanent-income hypothesis (conceptual) MPC = \frac{dC}{dY}

  • The idea that consumption smoothing leads to a stable MPC across income components

• Risk aversion and marginal utility (Rabin 2000 intuition)

  • If MU declines rapidly with wealth, then larger wealth increments have tiny marginal utility, leading to strong risk aversion for larger gambles

  • With a simple illustration: if MU(W)\gg MU(W+1), small increments in wealth are valued very differently depending on current wealth

• Framing and accounting: no single explicit risk-utility formula; conceptually, gains/losses are allocated to separate accounts and evaluated using adjusted utility functions per account

Connections to other lectures and real-world relevance

• Links to: sunk cost fallacy, risk preferences, and intertemporal choice

• Real-world relevance: consumer finance, tax policy, social welfare programs (rebates, withholdings, SNAP), and financial markets (investor behavior via hedonic editing and account management)

• Ethical and practical relevance: understanding mental accounting helps design better policies and financial products that align with actual decision processes rather than idealized rationality