Bounded rationality Study Guide

📖 Core Concepts Bounded Rationality – Decision‑makers operate under limited information, limited cognitive capacity, and limited time, so they settle for a “good enough” (satisficing) option rather than the mathematically optimal one. Satisficing – Choosing an alternative that meets an adequacy threshold; stops the search once the threshold is reached. Heuristics – Simple decision rules (e.g., anchoring, availability, representativeness) that replace exhaustive optimization when resources are scarce. Adaptive Toolbox (Gigerenzer) – A collection of fast‑and‑frugal heuristics that exploit regularities in the environment and can outperform complex optimization under constraints. Rational Inattention (Sims) – Agents deliberately limit information acquisition because processing the information is costly. Bounded Willpower – Self‑control is finite; short‑term desires often dominate long‑term plans (hyperbolic discounting). Bounded Selfishness – Social preferences (fairness, reciprocity) exist, but the scope of concern for others is limited and shaped by group boundaries. --- 📌 Must Remember Decision limits come from problem difficulty, cognitive capacity, time pressure, information‑acquisition costs, and environmental structure. Simon’s core claim: Humans use heuristics, not full optimization, because of bounded rationality. Key heuristics: Anchoring & adjustment – start from a reference point, adjust insufficiently. Availability – over‑estimate vivid or recent events. Representativeness – judge by similarity, ignore base rates. Nudge principle: Change the choice architecture (e.g., defaults) to guide behavior without restricting freedom. Loss aversion: Losses loom larger than equivalent gains; drives framing effects. Hyperbolic discounting: Preference for immediate rewards declines steeply over short delays, then flattens. --- 🔄 Key Processes Satisficing Decision Cycle Define adequacy threshold → Search options → Stop when an option meets threshold → Implement. Heuristic Selection (Adaptive Toolbox) Identify environmental regularity → Choose matching fast‑and‑frugal heuristic → Apply → Evaluate outcome. Nudge Implementation Identify cognitive limitation → Design choice architecture (default, salient cue, simplified info) → Deploy → Monitor behavior shift. Rational Inattention Process Weigh cost of acquiring/processing information vs. expected benefit → Choose limited information set → Make decision based on that set. --- 🔍 Key Comparisons Satisficing vs. Optimizing – Satisficer stops at “good enough”; Optimizer exhaustively searches for the best possible outcome. Anchoring vs. Availability – Anchoring starts from a reference point; Availability relies on memory vividness. Default‑bias nudge vs. Information provision – Default nudges exploit inertia; information nudges rely on clearer data to improve choice quality. Bounded Willpower vs. Unbounded Willpower – Bounded willpower leads to preference reversals over time; unbounded willpower yields consistent long‑term plans. --- ⚠️ Common Misunderstandings “Bounded rationality = irrationality.” – It describes systematic, resource‑constrained decision making, not random error. Heuristics are always “bad.” – In the right environment, fast‑and‑frugal heuristics can outperform complex optimization. Nudges force choices. – Nudges preserve freedom; they only change the default or presentation. Loss aversion means people hate any loss. – It means relative weighting: a loss of \$100 feels worse than a gain of \$100 feels good. --- 🧠 Mental Models / Intuition “Cognitive bandwidth is a budget.” – Think of attention and processing power as a limited budget you must allocate across information sources. “Heuristics are shortcuts that work when the road is straight.” – When the environment is predictable, a simple rule gets you close to the optimum with far less effort. “Default = inertia.” – People tend to stick with the status quo; setting the desired outcome as the default leverages this inertia. --- 🚩 Exceptions & Edge Cases Complex, highly variable environments – Fast heuristics may fail; more deliberative analysis becomes necessary. When information costs are negligible (e.g., with AI tools) – The rationality bound expands, allowing closer to optimal optimization. Strong emotional framing – Extreme fear or excitement can override typical loss‑aversion patterns. --- 📍 When to Use Which Use a heuristic when: Time is short, information is costly, and the environment shows regular patterns. Apply a full optimization model when: Decision stakes are high, ample time and computational resources are available, and the problem structure is well‑defined. Deploy a default‑bias nudge when: The desired behavior aligns with the majority’s best interest and you can set the default without restricting alternatives. Provide clear information (e.g., calorie labels) when: The barrier is lack of salient data rather than cognitive overload. --- 👀 Patterns to Recognize “Cue‑driven voting” – Look for answer choices that mention reliance on party affiliation, appearance, or single‑issue cues. Framing switches – Same numbers presented as “gain” vs. “loss” often predict opposite choices. Hyperbolic discounting – Preference for a smaller‑soon reward over a larger‑later reward that disappears as the delay shortens. Default‑effect – Sharp jumps in participation rates when the default option changes (opt‑in vs. opt‑out). --- 🗂️ Exam Traps Distractor: “Bounded rationality means people are completely irrational.” – Wrong; it means decisions are systematically limited, not random. Distractor: “Heuristics always lead to errors.” – Incorrect; many heuristics are adaptive and can be optimal in certain settings. Distractor: “Nudges remove choice.” – Misleading; nudges preserve freedom, they only restructure the choice architecture. Distractor: “Loss aversion applies only to financial losses.” – Too narrow; loss aversion influences any domain where outcomes can be framed as gains vs. losses. ---