Affective forecasting

The affective forecasting literature documents the systematic errors humans make when predicting their future emotional states. We consistently overestimate both the intensity and duration of our emotional reactions to events—a phenomenon known as the impact bias. We also fail to account for our remarkable capacity for emotional adaptation, which causes us to return to a baseline level of happiness far faster than we anticipate. The result is a persistent misalignment between the choices we make (based on predicted affect) and the experiences we actually have.

These errors are not merely individual quirks; they scale to institutional failure. Policymakers, marketers, and designers all rely on affective forecasts—whether explicit or implicit—to shape decisions that affect millions. When the forecasts are systematically wrong, the systems built on them inherit the error. Affect is not merely personal; it is infrastructural.

Network Contagion and Collective Error

The impact bias does not remain confined to individual minds. It propagates through social networks via emotional contagion—the phenomenon by which one person's emotional state influences the emotional states of others. When a community collectively overestimates the emotional impact of an event, the resulting shared misforecast can drive behaviors that amplify the very outcomes being predicted. Financial markets exhibit this dynamics in bubbles and panics: traders forecast their own future regret, overestimate its intensity, and trade accordingly, producing self-fulfilling cascades of fear or euphoria.

The network structure of affective forecasting matters. In tightly clustered networks, forecasting errors reinforce each other through echo-chamber dynamics; in bridge-rich networks, errors diffuse but also encounter corrective information. The small-world property of human social networks—high clustering with short path lengths—creates conditions where local consensus on emotional outcomes can rapidly globalize before encountering disconfirming evidence. Affective forecasting is not merely a cognitive bias; it is a network dynamical process with its own threshold behavior and phase transitions.

Institutional Amplification

Institutions do not merely inherit individual affective forecasting errors; they amplify them through organizational structure. Bureaucracies aggregate forecasts upward, and at each level of aggregation, the systematic components of error (overestimation of impact, underestimation of adaptation) are reinforced while the random components cancel. The result is institutional decisions that are confidently wrong in predictable directions. Urban planners overestimate the negative affect of density; policymakers overestimate the positive affect of tax cuts; product designers overestimate the hedonic return of feature proliferation.

Prediction markets were proposed as a corrective: aggregating many individual forecasts should produce wisdom-of-crowds accuracy. But prediction markets forecast outcomes, not affective reactions to outcomes. A market can correctly predict that a policy will pass while failing to predict that the policy's passage will produce less happiness than its supporters anticipated. The gap between outcome prediction and affective prediction is a blind spot in institutional decision-making that no voting system or market mechanism currently addresses.

The Measurement Problem

Affective forecasting research relies on self-report: subjects predict their future feelings and later report their actual feelings. But self-report is itself a constructive process, shaped by present context, memory reconstruction, and social desirability. The claim that humans are systematically wrong about their future emotions assumes that there is a true future emotional state against which the forecast can be measured—a claim that enactivists and constructionists about emotion would reject. If emotions are not pre-formed states waiting to be experienced but emergent properties of situated interaction, then the very concept of a forecasting error may be poorly posed.

This does not mean affective forecasting research is meaningless. It means its findings are better interpreted as evidence about the dynamics of affective representation—how humans construct narratives about their future selves—than as evidence about the accuracy of emotional prediction per se. The impact bias may tell us less about emotional futures and more about the present-tense stories we tell ourselves to justify our choices.

The affective forecasting literature treats its subject as a cognitive bias to be corrected—through better information, more realistic expectations, or nudges toward hedonic neutrality. This framing misses the structural function of affective misforecasting. Overestimating the emotional impact of choices is not a bug in human decision-making; it is a feature that produces the very motivation required to make choices at all. Without the impact bias, humans might never leave bad situations, never take risks, never invest in the future. The bias is the engine of change, and the literature's aspiration to eliminate it is an aspiration to eliminate the friction that makes human lives stories rather than equilibrium states.