The Augmented Behavioural Capacity (ABC) Model: Large Language Model Assistance and Calibrated Retail-Investor Decision-Making
Year: 2026
Type: preprint
Abstract: Retail investors increasingly bring large language models (LLMs) into cognitively demanding trading decisions - not as tools but as conversational reasoning partners that help frame problems, test scenarios, and judge feasibility. This creates a calibration problem with direct bearing on trading behaviour: the same assistance can yield nhancement or miscalibrated confidence, deference, and offloading. Existing adoption and intention theories explain why investors engage such systems but leave unspecified the decision-time mechanism through which dialogic assistance enters evaluative reasoning - a gap diagnosed here as a three-layer (temporal, ontological, phenomenological) insufficiency. The Augmented Behavioural Capacity (ABC) Model addresses it as a domain-bounded extension for LLM-assisted retail-investor decision-making, specifying a moderated process architecture of three elements: Capacity, operationalised through Perceived Cognitive Assistance (PCA); Calibration, a diagnostic overlay separating calibrated enhancement from miscalibrated confidence, deference, or offloading; and Choice, the proximal outcomes spanning complexity intention, strategy evaluation, structural complexity, and risk assertion. The method is theoretical model development - construct roles, boundary conditions, rival explanations, and falsifiable propositions. The contribution is bounded: ABC claims neither improved returns nor full causal validation, but offers a specified architecture, partially anchored by prior PCA validation, and a staged agenda for causal, calibration, longitudinal, and behavioural-footprint testing.
Source: Preprints.org
Authors: Dmitrii Gimmelberg, Iveta Ludviga
Language: English
Subfield: Finance
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7167697352