Anastasiia Morozova

We introduce a general test of capacity-constrained learning models. Learning has capacity constraints when the set of possible ways to learn is exogenously fixed, as in the widely used fixed-capacity versions of rational inattention (Sims 2003) and efficient coding (Woodford 2012). With such models, changes in incentives do not alter the extent of attention, only how individuals decide to allocate their scarce attention. We show that choice data are consistent with capacity-constrained learning if and only if they satisfy a No Improving (Action or Attention) Switches (NIS) condition. Based on existing experiments in which the incentives for being correct are varied, we find strong evidence that participants fail NIS for a wide range of standard perceptual tasks: identifying the proportion of ball colors, recognizing shapes, and counting the number of balls. We further show that violations of NIS occur systematically in response to higher incentives, suggesting that incentives often expand attention beyond what capacity-constrained models allow. However, we find that this is not true for all existing perceptual tasks in the literature, which offers insights into settings where we do or do not expect incentives to impact the extent of attention.

In a novel lab-in-the-field experiment within a real company, I leverage endogenous social and professional relations to assess the overweighting/under-weighting of others’ signals and beliefs relative to the hierarchical distance, tenure, perceived ability, professional prestige, and social distance of the source of the information. The design allows me to distinguish heuristic salience-based distortion from belief-based ability assessments and from strategic deference. Formal hierarchical measures, despite being prevalent in the literature, do not universally distort belief updating. By contrast, organic elements seem to matter. Other signals influence guesses more if their source belongs to the participant’s professional advice or strategic information network, consistent with an attention-based mechanism. Participants also respond more strongly to guesses of partners they perceive as higher ability. Network simulations keeping the organizational chart constant imply a 62% information loss and underscore the importance of endogenous social determinants of belief updating for information transmission quality, typically ignored in organizational economics research.

Recent surveys suggest that younger Americans are shifting toward riskier assets — cryptocurrencies, sports betting, prediction markets. We hypothesize that selective disclosure on financial social media contributes to this shift: spectacular gains and dramatic losses are amplified while ordinary returns go unshared, distorting the advice available to retail investors. We study both the supply and demand sides of retail financial advice in a laboratory experiment. In Individual sessions, conducted first, subjects complete a portfolio allocation task and choose whether to supply financial advice. In Social sessions, subjects observe the resulting leaderboard and are given the option to imitate. We study endogenous selection into both the supply of financial advice and the consumption of that advice through imitation, and how each interacts with task complexity and performance — the resulting portfolio selection and risk. Our main hypothesis is that limited understanding of market fundamentals, coupled with selective disclosure that goes unappreciated, results in excessive risk-taking on the part of consumers of retail financial advice.