EF Online Talks 11/2021

We finished four sessions of our first series of  EF Online Talks,  We enjoyed the talks which partially allowed for taking new perspectives in the area of (experimental) finance.
Thanks to all the speakers and their excellent talks. 

Elena Asparouhova (Incoming President), Luba Petersen (Conference Officer, North America),
Stefan Palan, Anita Kopanyi-Peuker,  and Sascha Fullbrunn (Managing Board)

Topic:  Choice Process Measures
November 11th, start at 8:00am (PDT) = 5:00pm (CET), Chair: Elena Asparouhova

Topic: Perception in Finance
November 11th, start at 10:00am (PDT) = 7:00pm (CET), Chair: Stefan Palan

Topic: Macro-Finance
November 17th, start at 8:00am (PDT) = 5:00pm (CET), Chair: Luba Petersen

Topics: Asset Markets
November 17th, start at 10:00am (PDT) = 7:00pm (CET): Asset Markets. Chair: Sascha Füllbrunn 


Sessions Nov 11th, Choice Process Measures

Decisions under risk are typically modeled as the interaction of stable preferences with payoffs and rewards. However, high stakes and uncertainty can also increase autonomic arousal. Arousal causes changes widespread in attention and cognition, and hence might affect not only attitudes towards risk but also the way decisions are made. This talk will cover recent research on autonomic arousal, attention, and the process of making decisions under risk.  

When people make decisions they often evaluate and compare the options, trying to discern which one is best.  This process can lead to hesitation when the best option is unclear.  Here I describe a model of the evaluation and comparison process and demonstrate how that model accounts both for what people choose and how long it takes them to choose.  This framework allows us to infer strength of preference from response times (RT).  There are also physical manifestations of this choice process.  Using mouse tracking we can measure how clearly a person prefers one option over the other.  This mouse-tracking data provides additional information on preferences, beyond choices and RT.  For example, the mouse trajectory from a single risky choice trial is highly correlated with a person's level of risk aversion, as measured from a full battery of choices.  Overall, these results illustrate the usefulness of process data for better understanding peoples' preferences.  

We apply a machine-learning algorithm, calibrated using general human vision, to predict visual salience of prices of stock price charts. We hypothesize that visual salience of adjacent prices increases decision weights on returns computed from those prices. We analyze the inferred impact of these weights in two experimental studies that use either historical price charts or simpler artificial sequences. We find that decision weights derived from visual salience are associated with experimental investments. The predictability is not subsumed by statistical features, and goes beyond established models.

Sessions Nov 11th, Perception in Finance

In three well-powered experiments (N=905), I study an application of the affect heuristic to the perception of financial assets. The affect heuristic predicts that people derive expectations of return and risk from a global affective impression of an asset, which leads to negative risk-return correlations. Experimental results confirm the general presence of an affect heuristic when evaluating individual stocks. Negative risk-return correlations emerge for pooled results on stock level, as well as within individuals. A weaker effect for asset classes suggests that the ease to recognize a positive risk-return correlation can curb the use of an affect heuristic. Asking for required returns instead of expected returns also makes the risk-return trade-off more salient. Allowing people to select their most liked or disliked stocks, however, exacerbates the use of the affect heuristic as participants presumably select stocks they feel most strongly about. 

We experimentally study how presentation formats for return distributions affect investors' diversification choices. We find that sampling returns alleviates correlation neglect and constitutes an effective way to improve financial decisions. When participants get a description of the probabilities for outcomes of the joint return distribution, we confirm the findings of others that investors neglect the correlation between assets in their diversification choices. However, when participants sample from the joint distribution, they change their allocation between two assets in response to a change in their correlation in the predicted direction. The results are robust across two experiments that have participants with varying experience (students vs. private investors). (with Christine Laudenbach and Martin Weber)

We study four fundamental components of financial agency settings: The perception of commonly used investment profile terminology, agents’ customization of portfolios to clients’ preferences, the effect of agents’ and clients’ preferences on investment levels, and the role of compensation schemes. We observe large heterogeneity in the perception of investment profiles, resulting in substantial miscommunication between clients and agents. Financial agents show a high willingness to implement their clients’ preferred investment profiles, yet appear to fail because of deviating perceptions. Agents’ own investment preferences matter, but take a back seat to clients’ preferences in determining investment shares. Different monetary incentive schemes hardly affect behavior. Our results point to moral constraints that limit agents’ discretion in the agency situation.

Session Nov 17th, Macro-Finance

We investigate the experimental environment where packages of goods can be acquired using dollars and bitcoin. The packages of goods have different transaction costs in terms of each currency. Trading takes place via two continuous double auctions, one for each currency. We ran three treatments, (1) with constant supply of both currencies,(2) with constant supply of both currencies and no transaction cost, and with increasing dollar money supply. In the first treatment, after initial adaptive dynamics, experimental economies converge to the equilibrium prices and packages traded in each currency. In the second treatment, we observe excessive volatility, which is in line with the theoretical prediction of indeterminacy of the exchange rate. Finally, in the third treatment, after the initial stage where subjects tend to 'hoard' the currency whose supply is increasing, there is a gradual adjustment towards the equilibrium real money holdings of each of the two currencies. 

We report on an experiment studying market reactions to stock splits and reverse splits. In the first environment, two assets have increasing fundamental values, and one asset is subject to a 2-for-1 share split while the other is not. In the second environment, the fundamental values of both assets are decreasing, and one asset is subject to a 1-for-2 reverse split while the other is not. We find that share prices do not fully adjust to changes in fundamental values per share following both types of splits and we relate this phenomenon to difficulties that traders have with proportional thinking.  (with Jean Paul Rabanal and Olga Rudd)

We propose a new experimental approach to assess inflation expectations anchoring using “strategic surveys.” Namely, we measure households’ revisions in long-run inflation expectations after they are presented with different economic scenarios. This approach has a causal interpretation and maps directly into policy makers concerns. We implement the method in the summer of 2019 and the spring-summer of 2021 when the anchoring of long-run inflation expectations was questioned. We find that the risk of un-anchoring was reasonably low in both periods, and that long-run inflation expectations were essentially as well anchored in August 2021 as in July 2019, before the Covid-19. 

Session Nov 17th, Asset Markets

How does public information regarding an asset affect the attention investors pay to their own private information? Our paper investigates by constructing a simple bond market in the laboratory. Human investors observe public information on default probability and then decide how much to invest in private information on the possible recovery rate in case of default. We compare investment choices, price efficiency and other bond market outcomes across different levels of default probability and across several different market formats. The data so far suggest that the rational inattention model does a fairly good job of predicting average private information investment and greater dispersion of investment for the Call market and continuous double auction market formats. It also has predictive power for price volatility and market liquidity. (with Grace Weishi Gu and Vivian Juehui Zheng)

Discussions about insider trading regulation often veer between the poles of forbidding insider trading to protect public confidence in the stock market's integrity, and allowing insider trading to foster the informational efficiency of prices. We study traders' preferences directly by offering them concurrent markets with different regulatory regimes in an experimental setting. We find that informed traders' preference for the unregulated market causes both informed and uninformed traders to be more active in the unregulated market. This market, thus, sees more trading volume, lower spreads and less mispricing. Nevertheless, uninformed traders suffer greater losses in unregulated markets, while informed traders profit from the lack of regulation. (with Robert Merl, Stefan Palan, and Dominik Schmidt)