Schürmann, O., Andraszewicz, S., & Rieskamp, J., Challenges of the
past and future directions for financial risk profile assessment.
In two experiments with test-retest data collection separetdy by one
month, we measured a number of the most relevat measures used for
assessing people's risk preferences. We evaluated test-retest
reliavility of each of these measures and we tested the consistency
among these scales. Next, we contrasted these results with a set of
six questions created based on the most desired features of the
existing scales and refined over a two-step test procedure. We
identify the main challenges related to assessment of people's risk
attitude and we propose future measures and direction.
Andraszewicz, S., Kaszás, B., Grosshans, D. , Zeisberger, S., Sornette, D., &
Hölscher, C. Influence of simulated market crashes on trading strategy and stress reaction.
In this project, we measure the effect of experiencing an exogenous market crash on the
trading strategies of novice traders. Participants see multiple trading sessions, lasting
about 2-3 minutes each, in which they view the price development of a risky asset. They can make
trades using quick-buy buttons at any time during the trading session, assuming that they have
enough cash for purchases and enough shares for selling. They observe online performance of their
portfolio, while the value of the shares changes. The players view real market historical
data that correspond to a 12-month period (i.e., 252 trading days). The players are informed
that they view the real-market data and that these correspond to the daily closing
prices. At the start of the trading session, the users are endowed with 1 000 000 units of Experimental
Francs and such that 50% of the units are distributed as cash and the remaining 50% of the units are
invested in the risky asset. After the end of the trading session, the user is informed about his/her
performance and can click a button to proceed to the next trading session. In order to avoid hedging,
one session is randomly chosen of the performed trading sessions and a bonus is paid out depending
on the performance in that session. The preliminary results reveal three major stylised
facts: 1) people who experienced a crash, generate lower earnings even in the "normal"
market conditions; 2) people who did not experience a crash learn to increase their earnings
over time; 3) when people experience an exogenous market crash, overall, they shift their
strategy such that they sell the risky asset right after the crash.
Sornette, D., Andraszewicz, S.,Wu, K., Murphy, M.O., Rindler, P., & Sanadgol, D. Resolving persistent
uncertainty by self-organized consensus to mitigate market bubbles.
We propose a new paradigm to study coordination in complex social systems,
such as financial markets, that accounts for fundamental uncertainty. This new
context has features from prediction markets that have been shown previously
to mitigate price bubbles in classical asset market experiments. Our setup is
more realistic as it others multiple securities that are continuously traded over
days and, importantly, there is no true underlying price. Nonetheless, the
market is designed such that its rationality can be evaluated. Quick consensus
emerges early yielding pronounced market bubbles. The overpricing diminishes
over time, indicating learning, but does not disappear completely. Traders' price
estimates become progressively more independent via a collective realization
of communal ignorance, pushing the market much closer to rationality, with
forecasts that are close to the realized outcomes. We tested this setup in three experiment:
two classroom experiments lasting four weeks each and one laboratory experiment that
replicated the classroom design within a much shorter time-frame (i.e. 10 minutes trading time).
We used a multi-use trading platform integrated in the
InnovWiki interactive platform developed by the
Chair of Entrepreneurial Risks of Prof. Sornette. In this platform, any number of users simultaneously
trade financial assets such that only one asset pays out the dividend equal 100 Experimental Francs,
while other securities will be valued at 0. Therefore, the point of the traders is to generate cash, using
market inefficiencies, and to hold as many shares of the security that pays out the dividend.
Wu, K., Andraszewicz, S., Sornette, D., Murphy, & Sanadgol, D. Behavioural experiment
Influence of of expert information and teaching of arbitrage strategy on price distribution.
This study uses the experimental setup of Sornette et al. (2016) to test two research questions.
First, we investigated the influence of the cascading effect of the information spread
among the market players. We distributed news to half of the students participating in
the market and we found that people with some trading experience anchor on the public
information, even if the information is obtained indirectly. This result was surprising
because the students were explicitly informed that the information may or may not be
accurate and they have already completed eight trading sessions in which they learned
about lack of predictability in the market. Second, we made participants be aware of the
overpricing in the market by providing them with the market index that should not exceed
the value of 100 to be rational. The overpricing diminished but was still present.
Some players managed to apply an arbitrage strategy in the situation of the market overpricing.
Andraszewicz, S., & Sornette, D. Mathematical description of the nature
of intrinsic probabilistic choice.
In this project, we investigate the mathematical derivation and evolutionary reasoning of
phenomena such as preference reversals, choice inconsistency and the probabilistic nature
of choice. We link our derivation to experimental results from previous studies and to the
well established choice models present in the literature for Judgment and Decision Making.
Klucharev, V., Andraszewicz, S., & Rieskamp, J. Neural underpinnings of the depletion of common
Why do people often exhaust unregulated common natural resources but
successfully sustain similar private resources? To answer this question the present work
combines a neurobiological, economic, and cognitive modeling approach. Using
functional magnetic resonance imaging we showed that sharp depletion of a common
(shared) and a private resource deactivated the ventral striatum, that is involved in the
valuation of outcomes. Across individuals the observed inhibition of the ventral striatum
negatively correlated with attempts to preserve the common resource, but the opposite
pattern was observed when individuals dealt with their private resource. The results
indicate that the basic neural value signals differentially modulate people's behavior in
response to the depletion of common versus private resources. The computational
modeling of the results suggests that the overharvesting of common resources is
facilitated by social comparison. Overall, the results could explain some aspects of
people's tendency to overexploit unregulated common natural resources.
Andraszewicz, S., von Helversen, B., & Rieskamp, J. Expected shortfall as a measure of risk
in risk-value models.
Models of decision making under risk propose different definitions of risk. The most
commonly used definition is variance of choice option's outcomes. Further developments of
the standard risk-value models incorporate skewness in the model specification. In contrast,
expected utility theory assumes people's risk aversion results from the curvature of the utility
curve. Here, we propose expected shortfall, as an established in finance measure of risk,
which provides a plausible psychological interpretation of risk, where a decision maker falls
short of their aspiration expected outcome. We integrate this measure in the standard riskvalue
models assume a trade-off between the expected gain and expected risk. We test the
new risk-value shortfall model against the existing models in two behavioral experiments.
Our results indicate that the proposed model can successfully predict people's preference for
options with the higher expected value, lower variance and more positively skewed
distribution. Also, we showed the advantage of the risk-value shortfall model over expected