Michal Lewandowski’s personal webpage
I conduct research on decision-making under risk, ambiguity, time delays, and strategic uncertainty. I am currently working as a PI in the OPUS research grant of the Polish National Science Center, measuring and characterizing attitudes toward uncertainty using indifference prices.
My research can be accesses via my Google Scholar, ResearchGate or ORCID profiles.
I am currenty working on several projects. Preliminary drafts are available below
Loss aversion or preference imprecision? What drives the WTA-WTP disparity? (submitted)
joint work with Łukasz Woźny and Michał Jakubczyk
Abstract: We propose a framework that, while eliminating the endowment effect, accounts for the two leading explanations of the disparity between willingness to accept (WTA) and willingness to pay (WTP): loss aversion and preference imprecision. Our approach applies to incomplete preferences under both risk and ambiguity. We introduce axioms that characterize the disparity and each of its two components. We show that the WTA-WTP disparity is a monetary measure (i.e., a premium) of uncertainty aversion, with hedging, rather than risk, serving as the neutrality benchmark. We derive comparative statics results for this measure with respect to both individuals and prospects. Our framework is general and encompasses several models, including multiple-utility multiple-prior models, as special cases.
Companion paper: Loss aversion or preference imprecision? What drives the WTA-WTP disparity? An experimental illustration
joint work with Łukasz Woźny and Michał Jakubczyk
Abstract: We report results of an experiment decomposing the disparity between willingness-to-pay (WTA) and willingness-to-pay (WTP) into two leading motives: loss aversion and preference imprecision. Our experiment is based on a method developed in a related paper Lewandowski M., Woźny Ł., Jakubczyk M., Loss aversion or preference imprecision? What drives the WTA-WTP disparity?
Biseparable representations of the certainty equivalents (submitted)
joint work with Jacek Chudziak
Abstract: We consider the following biseparable representation of the certainty equivalent: $F(x, y; p) = u^{−1}(w(p)u(x) + (1 − w(p))u(y))$, where $(x, y; p)$ is the binary monetary prospect, u is the utility function, and w is the probability weighting function. We provide a simple set of axioms characterizing this form for all binary prospects as well as for the subset of binary prospects, called simple prospects, in which one of the two payoffs is fixed. We consider both the case of general w and the case of expected utility, where w is the identity function. We discuss the extent to which such models can be identified, the issue of extending these models to a larger number of payoffs, and draw conclusions for model testing.
Individual and collective rationality in carpooling forthcoming in European Journal of Operational Research
joint work with Pawel Kalczynki
Abstract: We define carpooling as a coalition game and present a socially optimal solution that minimizes the overall cost of commuting and is stable and fair. Instead of transferring costs between players, individual rationality is achieved by appropriate composition and assignment of drivers within carpools. We develop a three-step solution procedure, where our final solution is based on the stable pre-nucleolus of the underlying game. The results of computational experiments show that our procedure guarantees substantial gains from carpooling. These gains increase with the number of commuters and are comparable to gains achieved by centralized systems, which ignore stability and fairness.
joint work with Mariusz Kozakiewicz, Bogusz Lewandowski, Pawel Kalczynski
Abstract: This paper presents a methodological and substantive critique of (Kontek K., 2025, Reassessing poland’s 2024 presidential runoff: Using spatially-grouped mad detection to recalculate the result. available at: https://ssrn.com/abstract=5296435.), which analyzes the impact of anomalies on the second round of Polish presidential election outcomes in 2025. We identify a key flaw: anomalies are detected only when favoring one candidate, with reverse cases omitted. First, we replicate the original results. Second, using the same method, we include reverse anomalies, showing how selective the original approach was. Third, we question the validity of the methodology and the conclusions drawn. We argue that the findings in the original study are methodologically biased and therefore unjustified. Keywords: outlier detection, robust statistics, election manipulations