Analytical Proof of Bounded Rationality Theory on Unitas SACCO Members in Kenya: Bayesian Approach

Jomo Kenyatta

Abstract


The first of the four fundamental assumptions in
economics and standard finance is that humans are rational
actors, which does not hold good all the time. This anomaly
begot behavioural finance that recognizes instances of irrational
decision making in human beings. Rationality bounds in
financial decision making as espoused in bounded rationality
theory need to be determined to reflect how humans actually
behave rather than how they should behave; to pave way for
modification of classical economics and standard finance
theories. This analytical proof of bounded rationality utilizes
LOT-R parameterized cumulative prospect theory decision
weights function to transform subjective to objective
probabilities. Human intrinsic perceptions and self proclaimed
prospects measured on a 9 point Likert scale are converted into
subjective probabilities. I construct a rationality measuring
instrument on a 0 – 1 scale using the probabilities of making a
rational decision, of observing economic well being increase
after an irrational decision and that of observing economic well
being increase after a rational decision as inputs. Thereafter, I
show that the multiperiod model forms an absolutely converging
sequence in the open interval (0, 1), hence bounded. The
instrument is a Bayesian learning model whereby wealth
movement is the observable dimension variable and the
rationality to be determined is the unobservable dimension
variable. Stochastic discrete time case in a binomial setting is
explored. The concept of entropy from the second law of
thermodynamics in physical chemistry – information theory
version is expected to feature prominently and to guide
recommendations likely to yield global lessons.

Full Text:

PDF