Behavioral Finance an der Universität Marburg

- Risk elimination has a particular value

- Agents focus on levels of outcomes instead of probabilites of occurence

- Agents are influenced by decision frames

- Agents relate decision outcomes to reference points

- Agents exhibit loss aversion

- Agents have differing risk attitudes in gains and losses

- Prospect Theory distinguishes two phases in the choice process

1. Editing phase: Organizing and reformulating choice options (prospects) in order to simplify subsequent valuation

2. Evalution phase: Assessment of Prospects

- Changes of welfare defined to a reference point

- Evaluation of prospects based on decision weights, not on probabilites of occurence

--> Transformation of probabilites

- Hypothesis: People maximize their expected utility instead of Expected Value of the sum of the money
- CE=U^-1(E(U(X)))
- Risk Premium = EX - CE
- Bernoullis reasonable utility function for this: U=ln(x)

- Game: Fair coin tossed at each stage. First time tail appears, game ends and player wins the pot. Willingness to pay should be any price, because theoretically infinite Game

- If two lotteries are combined with a third lottery L, where L has a probability of (1-p), then the preference relation between L1 and L2 is identical regardless of whether L1 and L2 are evaluated in isolation or as parts of combined lotteries

- complex may be stripped of their common components, without affecting the preference relation between these lotteries

- most deviations from EUT are violations of the independence axiom

- Completely specified set of procedures, including instructions, incentives and rule of play

Between Subject-Design:
- division in 2 subgroups, each assigned with a different treatment

- benefit: no learning effect 

- downside: subgroups may vary systematically regarding certain characteristics

Within Subject-Design:
- each member of a cohort receives a number of different treatments

- benefit and downside vice versa

- Possibility to repeat a given setup without having to account for a host of confounding effects

- Permits replication of results by other researchers

- allows for the possibility to conduct a given experiment in different cultures and countries

- written instructions helps presenting ambigous terminology

- experiments should capture real-world decision behavior (typically monetary incentives)

- stylized facts: 

1. Individuals are more generous when no real money is involved

2. Risk aversion increases considerably in the amount of real money at stake

- possibility to control for extraneous factors

- mentioning of negatively connotated terms can be avoided (important feature regarding test, e.g. pollution)

- real-world decision problems typically feature mulitple channels of influence

- experiments provide setup in which a single determinant is being analyzed

--> allows for an isolation of specific determinants

Income and wealth:

- Affluent individuals are less risk averse

Gender:

- Women are less risk averse when stakes are low, but no gender effects if stakes are high

Findings for Germany:

- Men, younger and taller individuals, as well as highly educated people are less risk averse

Certainty effect:

- certainty may be more desirable than higher expected utility

- EUT implies that individuals can subsitute a sure payoff with a lottery

Small Probability Events: 

- Subjects can be guided by either a probability or level of gains

- Subjects do not correctly perceive differences due to small absolute probabilities

Isolation effect:

- Common components of lotteries are being disregarded, instead focus on their distinctive components

Framing effect:

- Different formulations of choice options has a significant impact on the decision outcome

Reflection effect:

- Risk aversion in gains and risk seeking in losses

Reference point effect:

- EUT focuses on states of wealth

- Empirical evidence values changes of wealth relative to a reference point

- Differing reference points lead to different decision outcomes

- developed by von Neumann and Morgenstern

- utility function assigns utility value to each outcome

- the higher a given alternatives expected utility, the stronger the preference for this alternative

- no utility measurement, rather ranking of alternatives