«Kavitha Ranganathan* Srinivas Prakhya** *Kavitha Ranganathan is Research Officer at National Institute of Securities Markets (NISM), Mumbai; Email: ...»
Measuring Risk Attitudes and Personal Values: The Bounded Rationality Approach
*Kavitha Ranganathan is Research Officer at National Institute of Securities Markets (NISM),
Mumbai; Email: firstname.lastname@example.org. Prof. Srinivas Prakhya is Associate Professor at
Indian Institute of Management, Bangalore (IIMB); Email: email@example.com.
Measuring Risk Attitudes and Personal Values – The Bounded Rationality Approach 2
Measuring Risk Attitudes and Personal Values:
The Bounded Rationality Approach
JEL Classification: C91, G11, D81, A13 Keywords: Bounded Rationality, Investment Decisions, Decision under risk, Personal Values Measuring Risk Attitudes and Personal Values – The Bounded Rationality Approach 3
1. Introduction Attitude to risk is an important determinant of individual choice and decision making process.
People routinely make decisions in incomplete information, while most economic theory is built on the preposition that individuals have rational expectations and incorporate all information in an unbiased and coherent fashion. Individuals have different perception of risk because of their different interpretation of reality. Attitude to risk is formally modeled as the shape of the decision makers’ utility function (Keeney and Raiffa, 1976). The axiomatic approach of expected utility assumes decision maker has well defined preferences, follows a set of rationality axioms and evaluates all possible decision to choose the one that scores the best. However, the empathetic stream of behavioral finance accommodates deviations from rational expectations and takes into account limitations of knowledge, cognitive issues, behavioral biases and emotional factors. While it might be useful to measure attitudes to risk, it would also be more worthwhile to understand antecedents and correlates to risk. Research studies have identified a wide range of factors, some exogenous and some endogenous, that influence attitudes to risk.
Existing research supports the influence of exogenous factors such as socio-demographic variables on risk perceptions (Hartog, Ferre-i-Carbonnell and Jonker, 2002). However, attitude to risk is also recognized as a highly
constellation of psychological attributes, influenced by past experiences, beliefs, emotions (Loewenstein et al. 2001), personality (Nicholson et al.
2002)), the context and presentation formats (Kahneman & Tversky, 2000) which characterize real environments. Hence, there is a need to study financial behaviour patterns, rational or irrational, to reduce the distance between economic theories and the actual behavior that does not appear to be sufficiently linked to principles of rationality (Thaler, 1992; Shefrin, 1999; Shiller, 2000).
In this study, we examine attitudes to risk among individual investors and study the influence of personal value systems and demographics. We explore more natural ways to calibrate attitudes to risk using the bounded rationality approach (Simon, 1982). In an experimental study coined ‘Riskitude’; subjects invest in a portfolio that contains a risk-free but profitable bond and a risky asset with high or low return states. We predict the portfolio allocation by interrelating aspiration data of investors in low and high return states. The aspiration data also facilitates to calibrate a unique risk co-efficient that captures the degree of risk aversion. We further categorize investors Measuring Risk Attitudes and Personal Values – The Bounded Rationality Approach 4 based on their degree of risk aversion and importance to personal values (SVS; Schwartz et al.,
2001) such as conformity, power, achievement, stimulation and more. Personal values systems act as standards of conduct (Kluckhohn, 1951; Meglino and Ravlin, 1998) and influence decision making process. The study attempts to answer the question; do values like security, conformity, achievement, stimulation or more segment investors with different risk preferences? The study highlights two distinctive investor groups with individual interests and collective interests. The basic values that drive risk aversion are self-transcendental and those which drive risk seekers are individualistic.
The paper is organized as follows: Section 2 discusses briefly the decision making approaches and attitude to risk. Section 3 examines the bounded rationality and satisficing approach. The experimental design to formulate the bounded rationality approach and research method is discussed in Section 4 and 5 respectively. Section 6 describes the calibration of risk attitudes.
Section 7 explains segmentation of investor types and personal value systems that drive differences in risk perceptions. The paper concludes with a discussion in Section 8.
2. Decision Making Approaches and Risk Attitudes
Attitude towards risk is a fundamental facet of behavior influencing decision making in diverse settings. Decisions are made in uncertain environments on the basis of limited information and cognitive challenges. Attitude to risk is a core factor in models of choice and decision approaches (Kahneman & Tversky 1979; Camerer & Weber, 1987), it is observed as a personality trait (Zuckerman, 2000) which maybe domain specific (Weber, 2002). The expected utility approach provides a reigning basis for analysis of individual decision making under uncertainty. The utility framework provides key insights to empirically determine an individual investors’ attitude towards risk with specification of a utility function (Keeney and Raiffa, 1976).
The utility function accommodates different attitudes to risk; such as risk-averse, risk-seeking or risk-neutral types. The investor can determine the optimal investment strategy by maximizing his expected utility. However, the framework makes strong assumptions of rationality, each decision-maker is able to evaluate and maximize the utility function. The critics of the expected utility approach focus on its unrealistic assumptions about human analytical capabilities and in many situations these assumptions do not accurately describe how people make decisions (Camerer, 1995). Studies prove that deviations from the rational approach could arise due to Measuring Risk Attitudes and Personal Values – The Bounded Rationality Approach 5 various reasons such as social preference in decision making contexts (Guth, Schmittberger and Schwarze, 1982), loss aversion where people tend to weigh the possibility of a loss more heavily than that of a gain (Kahneman & Tversky, 1979) or even when utility is difficult to evaluate or maximize. Individuals seem to have limited cognitive resources (Simon, 1982) and choices appear to be lead by affective attitudes or subjective inclinations more than by economical reasons based on gain maximization (Kahneman, Ritov and Schkade, 1999). The view is that people use much simpler approaches or heuristics to arrive at decisions (Simon, 1955;
Gigerenzer, 2001). Hence, there is a need for a more psychologically plausible view of rationality that enables natural choice with limited mental resources. Therefore the paper explores the applicability of the bounded rationality approach to calibrate attitudes to risk where basic principles of aspiration formation and satisficing behavior are conformed by individual investors. Therefore, the decision maker is guided by aspiration adaptation rather than utility maximization in his decisions.
3. Bounded Rationality and Satisficing Approach
Bounded rationality is based on the premise that an individuals’ rationality is limited to cognitive ability and environment. The term “Satisficing” was coined by Herbert Simon in 1955 which means ‘satisfy + suffice’ where one finds sub-optimal solutions due to cognitive limitations and complexity of environment. The idea was that individuals do not seek the very best outcome but rather they stop searching once they find an outcome that is good enough. The concept of satisficing came originally from the realization that most maximizing problems are extremely complex and often lead to simple rules of thumb solutions. Hence, in complex environments such as financial markets, investors are boundedly rational as there would be various scenarios when they would just “satisfice” or find ‘good enough’ solutions rather than the ‘best solution’.
Intuitively, this lies between cognitive ability and adaptation to the environment, governed by simple and straightforward heuristics. In the spirit of bounded rationality, Gigerenzer (2001)
points out three distinct processes of the model:
Simple Search Rules
Simple Decision Rules Measuring Risk Attitudes and Personal Values – The Bounded Rationality Approach 6 Rationality itself is seen as an adaptive toolbox that is built on the building blocks provided by simple heuristics. Using these rules as a foundation, individuals develop an adaptive toolbox to deal with different problems in different circumstances; this is where the idea becomes relevant to investors. There are numerous investment options with different possible outcomes or returns that considering all of them would be implausible. For this reason, investors reflect over their aspiration levels, a lowest threshold that one wants to guarantee and a higher return level representing a real success. Lopes (1987) proposed a two-factor theory of risky choice in which she introduced a situational factor called Aspiration level. Aspiration is a link between goals and choice in the presence of uncertainty; goal for risk-averse people is security and risk-seeking is potential. Friedman and Savage (1948) noted that security and potential might co-exist in the same person; hence individuals buy both lottery tickets and insurance. Individual investors construct their portfolio as pyramid of assets (Statman and Shefrin, 1997) where they hold cash and bonds in the downside protection layer of their portfolio to prevent poverty and growth stocks in the upside potential layer of their portfolio to make them rich. Hence we use the bounded rationality approach to make investment decisions and predict portfolio allocation that satisfices investor aspirations. Fellner, Guth, and Martin (2007) examine whether individuals prefer satisficing over the optimizing approach in a simple investment decision. They view satisfying as more sensible and realistic, as it not only delivers an investment advice but also implies the outcome to be associated to their aspiration levels. However, the satisficing approach needs to be operationalized on a case specific basis and individuals have to explicitly learn what aspirations mean in specific task. In section 4, we detail the satisficing approach for making the investment decision in terms of aspiration setting and choosing the first alternative that exceeds the aspiration settings.
4. Experimental Design
Our experiment builds upon the study by Fellner, Guth, and Maciejovsky (2005). The financial decision environment and formalization of the bounded rationality approach are motivation for this study. The experiment illustrates an investment decision task where the subject has two financial assets, a risky one and a risk-free asset (bond), available to him as avenues for investment. The risk-free bond yields a fixed rate of return which is known to the investor prior to the investment decision. On the other hand, risky asset can land the investor is two states Measuring Risk Attitudes and Personal Values – The Bounded Rationality Approach 7 termed ‘high’ and ‘low’. The investor is aware of the rates of return in both states but does not know which state would occur at the time of investment decision. There are control questions to qualify participation in the experiment where subjects are expected to make simple calculations and identify different asset classes. Once qualified for participation, subjects are primarily classified into risk-averse, risk-neutral and risk-seeking category based on their investment choice. It is essential to categorise the subjects as characteristics of the assets are different for risk-averse and risk-seeking category as explained below.
The characteristics of the asset in possible states for risk-averse category are as follows:
R is the return offered by the bond, R = 1.10 H is the return offered by the risky asset in high state, H = 1.42 L is the return offered by the risky asset in low state, L = 0.80 E is the amount for investment decided by the investor himself I is the amount invested in the risky asset B = (E – I) is the amount invested in the bond p is the probability that the risky asset will attain a high state; p = 0.5 In the risk-averse category, L + H 2R; the expected value from the risky asset is greater than the return from the bond as the risk-averse investor needs a clear incentive to take risks.
The characteristics of the asset in possible states for risk-seeking category are as follows: