Chapter1. Introduction
1.1 Background and significance of the research
The question to be answered is the impact of incremental income on personalrisk attitude. Specifically, in this paper I will investigate whether the increase ofpersonal income will result in the change of personal risk attitude. If the personal riskattitude will change, will it increase or decrease? Surely these questions can beobserved in our daily life.During the past few years, we have witnessed the bull market5and also bearmarket6in Chinese stock market. There were many people caught the timing, enteredstock market and got the chance to earn revenue from the bull market. With theagitation of medias that the bull market could last for a period time, people believeinvest in stock market can bring revenues, and most of them also have tasted thesweetness. Within that period, we also witnessed a phenomenon that some peopleeven invest in stock market with using leverage. It is sure that using leverage willresult in the subject to have a higher revenue, however, meanwhile a higher level ofleverage also means a higher level of risk. Undoubtedly, tragedies resulted fromhigher level of leverage surely happened, a man who used eight times of leverage toinvest in stock market lost and as a result he chose to finish his life. There exists apuzzle behind this phenomenon, that is, will subject’s risk attitudes change when hisincome changes? By observing the phenomenon, we would find that some investorsstart to using leverage once he had tasted the sweetness of invest in stock market. Itseems some subject become less risk averse once his income increase. Thus in thiscase, to investigate the relation between risk attitudes and income is meaningful.
..........
1.2 Structure of the paper
This paper is pided into five sections:The first chapter is introduction. Firstly, the background and significance of topicis illustrated; secondly, it introduces the structure of this paper; thirdly, it introducesthe research methods used in this paper and finally it introduces the innovation andinsufficient of this paper.The second chapter includes literature review, theoretical consideration andresearch hypothesis. Firstly, it makes a review about the existing literatures, whichanalyze the relationship between income and risk attitudes; secondly, theoretical basisabout risk attitudes and income are introduced. Thirdly, basing on the existingliteratures and theories, I put forward the hypothesis of the relationship betweenpersonal income and risk attitudes, and the hypothesis will be test in the later chapter.The third chapter mainly focus on the sample and data that will be used to testhypothesis in the fourth chapter. In this chapter, I explained how the questions aredesigned and asked, how the sample is selected, where the data comes from and whatkind of variables will be used in the fourth chapter.The fourth section includes descriptive statistical analysis and empirical analysis.The empirical analysis includes simple regression and use of propensity scorematching methods.The last section makes a conclusion about the entire paper, points out theinnovation and limitations of this paper, and also bring up some questions which arenot answered within this paper but can be worked on later.
..........
Chapter2. Literature Review
2.1 Literature Review
Generally speaking, there are more literatures discussing about the influencingfactors of personal risk attitudes in international, while for the aspect of domesticresearch, there gradually showed some researches working on this field. A commonview is that personal risk attitude is influenced by some factors: for some factors suchas gender, the influential function is clear that male is less risk averse than women(Armin Falk, Anke Becker et al., 2015;), but for the other factor such as income, therelationship between risk attitude and income is still under debating.As for the experimental methods, it is more used for economical study ininternational, however in domestic research field, experimental method is graduallyaccepted but still not so much.Many literatures consider that the poor behave more risk averse than the rich(Bardhan et al., 2000). Arrow mentioned in his literature that “the degree of absoluterisk aversion is decreasing with the increase of income, and the degree of relative riskaverse is increasing with the increase of wealth” (Arrow, 1965). It is a simpleinduction drawn mainly based on daily observation of the different behavior of theincome of people, even though this setup is frequently used to predict economicbehavior under uncertain conditions (Gollier and Pratt, 1996). It is considered thestatement that absolute risk averse index is a concave function of function should beregarded as a natural hypothesis, because with this hypothesis, it means the moreincome an investor has, the reduction of risk premium brought by a certain amount ofincrease in wealth should be even reduced. However, there is no empirical evidence tosupport this idea. The earlier empirical evidence comes from experiment. For instance,experiments show that the utility of wealth with total wealth above $ 200,000 is in theform of logarithmic utility, which is often associated with risk aversion. However,investors with less than this level of wealth show an increasing relative risk aversion(Gordon et al., 1972). Cross-sectional survey data were used to study the relation between relative risk aversion index and wealth, and it is considered that the utilityfunction of constant relative risk aversion is accurate (Friend and Blume, 1975).Survey data is used to construct a measurement of absolute risk-averse index. And itis found that the degree of absolute risk averse is a decreasing function of wealth. Thefinding rejects the setting that inpiduals have constant relative risk aversion (CRRA)(Guiso and Paiella, 2008). Moreover, experimental techniques are also used tomeasure well trained monkey’s risk attitudes for water rewards as a function of bloodosmolality (an objective measure of how much water the subject possess). Monkeysare trained to be little risk averse to match the risk attitudes of human being. Andexperiment shows that risk attitudes of well-trained monkeys are strongly wealthdepend; as monkeys become “poorer”, risk aversion increased. (Hiroshi Yamada,Agnieszka Tymula et al., 2013; Thirst-dependent risk preference in monkeys identifya primitive form of wealth.)
........
2.2 Related Theories
Expected utility theory is the most basic and widely used theory about consumer choices ongoods with risk or uncertainty. In economics, game theory, and decision theory, theexpected utility hypothesis is about the assumption that people have a preference forchoices with uncertainties (gambling). This hypothesis suggests that, if satisfyingspecific axioms, subjective values associated with inpidual gambling are thestatistical expectation of the inpidual's valuation of the results of the gambling.In the von Nuemann-Morgenstern expected utility theory, there has four axiomsto give definition of how a rational decision maker is like. The four axioms arecompleteness, transitivity, independence and continuity.With the axiom of completeness, inpidual are assumed that they have welldefined preferences and inpiduals can always make decisions between any twochoices.Axiom (Completeness): For every A and B either A≥B or A≤B. This means that theinpidual either prefers A to B, or is indifferent between A and B, or prefers B to A.
.........
Chapter3. Research Methodology........21
3.1 Sample Selection .......21
3.2 Questionnaire Design .............22
3.3 Definition of Variables............24
3.3.1 Main Variables.............24
3.3.2 Control Variables .........26
3.4 Research Procedure................29
3.5 Research Model.........30
Chapter4. Statistical Analysis and Empirical Analysis.....33
4.1 Statistical Analysis .....33
4.2 Empirical Analysis......42
Chapter 4. Statistical Analysis and Empirical Analysis
4.1 Statistical Analysis
79.57% questionnaires are regarded as effective questionnaires. 602questionnaires delivered and all of the questionnaires collected back. Among the 602questionnaires, 537 of them are considered valid but not all the 537 questionnaires areused for main empirical analysis. As the experiment designed, people are supposed tochoose safe choices at first and switch to risk choices at a particular point and withoutswitching back. However, it is not possible for all the participants never switch backagain. The data collected back showed surely there are subjects switched more thanonce, among those 537 questionnaires, 58 respondents switched more than once. Forthe convenience of analysis, those 58 responds are eliminated. As a result, the left 479questionnaires are considered really valid.Degree of risk aversion is measured by the number of safe choice of ourrespondents. In this experiments, I evaluate respondents’ risk attitudes two times, oneis risk attitudes based on their reality, the other is risk attitudes based on theassumption of earning an amount of money ，（which equals to double of their annualsalary）by chance.Table 6 shows general description of the data and all the numbers shown in thetable are mean. For the risk attitude based on status quo Rd, it shows significantdifference between the variable Age, Edu and Inc. Take Inc as example, thoserespondents love risk have relative high income while risk-averse respondents havelower level income.
.........
conclusion
The conclusion drawn from this paper is that subject has increasing relative riskaversion (IRRA) and decreasing absolute risk aversion (DARA), which rejects thehypothesis that subjects are constant relative risk aversion (CRRA) and constantabsolute risk aversion (CARA). This result means, firstly, subjects’ personal riskattitude would not be consistent throughout his lifetime, it will change; secondly,subjects’risk attitudes will be influenced not only by his age, gender etc, but will alsobe influenced by his personal income; thirdly, subjects’ absolute risk attitudes willdecrease when his income increase, means from the view of absolute risk attitudes,when the subjects has more income, he will become less risk averse.The contribution of this paper is that, the conclusion implicates that, with theincrease of income, subjects would like to have riskier assets, but the percentage ofrisky assets to total assets may decrease.One of the shortcomings of this paper is the bidirectional change of income. Asprospect theory suggest, subject has different attitudes toward gain and loss, and it isalways related with a reference point. The function is concave in the domain ofgaining and is convex in the domain of convex. Moreover, subjects have differentdegree of feeling to the same amount of property in different domain. So, in theexperiment, I ought to ask subject how they will choose base on both gaining andlosing. However, when I gathered questionnaire which contains the entire threesituations, the quality of data is extremely bad. Subjects became bored with threeidentical groups of questions even these questions based on different background. Asa result, subjects just pick up options randomly. For this reason and in order to ensurethe quality of data, I just investigate the impact of incremental income on personalrisk attitude.
..........
References (abbreviated)