Ethical investment: How do moral considerations influence investment ...

12 downloads 0 Views 117KB Size Report
86 items - Eva Hofmann*, Erik Hoelzl & Erich Kirchler ... eva[email protected] ...... and policy implications of ethical investing, in P. Taylor-Goodby (ed.) ...
Ethical investment: How do moral considerations influence investment behaviour? Eva Hofmann*, Erik Hoelzl & Erich Kirchler University of Vienna, Austria

Abstract: In recent years, a change in shareholders’ demand can be noted: more and more investors ask for ethical investment opportunties. Investors are not only influenced by financial benefits but also by moral considerations. For example, shares of companies involved in the tobacco, alcohol or gambling industry are considered ‘sin stocks’ by some investors and therefore are not traded. Rational decision theorists argue that moral considerations would introduce inefficiency by reducing the number of investment options and therefore only financial considerations should govern investment decisions. However, the increasing demand for ethical investment indicates that investment decisions are influenced by additional factors, for example attitudes and moral values. The discrepancy between rational decision theory and actual shareholders’ demand opens a wide field of research for economic psychology. Several economic and psychological models can be applied to explain ethical behaviour. In the present study, we test suitability of: (a) multiple attribute utility theory, (b) theory of planned behaviour, and (c) issue-contingent model of ethical decision making in organisations. In an experimental setting, 141 participants traded company shares on a computer simulated asset market. Over 12 periods, companies varied in ethicalness (i.e., how the company treats their employees) and in profitability (i.e., whether the expected dividend for a share was low, medium or high). Participants’ bids and asks for shares were recorded. Results indicate that moral considerations influence investment decisions. Differences between the three models and implications for the understanding of ethical investment behaviour are discussed. Keywords: Behavioural finance, ethical decision making, ethical investment Acknowledgement: The authors would like to thank Nicole Fabbro and Silvia Hansbauer for helping with data collection and the ‘Verein zur Foerderung der Wirtschaftspsychologie’ for financial support.

*

Corresponding Author: Eva Hofmann, Department of Economic Psychology, Educational Psychology and Evaluation, Faculty of Psychology, University of Vienna, Universitaetsstrasse 7 1010 Vienna Austria T: +43 1 4277 47882 F: +43 1 4277 47889 [email protected]

Introduction Ethical investment is a relatively new catch phrase in the European world of finance. Sometimes also called sustainable investment or socially responsible investment, this kind of investment subsumes on the one hand the funding of companies and investing in funds that guarantee compliance to certain positive ethical criteria. On the other hand it represents the disregard of companies according to certain negative ethical criteria (Hofmann, Penz & Kirchler, under review; Lewis, Webley, Winnett & Mackenzie, 1998). Positive ethical criteria involve e.g., environmental protection and fair trade with the Third World, negative ethical criteria include e.g., the utilisation of nuclear power, the disdain of human rights, and so called ‘sin stocks’, i.e., shares of companies involved in the tobacco, alcohol or gambling industry (Anand & Cowton, 1993). Thus, since the 1980s, when the first ethical investment fund was launched by the life insurance company Friends Provident in the UK, more and more investors in Europe have been considering not only their financial benefits but also ethical issues when deciding on investments. In the view of rational decision theorists, the average human being – the rational homo economicus – is motivated exclusively by financial benefits; ethical thoughts are not relevant for behaviour. They argue that moral considerations would introduce inefficiency by reducing the number of investment options (Lewis, 2001; Cullis, Lewis & Winnett, 1992). Incorporating exclusively ethical investments in a portfolio and for this reason diminishing the portfolio, increases either the risk or reduces the profit of this portfolio and therefore makes it less efficient than conventional portfolios. Thus, only financial considerations should govern investment decisions. However, the increasing demand for ethical investment indicates that investment decisions are influenced not only by financial benefits but also by additional factors, such as attitudes and moral values. This discrepancy between rational decision theory and actual shareholders’ demand opens a wide field of research for economic psychology to discover those factors. Several theoretical models can be applied to explain ethical behaviour in general. In the current paper the suitability for investment behaviour of some of these models is tested. In particular, three approaches are compared: (a) multiple attribute utility theory, (b) theory of planned behaviour (Ajzen, 1985), and (c) issue-contingent model of ethical decision making in organisations (Jones, 1991). The multiple attribute utility theory (MAUT) is based on assumptions of a rational decision maker. It is presumed that when assessing different alternatives, homo economicus values the alternative with the highest utility most. In multiple attribute utility theory, utility is separated into attributes. For each alternative the utility (MAU) is determined by the sum over the utility of each attribute (ui) multiplied by the weight of importance of the attributes (wi) (Formula 1). Because of the separation of attributes, the multiple attribute utility theory represents a theoretical background for a practical method to measure utility (Baron, 2000). MAU = ∑ wi ui (1) i

The multiple attribute utility theory was already applied to study several decision problems concerning ethical issues. Studies investigated the evaluation of issues such as environmental planning (Kwak, Yoo & Kim, 2001; Kwak, Yoo & Kim, 1998) or health related matters (Chapman, Elstein, Kuzel, Nadler, Sharifi & Bennett, 1999; Kasubek & Aschenbrenner, 1978). These studies corroborate that the multiple attribute utility theory is a practical approach to simplify decision making by the evaluation of different alternatives. The theory of planned behaviour (Ajzen, 1985, 1991) is a social psychological concept for predicting human behaviour. It is based on the theory of reasoned action (Fishbein & Ajzen, 1975) which was extended by a factor. The theory of planned behaviour postulates that a person’s intention to perform a behaviour is the immediate determinant of that behaviour. Intention, in turn, is a function of three factors: Attitude toward the behaviour, subjective 1

norm and perceived behavioural control (Figure 1). The attitude toward the behaviour represents a personal factor which comprises the positive and negative evaluation of performing the behaviour. The factor subjective norm relates to social pressure to perform the behaviour. The perceived behavioural control covers the subjective controllability to perform the behaviour and depends on earlier experiences and expected obstructions. It represents the factor by which the theory of reasoned action was extended. Thus, the attitude toward the behaviour, the subjective norm and the perceived behavioural control influence, via the intention, behaviour. The relation between the factors of the theory of planned behaviour (Ajzen, 1985, 1991) was often supported and was also used to explain ethical behaviour. The theory was applied to such behaviours as environmental friendliness (e.g., Bamberg, Rolle & Weber, 2003; Montalvo-Corral, 2002), health promotion (e.g., Schifter & Ajzen, 1985) or money management (e.g., Bobek & Hatfield, 2003; East, 1993), once directly to ethical investment behaviour (Borrello, Morricone, Pedon & Benevene, 2004). Figure 1: The theory of planned behaviour (Ajzen, 1985, 1991; p.181) Attitude toward the behaviour Subjective norm

Intention

Behaviour

Perceived behavioural control

The issue-contingent model of ethical decision making in organisations (Jones, 1991) is a model that predicts ethical behaviour. Based on Rest’s model (1986) the issue-contingent model describes the decision process: Decision makers (1) recognise the moral issue, then (2) make a moral judgement and (3) establish a moral intent and finally, (4) engage in moral behaviour. These four components are influenced by two other factors: Moral intensity and organisational factors. Moral intensity represents the strength with which the moral issue influences the decision process and therefore determines all four decision components. It is composed of six subordinate factors: Magnitude of consequences, social consensus, probability of effect, temporary immediacy, proximity and concentration of effect. Organisational factors influence the decision process solely in the last two components, establishing moral intent and engaging in moral behaviour. Organisational factors are comprised by the subordinate factors group dynamics, authority factors and socialisation process (Figure 2). These manifold factors in the model generate complex interdependencies. Parts of the issue-contingent model of ethical decision making in organisations (Jones, 1991) were supported and explained ethical behaviour. The issue-contingent model was applied to ethical business decisions such as environmental pollution (Flannery & May, 2000; May & Pauli, 2002), product safety (May & Pauli, 2002; Weber, 1996) and fraud (Carlson, Kacmar & Wadsworth, 2002; Weber, 1996). According to these studies, moral intensity determines the factors “recognise moral issue” and “make a moral judgment”; it is weakly supported that it also influences the factor “establish moral intent”. To our knowledge, the model as a whole has never been tested.

2

Figure 2: The issue-contingent model of ethical decision making in organisations (Jones, 1991; p.379) Moral intensity Magnitude of consequences Social consensus Probability of effect Temporal immediacy Proximity Concentration of effect

Recognise moral issue

Make moral judgement

Establish moral intent

Engage in moral behaviour

Organisational factors Group dynamics Authority factors Socialisation process

It is the aim of the current study to answer three research questions: (1) Does ethicalness have an influence in investment decisions? If ethicalness influences investment decisions, (2) can this influence be explained by decision models? (3) Which model can explain the influence of ethicalness in investment behaviour best? Subsequently these questions will be answered by means of a questionnaire and a computer simulated asset market. Method Overall, 141 persons, 86 females and 55 males, took part in the study. Participants were mainly students of the University of Vienna and of the Vienna University of Economics and Business Administration. Therefore, participants were on average of 25 years (M=24.74, SD=4.12). Twenty-seven participants invested in shares, whereas 13 of them invested in ethical shares. Forty-five participants were members or donators of at least one charity. All participants were recruited either personally or via e-mail at the universities in June 2004. They were informed that they would take part in an experiment on investment behaviour lasting two hours for which they would receive 5 Euro as a compensation for their time. In addition they would have the chance, according to their performance in the experiment, to win a book voucher worth 50 Euro. When they agreed to take part in the experiment, they were given a questionnaire, which they had to complete and return before the computer simulated asset market. The questionnaire consisted of total 86 items and took 20 minutes to complete. First, the participants read a description of companies and for each company successively completed 19 items. They had to imagine that they had inherited a certain amount of money and shares of the company and were asked, whether they would like to add more shares of the companies to their portfolio. The descriptions were four fictitious scenarios describing company A as a personnel promoting company, which enhances projects for older employees, presenting company B also as a personnel promoting company, which enhances equal promotion prospects for all employees. Company C was described as a personnel neglecting company with the regular business practice is to release older employees at the age of 45 if legally possible, and company D was also presented as a personnel neglecting company, which refuses any promotion prospects for its employees. One quarter of the participants received a questionnaire in which they got information about company A first and then about company C, B and D, the others found a sequence of companies C-A-D-B, B-D-A-C or D-B-C-A. After each company description 19 questions had to be answered on a seven-point-scale. These 3

items measured the perceived ethicalness of the companies (1 item; utility), the factor intention (1 item; establishing moral intent), a semantic differential for an attitude scale (4 items; attitude toward the behaviour), a subjective norm scale (2 items), a control scale (2 items; perceived behavioural control), a moral intensity scale (6 items), a moral judgement scale (2 items; making a moral judgement) and the factor moral issue (1 item; recognising the moral issue)1. Following all four company descriptions and the respective questions, participants filled in a single item concerning personal importance of ethicalness of investments (weight of importance), which was worded generally and did not correspond to one of the four companies specifically. The final part of the questionnaire included nine items asking for socio-demographic data, such as gender, age and possession of conventional and ethical shares. The computer simulation was conducted with participants in groups of 14 persons in the computer laboratory of the Faculty of Psychology at the University of Vienna. Each group received an intensive training to learn the market mechanisms on the simulated asset market. In a continuous double auction (Smith, 1994), they were allowed to make bids and asks2 for company shares, and to accept existing bids and asks of other participants. As soon as all participants understood the mechanisms and were able to employ them, the simulation was started. Over 12 periods participants successively traded shares of the companies A, B, C and D which were described exactly as in the questionnaire. During one period only shares of one company were traded. Over the periods each company paid either low, medium or high dividends. Every period the participants were informed which share was going to be traded and which dividend was to be expected in the forthcoming period. On the information screen the participants were presented 5 different possible dividends which occurred with a 20% probability each.3 Again, the sequence of the companies in the computer simulation was balanced; the order of the companies were A-C-B-D, C-A-D-B, B-D-A-C or D-B-C-A. The sequences of the amount of expected values of dividends stayed the same for all companies. In the first period, when shares of a certain company were traded, a medium dividend had to be expected, in the second period a high dividend was anticipated and in the third a low dividend. After each period participants were informed about their earnings of the previous period and then received information concerning the following period. After the 12th period the profit earned over the entire computer simulation, was displayed and the lottery to gain the book voucher took place. The higher a participant’s profit compared to the other 13 participants was, the higher was the chance to win the book voucher. All participants were paid 5 Euro. Results For the statistical analysis, data from the questionnaire were matched with the bids and asks from the computer simulation via a personalised code. Although 12 sessions with 14 participants each were conducted, for two reasons the data of 27 participants could not be included in the analysis. First, some personalised codes on the questionnaire did not comply with the ones from the computer simulation, so that the data could not be matched; second, some participants did not fill in all items of the questionnaire, which was needed to undertake 1

The influence of organisational factors in the issue-contingent model of ethical decision making in organisations is not included, because investment behaviour in the experiment bases solely on individual decisions, which are not affected by organisational regulations. 2 A bid is a participant’s offer to buy a certain share at a pronounced price. An ask is a participant’s offer to sell a certain share at a pronounced price. 3 When a low profit had to be expected the screen displayed possible dividends of 0, 1, 3, 5 and 6 experimental currency units (expected value=3); at a medium expected profit it displayed 0, 39, 41, 53 and 67 experimental currency units (expected value=40); and at a high expected profit it displayed 0, 74, 82, 118 and 126 (expected value=80). The exact amount of the distributed dividend for each period was randomised depending on possible dividends and probability. 4

the analysis of the models, and are excluded for that reason. Thus, data from 141 participants remain for analysis. Table 1: Descriptives for model components (questionnaire data, n=141)

attribute utility (multiple attribute utility) intention (establish moral intent) attitude (attitude towards the behaviour) subjective norm control (perceived behavioural control) moral intensity social consensus moral judgement (make moral judgement) moral issue (recognise moral issue)

Personnel promoting companies Company A Company B M SD M SD 16.45 9.22 15.11 9.25 4.92 1.91 5.11 1.87 5.82 1.00 5.71 1.10 4.68 1.63 4.72 1.60 6.31 1.04 6.32 0.97 3.64 1.19 3.55 1.23 6.08 1.45 5.86 1.39 3.84 1.98 3.73 1.87 5.06 2.18 4.92 2.12

Personnel neglecting companies Company C Company D M SD M SD -7.16 6.39 -4.29 8.12 2.94 1.91 2.92 1.69 3.16 1.47 3.34 1.36 2.55 1.27 2.87 1.41 5.67 1.45 5.76 1.28 3.49 1.21 3.39 1.14 3.90 1.99 4.03 1.89 2.05 1.36 2.14 1.35 5.27 2.06 5.20 2.04

The scales from the questionnaire (Table1) were tested for their reliability. Whereas the scales attitude, subjective norm and moral judgement hold very good Cronbach-α (minimal α=.74, maximal α=.90), the scales control (minimal α=.29, maximal α=.40) and moral intensity (minimal α=.61, maximal α=.73) maintain less satisfying reliability. It is assumed that the disappointing reliability of the scale control emerges from ceiling effects which are due to the experimental approach of the study.4 The scale moral intensity is reduced by the item for social consensus, which measured moral intensity depending on how moral the companies are perceived while the other five items of the scale are independent of moral perception.5 According to the theoretical background the scale attribute utility is generated by a multiplication of the items utility and weight of importance for each company. Table 2: Descriptives for investment behaviour (asset market data)

low profit

bids asks

medium profit

bids asks

high profit

bids asks

Personnel promoting companies Company A Company B M SD M SD 23.42 15.87 23.94 16.30 n=79 n=87 40.61 19.81 35.68 30.68 n=72 n=72 43.90 19.73 41.07 14.89 n=92 n=85 60.86 37.23 57.21 15.70 n=94 n=84 59.08 18.83 55.08 21.04 n=97 n=79 77.95 27.95 77.71 33.41 n=93 n=84

Personnel neglecting companies Company C Company D M SD M SD 20.63 13.91 19.03 15.22 n=82 n=75 34.37 23.04 29.21 21.12 n=78 n=68 38.06 19.22 36.19 17.96 n=95 n=80 52.10 20.56 55.79 30.82 n=88 n=94 51.16 20.83 50.48 23.75 n=85 n=88 76.32 39.68 68.46 20.30 n=78 n=85

Participants’ investment behaviour (Table 2) was recorded in the computer simulation. Bids and asks of participants were collected in all periods. Overall 5337 bids were recorded. From these bids, 3593 were accepted by other participants and led to a trade. Additionally, 4590 asks were recorded. From these asks, 3090 were accepted by other participants and led to a trade. Because it is to be assumed that participants tend to bargain while selling and buying shares in the computer simulation, their bids and asks do not exactly represent their evaluation of the shares. Lower bids and higher asks are made, while the exact evaluation is in between their range. Therefore, for the subsequent analysis the highest bid and the lowest ask per 4

In the experiment participants should have felt objectively as well as subjectively to be in charge of the investment decision. 5 The exclusion of social consensus increases the reliability of moral intensity, especially for the personnel neglecting companies (minimal α=.74, maximal α=.77). 5

period of each participant are used. Interval regression (Stewart, 1993) is used for the statistical analysis. It allows the dependent variable to exist of intervals as well as lower and upper bounds, when participants gave solely bids or solely asks and also allows to account for repeated measures. The first research question (1), whether ethicalness has an influence in investment decisions, is tested with an interval regression using behaviour as dependent variable and ethicalness6 and low and high profit7 as predictors. The statistical test reveals that ethicalness as well as profit have significant influence (Chi2(5)=178.41, p