RIsk SCreening on the Financial Market (RISC-FM): A tool to assess investors’ financial risk tolerance

Abstract To advise investors on the financial market according to their financial risk tolerance it is necessary to apply a valid and reliable instrument measuring financial risk tolerance. We develop a screening instrument which assesses different facets of financial risk tolerance, namely, risk propensity, risk attitude, risk capacity, and risk knowledge. First, an item pool was generated and discussed with lay people as well as financial advisors to assure the questions’ understandability and answerability. Second, the most coherent and practice-oriented questions were tested empirically to determine four scales representing the four facets of risk tolerance. Third, resulting items were assessed using a representative sample of Austrian citizens interested in saving, stock trading, and investing, and psychometric quality of the instrument was determined.


PUBLIC INTEREST STATEMENT
Financial advisors are requested by legal regulations to assess their clients' risk tolerance to advise them accordingly. Many financial advisors currently use short item-scales on financial risk tolerance, which suffer empirical support and obey the law only in the broadest sense. These scales do also often lack minimum standards in test-psychology, a sub-discipline specialized in the construction of questionnaires. We as economic psychologists try to overcome this shortcoming and created a time-saving and easy-toapply questionnaire which is ideal for practical use. We called the questionnaire RISC-FM standing for RIsk SCreening on the Financial Market. With the RISC-FM, advisors can quickly learn about a client's financial risk tolerance and whether it is below, within, or above the population's average. This information helps starting a dialogue with clients about their financial needs and aspiration and meets legal requirements.

Introduction
Investment advisors need to know their clients' financial risk tolerance to be able to provide appropriate and well-tailored investment advice. To assure that financial advisors suggest suitable portfolios to their clients, legal regulations in many westerncountries (e.g., Australia: Corporations Act 2001 s912A(1)(h); European Union: Markets in Financial Instruments Directive 2014/65/EU Chapter 2 Section 2 Article25(2); United Kingdom: Conduct of Business sourcebook 2007 9.2.2; United States: Pension Protection Act of 2006 Section 601, Financial Industry Regulatory Authority Rule 2111, all as amended) demand the assessment of clients' risk tolerance and for them to be advised based on this tolerance. However, legal regulations are silent on how to assess risk tolerance.
Assessing investors' risk tolerance is difficult to estimate, it can vary over time, and value and development of securities, bonds, and stocks are subject to change overtime and the time horizon of investments plays a crucial role. Moreover, investors' self-ascribed risk tolerance may be affected by decision anomalies, such as framing effects and heuristics, or the investors may get emotionally overwhelmed while taking financial risks.
Investment advisors-although retained to assess their clients' risk tolerance-frequently lack validated instruments to assess risk tolerance. Consequently, investors often end up holding too conservative or too risky portfolios relative to their preferences (Cutler, 1995;Moreschi, 2005;Morse, 1998). In practice, financial advisors often provide their clients with information on investment options but do not incorporate their clients' needs and aspirations (Snelbecker, Roszkowski, & Cutler, 1990). Instead of judging their clients' risk tolerance, advisors often offer standardized rather than client-tailored solutions (Elsayed & Martin, 1998).
Investment institutions frequently use small sets of untested questions to assess risk tolerance with the aim to merely satisfy the legal requirements (Roszkowski, Davey, & Grable, 2005). Thus, instruments meeting psychometric standards, such as high reliability and validity, that can provide valid information on risk tolerance are long-needed (Dohmen et al., 2011;Grable, 2017). To the best of our knowledge, these criteria are so far met by  risk-tolerance scale (Kuzniak, Rabbani, Heo, Ruiz-Menjivar, & Grable, 2015). Furthermore, an instrument measuring financial risk tolerance should be easy to apply, economic and time costs involved in risk measurement should be low . Previous instruments interpret clients' individual scores, however, for an adequate interpretation of single values a comparison to norm values would be useful as this gives advisors and clients a comprehensive information about being below, within, or above the population's average. Individual risk tolerance resulting from a validated scale and related to population norms should be taken by advisors as a starting point to discuss and evaluate their clients' needs in depth (Davey, 2012;Grable, 2017;. Measuring risk tolerance is challenging as it is a psychological characteristic and as such not directly observable (Yao & Curl, 2011). Combining subjective data obtained from developed scales and objective risk measures based on previous behavior shall result in the most accurate assessment of clients' risk tolerance (Marinelli, Mazzoli, & Palmucci, 2017). However, gathering objective data is not always possible due to lack of data and privacy protection. Thus, accurate selfassessment measures are considered best practice to predict portfolio allocations (Guillemette, Finke, & Gilliam, 2012).
In the present paper, we develop a theoretically based and practically applicable screening instrument for the self-assessment of risk tolerance on the financial market. Psychometric standards are tested and norm values of the population's risk tolerance are developed. In the remainder of the paper we first define risk tolerance and describe a selection of biases in financial decision-making. Further, we describe the construction and selection of items for the risk tolerance scale and the final construction of the instrument on the basis of data from a representative sample of Austrian citizens interested in financial matters. Finally, we show how data collected regarding the risk assessment instrument can be interpreted in comparison with norm values.

Financial risk tolerance
Although people may be perceived generally as being either risk averse or risk seeking, predictions of risky decisions and behavior in a specific area cannot reliably be made by considering a person's general risk tolerance but must be based on risk tolerance in a specific field Dohmen et al., 2011). As risk tolerance cannot be generalized to other behavioral fields, risk tolerant financial investors may take risky decisions in different financial areas but not in other areas such as extreme sports. Research distinguishes between risk taking in the financial, physical, social, and ethical areas (Jackson, Hourany, & Vidmar, 1972;. A widespread definition of financial risk tolerance is that it is the maximum amount of uncertainty a person is willing to bear when making financial decisions (Grable, 2000). A more specific definition describes financial risk tolerance as risk-taking attitude in monetary affairs (Callan & Johnson, 2002;. This attitude varies on a continuum with extremes ranging from "low risk tolerance" to "high risk tolerance" . Although being quite constant over time Van de Venter, Michayluk, & Davey, 2012), critical life events such as getting married and having children can considerably affect people's financial situation and thus their financial risk tolerance (Davey, 2002). Cordell (2001Cordell ( , 2002 suggests the evaluation of the following four factors of clients' risk tolerance: (a) clients' past behavior in financial decisions (i.e., risk propensity), (b) clients' attitudes toward financial risks (i.e., risk attitude), (c) clients' ability to bear financial risks (i.e., risk capacity), and (d) clients' knowledge about financial risks (i.e., risk knowledge). Risk propensity and risk attitude especially reflect clients' subjective perception and acceptance of risk. The four factors are positively related to each other as depicted in Figure 1 (Cordell, 2001). Risk propensity is affected by risk attitude and risk knowledge and is interrelated with risk capacity. Furthermore, risk attitude is influenced by risk capacity and risk knowledge. Contrary to the other factors, risk knowledge is related to the other factors; however, it is not influenced itself by any of the other factors.
The most relevant factors for the evaluation of risk tolerance are risk attitude and risk capacity (Cordell, 2001;. Accordingly, risk seekers hold positive attitudes toward Figure 1. Associations between the four factors of financial risk tolerance (Cordell, 2001, p. 39 financial risks combined with a high financial capacity; whereas, risk avoiders hold negative risk attitudes and are unable to bear financial losses.
We assume that people face emotional strains as well when taking financial risks and that some people are emotionally overwhelmed by imagining or facing the risk of losing money. An instrument assessing financial risk tolerance thus needs to consider risk capacity as well as emotional aspects and should therefore measure objective and subjective risk capacity. Moreover, financial literacy, such as knowledge about investments was found to influence financial risk tolerance (Croy, Gerrans, & Speelman, 2010;Grable, 2000;Grable & Joo, 1999, 2004 and should therefore also be part of an instrument addressing financial risk tolerance. In our attempt to develop a valid financial risk tolerance assessment instrument, we consider the described four factors in the selection and construction of items. Moreover, in the following chapter we consider systematic biases occurring in financial decisions and explain how to formulate items accordingly. Furthermore, criterion-related validity (i.e., using situational and personal characteristics that are likely to influence a person's risk tolerance; cf. ) is assessed.

Financial decisions and biases
Financial decisions are predominantly decisions involving uncertainty and therefore often deviate from the neoclassical model of utility maximization and rationality. Especially deviations from rationality which yield biases resulting in a contorted perception of risk might influence financial decisions crucially. Examples for such systematic biases are framing effects, heuristics, and that the perception of risk depends on the presentation of probabilities.
Framing effects show that decision makers' preferences may alter if objectively identical decision alternatives are presented as either gains or losses (Tversky & Kahneman, 1981). Often when deciding between an alternative with a sure, but relatively small, gain and the alternative with a higher, but risky, gain, or no gain at all, people tend to choose the sure gain alternative. Thus, they show a low risk tolerance. However, when participants decide between a sure loss alternative and a lottery alternative with either no loss or an even higher loss, the risky option is more likely to be chosen. Thus, they show a high risk tolerance, as if they would try to repair the imminent loss. This effect was not only shown in gambling situations but also in financial decisions (Diacon & Hasseldine, 2007;. Accordingly, Guillemette et al. (2012) point out that amongst questions on self-assessment also questions incorporating loss aversion should be used to predict people's portfolio compositions. Consequently, measurement of financial risk tolerance needs to incorporate gain and loss situations as well as risk and certainty.
Heuristics are rules of thumb regarding decisions whose uncertain conclusions can be evaluated to ascertain and select an alternative that is expedited with a low cognitive effort (Tversky & Kahneman, 1974). While the application of heuristics often leads to good choices, sometimes they mislead decision makers. For example, the affect heuristic, which was discovered by psychologists, states that feelings may influence the outcome of risky decisions (Finucane, Alhakami, Slovic, & Johnson, 2000;Slovic, Finucane, Peters, & MacGregor, 2004). It is assumed that decision alternatives are evaluated and afflicted by either positive or negative feelings. Positive feelings toward alternatives lead to an underestimation of the contained risk and to an overestimation of the positive outcome of alternatives. Negative feelings toward alternatives cause an overestimation of the related risk and an underestimation of the utility. Thus, decision makers are more likely to select positively evaluated alternatives and neglect negatively evaluated alternatives. This effect shows even, if-from an objective point of view-the positive evaluated alternative yields worse results or the negatively evaluated alternative leads to better results than the other alternatives at stake. Johnson and Tversky (1983) showed that risk evaluations can be changed through a selective communication of positive or negative feelings. In addition, investors in the financial market showcase these biases. When estimating unknown stocks, positively evaluated stocks are perceived to have lower risks and higher chances of profits than the suggested objective criteria. In contrast, negatively evaluated stocks are perceived to include higher risks and lesser chances of profit than estimated objectively . Furthermore, investors report a biased tendency of investing in familiar instruments (Sahi, Arora, & Dhameja, 2013) and are more optimistic about the financial returns of companies with familiar product brands (Aspara, 2013). Accordingly, questions measuring financial risk tolerance should be formulated neutrally and should omit information that may trigger positive or negative feelings (e.g., trade names, company names).
In addition, the presentation of probabilities can influence decision makers' risk perception. Probabilities presented as frequencies (e.g., 20 out of 100) are perceived to be higher than the objectively similar probabilities presented as percentages (e.g., 20%; Slovic, Monahan, & MacGregor, 2000). Hence, when asking people about their financial risk tolerance, questions presenting probabilities both as frequencies and as percentages should be included.

Development of the item pool
Several steps were undertaken to develop items which suffice both the theoretical considerations on risk tolerance as well as the practical applicability in professional financial advising. First, to generate an extensive item pool on scientific and practically useful questions scientific as well as non-scientific literature was scanned for questions on financial risk taking (Appendix A shows the sources used). To identify questions suitable to measure financial risk tolerance we grouped the found questions according to Cordell's (2001) factors risk propensity, 1 risk attitude, risk capacity, and risk knowledge. Questions which did not reflect these factors were omitted. The so collected questions were translated into German and were reformulated for them to clearly refer to the used factors. For factors which were underrepresented, especially risk knowledge, additional items were composed. Second, the pool of 152 items so collected was discussed with two experienced financial advisors to see which items they consider as useful in the advisory practice. If an item was identified as not fulfilling the criteria of practical use, it was reformulated. Third, to learn whether the questions were comprehensible to laypeople, a focus group with laypeople was run. The group consisted of two women and two men aged between 31 and 65 years with an educational background ranging from school leaving examination to doctorate. All participants already invested on the financial market or were interested in investing in the near future. All items were thoroughly discussed and revised if necessary. Finally, to check which questions are relevant for the advisory routine, the item pool was discussed with two newly recruited financial advisors. The final item pool consisted of 60 coherent and practice-oriented questions.

Assessment of the item pool
An introduction letter with a link to the online questionnaire containing the selected 60 items, was distributed in the authors' professional and private networks. Participants were asked to forward the online questionnaire to friends and relatives (i.e., snowball sampling; Etter & Perneger, 2000). The online questionnaire was started by 124 participants. In total, 62 women and 41 men whose ages ranged between 23 and 76 years (M = 37.36; SD = 11.24; Md = 34.00) fully completed the online questionnaire. Of these participants, 4 had undergone apprenticeship training, 15 held a secondary education qualification, 82 held a university degree, and 2 did not indicate their education. Last year's gross income of participants amounted to less than 15,000 Euro for 14, between 15,001 and 30,000 Euro for 23, between 30,001 and 45,000 Euro for 21; between 45,001 and 60,000 Euro for 20; between 61,001 and 75,000 Euro for 6, and more than 75,000 Euro for 9; nine participants did not indicate their gross income. There were 67 participants who had prior experiences on the stock market; whereas, 36 participants indicated not having any prior experience whatsoever.
The set of 60 items included 19 items concerning the behavior in risky financial situations. These items referred either to previous gains or losses (answering format ranged from 1 = "very unlikely" to 7 = "very likely" and 1 = "very low-risk portfolio strategy" to 7 = "very high-risk portfolio strategy"). Additional 19 items assessed attitudes toward financial vulnerability and financial safety (the answering format ranged from complete disagreement (1) to full agreement (7); 15 items concerned financial and emotional risk capacity (answering format: 1 = "do not agree at all" to 7 = "fully agree"), and the remaining seven items examined participants' knowledge about financial risks (7-point answering format ranging from complete disagreement to full agreement). A final direct question about the general readiness to take risks was enclosed ("Please indicate how willing to take risks you estimate yourself"; answering format: 1 = "not willing to take risks at all" to 7 = "very willing to take risks"). Items with the same answering format were presented in randomized order. At the end, participants indicated their gender, age, last year's gross income, and prior experience in financial markets. Table 1 shows the descriptive statistics and the discriminatory power of all 60 items, as well as inter-item correlations. To construct the scales on risk taking behavior, items that were not normally distributed (skewness <−0.95 or skewness >0.95), and that showed floor or ceiling effects (median <2.00 or median >6.00) were excluded from further analyses. With the remaining 46 items, an exploratory principal component analysis with varimax rotation was conducted to test whether the resulting factor structure would represent the hypothesized factors risk propensity, risk attitude, risk capacity, and risk knowledge. Although, the sample size is small for running factor analyses, we used the procedure for a first test of the scales. Kaiser-Meyer-Olkin criterion for sampling adequacy (0.81) and Bartlett's test of Sphericity (χ 2 (1081) = 2801.09; p < .01) indicate that the data is adequate for factor analysis. The extracted factors' eigenvalues above 1.00 were 13.84, 3.01, 2.62, and 2.41; further, nine factors had eigenvalues between 1.81 and 1.07 suggesting a four-factor solution when depicted in a scree plot. Additionally, Velicer's (1976) MAP test conducted with O'connor's (2000) syntax for SPSS suggested a four-factor solution. The explained variances were 29.45%, 6.39%, 5.58%, and 5.12% for the first four factors and between 3.86% and 2.27% for the remaining nine factors.
The factor analysis was repeated with a restriction to four factors (explained variance = 46.55%). Items on propensity, attitudes, capacity, and knowledge loaded mainly on one of the factors. Items loading lower than .40 and items that could be assigned to more than one factor were disqualified from further analyses. In addition, to build reliable scales, the five items with the highest loadings on the respective factor were re-analyzed by a principal component analysis restricted to four factors. Explained variance amounted to 58.43%. Table 2 shows the descriptive statistics of the scales and the inter-scale correlations. Appendix B depicts all items of the questionnaire, labeling the selected items with an asterisk. Note: SK = skewness; DP = discriminatory power; * Items included in the scales financial risk taking behavior, attitudes towards financial risks, financial capacity, and financial knowledge; # items which are not normally distributed (skewness <-0.95 or > 0.95 or Md < 2.00 or Md > 6.00);~items loading lower than .40; + items, which loaded on more than one factor; correlations above .19 are significant on a level of p < .05; items with concluding "R" are recoded; PR = financial risk propensity, AT = financial risk attitude, CA = financial risk capacity, KL = financial risk knowledge. Finally, fit indices of confirmatory factor analyses with a general factor solution and a four-factor solution were compared. Critical values which indicate a good fitting model are a non-significant χ 2 -test with χ 2 /df < 2.00, RmSEA < 0.06, CFI > 0.90, and AGFI > 0.90 (cf. Byrne, 2001;Kline, 2011). The general factor solution with all items loading on one factor did not indicate a satisfactory fit (χ 2 (170) = 504.86, p < 0.01, RmSEA = 0.14, CFI = 0.57, AGFI = 0.64). The four-factor solution with the items on propensity, attitudes, capacity, and knowledge loading on a single latent factor each, however, showed a better but not fully sufficient model fit (χ 2 (164) = 259.09, p < 0.01, RmSEA = 0.08, CFI = 0.88, AGFI = 0.75). Only when the four-factor solution allowed for the correlation of error terms of items with similar concepts the model fit was satisfactory (χ 2 Figure 2. Factor structure of the four solution for the prelimiary scale development and the representative study.
Note: ***p < .001, **p < .01, *p < .05; the numbers preceding the slash show the coefficients of the preliminary scale development and the numbers following the slash depict the coefficients of the representative study; items with a closing "R" are recoded.
According to Cordell (2001) risk propensity, risk attitude, risk capacity, and risk knowledge are interrelated. Thus, a good construct validity of the scales goes along with positive inter-scale correlations. Furthermore, general risk tolerance is related to risk tolerance in more specific areas (Dohmen et al., 2011). Therefore, to test for construct validity, the correlations among the scales' propensity, attitude, capacity, and knowledge, as well as the general readiness to take risks were assessed (see Table 2). The scale on risk propensity and the scales on attitudes, capacity, and knowledge were positively related. In addition, attitudes toward financial risks and capacity and knowledge showed positive correlations. Furthermore, the scales on financial risk capacity and financial risk knowledge revealed a positive relation. As expected, the direct question on the readiness to take risks showed positive correlations with all constructed scales. Although the direct question on risk tolerance did not specifically ask about financial risk tolerance, participants were assumedly primed on financial risk tolerance.
The selection and first analyses of items yielded five useful items for each risk tolerance scale. In the next step of scale development, financial risk tolerance items were presented to a representative sample of investors.

Participants and procedure
An internationally operating market research institute was engaged for data collection. They sent an online questionnaire to a representative pool of Austrian residents interested in saving, stock trading, and investing. In total, data of 1,018 participants was obtained. However, some data had to be excluded due to odd response patterns (N = 18) and due to more than two missing responses (N = 64). Therefore, data of 936 participants (396 female; 540 male) was examined with an age range between 15 and 82 years (M = 46.81; SD = 15.75; Md = 47.00). As the highest educational level, 6.60% of the participants indicated to have completed compulsory education, 34.70% finished apprenticeship training, 30.40% indicated secondary education qualification, and 22.00% held a university degree. A non-specified other education was reported by 6.40% of the participants. The allocation of the personal monthly net income 2 was as follows: 16.8% earning less than 1,200 Euro; 14.50% between 1,201 and 1,650 Euro; 20.9% between 1,651 and 2,100 Euro, 12.20% between 2,101 and 2,700 Euro; 11.10% more than 2,701 Euro; 2.50% had no income of their own; 21.90% gave a blank response. One-fourth of the participants (25.70%) had no prior experience on the stock market, whereas, the majority of participants had already invested money at some point in time in the past. A small percentage (3.10%) did not respond to this question.

Material
The online questionnaire included the 4 × 5 questions on financial risk propensity, attitudes toward financial risks, financial risk capacity, and financial risk knowledge (answering formats: 1 = "very unlikely" to 7 = "very likely"; 1 = "do not agree at all" to 7 = "fully agree"; "I do not know"). In addition, a direct question on the general financial risk tolerance was included ("All in all, how risk averse or risk seeking do you evaluate yourself to be in financial matters?" answering format: 1 = "very risk averse" to 7 = "very risk seeking"). Finally, participants completed questions regarding their demographics and prior investment experiences. Table 2 shows the descriptive statistics and Cronbach's alphas of the scales propensity, attitude, capacity, and knowledge. To further examine the factor structure, confirmatory factor analyses with a general factor solution and a four-factor solution were conducted. Missing responses were replaced by the respective scale means (propensity: 4.06; attitude: 2.44; capacity: 3.91; knowledge: 3.96). The general factor solution revealed unsatisfactory fit indices (χ 2 (165) = 1,625.91, p < 0.01, RmSEA = 0.10, CFI = 0.77, AGFI = 0.78) and the four-factor solution showed better but still unsatisfactory fits (χ 2 (164) = 1,305.59, p < 0.01, RmSEA = 0.09, CFI = 0.82, AGFI = 0.82). However, allowing for the same correlations between error terms, as in the assessment of the item pool, resulted in a satisfactory model fit (χ 2 (158) = 647.16, p < 0.01, RmSEA = 0.06, CFI = 0.92, AGFI = 0.91). Regression coefficients and correlations are shown in Figure 2.

Results
To assess the scales' construct validity, the scale inter-correlations and the correlations between the scales and the direct question on financial risk tolerance were analyzed (see Table 2). Financial risk propensity was positively correlated with the attitude toward risk, capacity, and knowledge. Accordingly, positive relations were found between attitudes and capacity as well as between attitudes and knowledge. The relationship between capacity and knowledge was also positive. Finally, positive correlations were found between the direct question on general financial risk tolerance and the scales themselves. These results are a first indication of the existence of a good construct validity.
The correlation analyses included age, monthly net income, and the four developed scales on risk tolerance. Results show that with increasing age the propensity to take risks and the attitudes toward taking financial risks decrease (r = −.15, p < .001 and r = −.27, p < .001, respectively). However, no relation between the participants' age and their capacity and knowledge was found (r = −.05, p = .13, r = −.01, p = .68, respectively). The monthly net income was positively related to financial risk propensity (r = .16, p < .001), attitudes toward financial risks (r = .13, p < .001), financial capacity (r = .29, p < .001), and financial knowledge (r = .30, p < .001). In sum, these results are a first indication of a good criterion-related validity.

Computation and interpretation of data obtained by the instrument
As individual scores cannot be interpreted easily, psychometric personality tests compare the results of single participants to norm values obtained from a representative sample. Through this comparison the actual value of the single participant can be interpreted accordingly. To acquire the norm values for the financial risk tolerance scales at hand, data obtained from a representative sample interested in saving, stock trading, and investing was used. 3 To assess individuals' risk tolerance, data obtained by the questionnaire (see Appendix B for the items and scales) need to be processed using the following steps: (a) recoding the six questions which are formulated in the reversed direction (see Table 1); (b) calculating subscale means; (c) identifying the subscale means (i.e., raw scores) in the particular norm tables depicted in Appendix C and selecting the corresponding T-values. 4 A general value of risk tolerance should incorporate that the subareas are differently important for financial risk tolerance. Therefore, further steps include: (d) multiplying subscale means with their respective weighting factor 5 (propensity: 0.21; Table 3. Results of the criterion-related validity test of the scales financial risk propensity, financial risk attitude, financial risk capacity, and financial risk knowledge regarding prior experience, gender, age, and monthly net income On basis of the profile sheet, financial counselors can now assess a single investor's risk tolerance compared to a representative sample of Austrian citizens interested in financial matters (i.e., norm group). If the marked T-value is located within the gray area, it signifies an average value compared to other Austrian citizens interested in saving, stock trading, and investing in this subcategory. If the mark is on the left side of the gray area, it implies a below average value. Accordingly, if the mark is on the right side of the gray area, it indicates an above average value. This information can be easily understood by advisors as well as clients and can be used by advisors to start a well-founded discussion about an individual's investment intentions.

Discussion
The aim of the present study was to develop a reliable and valid screening instrument for assessing investors' risk tolerance. The 20-item-instrument is suitable to assess risk propensity, risk attitude, risk capacity, and risk knowledge within a few minutes. Following a theoretical and practical approach for item construction overcomes the shortcomings of ad hoc measures. After further validation the instrument can be used in the daily advisory routine.
Questions were collected according to the theoretical definition of financial risk tolerance and the empirically derived factors of risk propensity, risk attitude, risk capacity, and risk knowledge. Furthermore, when (re-)formulating items, systematic decision biases based on psychology were taken into account and omitted in item formulations. In addition, to receive relevant questions for the counseling practice, all questions were discussed with investment advisors. A further discussion with present and future investors assisted in obtaining a set of comprehensible questions, as requested by . Finally, psychometric principles were applied for scale construction, revealing four scales which when proved gave a first indication of its reliability and validity. Thus, the presented instrument is both theoretically founded as well as eligible for practical use. However, further validation of the found scales is needed and incremental validity compared to other existing scales should be examined.
A word of caution is as follows: when used for counseling, the instrument should not be used to replace an informative and constructive discussion between the client and the financial advisor (Davey, 2012;. Accordingly, the instrument at hand provides financial counselors with a first impression of a client's financial risk tolerance, which can function as the starting point of an informed dialog in which the client's needs and aspirations can also be identified. Since not all factors are perceived as equally important (Cordell, 2001;, risk attitude and risk capacity should have higher weights than risk propensity and risk knowledge. The instrument at hand takes this into account since it weighs the subscales differently before they are combined to obtain a general risk tolerance value. These calculations can easily be conducted either by manual calculation or via a computer program. To check whether a client is more or less risk tolerant than others, the results of individuals can be compared with the results of a norm group of the population. However, for translated questionnaires and for the application in different countries, new norms are needed. Also, different norms for comparison should be computed for men and women and for different age groups separately because gender and age influence risk tolerance (Bonsang & Dohmen, 2015;Dohmen et al., 2011;Grable, 2000;Hallahan et al., 2003Slovic, 1999;van Rooij et al., 2007;Wang & Hanna, 1997). Although, income and wealth also effect investors' risk tolerance (Cohn et al., 1975;Grable, 2000;Hallahan et al., 2003 we do not recommend generating separate norm groups, as these are sensitive topics and thus obtaining valid data is difficult. Moreover, norm values need to be renewed after some years to assure that they are still valid. Although a person's risk tolerance is rather stable over years , after experiencing significant life events (e.g., marriage, childbirth, etc.), changes may occur (Davey, 2002). Hence, if financial advisors are counseling clients from whom they do not have much background information, the questionnaire could be re-administered after a period of time to take possible changes into account.
Objective as well as subjective measures of risk tolerance are needed to accurately assess financial risk tolerance (Marinelli et al., 2017). The presented scale can be combined with clients' objective data such as investors' real-life portfolios to further prove its criterion validity. This could countervail the problems of subjectivity and self-representation underlying all psychological assessments. Thus, future research on the instrument should include objective risk measures. It would be of great interest to compare actual investment behavior of a relevant sample of investors with the data obtained through the instrument. This would also serve as a further validation of the instrument.

Correction
This article has been republished with minor changes. These changes do not impact the academic content of the article. A T-value of 50 indicates that 50% of the population has a lower result and 50% has a higher result than the corresponding norm value. A T-value of 30 represents that 2% of the population has lower results and 98% has higher results and a T-value of 80 is interpreted to indicate that 98% of the population has lower results and 2% has higher results. 5. Weighting factors include the beta-weights of the respective subscale regressed on the direct question on general financial risk tolerance in the representative study. The factors indicate the explanatory value of each scale regarding financial risk tolerance and show that the subscale attitude should have the highest influence in the aggregated scale; however, propensity, capacity, and knowledge are also needed to be considered to develop a complete understanding of an individual's financial risk tolerance. 6. Copies of no longer available online resources can be obtained from the authors. Imagine you could invest money in drilling a gold mine which has only a very low chance of success. If the drilling is successful you receive your investment hundredfold. If the drilling is not successful the invested money is lost. How likely would you be to invest in the drilling?
Imagine you could invest 10,000 euros in one of two types of investments. You know that the first type will be worth between 9,800 and 10,600 euros after one year. For the second type, you know that its value will be between 6,400 and 14,200 euros. How likely would you be to invest in the second type? Imagine you are working in a company that will go public in three years. You are now offered stocks which you cannot sell for the next three years and do not receive a dividend until then. There is, however, the possibility that the stocks are worth ten times more than what they are now after the initial public offering. How likely would you be to buy the stocks? I am ready to accept losses to achieve gains in the long run.
I am ready to reduce a little security to achieve higher gains.
High profits are attractive to me, although this means that I must also take a high risk.
I like to take the risk of losing money when there is the chance to win money.
I am ready to invest more than a quarter of my financial assets in a risky investment.
Hoping to achieve high gains, I am ready to bear high financial risks.
To get the chance to achieve high gains I would invest in products which lost in value over the past months and speculate that their value will increase in the future.