Investment Risk Appetite (IRA): Scale Development and Validation

Abstract

Investment Risk Appetite (IRA) is a pivotal concept in risk management, reflecting an investor’s willingness to tolerate financial risks within acceptable thresholds. As empirical investigations into this construct gain momentum, there is a growing need for a scientifically validated tool to facilitate in-depth examinations of risk appetite. There are no scales to measure risk appetite. The present study addresses this gap by developing and validating a scale on risk appetite. Leveraging data collected from 405 respondents and employing established methodologies, the study introduces the Investment Risk Appetite (IRA) Scale. The questionnaire had a five-point scale. The scale consists of two factors: risk tolerance (α = 0.837, composite reliability = 0.836) and risk aversion (α = 0.905, composite reliability = 0.906). The validation was done by exploratory and confirmatory factor analysis (EFA and CFA). The loadings for EFA and CFA exceeded the threshold limit of 0.40. The scale demonstrates robust internal consistency, content, and construct validity. Hence, this scale has all the required validity. Overall, this scale demonstrates robust validity and reliability. In addition, this study examined the differences based on the demographics of the respondents. The scale, poised to make a significant contribution to the literature on risk appetite, will provide a theoretical foundation for future in-depth investigations. This study is expected to inspire future empirical examinations of this compelling construct.

1. Introduction

Academics, financial experts, and policymakers seek to examine individual risk, financial, and investment behaviors thoroughly. There is a growing recognition of the importance of psychological and attitudinal factors, alongside traditional microeconomic and macroeconomic reasons, in shaping financial behaviour. Risk is now acknowledged as a crucial aspect influencing financial behaviour, reflecting a holistic understanding of investment decision-making (Bayar et al., 2020). What appears attractive to one investor may not be compatible with the aims or risk appetite of another. As a result, financial advisors must adopt a tailored strategy that considers individual preferences, risk appetite, and investment horizon. Furthermore, there may be significant differences based on intrinsic and observed risk-taking attitudes (Schoemaker, 1993). Risk-taking and risky behaviour have a long research history (Jennex et al., 2022; E. U. Weber et al., 2002; M. Weber et al., 2012). Prior studies have revealed that risk perception (Keller et al., 2006), risk tolerance (Cole et al., 2017), and risk appetite (Clark-Murphy & Soutar, 2004) significantly impact the investors’ ability to make decisions regarding investment. However, research on risk appetite (RA) based on a psychological backdrop and perspectives is deficient. Existing studies that examine individual-level psychological attributes that outline the risk propensity and appetite are scarce. Most scholarly works on risk and RA have stressed the appetite for corporate risks (Can Ergün et al., 2023; Marshall et al., 2019).

Exploring risk appetite will provide insight into investor sentiment and price movements, contributing to the existing literature. It will be helpful to know about investors’ risk preferences, acceptable risk levels, and their current mood. The scale will be beneficial for multiple stakeholders. Some of them include individual investors, portfolio managers, financial institutions, educators, and administrators. Precisely assessing risk appetite would enable portfolio managers to accurately examine their clients’ risk tolerance. It will also allow the provision of custom-made and tailored investment recommendations that align with an individual investor’s risk tolerance and aversion levels. Knowledge of risk appetite will enable financial institutions to develop tailored investment products that align with individual risk levels. It will also help educators design financial literacy programs and materials that facilitate investors’ understanding of their personal risk profiles, enabling them to make informed and robust investment decisions. Only scant efforts were made to study individual risk-taking propensities, willingness, and capacity to bear risk (Raut, 2020). Earlier studies examining psychological aspects have focused on bias and heuristics, acknowledging them as fundamental to behavioural decision-making (Kahneman & Tversky, 1979). These cognitive biases and heuristics have significantly influenced the financial markets, leading to herding (Youssef & Waked, 2022), overconfidence (Kuranchie-Pong & Forson, 2021), and risk and ambiguity aversion (Nisani et al., 2021). These attributes are the cornerstone of individual decision-making processes. Building on these propositions, prior research explored how individuals make decisions by drawing on their experiences, often biased in favour of past incidental reminders. In addition to psychological biases, social pressure, fear of criticism, technical knowledge constraints, and a lack of investment experience have also been identified as influential factors in decision-making, affecting investor attitudes and risk-taking propensity (Guiso et al., 2008). This work is prompted since individual psychological attributes associated with risk propensity and appetite in the dimension, as presented in this study, are not widely researched. Hence, the objective of this study is to develop and scientifically validate a tool to measure investment risk appetite. Since the primary focus of this study is on the psychological aspects of decision-making, cognitive biases, and risk-taking propensity, the findings are expected to have universal appeal, despite being collected in Saudi Arabia.

2. Literature Review

Investment risk appetite (IRA) is a critical concept in financial risk management, denoting the level of risk that investors can tolerate. It depends on their readiness to handle uncertainty, which is influenced by the unpredictability of the factors that determine asset prices (Hui et al., 2010). Additionally, variations in intrinsic risk-taking attitudes and observed risk behaviour further underscore the need for personalized strategies (Rodriguez & Edwards, 2010; Schoemaker, 1993), highlighting the requirement for IRA. Knowledge of IRA is essential for traditional and Fintech financial advisory to provide consistent financial advice with the actual level of risk acceptable to each investor (Lippi & Rossi, 2020). Despite the growing recognition of the importance of psychological and attitudinal factors in shaping financial behaviours, existing research on IRA is deficient. The absence of a thorough scale to evaluate IRA challenges financial advisors and policymakers in providing tailored investment advice that aligns with individual investors’ risk tolerance and aversion levels. The inability to accurately assess IRA may result in misaligned investment strategies, potentially leading to suboptimal investment outcomes. Given this context, there is a definite need for a scale to evaluate IRA, including their tolerance levels and aversion.

Due to the diverse profiles and unique goals of investors, financial advisors must proceed cautiously and offer customized investment advice. This caution is crucial because an individual’s risk appetite (RA) level may vary, and an investment strategy that appeals to one investor may not align with the risk tolerance or aversion of another. Hence, financial advisors must adopt a customized investment strategy, considering an IRA, that stems from individual preferences, risk tolerance, and aversion levels. Given these considerations, this study aims to bridge the current research gap by developing and validating a psychometric tool to measure IRA, providing a thorough insight into investors’ risk appetite, tolerance, and aversion levels. The validation of such a psychometric scale would offer valuable insights for financial practitioners and policymakers. This would significantly enhance the effectiveness and efficiency of financial advisory services, enabling better alignment with investors’ needs and objectives. The study also aims to investigate whether there are any differences in the variables under study based on the demographics of the respondents.

Individual RA is a multifaceted psychological construct that profoundly impacts decision-making processes, wherein risk-averse or risk-taking attitudes can have far-reaching consequences across various domains, including social, psychological, ethical, and financial realms (Lippi & Rossi, 2020). Aligned with the investigation of Nicholson et al. (2005), this study adopts an alternative perspective on investigating risk-taking. This perspective is based on the expected utility theories, such as prospect theory (Kahneman & Tversky, 1979, 2013). According to the prospect theory, risk-taking behaviour is contingent upon individuals’ perception of their position relative to a reference point. They tend to be risk-averse when positioned in a state of gain and risk-seeking when positioned in a state of loss.

The Utility theory proposes that an individual’s investment decision is based on rationality and trade-offs between risk and return (Newendorp, 1967). These theories encompass integrating psychology into financial behaviour, and decision-making processes. Ritter (2003) opined that financial behaviour is rooted in psychology and influences decision-making through cognitive processes and the exploitation of market inefficiencies, known as arbitrage. Another theory with application in the current study is the Personal Construct Theory, proposed by G. Kelly (1955). This theory postulates that all human thought patterns are dichotomous. In addition, B. Kelly (1966) suggested the presence of a theoretical bipolarity for almost all behavioural constructs, referred to as dyadic elicitation. Further, all constructs are bipolar, with almost all behavioural constructs having their respective opposites (Epting et al., 1993). This theory could help explain the simultaneous occurrence of (bipolar) risk-taking and aversion attitudes among investors.

Understanding the determinants of IRA will help Fintech and traditional financial advisory services and investors understand the provision of financial advice that could align accurately with each investor’s RT, allowing them to tailor products based on individual preferences. A substantial body of literature demonstrates how people consciously engage with and voluntarily accept risks. Risk-taking is socially constructed (Taylor-Gooby & Zinn, 2006; Zinn, 2015), and individuals who engage in risk-taking or risk-seeking understand that they are exposing themselves to potential harm. As a result, people take precautions to protect themselves from such risks and losses (Zinn, 2017). Hence, voluntary decision-making is dichotomous. Highlighting this dichotomy of voluntary risk-taking. Tulloch and Lupton (2003) defined it as an “activity in which individuals engage, is perceived by them to be in some sense risky but is undertaken deliberately and from choice.” This view was also expressed by Luhmann (1993), suggesting that risk encompasses certain advantages to be gained despite something being at stake. Risk-takers also consider probable steps to protect their stake and avert losses. This duality was also expressed in a few earlier studies (Breivik et al., 2017, 2020), which is reflected in the present study as well. According to Can Ergün et al. (2023), investors’ risk appetite is also influenced by their risk aversion and the macroeconomic environment.

People are risk-averse concerning physical, economic, ethical, and existential risks (Breivik et al., 2020). Investment RA is an investor’s acceptable or desirable level of risk, enabling consistent risk-taking. It is the willingness of investors to take risks and the tolerable or desirable level of risk that an individual is willing to accept (Can Ergün et al., 2023). Marshall et al. (2019) describe the concept of RA as discourse involving the assessment of “gaps” between individuals’ actual and ideal “risk exposure levels.” It explains the overall RT or the degree to which an individual is willing to make risky investment decisions. It gives investors boundaries, guidance, and expectations to make informed investment decisions. Multiple factors influence investment decisions, depending on the investors’ RA. In addition to RA, financial capability and overall socioeconomic profile influence investing decisions (Can Ergün et al., 2023).

Individuals have different RA levels. Risk appetite is a dynamic concept that defines the level of financial RT that changes over time. It determines the amount of risk that is tolerable or desirable, allowing for more consistent risk-taking when investing (Hui et al., 2010). Higher RA boosts investor confidence, encourages sensible and rational market behaviour, and helps them develop into assertive risk-takers. Individuals with a lower RA level have lower anticipated emotions and tolerance levels (Böhm & Pfister, 2008). This discussion points toward IRA involving RT on the one hand and RA on the other. This dichotomy/bipolarity in assuming the risk is well-established in behavioural research (B. Kelly, 1966). The Institute of Risk Management (Institute of Risk Management (IRM), 2011) proposes using “fight” versus “flight” terminology to explain how RA is reflected strategically to navigate risky environments. This perspective views choosing one strategic option as a willingness to “fight” by mitigating perceived risks, while avoiding another option that represents “flight” from associated risks. This metaphorical interpretation underscores the significance of tolerance and aversion components of RA, which are instrumental in guiding investment decisions. It is, hence, pertinent to discuss these two constructs. According to Grable (2000), RT is “the maximum amount of uncertainty an individual is willing to accept while making a financial decision.” It is a multidimensional and elusive concept influenced by several predisposing factors (Tigges et al., 2000), which involves the acceptable level of risk that an individual is willing to assume (Grable, 2000). Sometimes, “risk aversion” is used interchangeably with financial RT, although it may imply the opposite (Ryack & Sheikh, 2016). Furthermore, at times, RA is also added to this list, but it is not synonymous (Hui et al., 2010). Previous studies in financial planning have examined various factors that influence individual RT since they significantly impact their investment decisions (Bayar et al., 2020; Raut, 2020). Multiple factors influence an investor’s RT, leading to varying degrees of RA, which could change over time (Ravikumar et al., 2024). It also increases with excessive returns on risky assets.

Though there are multiple scales of risk propensity, risk tolerance, and the like, a scale envisaged in this study is lacking. M. S. Kumar and Persaud (2002) developed a measurement of risk aversion based on changes in excess returns of assets. Most scales in the existing literature are based on organizational risk appetite, with few scales available to measure IRA. A detailed literature review failed to find a generalized scale to measure IRA. The study intended to develop and validate an easy-to-administer and score, domain-specific measure of IRA. This was achieved in three steps: item generation, scale development, and evaluation of the IRA using psychometric tools. To accomplish this purpose, this study followed the guidelines outlined by T. Hinkin (1995) and Netemeyer et al. (2003), who recommended detailed scale development practices that many subsequent studies have followed. A few of them include Ahmad et al. (2020), Alon et al. (2016), Guleria and Kaur (2022), Horvath and Greenberg (1989), Jo and Hong (2023), Latif and Marimon (2019), and Sigmundsson and Haga (2024). The details are presented in the following sections.

3. Methodology

Developing a scale is a challenging process. The framework of T. R. Hinkin et al. (1997) and the recommendations of Hair et al. (2019) present explicit, often-used guidelines for developing and validating a psychometric tool. In particular, Hair et al. (2019) elucidate the complexity involved in tool development and validation, which is derived from various reference manuals. The study also followed the criteria stipulated by Ellis et al. (2008), who emphasized the need to adhere to rigorous techniques for developing new measures. Li et al. (2024), based on a review of four decades of literature, identified the development and validation procedures. This includes theoretical and empirical validation. Other extant literature that provides further guidelines for developing sound psychometric tools includes Clark and Watson (2016), T. R. Hinkin (1998), K. Kumar and Beyerlein (1991), Netemeyer et al. (2003), and Sulphey (2020). All the guidelines set by these social scientists were observed and followed in right earnest. The main stages proposed by the above studies, which were also observed in this study, include item generation, scale development, and evaluation. Thus, this study drew on the existing knowledge base to develop and validate the IRA.

The initial priority was to determine the appropriate scale length to elicit high-quality responses (Sulphey, 2015). There are divergent views about the length of the scale. T. Hinkin (1995) believes that scales that are too long or too short will negatively affect the results. Anastasi (1985, 1986) and Schriesheim et al. (1991) believe that a short scale helps deal with the “demands in terms of time.” It would also help minimize response biases arising from boredom and fatigue. Furthermore, short scales also help ensure internal consistency and reliability (Cook & Berrenberg, 1981; T. Hinkin, 1995). However, Nunnally (1978) presents a contrary argument, suggesting that short scales may lack validity, reliability, and consistency.

Domain sampling and parsimony are a precondition for validity (Cronbach & Meehl, 1955). Hence, attention was paid to ensuring the tool was neither excessively short nor excessively long, thereby creating a meaningful tool. The items were chosen based on a literature review, which helped evaluate the related tools and consultations with specialists. This would also help achieve content validity (Barrena-Martínez et al., 2017; Tossell et al., 2015). Content validity is the expert judgment (Hardesty & Bearden, 2004). Hence, proper domain sampling and parsimony were prioritized to identify initial items. Furthermore, Cronbach and Meehl (1955) stated that validity is impossible without parsimony.

To arrive at the domains of IRA, inputs from B. Kelly (1966), Böhm and Pfister (2008), and the Institute of Risk Management (Institute of Risk Management (IRM), 2011), particularly regarding the duality, were given due importance. The items for the IRA Scale were generated with inputs from various scales (Rahman et al., 2020; E. U. Weber et al., 2002). Since Levenson (1981) proposed that using a ‘personal’ conception when constructing measuring scales is better, all items are expressed in the first (“I”) rather than in the second/third (“He/She/They”) person. Hence, based on a detailed review, a 13-item pool with two domains was created. The two domains identified include risk-seeking and risk-aversion. The item pool was then refined with the assistance of experts, who include academics and investment managers. The experts acknowledged the identified domains and items. These are necessary for ensuring content and facial validity. Based on expert opinion, a pool of 10 items was retained. Three items had to be dropped as they seemed vague or repetitive. The full set of items is provided in the Appendix A. Example items from each domain are:

• If potential returns are substantial, I am willing to invest my funds despite the possibility of high losses.
• It is prudent to invest funds in secure investments despite low returns.

The tool had a five-point scale, ranging from “strongly agree” to “strongly disagree.” (Table A1). Utmost care was taken to ensure that items were understood as the researchers intended them to be, while also emphasizing aesthetics, thereby effectively ensuring face validity. A multifaceted approach involved quantitative methodologies, which included statistical analyses such as item-to-total correlation, factor analysis (T. Hinkin, 1995; Boyle, 1991), reliability testing, and validity assessments to ensure the robustness and psychometric properties of the measurement tool. These analyses helped to identify the underlying dimensions, evaluate the consistency of responses, and ascertain whether the instrument measures what it purports to measure. These steps ensured that the final instrument was comprehensive, relevant, and aligned with existing empirical works and established theoretical frameworks.

The questionnaire was administered online to potential respondents who are involved in investment activities, through various social media groups in which they are members. The link to the questionnaire, which was uploaded to Google Forms, was posted in the groups with the active help of the group administrators. An appeal to respond to the questionnaire and the purpose of the study was also posted. This helped to reach a broad range of respondents. To ensure privacy and maintain ethicality, the respondents were guaranteed anonymity and confidentiality, as proposed by Stanley and Wise (2010). Further, no identifying or personal questions were posed. This reduced potential psychological harm and stress to the respondents. Therefore, participation was strictly voluntary. The response to the link, which was purely voluntary, was gathered from retail and individual investors using a convenience sampling technique, with a focus on data collection in Riyadh, Saudi Arabia. Convenient sampling allows the respondents to have efficient access to respond to the questionnaire. The English and Arabic versions of the questionnaire were included to provide a better understanding and response. The forward and back translations were conducted according to the recommended method in translation guidelines (Sousa & Rojjanasrirat, 2010). This is a significant methodological approach when respondents are not proficient in the target language (Epstein et al., 2015). This method ensured that there was no difference between the two versions. The data were collected over approximately eight weeks, gathering responses from 405 individuals. There was no missing data, and all the responses were used for analysis. Sampling bias was effectively mitigated by ensuring diversity and heterogeneity in demographics. This included diversity in terms of gender, education levels, employment, and professional backgrounds. Of the 405 responses, 219 (54.1%) were from males and 186 (45.9%) from females. Most samples were employed (86.4%), and the least were retired (2.2%). The samples also had varying employment experience. The majority of respondents (59.7%) held diplomas, followed by those with a master’s degree (21.1%) and a Doctorate (16%). Hence, the collected data exhibit wide demographic diversity, are heterogeneous, and are reasonably representative of the population under study. The distribution of the sample can help gauge its representativeness and relevance (Table 1).

Table 1. Demographics of The Sample.

Demographics		Number	Percentage
Gender	Male	219	54.1
	Female	186	45.9
Employment	Unemployed	29	7.3
	Self-employed	17	4.3
	With employment	350	86.4
	Discharged	9	2.2
Educational Qualifications	Diploma	242	59.7
	Undergraduate	13	3.2
	Masters	85	21.1
	Doctorate	65	16.0
Employment Experience (in years)	Less than five	82	20.2
	5–10	92	22.7
	10–15	95	23.5
	15–20	75	18.5
	More than 20	61	15.1

Self-reporting was used for various reasons besides its equivalent validity and to capture inner preferences and orientations that an external observer cannot perceive. These inner preferences regarding investment behavior would be more closely related to the subsequent cognitive processes that generate perceptions and investment decisions (Oettingen & Seligman, 1990). Furthermore, they indicated that Eastern samples are less likely to overestimate their self-evaluation.

4. Results

The initial stages of the scale refinement involve examining inter-item and item-to-total correlations. Boyle (1991) considers that items can be eliminated if any inter-item correlation coefficient exceeds 0.70. However, no item required elimination in the present study, as all inter-item correlation coefficients were below 0.70. The results are reported in Table 2. Furthermore, according to Field (2013), Kim and Mueller (1978), and Nunnally (1978), items with item-to-total correlation coefficients greater than 0.30 should be retained for further analysis. This criterion assumes that items within a shared domain will have comparable correlations. Consequently, items with lower correlation coefficients may not accurately represent the intended domain and could indicate questionable reliability (Churchill, 1979). In this study, all item-to-total correlation coefficients exceeded this threshold, ranging from 0.391 to 0.933. Consequently, no items were removed following the item-to-total correlation analysis.

Table 2. Construct Reliability and Validity.

Factor	Item-to-Total Correlation	EFA Loading	CFA Loading	Alpha	CR (rho_c)	AVE
Risk Tolerance (RT)	0.448 **	0.811	0.796	0.837	0.836	0.565
	0.391 **	0.780	0.707
	0.451 **	0.740	0.716
	0.417 **	0.823	0.785
Risk Aversion (RA)	0.893 **	0.879	0.800	0.905	0.906	0.765
	0.933 **	0.892	0.924
	0.924 **	0.875	0.895

Note: EFA—Exploratory factor analysis. CFA—Confirmatory factor analysis. CR—Composite reliability. AVE—Average variance extracted. ** Significant at 0.01.

Following the initial steps, factor analysis (FA), including exploratory factor analysis (EFA) and confirmatory factor analysis (CFA), was also conducted. The Kaiser–Meyer–Olkin (KMO) value was 0.929, and Bartlett’s test of sphericity was significant at p < 0.000 (773.80). The EFA yielded a single-factor solution, capturing 63.81% of the total variance. After the EFA, the CFA was conducted using the partial least squares structural equation model. CFA helps confirm the factor structure and assess the model’s significance level (Kaur & Sharma, 2015). The factor loadings are presented in Table 2. T. Hinkin (1995) recommended a minimum loading of 0.40 for item retention. In this study, the loadings for EFA and CFA exceeded this threshold. The EFA loading ranged between 0.740 and 0.892, and the CFA values ranged between 0.707 and 0.924. Furthermore, Ford et al. (1986) proposed a minimum factor loading of 0.40 or higher for retaining items. All the factor loadings in this study meet this threshold, indicating that all items demonstrate satisfactory relationships with the underlying factor, which supports their inclusion in the measurement tool.

The FA used varimax rotation and extracted two factors without cross-loading (Annexure A). The first factor consisted of four items (Eigenvalue of 3.968, accounting for 56.691% of the variance). Based on the item properties, this factor is ‘Risk Tolerance (RT).’ The second factor has three items (with an Eigenvalue of 1.453, accounting for 17.895% of the variance). Based on the items, this factor is termed ‘Risk Aversion (RA).’ The factor identification is as expected and has a strong literature backing (J. R. Adams-Webber, 1998; J. Adams-Webber, 2003; B. Kelly, 1966; Reddy, 1999; Zinn, 2015). The results show a robust causal relationship and loadings between the factor (Investment risk appetite), the sub-factors (Risk tolerance and Risk aversion), and each item. The fit index values (Table 3) meet the thumb rules, indicating that the model enjoys a perfect fit.

Table 3. Fit Index.

Model Fit Criterion	Model Values	Acceptable Level	Reference
Goodness of fit index (GFI)	0.916	>0.90	Hair (2010)
Comparative fit index (CFI)	0.913	≥0.90	Hu and Bentler (1999)
Normed fit index (NFI)	0.905	≥0.90	Lohmöller (1989); and Bentler and Bonett (1980)
Adjusted goodness of fit index (AGFI)	0.832	≥0.80	Gefen et al. (2003)
Parsimony goodness of fit index (PGFI)	0.458	No limit. Close to 0.50	Kelloway (1998); Sharma (1996)

The reliability is examined with Alpha and item-to-total correlations (Table 2). It can be observed that all alpha values, 0.837 and 0.905, exceed the threshold of 0.70 (Nunnally, 1978), indicating reliability (Raes et al., 2011). Validity is “the degree to which an instrument covers the meaning of the concepts included in particular research” (Stone & Kotch, 1989). Validities could be content, convergent, discriminant, and criterion. There is no universally accepted quantitative examination for content validity, which can be determined through the investigator’s judgment, as Stone and Kotch (1989) argued. This measure has content validity because it was developed following a comprehensive literature assessment and subsequent revisions and validation. Kerlinger (1966) claimed that convergent validity can be inferred when evidence is gathered from multiple sources.

Hair et al. (2013) stipulate that AVE and item loadings are used to examine convergent validity. AVEs above 0.50 are acceptable (Barclay et al., 1995; Hair et al., 2013, 2020). Hair (2010) and Hair et al. (2013) prescribe a minimum composite reliability of 0.70. The AVE and CR values in Table 2 exceed the stipulated limits, confirming convergent validity. According to Hulland (1999), discriminant validity is the distinctiveness of measures or constructs. This uniqueness is essential for ensuring the validity of the constructs (Hair et al., 2013). In addition, Bagozzi and Kimmel (1995) argued that discriminant validity is assumed when there is a low correlation between factors. The analysis presented a p-value of 0.601. This confirms the presence of discriminant validity, as the correlation coefficient is less than 0.70 (Anderson & Gerbing, 1988). Additionally, the correlation values are lower than the square roots of AVE (0.752 and 0.875), thus meeting the criteria outlined by Fornell and Larcker (1981). According to Kerlinger (1966), construct validity links the theory to its psychometric properties. T. R. Hinkin (1998) emphasizes the importance of construct validity. A scale is robust with construct validity (Cronbach & Meehl, 1955; Schmitt et al., 1991). According to T. R. Hinkin (1998), a scale with internal consistency and content validity has construct validity. Hence, this scale has all the required validities. There is a possibility of Common method variance (CMV), since the data for the study were gathered through self-reporting questionnaires (Podsakoff & Organ, 1986). The Harman’s single-factor test (Podsakoff et al., 2012) was conducted, which revealed no CMV. The data collection was conducted in two stages, with a 15-day interval between them. Additionally, the items in the questionnaire were shuffled, along with the inclusion of dummy items. All these steps rejected the possibility of CMV.

Knowledge about IRA is essential in financial planning and will help make modern investment decisions. It will play a significant role in the success of financial planning and management processes. The IRA scale developed in this study offers several advantages for multiple reasons. Firstly, existing scales often lack rigorous psychometric validation procedures, which are essential for accurate measurement. Most current scales are not suitable for application in varying contexts. Additionally, the dimensions that define risk appetite are not adequately represented in the existing scales. The IRA scale, developed and validated in this study, comprises seven items organized into two factors. It is empirically sound and is valid, since it was constructed using rigorous psychometric tests (K. Kumar & Beyerlein, 1991; T. R. Hinkin, 1998; Clark & Watson, 2016; Sulphey, 2020).

A cursory examination of other Behavioural Finance (BF) questionnaires, including risk appetite and tolerance, reveals that most have single factors with very few items, some as few as a single item. For instance, Lippi and Rossi (2020), who investigated risk appetite, employed a single-item approach. They considered it an ordinal variable with three categories: low-risk, medium-risk, and high-risk, based solely on mean values. The Financial Risk Tolerance Questionnaire, developed by Grable and Joo (2004), consisted of four items (CR = 0.863). Another Financial risk tolerance questionnaire by Rahman (2019) had four items (CR of 0.81) under a single factor. The Financial Risk Tolerance Questionnaire used by Rahman et al. (2020) consisted of a single factor with only three items (CR = 0.921). The Financial Tranquility scale, developed by Vieira et al. (2023), consisted of ten items (Alpha = 0.918). The Risk Perception Scale by Ahmed et al. (2022) had six items (CR of 0.683). The details are presented in Table 4 to facilitate a better understanding of the various related scales.

Table 4. Details about related questionnaires.

No.	Questionnaire Name	Reference	No. of Factors	No. of Items
1	Risk appetite	Lippi and Rossi (2020)	1	1
2	Financial risk tolerance	Grable and Joo (2004)	1	4
3	Financial risk tolerance	Rahman (2019)	1	4
4	Financial risk tolerance	Rahman et al. (2020)	1	3
5	Financial tranquillity	Vieira et al. (2023)	1	10
6	Risk perception scale	Ahmed et al. (2022)	1	6

Behavioural scientists, such as Zinn (2015), have identified risk-taking as socially construed, and risk-takers recognize their exposure to potential harm. Risk aversion is the opposite of this, denoting the investors’ perceived attitudes and underlying distaste for uncertainties (Dow & da Costa Werlang, 1992). This bipolarity of risk attitude draws on the personal construct theory proposed by G. Kelly (1955), which postulates that all human thought patterns are dichotomous. Further, B. Kelly (1966) suggested a theoretical bipolarity (two-element) for all behavioral constructs. Reddy (1999) and Epting et al. (1993) also identified dyadic elicitation. According to them, all construing is bipolar, with behavioral personal constructs having their respective opposites. Accordingly, a contrasting psychological relationship exists between construct poles, which encompass two (or sometimes more) contrasting concepts, aligning with the semantic contrast in cognitive psychological Research. Risk appetite refers to the level of risk an individual is willing to take on in exchange for potential rewards, which varies in relation to the expected returns. While high RA investors prioritize significant gains and accept higher loss risks, low RA investors seek stability and capital preservation. The IRA scale developed in this study has two factors confirming these propositions. These aspects make the IRA scale superior to other existing scales. Additionally, the IRA scale was correlated with another independent measure, the Financial Risk Propensity scale developed by Manocha et al. (2023). The r-values were 0.825, 0.854, and 0.958, respectively, for RT, RA, and IRA. These values are significant at 0.01, denoting a strong parallel between the measures. This analysis helped examine the Criterion validity for IRA (Nga & Shamuganathan, 2010).

Following the validation and generalization of the scale, an effort was made to assess the extent of investor risk appetite across diverse categories of samples, based on gender, nationality, organizational type, and educational qualifications. t-test and ANOVA were used to determine the differences between the samples. This analysis would help identify potential variations in risk appetite, providing valuable insights for tailored strategies and interventions in investment decision-making and risk management practices. The normality and homogeneity assumptions necessary for t-tests and ANOVA were examined based on Park’s (2015) guidelines. Results indicated that all variables met the normality criteria, with skewness and kurtosis values falling within acceptable ranges. The Levene test showed a non-significant result for all variables, indicating homogeneous variance across groups. These findings align with the assertions of Baarda et al. (2019), suggesting that the dataset is suitable for further statistical analysis. Moreover, these initial assessments enhance the validity and reliability of subsequent inferential procedures. Table 5 shows the mean, SD, and t-values of RT, RA, and IRA variables based on gender, indicating significant differences (0.05 level) for the factors and the overall score. The mean values and SD for males were higher, denoting that males have higher RT, risk aversion, and IRA levels. This analysis and the results also facilitated the examination of measurement invariance. According to Davidov et al. (2014), measurement invariance involves comparing measured traits across different subgroups of survey respondents. It is a statistical property that indicates the questionnaire measures the same construct consistently across various subgroups of respondents (Millsap, 2011; Vandenberg, 2002).

Table 5. Figures and t-values based on Gender.

Variable	Gender	Number	Mean	Std. Deviation	t-Value
Risk tolerance	Male	219	13.5023	3.870	6.803 **
	Female	186	10.7581	4.243
Risk aversion	Male	219	7.6073	3.640	10.488 **
	Female	186	4.5269	1.813
Investment risk appetite	Male	219	21.110	6.895	9.783 **
	Female	186	15.215	4.653

Note: ** significant at 0.01 level.

ANOVA examined the difference based on qualifications, and the results (Table 6) show a significant difference (p < 0.000 level) for the two factors and IRA.

Table 6. Results of ANOVA Based on Educational Qualification.

		Sum of Squares	df	Mean Square	F	Sig.
Risk Appetite	Between Groups	2306.117	2	1153.058	29.579	0.000
	Within Groups	15,203.166	390	38.982
	Total	17,509.282	392
Risk Aversion	Between Groups	783.554	2	391.777	42.466	0.000
	Within Groups	3598.009	390	9.226
	Total	4381.562	392
Risk Tolerance	Between Groups	401.421	2	200.710	11.684	0.000
	Within Groups	6699.699	390	17.179
	Total	7101.120	392

Table 7 presents the ANOVA results based on occupational experience. The table shows significant differences based on the respondents’ occupational experience with the two factors and IRA.

Table 7. Results of ANOVA Based on Occupational Experience.

		Sum of Squares	df	Mean Square	F	Sig.
Risk appetite	Between Groups	736.207	4	184.052	4.319	0.002
	Within Groups	17,045.309	400	42.613
	Total	17,781.516	404
Risk aversion	Between Groups	199.029	4	49.757	4.681	0.001
	Within Groups	4251.948	400	10.630
	Total	4450.978	404
Risk tolerance	Between Groups	226.880	4	56.720	3.184	0.014
	Within Groups	7125.406	400	17.814
	Total	7352.286	404

5. Discussion

It is crucial to recognize that for the development of measures, the theoretical foundations and validation methods are not just procedural formalities. They are fundamental to ensuring the accuracy and reliability of new instruments (Li et al., 2024). The study has followed the development and validation procedures and frameworks proposed by various experts in the field. Some of the experts who provided valuable insights about the rigor in their construction process include T. R. Hinkin et al. (1997), Hair et al. (2019), and Ellis et al. (2008).

Investment risk appetite plays a crucial role in shaping investment decisions, contingent upon whether the investor exhibits a risk-tolerant or risk-averse attitude. Given the inherent significance of risk in decision-making, assessing an investor’s tolerance level becomes a critical precursor to selecting an appropriate investment option (Sayim et al., 2013; Corter & Chen, 2005). Risk appetite significantly impacts investors’ ability to make investment decisions (Keller et al., 2006; Kaufmann et al., 2013). Knowledge of IRA can assist Fintech advisors in providing professional financial advice by assessing the actual level of risk acceptable to each investor (Lippi & Rossi, 2020). It also helps to accurately check the optimum level of risk that investors can assume. There is a dearth of a professionally constructed risk appetite scale, making this study essential. This study constructed and validated an IRA scale through the rigorous application of psychometric techniques. The study has developed and standardized a two-factor, seven-item scale. The identified factors are investment risk tolerance and averting. Risk tolerance is associated with the investor’s risk-taking approach, typically based on the funds available for investment and the individual’s financial situation. According to Zinn (2015), risk-taking is socially construed, and risk-takers recognize their exposure to potential harm. Risk aversion is the polar opposite of the former, denoting investors’ perceived attitudes and underlying distaste for uncertainty (Dow & da Costa Werlang, 1992). Behavioral scientists, such as Zinn (2015), assert that risk-taking is socially constructed, while risk aversion reflects a distaste for uncertainty, based on G. Kelly’s (1955) Personal Construct Theory, which posits that all human thought patterns are dichotomous. This bipolarity of risk attitude, referred to as dyadic elicitation (Reddy, 1999; Epting et al., 1993), is drawn from Personal Construct Theory (G. Kelly, 1955), as supported by Reddy (1999) and Epting et al. (1993). Based on such complex bipolar construction, individuals make inferences to anticipate the outcomes of their thought patterns (J. R. Adams-Webber, 1998; J. Adams-Webber, 2003; Pflueger et al., 2018). The scale confirms these propositions, showing its superiority to existing scales by encompassing high- and low-risk appetite factors.

This study offers both theoretical application and practical implications, contributing to the understanding of financial risk. First, this study introduces a novel approach and tool for quantifying the concept of IRA. Investigating RA can help shed light on investor mood and market price movements. In addition, the results add value to the discussion of RA and contribute to previous material. This tool can serve as a compass for investors’ risk preferences, the risk levels they are willing to accept, and whether their mood is optimistic or pessimistic. A high value implies a willingness to take risks and vice versa. The validated IRA scale can help portfolio managers gauge their clients’ risk tolerance more precisely, providing personalized advice and tailored investment recommendations that align with an individual investor’s risk tolerance and aversion levels. Another implication is that financial institutions can develop investment products that cater to different risk appetite levels. This segmentation can attract more investors by meeting their specific risk preferences. In addition, educators and financial literacy programs can use insights to design courses and materials that help investors understand their risk profiles and the implications for their investment decisions. Although the study was conducted in Saudi Arabia with Saudi samples, it may have global appeal, as its primary focus was on the psychological aspects of decision-making, cognitive biases, and risk-taking propensity. Although this study was conducted in Saudi Arabia, its emphasis on cognitive biases, decision-making, and risk-taking propensity was drawn from psychological insights widely recognized in the Behavioural Finance (BF) literature. These aspects are not exclusive to any specific cultural or geographic context, which lends the findings broader theoretical appeal. However, this needs to be empirically examined with cross-cultural samples from other geographical localities. This aspect opens up new avenues for further research in this fascinating field.

While the IRA scale holds significant academic and practical relevance, the study suffers from a few limitations. One limitation of this study stems from the lack of longitudinal data, which constrained the examination to concurrent validity. A study based on longitudinal data would enable the assessment of predictive validity, offering more profound insights into the scale’s effectiveness over time. Another limitation is the observer effect (Rosenthal, 1976). While this bias may not have impacted this study due to adherence to prescribed values, it remains a factor to acknowledge in future research endeavours. Considering each study’s background conditions, this questionnaire may benefit from further refinement by incorporating variables such as socio-cultural and religio-political upbringing, which could enhance applicability across diverse populations. Furthermore, this study examined the differences in IRA for only a few demographics. Other variables, such as marital status, age of the investor, income levels, number of dependents, and several other factors, could influence the IRA, which could be examined in future studies. Future Research should explore these variations to ensure the scale’s effectiveness and applicability across diverse contexts. Although the study was conducted in Saudi Arabia, its findings have applicability beyond national borders. However, caution should be exercised when generalizing, given the study’s limited sample scope. Future research endeavours could explore the tool’s efficacy in other regions by expanding the sample to include diverse nationalities.

This work assumes that it would be idyllic to integrate rational finance and BF to present an all-inclusive risk assessment framework. Rational finance offers dynamic asset pricing models that show a rigorous framework for incorporating behavioral premises, such as sentiments, bounded rationality, and heuristic-based decisions (Avgouleas, 2012; Nau, 2025). This is possible even while conserving the no-arbitrage principle, which is a cardinal tenet of rational finance. The IRA scale developed in this study is a bridging mechanism, integrating the psychological and behavioral dimensions of risk perceptions. The scale does so while integrating them with the rational asset pricing requirements. An integrative framework that acknowledges investor heterogeneity and biases, while maintaining theoretical consistency and practicality, would provide a comprehensive and balanced understanding of investment risk appetite. The IRA operationalizes behavioral constructs in a manner that does not compromise internal consistency but expands the logic and scope of dynamic asset pricing models (Brav et al., 2002). Hence, BF is positioned so that it does not compete with rational finance but rather serves as an extension that enriches its predictive and diagnostic abilities. By demonstrating how psychological rationality and financial rigour can be successfully integrated, this work makes a meaningful contribution to the ongoing debate, enhancing both theory and practice.

6. Research Implications and Limitations

This study offers both theoretical application and practical implications. This research contributes to the understanding of financial risk in multiple ways. First, this study introduces a novel approach and tool for quantifying the concept of investment risk appetite (IRA). It helps to investigate risk appetite, which may offer new light on investor market mood and market price movements. In addition to the above points, the results of this study add value to the risk appetite discussions and contribute to previous material. This tool can serve as a compass for investors’ risk preferences and the level of risk they are willing to accept. It will also help determine whether the investor mood is optimistic or pessimistic. A high value implies a willingness to take risks and vice versa. As a result, when the market crashes, the risk appetite falls and fluctuates considerably more sharply. This scale would help gauge the overall market sentiment by aggregating individual risk appetites and provide valuable insights into market trends and potential shifts in investor behaviour.

6.1. Study Implications

The validated IRA scale can help portfolio managers gauge their clients’ risk tolerance more precisely and provide personalized advice and tailored investment recommendations that align with an individual investor’s risk tolerance and aversion levels. This customization would enable clients to receive personalized advice tailored to their financial goals and comfort level with risk, fostering robust and more trusting relationships with their clients, ultimately leading to higher client satisfaction and retention. Another implication is that financial institutions can develop investment products that cater to different risk appetite levels. This segmentation can attract more investors by meeting their specific risk preferences. Additionally, educators and financial literacy programs can utilize insights from the IRA scale to develop courses and materials that help investors understand their risk profiles and the implications for their investment decisions.

The IRA scale also facilitates more in-depth Research into the factors influencing risk appetite. Researchers can investigate how various factors, including economic and cultural influences, affect risk tolerance and aversion. In addition, insights from this work can be applied to other related fields, such as behavioural economics and psychology, to understand the interplay between risk attitudes and decision-making. While the scale provides behavioural insights and enables capturing investor risk, its utility as a risk assessment tool requires critical examination. Since risk appetite is unstable over time, it may not resolve issues of arbitrage due to deviating from the Asset Pricing Theorem. The scale would be of high utility by operationalizing behavioral inputs, enhancing explanatory power, and combining it with limits of arbitrage (Shleifer & Vishny, 1997). Although the IRA enhances the BF literature, its utility, theoretical rigour, and practicality depend on psychometric robustness, cross-cultural utility, and its integration with rational asset pricing models.

6.2. Limitations of the Study

Like other studies, the IRA scale developed in this study holds significant academic and practical relevance, offering a valuable contribution to understanding and measuring the concept. However, the study suffers from a few limitations. Although meticulous attention was paid to adhering to relevant guidelines and best practices in this Research, it is essential to acknowledge that every study inevitably faces limitations, and this study is no exception. One limitation of this study stems from the lack of longitudinal data, which constrained the examination to concurrent validity. Longitudinal data would have enabled the assessment of predictive validity, offering more profound insights into the scale’s effectiveness over time. Another limitation involves the use of convenience sampling for data collection. Though this method enabled the collection of adequate responses across multiple contexts, it could restrict generalizability. Future research could employ other methods, such as probability or stratified sampling techniques. This could facilitate representativeness of data and strengthen the external validity. The likelihood of observer effect cannot be entirely ruled out in this work. However, all possible efforts were made to minimize observer effect through the use of standardized procedures, including ensuring respondent anonymity and investigator neutrality. Another potential limitation is the experimenter-observer effect, as described by Rosenthal (1976). While this bias associated with the researcher’s interpretation of results may not have impacted this study due to adherence to prescribed threshold values and rules of thumb, it remains a factor to acknowledge in future research endeavours. Furthermore, the questionnaire used in this study may benefit from further refinement, considering the specific background conditions of each study.

Additionally, refinement of the scale could incorporate variables such as socio-cultural and religio-political upbringing to enhance its applicability across diverse populations. While the scale demonstrated reasonably good reliability and validity across a cross-section of the sample, variations may exist within specific sub-groups or populations. This study examined the differences in IRA based on gender, educational qualification, and occupational experience only. There may be other variables, such as marital status, age of the investor, income levels, number of dependents, and several other factors, that could influence the IRA and be examined in future studies. Therefore, future research should explore these variations to ensure the scale’s effectiveness and applicability across diverse contexts. Although the study was conducted in Saudi Arabia, its findings have applicability beyond national borders. However, caution should be exercised when generalizing, given the study’s limited sample scope. Future research endeavours could explore the tool’s efficacy in other regions by expanding the sample to include diverse nationalities.

This work adopts the behavioral approach to risk assessment. However, it is essential to recognize the innate methodological and theoretical limits. As mentioned earlier, since BF emphasizes psychological biases and heuristics, it implies heightened levels of risk that may not be entirely recognized or mitigated through tools like the IRA scale. This is because behavioral theories deviate from the basic assumptions of Asset Pricing, which emphasizes rational asset pricing, raising the possibilities of arbitrage options. Barberis and Thaler (2002) opined that, though BF provides insights into investor irrationalities, it cannot completely substitute rational asset pricing theory. This issue, which could cast doubt on the internal consistency and empirical applicability of behavioral frameworks (Shleifer & Vishny, 1997), highlights the limitations that future researchers could address. Additionally, this study employed a limited set of factors. Noting this limitation, there exists ample scope for future expansion of the study to include a broader range of behavioral, economic, and contextual variables, which could provide a comprehensive framework for examining investment risk appetite. Advancing the literature will also facilitate moving beyond exploratory factor construction and include sound, theoretically based models that incorporate both behavioral and rational finance perspectives. Future research can substantially contribute to understanding the various aspects of investment risk appetite and refining it.

7. Conclusions

The purpose of the study is to develop and validate the IRA scale. This study represents a significant advancement in behavioural finance and risk management by addressing the gap in measuring risk appetite and developing and validating a comprehensive questionnaire. A thorough literature review revealed a deficiency in suitable tools for assessing this concept, prompting the need for the present Research. The study successfully developed a psychometrically sound scale to measure risk appetite, employing a rigorous methodology and leveraging quantitative methods. This achievement fills a crucial void in the existing literature and lays the groundwork for future investigations. Researchers and financial service providers, such as fintech companies and asset managers, can utilize the scale to gauge investors’ risk appetite. The scale demonstrates strong internal consistency, indicative of its reliability and robustness. As such, it is a valuable tool for further exploration and analysis of risk appetite, providing researchers with a standardized means to assess and measure this important construct. In addition to being the first study of its kind in the literature, this Research stands out for its methodological rigour, rendering it suitable for replication and adaptation in various countries across the region. The scale holds significant utility for policymakers, researchers, and industry professionals, offering a means to assess investors’ risk appetite. Using the questionnaire, stakeholders can pinpoint the risk attitude in today’s dynamic investment scenario. The development and standardization of the questionnaire in this study paves the way for further empirical investigations in investment behaviour. As such, it is anticipated that this Research will inspire new avenues of inquiry and advance knowledge in this exciting field.