RESEARCH DOSSIER
Psychological Research on High-Net-Worth Individuals
Abstract. Psychological research on high-net-worth individuals investigates measured psychological characteristics, subjective experience, and situated financial and social decisions among people selected by an explicit wealth criterion. This critical narrative review examines direct studies of wealthy individuals and selected adjacent evidence on wealth shocks, social class, and offshore conduct. It distinguishes individual from household assets, investable assets from comprehensive net worth, reported preferences from observed choices, and observational associations from causal findings. Evidence on personality, entrepreneurial risk-taking, portfolio decisions, well-being, time use, and giving supports differentiated accounts conditioned by acquisition history, institutional circumstances, and sampling design. Administrative evidence of offshore financial activity establishes economic or legal facts within its scope and cannot by itself identify individual motives. The study develops an analytical synthesis of the evidentiary requirements for rigorous psychological inquiry into affluent populations, including representative recruitment, measurement comparability, longitudinal identification, replication, and confidentiality. No new theory or primary empirical dataset is introduced.
Keywords: High-net-worth individuals; wealth; economic psychology; personality; risk preference; sampling; subjective well-being; prosocial behavior; offshore finance; research ethics
Research object and scientific scope
High-net-worth individuals constitute a distinct population for psychological inquiry because the ownership and control of substantial financial resources change the circumstances in which choices are made. The scientific object is the empirical investigation of people selected by a specified wealth criterion, the psychological characteristics measured in that population, and the adequacy of the methods used to establish conclusions about them. The research concerns the distribution of psychological characteristics within an economically defined population, the relationship between those characteristics and financial life, and the effect of the population definition on the results. It is a study of the evidence concerning people who possess wealth, rather than an attempt to explain every psychological consequence of wealth across society.
Three questions organize the inquiry. What do direct studies establish about affluent respondents? Which features of the observed relationships arise from research design, population selection, and the history of asset accumulation? What additional evidence is required before conclusions about wealthy individuals can illuminate cross-border asset ownership? These questions are inseparable in practice. An ostensibly precise estimate of a personality difference may answer the first question poorly if recruitment has selected a distinctive occupational group. An excellent economic record may establish the presence of a foreign account while leaving the holder's psychological characteristics entirely unmeasured.
The dossier belongs to Consumer and Economic Psychology within Psychology by Bogdanova and to The Global Offshore System research program. Its concern is with evidence about people, their reported experiences, and measured behavior. The economic magnitude of international portfolios, the legal definition of beneficial ownership, and institutional rules of taxation belong to their respective disciplines. Research on those subjects enters this inquiry when it clarifies population definition or the observable basis for psychological analysis. Financial privacy, wealth-related identity, perceived control, social status, and tax behavior are treated here only as domains of measurement and inference. Each deserves its own substantive investigation.
The recurring mistake in public discussion is to move directly from a balance sheet to a personality portrait. Wealth is a property of a specified economic unit at a specified time. A personality score is the result of a measurement procedure administered to a person. A financial decision is a situated action, potentially involving other people and institutions. These objects can be related through evidence; their relationship requires investigation. It is therefore essential to identify the precise level at which each claim is made. Conclusions reported for a sample cannot be silently converted into individual predictions, and a finding about the ownership of assets cannot be silently converted into a claim about psychological motivation.
The research is relevant to global offshore arrangements because substantial private assets can create opportunities for cross-border investment, professional advice, and international ownership structures. These opportunities are conditions of choice. Whether a particular individual uses them, understands them, directs them, or even knows about a particular holding remains an empirical question. This distinction is maintained throughout the dossier.
Definitions and population boundaries
Net worth is the value of the assets attributable to an economic unit after deduction of its liabilities. The unit may be a person, family, or household. A research study that assigns wealth to a person must explain how jointly held assets and obligations are allocated. In an economic system where ownership and control are separated, legal title, beneficial entitlement, decision authority, and immediate access to resources can diverge.
A high-net-worth individual (HNWI) is defined operationally by the threshold and asset concept employed by a study. The million-unit boundary commonly used in the selected literature is a recruitment criterion, not a psychological discontinuity. One study may count investable financial assets, another total individual net worth, and a third assets held in an account at a particular bank. These definitions admit different people. In the German study of Leckelt and colleagues (2019), the financial asset criterion was at least one million euros; the subsequent study of millionaires used individual net wealth. In the US investment study by Bender and colleagues (2022), the investable asset criterion was at least one million US dollars. None is automatically equivalent to a household having one million in net assets, including the primary residence.
A person with a valuable privately held business, high borrowing obligations, and little liquid savings may meet one definition and fail another. Another person may control a substantial portfolio through a family arrangement without holding all its assets in an individual account. A retired household with valuable real estate can rank differently under an investable-asset criterion than under a comprehensive balance-sheet measure. The primary residence is especially consequential because its inclusion can transform the composition of a sample. Study 1 of Smeets and colleagues (2020) excluded the participant’s own home from the wealth measure. That methodological decision should travel with the finding whenever it is compared with a broader net-worth measure.
Absolute wealth criteria also differ from percentile-based criteria. A percentile locates an individual relative to a reference distribution and depends on the economic unit, geography, and date. The same nominal currency amount has different implications for access to goods and services, financial institutions, and economic rank over time and across jurisdictions. Currency conversion translates units of account but does not resolve those differences. Ultra-high-net-worth status must be defined separately wherever the term is used: this review does not adopt a universal cutoff or extrapolate millionaire averages to billionaires.
The acquisition history matters alongside current wealth. Earnings, business ownership, inheritance, gifts, investment returns, marital arrangements, and valuation changes may contribute simultaneously. A questionnaire that forces a respondent into an earned or inherited category can obscure the combination of initial advantage and subsequent activity. Financial independence, occupational authority, and subjective status must also be measured independently. Someone who is economically wealthy can report low security; someone with a high income can have low net assets. Consequently, a study comparing income groups or social classes provides adjacent evidence until the correspondence between its operational definition and the target wealth-defined population has been demonstrated.
Wealth should be recorded continuously where disclosure and measurement permit, with transparent bands used for privacy or response convenience. Analysis should identify whether effects depend on rank, absolute quantity, portfolio composition, or the comparison population. A threshold defines admission to a study; the relation between psychological outcomes and the amount of wealth within that sample remains a separate scientific question.
Evidence selection and methodological approach
This dossier uses a targeted critical narrative review. Searches combined terms for millionaires, high-net-worth individuals, wealthy investors, top wealth, personality, risk preference, happiness, time use, giving, sampling, and offshore tax evasion. The review prioritized primary empirical publications and official methodological documentation, with adjacent experimental and meta-analytic work used to test the limits of interpretation. The retained studies were selected for the relevance and clarity of their population criteria, measured outcomes, and study design. Searches were not a registered systematic review, and the resulting corpus is not represented as an exhaustive census.
The primary evidentiary distinction is between direct observation of a wealth-selected population and indirect evidence drawn from related populations or economic settings. A second distinction concerns the inferential design: cross-sectional comparison, within-population survey, randomized experimental condition, quasi-experimental wealth variation, and administrative record linkage. Sources are examined at the level actually available, ranging from publisher methods and results passages to abstracts and bibliographic records. Findings from restricted sources are limited to what their accessible summaries establish. No unavailable tables, coefficients, or supplementary materials are reconstructed from expectation.
The evidence is heterogeneous in financial definitions, sampling frames, outcome scales, and study purposes. Pooling the reported findings into one numerical estimate would conceal these differences. The review instead asks whether observations are consistent at the level of precisely specified relationships and whether apparent convergence could arise from common sampling methods or shared data. It also identifies the alternative explanations that survive each design. This approach is appropriate where the number of directly relevant studies is limited and their outcomes cannot reasonably be assumed to measure a single latent psychological dimension.
Interpretive claims in the dossier are analytical deductions from the design and scope of the sources. They are separated from original empirical claims. No participants were recruited, original data collected, or statistical model fitted for this investigation. The scientific contribution is the disciplined comparison of research objects and inferential limits. The source ledger identifies examined material, the claim supported, and important limitations. Some records require further full-text or direct-resolution work before a publication decision; access restrictions cannot be treated as completed methodological inspection.
Confidence is assessed in relation to the question being asked. A randomized behavioral experiment conducted with a narrow wealthy sample may offer a strong estimate of a contextual contrast within that sample and poor evidence about all affluent people. A well-weighted observational survey may estimate a population difference more effectively but remain unable to identify the causal direction. Administrative records may establish observed ownership with high precision, while disclosing nothing about the psychological states of the owners. There is no single hierarchy that ranks these studies independently of the claim under examination.
Access, sampling, and representativeness
Ordinary probability samples have difficulty yielding sufficient observations at the extreme upper end of a wealth distribution. If the target population constitutes one percent of the sampling frame, a simple random sample of one thousand people yields an expected ten members before nonresponse. For the top one-tenth of one percent, the corresponding expectation is one. This elementary calculation illustrates why substantial national surveys can still provide inadequate data for an investigation of differences within affluent groups.
Schröder and colleagues (2020) describe a shareholding-based sampling strategy in which corporate ownership records provide a frame for selecting relatively wealthy German residents. The frame is indirect: documented company ownership helps locate people likely to possess high wealth. The major advantage is that inclusion can be designed probabilistically within an identified frame. The major limitation is that the frame has its own economic structure. People with assets outside the observable ownership records may be less likely to be identified. In psychological research, this matters because occupation and organizational involvement may be connected to the traits under investigation.
A different approach is used by the US Survey of Consumer Finances. The Federal Reserve’s (2023) sampling and weighting documentation describes a geographic probability sample supplemented by a list designed to include wealthy families disproportionately, using statistical records derived from tax returns. Its 2022 response rate for the list sample was about 27 percent, with the likely wealthiest strata near 20 percent. The design excludes the Forbes 400. These explicit qualifications show that even an advanced wealth survey has a documented boundary at the top. Moreover, family financial data cannot directly supply individual-level psychological measures without a suitable linked instrument.
Commercially mediated research offers another route. A financial institution may invite customers who meet an asset threshold, making recruitment feasible without a public wealth register. Yet the resulting sample is conditioned on the institution's customer base, the particular products customers use, and willingness to participate. Clients with independent advisers, assets at multiple banks, or restricted disclosure preferences may differ from the observed participants. An institution's ability to identify affluent customers does not by itself establish that the sample represents the general wealthy population.
Lists of publicly recognizable rich people create an even narrower frame. Public visibility depends on how wealth is held, institutional reporting, media attention, and the public activities of the individuals concerned. Treating visibility as a random feature of ownership can introduce serious distortion. A psychological study of conspicuous founders or prominent philanthropists can establish valuable evidence about those particular subjects but should name its population accordingly.
Nonresponse presents a separate problem after frame coverage. A response rate gives the proportion invited who completed the procedure, not the magnitude or direction of resulting bias. Nonresponse bias depends on whether the conditional probability of responding varies with outcomes of interest after the available adjustments. If individuals especially concerned with disclosure decline research participation, the estimated average privacy concern may be too low. If those who particularly enjoy sharing their views accept invitations, some survey measures could be affected in another direction. These are plausible selection mechanisms requiring examination, not observed facts about all nonrespondents.
Oversampling must be accompanied by appropriate weighting for estimates of population means, and the uncertainty calculation must reflect the design. Where frame inclusion probabilities are unknown, the research should distinguish model-assisted adjustment from design-based representation. Report the eligible population, invitation denominator, achieved sample, recruitment channel, completion fraction, and missing-data handling. A large sample with unknown coverage is not necessarily more informative than a smaller sample with an intelligible probability design.
The rarest wealth strata create an additional problem of statistical precision and confidentiality. Sparse cells make estimates unstable and increase disclosure risk. Aggregating adjacent strata may be justified, but it changes the question from inference about the extreme tail to inference about a broader category. The appropriate response is to report uncertainty and redefine the estimand openly rather than suppress the boundary through a general label such as the rich.
Personality, stereotypes, and acquisition histories
Personality research on wealthy populations has addressed dispositional characteristics directly rather than relying exclusively on outsiders' impressions. Leckelt and colleagues (2019) compared a German high-net-worth sample of 130 respondents with population data and an additional panel measuring perceptions. Their study of personality and stereotypes reported that perceptions of the wealthy corresponded to some self-reported differences while tending to overstate them. Its population and recruitment criterion make it a valuable direct investigation, but the sample size calls for careful interpretation of estimates and subgroup comparisons.
The distinction between perceived and measured personality is critical. A stereotype is a representation of a group held by observers. A self-report is a person's description of habitual behavior and experience under a specified instrument. Agreement between these measures in aggregate does not establish accurate judgment of any particular individual. The relevant quantities include group means, distributions, within-group variation, and measurement reliability. The average difference may be substantially smaller than the variation among people meeting the same wealth criterion.
The later German study by Leckelt and colleagues (2022), self-made and inherited millionaires, provides additional population-oriented evidence. The reported wealthy profile includes greater risk willingness and higher levels of emotional stability, openness, extraversion, and conscientiousness. Differences were more pronounced among participants classified as self-made than among inheritors. This supports closer attention to how wealth was acquired rather than treatment of affluent respondents as a uniform psychological population. The finding remains observational.
At least three processes can contribute to such an association. Individuals with a given set of characteristics may select into occupations and opportunities through which substantial wealth is accumulated. The experience of managing resources and exercising organizational authority may, in turn, affect reported behavior and dispositions. Family background, education, health, historical setting, social networks, and institutional access may contribute both to financial outcomes and to measured psychological characteristics. A comparison of wealth groups at one time does not separate these processes. A difference between acquisition categories also inherits the limitations of the categories and the underlying selection mechanisms.
There is a further issue of survival. A sample of people who have accumulated wealth records the histories of those whose assets reached the inclusion threshold. Individuals with similar initial dispositions who incurred large losses or whose firms failed may be absent. The remaining sample can therefore display associations that characterize successful trajectories without identifying a general cause of success. Evidence about entrepreneurs with and without successful outcomes would be needed to evaluate the role of the measured trait across the full range of possible results.
Personality questionnaires require psychometric scrutiny. A response to an item about being organized or decisive may be influenced by the institutional responsibilities of the respondent. If respondents interpret an item differently because of occupation, the observed score difference combines the intended construct with differential item functioning. Tests of comparability should consider item properties across wealth groups, reliability within each group, and the distinction between broad traits and domain-specific behavior. A self-reported trait related to narcissism is not a clinical diagnosis of narcissistic personality disorder. Neither a wealth measure nor a group-level trait difference establishes clinical prevalence.
The robust analytical lesson is that acquisition histories and sample composition belong inside interpretation of personality findings. A stronger future design would measure characteristics before substantial wealth accumulation, link these measurements to subsequent wealth trajectories with appropriate consent, and record both losses and gains. Inherited wealth and unexpected changes in asset values offer potentially informative comparisons, provided the identification assumptions and effects of anticipation are modeled. Until such designs are applied directly to the target population, psychological characteristics observed among millionaires are evidence of association and heterogeneity.
Risk preferences, entrepreneurship, and portfolio behavior
The financial resources of an affluent individual change the set of feasible decisions. This fact makes risk research unusually difficult to interpret because psychological willingness, economic capacity, institutional requirements, and realized behavior can move together or apart. An investor may be willing to incur a loss but lack the liquidity to accept a particular exposure. Another may be cautious in subjective terms yet hold a concentrated business position for reasons of control, taxation, or contractual constraint.
Fossen, König, and Schröder (2024) investigated risk preference and entrepreneurial investment using the German SOEP high-wealth sample integrated with population data. Their estimates associate higher risk willingness with business ownership and the share of assets committed to one's own enterprise, with stronger relationships among high-wealth respondents. The analysis is particularly valuable because it combines a psychological self-rating with information about actual balance-sheet composition. Its cross-sectional basis limits claims about temporal direction. Moreover, the high-wealth sample derives from the shareholding-oriented sampling development examined earlier. It should not be counted as entirely independent confirmation of every personality finding based on related data.
Self-rated risk willingness measures a person's assessment of readiness to take risk. The portfolio share in a private enterprise reflects a financial position that may result from repeated decisions, changes in market value, financing arrangements, legal restrictions, inheritance, and business survival. Their association is substantively informative. It describes the correspondence between a reported preference and an economic outcome. To interpret the outcome as a revealed causal effect of preference, the timing of both variables and the construction of the available choice set would have to be established.
Bender and colleagues (2022) surveyed 2,484 US participants meeting an investable-asset threshold in Millionaires Speak. They report that affluent respondents' broad financial beliefs resemble those of ordinary households in important respects. Differences arise in the weight placed on financial constraints and other factors; advice, prior market experience, and the possibility of severe losses feature in respondents' explanations of investment. The findings illustrate that high wealth can coexist with recognizable financial judgment processes while changing the relative relevance of particular constraints. Survey responses establish what respondents report as important under the questionnaire, not the causal effect that changing an adviser or a belief would have on allocation.
Risk measures should be domain-specific when a question concerns a distinct financial context. General willingness to take risks can coexist with caution about financial disclosure, health, reputation, or family relationships. An entrepreneur may accept operating uncertainty while purchasing insurance against catastrophic loss. Evaluating a cross-border asset structure therefore requires specifying the risks involved, the available alternatives, transaction history, who makes the decision, and the objective probabilities or subjective beliefs relevant to that decision.
Absolute stakes and relative stakes require separate consideration. A monetary loss of a fixed size can represent very different proportions of disposable resources across participants. Conversely, a small proportional loss for a wealthy participant may represent large consequences for other people associated with a business or a charitable institution. Incentivized experiments should justify the scale and salience of their stakes. Hypothetical scenarios should disclose their hypothetical nature and distinguish willingness to endorse a choice from actual commitment of resources.
The comparison between direct investment surveys and the German entrepreneurship evidence yields an important distinction. Wealth can reduce some liquidity constraints while introducing or intensifying exposure to specialized assets, agency relationships, and concentrated ownership. Greater scope for action is compatible with a greater variety of risk management strategies. Consequently, neither conservative diversification nor entrepreneurial concentration can be treated as the universal financial psychology of high-net-worth individuals. Both the asset structure and the process by which it was acquired must enter explanation.
Subjective well-being, time use, and mental health
Well-being comprises outcomes that require different measures. Life satisfaction is an evaluation of one's life; affective experience concerns the frequency or intensity of emotional states; mental health encompasses symptoms, functioning, and clinical conditions that must be assessed using suitable instruments. A measure labeled happiness in one study may operationalize a different outcome from a life satisfaction measure in another. The interpretation of economic associations depends on these distinctions.
Donnelly and colleagues (2018) examined millionaires’ wealth and happiness in two samples comprising more than four thousand millionaires overall. They report modest positive differences within wealth strata at high levels of assets and an association between earned wealth and higher reported happiness compared with inherited wealth at comparable total amounts. The reported thresholds of eight and ten million US dollars in the respective samples belong to those analyses. They do not establish universal monetary thresholds of happiness. The abstract establishes the broad result; detailed sensitivity analyses and measurement properties require further full-text inspection.
The comparison of wealth sources is particularly susceptible to selection. An inheritor may be older or younger than a self-made participant, may have different family circumstances, and may face different responsibilities and expectations. Occupational engagement can contribute both to the acquisition of assets and to life evaluation. Social background can influence expectations, which in turn affect the meaning of a satisfaction rating. Cross-sectional comparisons cannot determine which of these mechanisms accounts for a reported difference.
Smeets and colleagues (2020) investigated time use and happiness in two Dutch studies. Their data associate wealth with differences in the composition of leisure and discretion over work activities. This finding directs attention to the organization of daily life rather than only to a monetary total. It also demonstrates why researchers should record the kinds of work and leisure performed, their voluntary or obligatory character, and the availability of alternatives. Time use may represent both a consequence of wealth and a result of preferences or occupational structures that precede wealth.
The causal evidence from wealth shocks provides a complementary perspective. Lindqvist, Östling, and Cesarini (2020) used randomized lottery prize variation to examine long-run psychological well-being in Swedish lottery participants. Their analysis supports persistent gains in life satisfaction after larger prizes, with smaller changes in happiness and mental health. The design improves identification of the effects of an unexpected increase in financial resources within the lottery population. Its relevance to established high-net-worth individuals is indirect. Inheritance anticipated for decades, the accumulated proceeds of a business, and a lottery windfall differ in the way they occur and the responsibilities accompanying them.
Taken together, the sources show why a general proposition about wealth and happiness would be scientifically inadequate. Studies must specify the outcome dimension, range of wealth, economic definition, and duration of the exposure. The possibility that resources improve subjective evaluations of financial circumstances is consistent with continued variation in emotional experience and mental health. Wealth may facilitate access to services and choice over daily activities; the presence or absence of a disorder must still be assessed using individual evidence.
Prevalence claims about depression, anxiety, addiction, loneliness, or personality disorders among affluent populations require appropriately sampled clinical or validated symptom data. The selected studies cannot establish those prevalence rates. Similarly, anecdotes about distressed billionaires and satisfied retirees illustrate individual variation without estimating its distribution. An ethically sound study records both beneficial experiences and difficulties, avoids language that treats wealth as a diagnosis, and distinguishes functioning from public reputation. The proper unit of psychological care is the individual, irrespective of financial status.
Prosocial decisions, politics, and moral inference
Research on the social behavior of wealthy people requires the distinction between observable action, available resources, interpersonal context, expressed preference, and moral evaluation. A donation is a transfer of resources. A statement about redistribution is a policy preference. A choice in a bargaining experiment is a response to structured incentives. These outcomes illuminate different aspects of social life and need not align within any given individual.
Smeets, Bauer, and Gneezy (2015) studied giving behavior of millionaires recruited through a Dutch bank. In an incentivized design, random allocation to different game environments and recipient categories made it possible to examine the role of bargaining power and recipient circumstances. The reported results show that giving varied with the situation; the participants were particularly generous toward low-income recipients in the unilateral allocation condition. The design supports conclusions about contextual contrasts among the recruited participants. It provides no universal comparison of moral standing, especially given self-selection into participation and the particular meaning of the experimental stakes.
The cost of a contribution differs across resource levels. A comparison based on absolute amounts may describe an economically important transfer while obscuring sacrifice relative to available resources. A comparison based on proportions may overlook fixed costs or differences in the utility of resources. A study asking about intentions may reveal attitudes while overestimating or underestimating actual giving. To understand prosocial behavior, investigators should record the action, the costs borne, the benefit to recipients, whether the action is public, and the constraints under which it occurs.
Adjacent social-class research provides useful counterevidence to simplified narratives. Jung and colleagues (2023) reported that two preregistered field replications of unethical behavior did not recover the originally proposed positive relationship with socioeconomic status. Wu and colleagues (2025), in their meta-analysis of class and prosociality, reported a small positive aggregate association across a broad corpus, r = .065, with a 95 percent confidence interval from .055 to .075. Neither investigation establishes the moral psychology of millionaires specifically. Their methodological relevance is that class-based hypotheses show sensitivity to design, definitions, and outcome choice, and an aggregate correlation is unsuitable as an individual moral classifier.
Political preferences can also be examined empirically among wealthy people. Page, Bartels, and Seawright (2013) conducted a pilot study of wealthy Americans focusing on approximately the upper one percent of wealth holders. The importance of that research for the present dossier is its explicit selection of wealth rather than merely moderately high income. Its pilot scope, national context, and political outcomes must be preserved. Political participation and policy preferences are legitimate subjects of investigation; they do not amount to comprehensive personality assessment.
A distinction between resource-enabled influence and motivation is essential. Economic resources can expand the actions available to a person or institution. The choice of one action from those available may result from several motives, including competing ones, and the motives may vary over time. Observing a donation, a business investment, or political activity establishes something about the action. Establishing why it occurred requires appropriate measurements, comparison conditions, and often longitudinal evidence. The scientific object remains a specified relation between resources, circumstances, and observed behavior.
Offshore research and the limits of psychological attribution
High-net-worth individuals are strategically important to the study of cross-border financial arrangements because substantial assets can make complex structures economically feasible. That relevance creates a risk of conflating financial evidence and psychological evidence. An account or corporate structure can be studied as a financial relationship, a legal arrangement, or a source of taxable income. It becomes evidence about a psychological process only when relevant characteristics, perceptions, decisions, or experiences are investigated.
Alstadsæter, Johannesen, and Zucman (2019), in Tax Evasion and Inequality, used offshore disclosure records linked to Scandinavian administrative wealth information to study the concentration of offshore tax evasion. Their results establish the importance of the upper wealth distribution in the investigated pattern of noncompliance. This is direct evidence about economic conduct and measurement of wealth, not an investigation of personality, fear, trust, or moral beliefs. A documented illicit arrangement is an appropriate subject for legal and economic analysis; the owner's psychological motivation remains open until independently supported.
Several distinctions are required. Cross-border asset ownership is a broad economic category. A legal entity established in another jurisdiction can serve many purposes. Tax compliance and noncompliance are determined by applicable obligations and facts. Observational data from a particular enforcement action or disclosure event are a selected record of activity; the determinants of inclusion differ from those of a probability sample of wealthy people. The absence of a name from such a record cannot establish the absence of offshore assets, and its presence cannot establish a general psychological disposition.
A reliable study of psychological determinants would define the relevant conduct in accordance with the legal context and recruit comparison groups with similar wealth, opportunity, and institutional access. It would measure actual decision authority, whether the participant received specialist advice, the kinds of alternatives available, beliefs about relevant rules, perceived privacy and security, and the sequence of decisions. The investigation would distinguish lawful international ownership, disputed interpretations of obligations, administrative errors, and established violations. Each category has a different evidentiary foundation and must be maintained as such.
Institutional mediation is especially consequential. An asset owner may authorize an adviser to execute a transaction, rely on a trustee's discretion, or possess rights under a structure whose detailed administration is carried out by others. The person who communicates with a researcher may not be the one who selected the relevant structure. Research that attributes the entire sequence to the psychological disposition of the economic beneficiary therefore risks an error in the unit of analysis. Decision-process evidence must identify who knew what, what options were available, and when authorization occurred.
Participants may have incentives to give socially desirable answers about tax and financial privacy; asking sensitive questions can also produce legitimate concern about confidentiality. Indirect survey techniques may reduce the pressure to disclose incriminating or identifying information, but their estimates concern populations or sets of responses rather than a named individual's conduct. Administrative linkage can strengthen factual verification only when legally and ethically permitted. Investigators must prevent financial records from being transformed into speculative psychological dossiers on identified individuals.
The program's central interdisciplinary opportunity is to connect rigorous accounting of asset structures with comparably rigorous measurement of decision processes. It is precisely the gap between the available administrative evidence and the psychological characteristics of individual actors that makes this research agenda necessary. The selected studies do not furnish a population-wide estimate of the psychological antecedents of offshore ownership. Establishing one would require evidence that directly joins the relevant population, measured psychological variables, and specified cross-border decisions.
Measurement, causality, and replication
A high-quality psychological study specifies its estimand before analysis. An investigator comparing wealthy and nonwealthy respondents may seek a population mean difference, a relationship conditional on age and occupation, an association within a narrow asset band, or a change within an individual after a financial event. These are distinct questions. Adjusting for occupation, for example, can remove a confounding difference if occupation precedes the relevant exposure; it can also remove part of an effect if occupational change follows wealth accumulation. The adjustment strategy requires an explicit temporal and causal rationale.
Consider an observed correlation between wealth and willingness to take financial risk. Educational and family circumstances may be upstream of both. Prior risk-taking can affect wealth accumulation, while current financial security can alter willingness to accept risk. Participants admitted by a wealth threshold are selected on an outcome influenced by many factors. If both entrepreneurial opportunities and risk willingness contribute to reaching that threshold, conditioning on inclusion may induce or alter associations between them. The direction and magnitude of the resulting distortion depend on the underlying data-generating process. Such selection cannot be repaired by simply adding more items to a questionnaire.
Longitudinal designs help establish order, distinguish changes within people from differences between people, and record entrants and exits from an affluent population. They do not eliminate every confounder. Wealth shocks can strengthen identification if their assignment is plausibly independent within the defined comparison group, but the type of shock matters. The analysis of randomized lottery prizes in Swedish participants illustrates a distinct causal question from the entrepreneurial trajectory studied in Germany. The two can contribute complementary evidence without their estimands being pooled.
Measurement quality must be separated from identification. A poorly measured psychological outcome can obscure a valid causal relationship. A reliable measurement can be embedded in a design that fails to establish causality. Instruments should be appropriate to the age and language of participants, and comparative studies should address comparability of item interpretation and scales. Where people report asset values approximately or through ranges, uncertainty in economic measurement should be propagated rather than ignored.
Financial measurement poses particular challenges for privately held assets, contingent obligations, foreign-currency valuation, and shared ownership. Record the valuation date, treatment of residence and pension claims, the nature of liabilities, and whether figures are self-reported or verified. Provide clear instructions for participants who cannot determine market value. When an asset threshold is used to assign people to groups, measurement error near the boundary can produce misclassification. Sensitivity analyses across alternative definitions or bands can show whether a reported result depends on the classification rule.
Statistical inference must also accommodate the possibility of numerous outcomes and small groups. An investigation measuring broad traits, additional narrow traits, subjective status, social preferences, and several financial decisions can generate many associations by chance. Pre-registration of primary hypotheses, appropriate intervals, transparent treatment of exploratory findings, and disclosure of all relevant analyses improve interpretability. Practical magnitude should be considered independently of statistical significance. A detectable difference between group averages can offer little precision in classifying a single person.
Replication requires attention to the independence of evidence. Studies sharing the German high-wealth survey may develop different analytical questions but are not necessarily independent population replications. A second paper may refine measurement or test a different outcome within the same underlying sample, which is valuable. The evidence for generalization across countries, sampling frames, or wealth definitions requires actual variation in those dimensions. Future synthetic reviews should code overlapping datasets explicitly and avoid giving repeated analyses of related samples the weight of separate replications.
Researchers should report whether their findings remain after reasonable changes to recruitment criteria, missing-data assumptions, covariate selection, and outlier treatment. When estimates depend heavily on a few extremely wealthy participants, show the sensitivity transparently within confidentiality limits. A prospective study should preserve an analysis plan, codebook, anonymized or securely accessible analysis materials where permissible, and a record of deviations. The purpose is to make the relation between the evidence and the conclusion inspectable.
Ethics, confidentiality, and a research agenda
The ethical obligation of confidentiality is especially demanding in studies of rare economic populations. Combining geography, detailed financial holdings, occupation, age, and family information can identify a participant even when names are removed. The wealthiest strata may have unusually distinctive public profiles. Researchers should use data minimization, secure separation of identifying information, controlled access, retention limits, and risk assessment before releasing tables or examples. Consent materials should explain who has access, what financial records may be linked, and which information may be published.
The terms of participation should preserve voluntariness even when a bank, employer, family office, or professional intermediary assists recruitment. An invitation transmitted by a trusted adviser can be interpreted as an obligation if the boundaries are unclear. A research arrangement should separate recruitment from service provision and ensure that refusing participation does not affect access to services. Sponsorship and potential conflicts of interest must be disclosed. A funder may have commercial reasons to prefer a particular interpretation of risk tolerance, generosity, or adviser influence. The right to publish appropriate results and report limitations should be preserved.
A study that uses financial or clinical information should acquire only what is justified by its question. Researchers should aggregate rare cells or provide controlled access rather than release combinations likely to identify participants. Qualitative interviews may contain biographical details that require careful de-identification; paraphrasing and consent for quotations can reduce some risks. Publicly available financial information does not create unrestricted permission to draw clinical conclusions about named people.
The substantive research agenda follows directly from the gaps identified. First, improve comparability of individual and household wealth measures and develop sampling methods that extend beyond a single banking channel or registered shareholder frame. Second, incorporate longitudinal measurements of personality, decision authority, perceived risk, and well-being alongside financial histories, including losses and exits from the target group. Third, compare domains of financial judgment using designs in which actual choices, perceived constraints, and advice can be distinguished. Fourth, investigate institutional mediation of cross-border financial decisions with rigorously defined legal categories and privacy-preserving linkage.
Age, gender, occupational pathway, country of residence, migration history, and family relationships are potentially relevant dimensions of heterogeneity. Their inclusion should be motivated by the scientific question and adequate sample size rather than an unrestricted search for group differences. Researchers should report the limits of representation and avoid making subgroup claims where small cells lead to unstable estimates or privacy risk. Comparative research across jurisdictions requires equivalent measurements and attention to local financial and social institutions.
Research with wealthy people can be made more informative by integrating psychological data with objective measures of financial choice, but the two forms of evidence have separate limits. Survey responses can identify participants' expressed reasons, diary and experience-sampling methods can illuminate daily life, and securely linked holdings can clarify the actual economic situation. No individual source can replace all the others. Triangulation is useful when studies address the same question with genuinely complementary evidence and their respective uncertainties remain visible.
The author's contribution is to establish the interpretive consequences of comparing these designs. Wealth measurement is a prerequisite for interpreting a wealthy sample, selection into that sample is part of the inferential problem, and psychological conclusions require a measured psychological object. These conclusions are analytical synthesis rather than a proposed theory, novel clinical method, or original statistical result.
Conclusion
Psychological research on high-net-worth individuals has developed through the combination of specialized sampling methods, direct personality studies, financial surveys, well-being research, and behavioral experiments. Direct investigations of millionaire populations provide substantive evidence of variation in measured traits, stated risk preferences, economic behavior, time use, and subjective experience. They also reveal the significance of asset definitions, acquisition histories, and recruitment channels.
The German evidence supports associations between high individual wealth and selected self-reported characteristics, with meaningful differences associated with wealth acquisition history. The financial research connects reported risk willingness with entrepreneurial positions and identifies the relevance of advice and experience in affluent investors' reported decisions. Well-being studies suggest that financial resources, acquisition histories, and organization of daily life warrant distinct analysis; the causal study of lottery wealth clarifies some effects of financial gains within its own population. Contextual experiments and adjacent class research show why social and moral behavior cannot be inferred from economic rank alone.
These observations warrant substantive research into the wealthy as a population. Their generalization must preserve the conditions under which they were obtained. Wealth classification determines who is eligible for investigation, but it does not determine who a participant is psychologically. The central evidence gap concerns trajectories and decisions observed before and after wealth changes, with comparable instruments and data spanning multiple acquisition pathways and economic settings.
For The Global Offshore System, the distinction between accessible financial conduct and measured psychological experience is essential. Administrative data may document financial relationships and, in suitable designs, established noncompliance. Psychological conclusions require evidence concerning the individuals and processes involved. Reliable interdisciplinary work joins these levels through a specified research design instead of using one as a substitute for the other.
The scientific contribution of this dossier is a structured interpretation of the direct HNW literature alongside the adjacent evidence most relevant to the boundaries of inference. Its findings justify additional research without establishing a new psychological theory or asserting original empirical discovery. A stronger empirical literature will emerge through transparent population definitions, better coverage of rare wealth strata, longitudinal and context-sensitive designs, independent replication, and confidentiality practices capable of protecting the people whose behavior is being investigated.
References
- Alstadsæter, A., Johannesen, N., & Zucman, G. (2019). Tax evasion and inequality. American Economic Review, 109(6), 2073–2103. Source
- Bender, S., Choi, J. J., Dyson, D., & Robertson, A. Z. (2022). Millionaires speak: What drives their personal investment decisions? Journal of Financial Economics, 146(1), 305–330. Source
- Board of Governors of the Federal Reserve System. (2023). Changes in U.S. family finances from 2019 to 2022: Evidence from the Survey of Consumer Finances. Source
- Donnelly, G. E., Zheng, T., Haisley, E., & Norton, M. I. (2018). The amount and source of millionaires’ wealth (moderately) predict their happiness. Personality and Social Psychology Bulletin, 44(5), 684–699. Source
- Fossen, F. M., König, J., & Schröder, C. (2024). Risk preference and entrepreneurial investment at the top of the wealth distribution. Empirical Economics, 66, 735–761. Source
- Jung, M. H., Smeets, P., Stoop, J., & Vosgerau, J. (2023). Social status and unethical behavior: Two replications of the field studies in Piff et al. (2012). Journal of Experimental Psychology: General, 152(5), 1368–1378. Source
- Leckelt, M., Richter, D., Schröder, C., Küfner, A. C. P., Grabka, M. M., & Back, M. D. (2019). The rich are different: Unravelling the perceived and self-reported personality profiles of high-net-worth individuals. British Journal of Psychology, 110(4), 769–789. Source
- Leckelt, M., König, J., Richter, D., Back, M. D., & Schröder, C. (2022). The personality traits of self-made and inherited millionaires. Humanities and Social Sciences Communications, 9, Article 94. Source
- Lindqvist, E., Östling, R., & Cesarini, D. (2020). Long-run effects of lottery wealth on psychological well-being. Review of Economic Studies, 87(6), 2703–2726. Source
- Page, B. I., Bartels, L. M., & Seawright, J. (2013). Democracy and the policy preferences of wealthy Americans. Perspectives on Politics, 11(1), 51–73. Source
- Schröder, C., Bartels, C., Grabka, M. M., König, J., Kroh, M., & Siegers, R. (2020). A novel sampling strategy for surveying high net-worth individuals—A pretest application using the Socio-Economic Panel. Review of Income and Wealth, 66(4), 825–849. Source
- Smeets, P., Bauer, R., & Gneezy, U. (2015). Giving behavior of millionaires. Proceedings of the National Academy of Sciences, 112(34), 10641–10644. Source
- Smeets, P., Whillans, A., Bekkers, R., & Norton, M. I. (2020). Time use and happiness of millionaires: Evidence from the Netherlands. Social Psychological and Personality Science, 11(3), 295–307. Source
- Wu, J., Balliet, D., Yuan, M., Li, W., Chen, Y., Jin, S., Luan, S., & Van Lange, P. A. M. (2025). Social class and prosociality: A meta-analytic review. Psychological Bulletin, 151(3), 285–321. Source