Journal of Economic Psychology 31 (2010) 435–443

Contents lists available at ScienceDirect

Journal of Economic Psychology

journal homepage: www.elsevier .com/ locate/ joep

Confidence mediates how investment knowledge influencesinvesting self-efficacy

James Forbes a,*, S. Murat Kara b

a Angelo State University, Department of Psychology, Sociology, and Social Work, ASU Station #10907, San Angelo, TX 76909-0907, USAb Angelo State University, Department of Accounting, Economics and Finance, ASU Station #10908, San Angelo, TX 76909-0908, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 2 May 2008Received in revised form 6 October 2009Accepted 27 January 2010Available online 6 February 2010

JEL classification:D03D12D14

PsycINFO classification:21002223234039003920

Keywords:KnowledgeConfidenceInvesting self-efficacyInvestor education

0167-4870/$ - see front matter � 2010 Elsevier B.Vdoi:10.1016/j.joep.2010.01.012

* Corresponding author. Tel.: +1 325 942 2068.E-mail address: james.forbes@angelo.edu (J. Forb

A comprehensive investment literacy questionnaire surveyed potential sources (viz.,knowledge, confidence) of investing self-efficacy in a large sample of working adults. Asexpected, the effect of investment knowledge on belief in one’s future capability of orches-trating a plan to achieve investment goals was mediated by confidence. Overall, employees’applied investment knowledge accuracy was low: 57%. In general, investment knowledgewas reliably related to confidence. However, confidence and investment knowledge accu-racy were completely independent for 9 of 21 items, implying an inability to inhibit poorinvestment decisions or an inability to exploit investment opportunities. A policy ofrequired investment training could be implemented so as to not impede individuals’ free-dom of choice, which would likely help the truly uninformed to become more informedand ultimately successful investors.

� 2010 Elsevier B.V. All rights reserved.

1. Introduction

Few domains of knowledge have the potential to be so literally enriching as investing. Yet, as is the case for most topicsfalling under the rubric of personal finance, investing knowledge is a cognitive accomplishment for which, like language,most people receive no direct formal instruction. Unlike language, where universal acquisition is the norm, most adults failto acquire competency in investment knowledge (e.g., Benish, 1998; Landstrom, 1995). Previous studies have surveyed thestock market knowledge and stock holding of adults in general (e.g., Bertaut, 1998), revealed gender differences in financialliteracy (e.g., Goldsmith, Goldsmith, & Heaney, 1997; Kirchler & Hubert, 1999), and examined how financial expertise affectsinvesting decisions (e.g., Hershey, Walsh, Read, & Chulef, 1990). For the present study, an original investment literacy ques-tionnaire was developed to evaluate working adults’ applied investment knowledge, as well as their self-reported level of

. All rights reserved.

es).

436 J. Forbes, S.M. Kara / Journal of Economic Psychology 31 (2010) 435–443

confidence about the accuracy of this knowledge. The questionnaire also measured participants’ investing self-efficacy – be-lief in one’s capability in achieving one’s ultimate financial goals.

1.1. Investment literacy surveys

Chen and Volpe (1998) developed a personal finance questionnaire which they sent to 1800 college students at 14 dif-ferent college campuses (51% response rate). The questionnaire surveyed college students’ knowledge about personal finance(24 items) and investing (7 items). Interestingly, the questionnaire solicited participants’ personal finance opinions and deci-sions. One closed-ended 5-point rating scale item measured participants’ opinion about the desirability of ‘‘planning andimplementing a regular investment program.” Thus, Chen and Volpe were able to determine the relationship of personal fi-nance and investment knowledge to personal finance opinions and decisions. Overall, the mean proportion of correct re-sponses to the personal finance questions was low (0.56); the mean proportion of correct responses to the investmentquestions was even lower (0.42). Furthermore, the researchers found that participants’ level of personal finance knowledgereliably influenced their investing opinions and decisions.

But what about the relationship between investment knowledge and investment behavior? Van Rooij, Lusardi, and Ales-sie (2007) developed a financial knowledge questionnaire which measured participants basic investment numeracy (fiveitems; e.g., ‘‘Suppose you had €100 in a savings account and the interest rate was 2% per year. After 5 years, how much do youthink you would have in the account if you left the money to grow?”) as well as their general investment knowledge (11 items;e.g., ‘‘Stocks are normally riskier than bonds. True or false?”). They used the instrument to survey financial knowledge and stockmarket participation in a large sample of adults representative of the Dutch population. Forty percent of the respondentscorrectly answered all five basic investment numeracy questions, whereas merely 5% of the respondents correctly answeredall 11 investment knowledge questions. Overall, respondents correctly answered 79% of basic investment numeracy items,but only 54% of the investment knowledge questions. These findings accord with other investment knowledge surveys com-missioned by the financial services industry (e.g., KPMG, 1995; Vanguard Group/Money Magazine, 1997) that have focusedon issues such as indexing, mutual funds, diversification, asset allocation, and retirement shelter participation. Typically,these surveys targeted working adults and found that participants answered fewer than 60% of the items correctly.

Overall, stock market participation among participants in the van Rooij et al. (2007) survey was low, which accords withreported low levels of direct stock ownership among adults in the United States and Europe (e.g., Guiso, Haliassos, & Jappelli,2002). Van Rooij et al. also found that stock ownership was positively related to investment knowledge. Forty-four percent ofparticipants scoring in the highest quartile of investment knowledge reported stock ownership, whereas merely 7.5% of par-ticipants scoring in the lowest quartile of investment knowledge reported owing stocks. The relationship between invest-ment knowledge and direct stock market participation held after van Rooij et al. controlled for variables such as age,education, gender, income, and wealth. Nonetheless, even after controlling for numerous demographic characteristics,investment knowledge accounted for only 12% of the variability in stock ownership.

One assumption underlying much of the investing literacy literature of which we are aware is that investing knowledge isan independent, virtually unmediated determinant of some objective investing behavior or outcome (e.g., defined contribu-tion plan participation, stock ownership). Moreover, much of the research in this area confounds knowledge with literacy,often using both terms as equivalent synonyms. However, investment knowledge refers to participants’ score on question-naires designed to assess investment terms and concepts. Investment literacy refers to the uses knowledge is put (viz., reg-ularly contributing to one’s defined contribution plan, evaluating intrinsic value, buying and selling stocks).

In numerous content domains (e.g., ecology, humor, logical reasoning, medicine) participants’ knowledge is influenced byconfidence in their performance on a wide variety of tasks (e.g., Ginkel, 2009; Kruger & Dunning, 1999). In these studies, con-fidence is directly measured. Typically, participants answer knowledge questions (e.g., Bornstein, 1999), or predict the like-lihood of future events (e.g., Paese & Sniezek, 1991), then rate the probability (confidence) that their answers or predictionsare accurate. Interest focuses on determining the extent to which people’s subjective judgments of accuracy exceed or fallbelow their observed accuracy.

Within the domain of investing literacy, previous studies have measured confidence indirectly by operationalizing con-fidence as different outcomes across experimental conditions (e.g., Rubaltelli, Rubichi, Savadori, Tedeschi, & Ferretti, 2005),or by inferring confidence from observations of brokerage account activity (e.g., Odean, 1999). Metacognitive skill in accu-rately assessing the level of one’s performance distinguishes the competent from the incompetent (Kruger & Dunning, 1999).The capacity to distinguish accurate from inaccurate investment knowledge may be an essential characteristic of successfulinvestors. Therefore, unlike previous research, the present study directly measured confidence by asking participants to self-report how confident they were that their responses to investment knowledge questionnaire items were accurate.

Also unlike previous surveys of investment knowledge, the ILQ developed for the present study was designed to measureapplied investment knowledge that could be used to improve an individuals’ investment returns. Some of the investmentknowledge items used in previous research appear to instead have surveyed participants’ awareness of macroeconomic is-sues affecting mutual fund returns (e.g., ‘‘If other factors remain the same, US dollar value of a Japan fund will be . . .,” Chen &Volpe, 1998), their regard for the credentials of financial advisors (e.g., ‘‘If a financial planner’s business card says that he orshe is a Registered Investment Advisor, the planner . . .,” (Volpe, Chen, & Pavlicko, 1996)), or whether participants conceiverisk as stock price volatility (e.g., van Rooij et al., 2007). In the present study, items used to survey mutual fund knowledgefocused on the inverse relationship between fund fees and fund returns, how turnover adversely affects returns, and the

J. Forbes, S.M. Kara / Journal of Economic Psychology 31 (2010) 435–443 437

expected difference in returns from index and actively managed funds. Additional items probed participants’ knowledgeabout retirement shelter issues specific to non-profit organizations, from where the study’s employed adults were recruited.One of these items determined whether employees knew whether 403(b) defined contribution contracts are administered byan insurance carrier, and 403(b)(7) defined contribution contracts are administered by a mutual fund custodian. Anotheritem determined whether employees knew why the net return would likely be meaningfully greater for contributions madeto a 403(b)(7) than to a 403(b) defined contribution plan.

1.2. Investing self-efficacy

Bandura (1986) argues that human behavior and motivation are affected by peoples’ self-beliefs about their capabilities.Self-efficacy refers to peoples’ beliefs about their ability to control their own behavior and influence events affecting theirlives (Bandura, 1997). Indeed, Bandura (1997) argues that self-efficacy is often a better predictor of ultimate performancethan measures of current performance levels. A number of domain specific self-efficacy measures have been developedfor use in a wide variety of research contexts (e.g., Gecas, 1989; McAvay, Seeman, & Rodin, 1996). For example, Dietz, Car-rozza, and Ritchey (2003) adapted items from the Pearlin Global Mastery Scale to measure financial efficacy, but found thatthis construct failed to account for gender differences in retirement saving strategies. Previous studies measuring domainspecific self-efficacy have used multiple item scales to measure multifaceted constructs. The present study used a single-item to measure a unidimensional construct. Investing self-efficacy was defined as participants’ level of agreement with asingle statement about their capability of achieving their long-term financial goals. Single-item measures of constructs suchas self-esteem are appropriate for numerous research contexts and have acceptable psychometric properties (Robins, Hen-din, & Trzensiewski, 2001).

The ILQ measured two potential sources of investing self-efficacy: knowledge and confidence. The knowledge surveyedfocused on applied investment issues across a broad range of investment topics (viz., stocks, stock market, mutual funds,retirement shelters, bonds, risk, investment return, valuation), using a large number (21) of items. For each investmentknowledge item, participants were asked to assess how confident they were about the accuracy of their response. Self-effi-cacy beliefs arise from varied sources, among which is performance or mastery experience (e.g., Bandura, 1986). Positivelyassessed performance tends to increase self-efficacy; negatively assessed performance tends to decrease self-efficacy (Bouff-ard-Bouchard, 1990; Pajares, 2003). Consequently, we expected that participants’ knowledge about investing, as well as theconfidence they reported in their investment knowledge predicted their investing self-efficacy. We further expected thatparticipants’ performance on the investment knowledge questionnaire would be mediated by their self-assessed perfor-mance accuracy.

2. Method

2.1. Participants

One hundred and eighty-nine university employees (41% return rate), of the 461 employees to whom the survey wasemailed, completed a 21-item investment literacy questionnaire (ILQ). The sample comprised 80 full time staff, 97 full timefaculty, and 12 administrators. Participants volunteered to complete the ILQ and were not offered any compensation. Partic-ipants ranged in age from 23 to 73 years (M = 45 years, SD = 11 years) and included 90 males and 99 females. Participantsdescribed their ethnic background as Caucasian (86%), Hispanic (8%), African American (1%), Asian (1%), and other (4%).

2.2. Investment literacy questionnaire (ILQ)

The questionnaire included 21 items which surveyed participants’ knowledge about stocks (3 items), mutual funds (4items), retirement shelters (6 items), bonds (3 items), as well as investment risk and return (5 items). All items had facevalidity in that items surveying stock knowledge featured stocks or the stock market as the subject, mutual fund questionsfeatured mutual funds as the subject, retirement shelter questions featured retirement shelters as the subject, and so on.Questionnaire items included 14 closed-ended multiple choice questions along with seven closed-ended true–falsequestions.

One closed-ended rating-scale item measured participants’ investing self-efficacy: ‘‘I am capable of achieving my long-term investment goals” – accompanied by a 5-point rating scale anchored at each end by the words ‘‘strongly disagree(1),” and ‘‘strongly agree (5)”.

Additionally, the ILQ included eight questions concerning employees’ age, sex, occupation, investment ownership, edu-cation level, retirement shelter participation, income, and ethnic background.

2.3. Inter-item difficulty and item genesis

Items selected for inclusion in the ILQ were obtained from a pool of 54 investment knowledge questions that 77 adultswere asked to answer and then rate for difficulty. Some of the items were adapted from a financial knowledge survey

438 J. Forbes, S.M. Kara / Journal of Economic Psychology 31 (2010) 435–443

(Vanguard Group/Money Magazine, 1997). The remaining items were developed by the investigators from materials ar-chived at a variety of personal finance web sites (e.g., Motley Fool, Vanguard), or from the investigators’ investment knowl-edge. Participants in this pilot study were asked to rate the difficulty of each item using a 5-point scale with 1 labeled ‘‘NotDifficult,” 3 labeled ‘‘Moderately Difficult,” and 5 labeled ‘‘Very Difficult.” The mean difficulty rating of the 21 items includedin the ILQ was 3.34 (SD = 0.76; range 2.36–3.97).

2.4. Procedures

The ILQ was administered entirely on-line after obtaining informed consent. Participants were asked to read each itemcarefully, then select the correct answer. For each item, participants were also asked to indicate how confident they werethat they had selected the correct answer by using a 5-point scale. The confidence scale was anchored at each end by thewords ‘‘confident (5),” and ‘‘unconfident (1)”. Intermediate confidence ratings were ‘‘moderately confident (4),” ‘‘neutral(3),” and ‘‘moderately unconfident (2)”.

3. Results

3.1. Investment knowledge

Table 1 shows employees’ mean proportion of correct responses to the ILQ overall, as well as for each category of invest-ment knowledge. Each of these means was compared to the mean proportion of correct responses that would have occurredif participants answered ILQ items by chance alone. Participants’ correctly answered investment knowledge items overall, aswell as items about mutual funds, retirement shelters, valuation, risk and return, and bonds at reliably greater than chancelevels: all single sample ts(191) P 3.33, p 6 0.001. For items comprising two categories of investment knowledge, stock mar-ket and 403(b) contracts, the proportion of correct responses employees gave was reliably below that expected by chance:both ts(191) 6 �3.33, p 6 0.001. Individual item analysis of the ILQ confirmed that employees’ mean proportion of correctresponses to 16 of 21 items was reliably greater than chance: all single sample ts(191) P 5.18, p = 0.0001. Employees’ meanproportion of correct responses to 2 items did not reliably differ from chance. Employees’ mean proportion of correct re-sponses to 3 items surveying knowledge of the stock market (M = 0.16; SD = 0.37), mutual funds (M = 0.39; SD = 0.27) and403(b) contracts (M = 0.18; SD = 0.27) was reliably less than chance: all ts(191) = 3.33, p = 0.001.

3.2. Full model: sources of investing efficacy

Multiple regression was used to determine whether investment knowledge score, investment knowledge confidence, age,sex, occupation, retirement shelter participation, income, and ethnicity predicted investing self-efficacy. All seven variables,entered simultaneously, predicted Investing Self-Efficacy: F(8, 171) = 14.24, p = 0.0001. Together, the seven predictors in thefull model accounted for 38% of the variability in self-reported Investing Self-Efficacy: adjusted R2 = 0.38. Three of the eightvariables, Investment Knowledge Confidence, Occupation, and Income were reliable individual predictors of Investing Self-Efficacy. Table 2 summarizes the regression coefficients from this analysis.

Table 1Mean proportion of correct responses, proportion of correct responses expected by chance, and mean confidence ratings to employee investment literacyquestionnaire items overall, and by categories of investment knowledge, Study 2.

Categories of investment knowledge Proportion of correct responses Proportion correct expected by chance Knowledge confidence

Stock market 0.16 0.25 3.07(0.37) (1.12)

403(b) Contracts 0.18 0.25 1.74(0.27) (0.94)

Mutual funds 0.39 0.25 2.70(0.27) (1.08)

Retirement shelters 0.70 0.44 3.78(0.23) (0.94)

Valuation 0.67 0.25 2.86(0.47) (1.42)

Risk and return 0.79 0.33 3.52(0.20) (0.97)

Bonds 0.52 0.42 3.10(0.28) (1.04)

Overall investment knowledge 0.57 0.33 3.13(0.15) (0.85)

Table 2Predictor variables, beta coefficients, b coefficients, t values, bivariate and partial correlations with global investing self-efficacy from the multiple regressionanalysis of predictors of investing self-efficacy, Study 2.

Predictor variable B b t Correlations

Zero-order Partial

Investment knowledge .39 .05 .69 .29 �.05Investment knowledge confidence .58 .44 5.78* .53 .41Age .004 .04 .63 �.17 .05Sex .17 .08 1.11 .30 .09Occupation status .28 .16 2.45* .21 .19Retirement shelter participation .006 .03 .79 .14 .06Income .22 .26 3.75* .41 .28Ethnicity .02 .12 .16 .03 .01

Adjusted R2 = .38.* p < 0.01.

* p = .0001

Investment Knowledge

Investing Self-Efficacy

Confidence

β = .32* [β = .04]

β = .55* β = .54*

Fig. 1. Summary of confirmatory mediation analysis showing confidence as a mediator of investment knowledge in predicting investing self-efficacy, withstandardized beta coefficients (b).

J. Forbes, S.M. Kara / Journal of Economic Psychology 31 (2010) 435–443 439

Working adults’ investment knowledge and knowledge confidence were reliably related: r(172) = 0.52, p = 0.0001. Cor-relations among the remaining predictors were 6 0.37. Examination of the tolerances for the seven predictor variables, allof which exceeded 1 � R2 (1 � .38), indicated that all of the variables contributed independently to the prediction of invest-ing self-efficacy. Moreover, the partial correlation between investing self-efficacy and investment knowledge confidence,controlling for investment knowledge score was reliably high: r(172) = 0.48, p = 0.0001. In contrast, the partial correlationbetween investing self-efficacy and investment knowledge, controlling for investment knowledge confidence was not reli-able. Consequently, it is reasonable to conclude that investment knowledge confidence and investment knowledge, as wellas the remaining predictor variables separately accounted for the variation in investing self-efficacy.

3.3. Mediated model: confidence mediates knowledge

Although investment knowledge was not a reliable individual predictor of investing self-efficacy in the full model, as ex-pected, knowledge was reliably related to confidence. Thus, the next step in the analysis was to confirm that confidenceacted as a mediator between knowledge and investing self-efficacy. Following Baron and Kenny (1986) approach for medi-ation analysis, we used linear regression to test the mediation model depicted in Fig. 1. First, investment knowledge reliablypredicted their investing self-efficacy: b = .32; F(1, 190) = 20.63, p = .0001. Second, investment knowledge reliably predictedconfidence: b = .55; F(1, 188) = 79.27, p = .0001. Third, multiple regression with investment knowledge and confidence en-tered simultaneously to predict investing self-efficacy was a reliable model: F(2, 188) = 42.05, p = .0001. Controlling forinvestment knowledge, confidence uniquely predicted variance in investing self-efficacy: b = .54; t = 7.37, p = 0.000. Finally,controlling for confidence, the relationship between investment knowledge and investing self-efficacy was no longer reli-able: b = .04, t = .56, p = .58; indicating that confidence mediates the relationship between investment knowledge and invest-ing self-efficacy. The conservative Sobel test confirmed that confidence reliably mediated the effect of investment knowledgeon investing self-efficacy: z = 5.65, p = .0001.

To explore whether alternative mediation models of the variables shown in Fig. 1 were reliable, we conducted a total ofsix additional exploratory analyses, one for each permutation of knowledge, confidence, and investing self-efficacy. No alter-native model resulted in a reliable complete mediation relationship among the variables.

3.4. Investment knowledge and investment knowledge confidence

To further explore the global reliable relationship between confidence and investment knowledge, a 5 (confidencelevel) � 5 (investment knowledge level) contingency table analysis was conducted. Participants were assigned to 1 of 5 over-all investment knowledge confidence percentile intervals and to 1 of 5 investment knowledge percentile intervals: <20th,

440 J. Forbes, S.M. Kara / Journal of Economic Psychology 31 (2010) 435–443

P20th and <40th, P40th and 660th percentile, P60th and <80th percentile, and P80th percentile. Investment knowledgeand confidence were reliably related: Pearson v2(25, N = 649) = 102.04, p = 0.0001, Cramér’s V = 0.18. Working adults withhigh overall investment knowledge were likely to report high knowledge confidence; working adults with low overall invest-ment knowledge were likely to report low knowledge confidence.

3.5. Specific item knowledge and item confidence

How well did participants’ accuracy on specific investment knowledge items accord with their confidence on those sameitems? To answer this question, a 2 (investment knowledge item: correct, incorrect) � 5 (knowledge confidence (unconfi-dent, moderately unconfident, neutral, moderately confident, confident) contingency table analysis was conducted for eachitem. The Bonferroni method was used to maintain Type I error at 0.05 across all 21 analyses. For 11 of the 21 ILQ items,investment knowledge and knowledge confidence were not independent: all Person v2(4, N = 186) = 16.18, p 6 0.05. Forthese items, low confidence was associated with low accuracy and high confidence was associated with high accuracy.But for 9 of 21 ILQ items (43%), investment knowledge accuracy and confidence about the accuracy were independent.Employees’ confidence in the accuracy of their responses to these 9 items was unrelated to whether they answered the itemscorrectly or incorrectly.

4. Discussion

A comprehensive investment literacy questionnaire surveyed applied investing knowledge among a large sample ofworking adults. Unlike previous investment literacy research, the ILQ also measured participants’ confidence in the accuracyof their responses to each item of investing knowledge. Because the overall purpose of the ILQ was to assess investing lit-eracy, or uses to which investment knowledge are put, the ILQ measured participants’ investing self-efficacy. The principalfinding was that the effect of investment knowledge on belief in one’s ultimate capability of achieving long-term investmentgoals was mediated by confidence.

4.1. Investing self-efficacy and confidence

The investing self-efficacy measure used in the present study assessed participants’ belief about their agentive capabilitiesin attaining their long-term investment goals. This subjective belief refers to a future capability of orchestrating a plan toachieve investment goals in an uncertain environment over which investors exert control over their decisions, but not allof the factors influencing those decisions (Bandura, 1986). Thus, investing self-efficacy is not necessarily derived from currentinvestment knowledge such as that measured in the present study. Because investing self-efficacy refers to a future capability,rather than a current ability, participants in the present study would presumably acquire the knowledge necessary to achievetheir long-term financial goals when sufficiently motivated to put forth the effort and perseverance to do so (Pajares, 2003).

Confidence, on the other hand, was measured by asking participants to rate how confident they were in selecting the cor-rect answer to each investment knowledge item. So, confidence came from individuals’ probabilistic assessment regardingthe accuracy of their close-ended responses to individual investment questions. Strength of confidence was derived fromparticipants’ current knowledge of specific items. Thus, confidence and investing self-efficacy are individually face-valid con-structs. Confidence accounted for 17% of the variability in investing self-efficacy, making it a reliable predictor of, rather thana proxy for investing self-efficacy.

The present study showed that one’s self-assessment of knowledge mediates one’s current investment knowledge in pre-dicting investing self-efficacy. At minimum, this finding undermines the assumption underlying much of the investing lit-eracy literature that investing knowledge is an independent, unmediated determinant of investing behavior. In fact, itappears that assessing one’s level of mastery plays a determining role in one’s eventual investing success. Since masteryand vicarious experience are widely considered among the most important sources of self-efficacy (e.g., Bandura, 1986;Schunk, 1985), future research could focus on varying experiences with applied investing principles and skills, or observingothers using skills necessary to achieve investing success, then observing effects on investing self-efficacy.

4.2. Global vs. specific applied investment knowledge and confidence

Overall, employees’ investment knowledge accuracy was low: 57%. Their investment knowledge accuracy varied fromlows of 16% (stock market) and 18% (403(b) contracts) to highs of 79% (risk and return) and 67% (valuation). This level of over-all knowledge accuracy appeared reasonably well aligned with overall investment knowledge confidence (M = 3.13 on a 5-point scale). Indeed, for 12 of 21 ILQ items, knowledge and confidence were reliably related. Working adults with low accuracyon these items also reported low confidence; those who had high accuracy on these same items also reported high confidence.

This global accord between overall investment knowledge and overall confidence was not observed on 9 of 21 (43%) ILQitems. For these nine items, the study revealed a reliable disconnect between investment knowledge and the confidence par-ticipants had about the accuracy of their responses. Indeed, confidence and investment knowledge accuracy were completelyindependent for these nine ILQ items. Participants expressing high confidence in the accuracy of their responses were nomore likely than those expressing low confidence to be correct or incorrect.

J. Forbes, S.M. Kara / Journal of Economic Psychology 31 (2010) 435–443 441

High confidence unrelated to high accuracy of specific investment knowledge items implies an inability to inhibit poorinvestment decisions (e.g., Camerer & Lovallo, 1999); low confidence regardless of accuracy implies an inability to exploitinvestment opportunities (e.g., Moore & Cain, 2006).

Two of the nine investing knowledge items for which working adult participants’ confidence was completely unrelated toaccuracy surveyed a fundamental distinction between actively managed and index funds. Merely 17% of the employeesknew that up to 80% of actively managed mutual funds under perform comparative broad based index funds over the longterm (e.g., Futrelle, 2004; Markiewicz, 2005). Only 55% of participants knew that one of the reasons for actively managedfunds’ underperformance relative to passively managed funds is fund turnover. Two additional items where participants’confidence was independent of accuracy surveyed employees’ knowledge about differences between 403(b) and403(b)(7) defined contribution plans. Seventy-five percent of the employees in our sample did not know that 403(b) con-tracts are administered by insurance companies, whereas 403(b)(7) contracts are administered by mutual fund companies.Furthermore, 90% of the employees in our sample did not know why 403(b) plans typically under perform 403(b)(7) plans. Aprimary reason for the underperformance of 403(b) plans is the higher fees charged by insurance carriers compared to mu-tual fund custodians.

4.3. Applications and public policy implications

Daniel, Hirshleifer, and Teoh (2002) adduce much evidence for what is termed ‘‘general investor credulity,” occasioned bylow investment knowledge, along with the instances of independence between specific item knowledge and confidence(among other sources) found in the present study. According to Daniel et al., investor credulity prevents investors from appro-priately discounting ‘‘for the incentives of interested parties such as firms, brokers, analysts, and other investors to manipulateavailable information” (p. 142), with systematic mispricing in capital markets as its consequence. It should be emphasizedthat the ubiquitous market mispricing is not addressed by the efficient markets hypothesis in finance theory. Moreover, over-generalization of the efficient markets hypothesis in public policy has influenced the adoption of a laissez-faire approach andfavored deregulation in domestic financial markets. The paradigmatic shift away from defined benefit plans (viz. pensions) todefined contribution plans (e.g., 401 K, 403(b)), as well as the persistent political rhetoric advocating privatization of SocialSecurity exemplify this laissez-faire approach. An intended consequence of these policies has been for individual credulousinvestors, with little investment knowledge, and low investing self-efficacy to shoulder greater responsibility for their invest-ment returns. At its best, laissez-faire policy allows market participants to exploit changing conditions and opportunities. Atits worst, laissez-faire facilitates the transfer of wealth from the uninformed to the unscrupulous informed.

Reminiscent of the dogged adherence to the efficient markets hypothesis in some quarters as the only market modelavailable, overconfidence as the default explanation for all manner of investor errors is too frequently promoted. Specula-tively, the observed independence of knowledge and confidence observed in the present study could imply over- as wellas underconfidence. One consequence of underconfidence among market participants may be immobility or investor behav-ior favoring a status quo. Observing the defined contribution plan participation behavior of their colleagues, Benartzi andThaler (2007) reported that ‘‘investors are relatively passive, . . . slow to join advantageous plans, make infrequent changes,and adopt naïve diversification strategies” (p. 102), suggesting status quo behavior by underconfident investors. Ironically,underconfident investors prone to a status quo bias, would also have difficulty exploiting investment opportunities and mar-ket mispricing occasioned by the excessive or ill advised trading by overconfident investors such as those studied by Barberand Odean (1999) (see also Berg & Lein, 2005).

Given that investor sentiment (Baker & Wengler, 2007) along with investor biases have significant adverse effects on indi-viduals, individual firms, and markets, an alternative to strictly laissez-faire public policy may be appropriate. Daniel et al.(2002) propose increased investor education, improved disclosure rules, and more transparent reporting conventions to helpinvestors avoid permanent loss of capital (see also Hilton, 2001). We believe that investor training regarding basic invest-ment knowledge such as that surveyed in the present study (e.g., the inverse relationship between fees and returns, the dif-ference between 403b and 403(b)(7) plans, diversification, risk, moral hazard, asymmetric information, investor biases)should be required for first time investors opening brokerage accounts and beginning participants in defined contributionplans or other retirement shelters. A policy of required investment training could be implemented so as to not impede indi-viduals’ freedom of choice (Thaler & Sunstein, 2003). Such a modest training intervention would not eliminate investinglosses, but may help the truly uninformed to become more informed and ultimately successful investors.

Appendix

Investment literacy questionnaire knowledge items (Correct responses in Bold).

(1) The Roth IRA allows your assets to grow: (i) Tax free; (ii) Tax deferred; (iii) Tax deductible; (iv) While being taxed at thelowest capital gains rate.

(2) The price-to-earnings (P/E) ratio is: (i) A company’s market capitalization divided by the last 12 months revenues; (ii)How much a company made in profits; (iii) A company’s total stock market value, calculated by multiplying the priceof a single share by the total number of shares outstanding; (iv) The stock price divided by the last four quarters’worth of earnings per share.

442 J. Forbes, S.M. Kara / Journal of Economic Psychology 31 (2010) 435–443

(3) If your overall rate of return from investments was 9% last year, and inflation was 3% during the same year, your real(inflation adjusted) return is: (i) 9%; (ii) 6%; (iii) 12%; (iv) �3%.

(4) A person cannot contribute to both a defined contribution plan (e.g., 401(k), 403(b)) and an Individual RetirementAccount (IRA) in the same tax year: (i) True; (ii) False.

(5) Taxes on investment earnings made inside of a 401(k) or 403(b) plan are deferred until retirement: (i) True; (ii) False.(6) Suppose you purchased a 10 year, $1000 bond with a coupon rate of 8% paid semi-annually. If interest rates for com-

parable bonds rose to 10%, and you were to sell your bond, which of the following is true? (i) Your bond would beworth more than $1000; (ii) Your bond would be worth less than $1000; (iii) Your bond would still be worth$1000; (iv) Changes in interest rates are irrelevant to bond prices so you could not predict how much your bond wouldbe worth.

(7) One difference between actively managed mutual funds and passively managed index funds is that: (i) Expenses are typ-ically lower for actively managed funds; (ii) Only actively managed funds are required to provide a prospectus topotential investors; (iii) Actively managed funds typically outperform index funds over the long term; (iv) Activelymanaged funds typically underperform index funds over the long term.

(8) The single greatest factor in growing your long term wealth is: (i) Investing fees; (ii) The initial amount invested; (iii) Therate of return; (iv) Your tax rate.

(9) Investment risk refers to the potential variability (increase as well as decline) of investment returns: (i) True; (ii) False.(10) Over the last 100 years, which of the following has achieved the highest average investment returns? (i) Stocks; (ii) Cor-

porate bonds; (iii) Money market funds; (iv) Long term government bonds (Treasury Bonds).(11) Employees who participate in employer-sponsored 401(k) or 403(b) plans defer part of their salary directly into their

401(k) or 403(b) account, which decreases the amount of money they pay in taxes: (i) True; (ii) False.(12) Bonds can be thought of as long term loans to government agencies and to corporations: (i) True; (ii) False.(13) Which of the following regarding 403(b) and 403(b)(7) defined contribution contracts is false? (i) The same investment

made within a 403(b) or a 403(b)(7) account will result in the same return over a given interval of time; (ii)403(b)(7) contracts are administered by a mutual fund custodian, whereas 403(b) contracts are administered by aninsurance carrier; (iii) Participants are prohibited from purchasing shares of individual stocks within 403(b) and403(b)(7) tax sheltered accounts; (iv) 403(b) and 403(b)(7) carriers both charge a management fee to administer par-ticipants’ investments, but 403(b) carriers also charge an additional mortality and expense fee, typically more than 1%.

(14) If the price of a stock increased from $40 to $50, what is the rate of return? (i) 5%; (ii) 10%; (iii) 20%; (iv) 25%.(15) If a mutual fund charges an expense ratio of 1%: (i) You pay a one-time fee of 1% after holding shares for one year; (ii) The

amount you invest in a fund is reduced by 1% at the time you buy shares; (iii) Your fund’s returns are reduced by 1%each year that you own the fund; (iv) You pay a 1% lead or sales charge to a broker at the time you buy shares.

(16) If you buy a $1000 bond with a coupon rate of 10 percent, you’ll receive payments of $100 per year: (i) True; (ii) False.(17) An index fund is a passively managed mutual fund that: (i) Tries to beat the investment return of a specified market

index; (ii) Seeks to track the investment returns of a specified market index; (iii) Buys only stocks in Standard& Poor’s 500 Stock Index; (iv) Seeks to invest in the best performing sectors of the stock market.

(18) Compounding occurs when an investment generates earnings on reinvested earnings: (i) True; (ii) False.(19) A mutual fund’s annual turnover ratio refers to: (i) The percentage of a fund’s holdings that change every year; (ii) The

number of stocks purchased during the year; (iii) The number of times per year that a dollar of assets is invested; (iv)How many times the manager has been replaced in a given year.

(20) Which market benchmark is the best gauge of the performance of the entire US stock market? (i) S&P 500 Index; (ii) Wil-shire 5000 Total Market Index; (iii) Dow Jones Industrial Average; (iv) NASDAQ Composite Index.

(21) Suppose $7000 was invested yearly for 25 years, in a whole market index fund, earning an annualized return of 10%,inside a defined contribution plan. After 25 years, the net return would: (i) Typically be higher in a 403(b) compared toa 403(b)(7) account because 403(b) accounts are safer and less risky: (ii) Be approximately $100,000 higher in a403(b) compared to a 403(b)(7) account because of lower fees; (iii) Be approximately $100,000 lower in a 403(b)compared to a 403(b)(7) account because of higher fees; (iv) Be the same regardless of whether the money wasinvested in a 403(b) or a 403(b)(7) account.

References

Baker, M., & Wengler, J. (2007). Investor sentiment in the stock market. Journal of Economic Perspectives, 21(2), 129–151.Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.Bandura, A. (1997). Self-efficacy: The exercise of control. New York: WH Freeman/Times Books/Henry Holt and Co.Barber, B. M., & Odean, T. (1999). Trading is hazardous to your wealth: Common stock investment performance of individual investors. Journal of Finance, 55,

773–806.Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic and statistical

considerations. Journal of Personality and Social Psychology, 51, 1173–1182.Benartzi, S., & Thaler, R. H. (2007). Heuristics and biases in retirement savings behavior. Journal of Economic Perspectives, 21(3), 81–104.Benish, R. A. (1998). The ongoing challenge of providing investment education. Compensation and Benefits Management, 14(1), 42.Berg, N., & Lein, D. (2005). Does society benefit from investor overconfidence in the ability of financial market experts? Journal of Economic Behavior and

Organization, 58(1), 95–116.

J. Forbes, S.M. Kara / Journal of Economic Psychology 31 (2010) 435–443 443

Bertaut, C. C. (1998). Stockholding behavior of US households: Evidence from the 1983–1989 survey of consumer finances. Review of Economics andStatistics, 80(2), 263–275.

Bornstein, B. (1999). ‘‘I know I know it, I know I saw it”: The stability of the confidence-accuracy relationship across domains. Journal of ExperimentalPsychology: Applied, 5(1), 76–88.

Bouffard-Bouchard, T. (1990). Influence of self-efficacy on performance in a cognitive task. Journal of Social Psychology, 130(3), 353–363.Camerer, C., & Lovallo, D. (1999). Overconfidence and excess entry: An experimental approach. The American Economic Review, 89(1), 306–318.Chen, H., & Volpe, R. P. (1998). An analysis of personal financial literacy among college students. Financial Services Review, 7(2), 107–128.Daniel, K., Hirshleifer, D., & Teoh, S. H. (2002). Investor psychology in capital markets: Evidence and policy implications. Journal of Monetary Economics,

49(1), 139–209.Dietz, B. E., Carrozza, M., & Ritchey, N. P. (2003). Does financial self-efficacy explain gender differences in retirement saving strategies? Journal of Women and

Aging, 14(4), 83–96.Futrelle, David (2004). The skinny on index funds. Money, 33(11), 82–84.Gecas, V. (1989). The social psychology of self-efficacy. Annual Review of Sociology, 15, 291–316.Ginkel, W. P. (2009). Knowledge about the distribution of information and group decision making: When and why does it work? Organizational Behavior and

Human Decision Processes, 108(2), 218–229.Goldsmith, R. E., Goldsmith, E. B., & Heaney, J. G. (1997). Sex differences in financial knowledge: A replication and extension. Psychological Reports, 81(3, Pt

2), 1169–1170.Guiso, L., Haliassos, M., & Jappelli, T. (2002). Household portfolios. Cambridge: MIT Press.Hershey, D. A., Walsh, D. A., Read, S. J., & Chulef, A. S. (1990). The effects of expertise on financial problem solving: Evidence for goal-directed, problem-

solving scripts. Organizational Behavior and Human Decision Processes, 46(1), 77–101.Hilton, D. (2001). The psychology of financial decision-making: Applications to trading, dealing, and investment analysis. The Journal of Psychology and

Financial Markets, 2(1), 37–53.Kirchler, E., & Hubert, A. C. (1999). Savings and investment decisions within private households: Spouses’ dominance in decisions ion various forms of

investment. Journal of Economic Psychology, 20(5), 499–519.KPMG Peat Marwick LLP (KPMG) (1995). Retirement benefits in the 1990s: 1995 survey data. June. New York, NY.Kruger, J., & Dunning, D. (1999). Unskilled and unaware of it: How difficulties in recognizing one’s own incompetence lead to inflated self-assessments.

Journal of Personality and Social Psychology, 77(6), 1121–1134.Landstrom, H. (1995). A pilot study on the investment decision-making behavior of informal investors in Sweden. Journal of Small Business Management, 33,

67–76.Markiewicz, A. (2005). Index funds outperform actively managed funds in the long run: S&P. Money Management Executive, 13(40), 3.McAvay, G., Seeman, T. E., & Rodin, J. (1996). A longitudinal study of change in domain-specific self-efficacy among older adults. Journals of Gerontology:

Series B: Psychological Sciences and Social Sciences, 51B(5), 243–253.Moore, D. A., & Cain, D. M. (2006). Overconfidence and underconfidence: When and why people underestimate (and overestimate) the competition.

Organizational Behavior and Human Decision Processes, 103(2), 197–213.Odean, T. (1999). Do investors trade too much? American Economic Review, 89, 1279–1298.Paese, P. W., & Sniezek, J. A. (1991). Influences on the appropriateness of confidence in judgment: Practice, effort, information, and decision-making.

Organizational Behavior and Human Decision Processes, 48(1), 100–130.Pajares, F. (2003). Self-efficacy beliefs, motivation, and achievement in writing: A review of the literature. Reading and Writing Quarterly, 19, 139–158.Robins, R. W., Hendin, H. M., & Trzensiewski, K. H. (2001). Measuring global self-esteem: Construct validation of a single-item measure and the Rosenberg

self-esteem scale. Personality and Social Psychology Bulletin, 27(2), 151–161.Rubaltelli, E., Rubichi, S., Savadori, L., Tedeschi, M., & Ferretti, R. (2005). Numerical information format and investment decisions: Implications for the

disposition effect and the status Quo bias. The Journal of Behavioral Finance, 6(1), 19–26.Schunk, D. H. (1985). Self-efficacy and classroom learning. Psychology in the Schools, 22, 208–223.Thaler, R. H., & Sunstein, C. R. (2003). Libertarian paternalism. American Economic Review, 93(2), 175–179.Vanguard Group/Money Magazine (1997). Financial literacy of mutual fund investors. January, New York, NY.Volpe, R. P., Chen, H., & Pavlicko, J. J. (1996). Personal Investment literacy among college students: A survey. Financial Practice and Education, 7(2), 107–128.Van Rooij, M., Lusardi, A., & Alessie, R. (2007). Financial literacy and stock market participation. DNB Working Papers 146. Netherlands Central Bank, Research

Department.